Predicting liver and/or lung metastases in colorectal cancer (CRC) patients remains a critical challenge, especially due to the strong class imbalance commonly found in clinical datasets such as SEER (Surveillance, Epidemiology, and End Results Program). To address this issue, this study presents a comparative analysis of eight data balancing techniques combined with seven machine learning algorithms for metastasis prediction using SEER data. The evaluated techniques include traditional SMOTE, ADASYN, Borderline-SMOTE, SMOTE-LOF, Radius-SMOTE, RDSMOTE, and CUSS, along with a baseline configuration without any balancing. A total of 53,463 CRC patients were analyzed, and model performance was assessed using F1-score and AUC metrics under five-fold stratified cross-validation. Among the techniques, SMOTE-LOF achieved the best overall results, particularly when combined with XGBoost, reaching an F1-score of 0.6642 and AUC of 0.8988. The results indicate that combining oversampling with noise filtering or structural awareness significantly enhances model sensitivity. This study highlights the importance of selecting appropriate resampling techniques for clinical prediction tasks and offers insights for improving the robustness and fairness of machine learning applications in oncology.

Chest radiography (CXR) is one of the first choices in epidemiological analyses such as tuberculosis, cancer, pneumonia, and, recently, COVID-19. It provides crucial information for decision making, treatment, and monitoring the evolution of clinical cases from small to high complexity. Thus, it is a valuable source of information for the study, training, research, and development of computational support to medical diagnoses. In this work, we introduce a new method for chest X-ray adjustment to identifying and correcting radiographic images orientation. So, they can be automatically rotated to a standard position. Our proposal uses structural characteristics and statistics of pixel intensity patterns of CXR images. Divided into three steps, our method begins with the preparation of the photos, followed by a feature extraction strategy, and it ends with the X-ray image orientation identification. We use three different databases that include pediatric and adult radiographic imaging. A result showed 99.4% accuracy in the databases in our experiments. The code prepared by the authors is publicly available.