Introduction: Many kidney tumors are benign and some malignant tumors behave in an indolent fashion. We sought to develop predictive models of kidney tumor malignancy and high-grade malignancy. Methods: Patients diagnosed with solitary renal masses were identified from the Canadian Kidney Cancer information system (CKCis). Specifically, we identified patients with clinical stage T1 and T2 disease. Demographic, clinical, and imaging data were compared to the pathologic diagnosis from surgery or biopsy. Tumors were categorized as malignant or benign, and aggressive (high-grade malignant) or indolent (low-grade malignant and benign). Logistic regression models were constructed to identify predictors of each category. Nomograms were created using statistically significant risk factors and were internally validated using bootstrap methods. Results: Of 5,517 CKCis patients with a solitary tumor between January 2011 and October 2022, 5,054 (92%) had malignant histology and 1,953 (40%) had high-grade disease. Factors associated with malignancy and high-grade malignancy were male sex (Odds Ratio [OR] 1.45; 95% confidence interval [CI] 1.19-1.77; OR 1.65, 95%CI 1.44-1.89, respectively) and tumor size (OR 1.27, 95%CI 1.20-1.33; OR 1.29, 95%CI 1.25-1.31, per increase in 1cm, respectively). An interaction between age and sex was identified for malignancy; odds of malignancy statistically decrease with older age in men and increase (though not significantly) with age in women (OR 0.89, 95%CI 0.84-0.95; OR 1.01, 95%CI 0.95-1.07, respectively). Older age was predictive of high-grade malignancy (OR 1.06, 95%CI 1.03-1.09, per increase in 5 years). The nomograms for malignant/benign tumors had moderate discrimination and excellent calibration (optimism corrected area under the curve [AUC]= 0.69, root mean square error [RSME]=0.02, calibration slope=0.99). The nomogram for aggressive/indolent tumors had good discrimination and excellent calibration (AUC=0.75, RSME=0.01, calibration slope=1.00). Conclusions: Patient and tumor characteristics are independently associated with cancer risk and high grade-cancer risk. The CKCis nomograms presented should be externally validated. These prediction tools can be used by physicians and patients with kidney tumors to help determine an optimal management plan. SOURCE OF Funding: No direct role or influence by sponsors. The Canadian Kidney Cancer information system (CKCis) is funded by the Kidney Cancer Research Network of Canada which receives funding support from industry sponsors.
