RESUMO
BACKGROUND & AIMS: Noninvasive variceal risk stratification systems have not been validated in patients with hepatocellular carcinoma (HCC), which presents logistical barriers for patients in the setting of systemic HCC therapy. We aimed to develop and validate a noninvasive algorithm for the prediction of varices in patients with unresectable HCC. METHODS: We performed a retrospective cohort study in 21 centers in the United States including adult patients with unresectable HCC and Child-Pugh A5-B7 cirrhosis diagnosed between 2007 and 2019. We included patients who completed an esophagogastroduodonoscopy (EGD) within 12 months of index imaging but before HCC treatment. We divided the cohort into a 70:30 training set and validation set, with the goal of maximizing negative predictive value (NPV) to avoid EGD in low-risk patients. RESULTS: We included 707 patients (median age, 64.6 years; 80.6% male; 74.0% White). Median time from HCC diagnosis to EGD was 47 (interquartile range, 114) days, with 25.0% of patients having high-risk varices. A model using clinical variables alone achieved an NPV of 86.3% in the validation cohort, whereas a model integrating clinical and imaging variables had an NPV 97.4% in validation. The clinical and imaging model would avoid EGDs in more than half of low-risk patients while misclassifying 7.7% of high-risk patients. CONCLUSIONS: A model incorporating clinical and imaging data can accurately predict the absence of high-risk varices in patients with HCC and avoid EGD in many low-risk patients before the initiation of systemic therapy, thus expediting their care and avoiding treatment delays.
RESUMO
PURPOSE: To determine if ancillary sonographic and Doppler parameters can be used to predict transplant renal artery stenosis in patients with renal graft dysfunction. MATERIALS AND METHODS: IRB-approved, HIPAA-compliant retrospective study included 80 renal transplant patients who had renal US followed by renal angiogram between January 2018 and December 2019. A consensus read of two radiologists recorded these parameters: peak systolic velocity, persistence of elevated velocity, grayscale narrowing, parvus tardus, delayed systolic upstroke, angle of the systolic peak (SP angle), and aliasing. Univariate analysis using t-test or chi-square was performed to determine differences between patients with and without stenosis. P values under 0.05 were deemed statistically significant. We used machine learning algorithms to determine parameters that could better predict the presence of stenosis. The algorithms included logistic regression, random forest, imbalanced random forest, boosting, and CART. All 80 cases were split between training and testing using stratified sampling using a 75:25 split. RESULTS: We found a statistically significant difference in grayscale narrowing (p = 0.0010), delayed systolic upstroke (p = 0.0002), SP angle (p = 0.0005), and aliasing (p = 0.0024) between the two groups. No significant difference was found for an elevated peak systolic velocity (p = 0.1684). The imbalanced random forest (IRF) model was selected for improved accuracy, sensitivity, and specificity. Specificity, sensitivity, AUC, and normalized Brier score for the IRF model using all parameters were 73%, 81%, 0.82, and 69 in the training set, and 78%, 58%, 0.78, and 80 in the testing set. VIMP assessment showed that the combination of variables that resulted in the most significant change of the training set performance was that of grayscale narrowing and SP angle. CONCLUSION: Elevated peak systolic velocity did not discriminate between patients with and without TRAS. Adding ancillary parameters into the machine learning algorithm improved specificity and sensitivity similarly in the training and testing sets. The algorithm identified the combination of lumen narrowing coupled with the angle of the systolic peak as better predictor of TRAS. This model may improve the accuracy of ultrasound for transplant renal artery stenosis.