- Email: [email protected]
Integrating the Genomic Prostate Score into Clinical Practice Workflow
urologypracticejournal.com
Integrating the Genomic Prostate Score into Clinical Practice Workflow David Albala,*,y Premal J. Desai,y Vahan Kassabian, Gerald Andriole, Jan C. Kennedy and Steven Canfieldy From the Associated Medical Professionals of New York, Syracuse, New York (DA), Tower Urology, Los Angeles, California (PJD), Georgia Urology, Atlanta (VK), and Dekalb Medicine, Decatur (JCK), Georgia, Washington University Urologic, St. Louis, Missouri (GA), and University of Texas Health Science Center at Houston, Houston, Texas (SC)
Abstract
Abbreviations and Acronyms
Introduction: A dilemma that urologists face is how to determine which patients with prostate cancer need immediate intervention and which patients can be safely placed on active surveillance. Gene expression profile analysis of biopsy tissue has been proposed as a means of providing more accurate risk stratification for low risk prostate cancer. However, there is a general lack of acceptance and standardization around the integration of genomic testing in clinical practice. The Oncotype DXÒ prostate cancer assay is a commercially available tissue based assay that assesses the expression of key genes across multiple biological pathways predictive of prostate cancer aggressiveness from the diagnostic biopsy specimen, and reports an individual Genomic Prostate Score. Methods: With the recommendations set forth in this article we aim to standardize operational best practices for the integration of the Genomic Prostate Score into clinical practice. Its purpose is to provide practical guidance to help physicians understand, run, interpret and communicate actionable results to patients.
AS = active surveillance GPS = Genomic Prostate Score NCCNÒ = National Comprehensive Cancer NetworkÒ PCa = prostate cancer PSA = prostate specific antigen
Results: The Genomic Prostate Score reflects the biology of the underlying tumor to help guide initial treatment decisions at the time of biopsy. This article is based on real-world evidence from the authors’ respective experiences at their institutions and practices. The authors were carefully selected based on their depth of experience and knowledge about the Genomic Prostate Score and, as such, it is their expertise that is being leveraged to support the best practices algorithm. Conclusions: This article provides easy to use, clear-cut and practical guidance for physicians on how to use the Genomic Prostate Score to inform decisions regarding active surveillance. Key Words: prostatic neoplasms, genomics, biological assay
Submitted for publication May 16, 2015. No direct or indirect commercial incentive associated with publishing this article. The corresponding author certifies that, when applicable, a statement(s) has been included in the manuscript documenting institutional review board, ethics committee or ethical review board study approval; principles of Helsinki 2352-0779/16/35-371/0 UROLOGY PRACTICE Ó 2016 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION
AND
Declaration were followed in lieu of formal ethics committee approval; institutional animal care and use committee approval; all human subjects provided written informed consent with guarantees of confidentiality; IRB approved protocol number; animal approved project number. * Correspondence: Duke University Medical Center, Syracuse, New York. y Financial interest and/or other relationship with Genomic Health.
RESEARCH, INC.
http://dx.doi.org/10.1016/j.urpr.2015.10.010 Vol. 3, 371-378, September 2016 Published by Elsevier
372
Genomic Prostate Score in Clinical Practice
Differentiating indolent from aggressive PCa can be difficult and consequently lead to overtreatment.1,2 Recent controversies and confusion about PSA screening have highlighted the unmet need for improved decision making tools for men with localized PCa.3 Although screening has resulted in a decrease in PCa mortality, it is clear that these gains have come at the cost of diagnosing and treating many men with indolent PCa, cancers with a very low probability of progressing and becoming lethal. There is now strong evidence from multiple studies that low risk disease can be safely managed with active surveillance, a strategy involving careful monitoring with serial PSA measurements, examinations and biopsies, with curative intervention offered to those men with evidence of disease progression. Molecular markers of tumor aggressiveness can improve the risk assessment of PCa,4 thereby helping to identify which patients may not need immediate treatment with surgery or radiation. The viability of AS as an initial strategy for men with low risk disease has now been recognized by societies in their guidelines, including those of the American Urological Association, European Association of Urology and NCCN. Despite this recognition and the evidence, AS remains underused, largely due to concerns about the accuracy of existing risk stratification tools available at biopsy (especially given the issues of prostate tumor heterogeneity and the potential for under sampling by conventional biopsy approaches). Several groups have studied AS outcomes across a variety of cohorts. Although each of these cohorts enrolled or selected men with favorable risk profiles, there were differences in patient selection, ie Gleason 3þ3 and 3þ4, as well as differences in surveillance protocols and triggers for intervention. Despite these differences, AS has been shown to be exceedingly safe, with a reported median followup of approximately 40 months and a PCa specific survival greater than 99%. Although we rely on risk stratification to identify men who have low risk features, 30% to 40% of men undergoing radical prostatectomy for presumed low risk disease will have higher grade, higher stage disease or both on pathological staging. Current tools used in prostate cancer staging and risk stratification (PSA, Gleason score, tumor stage) have limitations. Combining current tools with genomic information will improve risk stratification and the accuracy needed for decision making. Given the uncertainty of current risk stratification, it is not surprising that a majority of men today receive immediate therapy. Oncotype DX
The development strategy for the Oncotype DX PCa assay (which provides a more accurate and individualized risk
assessment for men at the time of PCa diagnosis) was based on the successful approach used to develop the Oncotype DX breast and colon cancer assays. Two challenges faced in developing the test were 1) addressing issues of tumor heterogeneity and multifocality, and 2) optimizing a technology platform to reliably analyze and provide a GPS result based on the small amounts of tumor typically found in prostate needle biopsies.5 Oncotype DX is a 17-gene expression assay that has been analytically validated using limited RNA inputs.5 The assay consists of 5 reference genes (which account for varying RNA quality/quantity) and 12 cancer related genes representing 4 distinct biological pathways (androgen signaling, stromal response, cellular organization and proliferation). These contribute to the predictive value of the assay. The assay provides a GPS, which ranges from 0 to 100, with higher scores representing less favorable pathology and lower scores representing more favorable pathology. The GPS is calculated by summing the weighted expression of genes associated with worse outcomes and subtracting the weighted expression of genes associated with better outcomes. The assay has been clinically validated in 2 separate independent cohorts confirming Oncotype DX as a predictor of the likelihood of adverse pathology from the prostate needle biopsy and also as a predictor of the risk of biochemical recurrence after surgery.6,7 The test was validated in patients with NCCN very low, low and a subset of intermediate risk disease referred to as low-intermediate risk (Gleason 3þ4 disease). The test has not been studied in higher risk patients and, therefore, should not be considered in that setting. The intention of the GPS is to be used with accepted clinical criteria (ie NCCN risk classification) to stratify biopsy diagnosed localized PCa according to biological aggressiveness and, thus, direct patient care. Beyond simply refining risk categorization for a patient, the value of the GPS lies in its ability to provide actionable information to support treatment decisions for patients with very low, low and low-intermediate risk disease. The GPS provides individual risk estimation based on the likelihood of favorable pathology, meaning a low Gleason grade and organ confined disease. The report also breaks out the elements of adverse pathology and notes separately the risk of high grade disease and risk of high stage disease. The cases shown in figure 1 demonstrate how the GPS can provide actionable feedback impacting treatment decisions. For some clinicians in certain clinical settings these individual components may be informative for treatment planning. We believe a standard approach to integrating Oncotype DX into clinical practice is warranted (fig. 2).
