Integrated clinical-genomic risk groups (Decipher + NCCN)

Early signal

Integrated clinical-genomic risk groups (Decipher + NCCN)

ForLocalized prostate cancer, NCCN low to high risk, with Decipher available

TL;DR10y DM 3.5%/29.4%/54.6% for clinical-genomic low/int/high vs NCCN 7.3%/9.2%/38.0%/39.5% in localized prostate cancer; 30% reclassified.

Why it mattersRadiation oncology

The pre-RT setting is the biopsy validation cohort (n=235), where 10y DM ran 0%/25.9%/55.2% across clinical-genomic tiers. The reclassification hits the NCCN favorable/unfavorable intermediate boundary hardest (56.8% and 36.0% moved in prospective cohort I), which is the line that gates adding ADT to RT.

Monday clinic

In NCCN intermediate-risk localized prostate cancer, a biopsy Decipher score can move the pt across the favorable/unfavorable line that frames the RT plus ADT discussion; it does not establish that treating to the new tier changes metastasis or survival.

13 details 3 trials watching

Design

Retrospective development (training n=756, RP genomic data) and validation (n=235, biopsy genomic data) with competing-risks DM estimates; training median follow-up 8 years. Two prospective cohorts (n=4,960 RP; n=977 biopsy) used for reclassification only.

Population & inclusion

Localized prostate cancer across NCCN low to high/very-high risk. Training cohort second-line therapy included adjuvant RT 7%, salvage RT 13%, salvage ADT 18%.

Endpoints

Outcome: distant metastasis at 10 years; PCSM secondary. Discrimination by time-dependent c-index; reclassification vs NCCN three-, four- and six-tier systems.

Results

Clinical-genomic tiers separated DM where NCCN fav-int and unfav-int did not split from their neighbors; c-index 0.77 vs 0.68 for NCCN four-tier in training.

Risk group	10y DM, training	10y DM, validation	HR DM, training (95% CI)
NCCN low	7.3%	n/a	Ref
NCCN fav-int	9.2%	n/a	1.2 (0.5-3.0)
NCCN unfav-int	38.0%	n/a	5.4 (2.8-12.0)
NCCN high	39.5%	n/a	6.0 (3.2-13.0)
CG low	3.5%	0%	Ref
CG int	29.4%	25.9%	9.3 (4.8-21.5)
CG high	54.6%	55.2%	21.9 (11.1-50.4)

System	Cohort	C-index (95% CI)
CG 6-tier	Training	0.77 (0.72-0.81)
NCCN 4-tier	Training	0.68 (0.64-0.72)
NCCN 3-tier	Training	0.65 (0.60-0.69)
CAPRA score	Training	0.68 (0.62-0.74)
CG 6-tier	Validation	0.84 (0.61-0.93)

vs leading data

Benchmarked against NCCN risk groups and CAPRA (continuous score 0.68, groups 0.64 in training), both outperformed on DM discrimination. The authors frame NCCN as built to predict biochemical recurrence, which ICECaP found is not a surrogate for survival.

Applies to

localized prostate cancer with a Decipher result from RP or biopsy tissue, NCCN low through high risk

Does not represent node-positive or metastatic disease, or a population in which reclassification has been shown to change treatment benefit.

Limitations

Adjusting for ADT and RT cuts clinical-genomic HRs from 9.3 and 21.9 to 3.5 and 6.1, so part of the gradient tracks treatment received. Training specimens date from 1997-2004, and many authors are GenomeDx employees.

Discussion

The system turns a continuous classifier score into categories that fit existing NCCN-based decision rules, which is its practical value. It does not answer whether treatment intensification or de-intensification by the new tier changes outcomes.

Rationale

Retrospective development with a 235-pt biopsy validation set and sparse low-risk events; prognostic only, no evidence that reclassification-driven treatment changes outcomes.

Open questions

Does genomic reclassification predict ADT benefit with RT?

active Two Studies for Patients With Unfavorable Intermediate Risk Prostate Cancer Testing Less Intense Treatment for Patients With a Low Gene Risk Score and Testing a More Intense Treatment for Patients With a Higher Gene Risk Score, The Guidance Trial Phase 3

n=2050 · primary completion 2026-11 · phase 3 Decipher-guided RT ± ADT/darolutamide by risk

active A Study of Stereotactic Body Radiation Therapy (SBRT) Without Androgen Deprivation Therapy (ADT) in People With Unfavorable Intermediate-Risk Prostate Cancer Phase 2

n=222 · primary completion 2027-12 · UIR SBRT without ADT, Decipher-guided selection
Independent validation in larger biopsy-based cohorts

active Validation Study on the Impact of Decipher® Testing - VANDAAM Study

n=240 · primary completion 2023-07 · Decipher accuracy in AAM vs non-AAM, 1st-line pts

📚 Sources · 📄 1 paper

📄 PAPER Spratt, Daniel E.; Zhang, Jingbin; Santiago-Jiménez, María et al. · Journal of Clinical Oncology (2018-02)

