As the incidence of prostate cancer has risen, clinicians have sought to improve treatment strategies that minimize the impact on quality of life while optimizing oncologic outcomes for those who require treatment.¹ Focal therapy has emerged as an attractive treatment option for both patients and providers, offering potential oncologic control with lesser effects on quality of life than whole gland treatment.² However, with the rise in focal therapy, pitfalls have been identified related in part to the difficulty in patient selection and target localization.³

Diagnostic imaging has improved dramatically over the past decade. MRI has become a standard in diagnosis and staging due to the established superiority of MRI-guided biopsy over transrectal ultrasound–guided biopsy alone.⁴ Additionally, the use of prostate-specific membrane antigen (PSMA) positron emission tomography (PET) imaging has increased, and is now recommended in those with high-risk disease at diagnosis to exclude regional or metastatic disease.^5,6 However, in men with localized disease, PSMA PET has been shown to have the additive benefit of identification of tumor burden when combined with MRI. Grading systems for PSMA PET have been developed to standardize reporting moving forward.^7,8

We aimed to determine whether PSMA imaging could reliably identify the laterality of disease, in the hope that it could assist in patient selection for focal therapy, potentially identifying those patients with occult bilateral disease not identified on MRI.

We included men who underwent radical prostatectomy (RP) at the University of California, San Francisco (UCSF), with localized disease defined as preoperative T stage, 1-3a, no cN1 or cM1. Participants had prostate MRI and PSMA imaging performed preoperatively and had positive findings on PSMA PET imaging. We compared MRI and PSMA to the final RP specimen as the gold standard.

We defined exact match as concurrent laterality on imaging and the final pathology specimen. A partial match was defined as unilateral imaging positivity with ipsilateral GG2 or higher disease and contralateral GG1 on RP. These patients were considered appropriate for focal therapy. Patients with discordance in laterality between the dominant lesion on imaging and on RP were considered poor candidates for focal therapy. Chi-squared analysis was used to compare imaging performance across grade groups. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated for good vs poor candidates for focal ablation based on MRI alone, PET alone, and combined MRI+PET compared to RP.

A total of 107 patients had positive PSMA PET imaging and MRI prior to RP and were included for analysis. The median age at diagnosis was 67 years (IQR 62–71), and the median PSA at diagnosis was 11.23 (ng/mL; IQR 6.50–18.60). At final pathology at RP, 39% of patients (42/107) were GG4-5, corresponding with 49% of patients who had high-risk surgical CAPRA-S scores.

When comparing MRI laterality to RP specimens, we found 42 patients (39%) with no or poor match, 9 patients (8%) with a partial match, and 56 patients (52%) with an exact match. When comparing uptake on PSMA PET imaging with RP specimens, we found 23 patients (21%) with no or poor match, 6 patients (6%) with a partial match, and 78 patients (73%) with an exact match (Table 1). Patients with partial or exact matches were considered good candidates for focal therapy.

Table 1. Concordance Between MRI Lesion and Prostate-Specific Membrane Antigen Uptake and Laterality of Disease on Final Radical Prostatectomy Specimen (n = 107)

When evaluating the degree of concordance across grade groups, there was no significant difference in MRI or PSMA’s ability to evaluate laterality across GG on biopsy or final prostatectomy specimen.

When looking at predicting candidacy for focality on imaging (good vs poor), compared to the final RP specimen, MRI had a better sensitivity (87% vs 61%) and negative predictive value (86% vs 75%) compared to PSMA PET. PSMA PET performed better than MRI in specificity (67% vs 44%) and positive predictive value (50% vs 46%). When MRI and PSMA PET were evaluated together, specificity (73%) and positive predictive value (53%) were higher than either individual modality (Table 2).

Table 2. Sensitivity, Specificity, Positive Predictive Value, Negative Predictive Value of MRI, Prostate-Specific Membrane Antigen Positron Emission Tomography, and MRI + Positron Emission Tomography as Compared to Radical Prostatectomy Specimen

In the intermediate- and high-risk patient population, we found that PSMA PET added to the specificity and positive predictive value of MRI in determining viable candidates for focal therapy, thereby allowing us to more confidently identify patients with true pathologic unilateral disease. In our study, which did not include a radiology re-review of PSMA PET imaging for dedicated laterality determination, neither modality performed particularly well in isolation. Consequently, a specificity of 73% for combined MRI–PSMA PET imaging is encouraging, compared to MRI (44%) and PSMA-PET (67%) separately.

The cohort for this analysis underwent PSMA PET/CT as well as PSMA PET/MRI. As PSMA PET/CT is the preferred imaging modality of the 2 for lesion visualization, the presence of PSMA PET/MRI in our cohort may decrease the perceived accuracy of PSMA PET, a potential limitation of this study. Additionally, we elected to evaluate how PSMA PET performed independent of dedicated MRI imaging, postulating that the larger gross tumor volume seen on PSMA PET compared to MRI and the improved performance of PSMA PET in disease identification over MRI found in prior studies would correlate into a clinically meaningful difference in PSMA PET ability to accurately stage laterality.^9,10 While we failed to prove this hypothesis in a small cohort of 107 men, we feel this topic is deserving of further investigation. As PSMA PET becomes increasingly available in the localized disease state, we must work to gain important information on how it can best be utilized. We believe that evidence such as ours highlights the need for prospective studies of PSMA PET in the focal-therapy-eligible population, intermediate-risk patients who previously were unlikely to undergo PSMA PET unless in a research setting.

Declarations: All authors declare no conflicts of interest regarding the material presented in this manuscript and approved the submission of the final article.

Funding: UCSF Goldberg-Benioff Program in Translational Cancer Biology.