GI Oncology Daily Digest

This systematic review pooled 13 comparative studies of postoperative immunotherapy versus observation or placebo after neoadjuvant therapy and esophagectomy, in populations made up mainly of patients with residual pathologic disease: two randomised trials and 11 retrospective cohorts. Hazard ratios were reconstructed from published Kaplan-Meier curves where they were not reported. Adjuvant immunotherapy was associated with longer OS (12 studies, 4,225 patients, HR 0.64, 95% CI 0.53–0.77) and longer DFS/RFS (11 studies, 2,363 patients, HR 0.74, 95% CI 0.66–0.84), with favourable estimates after neoadjuvant chemoradiotherapy, chemoimmunotherapy and chemotherapy; the interaction by neoadjuvant strategy was significant for OS but not DFS/RFS. The Bayesian predictive intervals crossed 1 for both outcomes and the authors rate certainty as low. Their own reading is the right one: randomised evidence supports a DFS benefit after neoadjuvant chemoradiotherapy, a randomised OS benefit has not been shown, the signal after neoadjuvant chemoimmunotherapy is entirely observational, and the chemotherapy-only estimate rests on two small retrospective studies.

This systematic review and meta-analysis searched PubMed, Embase, the Cochrane Library and trial registries to November 2025 for studies comparing indocyanine green (ICG)-guided with conventional lymphadenectomy in gastric cancer, and pooled seven studies with 7,518 patients. ICG guidance was associated with better 3-year OS (HR 0.635, 95% CI 0.511–0.790), 3-year DFS (HR 0.682, 95% CI 0.530–0.877) and 5-year OS (HR 0.700, 95% CI 0.507–0.966, P=0.030), while 2-year OS and DFS did not differ. Lymph node yield rose by a weighted mean of 6.6 nodes (95% CI 3.8–9.3), and postoperative complications did not increase (RR 0.899, 95% CI 0.647–1.250). The abstract does not report how many of the seven studies were randomised or give a randomised-only estimate, and the authors conclude that further high-quality randomised trials are needed.

This systematic review pooled 76 observational reports covering 10,243 patients with unresectable HCC treated with atezolizumab-bevacizumab (60 reports, 8,928 patients) or durvalumab-tremelimumab (16 reports, 1,315 patients). Sample-size-weighted median OS was 17.7 and 16.5 months, summarised descriptively and not formally compared. Pooled ORR was 26.6% vs 20.8% and DCR 71.4% vs 55.4%. In covariate-adjusted meta-regression, atezo-bev cohorts had longer median PFS (+1.96 months, 95% CI +0.91 to +3.02, P<0.001) and higher DCR (+16.15 points, P<0.001), while the OS difference (+3.57 months, 95% CI −0.38 to +7.52, P=0.076) and ORR difference (P=0.302) were not significant. The PFS association was not reproduced in first-line-predominant cohorts (+1.35 months, P=0.116) and attenuated after excluding four reports with possible patient overlap. The authors conclude that differences in treatment line, age, calendar period and the small durvalumab-tremelimumab evidence base preclude reliable comparative-effectiveness inference.

CRP below 100 mg/L on postoperative day 3 is widely used to exclude anastomotic leak, but enhanced-recovery pathways increasingly discharge patients before day 3. This two-cohort multicentre diagnostic-accuracy study tested a day-2 threshold in patients undergoing elective colon cancer resection with primary anastomosis, defining leak as Clavien-Dindo grade ≥3. Leak occurred in 4.0% of the test cohort (n=401) and 4.7% of the validation cohort (n=213). A day-2 CRP below 100 mg/L was present in 60% and 46% of patients, whose leak rates were 1.3% and 1.0%, for negative predictive values of 98.8% and 99.0%; with day-2 CRP of 100 or above, leak rates were 8.1% and 7.8%. In that higher-risk group, a day-3/day-2 CRP ratio below 1 gave NPVs of 98.9% after minimally invasive and 97.9% after open surgery, and with this two-step strategy nearly 90% of patients could be discharged by day 2 or 3 with a leak risk near 1%.