Expert deep dives Decipher, Prolaris, GPS, Artera AI--tests to guide prostate cancer decisions

[The National Comprehensive Cancer Network (NCCN) recently made news when it named the genomic/somatic test Decipher Prostate Genomic Classifier as the first test of its kind to be recognized as having high-level evidence of benefit. That seemed to me to give Decipher an edge in what I see as a competition for patients’ cores to be tested. I asked Decipher’s competitors, Myriad’s Prolaris, or Mdxhealth’s OncotypeDx GPS (Genomic Prostate Score), for comments.

Jonathan Tward, MD, Ph.D., of the University of Utah, took on the challenge in the essay below. He says each test has its own purpose, including Artera AI, which NCCN named as the first AI test in its guidelines.

He also is a member of the NCCN panel and so can’t comment directly on NCCN’s decisions.

Meanwhile, researchers at the University of Michigan are running a randomized controlled study comparing the three tests so we may have some answers in the future.

Tward discloses: He has led studies funded by Myriad.

My disclosure: As a freelance writer, not as editor of this newsletter, I helped research and prepare a white paper two years ago about genomic tests on behalf of Prolaris.

Dr. Tward responded to my article entitled “Decipher gets boost in guidelines in molecular genomics derby for testing prostate cancer.”

If GPS wants to respond, I am happy to run it, too. Even expert opinions can differ. Discuss with your urologist.

—Howard Wolinsky)

By Jonathan Tward, MD, PhD

I am delighted to respond. It is interesting that you characterize the biomarker space as a "derby.” I don't think men should look at these biomarkers as "which is best" or a race to a finish line, but they should look at them at "which is the right tool for the question I am asking?"

For full disclosure, I have served as the principal investigator for several of the critical studies funded by Prolaris to create and validate their treatment threshold recommendations. As such, I have served as a scientific advisor to Myriad, have received research funding from them, and continue to consult for them. Nevertheless, I have dedicated most of my career to developing better tools for men with prostate cancer, and I have also previously received compensation from Decipher and conducted research work with Veracyte (Decipher), Artera AI, and others.

I want to say that Decipher is a perfectly good test, in that it does give a prognosis (risk of metastasis) after therapy. However, the guidance on intensifying or deintensifying therapy that Decipher offers is not based on the actual risk of progression, but on a molecular score threshold (0.6) which was designed to ask what the risk of metastasis would be after treatment was rendered (radical prostatectomy). This is an important point: using it for biopsy tissue pre-treatment while also using the same risk scoring system developed for post-treatment tissue creates problems in interpreting what to do.

For example, a patient with NCCN Favorable Intermediate-Risk prostate cancer, with a decipher score of precisely 0.6 ("Decipher high risk") would see a score report that suggests the patient should consider "intensifying" treatment. Yet, that patient’s 10-year risk of metastasis following treatment in general is only 2.5%. If, for example, you chose to intensify treatment by adding ADT [Androgen Deprivation Therapy] to radiation therapy, prospective randomized trials would suggest that at best, you may cut that risk to 1.25%1. Would a man really want to endure six months of ADT for an approximately 1% smaller chance of getting mets? Perhaps some might, but most would not. Importantly, the Decipher test does not report what the risk reduction would be if they did thing A versus thing B (intensified treatment).

In contrast, Prolaris was specifically developed on biopsy tissue to directly address the question of "Is it safe for a man to go on active surveillance,” and it will reveal the precise risk of death if a man does that. Prolaris was tested on a prospectively followed of cohort men who actually received Active Surveillance for a death outcome 2,3. To the best of my knowledge, the evidence supporting Decipher in active surveillance utilized an endpoint of time to active treatment.4 The threshold for surveillance vs not that Prolaris puts on the score report for men to safely survey was shown to not lead to any death in an American study cohort 3,5.

As long as we are talking about active surveillance, there is something interesting about Decipher versus Prolaris. These data were obtained from the University of Michigan MUSIC Collaboration. The world of oncology knows that the vast majority of men with grade 1 prostate cancer should be surveyed. You quoted Dr. Todd Morgan from U Mich in your last article who was the senior author of a study from the Michigan Urological Surgery Improvement Collaborative (MUSIC) evaluating the performance of Decipher, Prolaris and GPS for men making active surveillance decisions. 5 In that study, the Prolaris, Decipher, and GPS were compared. They discovered that 58% of men receiving Decipher testing with GRADE GROUP 1 were above the Decipher's active surveillance threshold (compared to Prolaris at 14%). It seems highly odd that a test would imply that more than half of men with Gleason 6 should be routinely treated.