Genomic Prostate Score in Clinical Practice
373
Figure 1. Cases demonstrate how GPS can provide actionable feedback impacting treatment decisions, with GPS 31, NCCN low risk to low risk (A), GPS 6, NCCN low risk to very low risk (B) and GPS 12, NCCN intermediate risk to low risk (C).
374
Genomic Prostate Score in Clinical Practice
prostate cancer assay DISCUSS prostate cancer assay in detail (reference brochure)
Figure 2. Proposed algorithm for integration of GPS into clinical practice. DRE, digital rectal examination.
Ordering the Oncotype DX Prostate Cancer Assay
Running the Oncotype Prostate Cancer Assay
Key stakeholders should include the urologist, the pathologist, the radiation oncologist, registered nurses and physician assistants, as well as the patient and his family. All specialties should align on criteria to consider in making the decision on when and where to incorporate the test. There should be peer reviewed data to confirm that the test is valid and worthwhile for the patient, and the test should provide actionable information.8 Once the scientific merit of the test has been established, the clinical staff (eg the practice manager and nursing staff) should be made aware of the purpose of the test and its proposed integration into the practice. We believe that the physician should have a discussion with the patient about the rationale for ordering the assay. The assay can be ordered after a prostate biopsy has been completed and the clinician has had a chance to discuss the diagnosis of prostate cancer with the patient. Alternatively some practices may choose to order the test reflexively and have the results available at the initial discussion of the biopsy results. It is critical that the NCCN risk calculation be accurate and the appropriate box on the requisition form (NCCN very low, low, intermediate) be selected. We recommend identifying 2 individuals in the practice who can fill out the GPS order form. However, the urologist should take the responsibility of ensuring that the NCCN risk category selected is correct and accurate (as this will impact the test interpretation). Ordering can be done online through a secure web portal or by fax.
Pathologists have an important role in specimen retrieval and preparation.9 It is imperative that a standard protocol between pathologists and urologists be established to optimize the yield of prostate biopsies and proactively manage issues that may pose challenges to accurate, individualized risk assessment for treatment selection. Once the test is ordered, the pathologist will choose the tumor containing biopsy block with the greatest amount of the highest grade carcinoma (longest linear measurement). Approximately 1 mm of tumor (continuously or discontinuously) or approximately 5% or more of a standard transrectal ultrasound biopsy core is required to extract adequate RNA from the tumor. While blocks and unstained slides are acceptable, blocks are preferred. In both validation studies the biopsy examined was within 6 months of the radical prostatectomy and, therefore, the reported likelihood of favorable pathology reflects the pathology of the prostate within 6 months of the biopsy. In the event that recent biopsy material is not available, samples up to 3 years old may be submitted. Once the specimen is received by Genomic Health, each sample is centrally reviewed by 2 urological pathologists.5 The central review is not a second opinion or an attempt to regrade the Gleason score, but rather is a review of the submitted sections to verify the presence of tumor. In case a differing Gleason pattern is seen that was not appreciated by the diagnostic stains previously done, further discussion with the pathologist can be helpful. In the majority of cases
Genomic Prostate Score in Clinical Practice
(approximately 90%), unused tissue is returned to the ordering physician and available to run additional tests as needed. Unique considerations should be accounted for when communicating with in-house pathologists vs outsourcing to a reference laboratory. For example, ideally several pieces of information should be included in pathology reports, such as confirmation of PCa diagnosis, location of PCa, amount of tumor (linear length of tumor), Gleason score, fraction of positive cores (number of positive cores/number of cores submitted), percentage involvement by carcinoma and status of local invasion. The timing of tissue submission in turn impacts the turnaround time and receipt of the GPS. Open communication between the pathologist and the urologist can be helpful in expediting the processing and shipping of tissue, which will facilitate a rapid turnaround of the GPS.
Report Interpretation
The information in the report is personalized, based on the patient’s individual GPS and the NCCN clinical risk assessment category selected by the urologist when the test was ordered (fig. 1). The top of the report includes patient demographics, the GPS result, a brief description of the test and a description of the clinical validation study. The date is the date that the specimen was received at Genomic Health laboratories. The NCCN risk group, provided by the ordering physician, is indicated on the report and corresponds to the region of the bar graph shaded in purple. The GPS ranges from 0 to 100 and on every report the GPS is in a circle (fig. 1). The lower the GPS, the more favorable the tumor biology and the less aggressive the disease. Conversely, the higher the GPS, the more aggressive and less favorable the biology. The horizontal box above the GPS scale represents the NCCN clinical risk groups in which the test was validated. The GPS above the risk box will be to the left of the shaded segment when the biology is more favorable than that predicted by clinical criteria alone, or to the right when it is less favorable than would be predicted by clinical criteria alone. If the GPS result aligns in the shaded segment, the tumor biology is consistent with the clinical criteria confirming the clinical risk. In addition to providing an estimate of the overall likelihood of favorable pathology, the report breaks this end point out into its 2 components of the likelihood of freedom from high grade disease and the likelihood of freedom from nonorgan confined disease. For some clinicians in certain clinical settings these individual components may be informative for treatment planning. The incorporation of the GPS
375
and NCCN clinical risk provides a more individualized and precise risk assessment to allow more informed treatment recommendations and decisions. There should be an indepth discussion between the physician and the patient on all the components of the report (see Appendix). Once this has occurred, both parties can make a final informed treatment decision.
