Journal Briefs: The Journal of Urology: Specific Detection of Prostate Cancer Cells in Urine by RNA In Situ Hybridization

By: Jillian N. Eskra, PhD; Daniel Rabizadeh, BS; Jiayi Zhang, MD, PhD; William B. Isaacs, PhD; Jun Luo, PhD; Christian P. Pavlovich, MD | Posted on: 06 Aug 2021

Eskra JN, Rabizadeh D, Zhang J et al: Specific detection of prostate cancer cells in urine by RNA in situ hybridization. J Urol 2021; 206: 37.

Many contemporary problems in prostate cancer detection and management arise from the lack of precision of existing screening and detection methods.1-3 In the diagnostic setting, sensitivity for prostate cancer detection in men with indications for prostate needle biopsy is usually very high. However, among the roughly 1 million prostate biopsies performed annually in the United States alone, the majority find no cancer. In addition, prostate biopsies often result in the diagnosis of clinically insignificant cancers, contributing to prostate cancer overtreatment.4,5 Therefore, the lack of specificity in the traditional biopsy pathway for diagnosis of prostate cancer (particularly clinically significant prostate cancer) contributes to many clinical dilemmas, including the need for initial and repeat biopsy as well as the type and timing of definitive local and/or systemic treatment. Although adoption of targeted biopsy has resulted in some improvements, there remains a need to develop less invasive and more accurate methods of prostate cancer diagnosis, particularly of clinically significant cancers.

Urine represents an important biospecimen for the development of noninvasive prostate cancer detection methods. Many urine-based tests have been developed recently.6 A common feature of these urine tests, commercialized or currently under development, is the involvement of a “grind-and-bind” approach involving nucleic acid amplification that may be subject to many technical limitations as well as regulatory oversight. A cell-based test, in which a cancer cell can be definitively identified under a microscope, would overcome such limitations, and also achieve high specificity in cancer detection if cancer cells could be visualized.

We have developed technical approaches for a novel cell-based prostate cancer urine test employing molecular cytology. Our approach depends on shedding of target cells into urine facilitated by attentive digital rectal examination (DRE), and cancer cell detection was enabled by RNA in situ hybridization (RISH) designed to detect 3 prostate-specific RNA molecules: NKX3.1 (specific to cells of prostatic origin), PRAC-1 (specific to cells of prostatic origin but also increased in prostate cancer cells) and PCA3 (specific to prostate cancer cells). This method was analytically validated and enabled robust detection of malignant prostate cells in post-DRE urine sediments (see figure). Detailed analytical parameters and step-by-step protocols have been published.7

Figure. Representative image of single prostate cancer cell detected in post-DRE urine sediment by multiplex RISH for PRAC1 (red), NKX3-1 (white) and PCA3 (green), and counterstained with DAPI.

We applied this analytically validated PRAC1/NKX3.1/PCA3 multiplexed fluorescent RISH assay7 to a cohort of 98 post-DRE urine specimens.8 A total of 98 patients were recruited over a 5-month period at the Johns Hopkins Hospital. All patients enrolled in the study were seen by a single urologist to ensure uniform DRE and urine collection processes. Patients presenting to urology for prostate examination were instructed to collect the entire volume of urine with emphasis on catching the first-voided urine fraction. In this cohort, patient ages ranged from 41 to 84 years. Serum prostate specific antigen (PSA) levels ranged from 0.4 to 36.6 ng/ml. Among the biopsy-positive cancer patients (66), 32 (48.5%), 27 (40.9%) and 7 (10.6%) were found to have Gleason scores of 6, 7 and ≥8, respectively. According to Epstein criteria,9 48 cancer patients harbored clinically significant disease. The Epstein criteria (clinical stage T1c, PSA density <0.15 ng/ml/gm, absence of Gleason pattern 4 or 5, <3 positive biopsy cores and presence of <50% tumor per core) were defined more than 25 years ago and used to qualify patients for active surveillance.9 The biopsy positive rate in this preliminary study cohort is high because many patients were diagnosed patients under active surveillance. Biopsy was performed on 82 patients.

