MR Imaging In Soft Tissue Sarcoma:the Predict Value For Histological Grade Of Tumor

Purpose:To observe the magnetic resonance imaging (MR imaging) features of soft tissue sarcoma.(STS) and analyze the significant differences between different histological grade tumors. To identify the MR imaging characteristics that can differentiate high grade from low grade sarcomas.Materials and Methods:The study was approved by institutional review board, and informed consent was waived. The clinical data of patients who referred to Orthopedics or Oncology in January2010to January2013were reviewed. Patients were selected consecutively according to following criteria:patient had pathologic proved soft tissue sarcoma; definite histological grade of sarcoma can be obtained from pathologic results of surgical resection tumor; patient performed MR imaging exam prior to biopsy and any treatment and imaging can be reviewed. The histological grade was graded as grade1-3according to FNCLCC (French Federation of Cancer Centers Sarcoma Group) system. All MR images including non-enhanced and contrast enhanced images were reviewed by two experienced musculoskeletal radiologists in consensus. Firstly, the imaging quality of each MR imaging sequence was evaluated for diagnosis. Tumor tissue layer and tumor size were recorded. Observe MR imaging characteristics of tumors on non-enhanced T1/T2weighted imaging:signal intensity of tumor, signal heterogeneity, clear or not tumor margin, peritumoral high signal on T2weighted imaging, tumor invasion of surrounding structures and so on. Evaluated the level of enhancement of tumor and peritumoral enhancement on enhanced imaging. Grade1was considered as low grade while grade2and3as high grade. χ2test was used to evaluate significant difference of MR imaging features between high grade and low grade sarcomas. The sensitivity and specificity of MR imaging features for differentiating high grade sarcoma were calculated. Combining the MR imaging features with significant difference and the ones had high sensitivity/specificity, multivariate logistic regression model was established, from which the MR imaging features that can predict high grade tumor were recognized.Results:Ninety-five patients (1-95years, mean age55.5years) were chosen in156patients.16patients had grade1soft tissue sarcoma, and34patients had grade2while45had grade3tumor. The quality of all the non-enhanced MR imaging was diagnostic except one T2weighted sequence which was excluded afterwards. High grade sarcoma had significant differences from low grade sarcoma in tumor size (p<.01), tumor margin (p<.01), signal heterogeneity on T2weighted imaging (p<.01), peritumoral high signal on T2weighted imaging (p<.05), peritumoral enhancement (p<.01). Poorly and mixed defined margin had high sensitivity and specificity for differentiated high grade from low grade sarcoma (72%and69%,76%and73%,74%and86%on non-enhanced T1,T2weighted imaging, enhanced T1weighted imaging, respectively).82patients had contrast enhanced imaging and the quality were all diagnostic. Logistic regression model showed that peritumoral enhancement on enhanced imaging was the strongest indicator that can predict high grade soft tissue sarcoma independently (odds ratio13.6).Conclusion:Of the features of soft tissue itself and of MR imaging, tumor size, margin definition, characteristics on T2weighted and contrast enhanced imaging can help to differentiating high grade tumor; peritumoral enhancement can be used to predict high grade tumor independently. The MR imaging features of sarcoma provide clue of high grade sarcoma, as important supplement of biopsy.