|ECR 2003 / C-1042|
|Ultrasound with SHU 508A in the diagnosis of suspected lower limbs deep vein thrombosis|
Cases classified, as patients with non-enhanced CDU were 26, while classified DVT free were 24. True positive diagnoses were 25 and the true negatives were 20. In 10 patients the result of CDU without Levovist was considered as indeterminate. Among them, 4 patients were finally found to suffer from DVT and 6 were not. The use of contrast medium resulted in 31 positive diagnosis and 26 negative. Among them 30 were true positive and 24 true negative. Exams in which positive or negative diagnosis could not be made with absolute confidence were 3. In these cases ascending venography was the immediate next step, which in one out of these 3 cases revealed DVT. Some cases are seen in figures 1-2,3,4. Sensitivity of CDU was increased from 86,2% without Levovist to 93,8% with it, specificity from 95,2% to 96%, positive predictive value from 96,2% to 96,8%, negative predictive value from 83.3% to 92,3% and accuracy from 90% to 94,7%. The percentage of indeterminate scans was 16,7% with and 5% without Levovist. All the results mentioned above are summarized in tables1,2 and bar chart. Reduction in the number of indeterminate scans from 10 to 3 was the main reason, which caused the slight increase in parameters of diagnostic value detected in our study. The use of contrast medium seems to improve imaging quality in some cases with obesity or soft tissue swelling where non-enhanced color doppler imaging was not satisfactory. In some other cases, in which compression difficulty due to muscle spasm was the reason of false positive diagnosis, improved color doppler image with L-Vist led to correct diagnosis. The anatomic region where color doppler detection of DVT was inadequate included mainly calf veins. The probability that an individual, who tests positive for DVT, actually has the disease, is high without use of SHU 508A. This explains why positive predictive value did not show any significant increase. We should point out several limitations of this study. As far as patients enrolment is concerned, the only inclusion criterion was the existence of clinical symptoms or signs. This fact should not be considered as focusing on a subset of patients or some kind of convenience sampling, but part of our methodology. A second point is that, excluding indeterminate CDUs from calculation of diagnostic value parameters (PDV) can be considered some kind of methodological weakness-bias inherent in our study. For that reason we should notice that parameters with greatest calculation reliability in our study are PPV and NPV, because they are independent from the rate of indeterminate scans. The high initial number of indeterminate scans implies a limitation. This weakness is attributed to technical-equipment reasons, to patient factors, to operator exam dependence and to small number of patients. However the study shows that in cases where limitations lead to high initial rate of questionable results administration of contrast medium can be useful. Considering all the above factors as inevitable and not excluding indeterminate scans from PDV calculations would lead to the maximum increase of sensitivity, specificity and accuracy from about 75% without L-Vist to about 90% with it, showing a very important contribution of the contrast agent. On the other hand, considering that equipment or operator dependent factors, causing diagnostic pitfalls and leading to a high initial percentage of inconclusive scans could be almost completely eliminated, we can calculate the minimum increase in sensitivity, specificity and accuracy referred at the beginning of this section. Finally we should take under consideration that purpose of our study was to evaluate the possibility of confirmation or exclusion of DVT before and after administration of signal enhancing agent and search for possible reduction in the rate of inconclusive scans. In conclusion when veins are not compressible or when a thrombus is evident inside them diagnosis can be made with high confidence without need of a contrast agent, or further diagnostic procedures. On the other hand the probability of no DVT, given a negative CDU result, increases when the exam takes place using a signal-enhancing agent. Some more cases examined with CDU are seen in figures 5-6, 7-8, 9-10.