Of the 98 urine specimens collected, 20 were indeterminate and not suitable for analysis upon microscopic assessment. Reasons for indeterminate status were visual obstruction due to excessive debris, casts or acellular elements (5) and cellular overcrowding (15). This resulted in 78 specimens that were acceptable for our preliminary biomarker analysis (table 1). Among the 78 samples that were scored, 20 specimens were classified as RISH positive based on detection of prostate cancer cells (NKX3-1/PRAC1+ and PCA3+) and 58 patients were classified as RISH negative (table 1). The cancer negative category was comprised of 49 specimens positive for cells of prostatic origin (NKX3-1/PRAC1+ and PCA3—) and 9 specimens negative for prostate cells (NKX3-1/PRAC1— and PCA3—) (not shown in table 1). Notably, PCA3 positive cells were also positive for PRAC1/NKX3-1 in all 20 cases with positive detection of urinary prostate cancer cells.

Table 1. RISH results stratified according to patient characteristics and biopsy results

RISH positivity (NKX3-1/PRAC1+ and PCA3+) significantly correlated with biopsy detection of cancer (p=0.030), increasing Gleason score (p=0.003) and clinically significant disease by the Epstein criteria (p <0.0001). For detection of any prostate cancer, RISH performed with 35% sensitivity and 92% specificity (table 2). For detection of prostate cancer with Gleason ≥7, RISH performed with 54% sensitivity and 89% specificity (table 2). Stratification of patients according to Epstein criteria for clinically significant cancer resulted in 51% sensitivity and 95% specificity (table 2). The positive predictive value of the urine RISH test was 90% (95% CI 68.3 to 98.8) and the negative predictive value was 71% (95% CI 57.3 to 81.9) for clinically significant prostate cancer on biopsy. Detailed study findings from this pilot cohort have been published recently.8 Additional studies with an improved version of the test are ongoing.

Table 2. Assay performance

We have developed and analytically validated a multiplexed fluorescent RISH assay for specific detection and visualization of prostate cancer cells in post-DRE urine samples. The ability to directly visualize cancer cells in urine cytology specimens is considered a key advantage of this cell-based test. All current urine-based tests for prostate cancer involve in-solution polymerase chain reaction, an error-prone process that may adversely impact assay development subject to regulatory oversight. With some further improvement, we believe our test could be implemented in Clinical Laboratory Improvement Amendments or College of American Pathologists–certified labs, given its capacity to predict significant cancer on biopsy with high specificity.

  1. Walsh PC: Prostate cancer screening. N Engl J Med 2017; 376: 2401.
  2. Sharma V and Karnes RJ: Prostatectomy versus observation for early prostate cancer. N Engl J Med 2017; 377: 1302.
  3. Wallis CJD and Klotz L: Prostatectomy versus observation for early prostate cancer. N Engl J Med 2017; 377: 1301.
  4. Tosoian JJ and Carter HB: Active surveillance of localized prostate cancer: acknowledging uncertainty. J Clin Oncol 2016; 34: 4452.
  5. Carter HB: Optimizing active surveillance. Eur Urol 2016; 70: 909.
  6. Eskra JN, Rabizadeh D, Pavlovich CP et al: Approaches to urinary detection of prostate cancer. Prostate Cancer Prostatic Dis 2019; 22: 362.
  7. Eskra JN, Rabizadeh D, Mangold L et al: A novel method for detection of exfoliated prostate cancer cells in urine by RNA in situ hybridization. Prostate Cancer Prostatic Dis 2021; 24: 220.
  8. Eskra JN, Rabizadeh D, Zhang J et al: Specific detection of prostate cancer cells in urine by RNA in situ hybridization. J Urol 2021; 206: 37.
  9. Epstein JI, Walsh PC, Carmichael M et al: Pathologic and clinical findings to predict tumor extent of nonpalpable (stage t1 c) prostate cancer. JAMA 1994; 271: 368.