Traumatic injuries of the pelvis and thoracic and lumbar spine: does thin-slice multidetector-row CT increase diagnostic accuracy?

The objective is to evaluate different multidetector-row CT (MDCT) strategies for adequate classification of spinal and pelvic injuries. Seventy intubated patients after multiple trauma underwent conventional radiography (CR) and MDCT. Examinations included the pelvis (P), the lumbar spine (LS) and the thoracic spine (TS). Conventional radiographs, 3-mm (CT5) and 5-mm scans (CT3) and 3-mm and 5-mm scans combined with MPR (CT3R/CT5R) were compared to surgery, autopsy and clinical course. MDCT led to significantly better results than CR (P<0.01). Correlation coefficients were r=1.0 (CT3R), r=0.96 [TS] to r=1.0 [P/LS] (CT5R), r=0.8 [P] to r=1.0 [TS] (CT3), r=0.80 [P] to r=0.86 [TS] (CT5) and r=0.3 [TS] to r=0.69 [P] (CR). Fractures were identified by CT3R in 100% of cases, by CT5R in 95%, by CT3 in 90% [P]-100% [TS], by CT5 in 83.3% [LS]-90% [P] and by CR in 57.1% [TS]-87.2% [P]. Unstable fractures were identified in 100% by CT3R, CT5R and CT3, 85.7% [TS]-100% [P/LS] by CT5 and 57.1% [TS]-80% [P] by CR. Only overlapping thin-slice multiplanar reformation allows for an adequate classification of spinal and pelvic injuries and thus is highly emphasized in patients after severe blunt trauma. Copyright 2004 Springer-Verlag