Editor's Note: Transarterial chemoembolization (TACE) is the recommended treatment for intermediate-stage hepatocellular carcinoma (HCC) patients and a viable option for early-stage HCC patients who cannot undergo curative treatments. However, the prognosis for TACE varies significantly among individuals, with objective response rates ranging from 40% to 80% and overall survival times from 13 to 48 months. Given the advancements in molecular targeted therapy and immunotherapy for HCC, accurately predicting individual prognosis and identifying patients with poor prognosis is crucial for optimizing treatment choices and extending survival. On June 6, 2024, at the EASL Annual Meeting in Milan, Professor Guohong Han from Xi'an International Medical Center Hospital in China shared the latest clinical research. Their team developed a predictive model for TACE outcomes, helping guide treatment choices for intermediate HCC patients and improving their survival (Abstract No.: THU-475).

Understanding the Heterogeneity of HCC and TACE Prognosis

HCC is a highly heterogeneous malignant tumor. Factors such as tumor burden, liver function, tumor biology, and physical condition all influence TACE outcomes. Previous research has developed several early prediction models for TACE. However, most of these models are binary, leading to loss of prognostic information or poor validation in external datasets. Against this backdrop, our team recently proposed the “6-and-12 model” based on tumor burden [maximum tumor size (ts, cm) plus tumor number (tn)]. This model, suitable for TACE candidates recommended by guidelines, predicts TACE outcomes and stratifies patients based on survival time (cutoffs of 6 and 12).

Advancements in the “6-and-12 Model”

Notably, the predictive performance of the “6-and-12 model” can still be improved. Alpha-fetoprotein (AFP) is a marker reflecting tumor biology and an influencing factor for patient prognosis. However, the clinical application of AFP remains controversial. Firstly, AFP has been widely used in previous predictive models, but its cutoff value is still inconsistent. Secondly, the linear or nonlinear relationship between AFP as a continuous variable and post-TACE survival outcomes needs further study. Additionally, the interaction between AFP and other predictive factors is unclear.

Study Objective and Methodology

We aimed to develop and validate a novel, easily accessible TACE prognostic prediction model to improve the prognosis of intermediate HCC patients and guide patient stratification.

The study included 4,377 initially treated HCC patients from 39 centers across five countries. These patients were divided into a training dataset (1,604 patients from 24 centers in China), a domestic internal validation dataset (803 patients from 12 centers), a European validation dataset (1,130 patients from six centers in two countries), and an Asian validation dataset (840 patients from three centers in two countries). A Cox multivariate regression analysis was used to develop the new model, which was compared with our original “6-and-12 model” and other models.

Results

The Cox multivariate analysis identified ts, tn, and AFP values as prognostic factors, generating a linear model: “ts + tn + 1.5 × Log10 AFP,” referred to as the “6-and-12 Model 2.0.” The new model showed excellent discrimination (C-index 0.674) and calibration, outperforming existing models.

Using different AFP levels (≤100, 100-400, 400-2000, 2000-10,000, 10,000-40,000, and >40,000 ng/mL) and corresponding tumor burden thresholds (8/14, 7/13, 6/12, 5/11, 4/10, and any tumor burden), we proposed an easy-to-use stratification method. For example, if a patient’s AFP level is between 400-2000 ng/mL, they can be stratified into low risk (ts + tn ≤ 6), medium risk (6 < ts + tn ≤ 12), and high risk (ts + tn > 12) groups. Overall, the “6-and-12 Model 2.0” stratifies patients into three risk levels, with median overall survival times of 45.0 (95% CI: 40.1-49.9), 30.0 (95% CI: 26.1-33.9), and 15.4 (95% CI: 13.4-17.4) months, respectively (P < 0.001). These results were confirmed in validation cohorts and subgroup analyses.

Conclusion

In this international multicenter study, we developed and validated a novel, user-friendly model, “6-and-12 Model 2.0,” for recommended TACE candidates. This model demonstrated excellent discrimination and calibration in predicting outcomes, stratifying patients into three risk groups with significantly different survival times. The “6-and-12 Model 2.0” can be used as a tool for predicting the survival of HCC patients, aiding in better clinical decision-making and improving patient survival.