Interestingly, men with a higher NCCN risk group (favorable intermediate-risk prostate cancer) had a LOWER threshold (50%) for receiving treatment on Decipher testing than grade group 1 patients. Continuing on, Decipher was back at approximately 58% for grade group 2 patients.

It is not clear to me how Decipher scores can decrease in a patient population that should be at a higher risk (favorable intermediate) than most Grade Group 1, and have the same risk as grade group 2. The Prolaris' 14% estimator for treatment in this Grade Group 1 population is close to what you might expect for a low-risk Grade Group 1 population. It then incrementally increased in risk (~27% for favorable intermediate and 60% for grade group 2). Prolaris’ treatment guidance for not pursuing active surveillance is more proportional to the findings of the ProtecT randomized trial (nothing to do with biomarkers), which randomized low-and favorable intermediate-risk men to surveillance vs. RP vs. radiation. 6 In that study, 10% of the men in the surveillance group developed metastasis by 15 years, whereas 5% in the treatment arms did. Another way to look at the ProtecT trial is that 1 in 20 men get burned by choosing surveillance because they developed metastases that they otherwise would have averted.

Taking the findings of MUSIC and ProtecT together, one might extrapolate that Decipher could suggest as many 55% of men should avoid surveillance whereas Prolaris might suggest about 20% of people might do so. The ProtecT Trial proved unequivocally that the risk of death was equivalent in men with intermediate-risk or lower prostate cancer randomized to surveillance versus surgery or radiation and only a slightly worse risk of metastasis in men receiving active surveillance. Therefore, patients should be very cautious about actioning tests that suggest they should mostly receive treatment for low- and favorable-risk cancers.

One advantage of Prolaris is that it precisely reports the risk of metastasis in a man with localized prostate cancer choosing between monotherapy (RP or radiation) and precisely what the risk reduction would be if they actually intensified treatment by choosing multimodality radiation plus ADT instead. This risk calibration was performed by taking the risk of numerous retrospective/prospective registries to fix the monotherapy estimate and then data from many prospective randomized trials to ascertain the risk reduction 7,8. This was an extremely robust analysis that used Level 1 evidence from many prospective randomized studies to determine the risk reduction. Prolaris precisely describes the difference in metastasis outcomes for men receiving treatment A (monotherapy) versus treatment B (RT+ADT) and tells me precisely what the risk of death a man on surveillance has. This is something that none of the other classifiers can currently perform.

I think there is a lot of misunderstanding about what Simon Level 1 evidence means. The NCCN does indicate that Decipher has Level 1 evidence by Simon Criteria. However, it is important to understand that performing post-hoc biomarker testing on patients enrolled in completed prospective trials (which is Level 1 evidence for the trial question) is different from saying that the trials were designed to test the biomarker itself for its performance. No prostate biomarker has a trial completed where the performance of the biomarker itself is the trial question, although the NRG is currently testing Decipher in that way. We should have the results of those critical trials in 5 to 10 years. I would posit that any prognostic biomarker can be reliably tested on prospective or retrospectively collected data, as long as the data quality is high (this would be any study with Simon Levels 1-3). However, for a marker to be a predictive biomarker, it would require Simon Level 1 evidence. A prognostic biomarker provides information about the patient's overall cancer outcome, regardless of therapy, whereas a predictive biomarker provides information about the likely benefit from treatment or the likelihood of disease progression in response to a specific therapy. The trial that Decipher tested based on biopsy specimens was RTOG 0126.9

In this trial, the question asked was whether men with intermediate-risk prostate cancer should receive lower doses versus dose-escalated radiation therapy. This trial showed that Decipher could predict the risk of metastasis for everyone enrolled in the trial, and that persons who were Decipher low risk did not benefit as much from dose escalation as did patients with greater than Decipher low risk. As such, the evidence supports the idea that if a man chooses radiation therapy, they might consider de-intensifying radiation dose if they had a low Decipher score. However, testing a biomarker in a Level 1 evidence trial does not mean it is a great biomarker for every endpoint you can imagine...it merely means that it was tested on patients enrolled in a Level 1 evidence trial. Although RTOG 0126 did show Decipher prognosticates risk of metastasis, it doesn't really reveal who should get treatment intensification outside of the realm of radiation dose escalation. A more common question in this patient population is who should receive ADT, which this trial did not address. Another way of looking at RTOG 0126, a study of all intermediate-risk patients (including favorable and unfavorable) is that only 9% of the trial population tested was classified as Decipher High. The confidence interval for the risk of metastasis in the RTOG 0126 study for men with a Decipher "High" classification was 0%-36% risk at ten years. This difference is so broad that it is not clear to me how this biomarker can be used to make an ADT treatment decision if the risk of metastasis could be anywhere in that range. The contemporary care standard for men undergoing radiation therapy is that everyone should receive dose-escalated radiation therapy; therefore, Decipher may allow doctors to consider dose de-escalation in intermediate-risk patients based on this study. However, it does not inform the discussion of what the risk reduction would be if the decision was to add ADT to radiation or not.

