Imaging Pearls ❯ Deep Learning ❯ Clinical Applications (General)
|
-- OR -- |
|
- “Arguably the most impressive and distinguishing aspect of AI is its automated ability to search and extract arbitrary, complex, task-oriented features from data — so-called feature representation learning. Features are algorithmically engineered from data during a training phase in order to uncover data transformations that are correct for the learning task. Optimality is measured by means of an “objective function” quantifying how well the AI model is performing the task at hand. AI algorithms largely remove the need for analysts to prespecify features for prediction or manually curate transformations of variables. These attributes are particularly beneficial in large, complex data domains such as image analysis, genomics, or modeling of electronic health records.”
Where Medical Statistics Meets Artificial Intelligence
David J. Hunter, M.B., B.S., and Christopher Holmes, Ph.D.
N Engl J Med 2023;389:1211-9.
- “Integration of AI in the radiological clinical workflow is imminent. As these solutions are implemented, tools should be designed to train the radiologist, detect errors, and reduce bias and skill erosion. Potential solutions include external governing bodies, training platforms, competency-based learning portfolios, simulation, and changes to user interface and algorithm output. As these new solutions are developed, each type of user, from junior learners to experienced radiologists, should be considered so that our specialists can improve care and reduce potential harm.”
Are the Pilots Onboard? Equipping Radiologists for Clinical Implementation of AI
Umber Shafique ·Umar Shafique Chaudhry · Alexander J. Towbin
Journal of Digital Imaging (2023) 36:2329–2334 - “AI and its impact on learning should be a focus for resident and fellow training. We believe that the training of junior learners should include the basics of AI in addition to the training described above for each algorithm. This added level of initial training is needed so that the next generation of radiologists can address the concepts of training and skill erosion as AI increasingly becomes integrated within clinical practice. Given the rapidly changing landscape of radiology, we believe that this AI training is equivalent to the necessity of learning physics in that it impacts many elements of the daily workflow and is crucial to enable problem-solving.”
Are the Pilots Onboard? Equipping Radiologists for Clinical Implementation of AI
Umber Shafique · Umar Shafique Chaudhry · Alexander J. Towbin
Journal of Digital Imaging (2023) 36:2329–2334
- ”Integrating AI into healthcare holds excellent potential for improving disease diagnosis, treatment selection, and clinical laboratory testing. AI tools can leverage large datasets and identify patterns to surpass human performance in several healthcare aspects. AI offers increased accuracy, reduced costs, and time savings while minimizing human errors. It can revolutionize personalized medicine, optimize medication dosages, enhance population health management, establish guidelines, provide virtual health assistants, support mental health care, improve patient education, and influence patient-physician trust. ”
Alowais SA, et al. Revolutionizing healthcare: the role of artificial intelligence in clinical practice.
Alowais SA, et al.
BMC Med Educ. 2023 Sep 22;23(1):689 - ”AI can be used to diagnose diseases, develop personalized treatment plans, and assist clinicians with decision-making. Rather than simply automating tasks, AI is about developing technologies that can enhance patient care across healthcare settings. However, challenges related to data privacy, bias, and the need for human expertise must be addressed for the responsible and effective implementation of AI in healthcare. ”
Alowais SA, et al. Revolutionizing healthcare: the role of artificial intelligence in clinical practice.
Alowais SA, et al.
BMC Med Educ. 2023 Sep 22;23(1):689 - ”The integration of AI in healthcare has immense potential to revolutionize patient care and outcomes. AI-driven predictive analytics can enhance the accuracy, efficiency, and cost-effectiveness of disease diagnosis and clinical laboratory testing. Additionally, AI can aid in population health management and guideline establishment, providing real-time, accurate information and optimizing medication choices. Integrating AI in virtual health and mental health support has shown promise in improving patient care. However, it is important to address limitations such as bias and lack of personalization to ensure equitable and effective use of AI.”
Alowais SA, et al. Revolutionizing healthcare: the role of artificial intelligence in clinical practice.
Alowais SA, et al.
BMC Med Educ. 2023 Sep 22;23(1):689 - “Public perception of AI in healthcare varies, with individuals expressing willingness to use AI for health purposes while still preferring human practitioners in complex issues. Trust-building and patient education are crucial for the successful integration of AI in healthcare practice. Overcoming challenges like data quality, privacy, bias, and the need for human expertise is essential for responsible and effective AI integration.”
Alowais SA, et al. Revolutionizing healthcare: the role of artificial intelligence in clinical practice.
Alowais SA, et al.
BMC Med Educ. 2023 Sep 22;23(1):689
- “To translate artificial intelligence (AI) algorithms into clinical practice requires generalizability of models to real-world data. One of the main obstacles to generalizability is data shift, a data distribution mismatch between model training and real environments. Explainable AI techniques offer tools to detect and mitigate the data shift problem and develop reliable AI for clinical practice. Most medical AI is trained with datasets gathered from limited environments, such as restricted disease populations and center-dependent acquisition conditions. The data shift that commonly exists in the limited training set often causes a significant performance decrease in the deployment environment. To develop a medical application, it is important to detect potential data shift and its impact on clinical translation. During AI training stages, from premodel analysis to in-model and post hoc explanations, explainability can play a key role in detecting model susceptibility to data shift, which is otherwise hidden because the test data have the same biased distribution as the training data. Performance-based model assessments cannot effectively distinguish the model overfitting to training data bias without enriched test sets from external environments. In the absence of such external data, explainability techniques can aid in translating AI to clinical practice as a tool to detect and mitigate potential failures due to data shift.”
Translating AI to Clinical Practice: Overcoming Data Shift with Explainability
Youngwon Choi et al.
RadioGraphics 2023; 43(5):e220105 May 2023 - • One of the main challenges is the lack of AI applications that work well across multiple institutions or heterogeneous populations. It is often seen that a deep learning model trained with data from one hospital fails at other hospitals.
• In health care, with its limited datasets, explainability can offer tools that can be helpful to detect potential failures due to data shift in translating AI to clinical application.
•Data shift is one of the major obstacles because it is not easily detected or addressed with the classic techniques for preventing overfitting, which assume independent and identically distributed (IID) data.
Translating AI to Clinical Practice: Overcoming Data Shift with Explainability
Youngwon Choi et al.
RadioGraphics 2023; 43(5):e220105 May 2023 - • AI systems with proper explanations can detect data bias, which may cause model failures in external datasets. This section introduces some approaches to explainability based on the stage of model training: the (a) premodel approach, (b) in-model approach, and (c) postmodel approach.
• We introduced techniques for explainability and reviewed example uses of explanations for detecting the data shift problem. The examples demonstrate that explainability can reveal potential data shift issues at the model training stage. With domain knowledge, the explanations provide information for sanity checks and robust model selection
Translating AI to Clinical Practice: Overcoming Data Shift with Explainability
Youngwon Choi et al.
RadioGraphics 2023; 43(5):e220105 May 2023 - “The training sets of the models are usually collected from a few data sources with specific clinical conditions, then the trained models are applied to a different more variable environment in clinical practice. Both the training and test sets have probability distributions that may differ from the unknown real distribution of the target disease spectrum because they are collected from the restricted environments. This change in distribution is known as data shift and can degrade AI performance when transitioning from training to real-world application.”
Translating AI to Clinical Practice: Overcoming Data Shift with Explainability
Youngwon Choi et al.
RadioGraphics 2023; 43(5):e220105 May 2023 - “From both reviews, data shift was the predominant cause of the high risks of bias. Although many articles considered overfitting and showed their generalizability to the test set, none of the models were assessed as having potential data shift problems, for example, population difference, nonrepresentative selection of control patients, or exclusion of some patients. Without a rich number of external datasets from the various environments, data shift failure is difficult to detect with classic methods such as cross-validation. We propose use of explainable AI to detect potential model failures due to data shift.”
Translating AI to Clinical Practice: Overcoming Data Shift with Explainability
Youngwon Choi et al.
RadioGraphics 2023; 43(5):e220105 May 2023 - “Many state-of-the-art medical AI algorithms are susceptible to performance drops due to differences between the model training and real environments. This is known as the data shift problem and raises concerns about the reliability of AI in clinical practice. This data shift issue is challenging because it is often not apparent during model training. To make a model reliable in various environments, it is important to recognize possible data shift and to prevent the model from overfitting to the distribution bias. We described examples to crystallize the negative impact of the five common types of data shifts in medical imaging: population shift, prevalence shift, acquisition shift, annotation shift, and manifestation shift.”
Translating AI to Clinical Practice: Overcoming Data Shift with Explainability
Youngwon Choi et al.
RadioGraphics 2023; 43(5):e220105 May 2023 - “We highlighted the importance of explainability as a prerequisite for translating AI to clinical application. Explainability plays a central role in detecting this hidden model failure by providing human-understandable explanations about what contributes to the model’s decision making. We introduced techniques for explainability and reviewed example uses of explanations for detecting the data shift problem. The examples demonstrate that explainability can reveal potential data shift issues at the model training stage. With domain knowledge, the explanations provide information for sanity checks and robust model selection.”
Translating AI to Clinical Practice: Overcoming Data Shift with Explainability
Youngwon Choi et al.
RadioGraphics 2023; 43(5):e220105 May 2023
- “For some world problems, relying on AI is the only option we have. Diabetic retinopathy (DR) is the leading cause of vision loss globally, and out of the 451 million people worldwide with diabetes, a third of these people will develop DR. If treated, blindness due to DR is completely preventable, but the problem is that for many individuals a proper diagnosis of DR is impossible given the fact that there are only 200,000 ophthalmologists in the world. Thankfully, AI models have >97% accuracy (on par with ophthalmologists) in detecting DR, compensating for the lack of physicians capable of making this diagnosis. However, despite AI’s promising potential, it comes with challenges and limitations of its own.”
AI as a Public Service.
Lavista Ferres JM, Fishman EK, Rowe SP, Chu LC, Lugo-Fagundo E.
J Am Coll Radiol. 2023 Mar 30:S1546-1440(23) in press - “We forget that correlation or predictive power does not imply causation; the fact that two variables are correlated does not imply that one causes the other. In a Gallup poll several years ago, surveyors asked participants if correlation implied causation, and 64% of Americans answered “Yes.”. This occurs because humans learn from correlation, but we cannot observe causality. We have to understand that most people don’t know the difference.”
