MRI 信号强度比在脑恶性肿瘤鉴别诊断及 Ki-67 表达的应用价值.

Objective To investigate the application value of conventional MRI signal intensity ratio (SIR) in differential diagnosis of glioblastoma (GBM) and primary central nervous system lymphoma (PCNSL) and the correlation with Ki-67 expression. Methods Retrospective analysis of 72 cases of GBM and 32 cases of PCNSL MRI data, t test were used to analyze signal intensity ratio of tumors, four indexes were included: signal ratio of T1WI (rT1), signal intensity ratio of T2WI (rT2), enhanced signal ratio of T1WI (rT1CE) and signal ratio of T2WI/T1WI (T2/T1). The differential diagnostic performance was evaluated by receiver operating characteristic (ROC) curve. Spearman test was used to analyze the correlation between SIR and (Ki-67 label index, Ki-67 LI). Results The rT1 and rT1CE (rT1: 0.482±0.117, rT1CE: 1.319±0.243) of GBM were significantly lower than that of PCNSL (rT1: 0.557±0.478, rT1CE: 1.594±0.156), while rT2 and T2/T1 (rT2: 1.814±0.285, T2/T1: 3.995±1.189) were significantly higher than that of PCNSL (rT2: 1.435±0.135,T2/T1: 2.600±0.354), the comparison between the two groups was statistically significant (P＜0.001). ROC curve analysis showed that rT1, rT2, rT1CE and T2/T1 could be used to distinguish GBM from PCNSL. A single parameter of T2/T1 had the best diagnostic efficacy (AUC: 0.877), with a sensitivity of 93.8% and a specificity of 76.4%. rT1 +rT1CE +rT2 +T2/T1 combined parameters had the highest differential diagnostic efficiency (AUC: 0.966), with sensitivity of 87.5% and specificity of 91.7%. Correlation analysis showed that rT1 and rT1CE were positively correlated with Ki-67 LI, the correlation coefficients were 0.301 and 0.399, while rT2 and T2/T1 were negatively correlated with Ki-67 LI, and the correlation coefficients were-0.428 and-0.459 respectively. Conclusion Tumor SIR in MRI is important for discriminating GBM from PCNSL, and SIR as a quantitative parameter can provide an important complement to subjective judgment and preliminarily evaluation for the proliferation level of tumor cells. [ABSTRACT FROM AUTHOR]