Recently, a hepatic arterial infusion chemotherapy (HAIC)- associated combination therapeutic regimen, comprising HAIC and systemic therapies (molecular targeted therapy plus immunotherapy), referred to as HAIC combination therapy, has demonstrated promising anticancer effects. Identifying individuals who may potentially benefit from HAIC combination therapy could contribute to improved treatment.
This dual- center study was a retrospective analysis of prospectively collected data with advanced HCC patients who underwent HAIC combination therapy and pretreatment contrast- enhanced ultrasound (CEUS) evaluations from March 2019 to March 2023.
Two deep learning models, AE- 3DNet and 3DNet, along with a time- intensity curve- based model, were developed for predicting therapeutic responses from pretreatment CEUS cine images. Diagnostic metrics, including the area under the receiver- operating- characteristic curve (AUC), were calculated to compare the performance of the models. Survival analysis was used to assess the relationship.
South China, Guangdong Provincial Clinical Research Center for Cancer, Department of Hepatobiliary Oncology, Sun Yat- Sen University Cancer Center, Guangzhou, China | 4National- Regional Key Technology Engineering Laboratory for Medical Ultrasound, Guangdong Key Laboratory for Biomedical Measurements and Ultrasound Imaging, School of Biomedical Engineering, Shenzhen University Medical School,.
Keywords: combination therapy | contrast- enhanced ultrasound | deep learning | hepatic arterial infusion chemotherapy | hepatocellular carcinoma ABSTRACT Recently, a hepatic arterial infusion chemotherapy (HAIC)- associated combination therapeutic regimen, comprising HAIC and systemic therapies (molecular targeted therapy plus immunotherapy), referred to as HA
Liu: methodology, software. Huiying Wen: methodology, software. Wei Wang: resources. Jianhua Zhou: resources. Minhua Lu: meth - odology, software. Xin Chen: conceptualization, funding acquisition, resources. Ruhai Zou: conceptualization, data curation, formal anal - ysis, funding acquisition, project administration, supervision, writing – review and editing.
Zhong Liu: funding acquisition, methodology, project administration, visualization, writing – original draft, writing – review and editing. Acknowledgments The authors have nothing to report. Ethics Statement Approval of the research protocol by an Institutional Review Board. This retrospective analysis utilized data from a prospective clinical reg - istry (Ethical Approval no. B2016- 042- 01).
Written informed consent had been obtained from all participants prior to CEUS examinations, which included explicit permission for the utilization of their data in research upon anonymization. Conflicts of Interest The authors declare no conflicts of interest.
the results of which are consistent with those obtained in the internal validation. We also performed the DeLong test, which confirmed statistically significant differences between the ROC curves of AE- 3DNet and those of the other models ( p = 0.041 and p = 0.038 for 3DNet and TIC, respectively).
3.4 | Model Visualization Although AE- 3DNet demonstrated excellent performance, cli - nicians may be more concerned with understanding how the model infers its prediction outcomes.
To interpret the mod - el's behavior, we used Grad- CAM to convert the feature maps (generated by the AE- 3DNet model as it inferred the tumor re - sponse from the CEUS data) into pseudocolored maps, that is, attention maps or heatmaps. By reading the attention maps, we summarized several rules that might be used by AE- 3DNet for decision- making, most of which were related to the feeding ar -.
d external validation cohorts, respectively. AE- 3DNet's predicted response survival curves closely resembled actual clinical outcomes. The deep learning model of AE- 3DNet developed based on pretreatment CEUS cine performed satisfactorily in predicting the responses of advanced HCC to HAIC combination therapy, which may serve as a prom - ising tool for guiding combined therapy and individualized.
Trial Registration: NCT02973685. 1 | Introduction Primary liver cancer ranks as the sixth most commonly di - agnosed cancer and the third leading cause of cancer- related deaths globally [ 1]. Hepatocellular carcinoma (HCC) accounts for the majority (75%–85%) of primary liver cancer cases, with prognosis and treatment strategies varying significantly across stratified stages.
Early- stage HCC typically has a favorable prog - nosis, with curative treatment options such as surgical resec - tion, thermal ablation, and liver transplantation available. By contrast, advanced HCC exhibits a poor prognosis, and research continues to identify more effective therapies for improving the prognosis. Recent advancements in systemic therapy [ 2] and he - patic arterial infusion.
l., “Arterial Chemotherapy for Hepa - tocellular Carcinoma in China: Consensus Recommendations,” Hepa - tology International 18, no. 1 (2023): 4–31. 25. J. M. Llovet and R. Lencioni, “mRECIST for HCC: Performance and Novel Refinements,” Journal of Hepatology 72, no. 2 (2020): 288–306. 26. S. Ji, M. Yang, and K.
Yu, “3D Convolutional Neural Networks for Human Action Recognition,” IEEE Transactions on Pattern Analysis and Machine Intelligence 35, no. 1 (2013): 221–231. 27. R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, “Grad- CAM: Visual Explanations From Deep Networks via Gradient- Based Localization,” International Journal of Computer Vision 128, no. 2 (2020): 336–360. 28.
H.- C. Shin, H. R. Roth, M. Gao, et al., “Deep Convolutional Neural Networks for Computer- Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning,” IEEE Transactions on Medical Imaging 35, no. 5 (2016): 1285–1298. 29. M. Reig, A. Forner, J. Rimola, et al., “BCLC Strategy for Prognosis Prediction and Treatment Recommendation: The 2022 Update,” Jour - nal of Hepatology.