AI as a Public Service.
Lavista Ferres JM, Fishman EK, Rowe SP, Chu LC, Lugo-Fagundo E.
J Am Coll Radiol. 2023 Mar 30:S1546-1440(23) in press - “ are very good at cheating; if there is anything the model can use to cheat, it will learn it. An AI model was trained to distinguish between skin cancer and benign lesions and was thought to achieve dermatologist level performance. However, many of the positive cases had a ruler in the picture, while the negative cases did not. The model learned that if there was a ruler present in the image, there was a much higher chance of the patient having cancer.”
AI as a Public Service.
Lavista Ferres JM, Fishman EK, Rowe SP, Chu LC, Lugo-Fagundo E.
J Am Coll Radiol. 2023 Mar 30:S1546-1440(23) in press - Clinical Trials: Data Limitations
- Patient recruitment and enrollment: The process of traveling to trial sites can be burdensome and time consuming for trial participants, which can negatively affect enrollment.
- Patient monitoring, medical adherence, and retention: Frequent visits to trial sites may become invasive and unpleasant, leading participants to drop out of trials.
- Clinical-trial diversity: Companies often struggle to enroll diverse populations in clinical trials because trial sites may be inaccessible to underrepresented populations. - Clinical Trials: AI Advantages
- Tapping more participants and more diverse populations: AI-powered wearable devices can lessen the need for participants to travel to a physical site, which can enable organizations to recruit patients and diversify clinical trials participation.
- Boosting participant retention: Remote patient monitoring allows patients to participate in clinical trials with fewer potential hassles. AI algorithms can also be used to understand individual patient behaviors or needs, resulting in more patient-centric interactions and better retention.
- Producing faster trials at lower cost: Using AI, life sciences companies can reduce the cost and time required to process clinical-trial data through smart automation, improved efficiency, and less need for rework.
- Increasing reusable data: Organizations can use AI technologies to intelligently reuse existing data based on standards and metadata, reducing the need to start from scratch across trials. - AI in Healthcare: The Patient
- About six in 10 U.S. adults said they would feel uncomfortable if their provider used artificial intelligence tools to diagnose them and recommend treatments in a care setting, according to a survey from the Pew Research Center.
- Some 38% of respondents said using AI in healthcare settings would lead to better health outcomes while 33% said it would make them worse, and 27% said it wouldn’t make much of a difference, the survey found.
- Ultimately, men, younger people and those with higher education levels were the most open to their providers using AI.
- “Diagnostic AI refers to a broad range of applications that use learning strategies that mimic human approaches to learning. When clinicians understand the underlying mechanisms of diagnostic AI, they can become informed users of these tools,appreciating both their advantages and limitations.This Viewpoint outlines 3 learning methods that form the basis of many diagnostic AI systems—learning from experts, examples,and experience—and their parallels to clinicians’ existing approaches to learning.”
Decoding Artificial Intelligence to Achieve Diagnostic Excellence: Learning From Experts, Examples, and Experience.
Chen JH, Dhaliwal G, Yang D. JAMA. 2022 Aug 1. doi: 10.1001/jama.2022.13735.
Epub ahead of print. - “Advantages of AI approaches that learn by example include the ability to achieve expert-level performance on a wide variety of diagnostic tasks by learning from thousands of case examples more rapidly and consistently than humans have the capacity for. Such algorithms can estimate risks of future events (eg,myocardial infarction) using subtle diagnostic features (eg, retinal image patterns) that may be invisible or undiscovered by human experts.”
Decoding Artificial Intelligence to Achieve Diagnostic Excellence: Learning From Experts, Examples, and Experience.
Chen JH, Dhaliwal G, Yang D. JAMA. 2022 Aug 1. doi: 10.1001/jama.2022.13735.
Epub ahead of print. - “The disadvantage of such systems is that their diagnostic logic is often inscrutable to clinicians, which has introduced concerns about the “black box” nature of these diagnostic AI tools. Algorithms can now examine an electrocardiogram (ECG) and identify the signatures of previous atrial fibrillation episodes in a currently normal sinus rhythm tracing. If the algorithm cannot explain how it makes this diagnosis, will physicians be willing to act on it and prescribe anticoagulation? A broader limitation for many applications of supervised learning is the need for large amounts of manually labeled training data. This is not only resource intensive; it relies on the accuracy of human-designated labels, which may be biased or have poor interrater reliability(eg, diagnosing urinary tract infection vs asymptomatic bacteriuria).”
Decoding Artificial Intelligence to Achieve Diagnostic Excellence: Learning From Experts, Examples, and Experience.
Chen JH, Dhaliwal G, Yang D. JAMA. 2022 Aug 1. doi: 10.1001/jama.2022.13735.
Epub ahead of print. - “Game-playing AI algorithms that have defeated world champion human players in complex strategy games, including poker, chess, Go, and StarCraft, demonstrate the potential of reinforcement learning. These game-playing AI algorithms can learn from experience, playing millions of simulated games against themselves to experiment with different strategies. This allows the algorithms to explore and exploit unconventional choices against objective win conditions, offering the potential to discover novel moves that advance the state-of-the-art. Analogously, given a series of steps along the diagnostic pathway, a reinforcement learning approach could try different testing sequences to determine which lead to the most timely, accurate, and efficient diagnosis. For example, such an approach could simulate the thousands of different workups of a solitary pulmonary nodule, assessing the influence of different diagnostic moves both within and beyond the conventional sequencing of history, examination, laboratory testing, imaging, pathology, and genetic testing.”
Decoding Artificial Intelligence to Achieve Diagnostic Excellence: Learning From Experts, Examples, and Experience.
Chen JH, Dhaliwal G, Yang D. JAMA. 2022 Aug 1. doi: 10.1001/jama.2022.13735. Epub ahead of print. - “AI-trained systems cannot completely execute the diagnostic process, which involves human interactions, judgments, and social systems that are beyond what computers can model. AI systems will nonetheless become important tools—just like laboratory tests or imaging studies—in the increasingly complex quest to diagnose patients’ health problems. Diagnostic AI systems learn by mimicking experts, acquiring examples, or conducting experiments, much as clinicians do throughout their careers. As clinicians understand the ways in which diagnostic AI systems develop “intelligence,” they may recognize the strengths and limitations of their own learning practices and envision how to combine human and artificial intelligence to achieve diagnostic excellence better than either can alone.”
Decoding Artificial Intelligence to Achieve Diagnostic Excellence: Learning From Experts, Examples, and Experience.
Chen JH, Dhaliwal G, Yang D. JAMA. 2022 Aug 1. doi: 10.1001/jama.2022.13735. Epub ahead of print. - Key Points for Diagnostic Excellence
• Diagnostic artificial intelligence (AI) technologies represent a heterogeneous set of learning approaches that mimic human learning from experts, examples, and experience.
• Understanding this analogy can help clinicians demystify the “black box” of diagnostic AI and enable them to become
informed users of these systems.
• AI systems will eventually become important tools that augment the diagnostic process. Ideally, these tools will offload computational and data intensive work while enabling clinicians to focus on tasks they are uniquelywell-suited for, including history taking, communicating uncertainty, and understanding the patient’s context.
- "Several suggestions may be helpful for consideration by clinicians and decision makers who are designing and using AI tools. First, clinicians should not assume that traditional metrics, such as the area under the receiver operating characteristic curve, translate to clinical effects because such performance metrics are usually not optimized or evaluated for specific clinical contexts. Second, clinicians should be involved in guiding the design of metrics to ensure that the algorithms produce outputs that are clinically useful and patient-centered to minimize unintended harms.”
Rethinking Algorithm Performance Metrics for Artificial Intelligence in Diagnostic Medicine
Reyna MA et al.
JAMA Published online July 8, 2022 - "Third, clinicians should prioritize the use of AI tools with well documented and understandable explanations of performance metrics because doing so could enable informed decisions on whether and how best to use the algorithm. Fourth, clinicians should expect the prospective evaluation of algorithms in clinical settings. Evaluation in varied settings demonstrates the potential utility of an algorithm for actual clinical outcomes. Fifth, adopters of AI tools should require that Ai developers make available the full code for an algorithm, including the training data and code, so that the metrics used to develop the algorithms are explicit and modifiable.”
Rethinking Algorithm Performance Metrics for Artificial Intelligence in Diagnostic Medicine
Reyna MA et al.
JAMA Published online July 8, 2022 - "Clinicians and other health care decision makers have the responsibility to choose algorithms that are transparent, clinically useful, and effective across diverse patient populations. To facilitate an informed decision, algorithm development teams should also be diverse and work closely with clinicians to develop and implement AI performance metrics that incorporate clinical context. This process should also recognize and reflect the diversity of objectives and stakeholders in diagnostic medicine to improve the relevance and representation of AI tools in clinical practice.”
Rethinking Algorithm Performance Metrics for Artificial Intelligence in Diagnostic Medicine
Reyna MA et al.
JAMA Published online July 8, 2022
- “The gradual inclusion, or incursion as some see it, of artificial intelligence (AI) into various branches of medicine has ignited fierce debate in recent years, building bases of proponents and opponents in the process. We are firmly among those that believe AI will not replace practitioners and is most valuable as an adjunct in diagnostic radiology. We suggest a different approach to utilizing the technology, which may help even radiologists who may be averse to adopting AI. Here we consider early indications of AI’s potential as the ultimate quality assurance (QA) for radiologists..”
Is AI the Ultimate QA?
Edmund M. Weisberg, Linda C. Chu, Benjamin D. Nguyen, Pelu Tran, Elliot K. Fishman
J Digit Imaging. 2022 Feb 15. doi: 10.1007/s10278-022-00598-8. - “Fifteen studies (25%) addressed prediction of malignancy based on imaging findings, or differentiation of benign from malignant pancreatic conditions[280,301-315]. Marya et al[280] developed an EUS-based CNN model that distinguished AIP from normal pancreas with 99% sensitivity and 98% specificity, AIP from CP with 94% sensitivity and 71% specificity, and AIP from PDAC with 90% sensitivity and 93% specificity. Chu et al[301] conducted a study utilizing CT radiomics features to differentiate PDAC from normal pancreas tissue. The accuracy of the RF binary classification was 99.2%, with an AUC of 99.9%. All cases of PDAC were correctly identified, with a sensitivity of 100% and specificity of 98.5%.”
