This study aims to preoperatively predict DFS of patients with localised ccRCC using a deep learning prognostic model based on enhanced contrast CT images, validate it's predictive ability in multicentre data and compare it's predictive ability with traditional models.

Inclusion Criteria:

• underwent partial/radical nephrectomies
• histologically diagnosed as ccRCC
• with complete clinical data and preoperative CT image data

Exclusion Criteria:

• with incomplete clinic-pathological data
• lack of preoperative contrast-enhanced CT images or the image quality was unsuitable for analysis
• who received pre-surgery neoadjuvant or adjuvant therapies
• with multiple renal tumors or/and had synchronous metastasis

Patients admitted to urology departments at participating medical centres