骨盤内腫瘍悪性度診断への画像解析

概要

論文の詳細を見る
Diagnostic imaging is important in differentiating benign and malignant pelvic tumors and in staging malignant tumors. Many imaging techniques are now available. We describe computed tomographic (CT) and magnetic resonance imaging (MRI) features of gynecologic tumors. The following nine CT parameters were evaluated in 251 cases of cervical cancer (the incidence of each feature is given in parentheses): 1) enlargement of the cervix (58%), 2) low density area(s) (LDA) in the cervical region (28%), 3) presence of a necrotic cavity (11%), 4) pyometra (16%), 5) irregularity or indistinctness of the cervical margin (20%), 6) abnormalities of the parametrium (41%), 7) tumor extension to the vagina (9%), 8) tumor extension to the bladder (20%), 9) lymphadenopathy (8%). The more advanced the stage, the more features tended to be present. On T_2-weighted MRI, cervical cancer appeared as a high intensity image. There was a positive correlation (r=0.79) between MRI and pathologic findings concerning the thickness of the residual cervical myometrium. MRI was distinctly useful in both the staging of cervical cancer and the determination of the extent of tumor invasion of the vagina and bladder. We used three criteria to classify patients with endometrial cancer, which appeared as LDA within the uterus on contrast enhanced CT: 1) LDA occupied less than 50% of the uterine region, 2) the minimum thickness of the normal myometrium was over 0.5cm, 3) the ratio of maximum to minimum thickness of the normal myometrium was over 0.5. Patients who fulfilled all three criteria constituted group A (n=33), and those who failed to meet all three were designated group B (n=30). The rates of myometrial invasion through more than one third the thickness of the uterine wall were 15% in group A and 90% in group B. The rates of lymphatic or vascular invasion were 15% and 57%, respectively, and of extrauterine invasion or metastasis 9% and 47%. Each of these differences was significant (p<0.01). Metastasis was detectable by CT in four group B patients. On T_2-weighted MRI, endometrial cancer exhibited high intensity. A positive correlation (r=0.94) was obtained between MRI data and pathologic findings concerning the thickness of residual normal myometrium. Preoperative differentiation of benign and malignant ovarian tumors is important. Dermoid cysts appeared as low density masses on CT. In the 95 cases that were not dermoid cysts, we used five criteria to distinguish between benign and malignant tumors: 1) tumor shape, 2) tumor size, 3) presence of ascites, 4) enlargement of the contralateral ovary, 5) CT number. With these criteria, the rate of correct diagnosis was 91.7%. Diagnostic accuracy was even greater with MRI because endometrial cysts, which tend to be misdiagnosed on the basis of CT findings alone, appeared as high intensity images on both T_1- and T_2-weighted images. Malignant ovarian tumors (n=89) were staged on the basis of CT findings, according to seven criteria (the rate of accuracy for each criterion is given in parentheses): 1) tumor proliferation outside of the capsule or rupture of the capsule (80%), 2) bilateral ovarian tumors (78%), 3) abdominal ascites (80%), 4) spread to the fallopian tube or/and the uterus (78%), 5) peritoneal metastasis (68%), 6) bowel metastasis (77%), 7) omental metastasis (79%). The correspondence between CT and clinical staging was 64%.
社団法人日本産科婦人科学会の論文
1989-08-01

骨盤内腫瘍悪性度診断への画像解析

スポンサーリンク

概要

著者

関連論文

スポンサーリンク