ASCO GU 2020: Analysis of Small Non-Coding RNAs in Urinary Exosomes to Classify Prostate Cancer into Low-Grade (GG1) and Higher-Grade (GG2-5)

San Francisco, California ( Prostate cancer remains the only malignancy routinely diagnosed through systematic whole-organ tissue sampling. A non-invasive test to definitively diagnose or rule out prostate cancer is, therefore, a much sought-after paradigm-changing technology.

At the oral abstract session on prostate cancer at the 2020 Genitourinary Cancers Symposium, the venerable Dr. Derya Tilki (presenting on behalf of Dr. Laurence Klotz) presented the results of a trial leading to the development and validation of three new genomic tests in this field: miR Sentinel PCa, miR Sentinel CS, and miR Sentinel HG.

All of these tests analyze the small non-coding RNA (sncRNA) isolated from the urinary exosome. miR Sentinel PCa uses this data to determine the presence or absence of prostate cancer, miR Sentinel CS uses this data to distinguish patients with grade group (GG) 1 disease from those with GG 2-5 disease, and miR HG distinguishes GG1-2 from GG3-5.

Dr. Tilki explained the development of these tests as follows. Urine was collected from 836 men with known prostate cancer status as determined by TRUS biopsy. Biopsy data showed no cancer in 89 men, GG1 disease in 88 men, and GG2-5 disease in 56 men. Affymetrix miR 4.0 arrays and a custom OpenArray platform were used in conjunction with the pathologic biopsy data to identify the 230 most informative sncRNAs and train the miR Sentinel PCa, CS, and HG classifiers.

Unsurprisingly, the classifiers performed extremely well on the dataset on which they were trained.

Much more significantly, these classifiers were then validated on a second dataset which consisted of 600 men with elevated PSA who had both Sentinel and biopsy data.  

When tested in this new dataset, the performance characteristics for miR Sentinel PCa were, therefore, sensitivity 94%, specificity 92%, positive predictive value 92% and negative predictive value 94%. The performance characteristics for miR Sentinel CS were sensitivity 93%, specificity 90%, positive predictive value 91% and negative predictive value 92%. The performance characteristics for miR Sentinel HG were sensitivity 94%, specificity 96%, positive predictive value 91% and negative predictive value 97%.

These are truly impressive numbers. They suggest that these miR Sentinel tests are the most sensitive and specific tests currently available for the detection of clinically significant prostate cancer. If these findings are further validated, one can easily imagine a dramatic shift occurring in the way prostate cancer is diagnosed.

That said, further validation is certainly warranted. As Dr. Tilki pointed out, the gold standard used in this study is TRUS biopsy, which is known to have a significant false negative rate for the detection of prostate cancer and clinically significant prostate cancer. She explained the excellent performance characteristics of the Sentinel tests in light of this limitation of TRUS biopsy may be explained in part by the fact that many patients included in the dataset had undergone multiple biopsies, which would improve the performance of TRUS. Dr. Daniel Lin, who provided an invited commentary on the presentation also highlighted some significant heterogeneity in the training and validation cohorts, including a mean PSA in the GG4 cohort of 58, which suggests that many of the patients included in this trial would not actually have any use for a genomic classifier. 

With these caveats in mind, there is no doubt that the results of future validation of these tests will be eagerly awaited.

Presented By: Derya Tilki, MD attending urologist, Martini-Klinik Prostate Cancer Center, University Hospital Hamburg-Eppendorf, Hamburg, Germany

Written by: Marshall Strother, MD, Society for Urologic Oncology Fellow, Division of Urologic Oncology, Fox Chase Cancer Center, Philadelphia PA @mcstroth at the 2020 Genitourinary Cancers Symposium, ASCO GU #GU20, February 13-15, 2020, San Francisco, California


  1. Ahmed HU, El-shater bosaily A, Brown LC, et al. Diagnostic accuracy of multi-parametric MRI and TRUS biopsy in prostate cancer (PROMIS): a paired validating confirmatory study. Lancet. 2017;389(10071):815-822.
email news signup