The Sentinel™ PCC4 Assay is a new molecular classification technology that provides an assessment of risk of advanced disease based on the expression of a cohort of 442 small non-coding RNAs.
At the start of the study, miR Scientific will schedule a teleconference call with all of the participating physicians in Puerto Rico (San Juan Bautista Hospital) to outline the study, the basis for the test and to explain the differences between this Sentinel™ PCC4 Assay and the calculators commonly used in risk assessment for prostate cancer (the CAPRA and/or D'Amico risk calculators and any other risk assessment methodology the uro-oncologists utilize on a routine basis).
After the clinical data (see CRF form) and Sentinel™ result have been compiled for each of the first 50 patients who meet the inclusion/exclusion criteria, the CSO and/or Medical Director of miR Scientific will chair one or more teleconference call with the participating physicians to compare the results of the Sentinel™ PCC4 Assay and the CAPRA and other risk assessment calculators, discuss the relative strengths and weaknesses of the tests and answer any questions regarding the performance of the Sentinel™ Assay and discordances (if any) between the test results. It is important to note that while the clinical data and the Sentinel Score are affiliated with specific barcodes, these data are fully anonymized and cannot be associated with specific patients. In addition to preserving the PHI confidentiality, this also ensures that the information related to the Sentinel™ PCC4 Score will not influence the current standard of care for these patients.
At the end of each session, the participating physicians will complete a questionnaire designed to evaluate the quality of the training and information related to the test result (see appendix 1). This process will be repeated for every 50 new patients enrolled for the first 200 patients and then again after the data for 350 patients becomes available. The analysis of the complete data set for the 500-patient cohort will also be presented at the completion of the study, and a final questionnaire will be administered.
To validate the performance of the miR Sentinel™ PCC4 Assay for identifying patients with prostate cancer in men with suspicion of prostate cancer for whom a core-needle biopsy is performed, the miR Sentinel™ PCC4 results with low-, intermediate- and high-risk prostate cancer (histopathological Grade Group 1, 2 or 3-5) will be compared to the results of the CAPRA score. The sensitivity, specificity, positive and negative predictive values (PPV, NPV) will be established.
This clinical utility study is an observational, validation study. Patients with suspicion of prostate cancer and meet the inclusion/exclusion criteria listed below will be enrolled to provide a urine sample during their office visit prior to core-needle biopsy. The Pathological Grade Group system including the results of each biopsy will be grouped as either no cancer, CAPRA low-, intermediate- or high-risk by the pathologists at the participating centers.
The performance of the miR Sentinel™ PCC4 Assay to classify the disease status of a future patient as either prostate cancer or no prostate cancer and for patients with prostate cancer as having low-risk (nominally Grade Group 1), intermediate-risk prostate cancer (nominally Grade Group 2) or high-risk prostate cancer (nominally Grade Group 3-5) will be established. The test reports will be provided to the consulting urologists who will also have the clinical report associated with enrolled subjects.
The discordances between core needle biopsies and the Sentinel™ PCC4 Assay are well documented in Wang et al. (2020), see table below. At the start of the study the urologists will be briefed on the interpretation of the Sentinel™ PCC4 Assay, and will subsequently be asked to compare and contrast the information provided by the Sentinel™ PCC4 and the SOC (PSA +/- DRE and/or pathology report), using a validated questionnaire (see attached instrument appendix 1). To ensure that urologists associated with the project understand the Sentinel™ PCC4 platform, a review of the comparative results will be held when the first 50 patients have been accrued given that more than 50% of the patients (\> 25 patients) had undergone confirmatory core needle biopsy, and again for every additional 50 patients (regardless of percentage of patients with biopsy) until 200 patients are accrued, and when 350 and 500 patients have been accrued, at which time the trial will have completed its enrollment. The survey instrument will be administered after each review to determine whether familiarity with the Sentinel™ PCC4 Assay imbues confidence in the risk assessment.
