Developing a Multimodal Model for Detecting Higher-Grade Prostate Cancer Using Biomarkers and Risk Factors

Area of Science:

• Oncology
• Molecular Diagnostics
• Biostatistics

Background:

• Accurate prediction of clinically significant prostate cancer (PC) is crucial to minimize diagnostic errors and overdiagnosis.
• Existing risk assessment tools often lack the precision needed for definitive clinical decision-making.
• The integration of novel biomarkers with established clinical factors holds promise for enhanced diagnostic capabilities.

Purpose of the Study:

• To develop and validate a multimodal model for identifying individuals with high-grade PC from prostatic biopsies.
• To incorporate messenger RNA (mRNA) indicators and conventional risk variables into a predictive tool.
• To assess the clinical utility and cost-effectiveness of the developed multimodal model.

Main Methods:

• Two prospective multimodal investigations were conducted, collecting urinary samples for mRNA analysis.
• A multimodal risk score was developed in a cohort of 489 patients and validated in a separate cohort of 283 patients.
• Messenger RNA (mRNA) levels were determined using reverse transcription qualitative polymerase chain reaction; logistic regression was used for risk prediction, and model performance was evaluated using the area under the curve (AUC) and decision curve analysis (DCA).

Main Results:

• Specific mRNA markers, including sixth homeobox clustering and first distal-less homeobox, were highly predictive of high-grade PC.
• The multimodal model achieved an AUC of 0.90, with key predictors being mRNA features, PSA density, and prior negative tests.
• An additional model incorporating digital rectal examination (DRE) achieved an AUC of 0.86, with both models showing strong validation results and significant clinical utility via DCA.

Conclusions:

• The developed multimodal model, integrating mRNA biomarkers and clinical data, demonstrates high accuracy in predicting high-grade prostate cancer.
• This approach offers a significant improvement over traditional methods, enhancing diagnostic precision and potentially reducing healthcare costs.
• The findings support the clinical utility of this novel technique for improved prostate cancer risk stratification.