Optimization of ALPPS stage II timing with the APRI/ALBI score-an international, multicenter cohort study

Background: Primarily unresectable liver tumors may be approached by the Associating Liver Partition and Portal vein Ligation for Staged Hepatectomy (ALPPS) procedure. Post-hepatectomy liver failure (PHLF) poses the most significant risk factor for poor outcomes. The AST-to-platelets ratio index (APRI)/albuminto-bilirubin index (ALBI) score has been proposed as an easy and routinely available score to monitor liver function. Here, we explored the predictive capability of the APRI/ALBI score to determine PHLF and perioperative morbidity to help determine the optimal timing of the 2nd stage of ALPPS. Methods: Based on the international multicenter ALPPS registry, patients from 2012 to 2020 with an available APRI/ALBI score were included. Postoperative outcomes (clinically relevant PHLF B + C, 90-day mortality, and severe morbidity (>= Clavien-Dindo 3b) after ALPPS stage II were assessed. The APRI/ALBI score was monitored perioperatively, and the predictive value was evaluated using logistic regression and receiver operating characteristics. Performance of APRI/ALBI score was compared to the ALPPS futility risk score in this cohort study. Results: Overall, 464 patients from 16 participating centers were included. Clinically relevant PHLF (B + C) was observed in 7.5% of patients, of which 63% ultimately died. After stage I, the APRI/ALBI score gradually recovered. The pre-stage II APRI/ALBI score significantly predicted clinically relevant PHLF [area under the curve (AUC) =0.78; P<0.001], 90-day mortality (AUC =0.67; P=0.002), and severe morbidity (AUC =0.65; P<0.001). Three clinically relevant APRI/ALBI score risk groups were defined: clinically relevant PHLF occurred in 3.1% in the low-, 8.7% in the intermediate-, and 28.0% in the high-risk groups. 90-day mortality was 6.8% in the low-, 15.9% in the intermediate-, and 19.4% in the high-risk groups. Integrated assessment of the established futility risk score in combination with the APRI/ALBI score documented further increased predictive potential for clinically relevant PHLF (AUC 0.81; P<0.001). Conclusions: The APRI/ALBI score allows for simple and dynamic liver function recovery monitoring after the first ALPPS stage. Inadequate recovery of the APRI/ALBI score until ALPPS stage II was associated with PHLF B + C, 90-day mortality, and severe morbidity. With the proposed risk model, optimized timing of the second stage of ALPPS may further increase the safety of this procedure.