Application Of Conventional And Diffusion-weighted Magnetic Resonance Imaging For Diagnosis Of Spinal Tuberculosis

Objective:This study proposed to evaluate the clinical application and value of conventional sequences of magnetic resonance imaging combined with diffusion-weighted imaging (DWI) for detection spinal tuberculosis and the differentiation from malignant bone marrow lesions.Materials and methods:From June 2009 to January 2011, consecutive patients with spine pathologies proved clinically and pathologically were examined with conventional MR sequences and diffusion-weighted imaging. Thirty-six patients were diagnosed as tuberculosis and been appointed a case group, the rest were a randomly selected control group of 34 patients with malignant bone marrow diseases. 1, All images of conventional MR sequences were reviewed by two neuroradiologists blinded to clinical data and group. Sensitivity and specificity of each MR imaging features were calculated. 2, The apparent diffusion coefficient (ADC) values of all lesions been measured by two neuroradiologists alonely according the pattern of each lesion. After evaluating the reproducibility and agreement in the measurement of ADC values, using ADC value defined by receiver operating characteristic (ROC) curve analysis as the optimal threshold, the sensitivity and specificity of bone marrow and soft tissue ADC values to diagnose spinal tuberculosis were determined. 3, The values of conventional MR sequences, diffusion–weighted imaging and a combination of both in diagnosing spinal tuberculosis be evaluated by receiver operating characteristic curve analysis. Statistic analyses were completed by using SPSS 17.0. Results:1. The conventional MR imaging features with higher sensitivity and specificity ( > 80%) were endplate disruption (83.3%, 85.3%), disc involvement (80.6%, 82.4%). High sensitivity but low specificity signs in conventional MRI included bone and marrow destruction (91.7%, 50.0%), marrow edema (94.4%, 17.6%). Low sensitivity but high specificity features in conventional MRI were paravertebral thin and smooth abscess wall (63.9%, 94.1%), abscess subligamentous spread to three or more vertebral bodies (11.1%, 100.0%). Low sensitivity and specificity features in conventional MRI were paravertebral soft tissue entity (11.1%, 61.8%),paravertebral soft tissue swell (52.7%, 58.8%),involvement of Multiple Vertebral (41.7%, 55.9%), involvement of neural arch (22.2%, 64.7%). Overall,the sensitivity, specificity and accuracy of conventional MRI for spinal tuberculosis were 85.7%, 88.6% and 87.1% respectively.2. For spinal tuberculosis, the mean ADC values of vertebral destruction, paravertebral soft tissue masse and bone marrow edema were (0.53±0.11)×10^-3mm²/s, (0.66±0.12)×10^-3mm²/s and (0.49±0.08)×10^-3 mm²/s respectively at b-value of 300 s/mm². For malignant bone marrow lesions, the mean ADC values of vertebral destruction, paravertebral soft tissue masse and pathological bone marrow infiltrations were (0.43±0.14)×10^-3mm²/s, (0.39±0.07)×10^-3mm²/s and (0.46±0.16)×10^-3mm²/s respectively at b-value of 300 s/mm². There is significant difference between tuberculous spondylitis and malignant bone marrow lesions of ADC when b is 300 s/mm². The ADC values to differentiate between tuberculous spondylitis and malignant bone marrow lesions have yielded sensitivities and specificities of 88.6% and 74.2% respectively.For spinal tuberculosis, the mean ADC values of vertebral destruction, paravertebral soft tissue masse and bone marrow edema were (0.73±0.12)×10^-3mm²/s, (0.94±0.17)×10^-3mm²/s and (0.67±0.11)×10^-3mm²/s respectively at b-value of 500 s/mm². For malignant bone marrow lesions, the mean ADC values of vertebral destruction, paravertebral soft tissue masse and pathological bone marrow infiltrations were (0.60±0.15)×10^-3mm²/s, (0.53±0.09)×10^-3mm²/s and (0.61±0.12)×10^-3mm²/s respectively at b-value of 500 s/mm². There is significant difference between tuberculous spondylitis and malignant bone marrow lesions of ADC when b is 300 s/mm². The ADC values to differentiate between spinal tuberculosis and malignant bone marrow lesions have yielded sensitivities and specificities of 88.6% and 85.7% respectively.3. The mean ADC of spinal tuberculosis at b-value of 500 s/mm² is higher than the ADC at b-value of 300 s/mm². There is significant difference between the mean ADC of the spinal tuberculosis when b is 300 s/mm² and 500 s/mm². The mean ADC of malignant bone marrow lesions at b-value of 500 s/mm² is higher than the ADC at b-value of 300 s/mm². There is significant difference between the mean ADC of malignant bone marrow lesions when b is 300 s/mm² and 500 s/mm².The sensitivities and specificities using ADC values to differentiate between tuberculous spondylitis and malignant bone marrow lesions are higher at b-value of 500 s/mm², and there is significant difference between the area under the ROC curve (AUC) when b is 300 s/mm² and 500 s/mm².4. Conventional MR sequences combined with diffusion-weighted imagings to differentiate between spinal tuberculosis and malignant bone marrow lesions have yielded sensitivity, specificity and accuracy of 91.4%, 97.1% and 94.3% respectively.Conclusions:The conventional MR imaging is a useful tool for diagnosing spinal tuberculosis and discriminating from other malignant bone marrow lesions. endplate disruption and disc involvement are more reliable features in the diagnosis of tuberculous spondylitis. But conventional MR imaging has some limitations for differentiating some atypical forms of Spinal tuberculosis from malignant bone marrow lesions. In differentiating some atypical forms of Spinal tuberculosis from malignant bone marrow lesions, DWI also has obvious advantages as a sensitive, noninvasive, nonionizing and potential MR technique that is quick to perform and requires no injection of contrast medium. It is important for DW imaging in assessment of the spine. As a result, DWI in this clinical situation can be recommended as a worthwhile additional diagnostic sequence to conventional MR imaging. Combined use of DWI with conventional MRI can increase the diagnostic accuracy of MR Imaging in tuberculous spondylitis. Role of ADC values for the diagnosis of spinal tuberculosis is correlated with the b-value performed in a DW-MR, and should try to use a higher b value possibility.