Artificial intelligence in gastroenterology: A state-of-the-art review
Kröner, Paul T et al.
World journal of gastroenterology vol. 27,40 (2021): 6794-6824. - “Eleven studies (19%) evaluated differentiation of PCNs by classifying them into their respective subtypes based on their characteristics on imaging. Springer et al developed a multimodality ML model that integrated clinical, radiological and genetic/biochemical markers data to determine whether patients with pancreas cyst should undergo surgery, monitoring, or no further surveillance. The model correctly identified serous cystic neoplasms in 65% of the cases with 99% specificity, clearly outperforming the current standard of care of clinical identification in only 18% of cases. The authors conclude that these systems may serve an adjunct role in clinical practice, enabling the clinician to take better-informed clinical decisions[.”
Artificial intelligence in gastroenterology: A state-of-the-art review
Kröner, Paul T et al.
World journal of gastroenterology vol. 27,40 (2021): 6794-6824. - The latest advances in AI in gastroenterology and hepatology are promising for aspect many fields of clinical care, from detection of neoplastic lesions on endoscopic assessment and improving current survival models to predicting treatment response. The application of AI to large and complex datasets may assist in the identification of new associations between variables, potentially leading to changes in clinical practice. Furthermore, the use of AI-assisted technologies has the potential to dramatically improve the quality of care. Finally, the time for assisted precision medicine is at hand, with the AI being able to tailor a treatment regimen or potentially predict the response to treatment in a specific patient based on extensive amounts of clinical data from large patient datasets. It is important to realize that, while AI currently does not substitute human clinical reasoning, it has a bright future in the betterment of patient care.
Artificial intelligence in gastroenterology: A state-of-the-art review
Kröner, Paul T et al.
World journal of gastroenterology vol. 27,40 (2021): 6794-6824 - Introduction: Applications of artificial intelligence (AI) in health care have increased in the past decade, but little is known about how patients view these applications and whether they have concerns. We conducted a nationally representative survey to understand public perceptions of the use of AI in diagnosis and treatment.
Results: Comfort with AI varied by clinical application. For example, 12.3% of respondents were very comfortable and 42.7% were somewhat comfortable with AI reading chest radiographs, but only 6.0% were very comfortable and 25.2%were somewhat comfortable about AI making cancer diagnoses. Most respondents were very concerned or somewhat concerned about AI’s unintended consequences, including misdiagnosis (91.5%), privacy breaches (70.8%), less time with clinicians(69.6%), and higher health care costs (68.4%). A higher proportion of respondents who self identified as being members of racial and ethnic minority groups indicated being very concerned about these issues, compared with White respondents.
Perspectives of Patients About Artificial Intelligence in Health Care
Dhruv Khullar et al.
JAMA Network Open. 2022;5(5):e2210309. - “Most respondents had positive views about AI’s ability to improve care but had concerns about its potential for misdiagnosis, privacy breaches, reducing time with clinicians, and increasing costs, with racial and ethnic minority groups expressing greater concern. Respondents were more comfortable with AI in specific clinical settings, and most wanted to know when AI was used in their care. One limitation of this study was it involved a panel that had agreed to participate in surveys, which may limit generalizability. In addition, compared with nonrespondents, respondents were younger, but no significant differences by sex or race and ethnicity were found. Clinicians, policy makers, and developers should be aware of patients’ views regarding AI. Patients may benefit from education on how AI is being incorporated into care and the extent to which clinicians rely on AI to assist with decision-making. Future work should examine how views evolve as patients become more familiar with AI.”
Perspectives of Patients About Artificial Intelligence in Health Care
Dhruv Khullar et al.
JAMA Network Open. 2022;5(5):e2210309. - “Most respondents had positive views about AI’s ability to improve care but had concerns about its potential for misdiagnosis, privacy breaches, reducing time with clinicians, and increasing costs, with racial and ethnic minority groups expressing greater concern. Respondents were more comfortable with AI in specific clinical settings, and most wanted to know when AI was used in their care.”
Perspectives of Patients About Artificial Intelligence in Health Care
Dhruv Khullar et al.
JAMA Network Open. 2022;5(5):e2210309. - AI and Radiology: Applications
- Your clinical practice
- NO shows for appointment
- Prediction of future health care costs - “Patients’ no-shows, scheduled but unattended medical appointments, have a direct negative impact on patients’ health, due to discontinuity of treatment and late presentation to care. They also lead to inefficient use of medical resources in hospitals and clinics. The ability to predict a likely no-show in advance could enable the design and implementation of interventions to reducethe risk of it happening, thus improving patients’ care and clinical resource allocation. In this study, we develop a new interpretable deep learning-based approach for predicting the risk of no-shows at the time when a medical appointment is first scheduled. The retrospective study was conducted in an academic pediatric teaching hospital with a 20% no-show rate. Our approach tackles several challenges in the design of a predictive model by (1) adopting a data imputation method for patients with missing information in their records (77% of the population), (2) exploiting local weather information to improve predictive accuracy, and developing an interpretable approach that explains how a prediction is made for each individual patient.”
Machine learning approaches to predicting no-shows in pediatric medical appointment
Dianbo Liu et al.
Digital Medicine (2022) 5:50 ; https://doi.org/10.1038/s41746-022-00594-w - “Our proposed neural network-based and logistic regression-based methods outperformed persistence baselines. In an unobserved set of patients, our method correctly identified 83% of no-shows at the time of scheduling and led to a false alert rate less than 17%. Our method is capable of producing meaningful predictions even when some information in a patient’s records is missing. We find that patients’ past no-show record is the strongest predictor. Finally, we discuss several potential interventions to reduce no-shows, such as scheduling appointments of high-risk patients at off-peak times, which can serve as starting point for further studies on no-show interventions.”
Machine learning approaches to predicting no-shows in pediatric medical appointment
Dianbo Liu et al.
Digital Medicine (2022) 5:50 ; https://doi.org/10.1038/s41746-022-00594-w
- “The use of artificial intelligence (AI) has grown dramatically in the past few years in the United States and worldwide, with more than 300 AI-enabled devices approved by the U.S. Food and Drug Administration (FDA). Most of these AI-enabled applications focus on helping radiologists with detection, triage, and prioritization of tasks by using data from a single point, but clinical practice often encompasses a dynamic scenario wherein physicians make decisions on the basis of longitudinal information. Unfortunately, benchmark data sets incorporating clinical and radiologic data from several points are scarce, and, therefore, the machine learning community has not focused on developing methods and architectures suitable for these tasks. Current AI algorithms are not suited to tackle key image interpretation tasks that require comparisons to previous examinations. Focusing on the curation of data sets and algorithm development that allow for comparisons at different points will be required to advance the range of relevant tasks covered by future AI-enabled FDA-cleared devices.”
The Need for Medical Artificial Intelligence That Incorporates Prior Images
Julián N. Acosta et al.
Radiology 2022; 000:1–6 • https://doi.org/10.1148/radiol.212830 - “Most of these AI-enabled applications focus on helping radiologists with detection, triage, and prioritization of tasks by using data from a single point, but clinical practice often encompasses a dynamic scenario wherein physicians make decisions on the basis of longitudinal information. Unfortunately, benchmark data sets incorporating clinical and radiologic data from several points are scarce, and, therefore, the machine learning community has not focused on developing methods and architectures suitable for these tasks. Current AI algorithms are not suited to tackle key image interpretation tasks that require comparisons to previous examinations.”
The Need for Medical Artificial Intelligence That Incorporates Prior Images
Julián N. Acosta et al.
Radiology 2022; 000:1–6 • https://doi.org/10.1148/radiol.212830 - Essentials
• Most artificial intelligence (AI)–enabled devices that are approved by the U.S. Food and Drug Administration (FDA) and are avail- able to date address tasks by considering only a single point.
• Clinical tasks involve dynamic scenarios, and diagnostic and prognostic decisions often rely on the combination of prior and current information.
• The development of benchmark data sets and algorithms that leverage prior examinations have the potential to improve the range of tasks covered by FDA-approved medical AI devices.
The Need for Medical Artificial Intelligence That Incorporates Prior Images
Julián N. Acosta et al.
Radiology 2022; 000:1–6 • https://doi.org/10.1148/radiol.212830 - “In summary, even though physicians routinely perform comparisons with prior examinations when interpreting images in clinical practice, only a few artificial intelligence (AI) algorithms currently available are able to incorporate information from more than one point to help in these critical tasks. The curation of high-quality data sets with longitudinal clinical and imaging data, and the development of AI algorithms capable of solving a wider range of problems, will be essential to provide meaningful improvements in clinical workflows.”
The Need for Medical Artificial Intelligence That Incorporates Prior Images
Julián N. Acosta et al.
Radiology 2022; 000:1–6 • https://doi.org/10.1148/radiol.212830 - “For early cancer detection, the percentage of confidence that comes with a diagnostic decision made by the algorithm might appear straightforward (for example, 80% confidence that lung cancer is present), but the process behind this number is very complex and understandably may not be apparent to the user. It is therefore not difficult to understand that there might be resistance to adoption of such strategies and the fear of overdiagnosis. It is important to understand that AI will not remove the need for physicians and experts to interpret the findings, provide a global picture of patient health, spot related diseases, and come up with a final diagnosis.”
The future of early cancer detection
Rebecca C. Fitzgerald et al.
Nature Medicine| VOL 28 | APRIL 2022 | 666–677 - "What the introduction of AI algorithms might do, providing that data management and safety regulations are in place, is reduce the cost and time needed to diagnose the disease. This will enable health practitioners to spend more time developing efficient and holistic treatment protocols, and will make state-of-art diagnostics more affordable. Furthermore, AI can be a training tool that provides immediate specialist feedback to generalists so that, in time, they may perform at an expert level.”
The future of early cancer detection
Rebecca C. Fitzgerald et al.
Nature Medicine| VOL 28 | APRIL 2022 | 666–677 
The future of early cancer detection
Rebecca C. Fitzgerald et al.
Nature Medicine| VOL 28 | APRIL 2022 | 666–677
The future of early cancer detection
Rebecca C. Fitzgerald et al.
Nature Medicine| VOL 28 | APRIL 2022 | 666–677
The future of early cancer detection
Rebecca C. Fitzgerald et al.
