5 Key GI Cancer Updates From ESMO 2025
Key Points
-
The DYNAMIC-III and PEGASUS trials did not fully establish the feasibility of guiding treatment escalation or de-escalation in colon cancer based on circulating tumor DNA (ctDNA) status.
-
Data from the STELLAR-303 trial provided proof of concept for immunotherapy in patients with non-microsatellite instability-high metastatic colorectal cancer (mCRC).
-
Durvalumab significantly improved overall survival (OS) versus placebo for gastric and gastroesophageal junction cancers in the MATTERHORN trial.
-
The FORTITUDE-101 trial showed that anti-FGFR2b agents may improve survival for advanced gastric and gastroesophageal junction cancers with overexpression of FGFR2b.
ESMO 2025 GI Cancer Presentations
Cohosts of the Oncology Brothers podcast, Rahul Gosain, MD, MBA, of Wilmot Cancer Institute, and Rohit Gosain, MD, of Roswell Park Comprehensive Cancer Center, discussed five major presentations from colon and upper gastrointestinal (GI) cancer trials presented at the European Society of Medical Oncology (ESMO) Congress 2025 with guest, Rachna Shroff, MD, MS, FASCO, of University of Arizona Cancer Center. The doctors reviewed findings from the DYNAMIC-III, PEGASUS, STELLAR-303, MATTERHORN, and FORTITUDE-101 trials.
ctDNA-Guided Treatment Adjustment in DYNAMIC-III
The DYNAMIC-III trial explored using post-surgery ctDNA status to guide escalation or de-escalation of adjuvant chemotherapy in patients with stage III colon cancer. Participants had tumor-informed ctDNA testing 5 to 6 weeks after surgery and were randomized to either ctDNA-guided (n=353) or standard management (n=349) for adjuvant chemotherapy.
Patients with ctDNA negativity were de-escalated from 6 months of fluoropyrimidine (FP) to 3 months of FP or observation, from 3 months of FP plus oxaliplatin doublet therapy to single-agent FP, or from 6 months of doublet therapy to 3 months of doublet or single-agent FP.
After a median follow-up of 45 months, 319 (90.4%) ctDNA-negative patients had de-escalated treatment. Overall, 34.8% of patients in the ctDNA-guided arm received oxaliplatin-based chemotherapy versus 88.6% of those in the standard management arm (P < .001). De-escalation also reduced rates of grade 3 or higher adverse events of special interest (6.2% vs 10.6%; P = .037) and treatment-related hospitalization (8.5% vs 13.2%; P = .048).
The study could not confirm non-inferiority of ctDNA-guided de-escalation for 3-year recurrence-free survival (RFS), which was 85.3% with ctDNA guidance versus 88.1% with standard management (97.5% lower CI, –8.0%).
Conversely, in 266 patients with ctDNA-positivity, escalating treatment did not significantly improve the 2-year RFS rate compared with standard management. “DYNAMIC-III adds to our understanding of ctDNA, but I don’t necessarily think it has truly been proof of principle that we can use ctDNA to guide adjuvant therapy,” Dr. Shroff said.
PEGASUS: ctDNA Status to Guide Escalation or De-Escalation
In the PEGASUS trial, physicians evaluated the potential of ctDNA status on liquid biopsy to guide post-surgery and post-adjuvant management of patients with stage III or T4N0 stage II colon cancer. In the post-surgery setting, patients who tested positive for ctDNA received 3 months of CAPOX (oxaliplatin plus capecitabine) while those who tested negative received 6 months of capecitabine.
After retesting for ctDNA-status in the postadjuvant setting, patients treated with CAPOX were either escalated to FOLFIRI (leucovorin, fluorouracil [5-FU], and irinotecan) or de-escalated to capecitabine if they were ctDNA-positive or ctDNA-negative, respectively. Likewise, patients treated with capecitabine were escalated to CAPOX if ctDNA-positive or de-escalated to standard follow-up if ctDNA-negative.
A total of 135 patients were enrolled, of which 100 were ctDNA-negative and 35 were ctDNA-positive. At a median follow-up of 33.5 months, 12 of the 100 ctDNA-negative patients had relapsed, representing a 2-year disease-free survival rate of 88% (90% CI, 81-93). Comparatively, 13 of the 35 ctDNA-positive patients had relapsed. The 2-year time-to-relapse rate was 87.6% for ctDNA-negative and 61.5% for ctDNA-positive patients (hazard ratio [HR], 3.22; 95% CI, 1.20-7.96; P = .0013).
