Apppublic

drankitjoshi/NAHT-Model-Prostate_Cancer_Outcome_Predictor

sourceHugging Facecc-by-nc-4.0updated 2mo agoView on Hugging Face

0likes

App README

NAHT Outcome Predictor — High-Risk Prostate Cancer

Machine-learning models predicting outcomes after 6 months of Neoadjuvant Hormonal Therapy (NAHT) followed by Robot-assisted Radical Prostatectomy with bilateral pelvic lymph node dissection.

Models

Model	Timing	Cross-validated AUC (95% CI)
Positive surgical margin	Pre-operative	0.759 (0.685– 0.833)
Lymph node positivity	Pre-operative	0.842 (0.781– 0.894)
Biochemical recurrence	Post-operative	0.782 (0.710– 0.852)
BCR-free survival (Cox)	Post-operative	C-index 0.755 (0.702–0.807)

Performance was estimated by nested cross-validation (n=200, single centre): the algorithm and its hyperparameters were selected only inside training folds, so no test observation influenced model selection. These figures are therefore lower — but more trustworthy — than those from a single favourable train/test split.

Key finding

NAHT response metrics (PSA reduction %, volume reduction %, T-stage downstaging) add statistically significant predictive value beyond baseline characteristics:

—Biochemical recurrence: AUC 0.648 → 0.758 (ΔAUC +0.110, DeLong p = 0.0049)
—Positive margin: AUC 0.612 → 0.759 (ΔAUC +0.147, DeLong p = 0.0005)
—Node positivity: AUC 0.821 → 0.842 (p = 0.39, not significant)

Limitations

—Research tool only. Not a medical device; not validated for clinical decision-making.
—No external validation. Single-centre, retrospective, n=200.
—Pathological complete response (n=6, 3.0%) was too rare to model reliably and is not predicted.

Contact

[drankitjoshiurologist@gmail.com]