Comparative evaluation of multidetector CT and MR imaging in the differentiation of adnexal masses: authors’ reply

Abstract

Sir,We thank the authors of this letter fortheir thoughtful remarks.We agree that the performance ofunenhanced CT is necessary for thefull characterization of adnexalmasses, especially in ‘differentiatingenhancing lesion components fromhaemorrhage or mucinous content’.This could probably enable the correctcharacterization of one of the false-negative cases with multidetector CT(MDCT) in our study, by detecting theenhancing element within a borderlineovarian tumour with haemorrhagiccontent [1]. This study was one of thefirst conducted on our 16-row MDCTsystem, and unenhanced CT was notperformed due to concern about irra-diation. The effective dose for the CTprotocol used in this study was mea-sured at 7 mSv, i.e. not different fromthat for conventional CT of the abdo-men [2]. The protocol we use nowincludes unenhanced CTof the pelvis,covering the area from the iliac creststo the symphysis pubis, or the adnexalmass if larger, followed by enhancedCT of the abdomen and pelvis.RegardingtheareaofcoveragewithMR imaging this included the pelvis‘covering the area from the iliac creststo the symphysis pubis, or the ovarianmass, if larger’, as already mentionedin our paper in the ‘Material andmethods’ section [1].With regards to the statisticalanalysis,theROCcurvewasproducedusing the SPSS/PC (version 15.0)statistical software. ROC curves foreach diagnostic method were basedon 67 patients, who underwentmultidetector CT, MR imaging andhistopathological examination, whichwas used as gold standard. TheSPSS/PC (version 15.0) did notestimate statistical significancebetween two ROC curves. It justestimated the areas under the curve(0.921 and 0.982 for multidetector CTand MR imaging, respectively). Wejust compared these percentages, andthe difference was not statisticallysignificant for