Combining indocyanine green fluorescence imaging and ultrasound intraoperatively to visualize primary and secondary cancerous lesions in the liver: Systematic review and meta-analysis.

Abstract

Increasingly, published evidence is documenting survival advantages of resecting cancerous lesions in the liver, whether primary or metastatic. However, an average of 15% of such lesions are missed, especially when multiple lesions exist, since intraoperative ultrasound, the gold-standard intraoperative imaging technique, often misses lesions near the hepatic surface. In this meta-analysis, we examined whether adding fluorescence imaging with indocyanine green to intraoperative ultrasound increases the sensitivity of liver cancer detection. A thorough review of PubMed, EMBASE, and Scopus was conducted to identify all studies in which the sensitivity of adding indocyanine green to intraoperative ultrasound versus intraoperative ultrasound used alone was evaluated in patients with primary or secondary liver cancer. Strict study exclusion criteria were used, including a Modified Downs and Black Quality Assessment score. Forest and funnel plots were generated, and both odds ratios and effect sizes were calculated, with 95% confidence intervals, comparing ultrasound + indocyanine green against intraoperative ultrasound and indocyanine green-fluorescence imaging alone, and comparing indocyanine green-fluorescence imaging against intraoperative ultrasound. For this last comparison, a noninferiority analysis was also performed. In total, 2,276 abstracts were reviewed, revealing 11 studies, encompassing 558 patients and 894 cancerous liver lesions, eligible for analysis. Ultrasound + indocyanine green was statistically superior to both intraoperative ultrasound and indocyanine green-fluorescence imaging used alone (odds ratio, 15.07; 95% confidence interval, 6.65-34.15; and odds ratio, 5.34; 95% confidence interval, 1.83-15.56, respectively), achieving overall sensitivity of 99.78% vs 91.95% and 86.80% for indocyanine green fluorescence imaging and intraoperative ultrasound used alone, respectively. Indocyanine green-fluorescence imaging was statistically noninferior to intraoperative ultrasound. Adding indocyanine green-fluorescence imaging to intraoperative ultrasound markedly increases the sensitivity of detecting cancerous liver lesions from 86.8% to more than 99.5%.