EUS could detect ascites missed by CT scan

I read with great interest the article by Lee and colleagues (Gut 2005;54:1541–5) regarding the accuracy of endoscopic ultrasonography (EUS) in diagnosing ascites and predicting peritoneal metastases in patients with gastric cancer. I would like to thank the authors for quoting our study.¹ Lee and colleagues commented that the sensitivity of detection of ascites was lower in our study and that this might be due to the use of catheter probe. I would however like to point out that such a comparison was unfair as the two studies were fundamentally different in two ways.

(1) The patient populations of the two studies were different. Our study excluded all patients with evidence of ascites on physical examination or computed tomography (CT) scan. We believe that there is no need for an additional EUS to confirm the presence of ascites in such patients. Moreover, such patients should have paracentesis for cytological examination rather than a locoregional staging investigation like EUS.

During the study period of our paper (September 1995 to January 2002), 89 patients had evidence of ascites on CT scan or physical examination. There would be no difficulty in detecting ascites in these patients by EUS. If we included these patients, the detection rate by EUS would have been 25.5% (36+89/402+89 = 125/491). The overall incidence of ascites was 29.5% (56+89/402+89 = 145/491). The “adjusted” sensitivity would be 84.8% (34+89/56+89 = 123/145), not much lower than the 87.1% reported by Lee et al.

(2) CT scan was performed for all patients in our study. Of the 89 patients who were excluded from the study, 69 had evidence of ascites on CT scan. The sensitivity of CT scan for detection of ascites was 47.6% (69/145), higher than the 16.1% sensitivity (combined US and CT scan) reported by Lee et al. We suspect the lower sensitivity may be due to the predominant use of US scan (231 patients) rather than CT scan (99 patients) in their study. A direct comparison however is not possible as the study populations of the two studies may be different.

In summary, I would like to applaud Lee et al for their systematic study. The comment that the sensitivity of detection of ascites was lower in our study was unfair.