Hirokazu Ashida1, Takao Igarashi2, Kazuhiko Morikawa2, Kenji Motohashi2, Kunihiko Fukuda2, Naoto Tamai3. 1. Department of Radiology, The Jikei University School of Medicine, 3-25-8, Nishi-Shimbashi, Minato-ku, Tokyo, 105-8461, Japan. ashiyan@jikei.ac.jp. 2. Department of Radiology, The Jikei University School of Medicine, 3-25-8, Nishi-Shimbashi, Minato-ku, Tokyo, 105-8461, Japan. 3. Department of Endoscopy, The Jikei University School of Medicine, 3-25-8, Nishi-Shimbashi, Minato-ku, Tokyo, 105-8461, Japan.
Abstract
PURPOSE: This study aimed to assess the diagnostic performance of unenhanced computed tomography (CT) for distinguishing gastric anisakiasis from non-anisakiasis gastric conditions and the reproducibility of CT findings. METHODS: Fifty-six anisakiasis and 74 non-anisakiasis cases with gastric wall thickening on urgent observation using unenhanced CT were included. Using a κ analysis, two radiologists independently assessed the reproducibility of CT findings, including "circumferential gastric wall thickening," "gastric wall thickening extending more than two segments," "bulky and low-density gastric wall thickening," "increase in peri-gastric fat density," and "ascites." An anisakiasis diagnostic score (ADS) was developed for this study and was used to quantitatively evaluate the CT findings. An anisakiasis diagnostic prediction (ADP) with an appropriate cutoff value was used to further evaluate the ADS. Two radiologists reassessed the findings in consensus to determine the sensitivity, specificity, and accuracy of the CT findings, including the ADP and ADS area under the curve (AUC). RESULTS: Considering reproducibility, a substantial agreement (0.6 < κ < 0.8) was achieved for all findings except "circumferential gastric wall thickening" (κ = 0.499), whereas for diagnostic performance, all findings except ascites were significantly more frequent among the anisakiasis cases. "Bulky and low-density gastric wall thickening" had the highest sensitivity (98%), whereas "gastric wall thickening extending more than two segments" had the highest specificity (80%). The ADP sensitivity, specificity, and accuracy were 91%, 84%, and 87%, respectively. The AUC was 0.902 (p < 0.05). CONCLUSIONS: Unenhanced CT findings are useful for distinguishing anisakiasis from non-anisakiasis gastric conditions with sufficient reproducibility.
PURPOSE: This study aimed to assess the diagnostic performance of unenhanced computed tomography (CT) for distinguishing gastric anisakiasis from non-anisakiasis gastric conditions and the reproducibility of CT findings. METHODS: Fifty-six anisakiasis and 74 non-anisakiasis cases with gastric wall thickening on urgent observation using unenhanced CT were included. Using a κ analysis, two radiologists independently assessed the reproducibility of CT findings, including "circumferential gastric wall thickening," "gastric wall thickening extending more than two segments," "bulky and low-density gastric wall thickening," "increase in peri-gastric fat density," and "ascites." An anisakiasis diagnostic score (ADS) was developed for this study and was used to quantitatively evaluate the CT findings. An anisakiasis diagnostic prediction (ADP) with an appropriate cutoff value was used to further evaluate the ADS. Two radiologists reassessed the findings in consensus to determine the sensitivity, specificity, and accuracy of the CT findings, including the ADP and ADS area under the curve (AUC). RESULTS: Considering reproducibility, a substantial agreement (0.6 < κ < 0.8) was achieved for all findings except "circumferential gastric wall thickening" (κ = 0.499), whereas for diagnostic performance, all findings except ascites were significantly more frequent among the anisakiasis cases. "Bulky and low-density gastric wall thickening" had the highest sensitivity (98%), whereas "gastric wall thickening extending more than two segments" had the highest specificity (80%). The ADP sensitivity, specificity, and accuracy were 91%, 84%, and 87%, respectively. The AUC was 0.902 (p < 0.05). CONCLUSIONS: Unenhanced CT findings are useful for distinguishing anisakiasis from non-anisakiasis gastric conditions with sufficient reproducibility.