A Head-to-head Comparison of Four Prognostic Models for Prediction of Lymph Node Invasion in African American and Caucasian Individuals

Background: Four nomograms are available for the prediction of lymph node invasion (LNI) prior to radical prostatectomy (RP): the Cagiannos, the 2012-Briganti, the Godoy, and the online-Memorial Sloan Kettering Cancer Center (MSKCC). None was tested in African Americans (AAs). Objective: To perform a validation and head-to-head comparison of four nomograms for the prediction of LNI in AAs. Design, setting, and participants: Within the Surveillance Epidemiology and End Results database (2010–2014), we identified 14 077 Caucasians and 2668 AAs with clinically localised prostate cancer. Intervention: RP and pelvic lymph node dissection (PLND). Outcome measurements and statistical analysis: Predicted and observed LNI rates for the four nomograms were tested between AAs and Caucasians with Heagerty's concordance index (C-index), calibration plots, decision curve analyses (DCAs), and nomogram-derived cut-offs for avoided PLND and missed LNI cases. Results and limitations: All C-index values were lower in AAs than in Caucasians: Cagiannos (76.1% vs 79.5%), Godoy (73.0% vs 79.4%), 2012-Briganti (73.3% vs 81.3%), and MSKCC (72.6% vs 81.6%). All four nomogram calibration plots showed invariably worse performances in AAs. In DCAs focusing on AAs, the Cagiannos nomogram provided the highest net benefit relative to the remaining three nomograms. In nomogram cut-off analyses testing the number of avoided PLND against the number of missed LNI cases, the Cagiannos nomogram yielded the best results when tested in AAs. However, in DCAs and cut-off analyses that compared AAs with Caucasians, AAs yielded less favourable outcomes regardless of which nomogram was tested. We lack a central pathology review and standardisation of PLND templates. Conclusions: In AAs, the Cagiannos nomogram provides the optimal results relative to the remaining three nomograms. Nonetheless, all tested nomograms yielded worse metrics in AAs than in Caucasians. Patient summary: All four tested nomograms can be equally considered in Caucasians. Conversely, the Cagiannos nomogram should be preferred in African Americans.