Evaluation of Prognosis for Malignant Rectal Gastrointestinal Stromal Tumor by Clinical Parameters and Immunohistochemical Staining

Abstract

The aim of this study was to identify associated prognostic factors influencing the outcome of curative resection of rectal gastrointestinal stromal tumor. Diagnostic immunohistochemical staining with CD34, CD117, S-100, desmin, and muscle-specific actin was performed in 46 consecutive patients with previously diagnosed rectal leiomyosarcoma who underwent curative resection from 1979 to 1999. CD44, Bcl-2, P53, and Ki-67 staining were performed on tumors rediagnosed as gastrointestinal stromal tumor for the prognostic evaluation. There were 42 (91.3 percent) patients with rectal gastrointestinal stromal tumor (18 females and 24 males; mean age, 58.4 years). Twenty-nine patients underwent radical surgical resections, such as abdominoperineal resection or low anterior resection, whereas the other 13 patients underwent wide local excision, such as transrectal excision or Kraske's operation. Sixteen tumors were classified as high-grade gastrointestinal stromal tumors, and 26 as low-grade. No tumor had a positive P53 stain. Twenty-seven patients (64.3 percent) developed recurrence or metastasis postoperatively (median follow-up, 52 months). The one-year, two-year, and five-year disease-free survival rates were 90.2 percent, 76.7 percent, and 43.9 percent, respectively. Of these patients with recurrence, subsequent resections in 12 patients with local recurrence, transarterial tumor embolism or STI-571 chemotherapies in 3 patients with distant mestastases were performed. The one-year, two-year, and five-year overall survival rates were 97.4 percent, 94.3 percent, and 83.7 percent, respectively. Bcl-2 (P = 0.007) and histologic grade (P = 0.05) in disease-free survival analysis and age < 50 years (P = 0.03) and tumor size > 5 cm (P = 0.02) in overall survival analysis were independent prognostic factors. The group with wide local excision had a higher local recurrence rate than that of the radical resection group (77 percent vs. 31 percent, P = 0.006), despite smaller tumors (4.5 vs. 7.2 cm, P = 0.05). There was no difference in the incidence of distant metastasis between the two groups. Younger age (< 50 years), higher histologic tumor grade, positive Bcl-2 status, and larger tumors (> 5 cm) were factors associated with significantly poorer prognoses for rectal gastrointestinal stromal tumor. Radical resection was superior to wide local excision in the prevention of local recurrence, but not that of distant metastases.