Background The combination of atezolizumab and bevacizumab (AtezoBev) is the current first‐line treatment for patients with hepatocellular carcinoma (HCC). Our aim was to evaluate the prognostic role of alpha‐foetoprotein (AFP) early response and its combination with albumin–bilirubin (ALBI) in these patients. Methods Patients with HCC under AtezoBev with AFP > 20 ng/ml were included in three centres. The optimal threshold of AFP variation after 3 weeks of treatment was identified for overall survival (OS) and radiological response (RR) using RECIST 1.1 and mRECIST and its ability to predict progression‐free survival (PFS) and OS was tested using univariate and multivariate analysis in derivation and validation cohorts. Results Seventy‐five patients with AFP values >20 ng/ml were included. Fifty‐eight patients were male with a median age of 63.5 years; 73% had cirrhosis and HCC stage was classified as BCLC B (18.7%) or C (81.3%). In the derivation cohort (n = 38), a decline in AFP ≥ 20% at 3 weeks (AFP early response) was associated with RR using mRECIST criteria (OR: 13.09 95% CI: 1.44–19.34 p = .02), PFS (HR: 0.42; 95% CI: 0.19–0.93, p = .03) and OS (HR: 0.35; 95% CI: 0.15–0.83, p = .01). AFP early response was confirmed as predictor of RR (p = .02 for mRECIST) and OS (p = .03) in the validation cohort (n= 37). In the whole cohort, the combination of ALBI and AFP early response was significantly associated with OS (p = .046) and PFS (p = .012) with a poor prognosis in patients belonging to the ALBI2‐AFP non‐responders class. Conclusion AFP early response at 3 weeks predicts oncological outcomes in HCC patients treated with AtezoBev and combination with ALBI grade refines prognostic discrimination.

Background and Aims: The natural history of hepatocellular adenomas (HCAs) remains to be better described, especially in nonresected patients. We aim to identify the predictive factors of HCA evolution after estrogen‐based contraception discontinuation. Approach and Results: We retrospectively included patients with a histological diagnosis of HCA from three centers. Clinical, radiological, and pathological data were collected to identify predictive factors of radiological evolution per Response Evaluation Criteria in Solid Tumors, version 1.1, and occurrence of complications (bleeding, malignant transformation). We built a score using variables that modulate estrogen levels: body mass index and duration of estrogen‐based contraception. An external cohort was used to validate this score. 183 patients were included in the cohort, including 161 women (89%) using estrogen‐based contraception for a median of 12 years. Thirty percent of patients had at least one HNF1A‐inactivated HCA, 45.5% at least one inflammatory HCA, and 11% at least one HCA with activation of β‐catenin (bHCA). Twenty‐one symptomatic bleedings (11%) and eleven malignant transformations (6%) occurred. Ages < 37 years old (p = 0.004) and HCA > 5 cm at imaging were independently associated with symptomatic bleeding (p = 0.003), whereas a bHCA was associated with malignant transformation (p < 0.001). After a median follow‐up of 5 years, radiological regression was observed in 31%, stabilization in 47%, and progression in 22% of patients. Weight loss was associated with regression (p < 0.0001) and weight gain with progression (p = 0.02). The estrogen exposure score predicted radiological regression (odds ratio, 2.33; confidence interval 95%, 1.29–4.19; p = 0.005) with a linear relationship between the rate of estrogen exposure and the probability of regression. This result was confirmed in an external cohort of 72 female patients (p = 0.003). Conclusion Weight variation is strongly associated with radiological evolution after oral contraception discontinuation. A score of estrogen exposure, easily assessable in clinical practice at diagnosis, predicts regression of HCA.