Sensitivity and specificity of ultrasonography for the diagnosis of bowel infiltrative disorders (BID) in 35 horses: a retrospective study

AIM OF THE STUDY Chronic wasting syndrome (CWS) is frequently encountered in equine practice and it may be due to different causes, including bowel infiltrative disorders (BID). Histology represents the gold standard procedure for diagnosing BID and it can be performed either by surgical biopsy of the small intestine, rectal biopsy or endoscopic duodenal biopsy1. Although biopsy techniques remain the only gold standard accepted, they are invasive, therefore selecting patients that should be submitted to biopsy procedure with a screening diagnostic test such as ultrasonography is advisable. Little is known about the accuracy of ultrasonography in horses suspicious of being affected with BID: previous studies1,2,3 reported abdominal ultrasonography to be useful in the diagnostic approach for BID in equine patients but, to our knowledge, no studies assessed its accuracy. The aim of the present study was to assess the sensitivity and specificity (ROC curve elaboration) of ultrasonography, particularly of ultrasonographic measurement of the small intestinal wall thickness (USWT) to evaluate BID in horses presented for CWS, using histology of rectal biopsies as the reference standard. MATERIALS AND METHODS 35 horses with a history of unexplained CWS were retrospectively selected according to the availability of one or more USWT and a rectal biopsy histological diagnostic report. Repeatability and reproducibility of USWT in the selected clinical population were evaluated. All the USWT collected were analyzed, calculating an “ultrasonographic average loop thickness” (UALT) for each intestinal loop visualized. Then, all the UALT collected for each patient were used to create an “individual scoring system” (ISS) defining the extension of the tract of the small intestine involved by the abnormal thickening (>3 mm)4. According to the ISS, from the general group of horses (GG, 35 horses) a subgroup of patients showing a diffuse thickening of both duodenum and jejunum was created (SG, 24 horses). Furthermore, for each patient, the mean value of all the UALT calculated was defined as “ultrasonographic average individual thickness”(UAIT) and classified as “positive” (>3 mm) or “negative” (<3 mm)4. Rectal biopsy histological diagnostic report was classified according to Lindbergh et al. (1996)5 as “normal” or “abnormal”. Eventually, sensitivity (Se), specificity (Sp) of UAIT and ROC curve were statistically calculated both for GG and SG, using the rectal biopsy diagnostic report as the reference standard. RESULTS No significant differences were found among measurements made by the same operator (p<0.05) and by different operators (p<0.05) therefore the USWT was considered repeatable and reproducible.The ROC curve for the GG showed that a 5.7 mm small intestinal wall thickness (SWT) cut off value was the best compromise between sensitivity (36.8%) and specificity (87.5%). The ROC curve for the SG showed that a 5 mm SWT cut off value was the best compromise between sensitivity (50%) and specificity (100%). CONCLUSIONS Although repeatability and reproducibility of USWT have been already evaluated in a standardized population of normal horses 6, the results obtained in the present study seem to confirm that the same may be true also in a clinical population of horses presented for CWS. The ROC curve shows that the best compromise between sensitivity and specificity is set at a relatively high cut off value, i.e. 5.7 mm for the GG. This may imply that ultrasonography achieves the best reliability when the small intestinal wall is significantly thickened. However, if the horse population consisted of only patients with ultrasonographic evidence of diffuse thickening of the small intestine (SG), the best cut off value decreases to 5 mm. This may suggest that the presence of a diffuse small intestinal thickening increases the reliability of the measurements. According to the high specificity of ultrasonography, patients with unexplained CWS that present a UAIT greater than 5.7 mm should be submitted a for further investigations (such as biopsy, oral glucose tolerance test, endoscopy etc.), since they are likely to be affected with BID. According to the low-moderate sensitivity of ultrasonography, in patients with unexplained CWS that present a UAIT less or equal to 5.7 mm, the hypothesis of the presence of an infiltrative disorder should not be ruled out and they should be carefully monitored by periodic ultrasonography for possible progressive thickening. In conclusion, ultrasonography appears to be a valuable tool for the screening diagnostic procedure of BID in patients with CWS. BIBLIOGRAPHY 1. Mair TS, Pearson GR, Divers TJ (2006) Equine Vet. Ed. 18(6):299-308 2. Roberts MC (2000) Equine Vet. Ed. 12(4):214-219 3. Kaikkonen R, Niinisto K, Sykes B, Anttila M, Sankari S, Raekallio M (2014) Acta Veterinaria Scandinavica 56:35 4. Reef VB (Ed.1998) In: Equine Diagnostic Ultrasound, WB Saunders Co, Philadelphia. pp 273-363 5. Lindbergh R., Nygren A., Persson S.G.B. (1996) Equine Vet. J. 28 (4) 275-284 6. Bithell SJ, Habershon-Butcher JL, Bowen IM, Hallowell GD (2010) Vet. Rad. & Ultr. 51 (6): 647-651