Costs and Reimbursement
Cost is an important consideration in cancer care. By providing information that helps determine which patients need immediate treatment vs those who may benefit from active surveillance, Oncotype DX may help manage the overall cost of care.10 At present, Oncotype DX is covered by Medicare for patients with very low and low risk prostate cancer by NCCN classification. Genomic Health provides a Genomic Access Program (GAP) to help patients obtain coverage for Oncotype DX from their insurers or third-party payers as regional differences in coverage may exist.11 Financial assistance is also available to those with limited incomes. In cases in which all appeals have been exhausted and the claim is not reimbursed, the patient may be responsible for the balance or a portion of the remaining invoice when appropriate policies have been followed and indicate a balance is due. In such instances Genomic Health provides assistance for patients with financial hardship as well as a program for uninsured and underinsured patients (all programs based on financial eligibility). The GAP provides value in its ability to determine whether the patient’s insurance plan is likely to cover the Oncotype DX test, to obtain prior authorization, if required, and to assist in the appeals process if the claim is denied by the insurance company.
Patient Communication
Effective patient-clinician communication is crucial.12 When communicating with patients who have very low, low or intermediate risk PCa about the role of the GPS in their treatment plan, best practices should be followed. The physician and the patient should have a face-to-face discussion about the results (see Appendix). This may be done once the biopsy and Oncotype DX results have been received. However, the final approach is at the physician’s discretion. The physician should highlight the 3 important sections of the Oncotype DX report, namely 1) the overall likelihood of favorable pathology (defined as freedom from high grade [dominant Gleason pattern 4 or any pattern 5] and/or nonorgan confined disease), 2) the likelihood of
376
Genomic Prostate Score in Clinical Practice
freedom from high grade disease and 3) the likelihood of freedom from nonorgan confined disease. The incorporation of the GPS with a NCCN clinical risk assessment provides a more individualized and precise risk assessment to allow more informed treatment recommendations and decisions. The physician should explain the natural history of the disease. Selected men with low risk PCa can be safely treated with a strategy of AS, involving careful monitoring with serial PSA measurements, examinations and biopsies, with curative intervention provided to those men with evidence of disease progression. Nevertheless, AS remains underused, largely due to concerns about the accuracy of existing risk stratification tools available at the time of biopsy (especially given the issues of prostate tumor heterogeneity and the potential for under sampling by conventional biopsy approaches). The physician should provide supplemental educational materials on PCa treatments. Finally, the physician should inform the patient about available financial support services through GAP. Allow approximately 30 minutes for this conversation and address any additional questions
Conclusions
Several genomic tests are commercially available and others will soon enter the market. These tests were not all specifically designed for the same target population as Oncotype DX. Some are for high risk disease after prostatectomy while others are for negative biopsies. The Oncotype DX prostate cancer assay is validated and is specifically intended for men with newly diagnosed low to low-intermediate risk PCa. The assay assesses the biology of an individual’s prostate tumor from biopsy tissue, and measures the activity of key genes and biological pathways associated with and predictive of PCa aggressiveness. The gene expression information is then used to calculate a GPS. The GPS results reveal the underlying biology of a patient’s cancer, and combined with clinical risk, can help guide initial treatment decisions at the time of diagnosis. The GPS allows us to better understand the biology of each patient’s prostate cancer, providing more accurate risk assessment to guide treatment recommendations. Based on real-world evidence and our clinical experience, the biopsy based genomic assay provides additional information on the biology of the cancer at biopsy, when important treatment decisions regarding active surveillance or immediate treatment need to be made.
Acknowledgments
Karen Ventii provided editorial assistance.
Appendix. Patient Summary Box Important information that each patient needs to know: The rationale for the test (communicate that biopsy may provide equivocal results and, therefore, a genomic test should be considered to help provide more clarity) The decision to test The test results and implications for guiding the treatment approach What active surveillance is The actionable impact of the results Costs
References 1. Goodman M, Ward KC, Osunkoya AO et al: Frequency and determinants of disagreement and error in Gleason scores: a population-based study of prostate cancer. Prostate 2012; 72: 1389. 2. National Cancer Institute: Prostate-Specific Antigen (PSA) Test. Available at http://www.cancer.gov/types/prostate/psa-fact-sheet. 3. Summaries for patients. Screening for prostate cancer: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med 2012; 157: 1. 4. Crawford ED, Ventii K and Shore ND: New biomarkers in prostate cancer. Oncology (Williston Park) 2014; 28: 135. 5. Knezevic D, Goddard AD, Natraj N et al: Analytical validation of the Oncotype DX prostate cancer assayda clinical RT-PCR assay optimized for prostate needle biopsies. BMC Genomics 2013; 14: 690. 6. Klein EA, Cooperberg MR, Magi-Galluzzi C et al: A 17-gene assay to predict prostate cancer aggressiveness in the context of Gleason grade heterogeneity, tumor multifocality, and biopsy undersampling. Eur Urol 2014; 66: 550. 7. Cullen J, Rosner I, Brand T et al: A prospectively-designed study to determine the association of a 17-gene genomic prostate score with recurrence following surgery for localised prostate cancer. Presented at ESMO (European Society for Medical Oncology) 2014, Madrid, Spain, September 26-30, 2014. 8. Canfield SE, Kibel AS, Kemeter MJ et al: A guide for clinicians in the evaluation of emerging molecular diagnostics for newly diagnosed prostate cancer. Rev Urol 2014; 16: 172. 9. Montironi R, Lopez-Beltran A, Mazzucchelli R et al: Contemporary update on pathology-related issues on routine workup of prostate biopsy: sectioning, tumor extent measurement, specimen orientation, and immunohistochemistry. Anal Quant Cytopathol Histpathol 2014; 36: 61. 10. Badani K, Kemeter M, Febbo P et al: The impact of the Oncotype DX genomic prostate score (GPS) on initial treatment recommendations (TR) for men with newly diagnosed clinically low-risk prostate cancer (PCa). Presented at 15th Annual Meeting of the Society of Urologic Oncology, Bethesda, Maryland, December 3-5, 2014. 11. The Genomic Access Program (GAP) Brochure. Available at www.oncotypeDX.com. Accessed October 21, 2014. 12. Epstein RM and Street RL Jr: Patient-Centered Communication in Cancer Care: Promoting Healing and Reducing Suffering. NIH Publication No. 07-6225. Bethesda, Maryland: National Cancer Institute 2007.