Nature Medicine| VOL 28 | APRIL 2022 | 666–677
- "We sought to measure the accuracy of triage decisions provided by MayaMD. We compared triage decisions made by the AI-based application with those made by healthcare providers for various medical conditions. We simulated patients presenting with various medical conditions by developing 50 unique clinical vignettes, similar to prior studies. The vignettes were selected from a list of the most common clinical presentations to the emergency room, urgent care, and primary care services. Vignettes included the patient’s age, sex, medical history, and presenting symptoms with a variable amount of pertinent positive and negative symptoms. Vignettes were not associated with a definite diagnosis or expected standard of care. Vignettes were not associated with an available physical examination.”
Artificial Intelligence-Based Application Provides Accurate Medical Triage Advice When Compared to Consensus Decisions of Healthcare Providers.
Delshad S et al.
Cureus 13(8): e16956. DOI 10.7759/cureus.16956 - "Our study is the first to our knowledge to compare AI triage decisions with those of a consensus of a group of providers from multiple specialties working at various practice locations, including both academic and non-academic settings. Our findings are significantly strengthened by this utilization of a consensus group of diverse, actively practicing clinicians. Our study is also the first to our knowledge that demonstrates the ability of AI to discriminate between ER and urgent care triage, as previous studies only assessed the AI- based application’s ability to determine urgent versus non-urgent triage. Appropriate triage discrimination between urgent and emergency care services has significant implications for cost savings and efficient healthcare delivery for health insurers and health systems.”
Artificial Intelligence-Based Application Provides Accurate Medical Triage Advice When Compared to Consensus Decisions of Healthcare Providers.
Delshad S et al.
Cureus 13(8): e16956. DOI 10.7759/cureus.16956 - “Our findings have significant implications. The utilization of AI-based applications that improve the appropriateness and safety of medical triage has the potential to improve patient outcomes and experience as well as the efficiency of healthcare delivery. Payors, providers, and patients may benefit from cost savings and higher-value care. AI-based applications may also be able to provide triage assistance in more rural or underserved areas where access to traditional triage nursing services may be limited. This study has limitations. We utilized clinical vignettes that were developed by a team of clinicians; therefore, limiting the real-world implications of our study. Real patients’ interaction with the AI-based application may not generate the same triage decision as for a clinical vignette simulating the same presentation of symptoms. Future research in which the AI application faces real patients is needed to ensure the accurate triage decisions seen in the present study translate to a real-world setting.”
Artificial Intelligence-Based Application Provides Accurate Medical Triage Advice When Compared to Consensus Decisions of Healthcare Providers.
Delshad S et al.
Cureus 13(8): e16956. DOI 10.7759/cureus.16956
- “It is time to move beyond studies showing that AI can detect opacities at CT or chest radiography—this is now well established. Instead, there is a great need for AI systems, based on a combination of imaging, laboratory, and clinical information, that provide actionable predictions otherwise unavailable or less accurate without AI.”
Artificial Intelligence of COVID-19 Imaging: A Hammer in Search of a Nail
Ronald M. Summers
Radiology 2020; 00:1–3 - “More observer performance experiments are necessary to determine whether AI improves clinical interpretation according to reader experience level and reading paradigm (first, concurrent, or second reader). Prospective outcome studies are necessary to determine whether the use of AI leads to changes in patient care, shortened hospitalizations, and reduced morbidity and mortality. Nonradiology clinical information will need to be routinely incorporated into AI models. Assessment of risk and progression of the chronic sequela of COVID-19 infection is necessary. A prospective randomized controlled trial would be exemplary.”
Artificial Intelligence of COVID-19 Imaging: A Hammer in Search of a Nail
Ronald M. Summers
Radiology 2020; 00:1–3 - “How does one put this deluge of articles into context? It seems unlikely that an AI system would detect many patients with COVID-19 who had a negative reverse transcription polymerase chain reaction test. Anecdotes will occur. But from a general perspective, this is unlikely to propel dissemination of the AI technology. What about distinguishing COVID-19 from other viral pneumonias? It seems unlikely that clinical decision mak- ing would depend on the recommendations of AI, given more definitive laboratory tests are available. Could AI lead to a fully automated interpretation? This has not been the focus of COVID-19 imaging AI to date. Multitask approaches that identify multiple abnormalities at chest imaging besides opacities will be needed, such as universal lesion detection. What about mortality prediction? Hazard ratios on the order of 2 to 3, as found in the article by Mushtaq et al, are generally insufficient for clinical decision making. While it is possible that prediction of an adverse outcome could lead to more aggressive treatment, it could also lead to unnecessary costs and adverse effects.”
Artificial Intelligence of COVID-19 Imaging: A Hammer in Search of a Nail
Ronald M. Summers
Radiology 2020; 00:1–3 - "What are the current needs of AI systems for COVID-19 and CT and chest radiography? Public challenges or competitions pitting different AI systems against one another would enable “apples-to-apples” comparisons of performance. More observer performance experiments are necessary to determine whether AI improves clinical interpretation according to reader experience level and reading paradigm (first, concurrent, or second reader). Prospective outcome studies are necessary to determine whether the use of AI leads to changes in patient care, shortened hospitalizations, and reduced morbidity and mortality. Nonradiology clinical information will need to be routinely incorporated into AI models. Assessment of risk and progression of the chronic sequela of COVID-19 infection is necessary.”
Artificial Intelligence of COVID-19 Imaging: A Hammer in Search of a Nail
Ronald M. Summers
Radiology 2020; 00:1–3
- “Why aren’t better data available? One of our patients, a veteran, once remarked in frustration after trying to obtain his prior medical records:, “Doc, why is it that we can see a specific car in a moving convoy on the other side of the world, but we can’t see my CT scan from the hospital across the street?” Sharing data in medicine is hard enough for a single patient, never mind the hundreds or thousands of cases needed to reliably train machine learning algorithms. Whether in treating patients or building AI tools, data in medicine are locked in little silos everywhere.”
Health Care AI Systems Are Biased
Amit Kaushal, Russ Altman, Curt Langlotz
Scientific American November 2021 - "Medical data sharing should be more commonplace. But the sanctity of medical data and the strength of relevant privacy laws provide strong incentives to protect data, and severe consequences for any error in data sharing. Data are sometimes sequestered for economic reasons; one study found hospitals that shared data were more likely to lose patients to local competitors. And even when the will to share data exists, lack of interoperability between medical records systems remains a formidable technical barrier. The backlash from big tech’s use of personal data over the past two decades has also cast a long shadow over medical data sharing. The public has become deeply skeptical of any attempt to aggregate personal data, even for a worthy purpose.”
Health Care AI Systems Are Biased
Amit Kaushal, Russ Altman, Curt Langlotz
Scientific American November 2021 - “AI has reached health care, and radiology in particular,and it is here to stay. Careful evaluation and adoption of AI-based tools will allow radiologists to pioneer the transition toward AI- enabled, patient-centric health care delivery. In collaboration, radiology researchers, health care providers, industry partners, and policymakers have the potential to realize the promise of AI to provide equal access to high-quality care, overcome the challenge of ongoing performance monitoring, and achieve the development of socially beneficial AI solutions.”
Artificial Intelligence in Radiology: The Computer’s Helping Hand Needs Guidance
Evis Sala, Stephan Ursprung
Radiology: Artificial Intelligence 2020; 2(6):e200207 - “Solutions that improve the quality of radiologic diagnosis without generating immediate financial benefit are less likely to permeate the mass market rapidly. Nonetheless, their potential to promote health equity through increasing diagnostic quality and consistency in non-subspecialized settings could profoundly improve the overall performance of a health care system. Therefore, noncommer- cial stakeholders should pursue research and publication of socially desirable AI solutions.”
Artificial Intelligence in Radiology: The Computer’s Helping Hand Needs Guidance
Evis Sala, Stephan Ursprung
Radiology: Artificial Intelligence 2020; 2(6):e200207 - AI and Measuring Its Value
1. How will the solution integrate into the current clinical pathway?
a) Rule-in/rule-out/triage of patients b) First reader/second reader c) Equality of use case and approved use
Artificial Intelligence in Radiology: The Computer’s Helping Hand Needs Guidance
Evis Sala, Stephan Ursprung
Radiology: Artificial Intelligence 2020; 2(6):e200207 - AI and Measuring Its Value
2. How well will the algorithm generalize?
a) Comparability of patient demographics (age, sex, ethnicity) between testing dataset and intended patient cohort
b) Comparability of the clinical setting (screening, diagnostic, investigative, or therapeutic setting) between testing dataset and intended use case
c) Evidence from monitoring programs and ongoing trials
Artificial Intelligence in Radiology: The Computer’s Helping Hand Needs Guidance
Evis Sala, Stephan Ursprung
Radiology: Artificial Intelligence 2020; 2(6):e200207 - AI and Measuring Its Value
3. Does the solution have a clinical and/or cost-benefit in real use?
a) Equivalent or improved diagnostic accuracy: sensitivity, specificity...
b) Overall cost-benefit: clinician time, reducing or increasing additional tests, overdiagnosis and overtreatment of indolent findings
Artificial Intelligence in Radiology: The Computer’s Helping Hand Needs Guidance
Evis Sala, Stephan Ursprung
Radiology: Artificial Intelligence 2020; 2(6):e200207 - Results: A total of 119 software offerings from 55 companies were identified. There were 46 algorithms that currently have Food and Drug Administration and/or Conformité Européenne approval (as of November 2019). Of the 119 offerings, distribution of software targets was 34 of 70 (49%), 21 of 70 (30%), 14 of 70 (20%), and one of 70 (1%) for diagnostic, quantitative, repetitive, and explorative tasks, respectively. A plurality of companies are focused on nodule detection at chest CT and two-dimensional mammography. There is very little activity in certain subspecialties, including pediatrics and nuclear medicine. A comprehensive table is available on the website hitilab.org/pages/ai-companies.
Conclusion: The radiology AI marketplace is rapidly maturing, with an increase in product offerings. Radiologists and practice administrators should educate themselves on current product offerings and important factors to consider before purchase and implementation.
The State of Radiology AI: Considerations for Purchase Decisions and Current Market Offerings
Yasasvi Tadavarthi et al.
