Background
Management of advanced hepatocellular carcinoma (HCC) has been transformed by immune checkpoint inhibitors (ICIs), which are now first-line systemic therapies. However, while many HCC patients suffer from frailty and malnutrition as a result of underlying liver disease and/or their cancer burden in and of itself, major clinical trials included only highly optimized patients and excluded those with Eastern Cooperative Oncology Group (ECOG) scores over 1. We thus investigated how functional and nutritional statuses impact clinical outcomes of ICI-treated HCC patients in real-world practice.

Methods
We conducted a retrospective cohort study of 247 ICI-treated HCC patients. Functional status was assessed using ECOG scores recorded prior to treatment initiation. Pre-treatment nutritional status was quantified using the Prognostic Nutritional Index (PNI), a marker derived from albumin and lymphocyte count that has been validated in both liver disease and various cancers. Cox regression was employed to assess the impact of ECOG and PNI on overall survival after adjusting for potential confounders.

Results
In our cohort, 33 patients (13%) had an ECOG of 2 or greater, and the median PNI was 38 (IQR 32-42). The median follow-up time was 253 days, and 177 (72%) patients died during the study period. Patients with higher ECOG scores had significantly shorter survival times (median 445 vs 177 vs 69 days for ECOG 0 vs 1 vs 2+, log-rank p<0.001; Fig 1A), as did those with lower PNI scores (median 114 vs 389 vs 646 days for PNI tertile 1 vs 2 vs 3, log-rank p<0.001; Fig 1B). Age, MELD, and Barcelona Clinic Liver Cancer (BCLC) stage were all separately associated with ECOG and PNI scores. After adjusting for these covariates, ECOG was independently associated with survival (HR 1.37 for each one-point increase in ECOG, 95% CI 1.093-1.728, p=0.007). PNI was also independently associated with survival when adjusting for these same covariates (HR 0.95 for each one-point increase in PNI, 95% CI 0.926-0.986, p=0.006). When both ECOG and PNI were included in the same Cox regression model, they each remained independently associated with survival (ECOG HR 1.30, p=0.027; PNI HR 0.96, p=0.022).

Conclusions
ECOG and PNI are independent predictors of poor prognosis in HCC patients receiving immunotherapy. This supports the inclusion of ECOG scores in the BCLC staging system and suggests that pre-treatment nutritional status could similarly be used to refine staging criteria and treatment recommendations. These findings also highlight a need to identify more precise measures of nutritional and functional status in order to improve HCC prognostication. Finally, longitudinal study of those patients whose ECOG/PNI scores improve as a result of ICI-mediated reduction in tumor burden would further expand our understanding of how functional status and nutrition impacts prognosis.

Michael Li

Appearances