(Dr. Jonathan Tward, University of Utah.)

To summarize, one must choose a test based on the question being asked. For a newly diagnosed prostate cancer patient who has not yet received therapy, I, personally, have the following questions:

1) Is active surveillance appropriate? Since Prolaris was designed and tested around this endpoint it has the best performance and gives an exact risk of death from prostate cancer over the next decade if surveillance is chosen. The other tests do not provide that information.

2) What is the risk of metastasis after treatment if a patient chooses radiation alone versus radiation plus ADT? Prolaris will provide the precise risk of metastasis for both of these scenarios, so that the patient can decide if they want to accept the additional toxicity of ADT against its benefit.

3) Is there a chance that ADT is futile (will not work regardless of the risk of metastasis and death) in men with intermediate-risk prostate cancer? In this case, Artera AI has the best data as it can specifically predict the futility of ADT in intermediate-risk patients.

After prostatectomy, I have different questions:

Q) In a man who has undergone radical prostatectomy and is going to receive salvage radiation, should I also use ADT? In this case, Decipher is exceptional because it was developed specifically for estimating metastasis after prostatectomy and validated in a prospective trial to predict the benefit of ADT in this context.

So, you can see that there is no single test that suits all clinical scenarios. Each test has its own niche, and it is important to understand the evidence-based clinical utility of the tests within their study-designed context.

The table below summarizes the evidence for various commercially available biomarkers. It is not an exhaustive list, nor does it list all the study endpoints.  However, it may help readers understand how biomarkers can be used for various questions about the risk of metastasis or death from prostate cancer.

(Source: Dr. Jonathan Tward, University of Utah.)

(References below.)

Quick Q&A with Dr. Tward

By Howard Wolinsky

Since I had Dr. Jonathan Tward’s attention, I thought I’d ask some questions I’d been wondering about regarding genomic testing of prostate cancer in men with Grade Group 1 (Gleason 6) who are on Active Surveillance (AS). I had heard from some experts that genomic testing was unnecessary for men with Grade Group 1 (Gleason 6) prostate cancer. He helped me get straight why a repeat genomic test can make sense if your Gleason score has risen. Also, he clarifies why it can make sense to undergo a genomic test as a confirmatory move before going on AS.

The Active Surveillor: Which men with Gleason 6 do you think should have a genomic/somatic test? High-volume Gleason 6? Or other categories? Is high anxiety level a good reason? And do you recommend Prolaris for them? Or would other tests do?

Dr. Jonathan Tward (JT): Personally, I like to get Prolaris on Gleason 6 patients to validate that Active Surveillance is safe. The ProtecT Trial demonstrated that men on Active Surveillance have a higher rate of metastasis but not death by 15 years compared to men who sought immediate treatment. I believe the Prolaris test can help me identify this smaller percentage of men who benefit from immediate treatment. =

Keep in mind that Gleason grading is subjective, not objective. Therefore, one pathologist may score a patient Gleason 6, and another may call it 3+4. Prolaris has a nice balance of objective data (the molecular signature, PSA and age) as well as subjective criteria like Gleason score. 

Other tests did not test Active Surveillance populations for a prostate cancer specific death endpoint, therefore I don't think they are as informative. The other tests have been tested in AS populations for things like time to active treatment, but this is not a significant oncologic endpoint to me. Time to treatment may simply be due to anxiety, arbitrary PSA cutoffs, or again, heterogeneity of Gleason grading on repeat biopsies.

TAS: What do you think about patients undergoing the same test multiple times--that doesn't make sense, right?

Dr. JT: It does not make sense to do multiple tests on the same tissue for the same outcome. However, I do think there is value on repeating Prolaris testing on each biopsy of a man undergoing active surveillance if the Gleason score changes. Regarding different tests, this is something I would refrain from. If the only thing someone wants to know is a ballpark estimate of the risk of metastasis or death after treatment generically, then any of the above tests (Prolaris, Decipher, and GPS) would do. If they actually want actionable information that personalizes treatment decisions based on validated thresholds, they need to consider the tests for the indications I have provided in the table above.

References to Dr. Tward’s article.

