The promise of using reverse transcriptase–polymerase chain reaction (RT-PCR) technology for the detection of circulating prostate cancer cells in peripheral blood, although technically feasible at the molecular level, has proven clinically impractical for routine implementation in patient management. Reverse transcriptase–polymerase chain reaction has been successfully applied to detect and quantify (relatively speaking) genes that are differentially expressed in cells and tissues obtained from patients during various stages of malignant growth. In addition, the method has been applied to the detection of circulating cancer cells in peripheral blood using highly specific primer sets for specific molecular targets. These include epithelial cell cytokeratins for breast cancer, as well as enzymes, such as tyrosinase for melanoma and prostate-specific antigen (PSA) and prostate-specific membrane antigen (PSMA) for prostate cancer, using either nonnested or nested methodologies.
The promise of using reverse transcriptasepolymerase chain reaction (RT-PCR) technology for the detection of circulating prostate cancer cells in peripheral blood, although technically feasible at the molecular level, has proven clinically impractical for routine implementation in patient management. Reverse transcriptasepolymerase chain reaction has been successfully applied to detect and quantify (relatively speaking) genes that are differentially expressed in cells and tissues obtained from patients during various stages of malignant growth. In addition, the method has been applied to the detection of circulating cancer cells in peripheral blood using highly specific primer sets for specific molecular targets. These include epithelial cell cytokeratins for breast cancer, as well as enzymes, such as tyrosinase for melanoma and prostate-specific antigen (PSA) and prostate-specific membrane antigen (PSMA) for prostate cancer, using either nonnested or nested methodologies.
Clinical Sensitivity and Specificity of RT-PCR
The molecular sensitivity and specificity of RT-PCR have never been issues, since the methodology can detect a single cancer cell in a background of 107 to 108 white blood cells. It is the clinical sensitivity and specificity of the assay in prostate cancer that have been challenged, ie, the ability of RT-PCR to reproducibly detect circulating prostate cells in patients blood and then accurately relate those results to either stage or recurrence.
Most of the technical factors associated with performance of the RT-PCR test, such as primer design, RNA concentration, and RT and PCR conditions, can be rigorously standardized and controlled, rendering experimental molecular accuracy very reproducible. In contrast, other factors that compromise the quality and quantity of the RNA obtained from clinical samples have produced laboratory results from various institutions that do not reproducibly correlate with patient-specific stage and/or recurrence predictions. Hence, it appears that biological factors related to specimen processing and the patients prostate cancer status, as well as other underlying treatment and disease processes, are not so controllable and may contribute to variability in results.
Prostate cancer is an extremely heterogeneous disease biologically when assessed at the anatomic and molecular pathologic levels.[1,2] This heterogeneity also likely contributes to the variability in RT-PCR results mentioned in this review by de la Taille and colleagues.
Furthermore, the detection of a circulating prostate cell is somewhat irrelevant to the question of the malignant potential of that cell and its contribution to the ultimate development of metastasis and an androgen-independent state, leading to a fatal outcome for the patient. In my opinion, there currently is no single blood test that can make such critical clinical distinctions.
Alternative Technologies for Detecting Circulating Cancer Cells
In terms of the completeness of the authors review, it would have been useful if they had cited additional studies of RT-PCR and alternative technologies for the detection of circulating cancer cells in the blood. For example, there is evidence that normal individuals and men with benign prostatic hypertrophy demonstrate positive RT-PCR results.[3-6] Also, Moreno et al clearly demonstrated the value of RT-PCR for assessing the impact of manipulation of the prostate by either biopsy or transurethral resection of the prostate , and demonstrated that such manipulation results in a shower of prostate cells into the peripheral blood.
Recently, as cited in the article by the collaborating authors, Dr. Robert Vessella presented the results of an RT-PCR consortium at the 1998 American Urological Association (AUA) meeting held in San Diego, which clearly revealed the clinical impracticality of the RT-PCR-PSA technology. This more than 2-year study involved over 300 patients and employed seven clinical sites, as well as four different
RT-PCR-PSA testing sites, using only one central site to process all of the blood samples for RNA. In his presentation, Vessella noted that the four testing sites, employing different validated technologies of equal sensitivity, failed to produce data that even remotely correlated for the same samples of RNA. (These data are being prepared for publication.) Most likely, the RNA was severely compromised during the collection and shipping process.
