The Characteristic Analysis And Clinical Correlation Research Of Lisfranc Injury Based On CT Imaging Assisted By AI

Objective: To explore the possibility of the diagnosis of Lisfranc injury and the epidemiological analysis of Lisfranc injury with big data assisted by AI intelligence.Methods: from January 2017 to October 2018,326 cases of Lisfranc injured patients were collected from Shanghai Sixth People’s hospital.The Xinjian arigin3 d pro three-dimensional medical reconstruction software V5.0 was used to reconstruct and analyze 100 of the cases to improve the software algorithm,so that the software can identify the original CT data,describe the injury location and degree.After software identification and analysis,the remaining cases were compared with the identification results of imaging doctors and clinicians(foot and ankle surgery),the identification errors were corrected,the algorithm was improved,and Statistical analysis of characteristics of Lisfranc injuries was carried out.after detailed description of all cases.Results: the recognition rate of AI software was about 70%,but there were still some false positive and false negative rates.After artificial recognition,the 326 CT data included in this study included208 males(63.8%),118 females(37.2%),the most common age group was19-44 years old(42.9%),followed by middle-aged patients(32.2%)..According to the theory of three column classification,149 cases(45.7%)suffered the most from three column injury,and the incidence of medial column injury was the highest(68.1%).The proportion of medial and lateral column injuries was 19.6% and 21.8% respectively.Among all the patients,there were 7 cases of ligament injury,70 cases of fracture dislocation,1 case of simple dislocation and 248 cases of simple fracture,245 cases of cuneiform bone fracture,35 cases of cuboid compression fracture and 16 cases of Chopart joint involvement.The most important finding of this study is that Lisfranc injury combined with cuneiform bone fracture accounts for a large proportion(245 / 326,75.2%).Through single factor analysis,we found that there is a certain relationship between age and medial column injury,and Lisfranc injury has a gender distribution difference.The second metatarsal bone is related to the middle cuneiform bone injury.There is a relationship between the second and third metatarsals,the third metatarsals,the lateral cuneiform bone and the cuboid bone have a certain relationship.The damage of the middle column is related to the damage of the lateral column.Conclusion: AI intelligent software still needs further improvement before it is put into use.More data need to be identified and analyzed,incidence rate of Lisfranc injury is higher than that of literature,and there are differences in age and gender distribution.The injury of the medial column is more common.In clinical work,the injury of the medial column should be carefully evaluated to reduce the incidence of trauma sequelae.