1. Kishan AU, Sun Y, Hartman H, et al.: Androgen deprivation therapy use and duration with definitive radiotherapy for localised prostate cancer: an individual patient data meta-analysis. Lancet Oncol 23:304-316, 2022

2. Cuzick J, Berney DM, Fisher G, et al.: Prognostic value of a cell cycle progression signature for prostate cancer death in a conservatively managed needle biopsy cohort. Br J Cancer 106:1095-9, 2012

 3. Cuzick J, Stone S, Fisher G, et al.: Validation of an RNA cell cycle progression score for predicting death from prostate cancer in a conservatively managed needle biopsy cohort. British Journal of Cancer 113:382-389, 2015

 4. Vince RA, Jiang R, Qi J, et al.: Impact of Decipher Biopsy testing on clinical outcomes in localized prostate cancer in a prospective statewide collaborative. Prostate Cancer and Prostatic Diseases, 2021

 5. Kaul S, Wojno KJ, Stone S, et al.: Clinical outcomes in men with prostate cancer who selected active surveillance using a clinical cell-cycle risk score. Personalized Medicine, 2019

 6. Hamdy FC, Donovan JL, Lane JA, et al.: Fifteen-Year Outcomes after Monitoring, Surgery, or Radiotherapy for Prostate Cancer. N Engl J Med 388:1547-1558, 2023

7. Hutten RJ, Odei B, Johnson SB, Tward JD: Validation of the Combined Clinical Cell-Cycle Risk Score to Prognosticate Early Prostate Cancer Metastasis From Biopsy Specimens and Comparison With Other Routinely Used Risk Classifiers. JCO Precis Oncol 8:e2300364, 2024

8. Tward J, Lenz L, Flake DD, II, et al.: The Clinical Cell-Cycle Risk (CCR) Score Is Associated With Metastasis After Radiation Therapy and Provides Guidance on When to Forgo Combined Androgen Deprivation Therapy With Dose-Escalated Radiation. Int J Radiat Oncol Biol Phys 113:66-76, 2022

 9. Spratt DE, Liu VYT, Michalski J, et al.: Genomic Classifier Performance in Intermediate-Risk Prostate Cancer: Results From NRG Oncology/RTOG 0126 Randomized Phase 3 Trial. Int J Radiat Oncol Biol Phys, 2023

 10. Esteva A, Feng J, van der Wal D, et al.: Prostate cancer therapy personalization via multi-modal deep learning on randomized phase III clinical trials. npj Digital Medicine 5:71, 2022

11. Spratt DE, Tang S, Sun Y, et al.: Artificial Intelligence Predictive Model for Hormone Therapy Use in Prostate Cancer. Res Sq, 2023

12. Nguyen PL, Haddad Z, Ross AE, et al.: Ability of a Genomic Classifier to Predict Metastasis and Prostate Cancer-specific Mortality after Radiation or Surgery based on Needle Biopsy Specimens. Eur Urol 72:845-852, 2017

13. Spratt DE, Zhang J, Santiago-Jiménez M, et al.: Development and Validation of a Novel Integrated Clinical-Genomic Risk Group Classification for Localized Prostate Cancer. J Clin Oncol 36:581-590, 2018

14. Klein EA, Haddad Z, Yousefi K, et al.: Decipher Genomic Classifier Measured on Prostate Biopsy Predicts Metastasis Risk. Urology 90:148-152, 201

15. Spratt DE, Huang H-C, Michalski JM, et al.: Validation of the performance of the Decipher biopsy genomic classifier in intermediate-risk prostate cancer on the phase III randomized trial NRG Oncology/RTOG 0126, American Society of Clinical Oncology, 2022

16. Berlin A, Murgic J, Hosni A, et al.: Genomic Classifier for Guiding Treatment of Intermediate-Risk Prostate Cancers to Dose-Escalated Image Guided Radiation Therapy Without Hormone Therapy. International Journal of Radiation Oncology*Biology*Physics 103:84-91, 2019

17. Cullen J, Kuo HC, Shan J, et al.: The 17-Gene Genomic Prostate Score Test as a Predictor of Outcomes in Men with Unfavorable Intermediate Risk Prostate Cancer. Urology 143:103-111, 202

18. Canter DJ, Freedland S, Rajamani S, et al: Analysis of the prognostic utility of the cell cycle progression (CCP) score generated from needle biopsy in men treated with definitive therapy. Prostate Cancer and Prostatic Diseases 23:102-107, 2020

19. Canter DJ, Reid J, Latsis M, et al: Comparison of the Prognostic Utility of the Cell Cycle Progression Score for Predicting Clinical Outcomes in African American and Non-African American Men with Localized Prostate Cancer. European Urology 75:515-522, 2019

