Biparametric MRI Radiomics Combined with Serum Bone Turnover Biomarkers for Predicting Postoperative Bone Metastasis in Prostate Cancer: Precision Selection for Targeted Radionuclide Therapy
Synopsis
In a prospective single-center cohort of 143 patients with clinically localized prostate cancer (cT1-2N0M0) undergoing laparoscopic radical prostatectomy, the study measured preoperative serum bone metabolism indicators such as osteocalcin N-terminal mid-fragment and alkaline phosphatase, extracted biparametric MRI radiomics variables including apparent diffusion coefficient mean and K trans mean, and integrated serum and imaging biomarkers with a multivariate logistic regression model; the integrated model reached an area under the ROC curve of 0.984, significantly outperforming individual biomarkers and imaging features (all p < 0.001), showed clinical net benefit with strong calibration (Hosmer-Lemeshow test p = 0.
Interpretation
The study proposes and validates a multimodal predictive framework that integrates biparametric MRI radiomics features with serum bone turnover biomarkers to predict postoperative bone metastasis risk in early-stage prostate cancer patients. Whereas prediction often relies on a single information source, this work combines imaging variables reflecting the tumor microenvironment and perfusion characteristics (apparent diffusion coefficient mean, K trans mean) with preoperative serum bone metabolism indicators (osteocalcin N-terminal mid-fragment, alkaline phosphatase) in one logistic regression model. Prospective single-center cohort of 143 patients with clinically localized prostate cancer, evaluated with ROC analysis, calibration curves, decision curve analysis, and Kaplan-Meier survival analysis.
The integrated model showed markedly higher discrimination than individual biomarkers or imaging features alone. It reports an area under the ROC curve of 0.984, with differences from individual indicators all statistically significant (all p < 0.001), indicating that the discrimination gain from combining information was observable in this cohort. Area under the ROC curve served as the main discrimination metric, with p values reported for comparisons against each individual indicator.
The model demonstrated clinical net benefit and strong calibration and separated patients into risk groups with significantly different prognosis. Decision curve analysis showed significant clinical net benefit and the Hosmer-Lemeshow test gave p = 0.830, indicating strong calibration; model-based risk stratification yielded 53 high-risk and 90 low-risk patients with significantly different bone metastasis-free survival (log-rank p < 0.001). Supported jointly by calibration curves, decision curve analysis, and Kaplan-Meier bone metastasis-free survival analysis within the same prospective cohort.
Perspective
The results apply to patients with clinically localized prostate cancer (cT1-2N0M0) undergoing laparoscopic radical prostatectomy who have preoperative biparametric MRI radiomics variables and serum bone turnover biomarker measurements available; its intended use is postoperative bone metastasis risk stratification and early identification of candidates for targeted radionuclide therapy.
This is a single-center prospective cohort with 143 patients, so model performance and risk stratification thresholds still warrant further observation in other populations and multicenter settings; in addition, the available text is abstract-level content without figures or full methodological detail, leaving the radiomics feature selection pipeline, the specific model variables, and follow-up duration as open questions.