Radiology: Artificial Intelligence 2020; 2(6):e200004 - “In many cases, AI radiology software represents uncharted ter- ritory. The vast majority of AI radiology companies are nascent and working on their initial clinical offerings. Because of this, many companies seek partnerships with institutions for clinical feedback or even codevelopment. Questions about ownership of data and intellectual property often arise during codevelopment, which can be difficult to address. An institution commonly provides data and medical expertise, whereas the industry partner provides technical expertise, engineering support, and productization strategies that may incorporate several years of prior effort and investment. Clear delineation of ownership of any resultant algorithm, software, or datasets is important to address in advance and, if applicable, the route for approval through the institution’s technology transfer or licensing office.”
The State of Radiology AI: Considerations for Purchase Decisions and Current Market Offerings
Yasasvi Tadavarthi et al.
Radiology: Artificial Intelligence 2020; 2(6):e200004 - "Unlike quality improvements, efficiency improvement is easier to demonstrate as an ROI. For example, practices can measure the difference in study read times with and without the AI software. Time savings can be directly translated into person- hours and subsequent cost savings. In the nascent stages of radiology AI software, saved person-hours may be the quickest and most efficacious method to demonstrate value. As an added benefit, AI software geared toward increasing efficiency for repetitive or mundane tasks could be used to attract radiologists to join the practice in a competitive market.”
The State of Radiology AI: Considerations for Purchase Decisions and Current Market Offerings
Yasasvi Tadavarthi et al.
Radiology: Artificial Intelligence 2020; 2(6):e200004 - "Serving the needs of radiologists requires maintaining a balance between providing additional useful information while minimizing false-positive findings, unnecessary clicks, and mouse mileage. Length of time required for educating users on using new software must also be considered, as many radiology AI software packages may function differently than what the clinician is accustomed to. Launching a separate application window, which is a frequently used implementation for many breast and prostate MRI workflows, is generally viewed as inconvenient.”
The State of Radiology AI: Considerations for Purchase Decisions and Current Market Offerings
Yasasvi Tadavarthi et al.
Radiology: Artificial Intelligence 2020; 2(6):e200004 - "There are relatively few applications of AI in the abdomen and pelvis. Detection of free air, hem- orrhage, and aortic dissection and/or aneurysm are among the few in CT analysis. Other appli- cations include detection and characterization of hepatic tumors at MRI and automated process- ing and detection of lesions at CT colonography. Currently, none are FDA or CE approved.”
The State of Radiology AI: Considerations for Purchase Decisions and Current Market Offerings
Yasasvi Tadavarthi et al.
Radiology: Artificial Intelligence 2020; 2(6):e200004 - “The promise and potential benefits of radiology AI software continue to grow, and radiologists, practice administrators, and IT staff must continue to educate themselves on the potential ben- efits, drawbacks, and costs of implementation. We encourage the reader to consider using the guidelines in the Table when evalu- ating companies to ensure all aspects of a purchasing decision are considered. Deployed correctly, these software can be a boon to both patients and providers in an ever-evolving health care setting with increasing imaging volumes and complexity.”
The State of Radiology AI: Considerations for Purchase Decisions and Current Market Offerings
Yasasvi Tadavarthi et al.
Radiology: Artificial Intelligence 2020; 2(6):e200004 
The State of Radiology AI: Considerations for Purchase Decisions and Current Market Offerings
Yasasvi Tadavarthi et al.
Radiology: Artificial Intelligence 2020; 2(6):e200004- OBJECTIVE To assess the performance of artificial intelligence(AI) algorithms in realistic radiology workflows by performing an objective comparative evaluation of the preliminary reads of anteroposterior (AP) frontal chest radiographs performed by an AI algorithm and radiology residents.
CONCLUSIONS AND RELEVANCE These findings suggest that it is possible to build AI algorithms that reach and exceed the mean level of performance of third-year radiology residents for full- fledged preliminary read of AP frontal chest radiographs. This diagnostic study also found that while the more complex findings would still benefit from expert overreads, the performance of AI algorithms was associated with the amount of data available for training rather than the level of difficulty of interpretation of the finding. Integrating such AI systems in radiology workflows for preliminary interpretations has the potential to expedite existing radiology workflows and address resource scarcity while improving overall accuracy and reducing the cost of care.
Comparison of Chest Radiograph Interpretations by Artificial Intelligence Algorithm vs Radiology Residents
Joy T. Wu et al.
JAMA Network Open. 2020;3(10):e2022779. doi:10.1001/jamanetworkopen.2020.22779 - Question: How does an artificial intelligence (AI) algorithm compare with radiology residents in full-fledged preliminary reads of anteroposterior (AP) frontal chest radiographs?
Findings: This diagnostic study was conducted among 5 third-year radiology residents and an AI algorithm using a study data set of 1998 AP frontal chest radiographs assembled through a triple consensus with adjudication ground truth process covering more than 72 chest radiograph findings. There was no statistically significant difference in sensitivity between the AI algorithm and the radiology residents, but the specificity and positive predictive value were statistically higher for AI algorithm.
Meaning: These findings suggest that well-trained AI algorithms can reach performance levels similar to radiology residents in covering the breadth of findings in AP frontal chest radiographs, which suggests there is the potential for the use of AI algorithms for preliminary interpretations of chest radiographs in radiology workflows to expedite radiology reads, address resource scarcity, improve overall accuracy, and reduce the cost of care.
Comparison of Chest Radiograph Interpretations by Artificial Intelligence Algorithm vs Radiology Residents
Joy T. Wu et al.
JAMA Network Open. 2020;3(10):e2022779. doi:10.1001/jamanetworkopen.2020.22779 - “Overall, this study points to the potential use AI systems in future radiology workflows for preliminary interpretations that target the most prevalent findings, leaving the final reads performed by the attending physician to still catch any potential misses from the less-prevalent fine-grained findings. Having attending physicians quickly correct the automatically produced reads, we can expect to significantly expedite current dictation-driven radiology workflows, improve accuracy, and ultimately reduce the overall cost of care.”
Comparison of Chest Radiograph Interpretations by Artificial Intelligence Algorithm vs Radiology Residents
Joy T. Wu et al.
JAMA Network Open. 2020;3(10):e2022779. doi:10.1001/jamanetworkopen.2020.22779
- IMPORTANCE Machine learning (ML) algorithms can identify patients with cancer at risk of short-term mortality to inform treatment and advance care planning. However, no ML mortality risk prediction algorithm has been prospectively validated in oncology or compared with routinely used prognostic indices.
OBJECTIVE To validate an electronic health record–embedded ML algorithm that generated real-time predictions of 180-day mortality risk in a general oncology cohort
CONCLUSIONS AND RELEVANCE In this prognostic study, an ML algorithm was feasibly integrated into the electronic health record to generate real-time, accurate predictions of short-term mortality for patients with cancer and outperformed routinely used prognostic indices. This algorithm may be used to inform behavioral interventions and prompt earlier conversations about goals of care and end-of-life preferences among patients with cancer.
Validation of a Machine Learning Algorithm to Predict 180-Day Mortality for Outpatients With Cancer
Christopher R. Manz et al.
JAMA Oncol. doi:10.1001/jamaoncol.2020.4331 Published online September 24, 2020. - OBJECTIVE To validate an electronic health record–embedded ML algorithm that generated real-time predictions of 180-day mortality risk in a general oncology cohort
CONCLUSIONS AND RELEVANCE In this prognostic study, an ML algorithm was feasibly integrated into the electronic health record to generate real-time, accurate predictions of short-term mortality for patients with cancer and outperformed routinely used prognostic indices. This algorithm may be used to inform behavioral interventions and prompt earlier conversations about goals of care and end-of-life preferences among patients with cancer.
Validation of a Machine Learning Algorithm to Predict 180-Day Mortality for Outpatients With Cancer
Christopher R. Manz et al.
JAMA Oncol. doi:10.1001/jamaoncol.2020.4331 online September 24, 2020. - OBJECTIVE To validate an electronic health record–embedded ML algorithm that generated real-time predictions of 180-day mortality risk in a general oncology cohort.
DESIGN, SETTING, AND PARTICIPANTS This prognostic study comprised a prospective cohort of patients with outpatient oncology encounters between March 1, 2019, and April 30, 2019. An ML algorithm, trained on retrospective data from a subset of practices, predicted 180-day mortality risk between 4 and 8 days before a patient’s encounter. Patient encounters took place in 18 medical or gynecologic oncology practices, including 1 tertiary practice and 17 general oncology practices, within a large US academic health care system. Patients aged 18 years or older with outpatient oncology or hematology and oncology encounters were included in the analysis. Patients were excluded if their appointment was scheduled after weekly predictions were generated and if they were only evaluated in benign hematology, palliative care, or rehabilitation practices.
Validation of a Machine Learning Algorithm to Predict 180-Day Mortality for Outpatients With Cancer
Christopher R. Manz et al.
JAMA Oncol. doi:10.1001/jamaoncol.2020.4331 online September 24, 2020. - CONCLUSIONS AND RELEVANCE In this prognostic study, an ML algorithm was feasibly integrated into the electronic health record to generate real-time, accurate predictions of short-term mortality for patients with cancer and outperformed routinely used prognostic indices. This algorithm may be used to inform behavioral interventions and prompt earlier conversations about goals of care and end-of-life preferences among patients with cancer.
Validation of a Machine Learning Algorithm to Predict 180-Day Mortality for Outpatients With Cancer
Christopher R. Manz et al.
JAMA Oncol. doi:10.1001/jamaoncol.2020.4331 online September 24, 2020. - Key Points
Question: Can a machine learning algorithm prospectively identify patients with cancer at risk of 180-day mortality?
Findings: In this prognostic cohort study of 24582 patients seen in oncology practices within a large health care system, a machine learning algorithm integrated into the electronic health record accurately identified the risk of 180-day mortality with good discrimination and positive predictive value of 45.2%. When added to performance status– and comorbidity-based classifiers, the algorithm favorably reclassified patients.
Meaning: An integrated machine learning algorithm demonstrated good prospective performance compared with traditional prognostic classifiers and may inform clinician and patient decision-making in oncology.
Validation of a Machine Learning Algorithm to Predict 180-Day Mortality for Outpatients With Cancer
Christopher R. Manz et al.
JAMA Oncol. doi:10.1001/jamaoncol.2020.4331 online September 24, 2020. 