20. Tward J, Lenz L, Flake II DD, et al.: The clinical cell-cycle risk (CCR) score is associated with metastasis after radiation therapy and provides guidance on when to forgo combined androgen deprivation therapy with dose-escalated radiation. International Journal of Radiation Oncology* Biology* Physics 113:66-76, 2022

21. Tward JD, Schlomm T, Bardot S, et al: Personalizing localized prostate cancer: validation of a combined clinical cell-cycle risk (CCR) score threshold for prognosticating benefit from multimodality therapy. Clinical genitourinary cancer 19:296-304. e3, 2021

22. Brawer MK, Cooperberg MR, Freedland SJ, et al.: Development and validation of a mutivariate model combining cell cycle progression score with CAPRA to predict prostate cancer mortality in a conservatively managed cohort. Journal of Clinical Oncology 31:67-67, 2013

23. Spratt DE, Dess RT, Efstathiou JA, et al.: Two Years of Anti-Androgen Treatment Increases Other-Cause Mortality in Men Receiving Early Salvage Radiotherapy: A Secondary Analysis of the NRG Oncology/RTOG 9601 Randomized Phase III Trial. International Journal of Radiation Oncology • Biology • Physics 105:680, 2019

Jonathan Tward, MD, Ph.D., is a tenured Professor in the Department of Radiation Oncology. He is recognized as a major authority and key opinion leader in treating prostate, bladder, and penile cancers. He holds The Vincent P. and Janet Mancini Presidential Endowed Chair at Huntsman Cancer Institute in Genitourinary Malignancies.

Tward is the leader of the Huntsman Cancer Institute's (HCI) Genitourinary Cancers Center (GUCC), coordinating the activities of its approximately 100 members who represent 45 of the University of Utah or HCI Departments or Colleges, and 21 Interdisciplinary Programs. As the leader, Tward is responsible for the strategic vision of the GUCC, and its operations, by directing multidisciplinary collaboration of patient care, community engagement, research, and donor relations. Tward also serves as the physician lead on several operational committees and working groups within the Department of Radiation Oncology.

Dr. Tward is an expert at delivering highly precise and targeted radiation therapy to people afflicted with cancers of the prostate, bladder, kidney, testis, and penis. He uses the vast array of technologies available at the HCI to deliver intensity-modulated radiotherapy (IMRT), image-guided radiotherapy (IGRT), stereotactic body radiotherapy (SBRT), as well as low-dose-rate (LDR) and high-dose-rate (HDR) brachytherapy.

As a practicing physician and scientist, mentor, and educator, Tward has contributed to over 100 published-peer reviewed journal articles. He is the Principal Investigator (PI) on several ongoing investigator-initiated clinical trials and is the site PI on numerous cooperative group or industry-sponsored clinical trials. Tward has received grant funding for his research interests which include "Big Data" analytics, diagnostic biomarker development, comparative effectiveness outcomes studies, and clinical trials. Tward serves as the Utah State Captain for the American Society of Radiation Oncology (ASTRO) and is a strong advocate for cancer patients through federal and state advocacy efforts with Utah's elected federal representatives. He serves on the National Comprehensive Cancer Network (NCCN) clinical practice guidelines committees for prostate, bladder, and penile cancers. These clinical practice guidelines are internationally recognized standards for clinical direction and policy in cancer care. As a member of the NCCN Framework for Resource Stratification subcommittees, Tward helps define appropriate cancer treatment pathways of low- and middle-resource countries where specific diagnostic tests or treatment approaches may be unavailable.

Tward earned a bachelor's degree in biology at the University of California, Los Angeles, a Ph.D. in biochemistry, and an MD at Tufts University in Boston, Massachusetts.

ABCs of BPH: ASPI webinar

Active Surveillance Patients International is holding a webinar “ABCs of BPH” from noon-1:30 p.m. Eastern on Saturday, March 23,. Register here: https://zoom.us/meeting/register/tJYoduChrzIrH9dHBNZXqD_pOUCG85yv_KQF#/registration

Something completely different: a brag on my son David

By Howard Wolinsky

My son David Wolinsky is publishing his first book. I’ve been sworn to silence for years on his project. But on Wednesday, Amazon.com announced preorders for the book. https://www.amazon.com/Hivemind-Swarmed-Conversations-Gamergate-Aftermath/dp/0807017736

It’s not about prostate cancer. I’ve cornered that market.

David’s book covers the Gamergate scandal that hit the gaming industry 10 years ago. Famed documentarian Ken Burns wrote a blub.

Beacon will make a formal announcement about the book in April. The book will be released in August.

So, if you’re interested, please check it out,