Application of deep learning-based chest CT auxiliary diagnosis system in occult rib fractures patients
Objective To evaluate the diagnostic efficacy and potential application value of deep learning-based chest CT auxiliary diagnosis system in occult rib fractures patients.Methods In this retrospective study,95 patients with suspected rib fracture for emergency CT examination,who were randomly selected,and then introduced them into DL auxiliary diagnosis system and obtained the original report results of Radiologist.Based on the results of chest CT Reexamination in 2-3 weeks,the diagnostic efficiency of DL auxiliary diagnosis system and Radiologist was compared.Results among 95 patients,84 cases were diagnosed as rib fracture by reexamination.The sensitivity,specificity,positive predictive value and negative predictive value of DL auxiliary diagnostic system for rib fracture were 98.81%,81.82%,97.64%and 90.00%,respectively.The sensitivity,specificity,positive predictive value and negative predictive value of Radiologist were 80.95%,81.82%,97.14%and 36.00%respectively.The sensitivity and negative predictive value of DL auxiliary diagnostic system in the diagnosis of rib were significantly higher than those of radiologists(Z=4.248,P<0.001).Conclusion DL auxiliary diagnosis system can automatically identify rib fracture with low missed diagnosis rate,which plays an important role in optimizing the diagnosis and treatment process of emergency rib fracture.
万方数据
维普
