PSMA PET/CT-Based Multimodal Model for Predicting Response to First-Line Therapy in mHSPC.

Metastatic Hormone-Sensitive Prostate Cancer, Prostate Cancer

mHSPC, metastatic prostate cancer, PSMA PET/CT, tumor burden, PSMA tumor volume, TL-PSMA, PSA response, first-line therapy

This multicenter retrospective study developed and validated a prediction model based on PSMA PET/CT and routine clinical information to estimate early treatment response in patients with metastatic hormone-sensitive prostate cancer (mHSPC) receiving first-line standard therapy. Existing PSMA PET/CT scans were used to quantify tumor burden, and imaging metrics were combined with baseline clinical factors, including laboratory results and disease characteristics, to build an interpretable model for predicting the likelihood and timing of achieving a deep PSA response (PSA ≤ 0.2 ng/mL) after initiation of first-line treatment. All data were collected from medical records and imaging obtained as part of routine care; no additional tests or treatments were required. The objective was to improve risk stratification and support individualized follow-up and treatment planning for patients with mHSPC.

This multicenter retrospective cohort study was conducted across five tertiary hospitals in China. Patients diagnosed with metastatic hormone-sensitive prostate cancer (mHSPC) between September 2020 and May 2025 were identified from routine clinical practice. Distant metastasis was confirmed by conventional imaging and/or PSMA PET/CT. Treatment was not assigned by the study; all patients received first-line standard therapy as determined by their treating physicians. Baseline PSMA PET/CT scans obtained before treatment initiation were analyzed to derive quantitative imaging features, including measures of whole-body tumor burden such as PSMA PET tumor volume and lesion activity-weighted metrics. These imaging variables were integrated with baseline clinical factors to develop and validate a multimodal prediction model. Model development was performed in one cohort and externally tested in an independent cohort to evaluate generalizability. The primary endpoint was time to deep PSA response (PSA ≤ 0.2 ng/mL), defined as the time from initiation of first-line therapy to the first PSA measurement meeting this threshold. Participants without deep PSA response were censored at the date of the last available PSA measurement. Secondary endpoints included model discrimination and calibration metrics, such as time-dependent AUC and C-index, as well as model interpretability analyses. Key eligibility criteria included confirmed mHSPC, availability of complete baseline PSMA PET/CT imaging and clinical data, and initiation of first-line standard therapy within 3 months of diagnosis. Exclusion criteria included other concurrent malignancies, receipt of other prostate cancer therapies around the time of diagnosis, non-adenocarcinoma histologies, and insufficient follow-up duration. The study used de-identified data extracted from existing records and did not require additional procedures beyond routine care.

No related papers found yet.

China