Diagnostic accuracy of Chronic Kidney Disease Epidemiology Collaboration 2021 and Janowitz-estimated glomerular filtration rate for identifying renal dysfunction compared with 99mTc-DTPA in patients with locally advanced cervical cancer

Why clinicians should know about it

• Picked for Obstetrics and Gynecology (paper of the day, 8 September 2026): CKD‑EPI 2021 GFR accuracy for cisplatin eligibility

Abstract

OBJECTIVE: To assess the diagnostic accuracy of the Chronic Kidney Disease Epidemiology Collaboration 2021 and Janowitz equations for identifying renal dysfunction, defined using the pre-specified absolute glomerular filtration rate threshold of <60 mL/min, a treatment-relevant boundary for cisplatin eligibility assessment, using 99mTc-DTPA (Technetium-99m Diethylenetriaminepentaacetic acid) isotopic quantification as the reference standard in women with locally advanced cervical cancer. METHODS: Retrospective diagnostic accuracy study including women with International Federation of Gynecology and Obstetrics 2018 stage IB3-IVA cervical cancer evaluated at the Instituto Nacional de Cancerología in Bogotá, Colombia, between June 2018 and September 2023 and selected as candidates for primary concurrent chemoradiation. All glomerular filtration rate values were standardized to absolute mL/min before analysis. 99mTc-DTPA values reported as mL/min/1.73 m2 were de-indexed by multiplying by body surface area/1.73. Chronic Kidney Disease Epidemiology Collaboration 2021 was calculated as an indexed creatinine-based estimate and converted to absolute mL/min using the same body surface area adjustment. Janowitz estimates were analyzed as absolute mL/min. RESULTS: A total of 210 of 237 consecutively included women had complete data and were included in the analysis; 27 (12.9%) had a 99mTc-DTPA glomerular filtration rate <60 mL/min. Correlation with measured glomerular filtration rate was moderate for Chronic Kidney Disease Epidemiology Collaboration 2021 (rho = 0.53, 95% confidence interval 0.43 to 0.62) and Janowitz (rho = 0.51, 95% confidence interval, 0.41 to 0.61). Mean bias (estimated minus measured glomerular filtration rate) was 5.0 mL/min for Chronic Kidney Disease Epidemiology Collaboration 2021 and 0.2 mL/min for Janowitz, with wide limits of agreement. Chronic Kidney Disease Epidemiology Collaboration 2021 detected 12/27 cases and missed 15 (10 false positives; sensitivity 44%, 95% confidence interval 25% to 65%; specificity 95%, 95% confidence interval 90% to 97%; negative likelihood ratio 0.59). Janowitz detected 10 of 27 and missed 17 (7 false positives, sensitivity 37%, 95% confidence interval 19% to 58%; specificity 96%, 95% confidence interval 92% to 98%; negative likelihood ratio 0.65). CONCLUSION: Chronic Kidney Disease Epidemiology Collaboration 2021 and Janowitz were insufficiently sensitive to replace measured glomerular filtration rate when cisplatin eligibility depends on threshold-based glomerular filtration rate classification. These findings support measured glomerular filtration rate when accurate threshold classification would alter cisplatin eligibility; any estimate-based triage strategy requires prospective validation.

Abstract as published, via PubMed.