Main Session

Sep 29

PQA 06 - Genitourinary Cancer, Gynecological Cancer, and Health Care Access and Engagement

3226 - Association of 22-Gene Genomic Classifier with ADT Use with Secondary Radiotherapy: A National Real-World Analysis

02:15pm - 03:30pm ET

Poster Hall - Exhibit Hall A

Screen: 6

POSTER

Presenter(s)

Daniel Gorovets, MD - Memorial Sloan Kettering Cancer Center, New York, NY

A. R. Barsky¹, Q. Joslove Xu², T. P. Kole¹, J. Setton¹, E. Bent¹, Z. R. Moore¹, M. B. Bernstein¹, D. J. Gorovets¹, V. S. Brennan¹, S. M. McBride¹, J. Janopaul-Naylor¹, J. Haseltine¹, D. M. Guttmann³, M. A. Kollmeier⁴, X. Zhao², A. Moradi², Y. Liu², J. Proudfoot², E. Davicioni², and H. Nagar¹; ¹Department of Radiation Oncology, Memorial Sloan Kettering Cancer Center, New York, NY, ²Veracyte, Inc., South San Francisco, CA, ³Memorial Sloan Kettering Cancer Center, New York, NY, ⁴DepMemorial Sloan Kettering Cancer Center, New York, NY

Purpose/Objective(s):

Post-prostatectomy secondary radiation therapy (SRT) with or without androgen deprivation therapy (ADT) is a treatment option for men with adverse pathologic features and/or biochemical recurrence (BCR) after radical prostatectomy (RP). The Decipher genomic classifier (GC) demonstrated differential ADT benefit by post-op GC risk group in RTOG 9601. We examined real-world data (RWD) to assess the association between GC risks and ADT use with SRT, as well as GC-informed temporal prescribing patterns.

Materials/Methods:

We conducted a retrospective analysis of a national clinical–genomic linkage of Decipher GC with longitudinal RWD (Veracyte, San Francisco, CA). Between 2016-2024, 12,987 patients who underwent RP had post-RP GC testing, and subsequently received SRT were analyzed. ADT use was identified from pharmacy and claims records. Chi-square tests compared proportions. Multivariable logistic regression (MVA) identified independent predictors of ADT use, adjusting for GC risk, grade group (GG), surgical margin status (SM), extraprostatic extension (EPE), seminal vesicle invasion (SVI), lymph node invasion (LNI), log2 PSA, and age. Temporal trends in ADT use by GC risk group were assessed using logistic regression with year as a continuous predictor.

Results:

Among 12,987 men receiving SRT, 5,284 (40.7%) received concurrent ADT. ADT utilization increased with GC risk group: 26% in GC-low, 33% in GC-intermediate, and 47% in GC-high (p<0.001). ADT use also increased with GG: 18% in GG1, 32% in GG2, 41% in GG3, 50% in GG4, and 56% in GG5 (p<0.001). ADT use significantly was higher in those with SVI (50% vs 37%, p<0.001) and LNI (56% vs 40%, p<0.001).

On MVA (n=7,176), all GC risk groups were independently associated with ADT use vs. GC-low: GC-int OR=1.32 (1.15-1.52, p<0.001), GC-high OR=1.97 (1.77-2.19, p<0.001). GG, SVI, and LNI were also independently significant (all p <0.001).

Temporal analysis revealed diverging trends by GC risk group: ADT utilization in GC-low declined from 26% (2017) to 22% (2023, p-trend<0.05), while GC-high increased from 44% to 50% (p-trend<0.05), widening the spread from 18% to 28%. Comparing pre- vs. post-RTOG 9601 era analysis, ADT in GC-low decreased from 30% to 24% (p<0.01) while GC-high increased from 48% to 50% (p<0.05) with spread from 18% to 26%.

Conclusion:

In this large, real-world cohort, higher Decipher GC scores were independently and strongly associated with ADT utilization with SRT, with a continuous association across risk groups. Temporal analysis demonstrated widening separation in ADT use by GC risk group over time, with ADT use in GC-low declining, and ADT use in GC-high increasing. These findings suggest increasing temporal trends in GC-informed ADT decision-making for patients receiving SRT.