Prediction Models for Kidney Failure in Long-Term Survivors of Childhood Cancer

Get Permission

In a study reported in the Journal of Clinical Oncology, Wu et al developed models for predicting kidney failure among 5-year survivors of childhood cancer.

Study Details

In the study, predictive models were developed using data from 25,483 survivors from the Childhood Cancer Survivor Study (CCSS) without a history of kidney failure. The cohort was assessed for subsequent kidney failure (ie, dialysis, kidney transplantation, or kidney-related death) by age 40 years. A sibling cohort (n = 5,045) served as a comparator. Results were validated in a St. Jude Lifetime Cohort Study cohort of 2,490 survivors and a National Wilms Tumor Study cohort of 396 survivors.

Two predictive models yielding risk scores were developed, consisting of a simple binary Yes or No model and a dose-specific model. For the simple model, binary predictors of late kidney failure included in the final model were: Black race (vs all other race/ethnicity groups combined); age < 10 years at cancer diagnosis; history of nephrectomy within 5 years of cancer diagnosis; exposure to ifosfamide, platinum, and anthracycline chemotherapy; history of abdominal radiation; congenital genitourinary anomalies; and onset of hypertension within 5 years of cancer diagnosis. The final dose-specific model added cumulative ifosfamide dose and mean radiotherapy kidney dose to the factors in the simple model.   

Key Findings

Among CCSS survivors, 204 developed late kidney failure. The receiver operating characteristic area under the curve (AUC) and C-statistic values for development by 40 years were 0.65 (95% confidence interval [CI] = 0.62–0.69) and 0.68 (95% CI = 0.65–0.72), respectively, for the simple model, and 0.67 (95% CI = 0.63–0.70) and 0.69 (95% CI = 0.64–0.72), respectively, for the dose-specific model.

In the St. Jude Lifetime Cohort Study cohort, AUC and C-statistic values were 0.83 and 0.86 for the simple model and 0.88 and 0.88 for the dose-specific model. In the National Wilms Tumor Study cohort, AUC and C-statistic values were 0.62 and 0.63 for the simple model and 0.67 and 0.64 for the dose-specific model.

Risk scores in the CCSS cohort were aggregated to form distinct low- (n = 17,762), moderate- (n = 3,784), and high-risk (n = 716) groups. Cumulative incidence of kidney failure by age 40 years was: 0.6% (95% CI = 0.4%–0.7%) in both models in the low-risk group; 2.3% (95% CI = 1.6%–3.2%) in the simple model and 2.1% (95% CI = 1.5%–2.9%) in the dose-specific model in the moderate-risk group; and 9.4% (95% CI = 4.4%–16.7%) in the simple model and 7.5% (95% CI = 4.3%–11.6%) in the dose-specific model in the high-risk group. The cumulative incidence in the sibling cohort was 0.2% (95% CI = 0.1%–0.5%) in the sibling cohort. 

The low-risk survivor group had a significantly greater risk of kidney failure compared with the sibling cohort, with relative risks of 3.2 (95% CI = 1.6%–6.1%) in the simple model and 3.0 (95% CI = 1.6%–5.9%) in the dose-specific model (both P < .001).

The investigators concluded, “Prediction models accurately identify childhood cancer survivors at low, moderate, and high risk for late kidney failure and may inform screening and interventional strategies.”

Natalie L. Wu, MD, MS, of the University of California San Francisco Benioff Children’s Hospital, is the corresponding author for the Journal of Clinical Oncology article.

Disclosure: The study was supported by the National Cancer Institute and American Lebanese Syrian Associated Charities. For full disclosures of the study authors, visit

The content in this post has not been reviewed by the American Society of Clinical Oncology, Inc. (ASCO®) and does not necessarily reflect the ideas and opinions of ASCO®.