Study of China automatic staging model of hepatocellular carcinoma based on big data platform

Abstract

Objective To establish an automatic staging model of hepatocellular carcinoma (HCC)in China based on big data platform. Methods Based on the information system structured by Primary Liver Cancer Big Data (PLCBD) platform of Big Data Institute of Southeast Hepatobiliary Health Information,Meng Chao Hepatobiliary Hospital of Fujian Medical University, the data for HCC staging, such as performance status (PS) score, Child-Pugh grading, extrahepatic metastasis, vascular invasion, number of tumor and tumor size, were rapidly extracted in database visualization mode. The China HCC automatic staging model of was constructed by using CASE-WHEN conditional judgment statement, and was visualized through web page developing. In total, 100 cases with complete PLCBD data were randomly selected for testing. The China HCC automatic staging model was employed for automatic staging. Manual staging for the tested patients was done by 4 attending physicians and 6 resident physicians from Department of Hepatobiliary Surgery. Multidisciplinary consultation was taken as the gold standard to observe the accuracy and practicability of the model. The results among automatic staging and manual staging of two groups were statistically compared by one-way ANOVA. Results Through the database visualization mode, the extraction of PS score, Child-Pugh grading, extrahepatic metastasis, vascular invasion, number of tumor, tumor size and staging-related indexes can be performed. Based on the big data of the above 6 aspects, China HCC automatic staging model was successfully established. The time for conducting automatic staging was 3 s, the average time of manual staging by attending physicians was (40±6) min, and (100±8) min for resident physicians with significant difference between them (F=227.90, P 0.05). Conclusions The China HCC automatic staging model can be successfully established based on the big data platform. The automatic staging model is highly efficient and accurate. Key words: Carcinoma, hepatocellular; Neoplasm staging; Big data; Database