Rethinking Cancer Epidemiology, Aging and Machine Learning

The incidence of skin cancer is rising globally, posing a significant public health threat. An early and accurate diagnosis is crucial for patient prognoses. However, discriminating between malignant melanoma and benign lesions, such as nevi and keratoses, remains a challenging task due to their visual similarities. Image-based recognition systems offer a promising solution to aid dermatologists and potentially reduce unnecessary biopsies. This research investigated the performance of four unified convolutional neural networks, namely, YOLOv3, YOLOv4, YOLOv5, and YOLOv7, in classifying skin lesions. Each model was trained on a benchmark dataset, and the obtained performances were compared based on lesion localization, classification accuracy, and inference time. In particular, YOLOv7 achieved superior performance with an Intersection over Union (IoU) of 86.3%, a mean Average Precision (mAP) of 75.4%, an F1-measure of 80%, and an inference time of 0.32 s per image. These findings demonstrated the potential of YOLOv7 as a valuable tool for aiding dermatologists in early skin cancer diagnosis and potentially reducing unnecessary biopsies. Full article

Background: Choroid plexus tumors (CPTs) are rare neoplasms found in the central nervous system, comprising 1% of all brain tumors. These tumors include choroid plexus papilloma (CPP), atypical choroid plexus papilloma (aCPP), and choroid plexus carcinoma (CPC). Although gross total resection for choroid plexus papillomas (CPPs) is associated with long-term survival, there is a scarcity of prospective data concerning the role and sequence of neoadjuvant therapy in treating aCPP and CPC. Methods: From the years 2000 to 2019, 679 patients with CPT were identified from the Surveillance, Epidemiology, and End Result (SEER) database. Among these patients, 456 patients had CPP, 75 patients had aCPP, and 142 patients had CPC. Univariate and multivariable Cox proportional hazard models were run to identify variables that had a significant impact on the primary endpoint of overall survival (OS). A predictive nomogram was built for patients with CPC to predict 5-year and 10-year survival probability. Results: Histology was a significant predictor of OS, with 5-year OS rates of 90, 79, and 61% for CPP, aCPP, and CPC, respectively. Older age and African American race were prognostic for worse OS for patients with CPP. Older age was also associated with reduced OS for patients with aCPP. American Indian/Alaskan Native race was linked to poorer OS for patients with CPC. Overall, treatment with gross total resection or subtotal resection had no difference in OS in patients with CPP or aCPP. Meanwhile, in patients with CPC, gross total resection (GTR) was associated with significantly better OS than subtotal resection (STR) only. However, there is no difference in OS between patients that receive GTR and patients that receive STR with adjuvant therapy. The nomogram for CPC considers types of treatments received. It demonstrates acceptable accuracy in estimating survival probability at 5-year and 10-year intervals, with a C-index of 0.608 (95% CI of 0.446 to 0.77). Conclusions: This is the largest study on CPT to date and highlights the optimal treatment strategies for these rare tumors. Overall, there is no difference in OS with GTR vs. STR in CPP or aCPP. Furthermore, OS is equivalent for CPC with GTR and STR plus adjuvant therapy. Full article