Imaging Pearls ❯ Liver ❯ Texture Analysis
|
-- OR -- |
|
- Purpose: To explore the value of CT texture analysis (CTTA) for differentiation of focal nodular hyperplasia (FNH) from hepatocellular adenoma (HCA) on contrast-enhanced CT(CECT).
Results: On HAP images, mean, mpp, and skewness were significantly higher in FNH than in HCA on unfiltered images (p £ 0.007); SD, entropy, and mpp on filtered analysis (p £ 0.006). On PVP, mean, mpp, and skewness in FNH were significantly different from HCA (p £ 0.001) on unfiltered images, while entropy and kurtosis were significantly higher in FNH on filtered images (p £ 0.018). The multivariate logistic regression analysis indicated that the mean, mpp, and entropy of medium-level and coarse-level filtered images on HAP were independent predictors for the diagnosis of HCA and a model based on all these parameters showed the largest AUROC (0.824).
Conclusions: Multiple explored CTTA parameters are significantly different between FNH and HCA on CECT.
Evaluation of texture analysis for the differential diagnosis of focal nodular hyperplasia from hepatocellular adenoma on contrast-enhanced CT images
Roberto Cannella et al.
Abdom Radiol (2019) 44:1323–1330 - Purpose: To explore the value of CT texture analysis (CTTA) for differentiation of focal nodular hyperplasia (FNH) from hepatocellular adenoma (HCA) on contrast-enhanced CT(CECT).
Conclusions: Multiple explored CTTA parameters are significantly different between FNH and HCA on CECT.
Evaluation of texture analysis for the differential diagnosis of focal nodular hyperplasia from hepatocellular adenoma on contrast-enhanced CT images
Roberto Cannella et al.
Abdom Radiol (2019) 44:1323–1330 - “In conclusion, multiple explored CTTA parameters are significantly different in FNH compared to HCA. These results justify further larger-cohort studies, preferably multicenter, to validate the use of texture analysis in differentiating benign liver lesions and to increase the diagnostic value of CECT.”
Evaluation of texture analysis for the differential diagnosis of focal nodular hyperplasia from hepatocellular adenoma on contrast-enhanced CT images
Roberto Cannella et al.
Abdom Radiol (2019) 44:1323–1330
- “ The random forest model successfully distinguished the three lesion types and normal liver, correctly categorizing adenomas in 91.2% of cases, FNHs in 94.4% of cases, and HCCs in 98.6% of cases, with an overall error rate of 4.2%. When incorporating both texture analysis data and patient characteristics, the model demonstrated a sensitivity and specificity for adenoma of 91.2% and 98.3%, for FNH of 94.4% and 98.4%, and for HCC of 98.6% and 99.5%, respectively.”
Preliminary data using CT texture analysis for the classification of hypervascular liver lesions:
Generation of a predictive model on the basis of
quantitative spatial frequency measurements – A work in progress
Raman SP, Fishman EK et al.
J Comput Assist Tomogr (in press) - “ The accuracy rates for the two human readers are provided in Table 7. The overall accuracy rate for the senior reader was 72.2% and 65.6% for the more junior reader. Accuracy, sensitivity, and specificity rates in Table 7 reflect the ‘pooled’ performance of the two readers together. Notably, the accuracy rates for all three classes of lesions (HCC, adenoma, FNH) was lower than those derived from the texture analysis data, with the greatest discrepancy seen with adenomas.”
Preliminary data using CT texture analysis for the classification of hypervascular liver lesions:
Generation of a predictive model on the basis of
quantitative spatial frequency measurements – A work in progress
Raman SP, Fishman EK et al.
J Comput Assist Tomogr (in press) - “Texture analysis provides quantitative measures of heterogeneity from the distribution of pixel intensities at different spatial frequencies within a region of interest. While this technique has previously been used to primarily predict tumor prognosis and patient outcomes, CTTA may prove valuable in lesion discrimination and characterization.”
