Statistics from Altmetric.com
As diagnosis of sphincter of Oddi (SO) dysfunction may require perendoscopic manometric assessment of the sphincter, which is an invasive technique carrying a significant risk of pancreatitis, non-invasive scintigraphy has been proposed as an alternative diagnostic method. Craig et al(Gut 2003;52:352–7) reported their experience in assessing SO dysfunction in post-cholecystectomy patients with scintigraphy using the hepatic hilum-duodenum transit time (HHDT) of Cicala and colleagues1 and the scoring system of Sostre and colleagues.2 In the study of Craig et al, scintigraphic data after cholecystokinin octapeptide (CCK-OP) infusion were compared with SO basal pressure, the latter recorded at manometry in the absence of any stimulus.
Craig et al concluded that none of the above scintigraphic variables was sufficiently sensitive to diagnose SO dysfunction identified at SO manometry. Although it is evident that the scintigraphic method of Craig et al used in their study was indeed poorly sensitive, the conclusion of their study cannot be extended to interpret the validity of scintigraphy, as performed in other centres.
It is necessary to identify regions of interest (ROIs) that are repeatable. In the original description of the method it was pointed out that only HHDT performed with the correct ROIs and subtractions was repeatable, compared with other scintigraphic variables related to hepatic uptake and clearance of radiolabelled bile.1 Noticeably in the original HHDT method, ROIs were the heart, right external liver parenchyma, hepatic hilum, and duodenum, whereas in Craig’s study ROIs were the right lobe of the liver and the common bile duct (CBD), which are not sufficient reference sites to construct a transit measurement from the hepatic hilum to the duodenum. In addition, an ROI placed on the CBD cannot correctly discriminate CBD activity from that of the major bile ducts at the hepatic hilum.
Furthermore, analysis of time threshold performed on visual assessment of three minute composite images implies a measurement error of several minutes. In addition, assessment of radiolabelled bile at different sites cannot be performed correctly with a visual method that is significantly delayed in comparison with assessment based on time activity curves (TAC), provided that frame timing is adequate. For example, this methodological error is clearly described and illustrated in figure 3 of the article by Cicala and colleagues,1 with the original description of the HHDT method where the time period assessed with static images at 2.5 minutes apart was delayed by 2.5 minutes in comparison with the assessment based on the TAC constructed on 15 second frame timing. Use of a cholecystokinetic stimulus to assess HHDT is also questionable as CCK is known to affect hepatic bile secretion and SO motor activity, either accelerating the transit of bile under normal conditions or slowing it in the case of SO paradoxical response. Madacsy and colleagues,3 comparing measurement of HHDT without any stimulus and after caerulein administration, showed that the 89% sensitivity of the test without the cholecystokinetic stimulus decreased to 0% after the cholecystokinetic stimulus. In addition, it is not acceptable to derive any conclusions on test sensitivity from a comparison between a scintigraphic assessment performed after a cholecystokinetic stimulus and manometric recordings performed in the absence of a stimulus. Craig et al’s study refers to >9 minutes as an abnormal threshold of HHDT, as indicated in the study of Cicala and colleagues.1 Use of a reference threshold from another centre does not apply when a different technique is used. The technique of Craig et al should be validated with correct reference standards defined in a control group, which was lacking in their study. Several studies have used similar but not comparable scintigraphic techniques to assess SO dysfunction by means of the hepatic duodenum transit time or a score (see table 1).
All but one of the studies in table 1 have compared scintigraphy with manometry and have shown a high specificity of the test and, with the exception of Craig et al, have also shown a satisfactory sensitivity in the absence of a cholecystokinetic stimulus. Reliability of the HHDT test is supported by the following studies:
it discriminates asymptomatic controls from SO dysfunction patients (Corazziari and colleagues5).
Finally, in common with the validation process used for SO manometry, validity of a scintigraphic diagnostic test for SO dysfunction is proved if the tested variable (that is, HHDT or a score) normalises after, and predicts the outcome of, treatment of the abnormality that the test is designed to detect. Both of the above demonstrations for HHDT have been presented in the study by Cicala and colleagues.4
We are aware that information concerning reliability and outcome prediction of HHDT derive from a single group of investigators and confirmation studies with comparable techniques performed in other centres would be welcome. However, we would caution against making comparisons and drawing conclusions with techniques that are not comparable and have not been submitted to proper validation studies to ascertain their reliability.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.