Imaging Pearls ❯ April 2025
|
-- OR -- |
|
3D and Workflow
- ”MCD is a promising, non-invasive test that can augment traditional cancer screening. As the role of MCD in cancer detection evolves, further research is essential to establish how it will be integrated into clinical practice, ensuring informed, shared decision-making with patients.”
Multicancer Detection (MCD)Testing in Gastrointestinal Cancers: An Evolving Tool for Early Diagnosis
Aditya K. Ghosh · Kyle R. Stephens · Prem A. Kandiah · et al.
Current Gastroenterology Reports (2025) 27:19 - ”MCD screening assays are predominately multi-analyte, blood-serum assays that evaluate the epigenetic signature using targeted methylation-based cfDNA assays for cancers of various types. Some companies have called them multicancer early detection (MCED) tests. However, the early detection rates of MCDs depend on several factors, including the test used and the testing frequency. Figure 1 provides an overview of biomarker detection and analysis techniques used in current MCD tests, illustrating the multifaceted approaches that underpin multi-cancer detection. The analytical validation from an early MCD study demonstrates high specificity (99.3%) and accuracy in predicting cancer signal origin through machine learning analysis of over a million methylation sites, supporting the test’s robustness for clinical use .”
Multicancer Detection (MCD)Testing in Gastrointestinal Cancers: An Evolving Tool for Early Diagnosis
Aditya K. Ghosh · Kyle R. Stephens · Prem A. Kandiah · et al.
Current Gastroenterology Reports (2025) 27:19 - In another study, the CancerSEEK (Exact Sciences, Madison, Wisconsin) test was applied to 1005 patients with nonmetastatic clinically detected cancers of the ovary, liver, stomach, pancreas, esophagus, colorectal, lung, or breast. The median sensitivity of CancerSEEK for the eight cancer types was 70%. Notably, for several cancers that do not currently have available screening methodologies, such as cancers of the liver, stomach, pancreas, and esophagus, the test sensitivity ranged from 69 to 98%. The specificity of CancerSEEK was greater than 99%.
Multicancer Detection (MCD)Testing in Gastrointestinal Cancers: An Evolving Tool for Early Diagnosis
Aditya K. Ghosh · Kyle R. Stephens · Prem A. Kandiah · et al.
Current Gastroenterology Reports (2025) 27:19 
Multicancer Detection (MCD)Testing in Gastrointestinal Cancers: An Evolving Tool for Early Diagnosis
Aditya K. Ghosh · Kyle R. Stephens · Prem A. Kandiah · et al.
Current Gastroenterology Reports (2025) 27:19- The diagnostic evaluation for positive MCD signals is not standard at this time but is based on the best medical judgment. For foregut malignancies, the diagnostic evaluation may include EGD, EGD with EUS, MRCP, or mapping gastric biopsies, which may be in addition to imaging such as CT or PET-CT. Neuroendocrine signals may warrant dotatate-PET. For some tests, no signal of origin is provided, and the diagnostic evaluation may be more generic, starting with CT of the chest, abdomen and pelvis with or without PET-CT. For those MCD tests targeting GI malignancies but that provide signal positivity for other cancers, the evaluation may need to extend beyond a GI evaluation if no cancer is identified.
Multicancer Detection (MCD)Testingin Gastrointestinal Cancers: An Evolving Tool for Early Diagnosis
Aditya K. Ghosh · Kyle R. Stephens · Prem A. Kandiah · et al.
Current Gastroenterology Reports (2025) 27:19 - “The cost to evaluate the positive signal can be significant as the evaluation often includes expensive and/or invasive testing such as PET-CT and endoscopy. Moreover, these tests may or may not be covered by patients’ insurance, opening the door to additional financial burdens. Because MCD testing has not yet been approved by the Food and Drug Administration (FDA), many insurance companies do not cover the test nor the subsequent evaluation (all or parts). This can lead to high costs and, combined with the low positive predictive value, may not lead to a definitive diagnosis.”
Multicancer Detection (MCD)Testing in Gastrointestinal Cancers: An Evolving Tool for Early Diagnosis
Aditya K. Ghosh · Kyle R. Stephens · Prem A. Kandiah · et al.
Current Gastroenterology Reports (2025) 27:19 - “Facilitating shared decision-making by having risk-benefit discussions with patients is critical prior to obtaining an MCD test. The potential benefits that MCD can provide in early diagnosis of cancer must be balanced against the inherent risks that accompany for invasive testing to obtain a diagnosis, and the potential that evaluation may not be diagnostic. The risk of a false positive test and the potential for anxiety and diagnostic uncertainty should also be discussed. MCD testing does not replace age-appropriate cancer screenings such as colonoscopy and cancer screening recommended for genetic predispositions of cancer [38]. The role of MCD is evolving, and ongoing multi-center research is necessary to provide more guidance for clinical decision-making in the years to come. ”
Multicancer Detection (MCD)Testing in Gastrointestinal Cancers: An Evolving Tool for Early Diagnosis
Aditya K. Ghosh · Kyle R. Stephens · Prem A. Kandiah · et al.
Current Gastroenterology Reports (2025) 27:19 - Background: Multicancer early detection tests may expand cancer screening. Characterizing diagnostic resolution approaches following positive multicancer early detection tests is critical. Two trials employed distinct resolution approaches: a molecular signal to predict tissue of origin and an imaging-based diagnostic strategy. This modeling study characterizes diagnostic journeys and impact in a hypothetical population of average-risk multicancer early detection–eligible patients
Conclusions: An imaging-based diagnostic strategy is more efficient than a molecular tissue of origin–informed approach across almost all PPV and tissue of origin accuracy combinations. The use of an imaging-based approach for cancer localization can be efficient and low-risk compared with a molecular-informed approach.
Tumor localization strategies of multicancer early detection tests: a quantitative assessment.
Tyson C, Li KH, Cao X, O'Brien JM, Fishman EK, O'Donnell EK, Duran C, Parthasarathy V, Rego SP, Choudhry OA, Beer TM.
JNCI Cancer Spectr. 2025 Mar 3;9(2):pkaf011. - Methods: A mathematical expression for diagnostic burden was derived using positive predictive value (PPV), molecular tissue of origin localization accuracy, and numbers of procedures associated with each diagnostic outcome. Imaging-based and molecular tissue of origin–informed strategies were compared. Excess lifetime cancer risk due to futile radiation exposure was estimated using organ-specific diagnostic imaging radiation doses.
Results: Across all PPVs and localization performances, a molecular tissue of origin strategy resulted in a higher diagnostic burden (mean ¼ 3.6 [0.445] procedures vs mean ¼ 2.6 [0.100] procedures) for the imaging strategy. Estimated diagnostic burden was higher for molecular tissue of origin in 95.5% of all PPV and tissue of origin accuracy combinations; at least 79% PPV and 90% accuracy would be required for a molecular tissue of origin–informed strategy to be less burdensome than imaging. The maximum rate of excess cancer incidence from radiation exposure for multicancer early detection false-positive results (individuals aged 50-84 years) was 64.6 of 100 000 (annual testing, 99% specificity), 48.5 of 100 000 (biennial testing, 98.5% specificity), and 64.6 of 100 000 (biennial testing, 98% specificity).
Tumor localization strategies of multicancer early detection tests: a quantitative assessment.
Tyson C, Li KH, Cao X, O'Brien JM, Fishman EK, O'Donnell EK, Duran C, Parthasarathy V, Rego SP, Choudhry OA, Beer TM.
JNCI Cancer Spectr. 2025 Mar 3;9(2):pkaf011. 
Tumor localization strategies of multicancer early detection tests: a quantitative assessment.
Tyson C, Li KH, Cao X, O'Brien JM, Fishman EK, O'Donnell EK, Duran C, Parthasarathy V, Rego SP, Choudhry OA, Beer TM.
JNCI Cancer Spectr. 2025 Mar 3;9(2):pkaf011.- Radiological imaging is a component of molecular tissue of origin–informed and imaging-based approaches for localization, and although it does introduce the risk of radiation exposure to both approaches, our analyses show that the magnitude of that risk is minimal. A single false-positive event is calculated to represent on average, approximately a 0.2% increase in cancer risk over the patient’s lifetime. The probability of multiple false-positive events, and therefore multiple futile radiological imaging sequences, is calculated to be no higher than 5%. Taken together, the radiation risk presented by the diagnostic workup of a positive multicancer early detection test is minimal, lower than the estimated risk of screening mammography at a population level. Although our excess lifetime risk estimates are derived using data representative of an imaging-based approach, some 90% of patients in a clinical study using molecular tissue of origin–based multicancer early detection testing required imaging procedures, which suggests a similar radiation risk.
Tumor localization strategies of multicancer early detection tests: a quantitative assessment.
Tyson C, Li KH, Cao X, O'Brien JM, Fishman EK, O'Donnell EK, Duran C, Parthasarathy V, Rego SP, Choudhry OA, Beer TM.
JNCI Cancer Spectr. 2025 Mar 3;9(2):pkaf011. - A follow-up analysis of the prospective DETECT-A study suggests that an imaging-driven workup is reliable. Of those patients with a positive multicancer early detection result but a negative imaging evaluation, only 3 of 98 patients were diagnosed with cancer during more than 4 years of follow-up. The annual incidence rate of cancer during follow-up was 1.0% (95% confidence interval ¼0.2% to 2.8%),34which is comparable with the Surveillance, Epidemiology, and End Results annual incidence rate of 1.5% for women aged 65-74 years,35suggesting that imaging localization is an effective strategy to resolve a positive multicancer early detection result. Furthermore, DETECT-A showed that an imaging-driven approach successfully localized all cancers that were diagnosed after a positive multicancer early detection test result across many cancer sites.
Tumor localization strategies of multicancer early detection tests: a quantitative assessment.
Tyson C, Li KH, Cao X, O'Brien JM, Fishman EK, O'Donnell EK, Duran C, Parthasarathy V, Rego SP, Choudhry OA, Beer TM.
JNCI Cancer Spectr. 2025 Mar 3;9(2):pkaf011. - In conjunction with other reports characterizing the broad utility of radiological imaging in the peridiagnostic time even in the absence of multicancer early detection, there is growing evidence that an imaging-based diagnostic strategy for cancer localization can be efficient and low risk compared with a molecular approach of current generation multicancer early detection tests. In the future, as low-dose radiological imaging continues to improve and additional advanced imaging modalities become more available, there may be imaging-based approaches that pose even lower risk than reported in this study.
Tumor localization strategies of multicancer early detection tests: a quantitative assessment.
Tyson C, Li KH, Cao X, O'Brien JM, Fishman EK, O'Donnell EK, Duran C, Parthasarathy V, Rego SP, Choudhry OA, Beer TM.
JNCI Cancer Spectr. 2025 Mar 3;9(2):pkaf011.
- “This reimagination of authentication includes employing biometric data, such as face, fingerprint, eyes, and even voice confirmation, built around the idea of consumer-centralized connectedness. Such methodology strives to unify all relevant characteristics and services through the one consistent factor—the consumer. Gone are the days of hoisting around a bevy of cards that supposedly identify us—potentially to be replaced by our face or alternative biometric features, which would then become the key that grants access to relevant needs in our daily lives. This discussion of proper authentication serves to reinforce that identity matters—especially in the current world. For instance, in health care, there has been a 532% increase in medical identity theft, an 18% average increase in duplicate patient records due to manual entry, and a 279% increase in social engineering attacks.”
Identity Matters: The Potential for a Frictionless and Secure Patient Experience
Caryn Seidman Becker, Elliot K. Fishman, MD, Steven P. Rowe, MD, PhD, Linda C. Chu, MD, Charles K. Crawford
JACR 2025 (in press) - “In many fields, like health care, identity is foundational to connectivity, not only between organizations but also between consumers and products. Oftentimes, one patient is required to manually identify themselves repeatedly for different sectors involved in health (check-in, verification, insurance, payment, etc). Not surprisingly, efficiency is curbed because of those systems’ inability to “talk” to each other, requiring dedicated communication, wasting hours each day, and inevitably introducing possibilities for human error. There are many opportunities to replace inefficient systems of connection, migrating away from the analog versions and toward the safer, more connected digital versions, ultimately benefiting efficiency. Such innovative perspective on identity drives ease and security because everything relates to the consumer as the metaphorical “key” to their own life.”
