Traumatic diaphragmatic rupture: the diagnostic value of multiplanar reformation in multi-slice spiral CT examination

Abstract

Objective To investigate the diagnostic value of multiplanar reformation (MPR)reconstruction for the detection of traumatic diaphragmatic rupture (TDR) in multi-slice CT examination.Methods Thirty six cases with thoracoabdominal trauma, including 21 cases with and 15 cases without TDR confirmed by surgery, received multi-slice CT examination. They were enrolled in this study. Three experienced radiologists retrospectively analyzed the axial and MPR images. The diagnostic criteria for TDR included abnormally elevated hemidiaphragm, diaphragmatic discontinuity, the collar sign or dependent viscera sign. Referenced to surgical results, the sensitivity, specificity, positive predictive value, negative predictive value and accuracy of axial and MPR images in detection of TDR were calculated. The McNemar was used to investigate the differences between axial and MPR images in the detection of diaphragmatic discontinuity and collar sign, and the differences between axial and MPR images of these two signs in TDR diagnosis. Results The sensitivity, specificity, positive predictive value, negative predictive value and accuracy of axial images in detection of TDR were 71% ( 15/21 ), 80% ( 12/15 ), 83% ( 15/18 ),67% ( 12/18 ) and 75% ( 27/36 ), respectively; of MPR images, they were 86% ( 18/21 ), 93%(14/15), 95% ( 18/19 ), 82% ( 14/17 ) and 89% ( 32/36), respectively. By axial images, twelve diaphragmatic defects or interrupts were identified in nine cases, and collar sign was identified in six cases. By MPR, 20 diaphragmatic defects or interrupts were identified in 15 cases ( P = 0.125 ), and collar sign was identified in 14 cases (P =0.021 ). The sensitivity and specificity of diaphragmatic defects or interrupts for TDR diagnosis in axial images were 43% (9/21) and 80% ( 12/15 ), respectively;in MPRimages, they were71% (15/21) (P=0.125)and93% (14/15) (P=0.500), respectively.The sensitivity and specificity of collar sign for TDR diagnosis in axial images were 29% (6/21) and 100% ( 15/15), respectively; in MPR images, they were 67% ( 14/21 ) (P =0. 021 ) and 100% (15/15)( P = 1.000), respectively. Conclusions MSCT presented good sensitivity, specificity and accuracy for the diagnosis of TDR. MPR images were useful supplements for axial images in TDR diagnosis which improved the diagnosis. Key words: Wounds and injuries; Tomography,X-ray computed; Comparative study