Interobserver variability in the preoperative assessment of future liver remnant function using hepatobiliary scintigraphy

Background: Post-hepatectomy liver failure (PHLF) is a serious complication following hepatic resection and associated with a high mortality rate. The risk of PHLF can be estimated pre-resection by assessing the function of the future liver remnant (FLR) using hepatobiliary scintigraphy (HBS). Inaccurate estimation can have profound consequences, including an incorrect decision whether or not to proceed with hepatic resection. Therefore, it is essential to determine the reproducibility of preoperative assessment of the FLR function using HBS. The interobserver variability in the assessments of FLR function between two independent observers, blinded for each other’s results, was evaluated.

Results: In 24 out of 50 patients, the FLR function was predicted to be sufficient (> 2.69%/min/m²) to proceed with hepatic resection without preoperative FLR hypertrophy-inducing measures. In contrast, six patients were first subjected to portal vein embolization based on a predicted insufficient FLR function, which subsequently resulted in resection in four patients. Comparing the FLR function analyses of both observers, Bland-Altman plots demonstrated that most assessments lie within the 95% confidence interval and no pattern suggesting bias was observed. The interobserver level of agreement therefore appeared high for the FLR function (ICC = 0.996, Spearman’s ρ = 0.995 and Cohen’s κ = 0.948).

Conclusions: This study shows a high interobserver agreement and a negligible interobserver variability in the assessment of FLR function using HBS, regardless of the extent of observer experience. Therefore, the preoperative assessment of the FLR function is reproducible in the workup for patients planned to undergo (major) hepatic resections.