Genomic Prostate Score in Clinical Practice
377
Editorial Commentaries
There are several options for reducing the harms of over detection and overtreatment of localized prostate cancer while preserving the 50% reduction in mortality that occurred during the last 25 years. These include using smarter screening practices, new biomarkers, and imaging to maximize the detection of high grade disease and minimize the detection of low risk tumors, and using active surveillance when low risk disease is identified. There is ample evidence that our use of active surveillance for low risk patients is increasing.1,2 However, there remains a greater than 20% risk that we misclassify a “tiger” as a “pussycat” using the relatively crude tools at our disposal such as PSA, grade, Gleason score and other histopathological findings on biopsy and, more recently, magnetic resonance imaging findings.3 This generates anxiety and uncertainty, and dampens enthusiasm for this approach. Ultimately molecular markers will supplant the existing tools as they have in many other cancers. Albala et al describe the novel test Oncotype DX which refines risk stratification beyond the NCCN risk strata. The factors that will drive the adoption of this and other new tests are 1) the science, ie how well the test distinguishes between tigers and pussycats, and 2) the implementation, ie
The Genomic Prostate Score (Oncotype DX) is 1 of 2 commercially available tests (along with ProlarisÒ) that assess the molecular features of prostate biopsy tissue. Test results have the potential to influence the decision for surveillance vs treatment as the initial strategy for very low, low and low-intermediate risk prostate cancer. Albala et al aim to standardize operational best practices and suggest an algorithm for the integration of the GPS into clinical practice. We believe that an initial assessment of the appropriateness of surveillance vs treatment can be determined based on the clinical characteristics of tumor and patient, but the final decision must be made after adequate patient education, some sort of additional testing and due consideration of patient preferences. It remains to be determined what combination of molecular testing, magnetic resonance imaging and/or repeat prostate biopsy is best to inform shared decision making after initial biopsy. As the authors state, if a test is to be ordered, it should provide actionable information for the patient. At this time uncertainty remains regarding how much weight to give to genomic testing with regard to shared
the test’s marketing, ease of use, cost and extent to which it provides digestible, actionable information to the patient and provider. The greater our certainty in distinguishing tigers from pussycats, the more confidently we can promote active surveillance for appropriate candidates, and the more we can improve on the risk-benefit ratio of prostate cancer detection and treatment. Daniel A. Barocas Vanderbilt University Medical Center Nashville, Tennessee
References 1. Ritch CR, Graves AJ, Keegan KA et al: Increasing use of observation among men at low risk for prostate cancer mortality. J Urol 2015; 193: 801. 2. Cooperberg MR and Carroll PR: Trends in management for patients with localized prostate cancer, 1990-2013. JAMA 2015; 314: 80. 3. Tosoian JJ, Sundi D, Trock BJ et al: Pathologic outcomes in favorable-risk prostate cancer: comparative analysis of men electing active surveillance and immediate surgery. Eur Urol 2016; 69: 576.
decision making. Thus, test results may not be “actionable.” In addition, we caution against the standardized (or routine) use of any test (as implied in figure 2) because there are some patients for whom the treatment decision is almost certain not to change. We would encourage clinicians to order such testing only after face-to-face discussion of the initial biopsy results with the patient and a discussion regarding the merits of the GPS and its alternatives. Brian R. Lane Michigan State University College of Human Medicine Spectrum Health Medical Group Grand Rapids, Michigan and Michael L. Cher Karmanos Cancer Institute Department of Urology Wayne State University School of Medicine Detroit, Michigan
378
Genomic Prostate Score in Clinical Practice
Reply by Authors
Many localized prostate cancers have similar histological features at diagnosis. However, they often exhibit diverse biological behavior that may not be captured by traditional clinical risk categories (eg NCCN criteria). Traditional risk categories are based on clinical findings, including PSA and Gleason grade or stage on digital rectal examination. Uncertainty around whether to treat immediately has been caused in part by the divergent behavior of localized prostate cancers and the limitations of current diagnostic techniques. Given the lack of accurate risk assessment tools for early PCa as well as the potential inaccuracy of current biopsy techniques, additional prognostic methods are needed to biologically stratify this disease. Recent NCCN guidelines recognize that genomic testing should have a role in PCa treatment decision making.1 With such recognition there will likely be more awareness of the role of genomics in everyday clinical practice, helping physicians better decide on treatment vs observation.
Collectively the genomic tests are relatively new to PCa and are following a trajectory similar to that of prognostic breast cancer assays. The Oncotype DX prostate cancer assay is a commercially available tissue based assay that helps to confirm clinical judgment regarding the management of PCa (reference 6 in article). More information is needed on the practical application of the test. This review builds on our clinical expertise (depth of experience and knowledge about the assay) to provide easy to use, clear-cut and practical guidance for physicians on its use in patients with low and low-intermediate risk PCa.
