Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-1964
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-20T01:10:33.263781Z
status: published
lens: g_space
sector: health
headline: Machine Learning Models Combining [<sup>68</sup>Ga]Ga-PSMA-11 PET Radiomics and Clinical Features to Predict Overall Survival for Metastatic Castration-resistant Prostate Cancer Treated with [<sup>177</sup>Lu]Lu-PSMA Radioligand Therapy
dek: Purpose Metastatic castration-resistant prostate cancer (mCRPC) remains incurable despite recent therapeutic advances. [ 177 Lu]Lu-PSMA RLT (lutetium-177 labelled prostate-specific membrane antigen radioligand therapy) provides meaningful clinical benefit in selected patients. However, ∼30% of the patients do not respond to [ 177 Lu]Lu-PSMA RLT. Although several baseline prognostic factors are well known, reliable pretherapeutic predictive biomarkers are still lacking. We developed and compared several predictiv…
gain_title: Multimodal machine learning models using pretreatment PET radiomics plus clinical data predicted 12-month survival in mCRPC patients on Lu-PSMA therapy, with a best AUC of 0.778, which could help stratify patients.
problem_title: (none)
trace_subject: (none)
gain_reading: Multimodal machine learning models using pretreatment PET radiomics plus clinical data predicted 12-month survival in mCRPC patients on Lu-PSMA therapy, with a best AUC of 0.778, which could help stratify patients.
gain_evidence: The best-performing ML model achieved a 12-mo area under the curve (AUC) of 0.778 | The model leveraging PET data alone achieved strong performance, with an AUC comparable to the clinical model (0.715 vs 0.766)
problem_reading: (none)
problem_evidence: (none)
quick_read: Researchers built and compared machine learning models to estimate overall survival in 102 men with metastatic castration-resistant prostate cancer who received lutetium-177 PSMA radioligand therapy between January 2022 and March 2024. Models drew on clinical, biochemical, genomic and imaging inputs, including radiomic features from pretherapy gallium-68 PSMA-11 PET/CT scans.

The work matters because about 30% of patients do not respond to this therapy and reliable pretreatment predictive biomarkers are lacking. A PET-only model performing close to a clinical model suggests imaging could support patient stratification. Evidence comes from a single-center retrospective cohort, so how well it generalizes remains unestablished.
limitation: The study was retrospective and conducted at a single center with 102 patients, which limits how far the findings can be generalized.
tag: Evidence-backed gain
key_points: The cohort had 102 patients with mCRPC treated with Lu-PSMA between January 2022 and March 2024. | Five lesions per patient were segmented on pretherapy PET/CT scans, and radiomic features followed Image Biomarker Standardization Initiative guidelines. | Median overall survival of the whole cohort was 11.2 months. | Higher intralesional homogeneity, measured by normalized inverse difference moment, was associated with shorter survival (p = 0.025). | The best model combined 22 clinical, biochemical, genomic and radiomic features to reach a 12-month AUC of 0.778.
rundown: Metastatic castration-resistant prostate cancer remains incurable despite recent advances. Lu-PSMA radioligand therapy gives meaningful benefit to selected patients, but roughly 30% do not respond, and known baseline prognostic factors have not yielded reliable predictive biomarkers before treatment.

The authors segmented up to five lesions per patient on the pretherapy scans and extracted standardized radiomic features. They then trained several multimodal models to predict overall survival, comparing clinical-only, PET-only and combined approaches.

The PET-only model reached an AUC of 0.715 versus 0.766 for the clinical model, while the best combined model reached 0.778. The authors describe PET-based models as a noninvasive, standardized and widely accessible approach for stratifying patients in real-world practice.
sources:
- peer_reviewed | European Urology Oncology | https://doi.org/10.1016/j.euo.2026.09.015 | 2026-10-15
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

345e7c81d27e1fcfd568a0ab0bcb914f68f7a4a5daa1a0719d96b75e57a57624

previous

0000000000000000000000000000000000000000000000000000000000000000