The liver is an important metabolic organ in the human body, and the exploration and clinical research of liver diseases has always been the top priority of the development of my country's medical and health undertakings. Hepatocellular carcinoma (HCC) is the sixth most common malignancy worldwide and the third leading cause of cancer-related death.

Currently, surgical resection is the main curative method for HCC. However, about 60-70% of HCC patients experience recurrence within 5 years after surgery, and the long-term prognosis is still poor. How to evaluate and predict the recurrence and prognosis of HCC after surgical resection has become a research hotspot. HCC also has extremely high tumor heterogeneity, specifically, patients with the same stage and the same treatment can show very different survival conditions, while patients with different stages and different treatments can show similar survival conditions, which seriously restricts the patients. precise and standardized treatment.

In response to relevant research hotspots and topics, recently, Professor Hu Hongjie's team from the Department of Radiology, Run Run Run Run Shaw Hospital, Zhejiang University School of Medicine, together with Professor Yang Yunjun's team from the First Affiliated Hospital of Wenzhou Medical University, and Shangtang Technology, are online in the Journal of Hepatocellular Carcinoma (2021 IF: 4.962) Published a research paper titled "CT-Based Radiomics for the Recurrence Prediction of Hepatocellular Carcinoma After Surgical ResecTIon".

This paper explores objective imaging markers for the prediction of HCC recurrence. By mining radiographic features, it can predict recurrence-free survival (RFS) of HCC patients after surgical resection, and help the clinic to further screen for possible outcomes from postoperative adjuvant transarterial chemoembolization. patients benefiting from surgery (PA-TACE).

This study combines the local characteristics of the tumor itself and the overall health of the liver parenchyma to comprehensively evaluate the treatment plan, so that the prognosis prediction method can be transformed from macroscopic tumor morphological grading to refining the microbiological information of the tumor through artificial intelligence, which is helpful for clinical formulation. A reasonable and precise treatment plan can reduce the postoperative recurrence rate of HCC patients and improve the overall prognosis.

AI unleashes the value of CT images and effectively predicts recurrence

Based on the multi-center data of 364 cases, the two teams established a CT-based radiomics model, a clinical model based on clinical factors, and a combined model based on imaging and clinical factors to predict RFS. The consistency index (C-index) was used to predict RFS. The performance of the three models is evaluated.

In addition, the team divided all patients into high-risk and low-risk groups based on radiomic scores to analyze the benefit from PA-TACE.

The results of the study showed that the radiomics model showed good predictive ability for postoperative recurrence of HCC on the training set, internal validation set, and external test set, with C-index of 0.892, 0.812, and 0.809, respectively. Univariate and multivariate COX analysis showed that radiomics score was an independent prognostic factor. There were significant differences in RFS prediction between high and low recurrence risk groups in all three datasets.

Further analysis showed that patients in the high recurrence risk group benefited significantly from PA-TACE, while those in the low recurrence risk group had little benefit.

Figure 1: Kaplan-Meier survival analysis (A) and time-dependent ROC curve (B) based on radiomics score.

Figure 2: Nomogram (A) for predicting RFS, along with calibration curve (B) and decision curve (C).

Figure 3: Kaplan-Meier analysis curve used to assess the benefit of PA-TACE. (A) No PA-TACE group; (B) PA-TACE group; (C) Low recurrence risk group; (D) High recurrence risk group.

Paper DOI: 10.2147/JHC.S362772

In-depth cutting-edge research, enabling clinical applications

The choice of treatment options for HCC, such as surgical resection, anatomical resection or non-anatomical resection, etc., needs to be comprehensively considered from many aspects such as tumor biological behavior, liver function, and patient laboratory indicators. Depends on invasive pathology.

In recent years, many clinical researchers have tried to use radiological imaging methods for non-invasive preoperative prediction of tumor microbiological behavior and liver function. In this paper, the prediction and prognosis model based on artificial intelligence method can take into account the accuracy and practicability, and is consistent with the existing It is fully compatible with the clinical workflow of the system, which is of great significance in guiding clinical decision-making.

In fact, around the application of AI technology in the clinical field of liver diagnosis and treatment, SenseTime began to carry out in-depth cooperation with Run Run Run Shaw Hospital very early. The two teams' in-depth radiological research on β-catenin mutation and prognosis prediction of hepatocellular carcinoma based on multimodal imaging, and the research on the development of LI-RADS accurate diagnosis and efficacy evaluation system based on artificial collaborative intelligence, were selected for the National Natural Science Project and Zhejiang Province respectively. Provincial Science and Technology Department key projects.

SenseCare liver series products landed in the Department of Radiology and Hepatobiliary Surgery of Run Run Run Shaw Hospital

SenseCareLiver CT liver CT intelligent analysis system, SenseCareLiver MR liver MR intelligent analysis system, and SenseCareLiver Surgical Planning liver intelligent surgery planning system of SenseTime have been applied in radiology, hepatobiliary surgery and other departments, seamlessly embedded in the workflow, helping Doctors efficiently and quickly complete the diagnosis and treatment of abdominal diseases and serve the majority of patients.

Benefiting from the long-term accumulation of technical advantages and clinical practice experience, SenseTime’s global original integrated solution for abdominal diseases that integrates CT and MRI multimodality, covers the whole clinical process of liver focal lesion detection, qualitative, quantitative, and treatment recommendation. , has been successfully implemented in dozens of medical institutions across the country, providing radiologists with a full range of image reading assistance, helping hepatobiliary surgeons to improve the efficiency and accuracy of individualized preoperative assessment and planning.

Abdominal products are only one of the clinical applications carried by Shangtang Technology's original SenseCare intelligent diagnosis and treatment platform. The platform integrates leading artificial intelligence algorithms and rich 3D post-processing functions, aiming to provide convenient and efficient intelligent auxiliary diagnosis and treatment tools for different departments. It has formed a high-performance auxiliary diagnosis and treatment solution covering more than 10 human body parts and organs, with intelligent analysis capabilities for multiple diseases, and oriented to multiple departments. It provides intelligent tools covering the whole process for clinical diagnosis, treatment and recovery, and benefits doctors ,patient.

Original title: Predicting the recurrence rate of liver cancer after surgery, Shang Tang and Run Run Shaw Hospital jointly researched and released the latest AI results

Article source: [WeChat public account: SenseTIme] Welcome to add attention! Please indicate the source when reprinting the article.

Reviewing Editor: Peng Jing

Leave a Reply

Your email address will not be published.