Back

Performance of the iBox prognostication system in African American kidney transplant recipients: a multicenter study

Truchot, A.; Lombardi, Y.; Raynaud, M.; Aubert, O.; Divard, G.; Thalamas, T.; Astor, B.; Mandelbrot, D.; Parajuli, S.; Newell, K. A.; Orandi, B.; Friedewald, J. J.; Gupta, G.; Akalin, E.; Jordan, S. C.; Matas, A. J.; Molnar, M. Z.; Yamauchi, J.; Fornadi, K.; Bentall, A. J.; Stegall, M. D.; Mannon, R. B.; Segev, D. L.; Huang, E.; Fitzsimmons, W. E.; Loupy, A.

2025-12-02 transplantation
10.64898/2025.12.01.25341345 medRxiv
Show abstract

The iBox is a validated prognostication system that predicts graft loss in kidney transplant recipients (KTRs), but its performance in the specific population of African American recipients has not been fully studied. We conducted a multicenter study on 3,588 KTRs from North America, including 866 (24.1%) African American KTRs, to assess the impact of race on the iBoxs performance in predicting graft loss. Performance metrics for the prediction of graft loss were similar for both African American and non-African American KTRs in terms of discrimination (C-index 0.81 [95% CI: 0.78-0.84] and 0.83 [95%CI: 0.81-0.85], respectively, p=0.25), and calibration (observed/expected ratio 1.13 [95%CI: 1.03-1.24] and 1.03 [95%CI: 0.95-1.11], respectively, p=0.13). No significant interaction between iBox score values and race was found in multivariate analysis stratified by transplant center (p=0.80 for the interaction term). Results were consistent regardless of the equation used to estimate glomerular filtration rate (Kidney Recipient Specific, CKD-EPI, or MDRD). In conclusion, the iBox prognostication system accurately predicts graft loss up to 7 years post-risk evaluation in African American KTRs and confirms its relevance as a surrogate endpoint in this population.

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.