Identity Matters: The Potential for a Frictionless and Secure Patient Experience
Caryn Seidman Becker, Elliot K. Fishman, MD, Steven P. Rowe, MD, PhD, Linda C. Chu, MD, Charles K. Crawford
JACR 2025 (in press) - “In medicine, security and confidentiality are essential concerns in which biometric capabilities can provide significant advantages. Patients need assurance that their personal information and health details are protected from technological failures and human errors. In radiology specifically, the adoption of biometric authentication has the potential to streamline the entire health care process while ensuring secure connections. For instance, enabling realtime image transfer could resolve longstanding challenges by granting patients greater control while reducing the need for unsecure processing of sensitive information through third-party sources. By limiting the number of unnecessary data transfers, patients will likely feel more confident in the privacy of their care.”
Identity Matters: The Potential for a Frictionless and Secure Patient Experience
Caryn Seidman Becker, Elliot K. Fishman, MD, Steven P. Rowe, MD, PhD, Linda C. Chu, MD, Charles K. Crawford
JACR 2025 (in press) - “By limiting the number of unnecessary data transfers, patients will likely feel more confident in the privacy of their care. Furthermore, trust and security are crucial in academic medicine, in which the willingness of patients to share data often plays a role in research progression. The safety and convenience offered by biometric authentication could increase patient participation in data sharing, ultimately benefiting all areas of research.”
Identity Matters: The Potential for a Frictionless and Secure Patient Experience
Caryn Seidman Becker, Elliot K. Fishman, MD, Steven P. Rowe, MD, PhD, Linda C. Chu, MD, Charles K. Crawford
JACR 2025 (in press) 
- Trust is essential for advertisement because of its fundamental importance to basic communication. The relationship between consumer and advertisement is tainted by unnecessary factors that demoralize and annoy the consumer. For example, the average consumer sees 10,000 to 15,000 advertisements daily—a number that seems impossible until the speed at which consumers can scroll through their smartphone is considered. It is improbable that every advertisement is relevant or timely for the consumer—elevating the level of annoyance and degrading trust, especially when advertisements follow individuals around the Internet. Digital media done right takes these factors of trust and experience into account, reaching the right audience at the right time in the right place and rendering well-done advertising as a service instead of a nuisance.
Matching the Message to the Audience-Understanding What Your Customer Needs to Hear.
Reed Reed, Shenan S, Fishman EK, Chu LC, Rowe SP, Crawford CK.
J Am Coll Radiol. 2025 Feb 7:S1546-1440(25)00106-1. doi: 10.1016/j.jacr.2025.02.004. Epub ahead of print. PMID: 39924135. - ”The first step in achieving a high standard of communication is actively listening to the consumer in a way that translates their motives, desires, and needs. Furthermore, obtaining a full understanding of customers requires asking self-reflective questions designed to gauge reciprocal understanding, including what consumers expect of products and if they understand the message. By developing and solidifying a bidirectional understanding, industries reflect an authentic awareness of who their audience is. Then, they can begin to analyze how best to convey their message to that intended audience.”
Matching the Message to the Audience-Understanding What Your Customer Needs to Hear.
Reed Reed, Shenan S, Fishman EK, Chu LC, Rowe SP, Crawford CK.
J Am Coll Radiol. 2025 Feb 7:S1546-1440(25)00106-1. doi: 10.1016/j.jacr.2025.02.004. Epub ahead of print. - “In conclusion, although digital media has revolutionized the way advertisements reach consumers, its effectiveness is dependent on trust, connectivity, and understanding the diverse needs of the audience. The rapid evolution of digital platforms and the shift toward more personalized, data-driven marketing present both opportunities and challenges. To succeed, industries must not only leverage the power of digital connectivity but also actively connect consumers to content in the right moment.”
Matching the Message to the Audience-Understanding What Your Customer Needs to Hear.
Reed Reed, Shenan S, Fishman EK, Chu LC, Rowe SP, Crawford CK.
J Am Coll Radiol. 2025 Feb 7:S1546-1440(25)00106-1. doi: 10.1016/j.jacr.2025.02.004. Epub ahead of print. - “Advertisers “reaching the right audience at the right time in the right place” is a literal parallel to the promise of precision medicine to “treat the right patient with the right therapy at the right time.” As such, a number of the same principles will apply. Patients, as health care consumers, might approach medicine in the same way they engage with other industries. Consequently, the same erosion of trust observed in advertising is also present in health care. Physicians must return to the foundational practice of understanding each patient and delivering precision care, recognizing that the treatment plan for one patient may differ from another, even with similar conditions.”
Matching the Message to the Audience-Understanding What Your Customer Needs to Hear.
Reed Reed, Shenan S, Fishman EK, Chu LC, Rowe SP, Crawford CK.
J Am Coll Radiol. 2025 Feb 7:S1546-1440(25)00106-1. doi: 10.1016/j.jacr.2025.02.004. Epub ahead of print. - “In radiology, it is essential for physicians to understand the whole patient, not just their imaging results. Such a comprehensive approach provides a more accurate understanding of the patient’s condition and often leads to different procedures or diagnoses for seemingly similar presentations. Even in cases in which the final message is identical, factors such as demographics, lifestyle, or family history can influence how we approach situations for screening procedures like mammograms or CT lung cancer screenings. Our ability to tailor our approach depending on the audience will benefit rapport and increase the likelihood of our message connecting with the patient. Radiology, like all health care specialties, must meet patients where they are, fostering trust and comfort that enhances the overall effectiveness of medical practice.”
Matching the Message to the Audience-Understanding What Your Customer Needs to Hear.
Reed Reed, Shenan S, Fishman EK, Chu LC, Rowe SP, Crawford CK.
J Am Coll Radiol. 2025 Feb 7:S1546-1440(25)00106-1. doi: 10.1016/j.jacr.2025.02.004. Epub ahead of print.
Adrenal
- Purpose: To develop and test a deep learning model to automatically depict adrenal nodules on abdominal CT images and to simulate triaging performance in combination with human interpretation.
Materials and Methods: This retrospective study (January 2000–December 2020) used an internal dataset enriched with adrenal nodules for model training and testing and an external dataset reflecting real-world practice for further simulated testing in combination with human interpretation. The deep learning model had a two-stage architecture, a sequential detection and segmentation model, trained separately for the right and left adrenal glands. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC) for nodule detection and intersection over union for nodule segmentation.
Conclusion: The deep learning model demonstrated high performance and has the potential to improve detection of incidental adrenal nodules.
Two-Stage Deep Learning Model for Adrenal Nodule Detection on CT Images: A Retrospective Study
Chang Ho Ahn • Taewoo Kim • Kyungmin Jo et al.
Radiology 2025; 314(3):e231650 - “Deep learning algorithms have shown promising results in detecting radiologic abnormalities across various imaging modalities. However, deep learning algorithms for the segmentation of intraabdominal organs have demonstrated the lowest accuracy for adrenal glands because of their greater complexity and anatomic variation. Initial attempts at characterizing the adrenal gland were limited in scope. A recent study showed success in using deep learning for adrenal gland segmentation and classification but did not validate it in a real-world clinical setting.”
Two-Stage Deep Learning Model for Adrenal Nodule Detection on CT Images: A Retrospective Study
Chang Ho Ahn • Taewoo Kim • Kyungmin Jo et al.
Radiology 2025; 314(3):e231650 - ■ In this retrospective study of 995 patients, the deep learning model depicted adrenal nodules with areas under the receiver operating characteristic curve of 0.99 and 0.93 for right and left adrenal glands, respectively (internal test set 1, n = 153); model sensitivity was superior to that of historical reports (99% vs 21%–45%; P < .001).
■ When combined with radiologists’ interpretation in additional test sets, the model improved nodule detection; triaging performance (the percentage that could be used to confidently classify by deep learning) ranged from 77% (938 of 1214) to 98% (11 776 of 12 080).
Two-Stage Deep Learning Model for Adrenal Nodule Detection on CT Images: A Retrospective Study
Chang Ho Ahn • Taewoo Kim • Kyungmin Jo et al.
Radiology 2025; 314(3):e231650 - “This study had several limitations. First, because of the retrospective design, the original radiology reports were produced by several radiologists with varying levels of experience, and the CT scans or protocols were also heterogeneous. The higher accuracy of the deep learning algorithm compared with the radiology report does not imply superiority over human interpretation. Second, only contrast-enhanced CT scans were included, whereas noncontrast CT is often preferred for assessing CT attenuation of adrenal nodules. Finally, the clinical usefulness of the automatic detection of adrenal nodules was estimated without consideration of the downside of increased detection rates.”
Two-Stage Deep Learning Model for Adrenal Nodule Detection on CT Images: A Retrospective Study
Chang Ho Ahn • Taewoo Kim • Kyungmin Jo et al.
Radiology 2025; 314(3):e231650 - “In summary, most adrenal nodules are detected incidentally and can be overlooked during imaging interpretation. Our deep learning model showed promising results in a real-world setting. With low and high thresholds, about 90% of patients could be classified with high confidence as having adrenal nodules using the deep learning model. Further studies should focus on developing user-friendly interfaces and integration into clinical workflows.”
Two-Stage Deep Learning Model for Adrenal Nodule Detection on CT Images: A Retrospective Study
Chang Ho Ahn • Taewoo Kim • Kyungmin Jo et al.
Radiology 2025; 314(3):e231650 - “In conclusion, the potential of AI to enhance the detection of adrenal nodules is becoming more transparent, with promising results in automated detection, segmentation, and even classification. However, the clinical use of these systems remains a subject of ongoing research and debate. The study by Ahn and Kim et al offers an exciting step forward, showing that AI can detect nodules with high sensitivity and assist radiologists in improving diagnostic accuracy. Yet, as with any emerging technology, integrating AI into clinical practice will require further validation, particularly in prospective studies and real-world scenarios. Ultimately, the future of AI in adrenal nodule detection may not only lie in better diagnosis but also in more precise, personalized characterization and prediction strategies for patients with adrenal abnormalities.”
Unveiling the Future: A Deep Learning Model for Accurate Detection of Adrenal Nodules
Ashkan A. Malayeri • Baris Turkbey
Radiology 2025; 314(3):e250387 - “Finally, while AI-powered detection algorithms demonstrate ever-increasing accuracy, they must be implemented in diagnostic workflows considering the entire imaging study rather than one single phase. For example, because adrenal CT with washout calculation is the primary technique for differentiating benign versus malignant adrenal nodules, a future vision of AI in this context may entail not only detecting and segmenting nodules, but also analyzing them across multiple phases, extracting the radiomic data, and providing an objective predictability score on the malignant versus benign nature of the nodules.”
Unveiling the Future: A Deep Learning Model for Accurate Detection of Adrenal Nodules
Ashkan A. Malayeri • Baris Turkbey
Radiology 2025; 314(3):e250387
- “Additional evaluation of the adrenal incidentaloma may be accomplished via biopsy in certain specific scenarios. However, adrenal biopsy is not without risk, which includes the potential for bleeding, hemothorax, pneumothorax, and needle tract seeding in the case of ACC. Therefore, if strong suspicion for ACC exists prior to biopsy, the risks of biopsy may not warrant the procedure. Furthermore, in the case of a pheochromocytoma, tumor disruption can lead to potentially dangerous catecholamine spillage. As a result, it is essential to exclude pheochromocytoma from the differential prior to biopsy to avoid intraoperative hypertension.”
Rare Adrenal Tumors and Adrenal Metastasis
Steven Leonard, BSa, Marc C. Smaldone, MD, MSHPb
Urol Clin N Am - (2025) (in press) - “Rare adrenal tumors encompass a wide range of possible neoplasms, from metabolically active adenomas to adrenal metastases. The clinical presentation of these tumors varies, with many asymptomatic tumors discovered incidentally. Advanced imaging techniques and laboratory analysis are the hallmarks of adrenal tumor diagnosis. It is important to undertake a multidisciplinary approach to the management of patients with adrenal neoplasia, especially considering the variety of tumors and associated presentations. Endocrinologists, oncologists, neurosurgeons, radiologists, and urologists must remain up-to-date on medical and surgical diagnosis and management of such lesions in order to elicit superior outcomes.”
Rare Adrenal Tumors and Adrenal Metastasis
Steven Leonard, BSa, Marc C. Smaldone, MD, MSHPb
Urol Clin N Am - (2025) (in press)
- “Ganglioneuromas present with low attenuation on both contrast-enhanced and unenhanced CT. Furthermore, calcifications are a typical finding. Thus, a metabolically inactive lesion with calcifications and less than 40 HU enhancement may indicate the presence of a ganglioneuroma. If suspicion for ganglioneuroma is high, additional diagnostic workup maybe accomplished with a metaiodobenzylguanidine (MIBG) scan. MIBG is a norepinephrine analog tagged to iodine-123 or iodine-131. Its usefulness in diagnosis of ganglioneuromas is based on increased uptake of the MIBG molecule in the ganglioneuroma, which is then associated with increased signal on imaging.”
