Modified FOLFIRINOX is the standard adjuvant chemotherapy for eligible patients with resected pancreatic ductal adenocarcinoma. However, whether genomic alterations or molecular subtypes can identify patients who benefit differently from modified FOLFIRINOX or gemcitabine remains uncertain.
On August 3, 2026, the Journal of Clinical Oncology published the study titled “Impact of Tumor Genomic Profile on Adjuvant Chemotherapy Efficacy in Resected Pancreatic Adenocarcinoma: Results From the PRODIGE-24/CCTG PA6 Study.”
Authors: Andréa Witz, Thierry Conroy, Aurélien Lambert, Julia Salleron, Aboubacar Diallo, Marie Husson, Rémy Nicolle, Pascal Hammel, James Biagi, Daniel J. Renouf, Anthony Turpin, Corentin Richard, Jean-Baptiste Bachet, Juan Iovanna, Nelson Dusetti, Laure Monard, Marjorie Mauduit, Jérôme Cros, and Alexandre Harlé.
A Molecular Analysis of PRODIGE-24
This retrospective molecular analysis included tumor samples from patients enrolled in the multicenter, open-label, phase 3 PRODIGE-24/CCTG PA6 trial, registered as NCT01526135.
Tumor DNA sequencing was successfully performed in 317 of 350 tumors. Among the analyzed patients, 168 received modified FOLFIRINOX and 149 received gemcitabine. The investigators evaluated four major pancreatic cancer driver genes—KRAS, TP53, SMAD4, and CDKN2A—along with 24 homologous recombination repair-associated genes and single-base substitution mutational signatures.
Transcriptomic subtyping was performed using the Purity Independent Subtyping of Tumors, or PurIST, classifier. The primary endpoint was disease-free survival, while cancer-specific survival was the secondary endpoint. In the molecular cohort, median disease-free survival was 20.5 months with modified FOLFIRINOX and 12.1 months with gemcitabine, with a stratified hazard ratio of 0.66 (95% CI, 0.51–0.86; P = .002).
Median cancer-specific survival was 54.4 months and 35.5 months, respectively, with a stratified hazard ratio of 0.71 (95% CI, 0.53–0.96; P = .024).
PurIST Subtype
Of the 317 tumors, 253, or 79.8%, were classified as classical, while 64 were classified as basal-like. Among patients with classical tumors, modified FOLFIRINOX improved disease-free survival compared with gemcitabine, with a stratified hazard ratio of 0.62 (95% CI, 0.46–0.84). A similar benefit was observed for cancer-specific survival, with a stratified hazard ratio of 0.65 (95% CI, 0.46–0.91).
Within the modified FOLFIRINOX group, patients with classical tumors had longer disease-free survival than those with basal-like tumors, with a stratified hazard ratio of 0.48 (95% CI, 0.31–0.77).
However, the interaction between treatment and PurIST subtype was not statistically significant for disease-free survival (P for interaction = .298) or cancer-specific survival (P for interaction = .262).
These findings indicate that PurIST subtype was prognostic for disease-free survival among patients treated with modified FOLFIRINOX but did not predict a differential benefit from modified FOLFIRINOX over gemcitabine. The results also did not support greater gemcitabine sensitivity in basal-like tumors.
KRAS Status Showed a Potential Treatment Interaction
Oncogenic KRAS mutations were identified in 277 of 317 tumors, representing 87.4% of the cohort. KRAS G12D was the most common variant, accounting for 43.7% of KRAS-mutated tumors. A significant interaction was observed between KRAS status and adjuvant treatment for disease-free survival (P for interaction = .010).
Among patients with KRAS-mutated tumors, modified FOLFIRINOX improved disease-free survival compared with gemcitabine, with a stratified hazard ratio of 0.60 (95% CI, 0.45–0.79; P < .001). The benefit was particularly observed in tumors harboring KRAS G12D and G12V mutations. No benefit from modified FOLFIRINOX over gemcitabine was observed in patients with KRAS wild-type tumors or KRAS variants other than G12D or G12V.
Within the modified FOLFIRINOX group, patients with KRAS-mutated tumors had longer disease-free survival than those with KRAS wild-type tumors, with a stratified hazard ratio of 0.44 (95% CI, 0.25–0.78; P = .005). No significant difference according to KRAS status was observed in the gemcitabine group, with a stratified hazard ratio of 1.21 (95% CI, 0.71–2.08; P = .486).
The KRAS wild-type subgroup included only 40 patients. The authors also noted the biological and technical heterogeneity of KRAS wild-type tumors and stated that false-negative KRAS results related to analytical limitations could not be excluded.
The apparent absence of benefit from modified FOLFIRINOX in this subgroup should therefore be considered hypothesis-generating. No prognostic or predictive value was identified for TP53, SMAD4, or CDKN2A mutations.
BRCA and Other HRR Alterations
Alterations in homologous recombination repair-associated genes were detected in 94 of 317 patients, representing 29.7% of the cohort. Pathogenic or likely pathogenic BRCA mutations were identified in 37 patients, or 11.6%. Six were estimated to be germline and 31 were considered somatic based on the study’s classification method. BRCA1 mutations were found in 13 patients, BRCA2 mutations in 24, and PALB2 mutations in eight.
Neither homologous recombination repair-associated gene status nor BRCA status predicted a differential benefit from modified FOLFIRINOX compared with gemcitabine. The interaction P values were .568 and .785, respectively.
Other homologous recombination repair-associated genes were also not predictive of treatment benefit. The authors emphasized that alterations in homologous recombination repair-associated genes should not be considered equivalent to homologous recombination deficiency.
The sequencing approach did not assess the genomic scarring patterns required to formally define homologous recombination deficiency. Matched tumor-normal sequencing was also unavailable, meaning that germline and somatic variants could not be definitively distinguished. Allelic status could not be assessed.
Mutational Signatures
Single-base substitution mutational signatures were evaluated in 315 tumor samples. SBS1 was the most prevalent signature, detected in 67.6% of samples. No significant interaction was observed between mutational signature status and treatment for disease-free survival or cancer-specific survival. No individual signature was prognostic within either treatment group.
The benefit of modified FOLFIRINOX remained consistent across single-base substitution-positive and single-base substitution-negative subgroups.
SBS3 was not significantly associated with BRCA mutational status. The authors noted that the targeted 524-gene sequencing panel may have limited the reliable detection of mutational signatures, particularly SBS3.
Limitations
The study was retrospective and included patients according to the availability of tumor samples, which may have partially affected the balance created by randomization in the original trial. Several molecular subgroups were small, particularly the KRAS wild-type and estimated germline BRCA-mutated groups, limiting the statistical power of the analyses.
Whole-exome and whole-genome sequencing were not performed, and copy-number alterations were not assessed. These limitations may have affected the detection and interpretation of some genomic alterations and mutational signatures. However, adjusted multivariable and competing-risk sensitivity analyses were consistent with the primary findings.
Takeaway
This molecular analysis of PRODIGE-24/CCTG PA6 does not support changing current adjuvant treatment strategies for resected pancreatic ductal adenocarcinoma.
Modified FOLFIRINOX remains the standard adjuvant regimen for eligible patients, regardless of PurIST subtype, BRCA status, other homologous recombination repair-associated alterations, or single-base substitution mutational signatures.
KRAS status showed a potential predictive signal, as modified FOLFIRINOX improved disease-free survival in KRAS-mutated tumors but not in the small KRAS wild-type subgroup. However, the lack of benefit observed in KRAS wild-type tumors remains hypothesis-generating and requires further validation before it can guide treatment selection.
The full article is available in the Journal of Clinical Oncology.