“When looking at the feasibility of the study’s protocol… it doesn’t really change our practice,” said Dr. Shroff. However, it remains important to understand how ctDNA can play a role in clinical trials and how it can guide risk stratification, she added.
Zanzalintinib Outcomes for mCRC in STELLAR-303
The phase 3 STELLAR-303 trial compared zanzalintinib plus atezolizumab against regorafenib in patients with previously treated mCRC. The trial randomized 901 patients with relapsed or refractory mCRC without microsatellite instability-high or mismatch repair-deficient status. Anti-VEGF antibody therapy had been given in 82% of participants.
After a median follow-up of 18 months, zanzalintinib plus atezolizumab significantly improved median OS versus regorafenib at 10.9 months versus 9.4 months (HR, 0.80; 95% CI, 0.69-0.93; P = .0045), respectively. Zanzalitinib plus atezolizumab also improved median progression-free survival (PFS) at 3.7 months versus regorafenib at 2.0 months (HR, 0.68; 95% CI, 0.59-0.79). OS and PFS benefits were also observed across subgroups including liver metastases and prior anti-VEGF therapy.
“It’s important to recognize that this was a positive study, this is proof of principle that immunotherapy plus a multitargeting tyrosine kinase inhibitor may have a role in the treatment of refractory [microsatellite stable] colorectal cancer,” said Dr. Shroff. She noted that the rate of grade 3 treatment-related adverse events was about 60% in the zanzalintinib arm, so more work may be needed to understand the risk versus benefit balance for this treatment approach.
MATTERHORN Data Update With Final OS Results
MATTERHORN enrolled patients with resectable gastric or gastroesophageal junction adenocarcinoma to receive either durvalumab (n=385) or placebo (n=372) plus FLOT (5-FU, leucovorin, oxaliplatin, and docetaxel) for 4 cycles followed by durvalumab or placebo for 10 additional cycles. The ESMO Congress 2025 presentation for the MATTERHORN trial reported final findings for OS, OS per PD-L1 status, and associations between event-free survival (EFS) and pathological complete response, major pathological response, and nodal staging.
Overall, durvalumab significantly improved OS versus placebo (HR, 0.78; 95% CI, 0.63-0.96; P = .021). The OS benefit with durvalumab was present for both patients with PD-L1 tumor area positivity (TAP) less than 1% (HR, 0.79; 95% CI, 0.41-1.50) and with PD-L1 TAP of 1% or greater (HR, 0.79; 95% CI, 0.63-0.99). In evaluable patients, durvalumab showed higher rates of negative nodal involvement compared with placebo (58.2% vs 44.8%; odds ratio, 1.72; 95% CI, 1.30-2.27). Additionally, durvalumab improved EFS compared to placebo in patients with any pathologic response, regardless of nodal status.
“The fact that this was a positive study in terms of final OS is a really important milestone, and reiterates the fact that [durvalumab plus FLOT], should really be our new approach to perioperative management of upper GI malignancies,” Dr. Shroff said.
Anti-FGFR2b Treatment Improves Survival in FORTITUDE-101
The FORTITUDE-101 trial investigated the first-in-class anti-FGFR2b antibody, bemarituzumab, plus chemotherapy for patients with HER2-negative, unresectable, locally advanced or metastatic, FGFR2b-overexpressing gastric or gastroesophageal junction cancer. In the trial, 159 of 274 patients in the bemarituzumab arm and 165 of 273 in the placebo arm had FGFR2b expression of 10% or greater.
In the primary analysis, after a median follow-up of 11.8 months, patients with FGFR2b of 10% or greater had a median OS of 17.9 months with bemarituzumab versus 12.5 months with placebo (HR, 0.61; 95% CI, 0.43-0.86; P = .005). In the descriptive follow-up analysis with a median follow-up of 19.4 months, the median OS for this subset was 14.5 months with bemarituzumab versus 13.2 months with placebo (HR, 0.82; 95% CI, 0.62-1.08).
Though the follow-up descriptive analysis suggested the benefit of bemarituzumab may diminish over time, the post-frontline setting raises the question of what previous treatments were used in the population and whether that may attenuate the OS impact over time, Dr. Shroff said.