Rare Adrenal Tumors and Adrenal Metastasis
Steven Leonard, BSa, Marc C. Smaldone, MD, MSHPb
Urol Clin N Am - (2025) (in press)
- “Neuroblastomas are tumors that arise from neural crest stem cells. They are most commonly found in the adrenal medulla, but like ganglioneuromas, these tumors may develop anywhere along the sympathetic chain as well. This tumor is the most common extracranial solid tumor in children and the most common malignancy in infants, although very rarely, cases can occur in adults as well. About one-third of patients present with metastatic disease. Patients who present at the age of 18 months or less are more likely to have disease that spontaneously regresses or achieve remission with surgical treatment alone, whereas patients greater than age 18 are more likely to require medical therapy to assist with the management of their disease.45 Patients may not exhibit symptoms, but when symptoms are present, they may be related to mass effect from the tumor, or be nondescript such as fevers, weight loss, and fatigue if disease is advanced.”
Rare Adrenal Tumors and Adrenal Metastasis
Steven Leonard, BSa, Marc C. Smaldone, MD, MSHPb
Urol Clin N Am - (2025) (in press)
- Adenomas are the most common type of adrenal tumor, representing 90% of all adrenal incidentalomas, with the majority being metabolically inactive. However, 7.1% of benign adenomas can secrete metabolically active substances, including aldosterone (1%) and glucocorticoids (6%), leading to Conn’s syndrome and Cushing’s syndrome, respectively. Exceedingly rare is the ability of these adenomas to secrete sex hormones. The physiologic activity of these hormones and their associated medical syndromes are covered elsewhere in this issue, but the evaluation for secretion of metabolically active substances places tumors with such capability under the purview of this rare tumor article. Tumors that do not initially possess the ability to secrete metabolic substances rarely gain the ability to do so. The rate of transformation to metabolically active adenoma has been documented at 1.7% but is exceedingly rare.
Rare Adrenal Tumors and Adrenal Metastasis
Steven Leonard, BSa, Marc C. Smaldone, MD, MSHPb
Urol Clin N Am - (2025) (in press) - “Classic external symptoms of Cushing syndrome include central obesity, abdominal striae, buffalo hump, and moon facies. Laboratory findings may also include dyslipidemia, hyperglycemia, and hypertension. Laboratory evaluation for the diagnosis of Cushing syndrome involves either measurement of 24-hour urinary-free cortisol level or low-dose dexamethasone suppression testing. Any abnormal result warrants referral to an endocrine specialist for definitive diagnosis. In evaluation of the origin of Cushing syndrome, the lowdose dexamethasone suppression test can be useful.”
Rare Adrenal Tumors and Adrenal Metastasis
Steven Leonard, BSa, Marc C. Smaldone, MD, MSHPb
Urol Clin N Am - (2025) (in press) - “As with adenomas, myelolipomas are typically diagnosed with cross-sectional imaging, and CT scan offers excellent value in the characterization of the lesion. Like adenomas, these lesions are well defined with smooth borders. However, myelolipomas contain elements of myeloid tissue among their fatty components, and on CT scan, these areas display greater attenuation. The myeloid tissue will further enhance with contrast. Some myelolipomas display calcification or areas of hemorrhage. MRI can also be used to further characterize myelolipomas. If fat is visualized on MRI, then myelolipoma diagnosis is highly likely, although definitive diagnosis still requires tissue sampling.”
Rare Adrenal Tumors and Adrenal Metastasis
Steven Leonard, BSa, Marc C. Smaldone, MD, MSHPb
Urol Clin N Am - (2025) (in press) - “Ganglioneuromas are exceedingly rare neoplasms that may arise from the adrenal gland in 40% of cases. However, they may also develop anywhere along the sympathetic chain, including the cervical (5%), thoracic (15%), abdominal (25%), and pelvic (5%) chain. These tumors are most common in children, but when discovered in adults, it is usually incidentally in the fourth or fifth decade of life. These neoplasms are benign and can grow quite large, but most are asymptomatic. 3If symptoms exist, diarrhea or hypertension may be the presenting symptom secondary to vasoactive intestinal peptide or catecholamine secretion in a minority of ganglioneuromas. These neoplasms have also been shown to encase vessels without impingement or invasion of the lumen, and such cases pose a challenge to surgical resection.”
Rare Adrenal Tumors and Adrenal Metastasis
Steven Leonard, BSa, Marc C. Smaldone, MD, MSHPb
Urol Clin N Am - (2025) (in press)
Cardiac
- “Coronary artery plaque can consist of calcified and/or noncalcified material. Computed tomography (CT) can detect and measure coronary artery calcium (CAC) levels in calcified plaque, which increases with age and is more common at all ages in men than women. In a study of 19 725 asymptomatic individuals aged 30 to 45 years without known coronary disease from 3US studies, 21% had any detectable CAC (CAC score >0).3 In the Multi-Ethnic Study of Atherosclerosis (MESA) of 6110 asymptomatic US individuals without known cardiovascular disease aged 45 to 84 years (mean age, 62 years), the presence and amount of CAC was greater in older adults. Amon men and women younger than 55 years, more than half of asymptomatic individuals had no detectable CAC. However, among those 80 years or older, more than 80% had some detectable CAC.”
Cardiac CT Calcium Score
Peter Glynn, MD, MSc; Sadiya S. Khan, MD, MSc; Philip Greenland, MD
JAMA. 2025 Mar 5. doi: 10.1001/jama.2025.0610. Online ahead of print. - The standard method to detect and quantify CAC is gated cardiac CT, which uses the patient’s electrocardiogram(ECG) to time CT imaging acquisition to improve image quality. CAC can also be identified on non–ECG-gated CT chest imaging but may not be as precise as an ECG-gated CT. CAC is quantified by the Agatston score, a measure of total coronary atherosclerotic burden. The Agatston score is the sum of the attenuation (in Hounsfield units) and area of all CAC lesions in all coronary arteries. Scores range from 0 (no calcified plaque) to more than 1000(extensive calcified atherosclerosis), with no specific maximum score.
Cardiac CT Calcium Score
Peter Glynn, MD, MSc; Sadiya S. Khan, MD, MSc; Philip Greenland, MD
JAMA. 2025 Mar 5. doi: 10.1001/jama.2025.0610. Online ahead of print. - Incidentally detected CAC on nongated CT chest scans, which is generally reported qualitatively as absent, mild, moderate, or severe, also predicts coronary heart disease (CHD) risk. In patients referred for lung cancer screening due to tobacco use in the National Lung Screening Trial, CAC was present in more than 70% of individuals on nongated thoracic CT scans. In multivariable scores of 1 to100, 101 to1000,andgreater than1000 Agatston units were associated with hazard ratios of 1.27 (95%CI,0.69-2.53), 3.57 (95%CI, 2.14-7.48), and 6.63 (95%CI, 3.57-14.97), respectively.6 analysis of time to CHD death, compared with a CAC score of 0,CAC.
Cardiac CT Calcium Score
Peter Glynn, MD, MSc; Sadiya S. Khan, MD, MSc; Philip Greenland, MD
JAMA. 2025 Mar 5. doi: 10.1001/jama.2025.0610. Online ahead of print. - The most common clinical scenario in which CAC testing should be considered is in an asymptomatic patient undergoing risk assessment for ASCVD. The 2019 ACC/AHA Primary Prevention Guidelines stated that the main purpose of CAC measurement is to facilitate decision-making regarding statin therapy. Therefore, CAC testing is not recommended or individuals already taking a statin, those with known ASCVD, or for individuals with certain ASCVD risk factors (eg, familial hyperlipidemia, diabetes, current smoking) in whom statin therapy is recommended regardless of CAC score.”
Cardiac CT Calcium Score
Peter Glynn, MD, MSc; Sadiya S. Khan, MD, MSc; Philip Greenland, MD
JAMA. 2025 Mar 5. doi: 10.1001/jama.2025.0610. Online ahead of print. - For asymptomatic individuals, CAC testing is not covered by the Centers for Medicare & Medicaid Services and is typically not covered by private insurance. Out-of-pocket costs range from approximately $50 to $400. CAC scoring can also lead to increased costs from additional cardiac testing (stress testing, invasive angiography) and from evaluation of incidental findings (lung nodules, aortic dilation). However, the discovery of a high CAC score in an asymptomatic patient should not automatically lead to additional cardiac testing, since additional testing should be based on the presence of symptoms and any other high risk factors. Recently, many health care systems have made CAC testing available by self referral. Because of the increasing prevalence of self-referral, it will be important to evaluate characteristics of those who self-refer for CAC scoring, and the risk reclassification, downstream testing, and cost-effectiveness in this population.
Cardiac CT Calcium Score
Peter Glynn, MD, MSc; Sadiya S. Khan, MD, MSc; Philip Greenland, MD
JAMA. 2025 Mar 5. doi: 10.1001/jama.2025.0610. Online ahead of print. - “The CAC score improves ASCVD risk assessment, particularly in individuals at borderline or intermediate ASCVD risk who are not already taking statin medication (Table). A CAC score greater than 0 indicates the presence of calcified coronary atherosclerosis and portends higher ASCVD risk compared with those with a CAC score of 0. Statin therapy is recommended for asymptomatic individuals with a CAC score of 100 or greater Agatston units and may be considered for those with a CAC score of 1 to 99 Agatston units, especially in patients younger than 45 years.”
Cardiac CT Calcium Score
Peter Glynn, MD, MSc; Sadiya S. Khan, MD, MSc; Philip Greenland, MD
JAMA. 2025 Mar 5. doi: 10.1001/jama.2025.0610. Online ahead of print. - “As measuring CAC score has become more widely available, this article focuses on 3 situations where CAC testing may be omitted or deferred until a time when CAC testing can provide clinically useful information. Three clinical scenarios to facilitate the clinician-patient risk discussion are as follows: (1) when CAC testing is too early, (2) when CAC testing is too late, and (3) when CAC testing is repeated too often. The timing of CAC testing sits within the decision point of lipid-lowering therapy use. High-risk young adults may face an elevated lifetime risk of cardiovascular disease despite a CAC level of 0, whereas older adults may not see an expected benefit over a short time horizon or may already be taking lipid-lowering therapy, rendering a CAC score less valuable. Integrating a CAC score into the decision to initiate lipid-lowering therapy requires understanding of a patient’s risk factors, including age, as well as the natural history of atherosclerosis and related events.” .
Coronary Artery Calcium Testing-Too Early, Too Late, Too Often.
Zheutlin AR, Chokshi AK, Wilkins JT, Stone NJ
JAMA Cardiol. 2025 Mar 5. doi: 10.1001/jamacardio.2024.5644. Epub ahead of print. PMID: 40042828. - “Repeat imaging must be paired with consideration of how this would change management. The Society of Cardiovascular Computed Tomography expert consensus recommends that repeat imaging may be beneficial among adults with a CAC score of 0 or CAC score greater than 0 and less than 100 in 5 years or 3 to 5 years, respectively. The 2018AHA-ACC-MS guidelines provide a slightly longer range of 5 to 10 years between repeat imaging as a reasonable interval before measuring another CAC score if the initial CAC score was 0 and the patient has no major risk factors (ie, smoking or incident diabetes).” .
Coronary Artery Calcium Testing-Too Early, Too Late, Too Often.
Zheutlin AR, Chokshi AK, Wilkins JT, Stone NJ
JAMA Cardiol. 2025 Mar 5. doi: 10.1001/jamacardio.2024.5644. Epub ahead of print. PMID: 40042828. - “CAC represents overt atherosclerotic disease and conveys more than the probability of disease unlike a calculated risk score. After statin initiation, repeated CAC imaging is not warranted to guide titration of statin therapy. Clinicians should explain to patients why a CAC score in a statin-treated patient does not predict plaque composition as it does for a statin naive patient with a similar score. A CAC score is an important decision tool, and CAC testing should be performed to help guide patient- clinician decisions. However, without considering when it is too early, too late to be useful, or too often (eg, yearly), an opportunity is missed to use this important tool selectively, focusing on those who would most benefit.” .
Coronary Artery Calcium Testing-Too Early, Too Late, Too Often.
Zheutlin AR, Chokshi AK, Wilkins JT, Stone NJ
JAMA Cardiol. 2025 Mar 5. doi: 10.1001/jamacardio.2024.5644. Epub ahead of print.
Deep Learning
- “Generative artificial intelligence (AI), defined as AI capable of generating new content, has been increasingly used to support academic publishing. This has been driven by the widespread adoption of large language models, which can generate human-like responses to natural language inputs. For example, given sufficient prompting, an AI chatbot that uses natural language processing can generate a scientific manuscript within minutes. However, the quality, integrity, and value of these articles may be called into question given the potential for generative AI to hallucinate, whereby seemingly realistic content is generated that may be inaccurate or misleading. In addition, such generated texts cannot provide human insights, and the tools cannot be accountable for the integrity of the generated content.”
