Content Type: Poster
1Michael J Pishvaian, 2Edik Matthew Blais, 2Dzung Thach, 3Jonathan R Brody, 4Lynn M Matrisian, 2David C Halverson, 2Patricia DeArbeloa, 3Flavio G Rocha, 5Andrew E Hendifar, 6,2Emanuel F Petricoin III
1Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Washington, DC; 2Perthera AI, Mclean, VA; 3Oregon Health & Science University, Portland, OR; 4Pancreatic Cancer Action Network, Manhattan Beach, CA; 5Cedars-Sinai Medical Center, Los Angeles, CA; 6George Mason University, Manassas, VA
BACKGROUND:
- Nearly 50% of pts with mPDAC never receive a 2nd line of therapy for metastatic disease following frontline FFX or GA.
-
Genomic alterations in the DDR pathway2 (e.g. BRCA1/2) are associated with increased progression-free survival (PFS) on platinum-containing regimens (e.g. FFX), but other biomarkers that predict benefit from GA and/or FFX in mPDAC remain unexplored.
-
Here, we used a machine learning approach to gain new data-driven insights from the mutational landscape in mPDAC and validate the PDACai signature in predicting relative benefit from FFX and GA.
METHODS:
We analyzed real-world outcomes from 711 pts with mPDAC who underwent clinical genomic profiling via the Know Your Tumor® program or were referred to Pertherab treating oncologists. Chart-abstracted PFS data on either 1st line FFX or GA were split (60:40) into independent training and validation cohorts for each regimen. All models integrate a shared set of 33 clinical and lab agnostic molecular features derived from clinical NGS testing reports. PDACai benefit scores predicted by FFX or GA models were evenly binned into three relative prediction groups representing lower, middle, and upper tertiles. Statistical differences in median PFS/OS were evaluated using ordinal Cox regression in each cohort (hypothesis:upper>middle>lower?).
RESULTS:
KM curves of PFS on 1st line therapies from pts allocated to independent training and validation cohorts. Actual median PFS [plus 95% CI] in months were summarized in pts assigned to lower, middle, or upper thirds based on relative PDACai predictions. The predictive utility of PDACai was confirmed in the independent validation cohorts by comparing PFS across tertiles [plus 95% CI].
The landscape of FFX versus GA percentiles across all cohorts highlights how the most important variables for FFX (DDR Network2) and GA (WNT Network) are enriched in a treatment-specific manner respectively for patients with higher PDACai values. Top PDACai pathway-level features are highlighted here for patients with genomic alterations in DDR (BRCA1/2, PALB2, CHEK1/2, ATR/ATM, FANC/MRN, etc2), WNT (RNF43, APC, GNAS, CTNNB1), or CDK (CDKN2A, CDK4/6, CCND1/2/3, RB1) gene networks.
Time-averaged performance (higher is better) assessed within each cohort comparing PFS against both PDACai model predictions. FFX PDACai was generally more predictive of PFS for FFX outcomes than GA outcomes (and vice versa for GA PDACai).
CONCLUSIONS:
- Response to chemotherapy is heterogeneous and difficult to predict in pts with mPDAC.
- Using RWE, PDACai signatures successfully predicted relative differences in PFS for both FFX and GA.
- Further efforts to optimize predictions that distinguish response to 2nd line variations of FFX are underway
- The prognostic/predictive importance of molecular features driving PDACai warrant further exploration.
- This study provides a proof-of-concept framework for the prospective validation of AI/ML models that utilize clinical NGS results to deliver insights for treatment sequencing within standard of care.
Currently in peer review. Not Published.