- “In summary, research on AI-powered technologies in the medical domain was at early stage in the 1970s. However, associated deep learning algorithms significantly attracted and revolutionized the scientific community with subsequent evolution of research and exponential growth of multidisciplinary publications since that time. Work in this field has impacted radiology as an area of predominant interest and has been led by institutions in the United States, Spain, France, China, and England. The bibliometric study reported herein can provide a broad overview and valuable guidance to help medical researchers gain insights into key points and trace the global trends regarding the status of AI research in medicine, particularly in radiology and other relevant multispecialty areas.”
Artificial Intelligence in Various Medical Fields With Emphasis on Radiology: Statistical Evaluation of the Literature
Emre Pakdemirli, Urszula Wegner
Cureus 12(10): e10961. DOI 10.7759/cureus.10961 - "Academic literature on AI in medicine had little room for optimism through the late 1970s. However, AI- driven software soon influenced and inspired qualified staff across the globe with subsequent increase of numerous associated publications in the 1990s. Importantly, this positive research trend demonstrates continuous dramatic growth pattern over the recent years.”
Artificial Intelligence in Various Medical Fields With Emphasis on Radiology: Statistical Evaluation of the Literature
Emre Pakdemirli, Urszula Wegner
Cureus 12(10): e10961. DOI 10.7759/cureus.10961 
Artificial Intelligence in Various Medical Fields With Emphasis on Radiology: Statistical Evaluation of the Literature
Emre Pakdemirli, Urszula Wegner
Cureus 12(10): e10961. DOI 10.7759/cureus.10961
- “The decision of what imaging test is most appropriate in each situation is influenced by many factors, some of which are highly subjective. The issue of over- and under-utilization of imaging resources is something that every clinician and radiologist struggles with. The desire to not miss acute pathology is balanced with the potential detriment of excessive radiation dose to susceptible populations. There likely exists a combination of historical and objective factors which can predict outcomes with sufficient sensitivity and specificity to guide the ordering pattern of most physicians.”
Applications of artificial intelligence in the emergency department
Supratik K. Moulik, Nina Kotter, Elliot K. Fishman
Emergency Radiology (2020) 27:355–358 - "In the future, predictive models will be trained on clinical, treatment, laboratory, and genetic data of individuals to facilitate personalized treatments. Machine learning systems are uniquely equipped for finding groups and subgroups that require more aggressive management. The goal of incorporating viral and host genetic data will require significant advances in computing and genetic sequencing.”
Applications of artificial intelligence in the emergency department
Supratik K. Moulik, Nina Kotter, Elliot K. Fishman
Emergency Radiology (2020) 27:355–358 - "It is important to keep the evolution of the AI/ML technology in context so as not to become overly enthusiastic about the current capabilities and simultaneously not to become overly pessimistic about future developments. Though the promised delivery date of fully self-driving cars has continuously been pushed back for the past decade, it is undeniable that drivers in semiautonomous vehicles are safer than unassisted drivers. Similarly, there are tangible patient care and cost benefits to be obtained through staged development of AL/ML systems even if fully autonomous MD systems are not on the horizon.”
Applications of artificial intelligence in the emergency department
Supratik K. Moulik, Nina Kotter, Elliot K. Fishman
Emergency Radiology (2020) 27:355–358 - “The FDA approval process to date has focused on applications (apps) that affect patient triage and not necessarily apps that have the computer serve as the only or final reader. We have chosen a select group of apps to provide the reader with a sense of the current state of AI use in the ER setting. Because adoption of new technology and FDA approval are always works in progress, it is not our intention here to be comprehensive. For a more thorough review of approved AI applications, please see the American College of Radiology record available here (https:// www.acrdsi.org/DSI-Services/FDA-Cleared-AI-Algorithms).”
The first use of artificial intelligence (AI) in the ER: triage not diagnosis
Edmund M. Weisberg, Linda C. Chu, Elliot K. Fishman
Emergency Radiology (2020) 27:361–366 - “Digital technology, including its omnipresent connectedness and its powerful artificial intelligence, is the most recent long wave of humanity’s socioeconomic evolution. The first technological revolutions go all the way back to the Stone, Bronze, and Iron Ages, when the transformation of material was the driving force in the Schumpeterian process of creative destruction. A second metaparadigm of societal modernization was dedicated to the transformation of energy (aka the “industrial revolutions”), including water, steam, electric, and combustion power. The current metaparadigm focuses on the transformation of information. Less than 1% of the world's technologically stored information was in digital format in the late 1980s, surpassing more than 99% by 2012. Every 2.5 to 3 years, humanity is able to store more information than since the beginning of civilization. The current age focuses on algorithms that automate the conversion of data into actionable knowledge.”
Digital technology and social change: the digital transformation of society from a historical perspective
Martin Hilbert
DIALOGUES IN CLINICAL NEUROSCIENCE • Vol 22 • No. 2 • 2020 - “The current metaparadigm focuses on the transformation of information. Less than 1% of the world's technologically stored information was in digital format in the late 1980s, surpassing more than 99% by 2012. Every 2.5 to 3 years, humanity is able to store more information than since the beginning of civilization. The current age focuses on algorithms that automate the conversion of data into actionable knowledge.”
Digital technology and social change: the digital transformation of society from a historical perspective
Martin Hilbert
DIALOGUES IN CLINICAL NEUROSCIENCE • Vol 22 • No. 2 • 2020 - “Each technological revolution, originally received as a bright new set of opportunities, is soon recognized as a threat to the established way of doing things in firms, institutions, and society at large. The new techno-economic paradigm gradually takes shape as a different “common sense” for effective action in any area of endeavor. But while competitive forces, profit seeking, and survival pressures help diffuse the changes in the economy, the wider social and institutional spheres — where change is also needed — are held back by strong inertia stemming from routine, ideology, and vested interests. It is this difference in rhythm of change, between the techno-economic and the socio-institutional spheres, that would explain the turbulent period.”
Digital technology and social change: the digital transformation of society from a historical perspective
Martin Hilbert
DIALOGUES IN CLINICAL NEUROSCIENCE • Vol 22 • No. 2 • 2020 - “The first focused on the transformation of material, including stone, bronze, and iron. The second, often referred to as industrial revolutions, was dedicated to the transformation of energy, including water, steam, electric, and combustion power. Finally, the most recent metaparadigm aims at transforming information. It started out with the proliferation of communication and stored data and has now entered the age of algorithms, which aims at creating automated processes to convert the existing information into actionable knowledge.”
Digital technology and social change: the digital transformation of society from a historical perspective
Martin Hilbert
DIALOGUES IN CLINICAL NEUROSCIENCE • Vol 22 • No. 2 • 2020
- Use of AI in Radiology beyond Reading Scans
- IMAGE PRODUCTION AND QUALITY CONTROL
- IMPROVING RADIOLOGY WORKFLOW
- BUSINESS APPLICATIONS
- BILLING AND COLLECTIONS
- RESEARCH APPLICATIONS - “The ultimate goal of AI in medical imaging is to improve patient outcomes. In this review, we have summarized some of the many ways in which noninterpretive AI is relevant to radiologists and their patients. At this time, only a few of these techniques are ready to translate into clinical practice. Regardless of which of these techniques are ultimately adopted, we hope that this review will provoke thought in the wider community of academic radiologists, and to help lead us to even newer and more intriguing applications.”
Noninterpretive Uses of Artificial Intelligence in Radiology
Michael L. Richardson et al.
Acad Radiol 2020 (in press) - “Deep learning approaches can also assist radiologists by assigning higher priorities to cases on the worklist that may contain emergent abnormalities. Such prioritization has been proposed in the setting of triage or screening systems to detect abnormalities on chest radiographs, abdominal CT, or head CT. In these paradigms, there is an image interpretation component to the AI’s tasks, but the role of the AI is not to primarily render an interpretation but to alert radiologists to potential critical findings and improve turnaround time for reporting of potentially actionable abnormalities.”
Noninterpretive Uses of Artificial Intelligence in Radiology
Michael L. Richardson et al.
Acad Radiol 2020 (in press) - "AI models have been used to successfully localize and annotate organs such as the kidney, segmental anatomy such as lobes of the liver or lung, and automated detection and labeling of vertebral bodies. This is extremely useful when volumetric assessment of a lesion or organ is needed. Examples include automated estimate of renal volume in a potential donor, liver volumes in patients with potential seg- mental or lobar resection and volumetric assessment in tumor treatment response.”
Noninterpretive Uses of Artificial Intelligence in Radiology
Michael L. Richardson et al.
Acad Radiol 2020 (in press) - “In radiology, image-based search engines can provide valuable opportunities for education as well as research. Large volumes of medical imaging are accumulating in shared and public databases, and image-based search engines connected to these databases may allow easy discovery and comparison of visually similar cases. As opposed to text searches, which are likely to find cases with similar diagnoses, image searches may also find visually similar cases with different diagnoses. Correlation of visual and textual features of images found using image-based search engines may also provide interesting research opportunities.”
Noninterpretive Uses of Artificial Intelligence in Radiology
Michael L. Richardson et al.
Acad Radiol 2020 (in press) - “With rapidly advancing progress in the development of algorithms for detecting and classifying imaging findings, more attention has turned towards limitations of these algorithms and particularly to vulnerabilities in these algorithms. To date, adversarial algorithms have been developed that can systematically deceive a trained AI model or a human radiologist. Notable examples include one algorithm that tricked an AI model into misclassifying pneumothorax on chest radio- graphs and another that misled human radiologists by adding fake pulmonary nodules and removing real pulmonary nodules from chest CT exams.”
Noninterpretive Uses of Artificial Intelligence in Radiology
Michael L. Richardson et al.
Acad Radiol 2020 (in press) - “AI tools might also help in tracking radiology resident performance and evaluate competency. AI tools are being devel- oped and implemented across medical specialties to evaluate physician competence, and radiology training should be particularly amenable given the highly digitized nature of radiology practice. Metrics used for evaluation of resident competency, such as the ACGME/American Board of Radiology milestone project, could incorporate AI-based assessments in the future.”
Noninterpretive Uses of Artificial Intelligence in Radiology
Michael L. Richardson et al.
Acad Radiol 2020 (in press) - “The ultimate goal of AI in medical imaging is to improve patient outcomes. In this review, we have summarized some of the many ways in which noninterpretive AI is relevant to radiologists and their patients. At this time, only a few of these techniques are ready to translate into clinical practice. Regardless of which of these techniques are ultimately adopted, we hope that this review will provoke thought in the wider community of academic radiologists, and to help lead us to even newer and more intriguing applications.”
