Prediction Of The Breast Tumour Based On Image Processing And Machine Learning Techniques

Abstract

This study explored the prognostic value of a digital risk score (DRS) derived from computational analysis of breast tumor tissue images. A DRS model was developed and validated using a cohort of One thousand two hundred and ninety-nine patients with breast cancer were evaluated. The model showed good performance in detecting individual with higher risk pro-files. and low risk of breast cancer death based on tumor morphology, including size, grade, involvement of lymph nodes, and hormone receptor status. Survival analysis demonstrated significant predictive power of the DRS for both disease-specific and overall survival, particularly in specific tumor subtypes and Subgroups of hormone receptor status. The DRS exhibited a statistically significant difference in its prognostic model compared to a visual risk score assigned by experienced pathologists, highlighting the complementary nature of both approaches. Our findings suggest that the DRS, in conjunction with conventional clinicopathological factors, offers a valuable tool for risk stratification and personalized treatment guidance in breast cancer. This work used a machine learning algorithm trained on digital pictures of tumor tissue microarrays (TMAs) to build and validate a digital risk score (DRS) model for predicting patient survival, both overall and specific to breast cancer.