A Clinicopathological Study of Colorectal Stromal Tumor Based on New Definition

• Main Authors: Jan Yee-Jee, 詹以吉
• Other Authors: Chen Howard Tseng M.D. Ph.D.
• Format: Others
• Language: en_US
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/44354196009628339489

碩士 === 中山醫學大學 === 醫學研究所 === 92 === Background: Stromal tumors of the gastrointestinal tract having C-KIT proto-oncogene mutation, which can be detected by CD117 immunohistochemical staining, were recently defined as gastrointestinal stromal tumors (GISTs). Colorectal stromal tumors were relatively rare when compared to those of stomach and small intestine. Traditionally, they were diagnosed as smooth muscle tumors. We investigated whether those traditionally diagnosed as smooth muscle tumors of colorectum conformed to GISTs in the new definition, and whether there were specific clinicopathological features in predicting the malignant behaviors of colorectal GISTs. Material and Methods: We collected traditionally-diagnosed smooth muscle tumors of colorectum from the department of Surgical Pathology at Taichung Veterans General Hospital from 1982 to 2001. The clinical and demographic data including age, sex, location of tumor, gross appearance (polypoid or intramural), size of tumor (maximal diameter in centimeter), treatment, and outcome of the patients were collected Histological studies such as predominant cell type, cellularity, mitotic count (per ten high power fields), nuclear pleomorphism, tumor hemorrhage and necrosis (unrelated to mucosal ulceration) of hematoxylin and eosin (H&E) stain on paraffin-embedded blocks were performed. Several immunohistochemical stains on paraffin-embedded blocks were performed. The selected immunohistochemical stains including CD117, CD34, smooth muscle actin ( actin M851), desmin, neuron specific enolase (NSE), and S-100 protein were applied to the representative block for each case with proper internal positive control Tumor cells stained as strongly as internal positive control were defined as positive and further recorded as 4 semi-quantified scores based on the percentage of positive tumor cells (0-25%: 1+, 25-50%: 2+, 50-75%: 3+, 75-100%: 4+). The clinicopathological differences between tumors with different CD117 staining, with or without local recurrence, distant metastasis and lethality were analyzed by Pearson’s Chi-square test, Fisher’s exact test, independent T test, and Mann-Whitney U test. Log rank test was done for survival analysis (SPSS for Windows, version 10.1; SPSS Inc., Chicago, IL, USA). P-values less than 0.05 were considered statistically significant. Results: The clinicopathological differences between 25 tumors with different CD117 staining result was analyzed. The P-valves were less than 0.05 for differences in tumor size, mitotic count, location, symptomatic presentation, cellularity, type of surgery, and gross feature. The 7 CD117 negative cases had polypoid gross feature, smaller average size (0.5 cm), fewer mitotic count, lower cellularity, and were less symptomatic. They showed only smooth muscle differentiation immunohistochemically. After polypectomy, no malignant behavior (local recurrence, distant metastasis, and lethality) was found in these cases. The other 18 CD117 positive cases were grossly intramural masses and frequently symptomatic. They were located in rectum, had larger tumor size, higher cellularity, higher mitotic count, necrosis and hemorrhage, and showed diverse smooth muscle and neural differentiation immunohistochemically. They were treated with peranal excision or segmental resection, and 6 of them had malignant behaviors (recurrence, metastasis, and lethality). For those 18 CD 117 postitive 18 cases, the serial biological analyses revealed mitotic count (P=0.019), tumor size (P=0.049), nuclear pleomorphism (p=0.019) and tumor necrosis (p=0.045) were statistically significant factors in lethality. Mitotic count was the only statistically significant factor in survival analysis (P=0.0304). Although not statistically significant, nuclear pleomorphism were associated with local recurrence (P=0.087). Furthermore, tumor necrosis (P=0.071), and tumor hemorrhage (P=0.071) were positively correlated with metastasis. Discussion: Not all traditionally-diagnosed smooth muscle tumors conformed to GISTs. According to the expression of CD117 immunohistochemical stain, collected cases could be divided into two clinicopathologically different categories; intramural GISTs and polypoid submucosal leiomyomas. For colorectal GISTs, mitosis, tumor size, nuclear pleomorphism and tumor necrosis were statistically significant features in predicting the lethality in univariate analysis but mitotic count was the only statistically significant factor in survival analysis. However, more cases are needed for further multivariate analysis or more complex survival analysis in order to refine the true roles of these characteristics in predicting survival.