Endoscopic Rectal Ultrasound-Based Radiomics Analysis for the Prediction of Synchronous Liver Metastasis in Patients With Primary Rectal Cancer.

OBJECTIVES: To develop and validate an ultrasound-based radiomics model to predict synchronous liver metastases (SLM) in rectal cancer (RC) patients preoperatively. METHODS: Two hundred and thirty-nine RC patients were included in this study and randomly divided into training and validation cohorts. A total of 5936 radiomics features were calculated on the basis of ultrasound images to build a radiomic model and obtain a radiomics score (Rad-score) using logistic regression. Meanwhile, clinical characteristics were collected to construct a clinical model. The radiomics-clinical model was developed and validated by integrating the radiomics features with the selected clinical characteristics. The performances of three models were evaluated and compared through their discrimination, calibration, and clinical usefulness. RESULTS: The radiomics model was developed based on 13 radiomic features. The radiomics-clinical model, which incorporated Rad-score, CEA, and CA199, exhibited favorable discrimination and calibration with areas under the receiver operating characteristic curve (AUC) of 0.920 (95% CI: 0.874-0.965) in the training cohorts and 0.855 (95% CI: 0.759-0.951) in the validation cohorts. And the AUC of the radiomics-clinical model was 0.849 (95% CI: 0.771-0.927) for the training cohorts and 0.780 (95% CI: 0.655-0.905) for the validation cohorts, the clinical model was 0.811 (95% CI: 0.718-0.905) for the training cohorts and 0.805 (95% CI: 0.645-0.965) for the validation cohorts. Moreover, decision curve analysis (DCA) further confirmed the clinical utility of the radiomics-clinical model. CONCLUSIONS: The radiomics-clinical model performed satisfactory predictive performance, which can help improve clinical diagnosis performance and outcome prediction for SLM in RC patients.

Can Modifications of LR-M Criteria Improve the Diagnostic Performance of Contrast-Enhanced Ultrasound LI-RADS for Small Hepatic Lesions up to 3 cm?

OBJECTIVE: To assess the diagnostic performance of the contrast-enhanced ultrasound liver imaging reporting and data system (CEUS LI-RADS) version 2017 for small hepatic lesions of ≤3 cm before and after changing the LR-M criteria. METHODS: We retrospectively analyzed the CEUS examination of 179 patients who were at high risk of hepatocellular carcinoma (HCC) with focal hepatic lesions ≤3 cm (194 lesions in total) and evaluated the diagnostic capability of the American College of Radiology and modified CEUS LI-RADS algorithms. RESULTS: Revision of the early washout time to 45 seconds increased the sensitivity of LR-5 in predicting HCC (P = .004), with no significant decrease in specificity (P = .118). It also made better the specificity of LR-M in predicting non-HCC malignancies (P = .001), with no significant decrease in sensitivity (P = .094). However, using within 3 minutes as the criterion for marked washout time improved the LR-5 sensitivity (P < .001) but decreased its specificity (P = .009) in predicting HCC, whereas the specificity of LR-M in predicting non-HCC malignancies increased (P < .001), but the sensitivity decreased (P = .027). CONCLUSIONS: CEUS LI-RADS (v2017) is a valid method for predicting HCC risk in high-risk patients. The diagnostic performance of LR-5 and LR-M could boost when the early washout time is revised to 45 seconds.

An endorectal ultrasound-based radiomics signature for preoperative prediction of lymphovascular invasion of rectal cancer.

OBJECTIVE: To investigate whether radiomics based on ultrasound images can predict lymphovascular invasion (LVI) of rectal cancer (RC) before surgery. METHODS: A total of 203 patients with RC were enrolled retrospectively, and they were divided into a training set (143 patients) and a validation set (60 patients). We extracted the radiomic features from the largest gray ultrasound image of the RC lesion. The intraclass correlation coefficient (ICC) was applied to test the repeatability of the radiomic features. The least absolute shrinkage and selection operator (LASSO) was used to reduce the data dimension and select significant features. Logistic regression (LR) analysis was applied to establish the radiomics model. The receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis (DCA) were used to evaluate the comprehensive performance of the model. RESULTS: Among the 203 patients, 33 (16.7%) were LVI positive and 170 (83.7%) were LVI negative. A total of 5350 (90.1%) radiomic features with ICC values of ≥ 0.75 were reported, which were subsequently subjected to hypothesis testing and LASSO regression dimension reduction analysis. Finally, 15 selected features were used to construct the radiomics model. The area under the curve (AUC) of the training set was 0.849, and the AUC of the validation set was 0.781. The calibration curve indicated that the radiomics model had good calibration, and DCA demonstrated that the model had clinical benefits. CONCLUSION: The proposed endorectal ultrasound-based radiomics model has the potential to predict LVI preoperatively in RC.