Clinical Application Value Of Artificial Intelligence Detection Technology Based On CT Image In Rib Fracture Diagnosis

Background and purposeRib fractures are the most common chest trauma in the clinic,and early and accurate diagnosis is of great significance for the treatment of patients and reducing medical disputes.With the development of examination technology,CT examination has become the most common examination method for chest trauma.However,the diagnosis of rib fractures by multi-slice spiral CT is huge,time-consuming,and easy to miss and misdiagnose.In recent years,artificial intelligence technology has developed rapidly in the field of medical imaging,and its scope has continued to expand.Rib fractures have become a hotspot for many artificial intelligence companies due to their simple but time-consuming diagnosis.This study intends to explore the value of artificial intelligence(AI)detection system in the diagnosis of rib fractures in CT images.Provide preliminary verification for the application value of artificial intelligence(AI)technology in the field of rib fracture diagnosis.MethodsData from 190 cases of rib fracture-positive patients and 74 cases of rib fracture-negative patients examined in the Department of Medical Imaging,the Fifth Affilated Hospital of Zhengzhou University from January 2018 to January 2020 were collected as test set data.In the study,four methods were used to review the films of all the patients and record the results and the time of reviewing films.Method A:All image data were "reading" by uAI BoneCare software;After reviewing the film by the imaging doctor of Method B,the reviewing physician will review its results;Method D:The method of reporting by the reporting doctor combined with rib fracture assisted detection software(uAI BoneCare).The results of rib fracture diagnosis and radiographing time were recorded.The accuracy,sensitivity,specificity,number of false positive markers and number of false negative markers of the four methods were calculated and compared with the results of the standard group,and the methods B and method D were compared.ResultsMethods A:The diagnostic accuracy of rib fracture assisted detection software(uAI BoneCare)was 93.5%,the sensitivity was 78.3%,and the specificity was 96.6%.Methods B:The diagnosis rate of rib fractures by doctor was 95.6%,the sensitivity was 87.8%,and the specificity was 97.2%.MethodC:After the doctor B read the film,the review physician’s accuracy of diagnosis of rib fracture was 94.0%,the sensitivity was 74.4%,and the specificity was 98.0%.Methods D:Doctor combined with rib fracture assistant detection software(uAI BoneCare)suggested that the diagnosis accuracy of rib fractures was 97.1%,sensitivity was 94.5%,and specificity was 97.6%.Methods B The average time for doctor to diagnose rib fractures was 2.75 minutes/case.The average time for doctor to diagnose rib fractures combined with uAI BoneCare software was 2.46 minutes/case.ConclusionRib Fracture auxiliary detection software has a good ability to detect rib fracture in chest CT,which can approximately reach the level of attending physician in imaging department.Artificial intelligence soft rib fracture auxiliary detection as an auxiliary diagnosis in the daily work can play a role in reducing missed diagnosis and impoving the efficiency of diagnosis.But at present,the artificial intelligence software still has a high rate of missed diagnosis and misdiagnosis.We need to further train and optimize the software in combination with the various causes of error diagnosis.