Reference 1. NCCN: National Comprehensive Cancer Network. Prostate Cancer Early Detection. Version 1, 2016.
Integrating the Genomic Prostate Score into Clinical Practice Workflow David Albala,*,y Premal J. Desai,y Vahan Kassabian, Gerald Andriole, Jan C. Kennedy and Steven Canfieldy From the Associated Medical Professionals of New York, Syracuse, New York (DA), Tower Urology, Los Angeles, California (PJD), Georgia Urology, Atlanta (VK), and Dekalb Medicine, Decatur (JCK), Georgia, Washington University Urologic, St. Louis, Missouri (GA), and University of Texas Health Science Center at Houston, Houston, Texas (SC)
Abstract
Abbreviations and Acronyms
Introduction: A dilemma that urologists face is how to determine which patients with prostate cancer need immediate intervention and which patients can be safely placed on active surveillance. Gene expression profile analysis of biopsy tissue has been proposed as a means of providing more accurate risk stratification for low risk prostate cancer. However, there is a general lack of acceptance and standardization around the integration of genomic testing in clinical practice. The Oncotype DXÒ prostate cancer assay is a commercially available tissue based assay that assesses the expression of key genes across multiple biological pathways predictive of prostate cancer aggressiveness from the diagnostic biopsy specimen, and reports an individual Genomic Prostate Score. Methods: With the recommendations set forth in this article we aim to standardize operational best practices for the integration of the Genomic Prostate Score into clinical practice. Its purpose is to provide practical guidance to help physicians understand, run, interpret and communicate actionable results to patients.
AS = active surveillance GPS = Genomic Prostate Score NCCNÒ = National Comprehensive Cancer NetworkÒ PCa = prostate cancer PSA = prostate specific antigen
Results: The Genomic Prostate Score reflects the biology of the underlying tumor to help guide initial treatment decisions at the time of biopsy. This article is based on real-world evidence from the authors’ respective experiences at their institutions and practices. The authors were carefully selected based on their depth of experience and knowledge about the Genomic Prostate Score and, as such, it is their expertise that is being leveraged to support the best practices algorithm. Conclusions: This article provides easy to use, clear-cut and practical guidance for physicians on how to use the Genomic Prostate Score to inform decisions regarding active surveillance. Key Words: prostatic neoplasms, genomics, biological assay
Submitted for publication May 16, 2015. No direct or indirect commercial incentive associated with publishing this article. The corresponding author certifies that, when applicable, a statement(s) has been included in the manuscript documenting institutional review board, ethics committee or ethical review board study approval; principles of Helsinki 2352-0779/16/35-371/0 UROLOGY PRACTICE Ó 2016 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION
AND
Declaration were followed in lieu of formal ethics committee approval; institutional animal care and use committee approval; all human subjects provided written informed consent with guarantees of confidentiality; IRB approved protocol number; animal approved project number. * Correspondence: Duke University Medical Center, Syracuse, New York. y Financial interest and/or other relationship with Genomic Health.
RESEARCH, INC.
http://dx.doi.org/10.1016/j.urpr.2015.10.010 Vol. 3, 371-378, September 2016 Published by Elsevier
372
Genomic Prostate Score in Clinical Practice
Differentiating indolent from aggressive PCa can be difficult and consequently lead to overtreatment.1,2 Recent controversies and confusion about PSA screening have highlighted the unmet need for improved decision making tools for men with localized PCa.3 Although screening has resulted in a decrease in PCa mortality, it is clear that these gains have come at the cost of diagnosing and treating many men with indolent PCa, cancers with a very low probability of progressing and becoming lethal. There is now strong evidence from multiple studies that low risk disease can be safely managed with active surveillance, a strategy involving careful monitoring with serial PSA measurements, examinations and biopsies, with curative intervention offered to those men with evidence of disease progression. Molecular markers of tumor aggressiveness can improve the risk assessment of PCa,4 thereby helping to identify which patients may not need immediate treatment with surgery or radiation. The viability of AS as an initial strategy for men with low risk disease has now been recognized by societies in their guidelines, including those of the American Urological Association, European Association of Urology and NCCN. Despite this recognition and the evidence, AS remains underused, largely due to concerns about the accuracy of existing risk stratification tools available at biopsy (especially given the issues of prostate tumor heterogeneity and the potential for under sampling by conventional biopsy approaches). Several groups have studied AS outcomes across a variety of cohorts. Although each of these cohorts enrolled or selected men with favorable risk profiles, there were differences in patient selection, ie Gleason 3þ3 and 3þ4, as well as differences in surveillance protocols and triggers for intervention. Despite these differences, AS has been shown to be exceedingly safe, with a reported median followup of approximately 40 months and a PCa specific survival greater than 99%. Although we rely on risk stratification to identify men who have low risk features, 30% to 40% of men undergoing radical prostatectomy for presumed low risk disease will have higher grade, higher stage disease or both on pathological staging. Current tools used in prostate cancer staging and risk stratification (PSA, Gleason score, tumor stage) have limitations. Combining current tools with genomic information will improve risk stratification and the accuracy needed for decision making. Given the uncertainty of current risk stratification, it is not surprising that a majority of men today receive immediate therapy. Oncotype DX
The development strategy for the Oncotype DX PCa assay (which provides a more accurate and individualized risk
assessment for men at the time of PCa diagnosis) was based on the successful approach used to develop the Oncotype DX breast and colon cancer assays. Two challenges faced in developing the test were 1) addressing issues of tumor heterogeneity and multifocality, and 2) optimizing a technology platform to reliably analyze and provide a GPS result based on the small amounts of tumor typically found in prostate needle biopsies.5 Oncotype DX is a 17-gene expression assay that has been analytically validated using limited RNA inputs.5 The assay consists of 5 reference genes (which account for varying RNA quality/quantity) and 12 cancer related genes representing 4 distinct biological pathways (androgen signaling, stromal response, cellular organization and proliferation). These contribute to the predictive value of the assay. The assay provides a GPS, which ranges from 0 to 100, with higher scores representing less favorable pathology and lower scores representing more favorable pathology. The GPS is calculated by summing the weighted expression of genes associated with worse outcomes and subtracting the weighted expression of genes associated with better outcomes. The assay has been clinically validated in 2 separate independent cohorts confirming Oncotype DX as a predictor of the likelihood of adverse pathology from the prostate needle biopsy and also as a predictor of the risk of biochemical recurrence after surgery.6,7 The test was validated in patients with NCCN very low, low and a subset of intermediate risk disease referred to as low-intermediate risk (Gleason 3þ4 disease). The test has not been studied in higher risk patients and, therefore, should not be considered in that setting. The intention of the GPS is to be used with accepted clinical criteria (ie NCCN risk classification) to stratify biopsy diagnosed localized PCa according to biological aggressiveness and, thus, direct patient care. Beyond simply refining risk categorization for a patient, the value of the GPS lies in its ability to provide actionable information to support treatment decisions for patients with very low, low and low-intermediate risk disease. The GPS provides individual risk estimation based on the likelihood of favorable pathology, meaning a low Gleason grade and organ confined disease. The report also breaks out the elements of adverse pathology and notes separately the risk of high grade disease and risk of high stage disease. The cases shown in figure 1 demonstrate how the GPS can provide actionable feedback impacting treatment decisions. For some clinicians in certain clinical settings these individual components may be informative for treatment planning. We believe a standard approach to integrating Oncotype DX into clinical practice is warranted (fig. 2).
