Digital pathology-based study of cell- and tissue-level morphologic features in serous borderline ovarian tumor and high-grade serous ovarian cancer
Serous borderline ovarian tumor (SBOT) and high-grade serous ovarian cancer (HGSOC) are two distinct subtypes of epithelial ovarian tumors, with markedly different biologic background, behavior, prognosis, and treatment. However, the histologic diagnosis of serous ovarian tumors can be subjectively variable and labor-intensive as multiple tumor slides/blocks need to be thoroughly examined to search for these features. In this study, we aimed to evaluate technical feasibility of using digital pathological approaches to facilitate objective and scalable diagnosis screening for SBOT and HGSOC. Based on Groovy scripts and QuPath, a novel informatics system was developed to facilitate interactive annotation and imaging data exchange for machine learning purposes. Through this developed system, cellular boundaries were detected and expanded set of cellular features were extracted to represent cell- and tissue-level characteristics. According to our evaluation, cell-level classification was accurately achieved for both tumor and stroma cells with greater than 90% accuracy. Upon further re-examinations, 44.2% of the misclassified cells were due to over-/under-segmentations or low-quality of imaging areas. For a total number of 6,485 imaging patches with sufficient tumor and stroma cells (ten of each at least), we achieved 91-95% accuracy to differentiate HGSOC v. SBOT. When all the patches were considered for a WSI to make consensus prediction, 97% accuracy was achieved for accurately classifying all patients, indicating that cellular features digitally extracted from pathological images can be used for cell classification and SBOT v. HGSOC differentiation. Introducing digital pathology into ovarian cancer research could be beneficial to discover potential clinical implications.