Preliminary data using CT texture analysis for the classification of hypervascular liver lesions:
Generation of a predictive model on the basis of
quantitative spatial frequency measurements – A work in progress
Raman SP, Fishman EK et al.
J Comput Assist Tomogr (in press) - “In this study, incorporating only a patient’s age, gender, and selected texture parameters generated from a single ROI, a random forest model was able to correctly categorize adenomas in 91.2% of cases, FNHs in 94.4% of cases, and HCCs in 98.6% of cases, with sensitivity and specificity for adenoma of 91.2% and 98.3%, for FNH of 94.4% and 98.4%, and for HCC of 98.6% and 99.5%, respectively.”
Preliminary data using CT texture analysis for the classification of hypervascular liver lesions: Generation of a predictive model on the basis of
quantitative spatial frequency measurements – A work in progress
Raman SP, Fishman EK et al.
J Comput Assist Tomogr (in press) - Representative examples from each type of hypervascular liver lesion in the study cohort, selected from the subset of lesions for which there was a age and gender matched control in the normal liver cohort
- The result of a multiple-variable analysis of variance (MANOVA) displayed as canonical variables in the group-determined eigenspace. The axes of the eigenspace are determined by the combination of variables that maximize the distances between groups.
Explicit prediction model formula N for classifying normal liver parenchyma vs. hypervascular lesion (focal nodular hyperplasia, adenoma, or hepatocellular carcinoma). For an arbitrary cutoff of N < c (cutoff value), the resulting area under a receiver operator characteristic curve is 0.94. For the cutoff value of 0 (that is a positive output indicating a lesion and negative output indicating normal parenchyma), the sensitivity and specificity are 96% and 91% respectively, while the error rate is 5.3%
The variable importance plots corresponding the random forest analyses reported in Tables 3 (A), 4 (B), 5 (C), and 6 (D), respectively. Variable “mean0” is the mean of the ROI while “Mean” is the background liver intensity.
- “CT texture analysis (CTTA) is a method of quantifying lesion heterogeneity based on distribution of pixel intensities within a region of interest. This study investigates the ability of CTTA to distinguish different hypervascular liver lesions, and seeks to develop a predictive model utilizing CTTA parameters to distinguish different lesions.”
Classification of hypervascular liver lesions using CT texture analysis:
Generation of a predictive model on the basis of
quantitative spatial frequency measurements
Raman SP, Schroeder J, Huang P, Fishman EK
Radiology (in revision) - “The random forest model successfully distinguished the three lesion types and normal liver, correctly categorizing adenomas in 91.2% of cases, FNHs in 94.4% of cases, and HCCs in 98.6% of cases, with an overall error rate of 4.2%. Logistic regression was utilized to create models distinguishing normal liver from the three lesions, and malignant (HCC) from benign lesions using a small number of the texture parameters.”
Classification of hypervascular liver lesions using CT texture analysis:
Generation of a predictive model on the basis of
quantitative spatial frequency measurements
Raman SP, Schroeder J, Huang P, Fishman EK
Radiology (in revision) - “Texture analysis provides quantitative measures of heterogeneity from the distribution of pixel intensities at different spatial frequencies within a region of interest. While this technique has previously been used to primarily predict tumor prognosis and patient outcomes, CTTA may prove valuable in lesion discrimination and characterization. In this study, incorporating only a patient’s age, gender, and selected texture parameters generated from a single ROI, a random forest model was able to correctly categorize adenomas in 91.2% of cases, FNHs in 94.4% of cases, and HCCs in 98.6% of cases.”
Classification of hypervascular liver lesions using CT texture analysis:
Generation of a predictive model on the basis of
quantitative spatial frequency measurements
Raman SP, Schroeder J, Huang P, Fishman EK
Radiology (in revision) - “ CTTA may prove valuable in lesion discrimination and characterization, and is able to successfully categorize three common hypervascular lesions based on texture parameters.”
Classification of hypervascular liver lesions using CT texture analysis:
Generation of a predictive model on the basis of
quantitative spatial frequency measurements
Raman SP, Schroeder J, Huang P, Fishman EK 
