Urinary Molecular Biomarker Test and Prostate Biopsy Decision
Urinary Molecular Biomarker Test Impacts Prostate Biopsy Decision Making in Clinical Practice
The Journal of Urology
Neal Shore, Jason Hafron, Timothy Langford, Marshall Stein, Jessica DeHart, Michael Brawer, Daphne Hessels, Jack Schalken, Wim Van Criekinge, Jack Groskopf, and Kirk Wojno
Introduction: There is an unmet need for noninvasive methods to better identify patients at increased risk for clinically significant prostate cancer. SelectMDx® is a molecular urine test validated for the detection of Gleason score 7 and higher cancers (ISUP [International Society of Urological Pathology] Grade Group 2-5). In this multicenter trial we evaluated the test’s impact on prostate biopsy decision making in clinical practice.
Methods: The study involved 5 U.S. community urology practices which sequentially enrolled 418 patients who received a SelectMDx test between May 2016 and April 2017 while undergoing evaluation for initial prostate biopsy. All tests were ordered by the urologist for patient management. We determined biopsy and prostate cancer detection rates in patients with SelectMDx positive versus SelectMDx negative results.
Results: Of the 418 subjects evaluated with SelectMDx 253 (61%) had negative results and 165 (39%) had positive results. Subsequent biopsy rates for SelectMDx positive and negative cases were 60% (99) and 12% (32), respectively (p <0.001). Time from SelectMDx result to biopsy was shorter for those with positive vs negative results (median 2 vs 5 months, p=0.001). Of patients who underwent biopsy within 3 months of testing 71 (43%) with positive results underwent biopsy and 27 had cancers identified, including 10 greater than Grade Group 2. Of 9 patients with SelectMDx negative results (3.6%) who underwent biopsy 4 were diagnosed with cancer, all Grade Group 2 or less.
Conclusions: SelectMDx had a significant impact on initial prostate biopsy decision making. Biopsy rates in SelectMDx positive cases were fivefold higher than in SelectMDx negative cases. These results describe the clinical utility of SelectMDx in real-world community urology practice.
The early detection of prostate cancer is a topic of ongoing debate. There is ample evidence of survival benefit associated with the treatment of intermediate and high risk, early stage prostate cancers. 1–4However, the low specificity of PSA screening has led to excess prostate biopsies and over detection of lower risk cancers, 5,6the treatment of which lacks proven benefit. 7,8Prostate biopsy procedures and interventional cancer treatment may carry significant risks, especially for patients with comorbidities. 9,10There is an unmet need for improved detection of PCa with specific aggressive features, such as Gleason score 7 (GG2) or higher grade, to optimize cancer survival benefit while also avoiding unnecessary prostate biopsies for insignificant, indolent cancers.
SelectMDx is a urine based molecular test that targets mRNA from 2 genes known to be over expressed in aggressive PCa, HOXC6 (cell proliferation gene) and DLX1 (progression gene). 11Urinary HOXC6 and DLX1 mRNA levels are measured following DRE and the RNA results are combined with clinical risk factors to determine the patient’s individualized risk of GG2 or higher PCa histopathology at a subsequent biopsy. 12Analytical validation studies have demonstrated the robustness and reproducibility of SelectMDx results. 13The test was clinically validated for the detection of GS 7, GG2 or higher cancers in a prospective, multicenter study of 905 subjects. 12At a risk score threshold corresponding to a NPV of 98%, sensitivity was 96% and specificity was 53%, indicating that SelectMDx could help avoid approximately half of excess biopsies (ie biopsies that would have identified Gleason score 6 or no cancer) with minimal risk of missing high grade cancers. The cost-effectiveness of SelectMDx in men with an elevated PSA has been assessed in European and U.S. modeling studies. 14,15In both cases use of SelectMDx decreased prostate biopsies (and associated over diagnosis/overtreatment) compared to the current standard of care, which resulted in cost savings and increased quality adjusted life-years per patient. SelectMDx risk scores have also been associated with mpMRI outcome, suggesting that SelectMDx may be helpful for identifying patients most likely to benefit from mpMRI. 16
Clinical utility studies are critical to understanding the potential impact of novel diagnostic tests in a real-world setting. In addition to analytical validity and clinical validity, it is important to determine whether physicians use the results of the test in medical decision making. To our knowledge the clinical utility of SelectMDx for guiding prostate biopsy decisions has not been previously assessed. In this multicenter study we evaluated the effect of the SelectMDx assay on the treatment of men being considered for initial prostate biopsy.
