The challenges associated with the use of PSA have highlighted the unmet need to develop additional measures to identify clinically significant PCa while reducing unnecessary biopsies, overdiagnosis of low-risk disease, and subsequent overtreatment. These efforts have led to the development of several biomarker tests, most of which incorporate PSA and other existing clinical characteristics, making it difficult to assess biomarker performance.1,15,16. This underscores the importance of tests, such as the EPI-CE, which provide additional independent information to aid in clinical decision making.3.
Until now, the PPE test was only available in the United States as a central lab test. However, the different healthcare systems across Europe and differences in the delivery of clinical diagnostic tests make centralized laboratory testing impractical in European clinical practice. The study presented here demonstrates that the new EPI-CE IVD test provides similar performance to the previously published EPI-LDT, in a cohort of patients primarily from European clinical sites. This study is the first to report on the performance of the EPI-CE test and the first to demonstrate that exosome-based diagnostic tests can be developed to the highest CE-IVD quality standards and made accessible to patients and European clinicians.
Although the prevalence of HGPCa is higher at 35% in the current European cohort versus 30% in previous US cohorts3.4a chi-square test on mild cases, GG1 and ≥ GG2 revealed no significant difference (p= 0.585), with the overall distribution of grade groups being very similar (i.e., no particular GG was responsible for the higher prevalence of HGPCa).
The cutpoint performance measures of EPI-CE observed in the present study were also similar to those previously published for EPI-LDT in the United States.ten using the same threshold of 15.6%, the most important parameters being sensitivity (92% versus 92%, to identify as many cases of HGPCa as possible), NPV (89% versus 90%, which gives confidence that patients with a negative test result may delay biopsy) and the number of patients below the threshold (25% versus 23%, representing a significant fraction of patients benefiting from a delayed biopsy), for EPI- CE versus EPI-LDT, respectively.
The decision curve showed that the EPI-CE provides a higher net benefit than either multivariate risk calculator over a wide range of risk acceptance levels, indicating that using the score EPI-CE to determine who to biopsy would lead to improved clinical outcomes.
Similarly, in the rank-based analysis, EPI-CE had the highest AUC of 0.67, followed by the multivariate risk models ERSPC AUC of 0.60 and PCPT AUC of 0.59 and PSA with AUC of 0.54, although none of these comparisons reached statistical significance. The 95% confidence interval of the EPI-CE AUC was 0.56 to 0.77 in this study, indicating that at the current cohort size, the natural variability of the dataset is well within the 0.70 range that would be expected based on previous publications. EPI performance2,3,4,5. Additionally, as demonstrated in this PPE-CE publication as well as previous PPE publications, performance is consistently superior to existing alternatives, i.e. multi-parameter risk calculators. It is important to note the improved performance of EPI-CE compared to multiparameter models that do not include MRI, especially given the recent PSA screening recommendations of EAU, which call for patients to undergo “risk stratification” before the MRI.1. We deliberately recruited patients in the present study who did not undergo MRI, to estimate the performance of EPI in this setting. Analysis of a subcohort of patients who meet all current EAU recommendations for further risk stratification prior to MRI (N=78, ≥3 ng/mL, 50-70 years) n revealed no difference in performance compared to the EPI-CE intended use population (2–10 ng/mL, ≥50 years), confirming the relevance of the EPI test for pre-MRI risk stratification.
The main limitation of the present study is the limited cohort size, which is partially mitigated by comparison with previously published performance characteristics of the US EPI-LDT version of the test. Additionally, since the study was designed to show the performance of EPI-CE regardless of whether the patient underwent MRI, the lack of MRI data does not allow direct comparisons between EPI- CE and MRI. MRI is an important risk stratification tool for prostate cancer and further studies are needed to show the performance of PEI in various clinical settings compared to MRI.
A major strength of the present study is that it represents a prospective validation, on a clinically relevant European cohort, of the first CE-marked exosome-based urine test for use in clinical laboratories across Europe, adding an important tool in risk stratification – prostate cancer toolkit for patients and physicians.