Generative Artificial Intelligence in Surgical Publishing.
Li B, Kayssi A, McLean LJ.
JAMA Surg. 2025 Feb 5. doi: 10.1001/jamasurg.2024.6446. Epub ahead of print. PMID: 39908048. - Generative AI technologies will have evolving implications for surgical publishing. There is substantial potential for these tools to support key aspects of the scientific process. Currently, leading surgical journals have provided authors and reviewers with guidance on AI use in the editorial process, stating the need for clear disclosure of the tools used and the content generated, the importance of respecting journal policies on confidentiality, and confirmation that authors and reviewers take responsibility for AI-generated content. To stay ahead of the curve in the future, it will be important for journals to routinely use validated software that can reliably detect AI use in manuscripts and reviews, which will help corroborate declarations by authors and reviewers. Furthermore, established organizations, such as the International Committee of Medical Journal Editors9 and the Committee on Publication Ethics,10 could provide updated guidance on the role of AI in the conduct and reporting of scientific research. Many major academic institutions also have recommended generative AI programs and user guidelines that authors should be aware of as they develop, execute, and write their work. With appropriate safeguards, there is potential to harness the power of generative AI to support the advancement of surgical research in a transparent, safe, and ethical manner.
Generative Artificial Intelligence in Surgical Publishing.
Li B, Kayssi A, McLean LJ.
JAMA Surg. 2025 Feb 5. doi: 10.1001/jamasurg.2024.6446. Epub ahead of print. PMID: 39908048. - “Recruiting patients for clinical trials is a costly and time-consuming process that often delays studies from being completed and, in turn, slows new therapies from becoming available. Traditional methods for manually screening participants requires substantial time and labor, contributing up to one-third of clinical trial costs. Although structured electronic health record (EHR) queries using diagnostic codes or medications have improved efficiency, researchers still need to manually review the unstructured clinical data such as that in clinical notes.”
Study Finds AI Can Quickly Prescreen Patients for Clinical Trials, Speeding Enrollment.
Hswen Y, Collins N.
JAMA. 2025 Mar 14. doi: 10.1001/jama.2025.2590. Epub ahead of print. - “Human-in-the-loop ensures that—as opposed to a completely autonomous system, in which an AI model, agent, or computer system is able to complete processes start to finish—a human participant or manager steps in and acts as a confirmation check, a gatekeeper, to make sure this is not being completely autonomized. There is a human touch that allows progress to the next step. For the purposes of this study and for this use case, [this means] ensuring that prior to a patient being contacted or [initiating] outreach for clinical trial [participation], ensuring that those inclusion and exclusion criteria are appropriate, are met, and this person is appropriate to reach out to participate.”
Study Finds AI Can Quickly Prescreen Patients for Clinical Trials, Speeding Enrollment.
Hswen Y, Collins N.
JAMA. 2025 Mar 14. doi: 10.1001/jama.2025.2590. Epub ahead of print. PMID: 40085112. - “Large language models (LLMs) parse unstructured data, which enables a shift in how patients are screened and enrolled. We developed an LLM tool, the Retrieval Augmented Generation Enabled Clinical Trial Infrastructure for Inclusion Exclusion Review (RECTIFIER), and demonstrated retrospectively that it can assess patient eligibility for specific inclusion and exclusion criteria with sensitivity, specificity, and accuracy exceeding that of study staff.”
Manual vs AI-Assisted Prescreening for Trial Eligibility Using Large Language Models-A Randomized Clinical Trial.
Unlu O, Varugheese M, Shin J, et al.
JAMA. 2025 Feb 17:e2428047. doi: 10.1001/jama.2024.28047. Epub ahead of print. PMID: 39960745; PMCID: PMC11833652. - “The AI-assisted patient screening using the RECTIFIER tool significantly improved the rate of eligibility determination and enrollment compared with manual screening in a randomized clinical HF trial. These results provide clinical trials with technology that could transform study operations. Limitations include a single center and a randomized clinical HF trial as the use case. Future efforts will require deployment at numerous sites and the expansion of the number of diseases states while remaining within a secure infrastructure to protect patient data.”
Manual vs AI-Assisted Prescreening for Trial Eligibility Using Large Language Models-A Randomized Clinical Trial.
Unlu O, Varugheese M, Shin J, et al.
JAMA. 2025 Feb 17:e2428047. doi: 10.1001/jama.2024.28047. Epub ahead of print.
- “Radiology, as a highly technical and information-rich medical specialty, is well suited for AI development and is readily impacted by advances in interpretive AI methods. Convolutional neural networks (CNNs), a type of deep learning architecture designed to process pixel data, have shown exceptional performance in recognizing and interpreting medical images. Of the 950 AI-enabled medical devices that have been cleared through the U.S. FDA 510(k) premarket notification process as of June 2024, 76% were authorized for uses in radiology, indicating the particular enthusiasm for radiology applications of AI. These radiology AI products have been developed for a variety of tasks using medical images, including detection, localization, classification, segmentation, and prediction, to help radiologists meet the demand for prompt delivery of precise and accurate imaging results.”
Deployment of Artificial Intelligence in Radiology: Strategies for Success.
Jiang S, Bukhari SMA, Krishnan A, et al.
AJR Am J Roentgenol. 2025 Feb;224(2):e2431898. - “After successful testing, the tool may be subjected to FDA clearance and approval (as described later) before it is implemented in clinical settings. After implementation, continuous monitoring is essential to ensure that the tool maintains its performance and adapts to changes in clinical practice, the patient population under evaluation, or other factors. These steps require regular auditing of metrics, collection of user feedback, and iterative improvements driven by real-world data.”
Deployment of Artificial Intelligence in Radiology: Strategies for Success.
Jiang S, Bukhari SMA, Krishnan A, et al.
AJR Am J Roentgenol. 2025 Feb;224(2):e2431898. 
Deployment of Artificial Intelligence in Radiology: Strategies for Success.
Jiang S, Bukhari SMA, Krishnan A, et al.
AJR Am J Roentgenol. 2025 Feb;224(2):e2431898.- “Brady et al. provided five meaningful strategic components in the process of evaluating an AI tool before purchase from the vendor. First, the AI model’s performance is compared with results from the ongoing local radiologist- driven workflow, focusing on true-positive and true-negative predictive values. Second, the enhanced detection rate is calculated as the ratio of AI-identified positive cases to radiologist-identified positive cases in radiology reports. Third, the most impressive cases for showing the AI model’s efficacy to potentially impacted groups are identified. Fourth, AI model pitfalls are categorized into false-positive and false-negative cases, to set realistic expectations for radiologists and reduce bias in the tool’s use. Finally, these steps are considered collectively in making a final decision of whether to buy and deploy the tool.”
Deployment of Artificial Intelligence in Radiology: Strategies for Success.
Jiang S, Bukhari SMA, Krishnan A, et al.
AJR Am J Roentgenol. 2025 Feb;224(2):e2431898. - “Many deep learning AI systems function as black boxes, making it difficult for health care practitioners to understand AI-driven decisions and hindering trust and acceptance. Transparency of AI systems is crucial for addressing such issues . Transparency entails the clear communication of core information about the AI system to its individual users including the availability of information about how the model makes decisions, to ensure such decisions’ explainability and interpretability. Joint efforts of the FDA, Health Canada, and the U.K.’s Medicines and Healthcare products Regulatory Agency (MHRA) identified transparency as a guiding principle in the use of machine learning techniques in health care. Transparency as a guiding principle is particularly relevant during the phase of educating radiologists and other users about the newly deployed AI tool.”
Deployment of Artificial Intelligence in Radiology: Strategies for Success.
Jiang S, Bukhari SMA, Krishnan A, et al.
AJR Am J Roentgenol. 2025 Feb;224(2):e2431898. - “One future direction for clinical AI deployment entails the development of AI models that integrate imaging data with radiomics and clinical data to create comprehensive predictive models. This holistic approach has the potential to improve the performance of AI tools. Additionally, AI could be integrated with Internet of Things–enabled devices, to leverage such devices’ data collection, sensor fusion capabilities, and seamless connectivity. Stronger integration of AI tools into the PACS, including real-time AI-driven annotation of imaging observations, could also substantially improve radiologists’ workflow.”
Deployment of Artificial Intelligence in Radiology: Strategies for Success.
Jiang S, Bukhari SMA, Krishnan A, et al.
AJR Am J Roentgenol. 2025 Feb;224(2):e2431898. - “The successful deployment of AI in radiology has the potential to enhance diagnostic performance, improve workflow efficiency, and support clinical decision-making. The full realization of AI benefit and practical value requires regulatory adherence, rigorous performance validation, seamless workflow integration, end user education, and a culture of ongoing evaluation and adaptation. As AI continues to evolve, its role in radiology will likely expand, promising further advances in diagnostic performance and operational efficiency. Radiology must embrace the presented strategies of effective AI deployment, to allow AI to serve as radiologists’ partner in achieving high-quality patient care.”
Deployment of Artificial Intelligence in Radiology: Strategies for Success.
Jiang S, Bukhari SMA, Krishnan A, et al.
AJR Am J Roentgenol. 2025 Feb;224(2):e2431898.
- “Radiomics is the high-throughput extraction of quantitative features from radiological images through data characterization algorithms. These features contain information representing pathophysiological mechanisms imperceptible to the human eye. Radiomic features have been linked to metabolic dysfunction– associated steatotic liver disease, cirrhosis, and tumor biological processes like hypoxia, angiogenesis, and inflammation. Quantitative assessment of these features can be used to predict the behavior of cancer, postoperative course, and surgical morbidity.”
Radiomics for Treatment Planning in Liver Cancers.
Mian A, Kamnitsas K, Gordon-Weeks A.
JAMA Surg. 2025 Feb 26. doi: 10.1001/jamasurg.2024.4346. Epub ahead of print. PMID: 40009391. - “Radiomics enables the noninvasive profiling of tumors through routine radiological images. By characterizing the tumor biological phenotype through imaging data, radiomics can provide personalized pretreatment prognostication. It can eliminate invasive biopsies and outperforms somatic sequence variation testing for selection of biological therapies—a particularly important consideration for colorectal liver metastases (CRLMs), where biopsy significantly risks tumor cell seeding, and for cholangiocarcinoma, which is technically challenging to sample. Implementation does not require specialist diagnostic or biochemical expertise and is, in principal, cost-effective compared with competitor prognostics, including circulating tumor DNA or cancer cells.”
Radiomics for Treatment Planning in Liver Cancers.
Mian A, Kamnitsas K, Gordon-Weeks A.
JAMA Surg. 2025 Feb 26. doi: 10.1001/jamasurg.2024.4346. Epub ahead of print. PMID: 40009391. - “Similarly, in HCC, computed tomography (CT) radiomic texture analysis, maybe used to facilitate therapeutic decision-making alongside the Barcelona Clinic Liver Cancer (BCLC) criteria when considering surgical resection, transplant, or other liver-directed therapies. In CRLM, radiomics may also assist in determining optimal surgical resection margins by predicting KRAS sequence variations or identifying replacement-type growth, both of which require wider resection margins. Finally, radiomics approaches could aid in forecasting the volume and health of the future liver remnant post–major hepatectomy.”
Radiomics for Treatment Planning in Liver Cancers.
Mian A, Kamnitsas K, Gordon-Weeks A.
JAMA Surg. 2025 Feb 26. doi: 10.1001/jamasurg.2024.4346. Epub ahead of print. PMID: 40009391. - “Radiomics has demonstrated multiple uses in liver cancers. One large, prospective study demonstrated that a radiomics signature derived from CRLM CT scans outperformed the Response Evaluation Criteria in Solid Tumors (RECIST) criteria for predicting folinic acid, fluorouracil, and irinotecan (FOLFIRI) and bevacizumab response. These findings were supported by another study, where radiomics features outperformed RECIST and KRAS sequence variation status, with the radiomics signature predicting response to FOLFIRI and cetuximab with an area under the curve (AUC) of 0.80 compared with only 0.67 for KRAS status and 0.75 for 8-week tumor shrinkage measured by RECIST. This information can guide early chemotherapy regimen switching in nonresponders and identify patients for resection based on favorable biology. Futurework should explore if radiomics provides relevant insights into the biology of KRAS mutant vs wild-type CRLM.”
Radiomics for Treatment Planning in Liver Cancers.
Mian A, Kamnitsas K, Gordon-Weeks A.
