In a recent study presented at the 2021 AUA Annual Meeting, Laurence Klotz, MD, FRCSC, and a team of investigators determined the accuracy of the new miR Sentinel PCC4 Test, a urinalysis of low- to high-grade prostate cancer in men.
Although radiographic imaging and biopsy have been the top methods for diagnosing prostate cancer for many years, investigators are still trying to explore and innovate more noninvasive tests that maximize accuracy.
In a recent study presented at the 2021 American Urological Association Annual Meeting,¹ Laurence Klotz, MD, FRCSC, and a team of investigators determined the accuracy of the new miR Sentinel PCC4 Test, a uranalysis of low- to high-grade prostate cancer in men. Klotz is a professor of surgery at the University of Toronto, and the chair of prostate cancer research at the Sunnybrook Health Sciences Centre, Ontario, Canada.
The background is really the Sentinel platform. This is a new assay that's based on analysis of 442 individual microRNA sequences derived from urinary exosomes. To make a long story short, exosomes are micro vesicles that are secreted in the thousands by individual cells, both benign and malignant. And these exosomes contain important biological material that they're packaged with—microRNA, which is noncoding RNA, and they're much shorter segments than regular messenger RNA. The result is that they're quite stable. A great deal of research has gone into demonstrating over the last decade that these microRNAs are, to some degree, specific for cancer. And they're functional, so they often serve as regulatory molecules or as silencing molecules.
Several studies have demonstrated that individual microRNAs are predictive for, for example, the presence or absence of prostate cancer. The unique thing about the power of this assay is it uses 442 individual sequences. Each one of these individual micro RNAs has been demonstrated to be predictive for the presence of or absence of cancer, either low-grade or higher-grade cancer. By combining the predictive value of these 442 sequences, it really gives the assay a very extraordinary high level of sensitivity and specificity. The way the basic assay works is that it's based on pre-DRE urine. The requirements for storing and shipping are relatively modest; it just needs to be sent on ice. It's also a fairly short turnover to do the assay.
The essay is iterative in 3 steps. The steps vary according to how the individual sequences are weighted, so it's the same set of sequences that are analyzed. Step 1 is for the presence or absence of any cancer from grade group 1 to 5. Step 2 is for the presence of grade 1 versus grade 2 to 5 in those patients who are found to have cancer. Step 3 is for grade group 1 and 2 versus 3, 4, and 5. Again, it's not that the sequences are evaluated 3 times. It's that the weighting that is placed on the sequences is varied for each one of those iterative steps. So, what was shown in this presentation was the updated validation data. This data is based on 763 patients at risk for prostate cancer who are having a biopsy, most of whom had systematic biopsies because this study was begun before the era of routine MRI before biopsy. A proportion of them had systematic and targeted biopsy, and some of that data turns out to be quite important in terms of interpreting the data.
Obviously, there's a lot of assays now—at least 9 or 10—based on either blood or urine or tissue, that purport to identify whether significant cancer is present. Typically for those assays, the area under the curve varies between about 0.7 to 0.9. In the initial validation studies that we published, the area under the curve for this assay was more like 0.96 to 0.97. So, we think the advantage of this assay is that it's more accurate in terms of sensitivity and specificity. The other appeal is that it's based on urine. It's based on pre-DRE urine, so there's really no need for the patient to see a doctor to have an examination of his prostate before he gives the specimen. They can even give the specimens at home. So, we think given the accuracy that this is potentially a game changer.
Just to summarize the key findings, the sensitivity for clinically significant cancer was 93%. The total of patients who had significant cancer was 220, and 16 of them had what we call “negative molecular evidence” for prostate cancer, or the assay suggested low molecular evidence of cancer. One important point is that there's not a direct correlation between the readout of this assay and the pathology because the assay is really giving us a molecular characterization of the cancer. So, for example, there are grade group 2 cancers that are clearly indolent. We think it's very plausible that the molecular characterization of these cancers using this Sentinel platform might give a negative readout, indicating that it's actually an indolent cancer despite the presence of some grade group 4. That's something to bear in mind when you're looking at the results.
The sensitivity was 93%, the false negative was 7%, there were 4% of significant cancers that were misclassified as no evidence of cancer at all, and the negative predictive value for the not clinically significant cancer being present was 96%. Finally, amongst the patients whose biopsy was negative, 32% had molecular evidence for significant cancer, so that was essentially 172 of 543 patients. That is roughly the same rate at which you would expect a systematic biopsy to undergrade or to miss the presence of significant cancer. So really, the question was, is this a false positive of the test or is it a false negative of the biopsy? To get a handle on that, we looked at the men who had initially a negative systematic biopsy and then underwent an MRI targeted biopsy. There were 108 of these patients in the study. Of those, in about three quarters, the systematic and targeted biopsies were both negative. But in 29 of the 108 patients, which was 27%, the systematic biopsy was negative, and the targeted biopsy was positive. When we went and looked at the concordance between those patients the molecular characterization of the cancer, 100% of the patients whose targeted biopsy was positive for a grade group 2 or higher were predicted by the Sentinel platform. There were essentially no cases where the grade group 2 was found on the targeted biopsy and was missed on the molecular assay. And then of the grade group 1 cases on the targeted biopsy, 10 of the 14 had intermediate to high molecular risk classification. That also suggests that there may have been a pathologic miss with some of those targeted biopsies.
There's really no way to know, and I emphasize that it's a preliminary analysis with a small subset of cases. But overall, what it suggests is extraordinarily high negative predictive value for significant cancer, and in the roughly one third of cases where there was an apparent false positive of the test, a lot of those were actually a false negative of the biopsy. If the patient had a targeted biopsy or a radical prostatectomy, the higher grade cancer would have been present. But again, that is somewhat speculation that is supported by the close concordance between the targeted biopsy results and the Sentinel assay readout.
There's a tremendous amount of research going on. There's several steps to this. One is further validation, so there's an ongoing study that ultimately will involve 4000 patients having a biopsy to really nail down with a high level of confidence, the sensitivity and specificity for the biopsy outcome. We're also launching a study of the performance of the assay as a dynamic assay, meaning in men on surveillance, how does the assay perform over time? Does it identify those men who eventually have grade progression to higher grade cancer? And so, all of those are ongoing at the moment, and essentially, we are waiting for FDA approval. We think that will be forthcoming in the fairly near future, but of course, we can't predict that with certainty.
Stay tuned. An assay that is really accurate and that can really tell you with a high degree of confidence that the patient either doesn't have cancer or has low-grade, intermediate, or high-grade cancer would profoundly and favorably influence how we manage patients. It would allow us to reduce the amount of imaging and biopsies that we would do and it would allow us to manage patients on surveillance with a higher degree of confidence that we weren't missing anything. I think this is coming and it potentially could really influence how we manage these patients. But it's still early days as far as the validation studies are concerned.
The company is miR scientific. They are ramping up in a big way. This is a very appealing technology because it's short turnover. The assay can be done in a number of days, somewhere around 3 days. It's very scalable. So, the company is ramping up for very high throughput. I expect that this may become available and part of our diagnostic armamentarium in the fairly near future.
1. Klotz L, Olsson C, Kapoor D, et al. A high through-put test interrogating 442 small non-coding RNAs (sncRNA) extracted from urine exosomes accurately identifies and stratifies prostate cancer into low-, intermediate- or high-risk disease. Paper presented at: 2021 American Urological Association Annual Meeting; September 10-13, 2021; virtual. Abstract LBA02-09