Predicting early progression to atezolizumab-bevacizumab in hepatocellular carcinoma: a clinical and imaging-based...

To develop a predictive model incorporating both clinical and imaging findings to predict early progression in patients with advanced hepatocellular carcinoma (HCC) undergoing atezolizumab plus bevacizumab (Atezo-Bev) therapy.

A total of 140 consecutive patients with HCC who initiated Atezo-Bev therapy between January 2020 and May 2022 at two tertiary care centres were retrospectively enrolled. Early progression was defined as progressive disease in the first response evaluation conducted at 4 weeks and 12 weeks after treatment initiation using dynamic CT or MRI. Images were reviewed by two radiologists. Logistic regression analysis was performed to determine the early progression scores.

The first response evaluation of Atezo-Bev therapy was conducted at a median of 56 days (interquartile range, 42-64 days) after treatment initiation. Approximately 40% (56/140) of patients with HCC showed early progression. The early progression score was defined as follows: (age < 60 years; 1 point) + (serum alpha-fetoprotein level ≥ 300 ng/mL; 3 points) + (neutrophil-to-lymphocyte ratio ≥ 2.8; 1 point) + (infiltrative appearance; 2 points). At a score of 3 or higher, the early progression score showed sensitivity of 91.1% (95% confidence interval [CI]: 83.6-98.5%) and a specificity of 53.6% (95% CI: 42.9-64.2%). At a score of 6 or higher, the score demonstrated a sensitivity of 55.4% (95% CI: 42.3-68.4%) and a specificity of 91.7% (95% CI: 85.8-97.6%).

We developed an early progression score that integrates clinical and imaging factors with high specificity to accurately predict early progression in patients with advanced HCC undergoing Atezo-Bev therapy.

Question No validated imaging-based tool currently exists to predict early progression to Atezo-Bev therapy in advanced HCC. Findings Infiltrative tumour appearance, high AFP, high neutrophil-to-lymphocyte ratio, and younger age were significantly associated with early progression. Clinical relevance A composite early progression score integrating clinical and imaging features showed high diagnostic accuracy for predicting treatment failure.