Noninterpretive Uses of Artificial Intelligence in Radiology
Michael L. Richardson et al.
Acad Radiol 2020 (in press) 
- “With machine learning, the input is based on hand-engineered features, while unsupervised deep learning is able to learn these features itself directly from data. Multiple research groups are working on applying AI to improve the reconstruction of CT images. One application is image-space-based reconstructions in which convolutional neural networks are trained with low-dose CT images to recon- struct routine-dose CT images. Another approach is to optimize IR algorithms.”
The evolution of image reconstruction for CT—from filtered back projection to artificial intelligence
Willemink MJ, Noël PB
European Radiology (2019) 29:2185–2195 - “Generally, IR algorithms are based on manually designed prior functions resulting in low-noise images without loss of structures. Deep learning methods allow for implementing more complex functions, which have the potential to enable lower-dose CT and sparse-sampling CT. These AI techniques have the potential to reduce CT radiation doses while speeding up reconstruction times. Also, deep learning can be used to optimize image quality without reducing the radiation dose, e.g., by more advanced DECT monochromatic image reconstruction and metal artifact reduction.”
The evolution of image reconstruction for CT—from filtered back projection to artificial intelligence
Willemink MJ, Noël PB
European Radiology (2019) 29:2185–2195 - “These methods are not yet ready for clinical implementation; however, it is expected that AI will play, in the near future, a major role in CT image reconstruction and restoration. We expect that AI will fit in current clinical CT imaging workflow by enhancing current reconstruction methods, for example by significantly accelerating the reconstruction process since application of a trained network can be instantaneously.”
The evolution of image reconstruction for CT—from filtered back projection to artificial intelligence
Willemink MJ, Noël PB
European Radiology (2019) 29:2185–2195 - “In this context, artificial-intelligence tools have been designed to support radiologists in the identification of lung nodules since when chest radiography was the diagnostic imaging modality of choice to detect lung cancer. With the advent of low-dose CT and in particular with its implementation in screening trials, many computer-aided detection (CAD) systems for lung nodule identification have been developed . The CAD potential in improving radiologists’ performance has been deeply investigated, highlighting that the CAD can successfully be used as a second reader.”
The potential contribution of artificial intelligence to dose reduction in diagnostic imaging of lung cancer.
Retico A, Fantacci M
Journal of Medical Artificial Intelligence, North America, 2, mar. 2019 - “The research in lung cancer diagnosis is now advancing in two distinct fields: the improvement in the image acquisition instrumentation and reconstruction techniques based on iterative processes is allowing to obtain high-quality CT images even at low and ultra-low dose (i.e., a dose amount very similar to that of a chest radiography), whereas the recent acceleration in the implementation of deep-learning methods in the medical imaging field is leading to an enhancement of the performance of CAD systems across different imaging modalities, in both detection and diagnosis tasks.”
The potential contribution of artificial intelligence to dose reduction in diagnostic imaging of lung cancer.
Retico A, Fantacci M
Journal of Medical Artificial Intelligence, North America, 2, mar. 2019 - “As AI continues to evolve,health care as we know it will dramatically change. Radiologists have always served at the forefront in adapting new technologies in medicine, and it should be no different with the advent of the AI revolution. I will not replace radiologists; instead those radiologists who take advantage of AI may ultimately replace those who refuse to accept it .It is crucial we build an ecosystem of key players in technology, research, radiology, and the regulatory bodies who will work together to effectively and safely integrate AI into clinical practice. As a of this technology will expand our efficiency and decision making capabilities, leading to earlier and better detection of disease and improve outcomes for our patients.”
Artificial intelligence in radiology: the ecosystem essential to improving patient care
Sogani J, Allen B Jr, K Dreyer, McGintgy GY
Clinical Imaging (in press)
- Competency, Motive, Transparency
- Competency reflects both the extent to which physicians are perceived to have clinical mastery and patients’ knowledge and self-efficacy of their own health. Because much of AI is and will be used to augment the abilities of physicians, there is potential to increase physician competency and enable patient-physician trust. This includes not only AI-assisted clinical decision support (eg, by suggesting possible diagnoses to consider) but also the use of AI for physician training and quality improvement (eg, by providing automated feedback to physicians about their diagnostic performance). AI can also serve an important role in empowering patients to better understand their health and self-manage their conditions.
Promoting Trust Between Patients and Physicians in the Era of Artificial Intelligence.
Nundy S, Montgomery T, Wachter RM.
JAMA. Published online July 15, 2019. doi:10.1001/jama.2018.20563 - “ On the other hand, trust will be compromised by AI that is inaccurate, biased, or reflective of poor-quality practices as well as AI that lacks explainability and inappropriately conflicts with physician judgment and patient autonomy.”
Promoting Trust Between Patients and Physicians in the Era of Artificial Intelligence.
Nundy S, Montgomery T, Wachter RM.
JAMA. Published online July 15, 2019. doi:10.1001/jama.2018.20563 - “Motive refers to a patient’s trust that the physician is acting solely in the interests of the patient. Patients are likely to perceive motive through the lens of the extent of the open dialogue they have with their physicians. Through greater automation of low-value tasks, such as clinical documentation, it is possible that AI will free up physicians to identify patients’ goals, barriers, and beliefs, and counsel them about their decisions and choices, thereby increasing trust. Conversely, AI could automate more of the physician’s workflow, but then fill freed-up time with more patients with clinical issues that are more cognitively or emotionally complex. AI could also enable greater distribution of care across a care team (both human agents and computer agents).”
Promoting Trust Between Patients and Physicians in the Era of Artificial Intelligence.
Nundy S, Montgomery T, Wachter RM.
JAMA. Published online July 15, 2019. doi:10.1001/jama.2018.20563 - “Whether this would enhance or harm trust would depend on the degree of collaboration among team members and the information flow, and could compromise trust if robust, longitudinal relationships were impeded.”
Promoting Trust Between Patients and Physicians in the Era of Artificial Intelligence.
Nundy S, Montgomery T, Wachter RM.
JAMA. Published online July 15, 2019. doi:10.1001/jama.2018.20563 - “Well-designed AI that allows patients to appreciate and understand that clinical decisions are based on evidence and expert consensus should enhance trust. It can also process patient data (including health care and consumer data) to provide physicians’ insight on patients’ behaviors and preferences.”
Promoting Trust Between Patients and Physicians in the Era of Artificial Intelligence.
Nundy S, Montgomery T, Wachter RM.
JAMA. Published online July 15, 2019. doi:10.1001/jama.2018.20563 - “ Moreover, if patient data are routinely shared with external entities for AI development, patients may become less transparent about divulging their information to physicians, and physicians may be more reluctant to acknowledge their own uncertainties. AI that does not explain the source or nature of its recommendations (“black box”) may also erode trust.”
Promoting Trust Between Patients and Physicians in the Era of Artificial Intelligence.
Nundy S, Montgomery T, Wachter RM.
JAMA. Published online July 15, 2019. doi:10.1001/jama.2018.20563 - “Where health care AI is implemented by health systems, it should be directed toward automating the transactional, business, and documentation aspects of care; doing so may provide time to physicians to engage with their patients more deeply. If AI is effective in relieving physicians from the burdens of data entry and other clerical tasks, much of the reclaimed time should be made available for patient care, shared decision-making, and counseling, which are the cornerstones of effective health care that are often compromised today.”
Promoting Trust Between Patients and Physicians in the Era of Artificial Intelligence.
Nundy S, Montgomery T, Wachter RM.
JAMA. Published online July 15, 2019. doi:10.1001/jama.2018.20563 - “When health care AI is developed by health systems and third-party organizations using patient data, physicians should be mindful of the effect on patient-physician trust. It will be important to develop ethical approaches that allow for patient input into decisions by health systems to share data for the purposes of developing AI through some combination of individual patient consent and the involvement of patient advocacy groups.”
Promoting Trust Between Patients and Physicians in the Era of Artificial Intelligence.
Nundy S, Montgomery T, Wachter RM.
JAMA. Published online July 15, 2019. doi:10.1001/jama.2018.20563
- “While the health system believes acquisition and use of this data are in the best interest of its patients (after all, office visits are short, and this knowledge can help guide its physicians as to a patient’s greatest risks), many patients might perceive this as an invasion of privacy and worry that the data might paint an incomplete picture of their lives and lead to unnecessary or inaccurate medical recommendations.”
Leading your organization to responsible AI
Roger Burkhardt, Nicolas Hohn, and Chris Wigley
McKinsey & Company (May 2019) - As a result, leaders must ask data-science teams fairly granular questions to understand how they sampled the data to train their models. Do data sets reflect real-world populations? Have they included data that are relevant to minority groups? Will performance tests during model development and use uncover issues with the data set? What could we be missing?
Leading your organization to responsible AI
Roger Burkhardt, Nicolas Hohn, and Chris Wigley
McKinsey & Company (May 2019) - Additionally, leaders must encourage their organization to move from a compliance mind-set to a co-creation mind-set in which they share their company’s market and technical acumen in the development of new regulations. Recent work in the United Kingdom between the Financial Conduct Authority (FCA), the country’s banking regulator, and the banking industry offers a model for this new partnership approach. The FCA and banking industry have teamed in creating a “regulatory sandbox” where banks can experiment with AI approaches that challenge or lie outside of current regulatory norms, such as using new data to improve fraud detection or better predict a customer’s propensity to purchase products.
Leading your organization to responsible AI
Roger Burkhardt, Nicolas Hohn, and Chris Wigley
McKinsey & Company (May 2019) - “Today, some 80% of large companies have adopted machine learning and other forms of artificial intelligence (AI) in their core business. Five years ago, the figure was less than 10%. Nevertheless, the majority of companies still use AI tools as point solutions — discrete applications, isolated from the wider enterprise IT architecture. That’s what we found in a recent analysis of AI practices at more than 8,300 large, global companies in what we believe is one of the largest-scale studies of enterprise IT systems to date.”
Taking a Systems Approach to Adopting AI
by Bhaskar Ghosh, Paul R. Daugherty, H. James Wilson, and Adam Burden
Harvard Business Review - “Edge computing is also breaking boundaries by moving much of the processing out to the edge of networks, where they meet with the physical world, as with smartphones, robots, drones, security cameras, and IoT. For instance, blockchain company Filament is using data-efficient AI, blockchain, and the Internet of Things (IoT) to enable secure and autonomous edge-computing transactions through a decentralized network stack — independent of underlying infrastructure.”
