Routine intraoperative ICG perfusion assessment and anastomotic leak after esophagectomy: a before–after cohort study

Background: Anastomotic leak (AL) is a major complication after esophagectomy, and gastric conduit perfusion is considered a key determinant of anastomotic integrity. Indocyanine green (ICG) fluorescence angiography is increasingly used for intraoperative perfusion assessment, although its clinical effect remains uncertain. We evaluated whether routine implementation of ICG assessment was associated with a lower incidence of AL compared with a historical no-ICG cohort, and explored intraoperative factors potentially associated with AL. Methods: Single-center before–after cohort study of consecutive adults undergoing Ivor-Lewis esophagectomy for cancer between January 2023 and December 2025, using a prospectively maintained database. ICG entered routine practice in March 2025, defining a historical no-ICG cohort and an ICG cohort. ICG (2 mL of a 25 mg/10 mL solution) was administered at three intraoperative timepoints; time to fluorescence (TTF), arcade–conduit fluorescence pattern, arterial blood gas values, and hemodynamic parameters were recorded. AL was defined according to Esophagectomy Complications Consensus Group criteria. Results: Sixty patients were included (ICG, n = 17; no-ICG, n = 43). AL occurred in 4/17 (23.5%) versus 5/43 (11.6%) patients (odds ratio 2.34, 95% CI 0.54–10.05; p = 0.256). Overall postoperative morbidity, conduit necrosis, pulmonary and infectious complications, reintubation, and 90-day mortality did not differ; a single death occurred, in the no-ICG cohort. All TTF values were below 60 s (range 23–46 s) and did not differ between patients with and without AL, nor did the fluorescence pattern; ICG never modified the planned anastomotic site or prompted additional conduit resection. Exploratory within-ICG analyses showed lower thoracic-phase PaO2 and PaCO2 and higher pH in patients who developed AL. Conclusions: Routine qualitative ICG assessment of gastric conduit perfusion was feasible; however, this small before–after cohort was insufficient to determine its effect on AL. Larger studies using standardized quantitative fluorescence assessment are required.