Background: This study aimed to evaluate the significance of MRI-based radiomics model derived from high-resolution T2-weighted images (T2WIs) in predicting tumor pathological features of rectal cancer. Methods: A total of 152 patients with rectal cancer who underwent surgery without any neoadjuvant therapy between March 2017 and September 2018 were included retrospectively. The patients were scanned using a 3-T magnetic resonance imaging, and high-resolution T2WIs were obtained. Lesions were delineated, and 1029 radiomics features were extracted. Least absolute shrinkage and selection operator was used to select features, and multilayer perceptron (MLP), logistic regression (LR), support vector machine (SVM), decision tree (DT), random forest (RF), and K-nearest neighbor (KNN) were trained using fivefold cross-validation to build a prediction model. The diagnostic performance of the prediction models was assessed using the receiver operating characteristic curves. Results: A total of 1029 features were extracted, and 15, 11, and 11 features were selected to predict the degree of differentiation, T stage, and N stage, respectively. The best performance of the radiomics model for the degree of differentiation, T stage, and N stage was obtained by SVM [area under the curve (AUC), 0.862; 95% confidence interval (CI), 0.750–0.967; sensitivity, 83.3%; specificity, 85.0%], MLP (AUC, 0.809; 95% CI, 0.690–0.905; sensitivity, 76.2%; specificity, 74.1%), and RF (AUC, 0.746; 95% CI, 0.622-0.872; sensitivity, 79.3%; specificity, 72.2%). Conclusion: This study demonstrated that the high-resolution T2WI–based radiomics model could serve as pretreatment biomarkers in predicting pathological features of rectal cancer.

At the time of diagnosis, approximately 15%‐20% of patients with rectal cancer (RC) presented synchronous liver metastasis (SLM), which is the most common cause of death in patients with RC. Therefore, preoperative, noninvasive, and accurate prediction of SLM is crucial for personalized treatment strategies. Recently, radiomics has been considered as an advanced image analysis method to evaluate the neoplastic heterogeneity with respect to diagnosis of the tumor and prediction of prognosis. In this study, a total of 1409 radiomics features were extracted for each volume of interest (VOI) from high‐resolution T2WI images of the primary RC. Subsequently, five optimal radiomics features were selected based on the training set using the least absolute shrinkage and selection operator (LASSO) method to construct the radiomics signature. In addition, radiomics signature combined with carcinoembryonic antigen (CEA) and carbohydrate antigen 19‐9 (CA19‐9) was included in the multifactor logistic regression to construct the nomogram model. It showed an optimal predictive performance in the validation set as compared to that in the radiomics model. The favorable calibration of the radiomics nomogram showed a nonsignificant Hosmer‐Lemeshow test statistic (P > .05). The decision curve analysis (DCA) showed that the radiomics nomogram is clinically superior to the radiomics model. Therefore, the nomogram amalgamating the radiomics signature and clinical risk factors serve as an effective quantitative approach to predict the SLM of primary RC.

To explore the effect of b-value distributions on the repeatability and diagnostic performance of the ADC value in rectal cancer patients using multiple b-values and mono-exponential model diffusion-weighted imaging (DWI). Thirty-two preoperative rectal cancer patients, without receiving neoadjuvant therapy, were scanned on a 3 Tesla magnetic resonance imaging scanner using DWI with 10 b-values ranging from 0 to 2000 s/mm2. The apparent diffusion coefficient (ADC) value was calculated using a mono-exponential model and 31 b-value combinations consisting of 2 to 10 b-values were explored. Regions of interest with the maximum cross-sectional tumour size were outlined on the ADC map by two independent observers. Intraclass correlation coefficients (ICC), coefficient of variation (CV), and Bland-Altman plots between the two observers were calculated and evaluated to determine repeatability. Areas under receiver operating characteristic curves (AUCs) were evaluated for rectal cancer characterization. Correlations between the mean ADC values and T stage were assessed using the Spearman correlation coefficient (ρ). α (= ICC + AUC + |ρ|- CV - |bias|) was defined and used to assess the optimal b-value distribution. Postoperative pathology tests revealed 4 patients with T1, 10 patients with T2, and 18 patients with T3 stages. There were no significant difference in age and sex between the two groups (T1–2 vs. T3). Excellent reproducibility was observed for ADC values between two observers with ICC and CV values ranging from 0.920 to 0.998, and 1.475 to 5.568%, respectively. The mean percent difference and ρ between the paired measurements was ranged from − 2.7 to 1.2% and from − 0.759 to − 0.407, respectively. The b-value combinations with the top three α values were b(0, 1000 s/mm2), b(500, 1500, 2000 s/mm2) and b(100, 1000, 1500 s/mm2) for α = 2.581, 2.571 and 2.569, respectively. The number of b-values and their distributions influenced the repeatability of the ADC values and their diagnostic performance. The optimal b-value combination was 0 and 1000 s/mm2 for DWI examination of rectal cancer patients.

Background: To validate and compare various MRI-based radiomics models to evaluate treatment response to neoadjuvant chemoradiotherapy (nCRT) of rectal cancer. Methods: A total of 80 patients with locally advanced rectal cancer (LARC) who underwent surgical resection after nCRT were enrolled retrospectively. Rectal MR images were scanned pre- and post-nCRT. The radiomics features were extracted from T2-weighted images, then reduced separately by least absolute shrinkage and selection operator (LASSO) and principal component analysis (PCA). Four classifiers of Logistic Regression, Random Forest (RF), Decision Tree and K-nearest neighbor (KNN) models were constructed to assess the tumor regression grade (TRG) and pathologic complete response (pCR), respectively. The diagnostic performances of models were determined with leave-one-out cross-validation by generating receiver operating characteristic curves and decision curve analysis. Results: Three features related to the TRG and 11 features related to the pCR were obtained by LASSO. Top five principal components representing a cumulative contribution of 80% to overall features were selected by PCA. For TRG, the area under the curve (AUC) of RF model was 0.943 for LASSO and 0.930 for PCA, higher than other models (P < 0.05 for both). As for pCR, the AUCs of KNN for LASSO and PCA were 0.945 and 0.712, higher than other models (P < 0.05 for both). The DCA showed that LASSO algorithm was clinically superior to PCA. Conclusion: MRI-based radiomics models demonstrated good performance for evaluating the treatment response of LARC after nCRT and LASSO algorithm yielded more clinical benefit.