Background: Indocyanine green is a useful tool in colorectal surgery. Quantitative values may enhance and standardize its application.
Objective: To determine if quantitative indocyanine green metrics correlate with standard subjective indocyanine green perfusion assessment in acceptance or rejection of anastomotic margins.
Design: Prospective single arm, single institution cohort study. Surgeons viewed subjective indocyanine green images but were blinded to quantitative indocyanine green metrics.
Main outcome measures: Accepted perfusion and rejected perfusion of the intestinal margin were defined by absence or presence of ischemia by subjective indocyanine green and gross inspection. The primary outcomes were quantitative indocyanine green values, maximum fluorescence and time to maximum fluorescence in accepted compared to rejected perfusion. Secondary outcomes included maximum fluorescence and time to maximum fluorescence values in anastomotic leak.
Results: There were 89 perfusion assessments, comprising 156 intestinal segments. Nine segments were subjectively assessed to have poor perfusion by visual inspection and subjective indocyanine green. Maximum fluorescence (% intensity) exhibited higher intensity in accepted perfusion [accepted perfusion 161% (82-351) vs. rejected perfusion 63% (10 – 76); p = 0.03]. Similarly, time to maximum fluorescence (seconds) was earlier in accepted compared to rejected perfusion [10s (1 – 40) vs. 120s (90 – 120); p < 0.01]. Increased body mass index associated with higher maximum fluorescence. Anastomotic leak did not correlate with maximum fluorescence or time to maximum fluorescence.
Conclusion: We demonstrate blinded quantitative values reliably correlate with subjective indocyanine green perfusion assessment. Time to maximum intensity is an important metric in perfusion evaluation. Quantitative indocyanine green metrics may enhance intraoperative intestinal perfusion assessment. Future studies may attempt to correlate quantitative indocyanine green values with anastomotic leak.