Statistical Analysis:
Definitions. The general notation P (A \| B) = P(A , B)/P(B) is the probability of the event A given that the event B occurred; it is the probability (prevalence) of the event A among the subgroup of patients in a population encompassing all who have the event B.
1. Sensitivity = P (screening test + \| truly disease).
2. Specificity = P (screening test - \| truly no disease).
3. PPV = P (truly disease \| screening test +).
4. NPV = P (truly no disease \| screening test -).
Estimate of Sample Size. The proposed prospective study will serve to establish the performance of the miR Sentinel™ PCC4 Assay on a representative sample of patients with any suspicion of prostate cancer for whom a core needle biopsy is ordered. The miR Sentinel™ PCC4 Assay is trained on retrospective data as described above.
The current proposed study is targeted to enroll 500 subjects who visited urology clinics due to suspicions of prostate cancer for clinical workup as part of the SOC and who provided a urine sample during an office visit prior to biopsy.
For any subgroup with a minimum of 100 enrolled participants, if the observed (empirical) sensitivity of the PCC4 Assay for that subgroup is 95%, then the one-sided upper 95% confidence interval for the true population sensitivity will be 90% or greater.
This prospective study will continue to enroll participants until the first of either 500 total participants are enrolled, or at least 100 participants have enrolled in each of the four groups: NEPC; Grade Group 1; Grade Group 2; Grade Group 3 and above.
Statistical Methodology. As outlined below, this clinical study will enroll patients visiting urology clinics for clinical workup based on any suspicion of prostate cancer. The study will evaluate the properties of the Sentinel™ PCC4 Assay that is based on use of a proprietary Classification Algorithm to identify future patients with prostate cancer and to classify the risk of prostate cancer patients.
The proprietary Classification Algorithm employed controls sensitivity for both identifying patients with prostate cancer and identifying high-risk prostate cancer at, or above, a pre-specified level, denoted 1-α; for example, the value that has been assumed in this design is α=0.05 so that sensitivity is at least 95% in the population. Note that the value of α represents the false-negative rate of the assay.
To describe how the cut-offs for the Sentinel™ PCC4 Assay are calculated to control sensitivity, for each participant in the training dataset, the Sentinel™ PCC4 Assay result will be calculated using the remaining members of the training dataset and only the current participant's sncRNA sequence; that is, the true disease status of each participant in the training dataset will be blinded, thereby mimicking the setting for classification of a future patient. The tuning parameters used in the miR Sentinel™ PCC4 Assay are then calculated so that the empirical sensitivity over participants in the training dataset with prostate cancer and with high-grade prostate cancer correspond to the value that provides an upper one-sided 95% confidence interval for the corresponding population sensitivity for a future patient of at least 1-α.
With these cut-offs for the Sentinel™ PCC4 Assay determined a priori from the training dataset, the corresponding values of sensitivity, specificity, positive and negative predictive values will be calculated, along with a corresponding upper 95% confidence intervals, on the prospective patient data accrued in this proposed study, with each biopsy result blinded, that is using only the participant's sncRNA sequence. Note that these error rates refer to the classification of a future patient with unknown disease status.
The statistical analysis of the prospective data accrued that blinds the group label from biopsy will provide an estimate of the proportion of patients that are mis-labeled by biopsy in the prospective dataset. For example, patients labeled as having no prostate cancer by biopsy and whose blinded miR Sentinel™ PCC4 Assay result indicates prostate cancer may be a patient that was potentially a false negative by core-needle biopsy (biopsy label negative). The validation of this discordance and others are addressed below in the follow-up data analysis section. For secondary analysis of the prospective data accrued in the "prostate biopsy decline" subject cohort, the clinical impact of Sentinel™ PCC4 classification of low-risk, intermediate-risk, and high-risk prostate cancer will be evaluated based on established performance data from Primary Study Objective #2 and uro-oncologist feedback.
Follow-up Data Analysis. If a follow-up biopsy is scheduled for participants with negative biopsies as part of their standard clinical care, the follow-up Sentinel Assay and biopsy results will be deemed as the clinical truth for the outcome analysis.