Taking a Systems Approach to Adopting AI
by Bhaskar Ghosh, Paul R. Daugherty, H. James Wilson, and Adam Burden
Harvard Business Review - ”Artificial intelligence is a vital part of adaptable systems. Whether it’s virtual agents, natural language processing, machine learning, advanced analytics, or other forms of AI, companies have a host of opportunities to transform the way they do business once their architectures make AI an integral part of the transaction flow. By finding a responsible, transparent balance between human and machine intelligence, and combining it with more basic forms of robotic process automation, adaptable systems can create value in ways that were previously impossible.”
Taking a Systems Approach to Adopting AI
by Bhaskar Ghosh, Paul R. Daugherty, H. James Wilson, and Adam Burden
Harvard Business Review - As systems evolve, so must the IT workforce. Companies will need multidisciplinary talent that can bridge infrastructure, development tools, programming languages, AI, and machine learning. They’ll also need to combine human talent with a growing army of smart machines to create entirely new kinds of hybrid IT roles. And they’ll need to develop new ways to continuously evolve their workforce, using ongoing learning and organizational transformation to adapt to the relentless pace of systemic AI advances.
Taking a Systems Approach to Adopting AI
by Bhaskar Ghosh, Paul R. Daugherty, H. James Wilson, and Adam Burden
Harvard Business Review
- “In healthcare, faxes remain the most common method that practitioners use to communicate with each other, and therefore often contain important clinical information: lab results, specialist consult notes, prescriptions and so on. Because most healthcare fax numbers are public, doctors also receive scores of pizza menus, travel specials, and other “junk faxes.” Faxes don’t contain any structured text — so it takes medical practice staff an average of two minutes and 36 seconds to review each document and input relevant data into patient records. Through a combination of machine learning and business-process outsourcing that has automated the categorizing of faxes, we’ve reduced time-per-fax for our practices to one minute and 11 seconds. As a result, last year alone we managed to eliminate over 3 million hours of work from the healthcare system.”
- “In healthcare, faxes remain the most common method that practitioners use to communicate with each other, and therefore often contain important clinical information: lab results, specialist consult notes, prescriptions and so on. Because most healthcare fax numbers are public, doctors also receive scores of pizza menus, travel specials, and other “junk faxes.” Faxes don’t contain any structured text — so it takes medical practice staff an average of two minutes and 36 seconds to review each document and input relevant data into patient records.”
How AI Is Taking the Scut Work Out of Health Care Jonathan Bush Harvard Business Review (March 2018) - “Here’s just one example of the immediate opportunity: Each year, some 120 million faxes still flow into the practices of the more than 100,000 providers on the network of athenahealth, the healthcare technology company where I’m CEO. That’s right: faxes. Remember those?.”
How AI Is Taking the Scut Work Out of Health Care Jonathan Bush Harvard Business Review (March 2018) - “We have a similar opportunity in medicine now with the application of artificial intelligence and machine learning. Glamorous projects to do everything from curing cancer to helping paralyzed patients walk through AI have generated enormous expectations. But the greatest opportunity for AI in the near term may come not from headline-grabbing moonshots but from putting computers and algorithms to work on the most mundane drudgery possible. Excessive paperwork and red-tape is the sewage of modern medicine. An estimated 14% of wasted health care spending — $91 billion — is the result of inefficient administration. Let’s give AI the decidedly unsexy job of cleaning out the administrative muck that’s clogging up our medical organizations, sucking value out of our economy, and literally making doctors ill with stress.”
How AI Is Taking the Scut Work Out of Health Care Jonathan Bush Harvard Business Review (March 2018) - “Excessive paperwork and red-tape is the sewage of modern medicine. An estimated 14% of wasted health care spending — $91 billion — is the result of inefficient administration. Let’s give AI the decidedly unsexy job of cleaning out the administrative muck that’s clogging up our medical organizations, sucking value out of our economy, and literally making doctors ill with stress.”
How AI Is Taking the Scut Work Out of Health Care Jonathan Bush Harvard Business Review (March 2018) - Open Radiology?
- PURPOSE: Diabetic retinopathy (DR) is one of the leading causes of preventable blindness globally. Performing retinal screening examinations on all diabetic patients is an unmet need, and there are many undiagnosed and untreated cases of DR. The objective of this study was to develop robust diagnostic technology to automate DR screening. Referral of eyes with DR to an ophthalmologist for further evaluation and treatment would aid in reducing the rate of vision loss, enabling timely and accurate diagnoses.
CONCLUSIONS: A fully data-driven artificial intelligence-based grading algorithm can be used to screen fundus photographs obtained from diabetic patients and to identify, with high reliability, which cases should be referred to an ophthalmologist for further evaluation and treatment. The implementation of such an algorithm on a global basis could reduce drastically the rate of vision loss attributed to DR. Automated Identification of Diabetic Retinopathy Using Deep Learning. Gargeya R1, Leng T2. Ophthalmology. 2017 Mar 27. pii: S0161-6420(16)31774-2 - “In computerized detection of clustered microcalcifications (MCs) from mammograms, the traditional approach is to apply a pattern detector to locate the presence of individual MCs, which are subsequently grouped into clusters. Such an approach is often susceptible to the occurrence of false positives (FPs) caused by local image patterns that resemble MCs. We investigate the feasibility of a direct detection approach to determining whether an image region contains clustered MCs or not. Toward this goal, we develop a deepconvolutional neural network (CNN) as the classifier model to which the input consists of a large image window. The multiple layers in the CNN classifier are trained to automatically extract image features relevant to MCs at different spatial scales.”
Global detection approach for clustered microcalcifications in mammograms using a deep learning network. Wang J, Nishikawa RM, Yang Y J Med Imaging (Bellingham). 2017 Apr;4(2):024501
- “In the same manner that automated blood pressure measurement and automated blood cell counts freed clinicians from some tasks, artificial intelligence could bring back meaning and purpose in the practice of medicine while providing new levels of efficiency and accuracy. Physicians must proactively guide, oversee, and monitor the adoption of artificial intelligence as a partner in patient care.”
What This Computer Needs Is a Physician: Humanism and Artificial Intelligence Abraham Verghese, MD1; Nigam H. Shah, MBBS, PhD1; Robert A. Harrington, MD JAMA (in press) doi:10.1001/jama.2017.19198 - “Recently, these deep learning algorithms have been applied to medical imaging in several clinical settings, such as detection of breast cancer on mammograms, segmentation of liver metastases with computed tomography (CT), brain tumor segmentation with magnetic resonance (MR) imaging, classification of interstitial lung disease with high-resolution chest CT, and generation of relevant labels pertaining to the content of medical images.”
Deep Learning: A Primer for Radiologists Chartrand G et al. RadioGraphics 2017; 37:2113–2131
- “There may be resistance to merging 2 distinct medical specialties, each of which has unique pedagogy, tradition, accreditation, and reimbursement. However, artificial intelligence will change these diagnostic fields. The merger is a natural fusion of human talent and artificial intelligence. United, radiologists and pathologists can thrive with the rise of artificial intelligence.”
Adapting to Artificial Intelligence Radiologists and Pathologists as Information Specialists Jha S, Topol EJ JAMA. Published online November 29, 2016. doi:10.1001/jama.2016.17438 - “Information specialists should train in the traditional sciences of pathology and radiology. The training should take no longer than it presently takes because the trainee will not spend time mastering the pattern recognition required to become a competent radiologist or pathologist. Visual interpretation will be restricted to perceptual tasks that artificial intelligence cannot perform as well as humans. The trainee need only master enough medical physics to improve suboptimal quality of medical images. Information special- ists should be taught Bayesian logic, statistics, and data science and be aware of other sources of information such as genomics and bio- metrics, insofar as they can integrate data from disparate sources with a patient’s clinical condition.”
Adapting to Artificial Intelligence Radiologists and Pathologists as Information Specialists Jha S, Topol EJ JAMA. Published online November 29, 2016. doi:10.1001/jama.2016.17438
- “In recent years machine learning (ML) has revolutionized the fields of computer vision and medical image analysis. Yet, a number of doubts remain about the applicability of ML in clinical practice. Medical doctors may question the lack of interpretability of classifiers; Or it is argued that ML methods require huge amounts of training data. Here we discuss some of these issues and show:
1. how decision trees (a special class of ML models) can be understood as an automatically-optimized generalization of conventional algorithms, and
2. how the issue of collecting labelled data (e.g. images) applies to both manually-designed and learning-based algorithms.”
Machine learning for medical images analysis Criminisi A Medical Image Analysis Volume 33, October 2016, Pages 91–93 - “Ultimately, these researchers argue, the complex answers given by machine learning have to be part of science’s toolkit because the real world is complex: for phenomena such as the weather or the stock mar- ket, a reductionist, synthetic description might not even exist.“
There are things we cannot verbalize,” says Stéphane Mallat, an applied math- ematician at the École Polytechnique in Paris.
- For select cancer histologies, aggressive focal therapy of oligometastatic lesions is already the clinical standard of care (i.e. colorectal cancer and sarcomas), while for other tumor types the evidence is still emerging (i.e. prostate, breast, etc.). It is increasingly important, therefore, for the radiologist interpreting oncology patients’ staging or restaging examinations to be aware of those diseases for which targeted therapy of oligometastases may be undertaken to effectively guide such management. The improved imaging resolution provided by technological advances promise to aid in the detection of subtle sites of disease to ensure the identification of patients with oligometastases amenable to targeted treatment.
What the Radiologist Needs to Know to Guide Patient Management Steven P. Rowe, MD, Hazem Hawasli, Elliot K. Fishman, MD, Pamela T. Johnson, Acad Radiol 2016; 23:326–328 - “As such, some of the impetus for exploring aggressive and potentially curative treatment in patients with oligometastases can come from improvements in imaging technology and techniques. Thus, radiologists should not only understand the implications of the new paradigm of oligometastatic disease for how they interpret studies, but they should also ac- tively engage in the research necessary to optimize the selection of patients for aggressive therapy of oligometastases.“
What the Radiologist Needs to Know to Guide Patient Management Steven P. Rowe, MD, Hazem Hawasli, Elliot K. Fishman, MD, Pamela T. Johnson, Acad Radiol 2016; 23:326–328