Perioperative treatment is increasingly being explored in resectable pancreatic ductal adenocarcinoma (PDAC), where early systemic relapse, incomplete resections, and limited completion of postoperative chemotherapy remain major clinical challenges. The phase II nITRO trial previously showed that perioperative NALIRIFOX could be active and feasible in patients with resectable PDAC. This updated analysis provides longer follow-up and adds translational biomarker data, focusing on survival outcomes, circulating inflammatory markers, germline DNA damage repair and tumor suppressor gene variants, and pharmacogenomic predictors of toxicity.
The study, titled “Perioperative NALIRIFOX in patients with resectable pancreatic ductal adenocarcinoma: updated analysis of survival and circulating and genetic biomarkers from the phase II nITRO trial,” was published in ESMO Open, Volume 11, Issue 5, in May 2026.
Authors: E. Scarlato, V. Merz, S. Pietrobono, S. Casalino, A. Quinzii, C. Zecchetto, L. Mendo, S. Giacomazzi, A. Bonato, V. De Vita, G. Butturini, R. Salvia, and D. Melisi.
Why the study matters
Surgery followed by adjuvant chemotherapy remains a standard approach for patients with resectable PDAC, but outcomes remain limited. Many patients relapse early, some undergo non-curative resections, and a substantial proportion never receive postoperative chemotherapy. These limitations have supported the development of perioperative strategies, where systemic therapy is delivered before surgery to treat micrometastatic disease earlier and to better select patients for resection.
NALIRIFOX combines liposomal irinotecan, 5-fluorouracil, leucovorin, and oxaliplatin. In the nITRO trial, perioperative NALIRIFOX previously demonstrated activity in resectable PDAC, including an R0 resection rate of 65.3%. The present updated analysis aimed to confirm long-term outcomes and explore whether inflammatory, genomic, and pharmacogenomic biomarkers could help identify patients most likely to benefit from treatment or at higher risk of toxicity.
Read more about Remission Meaning in Pancreatic Cancer on OncoDaily.
Methods
The study enrolled 107 patients between May 2018 and May 2022, forming the intention-to-treat population. Clinical outcomes were updated after a median follow-up of 50.2 months. Tumor response was assessed before surgery using RECIST v.1, and treatment-related adverse events were graded according to NCI-CTCAE version 5.0.
The translational analyses included three main components. Plasma levels of 25 cytokines, chemokines, and growth factors were measured in 97 patients. Germline whole-exome sequencing was performed in 48 patients, while 10 additional patients underwent BRCA1/2 testing. Pharmacogenetic profiling of DPYD and UGT1A1 variants was assessed to explore associations with treatment-related toxicity.
Key Findings
After a median follow-up of 50.2 months, the median overall survival was 32 months in the intention-to-treat population. Among patients who underwent surgical resection, median disease-free survival was 19 months, and median overall survival reached 48 months. Resection margin status was strongly associated with outcome. Patients who achieved an R0 resection had longer median disease-free survival than those with R1 resection, 32 months versus 12 months. Median overall survival was also longer in the R0 group, 58 months versus 40 months.
The circulating biomarker analysis showed that baseline TNF-α was one of the most clinically relevant markers. Using a harmonized cut-off of 11.74 pg/mL, patients with low TNF-α had a higher overall response rate than those with high TNF-α, 29.6% versus 10.5%. They also had a higher resection rate, 83.6% versus 54.8%, and longer median overall survival, 44 months versus 22 months. High TNF-α was therefore associated with poorer response, lower likelihood of resection, and shorter survival.
Germline testing identified pathogenic or predicted damaging variants in DNA damage repair or tumor suppressor genes in 21 of 58 patients, classified as g.Deficient. These patients had numerically higher response and resection rates than g.Proficient patients, although these differences were not statistically significant. A trend toward longer median overall survival was also observed in g.Deficient patients compared with g.Proficient patients, 55 months versus 32 months, though this difference did not reach statistical significance (P = 0.13).
The most favorable outcomes were observed in patients with combined g.Deficient/TNF-α–low status. This subgroup achieved a 100% resection rate and a median overall survival of 55 months. In comparison, patients with g.Deficient/TNF-α–high status had a median overall survival of 23 months.
The pharmacogenetic analysis showed that 40% of evaluable patients developed severe adverse events during the first three preoperative cycles. The most common grade ≥3 chemotherapy-related adverse events were neutropenia, hypokalemia, diarrhea, and elevated liver enzymes. The DPYD c.496A>G variant was independently associated with early grade ≥3 gastrointestinal toxicity and with the need for significant chemotherapy dose reduction. Female sex was also independently associated with grade ≥3 gastrointestinal toxicity.
Read more about Pancreatic Cancer Treatment Choices on OncoDaily.
Expert Highlights
To provide clinical context for the translational findings from the updated nITRO analysis, Davide Melisi, Associate Professor of Medical Oncology at the University of Verona, Italy, shared his perspective with OncoDaily GI on the potential role of TNF-α and the combined g.Deficient/TNF-α–low biomarker profile in resectable PDAC.
The updated analysis showed that high baseline TNF-α levels were inversely associated with radiological response, resection probability, 6-month DFS, and overall survival. Asked how TNF-α should be interpreted as a biomarker of treatment sensitivity or resistance, Dr. Melisi explained:
“I think TNF-α captures a more inflammatory, treatment-resistant PDAC biology rather than simply acting as a prognostic marker. Mechanistically, it likely reflects activation of TAK1/NF-κB survival pathways associated with chemoresistance, so high levels may indicate primary resistance, while low levels could help identify patients more likely to benefit from perioperative NALIRIFOX. We and others have also recently shown that tumor-associated macrophages are the possible key source of TNF-α, supporting an inflamed yet immunosuppressive microenvironment. That said, these findings remain hypothesis-generating and require prospective validation.”
Davide Melisi, University of Verona, Italy
The study also reported that patients with combined g.Deficient/TNF-α–low status appeared to derive the most favorable outcomes, including a 100% resection rate and median overall survival of 55 months. Asked what validation would be needed before this biomarker profile could guide perioperative treatment selection, Dr. Melisi stated:
“Before using this profile to guide treatment selection, it would need prospective validation in larger, ideally randomized cohorts of resectable PDAC patients treated with perioperative NALIRIFOX compared with upfront surgery.
The key steps would be to confirm that the combined g.Deficient/TNF-α–low profile is not only prognostic, but predictive of benefit from perioperative therapy compared with upfront surgery. Only then could it be used confidently to personalize treatment selection.”
Conclusion
The updated nITRO trial analysis confirms the long-term activity of perioperative NALIRIFOX in patients with resectable PDAC. With a median follow-up exceeding four years, the regimen was associated with a median overall survival of 32 months in the full intention-to-treat population and 48 months among resected patients.
Beyond survival outcomes, the study highlights the potential value of biomarker-guided perioperative treatment. Low baseline TNF-α levels were associated with better response, higher resection rates, and longer survival. Germline DNA damage repair or tumor suppressor gene alterations appeared to identify patients with favorable outcomes, particularly when combined with low TNF-α levels. DPYD c.496A>G may also help identify patients at increased risk of early severe gastrointestinal toxicity.
These findings support further investigation of inflammatory, genomic, and pharmacogenomic biomarkers in resectable PDAC. However, the biomarker results remain exploratory and should be prospectively validated in larger randomized studies before being applied in routine clinical decision-making.
The full article is published in ESMO Open.