Systematic Review Questions Impact of Genomic Classifier Tests in Prostate Cancer Treatment Decisions
核心洞察
A comprehensive systematic review reveals inconsistent influence of genomic classifier tests on risk classification in newly diagnosed prostate cancer (搜索) patients, with varying reclassification rates across different testing platforms.
The ENACT trial showed genomic testing led to increased treatment intensity preferences among urologists, with 2.55 higher odds of recommending active treatment when GPS testing was used.
Research highlights significant gaps in understanding genomic testing utility across racial groups, particularly noting potential missed aggressive tumors in Black men using traditional clinical risk classifiers.
A new systematic review conducted by researchers at the Moffitt Cancer Center has revealed complex and sometimes contradictory findings regarding the impact of genomic classifier (GC) testing in prostate cancer (搜索) management. The study, published in the Annals of Internal Medicine, evaluated the effectiveness of three major genomic tests in influencing risk classification and treatment decisions.
Variable Risk Reclassification Patterns
The analysis of 10 studies examining three prominent genomic tests - Decipher (搜索), Prolaris (搜索), and Oncotype DX Genomic Prostate Score (搜索) (GPS) - showed that risk reclassification patterns varied significantly. In low risk-of-bias observational studies, the majority of very low- or low-risk patients maintained or received lower risk classifications:
- GPS: 88.1% to 100%
- Decipher: 82.9% to 87.2%
- Prolaris: 76.9%
However, contrasting data emerged from the randomized ENACT trial, where GPS testing resulted in risk escalation for a substantial portion of patients - 34.5% of very low-risk and 29.4% of low-risk patients were reclassified to higher risk categories.
Impact on Treatment Decisions
The review examined 14 studies investigating how GC testing influenced treatment choices. Twelve observational studies demonstrated increased recommendations for active surveillance following testing, with changes ranging from 7.5% to 61.8%. However, the ENACT trial revealed different patterns:
- Patients showed modest increases in preference for active treatment
- Urologists demonstrated substantially increased preference for active treatment (from 11.4% to 29.3%) for patients receiving GPS testing
- The odds of urologists recommending active treatment were 2.55 times higher for patients who underwent GPS testing
Racial Disparities and Research Gaps
Dr. Amir Alishahi Tabriz and colleagues identified significant limitations in current research regarding racial and ethnic diversity. Only two studies beyond the ENACT trial examined risk reclassification effects in Black men, with just one investigating the impact on active surveillance selection. This gap is particularly concerning given evidence suggesting that some Black men may harbor genomically aggressive tumors that traditional clinical risk classifiers might not detect.
Expert Commentary
In an accompanying editorial, Drs. Syed Arsalan Ahmed Naqvi and Irbaz Bin Riaz from Mayo Clinic Phoenix emphasized that evidence supporting GCs as predictive biomarkers remains limited. They noted that while similar genomic testing in breast cancer required over a decade of pivotal trials to establish clinical utility, prostate cancer (搜索) applications are still in early stages.
The editorial authors suggested that future developments may require integration of GCs with AI-driven multimodal approaches, tested through prospective randomized controlled trials, to address current gaps in clinical utility while advancing precision oncology.
Study Methodology
The systematic review analyzed 19 studies, including two analyses from the ENACT randomized trial and 17 observational studies. The distribution of genomic tests examined was:
- GPS: 10 studies
- Prolaris: 5 studies
- Decipher: 4 studies
All but one study were conducted in the United States, providing a predominantly U.S.-centric perspective on the clinical utility of these genomic classifiers.