Materials and Methods
Community urology practices with the largest SelectMDx testing volume from May 2016 (earliest commercial availability) to April 2017 were contacted by e-mail in March 2018 and invited to participate in this retrospective study. Five community urology practices agreed to participate. All patients who received a SelectMDx test while under consideration for a possible initial prostate biopsy were eligible. No additional exclusion criteria were used. All SelectMDx tests were ordered by the treating urologist for patient management (ie not as part of a planned or ongoing study) and, thus, the results reflect real-world decision making. Medical record review was conducted in April 2018 to collect followup clinical information. For patients who underwent biopsy, histopathology findings including GS and GG as well as the time between the SelectMDx test and the biopsy were documented. We compared biopsy and cancer detection rates for SelectMDx positive and SelectMDx negative cases with initial prostate biopsy. For statistical analyses the chi-square test was used to assess differences in proportions and the Mann-Whitney test was used for continuous variables.
SelectMDx testing was performed at MDxHealth (Irvine, California). First catch urine is collected following standardized DRE (3 sweeps per prostatic lobe) and then mixed with a preservative to stabilize mRNA for room temperature transport to the laboratory. HOXC6 and DLX1 mRNA are quantified using reverse transcriptase polymerase chain reaction, then normalized for prostate RNA recovery using the level of urinary KLK3 mRNA ( KLK3 is the gene that encodes for PSA). The RNA results are combined in a clinical model with other risk factors including age, serum PSA, DRE (normal/suspicious), family history and prostate volume. The resulting risk score is then converted to the percentage likelihood that subsequent biopsy will identify high grade cancer (GG2 or higher). 12An independent institutional review board reviewed the protocol and provided ethical approval for the study (Western IRB, Puyallup, Washington, No. 1185030).
Five U.S. community urology practices participated in the study and 418 eligible subjects were sequentially enrolled. Table 1 provides an overview of subject characteristics for the entire study population, including those with SelectMDx negative and SelectMDx positive results. For the overall cohort the average age was 67 years (median 67, IQR 62 to 72) and average serum PSA was 5.8 ng/ml (median 5.1, IQR 3.8 to 7.1). Of the 418 subjects enrolled 375 (90%) were Caucasian, 37 (9%) African American and 7 (1%) undisclosed race.
A total of 255 (61%) patients had a negative SelectMDx test result and 163 (39%) had a positive result ( table 1). Compared to men with negative assay results, those with positive results were older (median age 69 vs 66 years), had higher serum PSA (7.1 vs 4.4 ng/ml) and were more likely to have abnormal DRE results (37% vs 5%). Between the SelectMDx positive and negative groups there were no significant differences in clinical monitoring after SelectMDx testing (median 16 months for both), race or percent of men with a family history of prostate cancer.
The biopsy rate for the SelectMDx positive group was 60% (99), compared to 12% (32) for the SelectMDx negative group (p <0.001, table 1). For subjects who underwent biopsy the median time from SelectMDx results report to biopsy was significantly shorter for men with SelectMDx positive vs negative results (median 2 vs 5 months, p=0.001). The correlation between SelectMDx result and interval between SelectMDx testing and biopsy is further illustrated in the figure. More than 40% of patients with SelectMDx positive results (more than 70% of total biopsies for this group) underwent prostate biopsy within 3 months of SelectMDx testing, compared to less than 4% of patients with SelectMDx negative results.
To increase the probability that SelectMDx test results were incorporated into the biopsy decision, we evaluated biopsy and cancer detection rates for subjects who underwent biopsy less than 3 months after receiving SelectMDx results ( table 2). In this subgroup 71 (44%) patients with SelectMDx positive results underwent biopsy and 27 had PCa identified, including 17 GG1 and GG2, and 10 higher grade disease, comprising 3 GS 4+3 (GG3) and 7 GS 8 or greater (GG4 or greater). Among the 9 (3.6%) men with SelectMDx negative results who underwent biopsy within 3 months of test results, 4 cancers (2 GG1 and 2 GG2) were identified.
For most patients with SelectMDx negative results who underwent biopsy, the interval between test results and diagnostic procedure was more than 5 months. Therefore, we evaluated changes in clinical risk factors that may have further influenced the biopsy decision. Followup serum PSA measurements were available for 151 patients, and PSA levels determined before biopsy were higher than the values from the patients with SelectMDx negative results who did not undergo biopsy during the study period (median 5.6 and 3.9 ng/ml, respectively, p=0.001). For patients with positive SelectMDx results there was no difference in PSA values for men who did vs did not undergo biopsy (median 7.1 ng/ml for both groups, p=0.732). SelectMDx percent likelihood values for GG2-GG5 cancers were higher for patients who underwent biopsy (median 22%, IQR 13 to 33) compared to those who did not (median 17%, IQR 14 to 23), although this difference did not reach statistical significance (p=0.057).
