Artificial Intelligence Enhances Quality of Physician Communication in Prostate Cancer Consultations

Researchers from Cedars-Sinai Medical Center have developed natural language processing (NLP) models to assess the quality of physician communication in prostate cancer consultations. The study, published in Prostate Cancer and Prostatic Diseases, aimed to identify key concepts discussed during consultations and evaluate the accuracy of physician communication. The research team used 50 consultation transcripts to train and validate NLP models, which successfully identified topic-concordant sentences with high accuracy. The models were tested in 20 separate consultations, demonstrating the potential for scalable quality assessment of risk communication in prostate cancer care.

Key Takeaways:

• Researchers developed NLP models to assess the quality of physician communication in prostate cancer consultations.
• The models were trained on 28,927 total sentences from 50 consultation transcripts and validated with an accuracy of 0.98.
• The models identified key concepts discussed during consultations, including cancer severity, life expectancy, cancer prognosis, baseline urinary and erectile function, and treatment side effects.
• The models demonstrated high accuracy in identifying topic-concordant sentences, with area under the curve (AUC) values ranging from 0.84 to 0.99.
• The models were tested in 20 separate consultations with high accuracy, demonstrating the potential for scalable quality assessment of risk communication in prostate cancer care.
• The study highlights the importance of accurate communication between physicians and patients in prostate cancer care.
• The NLP models may provide a useful tool for evaluating the quality of physician communication in prostate cancer consultations.

Statistics:

• 28,927 total sentences used to train and validate NLP models.
• 50 consultation transcripts used to train and validate NLP models.
• 20 separate consultations tested with high accuracy.
• AUC values ranging from 0.84 to 0.99 for identifying topic-concordant sentences.
• 100%, 90%, 95%, 95%, 80%, 95%, 85%, 100%, and 95% accuracy in identifying key concepts in separate consultations.

Sources:

• Development and validation of a natural language processing system to assess quality of physician communication in prostate cancer consultations. Prostate Cancer and Prostatic Diseases, 2025.
• NewsRx. Findings from Cedars-Sinai Medical Center Provides New Data about Erectile Dysfunction (Development and validation of a natural language processing system to assess quality of physician communication in prostate cancer consultations). Cancer Weekly. September 16, 2025; p 510.