INTERPRETING ABDOMINAL CTs ON A PC BASED WORKSTATION. HOW ACCURATE IS IT?

Maria Tzalonikou,M.D., William Bennett, M.D., Thomas Vrachliotis, M.D., Dimitrios Spigos,M.D.

Radiology Department, Ohio State University Hospitals,
Columbus, Ohio, U.S.A.

Purpose:

We evaluated the accuracy of a PC versus the 2048Kx2048K workstation in the interpretation of abdominal CTs in our totally filmless, digital Emergency Department.

Materials and Methods:

We reviewed 198 abdominal CTs performed during a two-month period of time, in a CT scanner installed in the Emergency Department of our hospital. The images were stored on a Dicom compatible PACS system.The images were sent to the PC by network. Images were displayed using software provided by the PACS vendor. All these CTs were viewed with a 3x2 display mode. The findings from that reading were compared to those from the 2048Kx2048K workstation, which was used as reference.

Results:

In 49 out of the 198 abdominal CTs, there were negative reports. In the remainder of the cases, there were positive findings.   Of the 149 positive cases, 30 had benign or malignant masses or other liver disease, 22 had lesions related to the gallbladder or the biliary tree or were status post cholecystectomy, 2 had pancreatic lesions, 38 had renal or ureter or bladder calculi, 4 had urinary tract calcifications and 26 had accompanying lesions, such as dilatation of the pelvis or the ureter. Eleven cases presented splenic lesions, while 20 had pelvic pathology. In 25 cases there were lesions related to the duodenum, the small or the large bowel, such as complete or partial obstruction, bowel wall thickening or inflammatory disease of the bowel. Six cases had findings consistent with appendicitis. Seven cases presented with abdominal hemorrhage, 7 had ascitic fluid and 10 had benign or malignant bone lesions, mainly fractures or metastases.Twenty-nine cases presented chest findings and in 48 cases other findings such as hernias or vascular lesions were identified.

Out of these positive cases, the diagnosis was not made in 3 cases from the PC. One was a cases of fatty replacement of the liver and two were cases of haziness of the mesenteric fat. This results in a missed-case rate of 2.01% of the positive cases.

Conclusion:

We found that the images of CT scans of the abdomen viewed on a 1024x768 monitor were of sufficient quality to render an accurate diagnosis. A 2.01% error rate was considered acceptable.

Corresponding author:

Maria Tzalonikou,M.D., W
Radiology Department, Ç
Ohio State University Hospitals,
Columbus, Ohio, U.S.A.

Oral presentation at EuroPACS'98, Barcelona, Spain