佐々木 賀広¹, 栗田 武彰², 棟方 昭博¹, 羽田 隆吉³, 遠藤 正章³, 加藤 智²

Yoshihiro SASAKI¹, Takeaki KURITA², Akihiro MUNAKATA¹, Ryukichi HADA³, Masaaki ENDOH³, Satoru KATOH²

¹弘前大学第1内科, ²弘前大学第2外科, ³木造成人病センター外科

¹First Department of Internal Medicine Hirosaki University School of Medicine, ²Second Department of Surgery Hirosaki University School of Medicine, ³Department of Surgery, Kizukuri Adult Disease Center Hospital

キーワード : Thyroid nodules, Brightness analysis, Determination of the nodule border, Clarity of the nodule border

Using an image processor, we defined the border of a thyroid nodule on an ultrasonic image as a set of points, where the local brightness gradient (隆B) indicates a maximal value. The determined border line appeared to be compatible with the visual one. Furthermore, we introduced a clarity index (CI), defined as the ratio of the number of points indicating a high 隆B value to the number of points indicating a low 隆B value, in order to quantify border clarity. The CI exhibited a value of more than 5, 0.5 to 2, and 2.58 in thyroid nodules presenting a cystic pattern, solid pattern, and mixed pattern, respectively. The border determination process, which we call brightness analysis, can be extensively applied to estimate nodular lesions in other organs.