JAMA Surg. 2025 Feb 26. doi: 10.1001/jamasurg.2024.4346. Epub ahead of print. PMID: 40009391. - “In HCC, radiomics and clinical features predicted transarterial chemoembolization (TACE) response with an AUC of 0.90, which was linked to improved survival in the external validation cohort (hazard ratio, 2.43; P < .05). This study, which focused on patients with BCLC stage B disease, raises the question of whether a selection of patients may benefit from liver transplant outside the Milan criteria vs palliative TACE alone. For cholangiocarcinoma, a combination of radiomics and clinical features improved recurrence-free survival prediction (C index: 0.75) vs clinical features alone (C index: 0.69) in external validation. Given high rates of postsurgical recurrence in this patient cohort, further improvements will identify a subset of high-risk patients who should be offered neoadjuvant systemic therapy as a test of biological behavior rather than up-front resection.”
Radiomics for Treatment Planning in Liver Cancers.
Mian A, Kamnitsas K, Gordon-Weeks A.
JAMA Surg. 2025 Feb 26. doi: 10.1001/jamasurg.2024.4346. Epub ahead of print. PMID: 40009391. - “A major challenge in radiomics is limited generalizability of models due to reliance on single-center data, which fails to account for variability in patient characteristics (eg, ethnicity, age, tumor biology), imaging protocols, and the scanners differences. Some radiomic features are sensitive to these variations, reducing reproducibility across institutions. Solutions include using robust features, multiinstitutional data, standardized imaging protocols, data augmentation (eg, generative adversarial networks), and harmonization techniques likeComBat. The Image Biomarker Standardization Initiative has played a key role in standardizing radiomic biomarker nomenclature and definitions. It has also created benchmark datasets and benchmarking values to verify image processing and biomarker calculations, while also developing robust reporting guidelines high-throughput image analysis. This is now helping drive the reliability and reproducibility of radiomic analyses across diverse datasets to develop generalizable radiomic models.”
Radiomics for Treatment Planning in Liver Cancers.
Mian A, Kamnitsas K, Gordon-Weeks A.
JAMA Surg. 2025 Feb 26. doi: 10.1001/jamasurg.2024.4346. Epub ahead of print. PMID: 40009391.
- “Advancements in Artificial Intelligence (AI) are driving a paradigm shift in the field of medical diagnostics, integrating new developments into various aspects of the clinical workflow. Neuroendocrine neoplasms are a diverse and heterogeneous group of tumors that pose significant diagnostic and management challenges due to their variable clinical presentations and biological behavior. Innovative approaches are essential to overcome these challenges and improve the current standard of care. AI-driven applications, particularly in imaging workflows, hold promise for enhancing tumor detection, classification, and grading by leveraging advanced radiomics and deep learning techniques.”
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28: - “AI has proven capabilities in enhancing the clinical management of NENs from multiple aspects, including analysis of serum biomarkers, pathological specimens, and genomics signatures. This review focuses on the impact of AI on diagnostic imaging of NENs, including both anatomical and functional imaging. The review also highlights emerging AI methodologies, their potential integration into clinical workflows, and the opportunities and challenges associated with leveraging AI to improve outcomes in NET management.”
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28: - “Although different approaches exist for AI models to find informative patterns, radiomics models and deep learning architectures are among the most widely used techniques for medical image analysis within the scope of computer vision. Radiomics refers to the high-throughput extraction of quantitative imaging features that characterize tissue phenotypes, such as texture, shape, and intensity, often revealing patterns beyond human perception. Deep learning employs complex neural network architectures, resembling biological neurons and synapses, to identify connections and patterns and extract meaningful features and associations directly from imaging data.”
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28: - “Alternatively, in scenarios where a tumor has already been identified, AI can be used to detect imaging patterns associ ated with different tumor types and help the diagnostician differentiate between NENs and other potential clinical entities. In this context, there is awareness of the tumor and go to AI to assist in selecting the most likely option among the differential diagnoses. However, a significant barrier to clinical implementation remains the labor-intensive process of manual segmentation required to define regions of interest (ROIs) for feature characterization. Advances in deep learning-based automated segmentation have partially addressed this challenge, offering a more efficient and scalable clinical integration.”
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28: - “In the context of a known NEN, AI models are commonly evaluated in clinical settings to characterize the tumor and predict clinicopathological characteristics that are relevant for determining prognosis. Perhaps due to the high relevance of tumor grade in NENs, along with the high morbidity associated with pancreatic surgery, the most common tumor characterization task among published studies is the prediction of tumor grade in PNETs. Despite the high clinical relevance of differentiating NETs from NECs, with previous exploratory analyses suggesting that imaging features could potentially distinguish between these tumor types, 46 very few studies have explored this task. Most effort has instead concentrated on differentiating between grades of NETs, likely due to limitations in sample size and the underrepresentation of NEC in institutional databases.”
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28: - Despite encouraging results, many barriers remain before AI models can be widely applied to support the clinical management of NENs.While most challenges are common across AI applications in medicine, the relative rarity of NENs and high heterogeneity in tumor biology and disease presentation amplify these difficulties. A major challenge with AI lies in the generalization of algorithms beyond he training cohort, often resulting in significant performance reductions for underrepresented groups. Most current NEN studies have sample sizes too small to develop robust, generalizable models, emphasizing the urgent need for collaborative efforts to create publicly available annotated imaging datasets in NENs. These datasets would enable large-scale training of radiomics and deep learning models, which are critical for improving AI performance and generalizability. Additionally, studies must prioritize demonstrating the clinical value of AI-prediction through standardized prospective studies compared with current diagnostic gold standards, a key step toward ensuring clinical relevance.
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28: - AI holds immense promise in revolutionizing the diagnosis, characterization, and management of neuroendocrine neoplasms. By integrating advanced imaging techniques, predictive analytics, and workflow optimization, AI enhances clinical decision-making and paves the way for personalized medicine. These tools can enhance classification accuracy, support the diagnosticians by reducing diagnostic uncertainty, and provide more reliable assessment to guide treatment planning. Despite challenges in standardization, validation, and limited training data due to the rarity of the disease, ongoing advancements in AI methodologies are set to transform management of neuroendocrine neoplasms, improving patient outcomes through more precise and efficient care.
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28: 
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28:
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28:
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28:- Cinematic rendering is post-processing technique based on a complex global lighting model that results in highly photorealistic anatomic images. The global lighting model results in high degrees of surface detail and shadowing effects in the final three-dimensional display. Cinematic rendering is regarded as a marked advance in the evaluation of abdominal cancers and more specifically in pancreatic cancers. In patients amenable to surgery, preoperative planning of PDAC requires accurate visualization of vascular structures in relation to the tumour and also identification of anatomical variants, such as presence of arcuate ligament or celiomesenteric trunk, that may make surgery difficult or increase the risk of intraoperative vascular complications. In this regard, cinematic rendering helps surgeon identify arterial anatomic variants and vascular involvement with high degrees of confidence.13 Regarding PanNET and cystic pancreatic tumours, cinematic rendering allows for more comprehensive visualization and description, and interpretation of anatomical structures along with a global assessment of the burden of metastatic spread.
CT Imaging of the Pancreas: A Review of Current Developments and Applications.
Barat M, Greffier J, Si-Mohamed S, Dohan A, Pellat A, Frandon J, Calame P, Soyer P.
Can Assoc Radiol J. 2025 Feb 22:8465371251319965. doi: 10.1177/08465371251319965. Epub ahead of print. PMID: 39985297. - Despite advances in MRI, CT remains a major diagnostic tool for pancreatic imaging. The benefit of spectral PCCT for image quality is already suggested in many applications, particularly those requiring high spatial resolution and contrast. Spectral PCCT expands the capabilities of spectral CT for several applications such as diagnosis, characterization, and staging of diseases and the anticipated development of colour K-edge contrast agents should open new research areas. Radiomics shows encouraging results in terms of tumour resectability prediction. AI is a field of major ongoing research. One goal for AI could be the detection of small PDAC at an early stage, as a substantial proportion of these cancers remain not visible to the human eyes. Future studies should also include sophisticated models that associate imaging data with clinical and biological data. The recent introduction of new therapies should also stimulate further studies for a better evaluation of tumour response.
CT Imaging of the Pancreas: A Review of Current Developments and Applications.
Barat M, Greffier J, Si-Mohamed S, Dohan A, Pellat A, Frandon J, Calame P, Soyer P.
Can Assoc Radiol J. 2025 Feb 22:8465371251319965. doi: 10.1177/08465371251319965. - “One difficulty in pancreatic cancer is the detection of tumours that are isoattenuating to the adjacent pancreatic parenchyma or tumours that are <20 mm. Chen et al have developed a deep learning-based tool that achieved 89.7% (600/669) sensitivity, 92.8% specificity (746/804), and an AUC of 0.95, with 74.7% sensitivity for cancers <20 mm in a test set of 1473 CT examinations.82 Korfiatis et al have developed a three-dimensional convolutional neural network classification system trained on a large data set that achieved 84% accuracy and an AUC of 0.91 for the diagnosis of visually occult PDAC on prediagnostic CT examinations.”
CT Imaging of the Pancreas: A Review of Current Developments and Applications.
Barat M, Greffier J, Si-Mohamed S, Dohan A, Pellat A, Frandon J, Calame P, Soyer P.
Can Assoc Radiol J. 2025 Feb 22:8465371251319965. doi: 10.1177/08465371251319965.
- Purpose: To develop and test a deep learning model to automatically depict adrenal nodules on abdominal CT images and to simulate triaging performance in combination with human interpretation.
Materials and Methods: This retrospective study (January 2000–December 2020) used an internal dataset enriched with adrenal nodules for model training and testing and an external dataset reflecting real-world practice for further simulated testing in combination with human interpretation. The deep learning model had a two-stage architecture, a sequential detection and segmentation model, trained separately for the right and left adrenal glands. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC) for nodule detection and intersection over union for nodule segmentation.
Conclusion: The deep learning model demonstrated high performance and has the potential to improve detection of incidental adrenal nodules.
Two-Stage Deep Learning Model for Adrenal Nodule Detection on CT Images: A Retrospective Study
Chang Ho Ahn • Taewoo Kim • Kyungmin Jo et al.
Radiology 2025; 314(3):e231650 - “Deep learning algorithms have shown promising results in detecting radiologic abnormalities across various imaging modalities. However, deep learning algorithms for the segmentation of intraabdominal organs have demonstrated the lowest accuracy for adrenal glands because of their greater complexity and anatomic variation. Initial attempts at characterizing the adrenal gland were limited in scope. A recent study showed success in using deep learning for adrenal gland segmentation and classification but did not validate it in a real-world clinical setting.”
Two-Stage Deep Learning Model for Adrenal Nodule Detection on CT Images: A Retrospective Study
Chang Ho Ahn • Taewoo Kim • Kyungmin Jo et al.
Radiology 2025; 314(3):e231650 - ■ In this retrospective study of 995 patients, the deep learning model depicted adrenal nodules with areas under the receiver operating characteristic curve of 0.99 and 0.93 for right and left adrenal glands, respectively (internal test set 1, n = 153); model sensitivity was superior to that of historical reports (99% vs 21%–45%; P < .001).
■ When combined with radiologists’ interpretation in additional test sets, the model improved nodule detection; triaging performance (the percentage that could be used to confidently classify by deep learning) ranged from 77% (938 of 1214) to 98% (11 776 of 12 080).
Two-Stage Deep Learning Model for Adrenal Nodule Detection on CT Images: A Retrospective Study
Chang Ho Ahn • Taewoo Kim • Kyungmin Jo et al.
Radiology 2025; 314(3):e231650 - “This study had several limitations. First, because of the retrospective design, the original radiology reports were produced by several radiologists with varying levels of experience, and the CT scans or protocols were also heterogeneous. The higher accuracy of the deep learning algorithm compared with the radiology report does not imply superiority over human interpretation. Second, only contrast-enhanced CT scans were included, whereas noncontrast CT is often preferred for assessing CT attenuation of adrenal nodules. Finally, the clinical usefulness of the automatic detection of adrenal nodules was estimated without consideration of the downside of increased detection rates.”
Two-Stage Deep Learning Model for Adrenal Nodule Detection on CT Images: A Retrospective Study
Chang Ho Ahn • Taewoo Kim • Kyungmin Jo et al.
Radiology 2025; 314(3):e231650 - “In summary, most adrenal nodules are detected incidentally and can be overlooked during imaging interpretation. Our deep learning model showed promising results in a real-world setting. With low and high thresholds, about 90% of patients could be classified with high confidence as having adrenal nodules using the deep learning model. Further studies should focus on developing user-friendly interfaces and integration into clinical workflows.”
