Background We attempted to train and validate a model of deep learning for the preoperative prediction of the response of patients with intermediate-stage hepatocellular carcinoma (HCC) undergoing transarterial chemoembolization (TACE). Method All computed tomography (CT) images were acquired for 562 patients from the Nan Fang Hospital (NFH), 89 patients from Zhu Hai Hospital Affiliated with Jinan University (ZHHAJU), and 138 patients from the Sun Yat-sen University Cancer Center (SYUCC). We built a predictive model from the outputs using the transfer learning techniques of a residual convolutional neural network (ResNet50). The prediction accuracy for each patch was revaluated in two independent validation cohorts. Results In the training set (NFH), the deep learning model had an accuracy of 84.3% and areas under curves (AUCs) of 0.97, 0.96, 0.95, and 0.96 for complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD), respectively. In the other two validation sets (ZHHAJU and SYUCC), the deep learning model had accuracies of 85.1% and 82.8% for CR, PR, SD, and PD. The ResNet50 model also had high AUCs for predicting the objective response of TACE therapy in patches and patients of three cohorts. Decision curve analysis (DCA) showed that the ResNet50 model had a high net benefit in the two validation cohorts. Conclusion The deep learning model presented a good performance for predicting the response of TACE therapy and could help clinicians in better screening patients with HCC who can benefit from the interventional treatment. Key Points • Therapy response of TACE can be predicted by a deep learning model based on CT images. • The probability value from a trained or validation deep learning model showed significant correlation with different therapy responses. • Further improvement is necessary before clinical utilization.

A ruptured hepatocellular carcinoma (HCC) is often fatal. In addition to surgery and transarterial embolization, radiofrequency ablation (RFA) might be another option for treating a ruptured HCC. Unfortunately, conventional RFA has a limited ablation zone; as such, it is rarely used to treat ruptured tumors. This case was a 60-year-old man who had a large, ruptured HCC in which hydrochloric acid (HCl)-enhanced RFA successfully controlled the bleeding and made the tumor completely necrotic. Considering the effectiveness of HCl-enhanced RFA in achieving hemostasis and tumor ablation, it might be a new option for treating large, ruptured HCCs.

Objectives: We aimed to compare the therapeutic outcomes of radiofrequency ablation (RFA) and microwave ablation (MWA) as first-line therapies in patients with small single perivascular hepatocellular carcinoma (HCC).Methods: A total of 144 eligible patients with small (≤ 3 cm) single perivascular (proximity to hepatic and portal veins) HCC who underwent RFA (N= 70) or MWA (N= 74) as first-line treatment were included. The overall survival (OS), disease-free survival (DFS), and local tumor progression (LTP) rates between the two ablation modalities were compared. The inverse probability of treatment weighting (IPTW) method was used to reduce selection bias. Subgroup analysis was performed according to the type of hepatic vessels.Results: After a median follow-up time of 38.2 months, there were no significant differences in OS (5-year OS: RFA 77.7% vs. MWA 74.6%;p= 0.600) and DFS (5-year DFS: RFA 24.7% vs. MWA 40.4%;p= 0.570). However, a significantly higher LTP rate was observed in the RFA group than the MWA group (5-year LTP: RFA 24.3% vs. MWA 8.4%;p= 0.030). IPTW-adjusted analyses revealed similar results. The treatment modality (RFA vs. MWA: HR 7.861, 95% CI 1.642–37.635,p= 0.010) was an independent prognostic factor for LTP. We observed a significant interaction effect of ablation modality and type of peritumoral vessel on LTP (p= 0.034). For patients with periportal HCC, the LTP rate was significantly higher in the RFA group than in the MWA group (p= 0.045). However, this difference was not observed in patients with perivenous HCC (p= 0.116).Conclusions: In patients with a small single periportal HCC, MWA exhibited better tumor control than RFA.Key Points: • Microwave ablation exhibited better local tumor control than radiofrequency ablation for small single periportal hepatocellular carcinoma.• There was a significant interaction between the treatment effect of ablation modality and type of peritumoral vessel on local tumor progression.• The type of peritumoral vessel is vital in choosing ablation modalities for hepatocellular carcinoma.

