The Positive Predictive Values of the Breast Imaging Reporting and Data System (BI-RADS) 4 Lesions and its Mammographic Morphological Features.

The Breast Imaging Reporting and Data System (BI-RADS) is a comprehensive guideline to systematize breast imaging reporting, and as per its recommendations, any lesion with likelihoods of malignancy greater than 2% is deemed as suspicious and tissue diagnosis is recommended. The aim of the study is to determine the positive predictive value (PPV) of BI-RADS categories 4a, 4b, and 4c for malignancy and association of mammographic morphological features of BI-RADS 4 subgroups with malignant outcomes. We retrospectively reviewed all the patients undergoing mammography with BI-RADS score of 4 followed by biopsy from May 2019 to April 2020. The predictive values of BI-RADS 4 subcategories and morphological features with malignancy are performed taking histopathology report as the gold standard. The PPV of BI-RADS subcategories 4a, 4b, and 4c for malignancies were 34, 89, and 97%, respectively. BI-RADS 4c patients tend to be older (50.2 ± 12.2 vs. 44.6 ± 10.3 years) with larger mass (44 ± 16 vs. 32.9 ± 16.8 mm) at presentation than 4a. Postmenopausal state (P = 0.03) and older age (P = 0.019) were significantly associated with malignancy. There is no meaningful difference observed in the predictability of BI-RADS category 4c lesions among different breast density patterns. The overall higher PPV for BI-RADS 4a and 4b reflects subjectivity in subcategory assignments of BI-RADS 4. In patients, less than 40 years with the BI-RADS 4a category on mammograms may undergo supplementary imaging with MRI which may downscale the lesion classification in turn reducing unnecessary biopsy and surgery.

Prediction of Colon Cancer Stages and Survival Period with Machine Learning Approach.

The prediction of tumor in the TNM staging (tumor, node, and metastasis) stage of colon cancer using the most influential histopathology parameters and to predict the five years disease-free survival (DFS) period using machine learning (ML) in clinical research have been studied here. From the colorectal cancer (CRC) registry of Chang Gung Memorial Hospital, Linkou, Taiwan, 4021 patients were selected for the analysis. Various ML algorithms were applied for the tumor stage prediction of the colon cancer by considering the Tumor Aggression Score (TAS) as a prognostic factor. Performances of different ML algorithms were evaluated using five-fold cross-validation, which is an effective way of the model validation. The accuracy achieved by the algorithms taking both cases of standard TNM staging and TNM staging with the Tumor Aggression Score was determined. It was observed that the Random Forest model achieved an F-measure of 0.89, when the Tumor Aggression Score was considered as an attribute along with the standard attributes normally used for the TNM stage prediction. We also found that the Random Forest algorithm outperformed all other algorithms, with an accuracy of approximately 84% and an area under the curve (AUC) of 0.82 ± 0.10 for predicting the five years DFS.