Abstract | March 24, 2024

Lesion Density is a Novel Pre-Biopsy Predictor of Clinically Significant Prostate Cancer Detection

Tivoli Nguyen, Urology, MS2, Tulane University School of Medicine, New Orleans, LA

Christine Lightfoot, Urology, MS2, Tulane University School of Medicine, New Orleans, LA; Jacob Greenberg, MD, Urology, PGY1, Tulane University School of Medicine, New Orleans, LA; Garrett Brinkley, MD, Urology, PGY2, Tulane University School of Medicine, New Orleans, LA; L. Spencer Krane, MD, Chief of Urology, Urology, New Orleans VA Medical Center, New Orleans, LA

Learning Objectives

  1. Identify the limitations of PIRADS for predicting csPCa detection
  2. Describe the consequences of overdiagnosis of low-risk prostate cancer
  3. Discuss the clinical utility of lesion density as a new pre-biopsy predictor of sucessful csPCa detection

Background/Knowledge Gap: The likelihood that a suspicious lesion on MRI is associated with clinically significant prostate cancer (csPCa) is currently assessed using the Prostate Imaging Reporting and Data System (PIRADS). However, there is disagreement over whether it is useful to biopsy particular lesions due to inconsistent diagnostic sensitivity. Patient selection for MRI-transrectal ultrasound (MRI-TRUS)-guided biopsy can be assisted by other supporting MRI predictors, thereby reducing the number of unnecessary biopsies. The aim of this study is to investigate whether lesion density, defined as lesion size divided by overall prostate volume, is clinically useful as a pre-biopsy predictor of csPCa detection.

Methods/Design: A prospective chart review study was completed between October 2017 and June 2023. This study enrolled men who underwent MRI-TRUS biopsy to identify csPCa, defined as Gleason Grade ≥ 2. Statistical analysis and generation figures was performed through R. All tests were two-sided with significance bar set to 0.05.

Results/Finding: 753 total MRI-TRUS biopsies were performed. 313 (41.57%) yielded csPCa. 479 patients were African American (63.6%). The median BMI, age, PSA, and TRUS volume were 28.4 (25.3-32.1), 68 (IQR 63-72), 6.04 (IQR 4.52-8.94), and 43.0 (IQR 32.0-61.6), respectively. Patients had 11(1.5%) PIRADS 2, 237(32.2%) PIRADS 3, 318(43.2%) PIRADS 4, and 237(32.2%) PIRADS 5 lesions. An ROC curve was generated to determine the accuracy of lesion density for predicting csPCa showing AUC = 0.66 (95% CI 0.61-0.66) and a cutoff value of 0.0174. For lesion densities above the cutoff, a higher percentage of clinically significant cases was detected (54.9% [359 cases] vs 29.4% [309 cases]). When compared to PSA and PIRADS, lesion density was a statistically better predictor of csPCA than PSA (p=0.0012) and comparable to PIRADS (p=0.86).

Conclusions/Implications: Multiparametric MRI lesion density with a cutoff determined to be 0.0174 was found to be a useful pre-biopsy predictor of csPCa detection. In fact, lesion density performed better at predicting csPCa than PSA and was comparable to PIRADS. Applying this to a clinical setting, lesion density can assist patient selection by differentiating patients to reduce the number of unnecessary biopsies that lead to overdiagnosis of low-grade prostate cancer.

References and Resources

N/A

Back to Abstracts