Background: Tumour-targeted fluorescence imaging has the potential to advance current practice of oncological surgery by selectively highlighting malignant tissue during surgery. Carcinoembryonic antigen (CEA) is overexpressed in 90% of colorectal cancers and is a promising target for colorectal cancer imaging. We aimed to assess the tolerability of SGM-101, a fluorescent anti-CEA monoclonal antibody, and to investigate the feasibility to detect colorectal cancer with intraoperative fluorescence imaging.
Methods: We did an open-label, pilot study in two medical centres in the Netherlands. In the dose-escalation cohort, we included patients (aged ≥18 years) with primary colorectal cancer with increased serum CEA concentrations (upper limit of normal of ≥3 ng/mL) since diagnosis, who were scheduled for open or laparoscopic tumour resection. In the expansion cohort, we included patients (aged ≥18 years) with recurrent or peritoneal metastases of colorectal cancer, with increasing serum concentrations of CEA since diagnosis, who were scheduled for open surgical resection. We did not mask patients, investigators, or anyone from the health-care team. We assigned patients using a 3 + 3 dose design to 5 mg, 7·5 mg, or 10 mg of SGM-101 in the dose-escalation cohort. In the expansion cohort, patients received a dose that was considered optimal at that moment of the study but not higher than the dose used in the dose-escalation cohort. SGM-101 was administered intravenously for 30 min to patients 2 or 4 days before surgery. Intraoperative imaging was done to identify near-infrared fluorescent lesions, which were resected and assessed for fluorescence. The primary outcome was tolerability and safety of SGM-101, assessed before administration and continued up to 12 h after dosing, on the day of surgery, the first postoperative day, and follow-up visits at the day of discharge and the first outpatient clinic visit. Secondary outcomes were effectiveness of SGM-101 for detection of colorectal cancer, assessed by tumour-to-background ratios (TBR); concordance between fluorescent signal and tumour status of resected tissue; and diagnostic accuracy in both cohorts. This trial is registered with the Nederlands Trial Register, number NTR5673, and ClinicalTrials.gov, number NCT02973672.
Findings: Between January, 2016, and February, 2017, 26 patients (nine in the dose-escalation cohort and 17 in the expansion cohort) were included in this study. SGM-101 did not cause any treatment-related adverse events, although three possibly related mild adverse events were reported in three (33%) of nine patients in the dose-escalation cohort and five were reported in three (18%) of 17 patients in the expansion cohort. Five moderate adverse events were reported in three (18%) patients in the expansion cohort, but they were deemed unrelated to SGM-101. No changes in vital signs, electrocardiogram, or laboratory results were found after administration of the maximum dose of 10 mg of SGM-101 in both cohorts. A dose of 10 mg, administered 4 days before surgery, showed the highest TBR (mean TBR 6·10 [SD 0·42] in the dose-escalation cohort). In the expansion cohort, 19 (43%) of 43 lesions were detected using fluorescence imaging and were not clinically suspected before fluorescent detection, which changed the treatment strategy in six (35%) of 17 patients. Sensitivity was 98%, specificity was 62%, and accuracy of fluorescence intensity was 84% in the expansion cohort.
Interpretation: This study presents the first clinical use of CEA-targeted detection of colorectal cancer and shows that SGM-101 is safe and can influence clinical decision making during the surgical procedure for patients with colorectal cancer.
Copyright © 2018 Elsevier Ltd. All rights reserved.