Soler R, Castro J M, Rodríguez E
Department of Radiology, Hospital Juan Canalejo, La Coruña, Spain.
Comput Med Imaging Graph. 1996 May-Jun;20(3):163-9. doi: 10.1016/0895-6111(96)00049-3.
To evaluate the value of each MRI findings in differentiating the nature of soft tissue lesions.
We performed a blind retrospective review of MR imaging in 65 consecutive soft-tissue lesions (cystic lesions in or around the joints, and the soft tissue abnormalities directly related to a known trauma were excluded). Morphology and signal intensity characteristics were analyzed. Each lesion was considered as benign tumor malignant tumor, benign but locally aggressive lesion, and undetermined. The final diagnosis was established pathologically (n = 45), and by the association of other imaging studies, clinical findings and follow-up over 2 yr (n = 20).
Involvement of one compartment was equal in malignant (57.1%) and benign (56.2%) tumors, whereas multicompartmental involvement was most common in non-tumoral lesions (39.1%) (P < 0.05). Well-defined margins were most frequent in benign tumors (189.2%) (P < 0.05). The change from homogeneous to heterogeneous pattern on T1- and T2-weighted sequences as a predictor of malignancy showed a sensitivity of 77.7% and a specificity of 20%. Based on the MRI findings a diagnosis of benign lesion was established with a sensitivity of 60.7% and a specificity of 77.7%; malignant tumor with a sensitivity of 78.5% and a specificity of 96% and benign locally aggressive lesions with 54.5% and 88.1%, respectively. With the addition of the clinical data, an etiologic diagnosis was performed in 78.5% benign tumors, in 85.7% malignant tumors, and in 95.6% benign non-tumoral lesions. Undetermined lesion was diagnosed in eight masses.
Soft tissue lesions can be diagnosed with certainty in many benign tumors based on the integrated evaluation of morphology and signal intensity MR findings. The association of MR and clinical data allowed us to identify benign but locally aggressive lesions, mostly related to infection. MR findings are highly specific for malignant tumor, although a histologic diagnosis cannot be performed based only on image analysis. When a mass is undetermined and no criteria for benignity or malignancy can be established, pathologic analysis should be always performed.
评估各MRI表现对鉴别软组织病变性质的价值。
我们对65例连续的软组织病变的MR成像进行了盲法回顾性研究（排除关节内或关节周围的囊性病变以及与已知外伤直接相关的软组织异常）。分析了形态学和信号强度特征。每个病变被视为良性肿瘤、恶性肿瘤、良性但局部侵袭性病变以及性质未明的病变。最终诊断通过病理确定（n = 45），以及结合其他影像学检查、临床表现和2年以上的随访确定（n = 20）。
单区域受累在恶性肿瘤（57.1%）和良性肿瘤（56.2%）中相当，而多区域受累在非肿瘤性病变中最为常见（39.1%）（P < 0.05）。边界清晰在良性肿瘤中最为常见（189.2%）（P < 0.05）。在T1加权和T2加权序列上从均匀模式变为不均匀模式作为恶性肿瘤的预测指标，其敏感性为77.7%，特异性为20%。基于MRI表现，诊断良性病变的敏感性为60.7%，特异性为77.7%；诊断恶性肿瘤的敏感性为78.5%，特异性为96%；诊断良性局部侵袭性病变的敏感性分别为54.5%和88.1%。加上临床资料后，78.5%的良性肿瘤、85.7%的恶性肿瘤以及95.6%的良性非肿瘤性病变做出了病因诊断。8个肿块性质未明。
基于形态学和信号强度的MR表现综合评估，许多良性肿瘤中的软组织病变可明确诊断。MR与临床资料相结合使我们能够识别良性但局部侵袭性的病变，大多与感染有关。MR表现对恶性肿瘤具有高度特异性，尽管不能仅基于图像分析进行组织学诊断。当肿块性质未明且无法确立良性或恶性标准时，应始终进行病理分析。
