Multicentre phase II trial of near-infrared imaging in elective colorectal surgery

Abstract

Background: Decreasing anastomotic leak rates remain a major goal in colorectal surgery. Assessing intraoperative perfusion by indocyanine green (ICG) with near-infrared (NIR) visualization may assist in selection of intestinal transection level and subsequent anastomotic vascular sufficiency. This study examined the use of NIR-ICG imaging in colorectal surgery.

Methods: This was a prospective phase II study (NCT02459405) of non-selected patients undergoing any elective colorectal operation with anastomosis over a 3-year interval in three tertiary hospitals. A standard protocol was followed to assess NIR-ICG perfusion before and after anastomosis construction in comparison with standard operator visual assessment alone.

Results: Five hundred and four patients (median age 64 years, 279 men) having surgery for neoplastic (330) and benign (174) pathology were studied. Some 425 operations (85·3 per cent) were started laparoscopically, with a conversion rate of 5·9 per cent. In all, 220 patients (43·7 per cent) underwent high anterior resection or reversal of Hartmann's operation, and 90 (17·9 per cent) low anterior resection. ICG angiography was achieved in every patient, with a median interval of 29 s to visualization of the signal after injection. NIR-ICG assessment resulted in a change in the site of bowel division in 29 patients (5·8 per cent) with no subsequent leaks in these patients. Leak rates were 2·4 per cent overall (12 of 504), 2·6 per cent for colorectal anastomoses and 3 per cent for low anterior resection. When NIR-ICG imaging was used, the anastomotic leak rates were lower than those in the participating centres from over 1000 similar operations performed with identical technique but without NIR-ICG technology.

Conclusion: Routine NIR-ICG assessment in patients undergoing elective colorectal surgery is feasible. NIR-ICG use may change intraoperative decisions, which may lead to a reduction in anastomotic leak rates.

Figure 1

Near‐infrared (NIR) assessment of level of transection. Images of a planned transection before indocyanine green (ICG) injection (arrow) and b visible transection area after ICG injection (arrow) are shown in normal view, NIR view and enhanced reality view. There is no change in transection area if the perfusion signal reaches the planned area for transection

Figure 2

Near‐infrared (NIR) perfusion assessment after a side‐to‐end colorectal anastomosis had been constructed. Intraoperative images of the anastomosis a before and b after indocyanine green (ICG) injection are shown in normal view, NIR view and enhanced reality view. After ICG injection, there was a good signal on the rectal stump and colon

Figure 3

Near‐infrared (NIR) perfusion assessment with change of plan owing to a lack of perfusion at the level of the section originally planned. Intraoperative images are shown in normal view, NIR view and enhanced reality view. a Image before indocyanine green (ICG) injection showing the planned area for proximal transection (yellow arrow) in a segment of descending colon after its mobilization (including high vascular ligation) and mesocolic preparation. b After ICG injection, a clear demarcation line appeared (white arrow) that was 4 cm more proximal (vertical yellow arrow on the initial transection area) and led to more proximal transection (horizontal arrow shows distance that has been assessed as well perfused) being undertaken. c A second injection of ICG in the same patient showed satisfactory perfusion of the constructed colorectal anastomosis in situ