Two-Stage Deep Learning Model for Adrenal Nodule Detection on CT Images: A Retrospective Study
Chang Ho Ahn • Taewoo Kim • Kyungmin Jo et al.
Radiology 2025; 314(3):e231650 - “In conclusion, the potential of AI to enhance the detection of adrenal nodules is becoming more transparent, with promising results in automated detection, segmentation, and even classification. However, the clinical use of these systems remains a subject of ongoing research and debate. The study by Ahn and Kim et al offers an exciting step forward, showing that AI can detect nodules with high sensitivity and assist radiologists in improving diagnostic accuracy. Yet, as with any emerging technology, integrating AI into clinical practice will require further validation, particularly in prospective studies and real-world scenarios. Ultimately, the future of AI in adrenal nodule detection may not only lie in better diagnosis but also in more precise, personalized characterization and prediction strategies for patients with adrenal abnormalities.”
Unveiling the Future: A Deep Learning Model for Accurate Detection of Adrenal Nodules
Ashkan A. Malayeri • Baris Turkbey
Radiology 2025; 314(3):e250387 - “Finally, while AI-powered detection algorithms demonstrate ever-increasing accuracy, they must be implemented in diagnostic workflows considering the entire imaging study rather than one single phase. For example, because adrenal CT with washout calculation is the primary technique for differentiating benign versus malignant adrenal nodules, a future vision of AI in this context may entail not only detecting and segmenting nodules, but also analyzing them across multiple phases, extracting the radiomic data, and providing an objective predictability score on the malignant versus benign nature of the nodules.”
Unveiling the Future: A Deep Learning Model for Accurate Detection of Adrenal Nodules
Ashkan A. Malayeri • Baris Turkbey
Radiology 2025; 314(3):e250387
Kidney
- IMPORTANCE Testicular cancer is the most common solid malignancy among males aged 15 to40 years in the US, with approximately 10 000 new cases diagnosed each year. Between 90% and 95%of testicular cancers are germ cell tumors (GCTs).
CONCLUSIONS AND RELEVANCE Testicular cancer is the most common solid malignancy in young men in the US, and 90% to 95%are GCTs. Patients with testicular GCT have a 5- year survival rate of 99%, 92%, and 85%for stages I, II, and III, respectively. Prompt diagnosis and treatment are important to optimize outcomes, and treatment decisions should balance oncologic control with survivorship concerns to minimize long-term adverse effects of treatment.
Testicular Germ Cell Tumors: A Review.
Singla N, Bagrodia A, Baraban E, Fankhauser CD, Ged YMA.
JAMA. 2025 Mar 4;333(9):793-803. - The mean age at diagnosis for testicular cancer is 33 years. GCTs are categorized as seminomas and nonseminomatous GCTs (NSGCTs) based on their embryonic origins and path of differentiation. Risk factors include cryptorchidism, family history of testicular cancer, gonadal dysgenesis, infertility, cannabis use, and genetic conditions such as Klinefelter syndrome. The most common presenting symptom of testicular cancer is a painless testicular mass. History, physical examination, scrotal ultrasound, laboratory assessment of GCT-associated serum tumor markers (α-fetoprotein, β-human chorionic gonadotropin, and lactate dehydrogenase), and prompt referral to a urologist are indicated when testicular cancer is suspected.
Testicular Germ Cell Tumors: A Review.
Singla N, Bagrodia A, Baraban E, Fankhauser CD, Ged YMA.
JAMA. 2025 Mar 4;333(9):793-803. - “The most common presenting symptom of testicular cancer is a painless testicular mass. History, physical examination, scrotal ultrasound, laboratory assessment of GCT-associated serum tumor markers (α-fetoprotein, β-human chorionic gonadotropin, and lactate dehydrogenase), and prompt referral to a urologist are indicated when testicular cancer is suspected. Early diagnosis and treatment, starting with a radical inguinal orchiectomy, are important to optimize outcomes. At diagnosis, GCT is stage I (localized to the testicle) in 70%to 75%of patients, stage II (metastatic only to the retroperitoneal lymph nodes) in 20%, and stage III (widely metastatic) in 10%. Treatment of GCTs is guided by histology, clinical staging, and risk classification, with 5-year survival rates of 99%, 92%, and 85%for those diagnosed at stages I, II, and III, respectively. Optimal treatment often involves a multidisciplinary team at high-volume, experienced medical centers and may include surveillance (serum tumor markers [α-fetoprotein, β-human chorionic gonadotropin, and lactate dehydrogenase] and imaging of the chest, abdomen, and pelvis), surgery (retroperitoneal lymph node dissection), chemotherapy, and/or radiation.”
Testicular Germ Cell Tumors: A Review.
Singla N, Bagrodia A, Baraban E, Fankhauser CD, Ged YMA.
JAMA. 2025 Mar 4;333(9):793-803. - “In 90% of patients with testicular cancer, the presenting symptom is a testicular abnormality,34 most commonly a painless testicular mass identified incidentally or after minor testicular trauma that may prompt a patient to become aware of a testicular mass. Approximately 10% of patients with testicular cancer have acute testicular pain attributable to rapid tumor growth that may cause intratesticular hemorrhage, testicular infarction, or rarely testicular torsion. Scrotal swelling, heaviness, discomfort, and testicular shrinkage are less common (<10%) manifestations of testicular cancer.”
Testicular Germ Cell Tumors: A Review.
Singla N, Bagrodia A, Baraban E, Fankhauser CD, Ged YMA.
JAMA. 2025 Mar 4;333(9):793-803. - “In 10% to 20%of patients, initial symptoms of testicular cancer result from metastatic disease. Spread of testicular cancer to the retroperitoneal lymph nodes may present as a palpable abdominal mass, which can obstruct the ureters causing flank pain; compress or invade the bowel causing gastrointestinal symptoms such as nausea, bloating, or constipation; impede venous blood flow from the testis causing a varicocele; or invade the spinous muscles or nerves causing back pain. Pulmonary metastases, which occur in 5%of patients with testicular cancer,may present as dyspnea, cough, chest pain, or hemoptysis. Palpable neck masses in young men may be caused by testicular cancer metastasis to supraclavicular lymph nodes.”
Testicular Germ Cell Tumors: A Review.
Singla N, Bagrodia A, Baraban E, Fankhauser CD, Ged YMA.
JAMA. 2025 Mar 4;333(9):793-803. - “The life expectancy of men diagnosed with testicular cancer at 30 years of age is estimated at 75.2 years, which is 1.3 years less than men without cancer. Across all stages of testicular cancer, 5-year overall survival is 95% (99%for stage I, 92%for stage II, and 85% for stage III). Among patients with metastatic disease, 5-year progression-free and overall survival vary by IGCCCG prognostic group: for the good risk IGCCCG group, progression-free survival is 89% to 90% and overall survival is 96%; intermediate risk has 77% progression-free and 88% overall survival; and poor risk has 54% progression-free and 67% overall survival.”
Testicular Germ Cell Tumors: A Review.
Singla N, Bagrodia A, Baraban E, Fankhauser CD, Ged YMA.
JAMA. 2025 Mar 4;333(9):793-803. - Testicular cancer is the most common solid malignancy in young men in the US, with a mean age of 33 years at diagnosis. Approximately 90% to 95% are GCTs. Management and prognosis of testicular cancer are determined by histology and clinical stage. Patients with testicular GCTs have a 5-year survival rate of 99%, 92%, and 85%for stages I, II, and III, respectively. Prompt diagnosis and treatment are paramount to optimizing outcomes, and treatment decisions should balance oncologic control with survivorship concerns to minimize long-term adverse effects of treatment.
Testicular Germ Cell Tumors: A Review.
Singla N, Bagrodia A, Baraban E, Fankhauser CD, Ged YMA.
JAMA. 2025 Mar 4;333(9):793-803. 
Testicular Germ Cell Tumors: A Review.
Singla N, Bagrodia A, Baraban E, Fankhauser CD, Ged YMA.
JAMA. 2025 Mar 4;333(9):793-803.
- “Retroperitoneal fibrosis (RPF) is a rare fibroinflammatory disease with idiopathic and secondary causes. Idiopathic disease is more common and is believed to be immune mediated; associations with autoimmune diseases and/or inflammatory disorders such as IgG4 related disease are often present. Common complications include hydronephrosis and venous stenosis and/or thrombosis. Due to its nonspecific clinical presentation, imaging is vital for diagnosis; in addition, imaging may help distinguish idiopathic from secondary causes and can aid in distinguishing early/active disease from chronic/inactive disease.”
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - “Retroperitoneal fibrosis (RPF) is a rare disease characterized by the presence of inflammatory tissue and fibrosis centered in the retroperitoneum. The fibroinflammatory infiltrate classically develops along the anterolateral aspects of the infrarenal abdominal aorta extending inferiorly to the common iliac arteries; adjacent structures such as the ureters and other vessels are often encased resulting in hydronephrosis, venous thrombosis, and/or arterial compromise.”
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - “Both idiopathic and secondary forms of RPF exist, with the former accounting for approximately two-thirds of cases . Idiopathic RPF is an immune-mediated process that is included as part of the chronic periaortitis disease spectrum; however, increasingly, this has been associated with autoimmune diseases and/or inflammatory disorders such as IgG4 related disease. Secondary causes of RPF include malignant disease, medications (such as ergot alkaloids), and biological agents (monoclonal antibodies such as infliximab) amongst other etiologies .”
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - The pathogenesis of idiopathic RPF remains unclear though is likely multifactorial . The current hypothesis is that it may be a manifestation of a systemic autoimmune disease which arises as a primary aortitis; this elicits a fibroinflammatory response which subsequently extends to the adjacent retroperitoneal tissues . Inflammation in the adventitia of affected aortic segments coupled with the presence of systemic symptoms, as well as an association with autoimmune disorders and a good response to immunosuppressive therapies all support this hypothesis.
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - “Computed tomography (CT) has become one of the preferred imaging modalities in assessing RPF, allowing a thorough evaluation of disease extent, location and morphology . In addition, CT can be used to assess for other conditions associated with RPF (IgG4 related diseases such as autoimmune pancreatitis or malignant etiologies). Even so, RPF has been reported to have no CT correlation in one third of patients with surgically proven disease . At present, the main utility of CT imaging lies in the ability to assess changes in size of the fibroinflammatory mass on serial imaging studies.”
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - The most common imaging findings include a confluent sheet-like mass surrounding the anterior and lateral aspects of the aorta, typically centered at the aortic bifurcation with caudal extension to the common iliac arteries. Typically, there is no lateral extension beyond the lateral aspects of the psoas muscle. Cephalad extension to the level of the renal arteries or caudal extension to the sacrum may be present; additional extension to retroperitoneal organs/spaces (such as the pancreas, perinephric fat, and the retroperitoneal portions of the duodenum) has also been reported.
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - “Radiomics is an emerging image processing technique which leverages artificial intelligence and machine learning algorithms to extract quantitative textures/features from regions of interest on radiology images. One study demonstrated that a radiomics algorithm had a discriminative accuracy of 72% in detecting residual fibrosis in retroperitoneal lymphadenopathy from testicular germ cell tumors after chemotherapy; this improved to 88% when combined with clinical predictors (prechemotherapy tumor markers, residual mass size, percentage of mass shrinkage, and the presence of teratoma elements in orchiectomy specimen.”
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - “Confidently distinguishing between idiopathic and malignant etiologies of RPF on FDG-PET has also proved challenging. Some work on the topic, however, has shown promise; in one study which compared idiopathic RPF with lymphoma and metastatic disease to the retroperitoneum, RPF was found to have a lower frequency of high FDG uptake as well as a lower mean maximum standardized uptake value (mean SUVmax 4.8 for RPF versus 13.5 and 8.7 for lymphoma and metastases respectively). Patients with lymphoma and metastatic disease were also found to have adenopathy located at distant sites including axillary, supraclavicular, inguinal and the peritoneum.”
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. 
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807.
Liver
- “Radiomics is the high-throughput extraction of quantitative features from radiological images through data characterization algorithms. These features contain information representing pathophysiological mechanisms imperceptible to the human eye. Radiomic features have been linked to metabolic dysfunction– associated steatotic liver disease, cirrhosis, and tumor biological processes like hypoxia, angiogenesis, and inflammation. Quantitative assessment of these features can be used to predict the behavior of cancer, postoperative course, and surgical morbidity.”
Radiomics for Treatment Planning in Liver Cancers.
Mian A, Kamnitsas K, Gordon-Weeks A.