To our knowledge this study is the first to evaluate the clinical utility of SelectMDx. We compared the test results with clinical outcome data to determine the impact on biopsy decisions. With a median followup of 16 months from SelectMDx testing, biopsy rates in the SelectMDx positive group were fivefold higher than in the SelectMDx negative group. These results indicate that the test provided actionable information for decision making, as 60% of SelectMDx positive cases involved biopsy and 88% of SelectMDx negative cases did not involve biopsy during the study period. In the subset of patients who underwent biopsy within 3 months of receiving test results, 27 of 71 biopsies performed in the SelectMDx positive group were PCa positive, including 10 GG3 or higher disease. In the SelectMDx negative group cancers were identified in 4 of 9 men who underwent biopsy within 3 months of testing, and all were low/intermediate grade (GG1, GG2). In addition, biopsies performed in men with SelectMDx negative results tended to occur at least several months after testing. This provides sufficient time to detect changes in clinical risk factors, and we found that serum PSA levels were elevated relative to index values and/or levels found in men with SelectMDx negative results who did not undergo biopsy during the study period. This finding may help explain some of the biopsies performed in men in the SelectMDx negative group, and underscores the fact that these men should be actively monitored. For patients with SelectMDx positive results there was a trend toward higher percent likelihood values in men who underwent biopsy compared to those who did not.
Although retrospective, this study provided a stringent test of SelectMDx clinical utility. All data were obtained from community urology practices during a time frame when the test had first become commercially available and tests were ordered for patient care with no specific restrictions. Therefore, the study could be viewed as an optimal assessment of SelectMDx for biopsy decision making in routine clinical practice. In addition, the results are complementary to the recently published U.S. health economic study that showed potential cost savings of roughly $1,600 per patient and quality adjusted life-years increase compared to the standard of care biopsy procedure. 15Thus, the test could help avoid potentially unnecessary biopsies while increasing the selection rate for identification of clinically significant cancer at biopsy.
Several options are available to help improve initial biopsy decision making. mpMRI is increasingly used for PCa detection. However, a recent meta-analysis highlighted mpMRI variability, 17with reported NPVs of 64% to 88%, which are lower than SelectMDx (98% NPV for detection of GG2 or higher PCa). 12In addition, the American Urological Association currently recommends the use of mpMRI only in the repeat biopsy setting. 18Blood based PSA derivative assays including the Prostate Health Index (phi) and 4Kscore® have been shown to improve the prediction of malignancy for initial biopsy outcome. 19Although head-to-head studies have yet to be performed (ie testing performed on same subjects), SelectMDx has demonstrated the greatest accuracy for the detection of GG2 or higher PCa, with a sensitivity and specificity of 96% and 53%, respectively. 12
The limitations of this study include the fact that it was retrospective, involved only 5 sites and was early in the clinical experience of the investigators (ie the time frame when the test was first made available to clinicians for patient management). Strengths include the “after the fact” decision to examine the effect of SelectMDx test results on biopsy decision making. This obviates a potential bias in clinical decision making that participating in a traditional prospective utility study may engender.
In this multicenter U.S. community urology based utility study SelectMDx had a significant impact on initial prostate biopsy decision making. Biopsy rates in men with SelectMDx positive results were fivefold higher than in men with SelectMDx negative results. In the subset of patients who underwent biopsy within 3 months of receiving test results, high grade cancers (GG3 or higher) were found only in the SelectMDx positive group, and men in the SelectMDx negative group who underwent biopsy within 3 months of testing were diagnosed with GG2 or lower grade disease. These results provide evidence that SelectMDx can help identify men with high grade cancer who would benefit from prostate biopsy and subsequent treatment options. Implementation of this biomarker test in a routine clinical pathway could also lead to a decrease in potentially unnecessary diagnostic procedures and treatments. In summary, this clinical utility study shows the value of SelectMDx for biopsy decision making in real-world clinical practice.
1. Crawford ED, Moul JW, Rove KO et al: Prostate-specific antigen 1.5-4.0 ng/mL: a diagnostic challenge and danger zone. BJU Int 2011; 108: 1743.
2. Crawford ED, Rosenberg MT, Partin AW et al: An approach using PSA levels of 1.5 ng/mL as the cutoff for prostate cancer screening in primary care. Urology 2016; 96: 116.