Primary gastrointestinal sarcomas: analysis of prognostic variables.

BACKGROUND: Primary gastrointestinal sarcomas are uncommon, and the clinicopathological determinants of survival remain unclear. In order to correlate clinical presentation, pathological assessment, and treatment with outcome, we have analyzed our institution's recent experience with these tumors.
METHODS: Records of adult patients admitted to our institution between July 1982 and December 1991 were reviewed.
RESULTS: During this period, 38 adult patients (> 16 years of age) were admitted to our institution with a primary gastrointestinal sarcoma. They accounted for 2% of all adult sarcoma admissions during that period. The study population was composed of 26 men and 12 women. Ages ranged from 29 to 82 years (mean 59). Disease was localized to the primary site in 30 patients (81%). The stomach was the most frequent site of disease (20 cases). The small bowel was affected in nine cases (five duodenum, four jejunum) and the large bowel in nine cases (two colon, seven rectum). Ninety-two percent of patients were symptomatic at presentation. A complete resection was performed in 27 cases, incomplete resection in seven cases, and biopsy only in the remaining three patients. Nine patients received doxorubicin-based chemotherapy. Leiomyosarcoma (n = 35) was the predominant histological diagnosis. Twenty-six tumors were classified as high grade (68%) and 12 as low grade (32%). Overall actuarial 5-year survival was 28% (median follow-up 26 months). Weight loss (p = 0.02) and pain at presentation (p = 0.05) were adverse prognostic factors. Histological grade (p = 0.0002), completeness/extent of surgical resection (p = 0.005), or small bowel primary site were significant determinants of overall survival. The resection of contiguous organs did not affect survival if the primary tumor was completely excised (p = 0.422). Age, race, sex, presentation (prior surgery), tumor size, or adjuvant therapy were not significant prognostic factors. Recurrence was noted in 44% after complete resection, and mean time to recurrence was 9 months (median 7, range < 1-37). Hepatic metastases (42%) and local recurrence (42%) were the predominant sites of initial failure. For patients with a complete resection, grade was the major prognostic determinant (5-year survival: high grade/complete resection 18% vs. low grade/complete resection 72%, p = 0.002).
CONCLUSION: The prognosis of gastrointestinal sarcomas is poor. Complete surgical excision is the optimal therapy. However, our results suggest that surgery alone is inadequate for high-grade tumors. We believe that these patients should be considered candidates for investigational adjuvant therapies.