More recently, investigators have turned to more direct physical isolation methods to procure purified prostate epithelial cells from the blood.[7,8] Such direct isolation methods permit analysis of the actual numbers of circulating cells and determine the malignant phenotype via molecular probes.[7,8] Although these techniques are currently labor-intensive and commercially challenging, they specifically address the key issue of whether or not one is dealing with a malignant cell. Direct physical isolation methods also offer the opportunity to directly study individual cells for genomic alterations that are suspected to cause androgen-independence and/or metastasis, as well as a fatal outcome for prostate cancer.
Statistical and Neural Network Patient Outcome Predictive Models
Dr. de la Taille and colleagues do allude to the importance of pathologic variables that can be determined from the biopsy, such as the Gleason pattern and score, both univariately and multivariately, for predicting stage and/or risk of progression. It might have been helpful if the authors had continued with this theme and had mentioned the use of published statistical and neural network patient outcome predictive models to accomplish pretreatment staging. This approach combines clinical and pathologic information available from the patient and his biopsy, as well as laboratory data, to construct a training set from which patient-specific outcomes may be predicted. I believe that such an approach may eventually result in decision support tools with superior clinical efficacy that will enhance the patients ability to elect treatment options that relate specifically to the biology of his cancer.
This review article concludes that more multi-institution studies are needed to determine the clinical validity of RT-PCR. Based on a thorough, critical analysis of the literature and the failure multiple studies to prove significant clinical efficacy of RT-PCR, I strongly argue that this approach is inadvisable.
Proposed future studies of RT-PCR would require the resolution of preservation issues for cancer cells and/or RNA. Instead, I would rather support and encourage the development of rapid methods for direct physical isolation and molecular diagnosis of the malignant phenotype of such cells as a possible replacement technology. Finally, I favor the development of highly accurate decision algorithms that combine all pretreatment clinical, pathologic, and laboratory information to construct patient-specific outcome predictions that can guide the patient and urologist in the selection of the most appropriate treatment options.[1,9]
1. ODowd GJ, Veltri RW, Orozco R, et al: Update on the appropriate staging evaluation for newly diagnosed prostate cancer. J Urol 158:687-698, 1997.
2. Cher ML, Bova GS, Moore DH, et al: Genetic alterations in untreated metastases and androgen-independent prostate cancer detected by comparative genomic hybridization and allelotyping. Cancer Res 56:3090-3102, 1996.
3. Gomella LG, Raj GV, Moreno JG: Reverse transcriptase polymerase chain reaction for prostate specific antigen in the management of prostate cancer. J Urol 158:326-337, 1997.
4. Moreno JG, OHara SM, Long JP, et al: Transrectal ultrasound-guided biopsy causes hematogenous dissemination of prostate cells as determined by RT-PCR. Urology 49:515-520, 1997.
5. OHara SM, Veltri RW, Skirpstunas P, et al: Basal PSA mRNA levels detected by quantitative RT-PCR-PSA in blood from subjects without prostate cancer (abstract). Proc Am Urol Assoc 155(4):430A, 1996.
6. Gala JL, Heusterspreute M, Loric S, et al: Expression of prostate-specific antigen and prostate-specific membrane antigen transcripts in blood cells: Implications for the detection of hematogenous prostate cells and standardization. Clin Chem 44(3):472-481, 1998.
7. Tso POP, Pannek J, Wang ZP, et al: Detection of intact prostate cancer cells in the blood of men with prostate cancer. Urology 49:881-885, 1997.
8. Racila E, Euhus D, Weiss AJ, et al: Detection and characterization of carcinoma cells in the blood. Proc Natl Acad Sci 95(8):4589-4594, 1998.
9. Veltri RW, ODowd GJ, Orozco R, et al: The role of biopsy pathology, quantitative nuclear morphometry, and biomarkers in preoperative prediction of prostate cancer staging and prognosis. Semin Urol Oncol 16:106-117, 1998.