Automated clear cell renal carcinoma grade classification with prognostic significance

Area of Science:

• Computational pathology
• Renal cell carcinoma research
• Digital pathology image analysis

Background:

• Clear cell renal cell carcinoma (ccRCC) grading is crucial for prognosis.
• Traditional grading relies on manual histopathological assessment, which can be subjective.
• Automated systems offer potential for objective and reproducible cancer grading.

Purpose of the Study:

• To develop and validate an automated 2-tiered Fuhrman's grading system for ccRCC using whole slide images (WSI).
• To assess the prognostic efficacy of the automated grading system compared to manual grading.

Main Methods:

• Utilized 395 The Cancer Genome Atlas (TCGA) ccRCC cases with WSIs and clinical data.
• Extracted 72 quantitative morphological, intensity, and texture features from nuclear regions.
• Developed a Lasso model to identify features associated with grade and predict grading outcomes.

Main Results:

• The Lasso model, using 26 features, achieved 84.6% sensitivity and 81.3% specificity in predicting ccRCC grade.
• The automated predicted grade was significantly associated with overall survival in an extended test set (HR 2.05).
• Manually assigned grades did not demonstrate prognostic significance in the study cohort.

Conclusions:

• An automated computational system can effectively perform 2-tiered Fuhrman's grading for ccRCC.
• The developed automated grading system shows prognostic value for patient survival.
• Future research should focus on adapting and validating this system for WHO/ISUP grading and diverse ccRCC cohorts.