The Prostate · 2026
Incremental Predictive Value of the Oncotype Genomic Prostate Score for Adverse Pathology in Active Surveillance Candidates.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Ozawa Yu, Moschovas Marcio Covas, Sandri Marco, Sharma Rohan, Saikali Shady, Rogers Travis, Patel Vipul
Аннотация
BACKGROUND: Genomic prostate score (GPS) may aid clinical decision-making for active surveillance. We assessed whether GPS adds predictive value beyond established clinical variables for adverse pathology at radical prostatectomy (RP) in active surveillance-eligible patients.
METHODS: We retrospectively analyzed 387 men with National Comprehensive Cancer Network very-low- to favorable intermediate-risk prostate cancer who underwent Oncotype DX testing followed by RP without active surveillance. Multivariable logistic regression models were constructed to develop prediction models for adverse pathology at RP (Grade Group ≥ 3 and/or ≥ pT3a), including clinical variables (age, PSA, PSA density, biopsy Grade Group 2 [vs. 1], and PI-RADS 4/5) with and without GPS. Model performance was assessed using the AUC with 10-fold cross-validation and compared using DeLong's test, continuous net reclassification improvement (NRI), and decision curve analysis. The incremental predictive value of GPS was also separately evaluated in very low/low-risk and favorable intermediate-risk subgroups.
RESULTS: GPS independently predicted adverse pathology (p < 0.001). The addition of GPS increased the AUC from 0.69 to 0.73 (ΔAUC = 0.036; 95% CI: 0.006-0.065; p = 0.018) and improved reclassification (NRI 0.41, 95% CI: 0.20-0.61). Decision curve analysis demonstrated added net benefit at intermediate threshold probabilities (0.40-0.70), with limited benefit at low thresholds (0.0-0.40). Improvement was significant in the favorable intermediate-risk subgroup (ΔAUC = 0.039; 95% CI: 0.011-0.124; p = 0.026) but not in the very low/low-risk subgroup (ΔAUC = 0.018; 95% CI: -0.043-0.049; p = 0.20). Surgical selection bias was the main limitation.
CONCLUSIONS: In this RP cohort, GPS modestly improved prediction of adverse pathology beyond PSA density and MRI, with clinical utility primarily in favorable intermediate-risk patients, where treatment decisions between active surveillance and definitive therapy are uncertain. These findings suggest selective, risk-adapted application of GPS to guide treatment decision-making. Validation in prospective, diverse cohorts is warranted.
METHODS: We retrospectively analyzed 387 men with National Comprehensive Cancer Network very-low- to favorable intermediate-risk prostate cancer who underwent Oncotype DX testing followed by RP without active surveillance. Multivariable logistic regression models were constructed to develop prediction models for adverse pathology at RP (Grade Group ≥ 3 and/or ≥ pT3a), including clinical variables (age, PSA, PSA density, biopsy Grade Group 2 [vs. 1], and PI-RADS 4/5) with and without GPS. Model performance was assessed using the AUC with 10-fold cross-validation and compared using DeLong's test, continuous net reclassification improvement (NRI), and decision curve analysis. The incremental predictive value of GPS was also separately evaluated in very low/low-risk and favorable intermediate-risk subgroups.
RESULTS: GPS independently predicted adverse pathology (p < 0.001). The addition of GPS increased the AUC from 0.69 to 0.73 (ΔAUC = 0.036; 95% CI: 0.006-0.065; p = 0.018) and improved reclassification (NRI 0.41, 95% CI: 0.20-0.61). Decision curve analysis demonstrated added net benefit at intermediate threshold probabilities (0.40-0.70), with limited benefit at low thresholds (0.0-0.40). Improvement was significant in the favorable intermediate-risk subgroup (ΔAUC = 0.039; 95% CI: 0.011-0.124; p = 0.026) but not in the very low/low-risk subgroup (ΔAUC = 0.018; 95% CI: -0.043-0.049; p = 0.20). Surgical selection bias was the main limitation.
CONCLUSIONS: In this RP cohort, GPS modestly improved prediction of adverse pathology beyond PSA density and MRI, with clinical utility primarily in favorable intermediate-risk patients, where treatment decisions between active surveillance and definitive therapy are uncertain. These findings suggest selective, risk-adapted application of GPS to guide treatment decision-making. Validation in prospective, diverse cohorts is warranted.