A Precision Diagnostic Framework of Renal Cell Carcinoma on Whole-Slide Images using Deep Learning

Diagnostic pathology, which is the basis and gold standard of cancer diagnosis, provides essential information on the prognosis of the disease and vital evidence for clinical treatment. However, pathological diagnosis is subjective, and differences in observation and diagnosis between pathologists are common. This phenomenon is more evident in hospitals with insufficient medical resources. Deep learning (DL) can be used to identify and classify structures in digital pathology. In order to solve the above difficulties, in this work, we propose a DL framework for generating pathological diagnosis by analyzing histopathological images of renal cell carcinoma. A deep neural network is trained on a large high-quality annotated dataset for accurate tumor area detection, subtyping, and grading. The results show that our framework has achieved pathologist-level accuracy in diagnosis, can generate pathology reports with tumor indicators, and provide pathologists with interpretable auxiliary diagnoses

I. Introduction

Renal cell carcinoma (RCC) is one of the most common types of renal cancer, accounting for 80% of early renal cancer, which is the third most common urological tumor after prostate cancer, and bladder cancer [1]. With the development of medical imaging, the detection rate of early kidney cancer has increased. The most common histopathological type of renal cell carcinoma is clear cell carcinoma (ccRCC), papillary renal cell carcinoma (pRCC), and chromophobe cell carcinoma (chRCC) [2]. In the pathological diagnosis, the subtypes and grades are the critical diagnostic results, different subtypes of renal cell carcinoma can be treated with different regimens (including chemotherapy and targeted therapy), the higher the grading, the worse the prognosis, the greater the possibility of recurrence and tumor metastasis, and the faster the course of the disease.