Our hypothesis is that a dedicated automated processing server can be used to automatically generate multiplanar reformations of trauma-related CT examinations of the cervical spine, chest, abdomen, and pelvis and send these reformations to PACS significantly reducing the time delay between when the study is performed and when the reformations are made available on PACS to be interpreted by a radiologist and viewed by clinical staff.

There are numerous studies detailing the clinical benefits of the routine use of coronal and/or sagittal reformations in CT examinations. In trauma-related CT examinations of the cervical, thoracic, and lumbar spine at our institution, the technologist performs coronal and sagittal reformations at the scanner console and sends these reformations to PACS for interpretation by a radiologist and clinical viewing. Although certain PACS systems, as well as thin client 3D systems, are capable of performing these reformations on the fly, many PACS systems, including the one at our institution, does not support this functionality. Certain CT scanners are also capable of automatically generating reformatted images. However, this functionality is also not widely available.

CT examinations are completed only when both the axial and multiplanar reformatted images are available for viewing. Although generating reformations usually takes the technologist less than 2 minutes, there can be substantial time delays in generating reformations because technologists are responsible for patient safety and throughput, which frequently take priority over creating reformatted images. In addition, enterprise distribution of reformatted images to our trauma colleagues is simplified when they are available as separate series on the PACS.

We have recently integrated a dedicated server that is capable of automatically generating multiplanar reformations from source axial images and sending these reformations to PACS without requiring technologist input. This dedicated server is currently used to generate coronal and sagittal reformations for trauma-related studies in parallel with technologist-generated reformations. The purpose of this study was to evaluate the average time delay attributable to the technologist generating coronal and sagittal reformations, then compare that time delay to the time required for the dedicated server to automatically generate the reformations.

The PACS database was queried using the exam codes for cervical, thoracic and lumbar spine reformations to generate a list of appropriate studies. The “acquisition time” associated with each study represents the time the axial acquisition was completed. The “study time” for a specific series (coronal or sagittal) represents the time the reformation was completed. These times were confirmed by manual observation of the technicians and comparing times recorded from the scanner to the times provided in the DICOM headers. To evaluate the time delay for the technologist to generate the reformations, the “acquisition time” for the axial images was recorded and compared to the “study time” for the coronal and sagittal reformations. To evaluate the time required for the dedicated server to automatically generate reformations for cervical spine studies, we reviewed the server log to determine when the reformations were completed and sent to the PACS server.

The average time delay attributable to the technologist for generating coronal and sagittal reformations of the cervical spine was 15 minutes and ranged from 3 to 57 minutes. The average time required for the dedicated server to automatically generate coronal and sagittal reformations for the cervical spine studies was 1 minute, with a range of 1 to 2 minutes. The difference was statistically significant with a p < 0.0001. The average time delay for sending the axial images to the dedicated server was 3 minutes, with a range of 1 to 10 minutes.

The average time delay attributable to the technologist for generating coronal and sagittal reformations of the thoracic and lumbar spines was 33 minutes and ranged from 8 to 127 minutes. Again, the average time required for the dedicated server to automatically generate coronal and sagittal reformations for the thoracic and lumbar spine studies was 1 minute, with a range of 1 to 2 minutes. The difference was statistically significant, with a p < 0.0001.

When comparing the two methods for generating reformations in trauma-related CT examinations of the cervical spine, using the dedicated server to automatically generate reformations resulted in substantial time savings when compared to technician-generated reformations. Even when considering the additional time required to send the axial images from PACS to the dedicated server, the time savings was 11 minutes per study, for an average of 10 studies per day. Given the significant average time delay of 33 minutes for technician-generated reformations in trauma-related thoracic and/or lumbar CT examinations, we see an even greater time savings when using the dedicated server. One important difference between the two methods for generating coronal and sagittal reformations is that the technologist adjusts the plane of reformation to correct for patient obliquity on the axial images. It is unclear if these adjustments significantly affect the quality of the reformations, and will be the subject of future studies.