Home  >  Medical  >  Frontiers of Medicine  >  New model for predicting peritoneal metastasis of gastric cancer

New model for predicting peritoneal metastasis of gastric cancer

04-03 BigMediumSmall I want to comment

gastric cancer.jpg

With the cooperation of the Institute of Automation of the Chinese Academy of Sciences, the Key Laboratory of Molecular Imaging of the Chinese Academy of Sciences, Ji Jiafu Team of the Cancer Hospital of Peking University, Gao Jianbo Team of the First Affiliated Hospital of Zhengzhou University, Shan Xiuhong Team of the Medical Imaging Department of the First People's Hospital of Zhenjiang City, Jiangsu Province, China and the Radiology Team of Yunnan Cancer Hospital Effective prediction of occult peritoneal metastasis in patients with gastric cancer. Especially for the occult peritoneal metastasis missed by CT, this model has a very high detection rate. This paper was published in the clinical oncology Journal Oncology Yearbook recently.

About half of all new and fatal cases of gastric cancer worldwide come from China. Distant metastasis of gastric cancer is one of the main causes of death. According to statistics, 53% to 66% of patients with distant metastasis of gastric cancer are peritoneal metastasis to adjacent areas. Surgery can not prolong the survival of these patients. The guidelines of the United States, China and Europe do not recommend surgical treatment. Therefore, accurate preoperative diagnosis of peritoneal metastasis can effectively assist doctors to make treatment decisions and avoid unnecessary surgery. At present, CT imaging is a common non-invasive method for the diagnosis of peritoneal metastasis before operation, but this method can not observe the obvious signs of peritoneal metastasis, which can easily lead to missed diagnosis and affect the treatment decision-making and prognosis of patients.

The study collected imaging and clinical data of 554 patients with preoperative CT negative peritoneal metastasis in four hospitals in China, 122 of whom were confirmed to be peritoneal metastasis positive by laparoscopic pathology, and these patients were usually misdiagnosed clinically. At the same time, the Chinese team extracted 266 quantitative imaging histological features based on the largest primary image of gastric cancer and a peritoneal image adjacent to the lesion in patients'venous CT images. Researchers found that peritoneal metastasis was significantly correlated with imaging features of 2 primary lesions, 2 peritoneal imaging features, and Lauren classification of 1 clinical factor. Based on this, an imaging histology prediction model was constructed. Based on the data of four hospitals, the research team validated that the model had good prediction effect, and the AUC of model performance evaluation index was above 0.92.

It is understood that previous studies have focused on peritoneal imaging itself. This study found that peritoneal metastasis of gastric cancer is related to both the primary lesion and peritoneum of gastric cancer. It also verified the classical "soil-seed theory" of tumor metastasis, that is, tumor metastasis is not only related to the tumor cells themselves (seeds), but also to the microenvironment (soil) of the metastasis area. The study also found that the heterogeneity of peritoneal imaging has good predictive performance, or can reflect some subtle changes of early peritoneal metastasis that are not easily detected by human eyes. Imaging histology can quantitatively extract these subtle signs by digging in-depth image data to assist doctors in diagnosing peritoneal metastasis of gastric cancer.