A urine-based Exosomal gene expression test stratifies risk of high-grade prostate Cancer in men with prior negative prostate biopsy undergoing repeat biopsy

Prostate cancer (PCa) is a leading cause of cancer death among men in the United States, with more than 3.6 million men living with prostate cancer. It is estimated that 174,650 newly diagnosed cases occurred in 2019 [1]. Prostate needle biopsies are typically recommended for men with elevated serum PSA levels and/or a suspicious digital rectal exam (DRE) with added considerations based on family history, age, and race. The anxiety, pain, and potential complications associated with prostate biopsy are well documented [2‐6]. Furthermore, a large percentage of newly diagnosed prostate cancers are indolent, clinically insignificant, and with low metastatic potential. These cancers typically do not require definitive treatment and may be managed most effectively with Active Surveillance (AS). The low specificity of PSA which contributes to the high frequency of newly-diagnosed low-risk PCa suggests that 60–70% of men may be able to avoid biopsy [7‐10]. Compounding the challenge, the majority of initial biopsy tests do not find cancer [11]. For each negative initial biopsy, it is unknown whether elevated PSA levels and/or a suspicious DRE facilitated the biopsy decision or if the biopsy simply missed a cancerous lesion due to sampling error, tumor heterogeneity and multifocality [12].

The European Randomized Study of Screening for Prostate Cancer (ERSPC) demonstrated that the overdiagnosis and overtreatment of PCa resulting from current practice requires more contemporary and improved methods to identify high-grade disease [13, 14]. Novel diagnostics such as the current test that provide additional clinical value to risk stratify high-grade PCa (Gleason Grade, or GG ≥2) have been developed with a primary objective to avoid biopsies for patients with an increased likelihood of having benign or non-aggressive (GG ≤1) disease [15‐17].

The ExoDx Prostate assay relies on the isolation and analysis of urinary exosomes, which are small lipid bilayer membrane extra-cellular microvesicles (typically 30–200 nm in diameter) that are secreted from all living cells. Exosomes contain RNA, DNA, and protein, and have been identified in a variety of biofluids such as blood, cerebrospinal fluid, and urine. They are particularly promising for RNA expression profiling given their protected microanatomic environment [18‐21]. The ExoDx Prostate (EPI) test has been extensively validated in two prospective multi-center US studies for biopsy naive patients and the signature is generated from quantitative analysis of PCA3 (prostate cancer antigen 3), ERG (V-ets erythroblastosis virus E26 oncogene homologs) and SPDEF RNA [16, 17].

The present analysis was designed to evaluate the performance of the ExoDx Prostate gene signature in a cohort of men with a history of prior negative biopsy and now presenting for a repeat biopsy.

Over 1 million transrectal ultrasonography guided (TRUS) prostate biopsies are performed annually in the US and Europe [6]. Although clinical information such as age, race, family history, or a suspicious DRE are triggers, most TRUS biopsies are driven by PSA screening. Biopsies are often associated with pain, bleeding, sepsis and, of significant concern, an increasing rate of antibiotic resistant infection [6, 25]. A significant number (approximately 70% of men) are not found to have prostate cancer on initial biopsy, and this leads to patient anxiety because of the false negative rate as a result of prostatectomy under-sampling and tumor heterogeneity/multifocality. These concerns drive many men to undergo repeat biopsy. In fact, the Surveillance, Epidemiology, and End Results (SEER) data indicates that ~ 12% of men with a prior negative biopsy have a repeat biopsy within 1 year and 44% of men younger than 70 years old have a repeat biopsy [26].

The ExoDx (EPI) Prostate test is a urine-based genomic assay which uses the expression levels of three genes (ERG, PCA3 and SPDEF) involved in PCa initiation and progression, to predict risk of HGPCa, independent of clinical features or standard of care (SOC) [27‐33].. The test algorithm was developed and validated on the intended use cohort, i.e. men presenting for their initial biopsy with a PSA 2–10 ng/mL where it achieved an NPV of 91%, a sensitivity of 91 and 34% specificity [16]. A recent blinded control arm utility study demonstrated that the EPI test reduced the number of biopsies when the test was negative, and helped to identify 30% more high-grade prostate cancers compared to the control arm without the EPI test result [34]. It is well established in the literature that prevalence of prostate cancer is significantly lower in a prior negative biopsy population and that this clinical feature will favorably impact test performance when included in risk assessment models (and some commercial assays) [35, 36]. To address this issue, we evaluated the accuracy of EPI in the prior negative biopsy population using the AUC and compared results with PSA alone and a well-established risk calculator, the European Randomized Study of Screening for Prostate Cancer (ERSPC) [13, 14]. EPI had a superior AUC of 0.66 vs PSA AUC 0.54, and ERSPC AUC 0.47. As EPI is independent of all clinical features it represents an independent biological assessment of prostate cancer risk.

In addition to the AUC we also employed a decision curve analysis which considers overall net health benefit of various biomarkers and risk assessment tools. When EPI was compared with PSA and clinical-only models such as the ERSPC, a higher net benefit was observed at a biopsy threshold probability of 10%, and maintained up to a maximum biopsy rate of 30%. In contemporary practice the decision to proceed with a biopsy is a delicate balance between the risk associated with the biopsy procedure vs. the risk of missing a potentially clinically significant cancer. The benefit for a test such as EPI is that the risk score is a biological construct independent of all clinical factors thereby allowing for combination with other variables such as race, family history and an underlying disease.

In the current prior negative biopsy population, an EPI risk score less than 15.6 would have potentially avoided biopsy in 26% of men based on total number of biopsies in the study (i.e. 229) and a score of less than 20 would have avoided a biopsy in 35% of men without missing additional GG2 or higher PCa. Furthermore, if we equate benign and Grade Group 6 biopsy outcomes as unnecessary biopsies, the < 15.6 and < 20 cut-points would represent 27 and 37% unnecessary biopsies, respectively. Furthermore, as per the 2020 NCCN Early Detection of Prostate Cancer Guidelines [37], it is clinically important to differentiate GG2 vs GG3 PCa which is relevant for the performance of EPI as only 3 patients with GG3 or higher disease were missed at either the 15.6 or the 20 cut-point vs. 6 individuals (3%) with the 29.6 cut-point. The EPI assay also generated a comparable NPV of 92 and 94% in the prior negative biopsy population (cut-points 15.6 and 20, respectively) vs. 91 and 89% in the two prior initial biopsy validation studies [16, 17].

Study limitations include the sample size and lack of a central pathology review, which may have introduced some variability, specifically when reporting small volume cancers and the fact that there were no men in the study who underwent multi-parametric MRI imaging pre-biopsy. During the study period (2014 and 2015), MRI imaging was not standard of care in the USA. Nevertheless, all the study participating centers represent large urology group practices and academic centers with highly experienced uropathologists. A second prospective study in prior negative biopsy men is underway in the US to address the increasing use of mpMRI imaging and fusion biopsy in this population.