Validation of PDACai v2.0 in predicting relative benefit from frontline FOLFIRINOX (FFX) and gemcitabine/nab-paclitaxel (GA) for patients (pts) with metastatic pancreatic cancer (mPDAC).

M Michael J. Pishvaian E Edik Matthew Blais (Perthera, South San Francisco, CA) L Lynn M. Matrisian (Pancreatic Cancer Action Network, Manhattan Beach, CA) J Jonathan R Brody (Oregon Health & Science University, Portland, OR) D Dzung Thach (Perthera, Mclean, VA) P Patricia Miren de Arbeloa (Perthera, Mclean, VA) F Flavio G. Rocha J Jennifer W. Chuy (Montefiore Medical Center, Bronx, NY) R Raymond Couric Wadlow (Inova Schar Cancer Institute, Fairfax, VA) A Andrew Eugene Hendifar (Samuel Oschin Comprehensive Cancer Institute, Cedars-Sinai Medical Center, Los Angeles, CA) E Emanuel Petricoin (George Mason University, Manassas, VA)

Abstract

776 Background: Novel biomarkers beyond genomic alterations in DDR pathway genes (e.g. BRCA1/2 ) have the potential to optimize frontline and subsequent treatment decisions for mPDAC. This real-world evidence (RWE) study validates PDACai v2.0, a machine learning framework that leverages clinical NGS data to predict progression-free survival (PFS) on FFX vs GA in individual pts with mPDAC. Methods: We analyzed outcomes from 774 pts with mPDAC who underwent genomic profiling via the Know Your Tumor program or were referred to Perthera by treating oncologists. Independent training and validation cohorts receiving either 1st line FFX or GA were split (60:40). PDACai v2.0 integrates an updated set of clinical (age < 63, sex) and lab-agnostic molecular features derived from genomic testing reports (specific KRAS variants, TP53 GOF vs LOF, DDR pathway redefined). Relative benefit scores predicted by FFX or GA models were evenly binned into three PDACai signature categories representing lower, middle, and upper tertiles for each independent cohort. Statistical differences in median PFS were evaluated using ordinal Cox regression. Results: Median PFS followed PDACai predicted trends for each therapy’s training and validation cohorts. The predictive utility of PDACai was confirmed in the independent validation cohorts when comparing PFS on FFX (p = 0.01637; HR = 0.73 [95% CI: 0.56-0.94]) and GA (p = 0.0007506; HR = 0.62 [95% CI: 0.47-0.82]) across tertiles. Conclusions: Using RWE, the PDACai v2.0 signature successfully predicted PFS differences in pts treated with 1st line FFX or GA in mPDAC. Prospective validation efforts and additional data-driven insights into 2nd line PFS on FOLFOX vs 5FU/nal-irinotecan are underway. Summary of actual PFS in months on 1st line therapies in pts assigned to lower, middle, and upper thirds based on relative PDACai v2.0 scores. Independent Cohort (# Pts) Lower TertilemPFS [95% CI] Middle TertilemPFS [95% CI] Upper TertilemPFS [95% CI] 1st line FFX Training (243) 6.5 [5.5-8.1] 9.3 [6.5-13.3] 15.8 [14.1-N/R] 1st line FFX Validation (165) 9.4 [7-13.3] 8.6 [5.6-11.4] 13.2 [9.9-39.6] 1st line GA Training (218) 5.6 [4.7-6.3] 6.9 [5.2-8.2] 11.2 [8.8-13.4] 1st line GA Validation (148) 5.6 [4.2-8.4] 7.5 [6.5-13.1] 8.5 [7.7-11.5]

Article Details

Volume / Issue Vol. 43, Issue 4_suppl
Published February 01, 2025
Pages 776-776
ISSN 0732-183X
Publisher Lippincott Williams & Wilkins

Journal Info

Journal of Clinical Oncology

Lippincott Williams & Wilkins

ISSN: 0732-183X Health Sciences

Authors (11)

M

Michael J. Pishvaian

E

Edik Matthew Blais

Perthera, South San Francisco, CA

L

Lynn M. Matrisian

Pancreatic Cancer Action Network, Manhattan Beach, CA

J

Jonathan R Brody

Oregon Health & Science University, Portland, OR

D

Dzung Thach

Perthera, Mclean, VA

P

Patricia Miren de Arbeloa

Perthera, Mclean, VA

F

Flavio G. Rocha

J

Jennifer W. Chuy

Montefiore Medical Center, Bronx, NY

R

Raymond Couric Wadlow

Inova Schar Cancer Institute, Fairfax, VA

A

Andrew Eugene Hendifar

Samuel Oschin Comprehensive Cancer Institute, Cedars-Sinai Medical Center, Los Angeles, CA

E

Emanuel Petricoin

George Mason University, Manassas, VA