Cyclin dependent kinase inhibitor 2A (CDKN2A) is an essential regulator of immune cell functionality, but the mechanisms whereby it drives immune infiltration in hepatocellular carcinoma (HCC) remain unclear. In the present study, we studied the association with CDKN2A expression and immune invasion with the risk of developing HCC. A totally of 2207 different genes were found between HCC and adjacent liver tissues from TCGA and GEO databases. CDKN2A was highly expressed in HCC and associated with poorer overall survival and disease-free survival. Notably, CDKN2A expression was positively correlated with infiltrating levels into purity, B cell, CD+8 T cell, CD+4 T cell, macrophage, neutrophil, and dendritic cells in HCC. CDKN2A expression showed strong correlations between diverse immune marker sets in HCC. These findings suggest that CDKN2A expression potentially contributes to regulation of tumor-associated macrophages and can be used as a prognostic biomarker for determining prognosis and immune infiltration in HCC.

BACKGROUND. Despite an increasing appreciation of the roles that myeloid cells play in tumor progression and therapy, challenges remain in interpreting the tumor-associated myeloid response balance and its translational value. We aimed to construct a simple and reliable myeloid signature for hepatocellular carcinoma (HCC). METHODS. Using in situ immunohistochemistry, we assessed the distribution of major myeloid subtypes in both peri- and intratumoral regions of HCC. A 2-feature-based, myeloid-specific prognostic signature, named the myeloid response score (MRS), was constructed using an L1-penalized Cox regression model based on data from a training subset (n = 244), a test subset (n = 244), and an independent internal (n = 341) and 2 external (n = 94; n = 254) cohorts. RESULTS. The MRS and the MRS-based nomograms displayed remarkable discriminatory power, accuracy, and clinical usefulness for predicting recurrence and patient survival, superior to current staging algorithms. Moreover, an increase in MRS was associated with a shift in the myeloid response balance from antitumor to protumor activities, accompanied by enhanced CD8⁺ T cell exhaustion patterns. Additionally, we provide evidence that the MRS was associated with the efficacy of sorafenib treatment for recurrent HCC. CONCLUSION. We identified and validated a simple myeloid signature for HCC that showed remarkable prognostic potential and may serve as a basis for the stratification of HCC immune subtypes. FUNDING. This work was supported by the National Science and Technology Major Project of China, the National Natural Science Foundation of China, the Science and Information Technology of Guangzhou, the Fundamental Research Funds for the Central Universities, the Guangdong Basic and Applied Basic Research Foundation, and the China Postdoctoral Science Foundation.

Objectives: We aimed to develop radiology-based models for the preoperative prediction of the initial treatment response to transarterial chemoembolization (TACE) in patients with hepatocellular carcinoma (HCC) since the integration of radiomics and deep learning (DL) has not been reported for TACE.Methods: Three hundred and ten intermediate-stage HCC patients who underwent TACE were recruited from three independent medical centers. Based on computed tomography (CT) images, recursive feature elimination (RFE) was used to select the most useful radiomics features. Five radiomics conventional machine learning (cML) models and a DL model were used for training and validation. Mutual correlations between each model were analyzed. The accuracies of integrating clinical variables, cML, and DL models were then evaluated.Results: Good predictive accuracies were showed across the two cohorts in the five cML models, especially the random forest algorithm (AUC = 0.967 and 0.964, respectively). DL showed high accuracies in the training and validation cohorts (AUC = 0.981 and 0.972, respectively). Significant mutual correlations were revealed between tumor size and the five cML models and DL model (each P < 0.001). The highest accuracies were achieved by integrating DL and the random forest algorithm in the training and validation cohorts (AUC = 0.995 and 0.994, respectively).Conclusion: The radiomics cML models and DL model showed notable accuracy for predicting the initial response to TACE treatment. Moreover, the integrated model could serve as a novel and accurate method for prediction in intermediate-stage HCC.