Nomogram Predicts Risk of Clinically Significant Prostate Cancer on MRI-Fusion Biopsy of PI-RADS 4 and 5 Lesions

The first validated nomogram that may aid in clinical decision making in conjunction with the Prostate Imaging – Reporting and Data System (PI-RADS) scoring was presented by Ayat A Shah, MBBS, from Roswell Park Comprehensive Cancer Center, Buffalo, NY.1

It has been suggested that certain clinical variables are associated with clinically significant prostate cancer on fusion biopsy of PI-RADS 4 or 5 lesions. Dr Shah and her colleagues aimed to identify these variables and subsequently develop and validate a nomogram to improve the specificity of PI-RADS scoring. They performed a retrospective review of patients over a 4-year period between the years of 2017 and 2021 who had undergone MRI or ultrasound-guided fusion biopsy after having undergone multi-parametric MRI (mpMRI) with findings of either PI-RADs 4 or 5 lesions. Prostate fusion biopsies were done utilizing the UroNav (Philips) system. Based on the pathology results from the prostate biopsy, patients were divided into separate cohorts according to the presence of clinically significant disease (Gleason grade group ≥2 or those with benign or clinically insignificant Gleason Grade Group 1) findings.

Dr Shah presented the results from 348 patients (median age 66 years) with 425 PI-RADS 4 or 5 lesions, of which 239 (56%) patients were PI-RADS 4 and 186 (44%) were PI-RADS 5. After undergoing fusion biopsy, clinically significant disease was shown to be present in 33% of PI-RADS 4 and 67% of PI-RADS 5 lesions.

Compared with patients with clinically significant disease on biopsy, patients with clinically insignificant disease had larger prostate volume on MRI (median 52 vs 42 cm3, P<0.01), lower PSA density (0.15 ng/mL2, P<0.01), multiple PI-RADS 4 or 5 lesions (102 vs 81, P<0.01), prior negative biopsy (53 vs 41, P=0.02), lower PSA velocity (median 0.97 vs 1.42, P<0.01), and longer PSA doubling time (median 5.6 vs 3.8 years, P<0.01). Multivariate analysis revealed that clinically insignificant disease was associated with longer PSA doubling time ≥4 years (odds ratio [OR] 1.71, P=0.02), a prior negative biopsy (OR 2.15, P<0.01), and multiple lesions (OR 2.25, P<0.01), while it was negatively associated with higher PSA density ≥15 ng/mL2 on MRI (OR 0.40, P<0.01) and PI-RADS 5 versus 4 lesions (OR 0.62, P=0.04).

Other factors that had an association with clinically significant prostate cancer included total number of cores (P<0.01), extracapsular extension (p=0.01), neurovascular bundle involvement (p=0.02), maximum diameter of the PI-RADS lesion (p<0.01).

Dr Shah and her colleagues used the data obtained from the multivariate analysis to develop a 100-point nomogram to assess individual risk of clinically significant disease from UroNav fusion biopsy of PI-RADS 4 and 5 lesions. Internal validation of the nomogram using the bootstrapping method, in which the original data set was resampled 500 times, revealed a mean of <0.01, with an area under the receiver operating curve (ROC) of 71% for the nomogram compared with 59% under PI-RADS, alone. “The nomogram proved to be more accurate in determining risks of having clinically significant disease,” Dr Shah said.

Dr Shah and her colleagues believe that their nomogram may improve the specificity of PI-RADS scoring and may also lead to improvement in patient counseling.

References

Houenstein H, Shah AA, Iqbal U, et al. Identifying risk of clinically significant disease from MRI-fusion prostate biopsy of PIRADS 4 and 5 lesions. Poster #184 presented the 22nd Annual Meeting of the Society of Urologic Oncology (SUO), December 103, 2021, Orlando, FL.