Results: CT Clinical Practice: Data acquisition protocols


Trauma

1. I am developing a whole body, vertex to pubis, protocol for severely traumatized patients on a Philips MX8000 (Quad) MDCT. I am aware of MGH and UT Houston protocols for GE lightspeed QX/I MDCT based on 5 mm slice thickness. On the MX8000, I am concerned 6.5 slice thickness for the thoracic aorta is too thick to allow 3D VR, but for the abdomen, 3.2 slice thickness is too thin. If we need MPRs of the thoracic and lumbar spine, 6.5 mm is too thick. Even with a 4 channel scanner, scanning the head and c-spine, then doing 3.2s of the C/A/P is challenging to the tube, and generates over a thousand images. Can you direct me to any literature on MX8000 protocols for blunt trauma whole body CT?

Answer: Call Phillips directly, and ask for Irene. She usually has the Phillips answers. Remember that even 3 mm is not optimal if detailed MPR's or 3D's are done.


2. What is the best CT protocol to demonstrate vascular damage to the kidneys? For a trauma patients with retroperitoneal bleed, in order to determine if renal perfusion was adequate (i.e. active bleeding). Literature (Harris, Radiographics 2001) states that only venous (60 second) and late images (3-5 mm) are necessary. What is your opinion?

Answer: Assuming that the patient is not hypotensive, a good protocol would be arterial phase at 25 seconds, late corticomedullary phase at 50-60 seconds and a 3-4 minute delayed phase. If you use thin sections, you can do excellent CT angiography of the renal artery and make the diagnosis of renal artery injury. If the patient is in shock/hypotensive then the delay may need to be longer.


3. What is your protocol for evaluating the chest, abdomen and pelvis in trauma patients- do you perform a noncontrast CT for the upper abdomen before performing the contrast CT?

Answer: We go straight to contrast enhanced CT and do not do non-contrast studies. I have never been convinced the added time and radiation provides any useful information that the contrast scans will not provide. Let me know if you feel otherwise?


4. Blunt abdominal trauma and vascular lesions of the spleen are strongly associated. Traumatic intrasplenic pseudoaneurysms are acknowledged mechanism of delayed splenic rupture. In our experience, such lesions can be difficult to detect with standard predelay (ca. 60 sec). Would you recommend to scan earlier, say after 35 seconds, in addition, if extravasation or splenic injury is suspected?

Answer: If splenic artery aneurysm or bleed is suspected I would use arterial phase imaging at 25 seconds for optimal detection.


References


Roos JE, Hilfiker P, Platz A et al. MDCT in emergency radiology: Is a standardized chest or abdominal protocol sufficient for evaluation of thoracic or lumbar spine trauma? AJR 2004; 182: 959-968.

  • Summary: In this series of 82 trauma patients, 4 slice MDCT was performed with 2.5 mm collimation. In patients with spine fractures, an additional acquisition was performed to evaluate the spine, using 4 x 1 mm collimation. From the original datasets using 2.5 mm collimation, targeted spine reconstructions were performed with 3 mm slice thickness, 1.5 mm reconstruction interval, and compared to the higher resolution 1 mm datasets which served as the standard for spinal fracture classification. These studies were added to 50 scans of patients without fractures and reviewed by 2 observers. Using the 2.5 mm collimation datasets, the reviewers demonstrated 97-98% sensitivity and 97% specificity for major spinal fractures, and all minor fractures were detected by both observers. A higher percentage of the 1 mm collimation datasets were graded as excellent (80%) compared to the 2.5 mm datasets (68%). Furthermore, the multiplanar reformations from the 1 mm datasets were superior, due to statistically increased image degradation in the 2.5 mm datasets. The authors conclude that the targeted reconstructions from the 2.5 mm datasets are accurate for detection and classification of spine fractures.

Wintermark M, Mouhsine E, Theumann N, Mordasini P, van Melle G, Leyvraz PF and Schnyder P. Thoracolumbar spine fractures in patients who have sustained severe trauma: Depiction with MDCT Radiology 2004; 227: 681-689.

  • Summary: In this study comparing conventional radiography to MDCT of the thoracolumbar spine, retrospective review of the images was performed in 100 consecutive patients, 26 of whom manifested 66 vertebral fractures. Conventional radiography was 33% sensitive, and MDCT 97% sensitive. The 4 slice MDCT technique included thoracic caudocranial CT with 4 x 2.5 mm collimation, and abdominal-pelvic craniocaudal CT with 4 x 2.5 mm collimation. From these datasets, the spine was reconstructed with 2.5 mm thick sections and 2 mm interval, supplemented by sagittal and coronal reformations of the thoraco-lumbar spine.

Alkadhi H, Wildermuth S, Desbiolles L et al. Vascular emergencies of the thorax after blunt and iatrogenic trauma: Multidetector row CT and 3D imaging. Radiographics 2004; 24: 1239-1255.

  • Summary: A 4 slice MDCT scanner was used in this study, and 150 mL of 320 mgI/ml isosomolar contrast infused at 3 mL/second and a bolus tracking timing protocol . The data acquisition protocol is reported as 4 x 1 collimation for arterial phase imaging, using 6 mm feed per rotation, 1 mm reconstruction thickness and a .6 mm reconstruction increment. For venous phase, acquisition was initiated 65-75 seconds after injection, using 4 x 2.5 mm collimation, 15 mm feed per rotation, 3 mm reconstruction thickness and a 2 mm reconstruction interval. This article provides CME credit.

Harris AC, Zwirewich CV, Lyburn ID, Torreggiani WC and Marchinkow LO
CT findings in blunt renal trauma Radiographics 2001; 21: S201-214.

  • Summary: This review article begins with an overview of the classification of renal injuries, followed by a discussion of CT imaging protocol and descriptions of CT findings in specific types of renal trauma. The authors provide the following recommendations: Using 320 mgI/ml contrast medium, 120 to 150 mL is infused at 2-4 mL/second. Scan delays of 60-70 seconds for corticomedullary phase imaging and 3 - 5 minutes for excretory-phase imaging are suggested.