Correlation Between Computed Tomography Findings And Risk Assessment Of Gastric Stromal Tumors

Objective: 1. To investigate the correlation between computed tomography(CT) findings and risk assessment of gastric stromal tumors(GST). 2. To investigate the expressions and clinical significances of immunohistochemistry in gastric stromal tumors(GST).Methods: 1. A retrospective analysis was conducted on 106 patients with GST which confirmed by CT and pathological diagnosis. The age, gender, location of tumor, size of tumor, morphology, growth pattern, boundary, density, enhancement and distant metastasis of the tumors were analyzed. Grading of GST was according to the 2008 United States National Institutes of Health(NIH) standard, which classified GST as low risk group, medium risk group and high risk group. All datas were analyzed using SPSS 20.0. Single variable analysis and multivariate stepwise Logistic regression were employed to reveal independent variable X(various CT characteristics) in predicting the dependent variable Y(low risk group, medium risk group and high risk group benign). First of all, the single variable analysis was performed and the independent variables which had the statistical significance entered the further multivariate Logistic regression analysis. P<0.05 was regarded statistically significant. 2. A retrospective analysis was conducted on HE staining sections and Envision two-step immunohistochemical sections in 81 cases of GST which confirmed by pathological diagnosis. The age, gender, location of tumor, size of tumor, mitotic figure and immunohistochemical indexs(CD117, CD34, DOG-1, Desmin, Ki-67, S-100 and SMA) of the tumors were analyzed. Grading of GST was according to the 2008 United States National Institutes of Health(NIH) standard, which classified GST as lowest risk group, low risk group, medium risk group and high risk group. All datas were analyzed using SPSS 20.0, with χ2 univariate analysis. P<0.05 was regarded statistically significant.Results: 1. In included 106 cases, the patients were distributed in 3 groups, 46 in low-risk group, 27 in medium-risk group and 33 in high-risk group. Of those, aged less than 50 years old in 21 cases, 51~70 years old 66 cases, 19 cases of over 70 years old. Male 43 cases, female 63 cases. Occurred in gastric fundus of 40 cases, 52 cases of gastric body, 14 cases of gastric antrum. There were 52 cases smaller than 5cm, 5~10cm 43 cases and 11 cases larger than 5cm. 69 cases of morphological rules, 37 cases of irregular shape, lobulated change. Intraluminal growth in 54 cases, extraluminal growth in 30 cases, mixed growth in 22 cases. The boundary is clear in 83 cases, and the boundary is fuzzy in 23 cases. There were 44 cases with uniform density, and 62 cases with uneven density. There were 33 cases with mild enhancement, 22 cases with moderate enhancement and 51 cases with severe enhancement. Distant metastasis occurred in 8 cases, 6 cases of peritoneal metastasis and 2 cases of lymph node metastasis. Among mentioned three GST groups, done by univariate analysis, size of tumor, morphology, growth pattern, boundary, density, enhancement and distant metastasis were different, the differences were statistical significance(P<0.05), however, age, gender and location of tumor were not statistically significant(P>0.05). With the use of multinomial logistic regression analysis, size of tumor, morphology and growth pattern had significant difference(P<0.05). 2. In included 81 cases, the positive rates of CD117, CD34, DOG-1, Desmin, Ki-67, S-100 and SMA were 98.8%, 98.7%, 91.5%, 21.7%, 40.3%, 25.0% and 31.4% respectively. The expression of CD117, CD34, DOG-1, Desmin, S-100 and SMA were not associated with age, gender, location of tumor, size of tumor, mitotic figure and malignancy potential grading(P>0.05). The expression of Ki-67 was not associated with age, gender and location of tumor(P>0.05). But the expression of Ki-67 was associated with size of tumor, mitotic figure and malignancy potential grading(P<0.05).Conclusion: 1. Computed tomography(CT) as a fast, effective method of examination, in size of tumor, morphology, growth pattern, boundary, density, enhancement and distant metastasis, can be used for risk assessment of GST. 2. Size of tumor, morphology and growth pattern can be used to predict the malignant degree of GST, thus, these indicators are useful for clinical diagnosis, preoperative evaluation, treatment plan and prognosis evaluation. 3. The combined detection of CD117, CD34 and DOG-1 has important value for the diagnosis of GST. 4. Desmin, S-100 and SMA can be used in the differential diagnosis of GST. 5. The expression of Ki-67 in combination with size of tumor, mitotic figure and malignancy potential grading, can be used to evaluate the prognosis of GST.