JAMA Surg. 2025 Feb 26. doi: 10.1001/jamasurg.2024.4346. Epub ahead of print. PMID: 40009391. - “Radiomics enables the noninvasive profiling of tumors through routine radiological images. By characterizing the tumor biological phenotype through imaging data, radiomics can provide personalized pretreatment prognostication. It can eliminate invasive biopsies and outperforms somatic sequence variation testing for selection of biological therapies—a particularly important consideration for colorectal liver metastases (CRLMs), where biopsy significantly risks tumor cell seeding, and for cholangiocarcinoma, which is technically challenging to sample. Implementation does not require specialist diagnostic or biochemical expertise and is, in principal, cost-effective compared with competitor prognostics, including circulating tumor DNA or cancer cells.”
Radiomics for Treatment Planning in Liver Cancers.
Mian A, Kamnitsas K, Gordon-Weeks A.
JAMA Surg. 2025 Feb 26. doi: 10.1001/jamasurg.2024.4346. Epub ahead of print. PMID: 40009391. - “Similarly, in HCC, computed tomography (CT) radiomic texture analysis, maybe used to facilitate therapeutic decision-making alongside the Barcelona Clinic Liver Cancer (BCLC) criteria when considering surgical resection, transplant, or other liver-directed therapies. In CRLM, radiomics may also assist in determining optimal surgical resection margins by predicting KRAS sequence variations or identifying replacement-type growth, both of which require wider resection margins. Finally, radiomics approaches could aid in forecasting the volume and health of the future liver remnant post–major hepatectomy.”
Radiomics for Treatment Planning in Liver Cancers.
Mian A, Kamnitsas K, Gordon-Weeks A.
JAMA Surg. 2025 Feb 26. doi: 10.1001/jamasurg.2024.4346. Epub ahead of print. PMID: 40009391. - “Radiomics has demonstrated multiple uses in liver cancers. One large, prospective study demonstrated that a radiomics signature derived from CRLM CT scans outperformed the Response Evaluation Criteria in Solid Tumors (RECIST) criteria for predicting folinic acid, fluorouracil, and irinotecan (FOLFIRI) and bevacizumab response. These findings were supported by another study, where radiomics features outperformed RECIST and KRAS sequence variation status, with the radiomics signature predicting response to FOLFIRI and cetuximab with an area under the curve (AUC) of 0.80 compared with only 0.67 for KRAS status and 0.75 for 8-week tumor shrinkage measured by RECIST. This information can guide early chemotherapy regimen switching in nonresponders and identify patients for resection based on favorable biology. Futurework should explore if radiomics provides relevant insights into the biology of KRAS mutant vs wild-type CRLM.”
Radiomics for Treatment Planning in Liver Cancers.
Mian A, Kamnitsas K, Gordon-Weeks A.
JAMA Surg. 2025 Feb 26. doi: 10.1001/jamasurg.2024.4346. Epub ahead of print. PMID: 40009391. - “In HCC, radiomics and clinical features predicted transarterial chemoembolization (TACE) response with an AUC of 0.90, which was linked to improved survival in the external validation cohort (hazard ratio, 2.43; P < .05). This study, which focused on patients with BCLC stage B disease, raises the question of whether a selection of patients may benefit from liver transplant outside the Milan criteria vs palliative TACE alone. For cholangiocarcinoma, a combination of radiomics and clinical features improved recurrence-free survival prediction (C index: 0.75) vs clinical features alone (C index: 0.69) in external validation. Given high rates of postsurgical recurrence in this patient cohort, further improvements will identify a subset of high-risk patients who should be offered neoadjuvant systemic therapy as a test of biological behavior rather than up-front resection.”
Radiomics for Treatment Planning in Liver Cancers.
Mian A, Kamnitsas K, Gordon-Weeks A.
JAMA Surg. 2025 Feb 26. doi: 10.1001/jamasurg.2024.4346. Epub ahead of print. PMID: 40009391. - “A major challenge in radiomics is limited generalizability of models due to reliance on single-center data, which fails to account for variability in patient characteristics (eg, ethnicity, age, tumor biology), imaging protocols, and the scanners differences. Some radiomic features are sensitive to these variations, reducing reproducibility across institutions. Solutions include using robust features, multiinstitutional data, standardized imaging protocols, data augmentation (eg, generative adversarial networks), and harmonization techniques likeComBat. The Image Biomarker Standardization Initiative has played a key role in standardizing radiomic biomarker nomenclature and definitions. It has also created benchmark datasets and benchmarking values to verify image processing and biomarker calculations, while also developing robust reporting guidelines high-throughput image analysis. This is now helping drive the reliability and reproducibility of radiomic analyses across diverse datasets to develop generalizable radiomic models.”
Radiomics for Treatment Planning in Liver Cancers.
Mian A, Kamnitsas K, Gordon-Weeks A.
JAMA Surg. 2025 Feb 26. doi: 10.1001/jamasurg.2024.4346. Epub ahead of print. PMID: 40009391.
Pancreas
- “Advancements in Artificial Intelligence (AI) are driving a paradigm shift in the field of medical diagnostics, integrating new developments into various aspects of the clinical workflow. Neuroendocrine neoplasms are a diverse and heterogeneous group of tumors that pose significant diagnostic and management challenges due to their variable clinical presentations and biological behavior. Innovative approaches are essential to overcome these challenges and improve the current standard of care. AI-driven applications, particularly in imaging workflows, hold promise for enhancing tumor detection, classification, and grading by leveraging advanced radiomics and deep learning techniques.”
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28: - “AI has proven capabilities in enhancing the clinical management of NENs from multiple aspects, including analysis of serum biomarkers, pathological specimens, and genomics signatures. This review focuses on the impact of AI on diagnostic imaging of NENs, including both anatomical and functional imaging. The review also highlights emerging AI methodologies, their potential integration into clinical workflows, and the opportunities and challenges associated with leveraging AI to improve outcomes in NET management.”
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28: - “Although different approaches exist for AI models to find informative patterns, radiomics models and deep learning architectures are among the most widely used techniques for medical image analysis within the scope of computer vision. Radiomics refers to the high-throughput extraction of quantitative imaging features that characterize tissue phenotypes, such as texture, shape, and intensity, often revealing patterns beyond human perception. Deep learning employs complex neural network architectures, resembling biological neurons and synapses, to identify connections and patterns and extract meaningful features and associations directly from imaging data.”
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28: - “Alternatively, in scenarios where a tumor has already been identified, AI can be used to detect imaging patterns associ ated with different tumor types and help the diagnostician differentiate between NENs and other potential clinical entities. In this context, there is awareness of the tumor and go to AI to assist in selecting the most likely option among the differential diagnoses. However, a significant barrier to clinical implementation remains the labor-intensive process of manual segmentation required to define regions of interest (ROIs) for feature characterization. Advances in deep learning-based automated segmentation have partially addressed this challenge, offering a more efficient and scalable clinical integration.”
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28: - “In the context of a known NEN, AI models are commonly evaluated in clinical settings to characterize the tumor and predict clinicopathological characteristics that are relevant for determining prognosis. Perhaps due to the high relevance of tumor grade in NENs, along with the high morbidity associated with pancreatic surgery, the most common tumor characterization task among published studies is the prediction of tumor grade in PNETs. Despite the high clinical relevance of differentiating NETs from NECs, with previous exploratory analyses suggesting that imaging features could potentially distinguish between these tumor types, 46 very few studies have explored this task. Most effort has instead concentrated on differentiating between grades of NETs, likely due to limitations in sample size and the underrepresentation of NEC in institutional databases.”
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28: - Despite encouraging results, many barriers remain before AI models can be widely applied to support the clinical management of NENs.While most challenges are common across AI applications in medicine, the relative rarity of NENs and high heterogeneity in tumor biology and disease presentation amplify these difficulties. A major challenge with AI lies in the generalization of algorithms beyond he training cohort, often resulting in significant performance reductions for underrepresented groups. Most current NEN studies have sample sizes too small to develop robust, generalizable models, emphasizing the urgent need for collaborative efforts to create publicly available annotated imaging datasets in NENs. These datasets would enable large-scale training of radiomics and deep learning models, which are critical for improving AI performance and generalizability. Additionally, studies must prioritize demonstrating the clinical value of AI-prediction through standardized prospective studies compared with current diagnostic gold standards, a key step toward ensuring clinical relevance.
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28: - AI holds immense promise in revolutionizing the diagnosis, characterization, and management of neuroendocrine neoplasms. By integrating advanced imaging techniques, predictive analytics, and workflow optimization, AI enhances clinical decision-making and paves the way for personalized medicine. These tools can enhance classification accuracy, support the diagnosticians by reducing diagnostic uncertainty, and provide more reliable assessment to guide treatment planning. Despite challenges in standardization, validation, and limited training data due to the rarity of the disease, ongoing advancements in AI methodologies are set to transform management of neuroendocrine neoplasms, improving patient outcomes through more precise and efficient care.
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28:
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28:
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28:
The Role of AI in the Evaluation of Neuroendocrine Tumors: Current State of the Art.
Lopez-Ramirez F, Yasrab M, Tixier F, Kawamoto S, Fishman EK, Chu LC.
Semin Nucl Med. 2025 Feb 28:- Cinematic rendering is post-processing technique based on a complex global lighting model that results in highly photorealistic anatomic images. The global lighting model results in high degrees of surface detail and shadowing effects in the final three-dimensional display. Cinematic rendering is regarded as a marked advance in the evaluation of abdominal cancers and more specifically in pancreatic cancers. In patients amenable to surgery, preoperative planning of PDAC requires accurate visualization of vascular structures in relation to the tumour and also identification of anatomical variants, such as presence of arcuate ligament or celiomesenteric trunk, that may make surgery difficult or increase the risk of intraoperative vascular complications. In this regard, cinematic rendering helps surgeon identify arterial anatomic variants and vascular involvement with high degrees of confidence.13 Regarding PanNET and cystic pancreatic tumours, cinematic rendering allows for more comprehensive visualization and description, and interpretation of anatomical structures along with a global assessment of the burden of metastatic spread.
CT Imaging of the Pancreas: A Review of Current Developments and Applications.
Barat M, Greffier J, Si-Mohamed S, Dohan A, Pellat A, Frandon J, Calame P, Soyer P.
Can Assoc Radiol J. 2025 Feb 22:8465371251319965. doi: 10.1177/08465371251319965. Epub ahead of print. PMID: 39985297. - Despite advances in MRI, CT remains a major diagnostic tool for pancreatic imaging. The benefit of spectral PCCT for image quality is already suggested in many applications, particularly those requiring high spatial resolution and contrast. Spectral PCCT expands the capabilities of spectral CT for several applications such as diagnosis, characterization, and staging of diseases and the anticipated development of colour K-edge contrast agents should open new research areas. Radiomics shows encouraging results in terms of tumour resectability prediction. AI is a field of major ongoing research. One goal for AI could be the detection of small PDAC at an early stage, as a substantial proportion of these cancers remain not visible to the human eyes. Future studies should also include sophisticated models that associate imaging data with clinical and biological data. The recent introduction of new therapies should also stimulate further studies for a better evaluation of tumour response.
CT Imaging of the Pancreas: A Review of Current Developments and Applications.
Barat M, Greffier J, Si-Mohamed S, Dohan A, Pellat A, Frandon J, Calame P, Soyer P.
Can Assoc Radiol J. 2025 Feb 22:8465371251319965. doi: 10.1177/08465371251319965. - “One difficulty in pancreatic cancer is the detection of tumours that are isoattenuating to the adjacent pancreatic parenchyma or tumours that are <20 mm. Chen et al have developed a deep learning-based tool that achieved 89.7% (600/669) sensitivity, 92.8% specificity (746/804), and an AUC of 0.95, with 74.7% sensitivity for cancers <20 mm in a test set of 1473 CT examinations.82 Korfiatis et al have developed a three-dimensional convolutional neural network classification system trained on a large data set that achieved 84% accuracy and an AUC of 0.91 for the diagnosis of visually occult PDAC on prediagnostic CT examinations.”
CT Imaging of the Pancreas: A Review of Current Developments and Applications.
Barat M, Greffier J, Si-Mohamed S, Dohan A, Pellat A, Frandon J, Calame P, Soyer P.
Can Assoc Radiol J. 2025 Feb 22:8465371251319965. doi: 10.1177/08465371251319965.
Small Bowel
- Meckel’s Diverticulum: Complications
- GI bleeding (usually in pediatric patients)
- Small bowel obstruction
- Diverticulitis
- May be difficult to distinguish from processes like acute appendicitis, inflammatory bowel disease (IBD), other causes of small bowel obstruction (SBO) - Meckels Diverticulum: Facts
- Most common congenital anomaly of GI tract and occurs in up to 2% of the population
- Occurs within 60 cm of ileocecal valve and is 6 cm in length
- Up to 57% of these contain ectopic gastric mucosa which is responsible for clinical symptoms
- Detection by Tc-99m pertechnetate scan which is actively accumulated and secreted by mucoid cells of gastric mucosa 
- OBJECTIVE. The objective of our study was to determine how often symptomatic Meckel diverticulum and asymptomatic Meckel diverticulum are detected on CT in patients with known Meckel diverticulum and to evaluate factors that influence detection.