Genomic Prostate Score in Clinical Practice
373
Figure 1. Cases demonstrate how GPS can provide actionable feedback impacting treatment decisions, with GPS 31, NCCN low risk to low risk (A), GPS 6, NCCN low risk to very low risk (B) and GPS 12, NCCN intermediate risk to low risk (C).
374
Genomic Prostate Score in Clinical Practice
prostate cancer assay DISCUSS prostate cancer assay in detail (reference brochure)
Figure 2. Proposed algorithm for integration of GPS into clinical practice. DRE, digital rectal examination.
Ordering the Oncotype DX Prostate Cancer Assay
Running the Oncotype Prostate Cancer Assay
Key stakeholders should include the urologist, the pathologist, the radiation oncologist, registered nurses and physician assistants, as well as the patient and his family. All specialties should align on criteria to consider in making the decision on when and where to incorporate the test. There should be peer reviewed data to confirm that the test is valid and worthwhile for the patient, and the test should provide actionable information.8 Once the scientific merit of the test has been established, the clinical staff (eg the practice manager and nursing staff) should be made aware of the purpose of the test and its proposed integration into the practice. We believe that the physician should have a discussion with the patient about the rationale for ordering the assay. The assay can be ordered after a prostate biopsy has been completed and the clinician has had a chance to discuss the diagnosis of prostate cancer with the patient. Alternatively some practices may choose to order the test reflexively and have the results available at the initial discussion of the biopsy results. It is critical that the NCCN risk calculation be accurate and the appropriate box on the requisition form (NCCN very low, low, intermediate) be selected. We recommend identifying 2 individuals in the practice who can fill out the GPS order form. However, the urologist should take the responsibility of ensuring that the NCCN risk category selected is correct and accurate (as this will impact the test interpretation). Ordering can be done online through a secure web portal or by fax.
Pathologists have an important role in specimen retrieval and preparation.9 It is imperative that a standard protocol between pathologists and urologists be established to optimize the yield of prostate biopsies and proactively manage issues that may pose challenges to accurate, individualized risk assessment for treatment selection. Once the test is ordered, the pathologist will choose the tumor containing biopsy block with the greatest amount of the highest grade carcinoma (longest linear measurement). Approximately 1 mm of tumor (continuously or discontinuously) or approximately 5% or more of a standard transrectal ultrasound biopsy core is required to extract adequate RNA from the tumor. While blocks and unstained slides are acceptable, blocks are preferred. In both validation studies the biopsy examined was within 6 months of the radical prostatectomy and, therefore, the reported likelihood of favorable pathology reflects the pathology of the prostate within 6 months of the biopsy. In the event that recent biopsy material is not available, samples up to 3 years old may be submitted. Once the specimen is received by Genomic Health, each sample is centrally reviewed by 2 urological pathologists.5 The central review is not a second opinion or an attempt to regrade the Gleason score, but rather is a review of the submitted sections to verify the presence of tumor. In case a differing Gleason pattern is seen that was not appreciated by the diagnostic stains previously done, further discussion with the pathologist can be helpful. In the majority of cases
Genomic Prostate Score in Clinical Practice
(approximately 90%), unused tissue is returned to the ordering physician and available to run additional tests as needed. Unique considerations should be accounted for when communicating with in-house pathologists vs outsourcing to a reference laboratory. For example, ideally several pieces of information should be included in pathology reports, such as confirmation of PCa diagnosis, location of PCa, amount of tumor (linear length of tumor), Gleason score, fraction of positive cores (number of positive cores/number of cores submitted), percentage involvement by carcinoma and status of local invasion. The timing of tissue submission in turn impacts the turnaround time and receipt of the GPS. Open communication between the pathologist and the urologist can be helpful in expediting the processing and shipping of tissue, which will facilitate a rapid turnaround of the GPS.
Report Interpretation
The information in the report is personalized, based on the patient’s individual GPS and the NCCN clinical risk assessment category selected by the urologist when the test was ordered (fig. 1). The top of the report includes patient demographics, the GPS result, a brief description of the test and a description of the clinical validation study. The date is the date that the specimen was received at Genomic Health laboratories. The NCCN risk group, provided by the ordering physician, is indicated on the report and corresponds to the region of the bar graph shaded in purple. The GPS ranges from 0 to 100 and on every report the GPS is in a circle (fig. 1). The lower the GPS, the more favorable the tumor biology and the less aggressive the disease. Conversely, the higher the GPS, the more aggressive and less favorable the biology. The horizontal box above the GPS scale represents the NCCN clinical risk groups in which the test was validated. The GPS above the risk box will be to the left of the shaded segment when the biology is more favorable than that predicted by clinical criteria alone, or to the right when it is less favorable than would be predicted by clinical criteria alone. If the GPS result aligns in the shaded segment, the tumor biology is consistent with the clinical criteria confirming the clinical risk. In addition to providing an estimate of the overall likelihood of favorable pathology, the report breaks this end point out into its 2 components of the likelihood of freedom from high grade disease and the likelihood of freedom from nonorgan confined disease. For some clinicians in certain clinical settings these individual components may be informative for treatment planning. The incorporation of the GPS
375
and NCCN clinical risk provides a more individualized and precise risk assessment to allow more informed treatment recommendations and decisions. There should be an indepth discussion between the physician and the patient on all the components of the report (see Appendix). Once this has occurred, both parties can make a final informed treatment decision.
