Anzeige
Erschienen in:
06.11.2017
Reproducible imaging features of biologically aggressive gastrointestinal stromal tumors of the small bowel
Erschienen in: Abdominal Radiology | Ausgabe 7/2018
Einloggen, um Zugang zu erhaltenAbstract
Purpose
To identify reproducible CT imaging features of small bowel gastrointestinal stromal tumors (GIST) that are associated with biologic aggressiveness.
Methods
Patients with histologically proven small bowel GISTs and CT enterography or abdominopelvic CT were included. Biologic aggressiveness was established based on initial histologic grading (very low risk to malignant), with “malignant” assigned if recurrence or metastases developed subsequently. CT exams were independently evaluated by three gastrointestinal radiologists for tumor size, growth pattern, enhancement, tumor borders, necrosis, calcification, ulceration, multiplicity, internal air or enteric contrast, nodal metastasis, liver metastasis, peritoneal metastasis, ascites, and draining vein size. Inter-observer variability and imaging features associated with high-grade and malignant small bowel GISTs were determined.
Results
Of 78 patients with small bowel GISTs, 10/78 (13%) were high grade and 18/78 (23%) were malignant. There was moderate to substantial inter-observer agreement (Kappa > 0.4) for all findings except tumor border, ulceration, and nodal metastases. Tumor size, irregular or invasive tumor border, necrosis, liver metastasis, ascites, and iso-enhancement were associated with high-grade/malignant small bowel GISTs (p < 0.04). Internal air or enteric contrast and peritoneal metastases additionally predicted malignant behavior (p < 0.03). When imaging features predicting malignant small bowel GISTs were absent and size was ≤ 3 cm, 0% (0/16), 5% (1/19), and 5% (1/17) of patients had high grade, and 0% (0/16, 0/19, and 0/17) had malignant tumors for the three readers, respectively.
Conclusion
Multiple, reproducibly identified, small bowel GIST imaging features suggest biologic aggressiveness. The absence of these imaging features may identify small tumors that can be followed in asymptomatic or high-risk patients.