CONCLUSION. CT can detect Meckel diverticulum in up to 47.5% of all patients with known Meckel diverticulum. Meckel diverticulum is more commonly detected in symptomatic patients than in asymptomatic patients, and detection is related to the amount of peritoneal fat.
CT Detection of Symptomatic and Asymptomatic Meckel Diverticulum.
Kawamoto S, Raman SP, Blackford A, Hruban RH, Fishman EK.
AJR Am J Roentgenol. 2015 Aug;205(2):281-91. doi: 10.2214/AJR.14.13898. PMID: 26204277. - “Meckel diverticulum, which is also called “Meckel’s diverticulum,” is estimated to occur in approximately 1–2% of the population. Although most people with Meckel diverticulum remain asymptomatic during their life, 16–20% of Meckel diverticula are symptomatic . Well-known complications of Meckel diverticulum include hemorrhage from a peptic ulcer related to ectopic gastric mucosa, small-bowel obstruction, intussusception, and diverticulitis without or with perforation. However, symptomatic Meckel diverticulum is often difficult to differentiate from a variety of other intraabdominal processes that cause acute abdominal pain or gastrointestinal hemorrhage, and preoperative diagnosis of symptomatic Meckel diverticulum is often difficult to establish.”
CT Detection of Symptomatic and Asymptomatic Meckel Diverticulum.
Kawamoto S, Raman SP, Blackford A, Hruban RH, Fishman EK.
AJR Am J Roentgenol. 2015 Aug;205(2):281-91. doi: 10.2214/AJR.14.13898. PMID: 26204277. - “Meckel diverticulum may cause bowel obstruction by several mechanisms including an intussusception, volvulus and internal hernia due to persistent attachment of the diverticulum to the umbilicus, mesodiverticular band, adhesion, luminal obstruction from an inverted diverticulum, diverticulitis, foreign body impaction in the diverticulum, inclusion of the diverticulum into a hernia, or neoplastic obstruction. In the cases of small-bowel obstruction associated with Meckel diverticulum in our study, the small-bowel obstruction was clearly seen on CT, but the underlying Meckel diverticulum was often difficult to detect. “
CT Detection of Symptomatic and Asymptomatic Meckel Diverticulum.
Kawamoto S, Raman SP, Blackford A, Hruban RH, Fishman EK.
AJR Am J Roentgenol. 2015 Aug;205(2):281-91. doi: 10.2214/AJR.14.13898. PMID: 26204277.
- “Mucinous neoplasms of the appendix are mucin-producing epithelial tumors classified into four types: adenoma, low-grade appendiceal mucinous neoplasm (LAMN), high-grade appendiceal mucinous neoplasm (HAMN), and mucinous adenocarcinoma. The peak prevalence is in middle-aged adults, with a slight female predominance. These tumors are typically asymptomatic and discovered incidentally at imaging, but occasionally patients present with acute appendicitis or other nonspecific abdominal symptoms related to peritoneal spread.”
Approach to Cystic Lesions in the Abdomen and Pelvis, with Radiologic-Pathologic Correlation.
Yacoub JH, Clark JA, Paal EE, Manning MA.
Radiographics. 2021 Sep-Oct;41(5):1368-1386. - “These neoplasms commonly appear as an appendiceal mucocele, a macroscopic description of an appendix abnormally distended with mucin Appendiceal mucoceles can be caused by nonneoplastic entities or by mucin-secreting epithelial neoplasms of the appendix. Non-neoplastic entities are most commonly mucous retention cysts caused by luminal obstruction and tend to be smaller than mucinous neoplasms. “
Approach to Cystic Lesions in the Abdomen and Pelvis, with Radiologic-Pathologic Correlation.
Yacoub JH, Clark JA, Paal EE, Manning MA.
Radiographics. 2021 Sep-Oct;41(5):1368-1386.
- Cystic neoplasms (mucinous cystic neoplasm[MCN], multicystic peritoneal mesothelioma[MCPM], pseudomyxoma peritonei [PMP])are more complex, with multilocularity, septa, and calcifications. Solid nodules are rare, and when present, are worrisome. Thickened calcified walls are most commonly associated with inflammatory, infectious, and iatrogenic cystic lesions. Finally, soft-tissue neoplasms or lymph nodes with central necrosis or cystic change can mimic a cystic lesion and should be considered for thick-walled complex cystic lesions with large solid components.”
Approach to Cystic Lesions in the Abdomen and Pelvis, with Radiologic-Pathologic Correlation.
Yacoub JH, Clark JA, Paal EE, Manning MA.
Radiographics. 2021 Sep-Oct;41(5):1368-1386. - “At CT, a cyst appears as a well-defined spheroid lesion with homogeneous near-water attenuation (0–20 HU). Cyst contents can be divided into three categories: homogeneous near-water attenuation, suggesting simple fluid; lower-than-water attenuation and fluid-fluid levels, suggesting chylous fluid and lipid contents; and higher-than-water attenuation, which can be homogeneous or heterogeneous owing to the presence of proteinaceous material, hemorrhage, or necrotic tissue.”
Approach to Cystic Lesions in the Abdomen and Pelvis, with Radiologic-Pathologic Correlation.
Yacoub JH, Clark JA, Paal EE, Manning MA.
Radiographics. 2021 Sep-Oct;41(5):1368-1386.
- “When approaching a cystic lesion, the key imaging features to assess include cyst content, locularity, wall thickness, and presence of internal septa, solid components, calcifications, or any associated enhancement. While definitive diagnosis is not always possible with imaging, careful assessment of the imaging appearance, location, and relationship to adjacent structures can help narrow the differential diagnosis.”
Approach to Cystic Lesions in the Abdomen and Pelvis, with Radiologic-Pathologic Correlation.
Yacoub JH, Clark JA, Paal EE, Manning MA.
Radiographics. 2021 Sep-Oct;41(5):1368-1386. 
Approach to Cystic Lesions in the Abdomen and Pelvis, with Radiologic-Pathologic Correlation.
Yacoub JH, Clark JA, Paal EE, Manning MA.
Radiographics. 2021 Sep-Oct;41(5):1368-1386.- “Peritoneal Simple Mesothelial Cyst.—PSMC is a congenital cyst formed in the small bowel, mesentery, mesocolon, or omentum, presumably after failure of the mesothelium-lined peritoneal surfaces to coalesce. Mesothelium-lined cysts have also been reported to arise from the pericardium and round ligament of the liver. PSMC is typically reported in children and young adults but has been reported in older adults. Usually asymptomatic, it can manifest with nonspecific abdominal symptoms including pain, distention, bloating, constipation, and vomiting. Children can present with acute abdomen due to rupture, inflammation, infection, torsion, or hemorrhage within the cyst.”
Approach to Cystic Lesions in the Abdomen and Pelvis, with Radiologic-Pathologic Correlation.
Yacoub JH, Clark JA, Paal EE, Manning MA.
Radiographics. 2021 Sep-Oct;41(5):1368-1386.
Syndromes in CT
- Castlemans Disease
Castleman disease is easily confused with lymphoma or other solid tumors. Therefore, it is essential to properly diagnose the exact type of CD and distinguish it from other diseases by clinical history and laboratory diagnostic measures with additional imaging techniques for prompt treatment and management procedures. Even though CD was defined as a benign lymphoproliferative disease, the systemic forms are particularly associated with related neoplasms and autoimmune disorders like Kaposi sarcoma and Follicular dendritic cell (FDC) tumors.[4] Lymphomas, especially Hodgkin and angioimmunoblastic T-cell lymphoma, are also known for polymorphous infiltrate and hypervascularity, are well-known to mimic Castleman disease.
The hyaline vascular CD can be strongly connected to FDC tumors, including dendritic cell sarcomas and neoplasms of the stroma. In addition, IL-6 release in CD may invoke molecular mimicry and epitope spreading, initiating a lichenoid interface dermatitis, leading to autoantibody production. Finally, it can be associated with a fatal autoimmune blistering disease called paraneoplastic pemphigus in patients with treatment-resistant erosive mucosal lesions.[16]
Vascular
- “Retroperitoneal fibrosis (RPF) is a rare fibroinflammatory disease with idiopathic and secondary causes. Idiopathic disease is more common and is believed to be immune mediated; associations with autoimmune diseases and/or inflammatory disorders such as IgG4 related disease are often present. Common complications include hydronephrosis and venous stenosis and/or thrombosis. Due to its nonspecific clinical presentation, imaging is vital for diagnosis; in addition, imaging may help distinguish idiopathic from secondary causes and can aid in distinguishing early/active disease from chronic/inactive disease.”
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - “Retroperitoneal fibrosis (RPF) is a rare disease characterized by the presence of inflammatory tissue and fibrosis centered in the retroperitoneum. The fibroinflammatory infiltrate classically develops along the anterolateral aspects of the infrarenal abdominal aorta extending inferiorly to the common iliac arteries; adjacent structures such as the ureters and other vessels are often encased resulting in hydronephrosis, venous thrombosis, and/or arterial compromise.”
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - “Both idiopathic and secondary forms of RPF exist, with the former accounting for approximately two-thirds of cases . Idiopathic RPF is an immune-mediated process that is included as part of the chronic periaortitis disease spectrum; however, increasingly, this has been associated with autoimmune diseases and/or inflammatory disorders such as IgG4 related disease. Secondary causes of RPF include malignant disease, medications (such as ergot alkaloids), and biological agents (monoclonal antibodies such as infliximab) amongst other etiologies .”
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - The pathogenesis of idiopathic RPF remains unclear though is likely multifactorial . The current hypothesis is that it may be a manifestation of a systemic autoimmune disease which arises as a primary aortitis; this elicits a fibroinflammatory response which subsequently extends to the adjacent retroperitoneal tissues . Inflammation in the adventitia of affected aortic segments coupled with the presence of systemic symptoms, as well as an association with autoimmune disorders and a good response to immunosuppressive therapies all support this hypothesis.
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - “Computed tomography (CT) has become one of the preferred imaging modalities in assessing RPF, allowing a thorough evaluation of disease extent, location and morphology . In addition, CT can be used to assess for other conditions associated with RPF (IgG4 related diseases such as autoimmune pancreatitis or malignant etiologies). Even so, RPF has been reported to have no CT correlation in one third of patients with surgically proven disease . At present, the main utility of CT imaging lies in the ability to assess changes in size of the fibroinflammatory mass on serial imaging studies.”
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - The most common imaging findings include a confluent sheet-like mass surrounding the anterior and lateral aspects of the aorta, typically centered at the aortic bifurcation with caudal extension to the common iliac arteries. Typically, there is no lateral extension beyond the lateral aspects of the psoas muscle. Cephalad extension to the level of the renal arteries or caudal extension to the sacrum may be present; additional extension to retroperitoneal organs/spaces (such as the pancreas, perinephric fat, and the retroperitoneal portions of the duodenum) has also been reported.
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - “Radiomics is an emerging image processing technique which leverages artificial intelligence and machine learning algorithms to extract quantitative textures/features from regions of interest on radiology images. One study demonstrated that a radiomics algorithm had a discriminative accuracy of 72% in detecting residual fibrosis in retroperitoneal lymphadenopathy from testicular germ cell tumors after chemotherapy; this improved to 88% when combined with clinical predictors (prechemotherapy tumor markers, residual mass size, percentage of mass shrinkage, and the presence of teratoma elements in orchiectomy specimen.”
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. - “Confidently distinguishing between idiopathic and malignant etiologies of RPF on FDG-PET has also proved challenging. Some work on the topic, however, has shown promise; in one study which compared idiopathic RPF with lymphoma and metastatic disease to the retroperitoneum, RPF was found to have a lower frequency of high FDG uptake as well as a lower mean maximum standardized uptake value (mean SUVmax 4.8 for RPF versus 13.5 and 8.7 for lymphoma and metastases respectively). Patients with lymphoma and metastatic disease were also found to have adenopathy located at distant sites including axillary, supraclavicular, inguinal and the peritoneum.”
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807. 
Multimodality imaging review of retroperitoneal fibrosis.
Czerniak S, Mathur M.
Abdom Radiol (NY). 2025 Mar 4. doi: 10.1007/s00261-025-04847-6. Epub ahead of print. PMID: 4035807.