Costs and Reimbursement
Cost is an important consideration in cancer care. By providing information that helps determine which patients need immediate treatment vs those who may benefit from active surveillance, Oncotype DX may help manage the overall cost of care.10 At present, Oncotype DX is covered by Medicare for patients with very low and low risk prostate cancer by NCCN classification. Genomic Health provides a Genomic Access Program (GAP) to help patients obtain coverage for Oncotype DX from their insurers or third-party payers as regional differences in coverage may exist.11 Financial assistance is also available to those with limited incomes. In cases in which all appeals have been exhausted and the claim is not reimbursed, the patient may be responsible for the balance or a portion of the remaining invoice when appropriate policies have been followed and indicate a balance is due. In such instances Genomic Health provides assistance for patients with financial hardship as well as a program for uninsured and underinsured patients (all programs based on financial eligibility). The GAP provides value in its ability to determine whether the patient’s insurance plan is likely to cover the Oncotype DX test, to obtain prior authorization, if required, and to assist in the appeals process if the claim is denied by the insurance company.
Patient Communication
Effective patient-clinician communication is crucial.12 When communicating with patients who have very low, low or intermediate risk PCa about the role of the GPS in their treatment plan, best practices should be followed. The physician and the patient should have a face-to-face discussion about the results (see Appendix). This may be done once the biopsy and Oncotype DX results have been received. However, the final approach is at the physician’s discretion. The physician should highlight the 3 important sections of the Oncotype DX report, namely 1) the overall likelihood of favorable pathology (defined as freedom from high grade [dominant Gleason pattern 4 or any pattern 5] and/or nonorgan confined disease), 2) the likelihood of
376
Genomic Prostate Score in Clinical Practice
freedom from high grade disease and 3) the likelihood of freedom from nonorgan confined disease. The incorporation of the GPS with a NCCN clinical risk assessment provides a more individualized and precise risk assessment to allow more informed treatment recommendations and decisions. The physician should explain the natural history of the disease. Selected men with low risk PCa can be safely treated with a strategy of AS, involving careful monitoring with serial PSA measurements, examinations and biopsies, with curative intervention provided to those men with evidence of disease progression. Nevertheless, AS remains underused, largely due to concerns about the accuracy of existing risk stratification tools available at the time of biopsy (especially given the issues of prostate tumor heterogeneity and the potential for under sampling by conventional biopsy approaches). The physician should provide supplemental educational materials on PCa treatments. Finally, the physician should inform the patient about available financial support services through GAP. Allow approximately 30 minutes for this conversation and address any additional questions
Conclusions
Several genomic tests are commercially available and others will soon enter the market. These tests were not all specifically designed for the same target population as Oncotype DX. Some are for high risk disease after prostatectomy while others are for negative biopsies. The Oncotype DX prostate cancer assay is validated and is specifically intended for men with newly diagnosed low to low-intermediate risk PCa. The assay assesses the biology of an individual’s prostate tumor from biopsy tissue, and measures the activity of key genes and biological pathways associated with and predictive of PCa aggressiveness. The gene expression information is then used to calculate a GPS. The GPS results reveal the underlying biology of a patient’s cancer, and combined with clinical risk, can help guide initial treatment decisions at the time of diagnosis. The GPS allows us to better understand the biology of each patient’s prostate cancer, providing more accurate risk assessment to guide treatment recommendations. Based on real-world evidence and our clinical experience, the biopsy based genomic assay provides additional information on the biology of the cancer at biopsy, when important treatment decisions regarding active surveillance or immediate treatment need to be made.
Acknowledgments
Karen Ventii provided editorial assistance.
Appendix. Patient Summary Box Important information that each patient needs to know: The rationale for the test (communicate that biopsy may provide equivocal results and, therefore, a genomic test should be considered to help provide more clarity) The decision to test The test results and implications for guiding the treatment approach What active surveillance is The actionable impact of the results Costs
References 1. Goodman M, Ward KC, Osunkoya AO et al: Frequency and determinants of disagreement and error in Gleason scores: a population-based study of prostate cancer. Prostate 2012; 72: 1389. 2. National Cancer Institute: Prostate-Specific Antigen (PSA) Test. Available at http://www.cancer.gov/types/prostate/psa-fact-sheet. 3. Summaries for patients. Screening for prostate cancer: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med 2012; 157: 1. 4. Crawford ED, Ventii K and Shore ND: New biomarkers in prostate cancer. Oncology (Williston Park) 2014; 28: 135. 5. Knezevic D, Goddard AD, Natraj N et al: Analytical validation of the Oncotype DX prostate cancer assayda clinical RT-PCR assay optimized for prostate needle biopsies. BMC Genomics 2013; 14: 690. 6. Klein EA, Cooperberg MR, Magi-Galluzzi C et al: A 17-gene assay to predict prostate cancer aggressiveness in the context of Gleason grade heterogeneity, tumor multifocality, and biopsy undersampling. Eur Urol 2014; 66: 550. 7. Cullen J, Rosner I, Brand T et al: A prospectively-designed study to determine the association of a 17-gene genomic prostate score with recurrence following surgery for localised prostate cancer. Presented at ESMO (European Society for Medical Oncology) 2014, Madrid, Spain, September 26-30, 2014. 8. Canfield SE, Kibel AS, Kemeter MJ et al: A guide for clinicians in the evaluation of emerging molecular diagnostics for newly diagnosed prostate cancer. Rev Urol 2014; 16: 172. 9. Montironi R, Lopez-Beltran A, Mazzucchelli R et al: Contemporary update on pathology-related issues on routine workup of prostate biopsy: sectioning, tumor extent measurement, specimen orientation, and immunohistochemistry. Anal Quant Cytopathol Histpathol 2014; 36: 61. 10. Badani K, Kemeter M, Febbo P et al: The impact of the Oncotype DX genomic prostate score (GPS) on initial treatment recommendations (TR) for men with newly diagnosed clinically low-risk prostate cancer (PCa). Presented at 15th Annual Meeting of the Society of Urologic Oncology, Bethesda, Maryland, December 3-5, 2014. 11. The Genomic Access Program (GAP) Brochure. Available at www.oncotypeDX.com. Accessed October 21, 2014. 12. Epstein RM and Street RL Jr: Patient-Centered Communication in Cancer Care: Promoting Healing and Reducing Suffering. NIH Publication No. 07-6225. Bethesda, Maryland: National Cancer Institute 2007.