Development and Validation of a Novel Integrated Clinical-Genomic Risk Group Classification for Localized Prostate Cancer

doi

Abstract

Purpose It is clinically challenging to integrate genomic-classifier results that report a numeric risk of recurrence into treatment recommendations for localized prostate cancer, which are founded in the framework of risk groups. We aimed to develop a novel clinical-genomic risk grouping system that can readily be incorporated into treatment guidelines for localized prostate cancer. Materials and Methods Two multicenter cohorts (n = 991) were used for training and validation of the clinical-genomic risk groups, and two additional cohorts (n = 5,937) were used for reclassification analyses. Competing risks analysis was used to estimate the risk of distant metastasis. Time-dependent c-indices were constructed to compare clinicopathologic risk models with the clinical-genomic risk groups. Results With a median follow-up of 8 years for patients in the training cohort, 10-year distant metastasis rates for National Comprehensive Cancer Network (NCCN) low, favorable-intermediate, unfavorable-intermediate, and high-risk were 7.3%, 9.2%, 38.0%, and 39.5%, respectively. In contrast, the three-tier clinical-genomic risk groups had 10-year distant metastasis rates of 3.5%, 29.4%, and 54.6%, for low-, intermediate-, and high-risk, respectively, which were consistent in the validation cohort (0%, 25.9%, and 55.2%, respectively). C-indices for the clinical-genomic risk grouping system (0.84; 95% CI, 0.61 to 0.93) were improved over NCCN (0.73; 95% CI, 0.60 to 0.86) and Cancer of the Prostate Risk Assessment (0.74; 95% CI, 0.65 to 0.84), and 30% of patients using NCCN low/intermediate/high would be reclassified by the new three-tier system and 67% of patients would be reclassified from NCCN six-tier (very-low– to very-high–risk) by the new six-tier system. Conclusion A commercially available genomic classifier in combination with standard clinicopathologic variables can generate a simple-to-use clinical-genomic risk grouping that more accurately identifies patients at low, intermediate, and high risk for metastasis and can be easily incorporated into current guidelines to better risk-stratify patients.

Retrospective Early signal

The longer read

The core argument of this paper is less that Decipher adds prognostic information, which earlier work had already suggested, and more that a continuous genomic score is clinically inert unless it maps onto the categories treatment recommendations are written in. Collapsing NCCN points and Decipher points into a 0 to 5 scale is a pragmatic answer to that, and the training data show why it was worth attempting: in this RP cohort the NCCN four-tier system did not distinguish low from favorable-intermediate risk (HR 1.2, 95% CI 0.5 to 3.0) or unfavorable-intermediate from high risk (10-year DM 38.0% vs 39.5%). A risk system whose middle boundaries carry little prognostic separation for metastasis is a weak foundation for decisions such as whether to add ADT to RT, and the clinical-genomic tiers spread DM from 3.5% to 54.6%.

The discrimination headline deserves a closer read. The abstract sets the clinical-genomic c-index of 0.84 against NCCN, but 0.84 comes from the 235-patient biopsy validation cohort with a confidence interval of 0.61 to 0.93. The within-cohort comparison in training, 0.77 for the six-tier system vs 0.68 for NCCN four-tier, is the fairer contrast and still favors the new system, but by a smaller margin than the abstract implies. The validation hazard ratios, with upper bounds of 2,727.6 and 7,969.6, reflect a low-risk group with no metastatic events at 10 years; that is reassuring directionally and nearly uninformative about magnitude.

The more important methodological signal is in the treatment-adjusted model. Once ADT, RT and institution are accounted for in training, the intermediate and high clinical-genomic HRs fall from 9.3 and 21.9 to 3.5 and 6.1. They remain significant, and the NCCN HRs shrink too (to 2.5 for both unfavorable-intermediate and high), so the relative advantage holds. But it means part of the raw gradient reflects who received salvage and adjuvant therapy, and a reader should anchor on the adjusted magnitudes when explaining risk to a patient.

For a radiation oncologist, the most relevant finding is where reclassification concentrates. In the large prospective RP cohort, 56.8% of NCCN favorable-intermediate and 36.0% of unfavorable-intermediate patients moved tier, which is exactly the boundary that frames short-course ADT with RT. The training set, however, is post-prostatectomy tissue from 1997 to 2004, and the biopsy validation cohort that mirrors a pre-RT decision is small. The prospective cohorts contribute reclassification counts but no outcomes, so the 30% figure describes how often the label changes, not how often the label change is correct.

What the study cannot settle is prediction. A prognostic tier tells you baseline risk; it does not tell you that a reclassified favorable-intermediate patient benefits from ADT, or that a reclassified unfavorable-intermediate patient can safely omit it. That requires randomized data stratified or selected by the score. The extensive author employment by the manufacturer does not invalidate the analysis but argues for independent replication, ideally in biopsy cohorts treated with contemporary RT.