Genomic Prostate Score in Clinical Practice
377
Editorial Commentaries
There are several options for reducing the harms of over detection and overtreatment of localized prostate cancer while preserving the 50% reduction in mortality that occurred during the last 25 years. These include using smarter screening practices, new biomarkers, and imaging to maximize the detection of high grade disease and minimize the detection of low risk tumors, and using active surveillance when low risk disease is identified. There is ample evidence that our use of active surveillance for low risk patients is increasing.1,2 However, there remains a greater than 20% risk that we misclassify a “tiger” as a “pussycat” using the relatively crude tools at our disposal such as PSA, grade, Gleason score and other histopathological findings on biopsy and, more recently, magnetic resonance imaging findings.3 This generates anxiety and uncertainty, and dampens enthusiasm for this approach. Ultimately molecular markers will supplant the existing tools as they have in many other cancers. Albala et al describe the novel test Oncotype DX which refines risk stratification beyond the NCCN risk strata. The factors that will drive the adoption of this and other new tests are 1) the science, ie how well the test distinguishes between tigers and pussycats, and 2) the implementation, ie
The Genomic Prostate Score (Oncotype DX) is 1 of 2 commercially available tests (along with ProlarisÒ) that assess the molecular features of prostate biopsy tissue. Test results have the potential to influence the decision for surveillance vs treatment as the initial strategy for very low, low and low-intermediate risk prostate cancer. Albala et al aim to standardize operational best practices and suggest an algorithm for the integration of the GPS into clinical practice. We believe that an initial assessment of the appropriateness of surveillance vs treatment can be determined based on the clinical characteristics of tumor and patient, but the final decision must be made after adequate patient education, some sort of additional testing and due consideration of patient preferences. It remains to be determined what combination of molecular testing, magnetic resonance imaging and/or repeat prostate biopsy is best to inform shared decision making after initial biopsy. As the authors state, if a test is to be ordered, it should provide actionable information for the patient. At this time uncertainty remains regarding how much weight to give to genomic testing with regard to shared
the test’s marketing, ease of use, cost and extent to which it provides digestible, actionable information to the patient and provider. The greater our certainty in distinguishing tigers from pussycats, the more confidently we can promote active surveillance for appropriate candidates, and the more we can improve on the risk-benefit ratio of prostate cancer detection and treatment. Daniel A. Barocas Vanderbilt University Medical Center Nashville, Tennessee
References 1. Ritch CR, Graves AJ, Keegan KA et al: Increasing use of observation among men at low risk for prostate cancer mortality. J Urol 2015; 193: 801. 2. Cooperberg MR and Carroll PR: Trends in management for patients with localized prostate cancer, 1990-2013. JAMA 2015; 314: 80. 3. Tosoian JJ, Sundi D, Trock BJ et al: Pathologic outcomes in favorable-risk prostate cancer: comparative analysis of men electing active surveillance and immediate surgery. Eur Urol 2016; 69: 576.
decision making. Thus, test results may not be “actionable.” In addition, we caution against the standardized (or routine) use of any test (as implied in figure 2) because there are some patients for whom the treatment decision is almost certain not to change. We would encourage clinicians to order such testing only after face-to-face discussion of the initial biopsy results with the patient and a discussion regarding the merits of the GPS and its alternatives. Brian R. Lane Michigan State University College of Human Medicine Spectrum Health Medical Group Grand Rapids, Michigan and Michael L. Cher Karmanos Cancer Institute Department of Urology Wayne State University School of Medicine Detroit, Michigan
378
Genomic Prostate Score in Clinical Practice
Reply by Authors
Many localized prostate cancers have similar histological features at diagnosis. However, they often exhibit diverse biological behavior that may not be captured by traditional clinical risk categories (eg NCCN criteria). Traditional risk categories are based on clinical findings, including PSA and Gleason grade or stage on digital rectal examination. Uncertainty around whether to treat immediately has been caused in part by the divergent behavior of localized prostate cancers and the limitations of current diagnostic techniques. Given the lack of accurate risk assessment tools for early PCa as well as the potential inaccuracy of current biopsy techniques, additional prognostic methods are needed to biologically stratify this disease. Recent NCCN guidelines recognize that genomic testing should have a role in PCa treatment decision making.1 With such recognition there will likely be more awareness of the role of genomics in everyday clinical practice, helping physicians better decide on treatment vs observation.
Collectively the genomic tests are relatively new to PCa and are following a trajectory similar to that of prognostic breast cancer assays. The Oncotype DX prostate cancer assay is a commercially available tissue based assay that helps to confirm clinical judgment regarding the management of PCa (reference 6 in article). More information is needed on the practical application of the test. This review builds on our clinical expertise (depth of experience and knowledge about the assay) to provide easy to use, clear-cut and practical guidance for physicians on its use in patients with low and low-intermediate risk PCa.
Reference 1. NCCN: National Comprehensive Cancer Network. Prostate Cancer Early Detection. Version 1, 2016.