Clinical Study Of Whole Body Diffusion Weighted MR Imaging In Malignant Tumors

Background and aim:Malignanct tumor now is presenting a elevating incidence worldwide. And in our nation, the tumors of top ten highest incidence include lung cancer, gastric carcinoma, HCC, colorectal carcinoma, esophageal carcinoma, bladder carcinoma, pancreatic carcinoma, luekemia and brain tumor for man and breast cancer, lung cancer, colorectal carcinoma, gastric caicinoma, HCC, ovary cancer, pancreatic carcinoma, esophageal carcinoma, endometrial carcinoma and brain tumor. The common tumor of the highest mortality is lung cancer.Imaging diagnosis is playing more and more important role in the diagnosis, treatment and follow-up procedure for patients with malignant tumor. Imaging modalities, include roentgenography, computed tomography, magnetic resonance imaging, ultrasonography, SPECT and PET.MR has become one of important oncological imaging modality due to its good tissu resolution, large coverage and multi -planar scan. Diffusion weighted imaging (DWI) is one of the functional MR imaging, which enables not only size and morphology assessment of tumor but also functional evaluation at a cellular level of lesions. The earlest application of DWI is in the CNS imaging, especially in the early detection of acute ischemic stroke. In recent years, the application of DWI has expanded to body imaging, such as imaging of the abdominal and pelvid organs, especially tumors. Whole-body diffusion weighted imaging with background body signal suppression (WB-DWIBS) is another new MR imaging method newly invented and applied, which overcomes the traditional limitation of DWI in body imaging such as breathhold scan and limited scan coverage. It enables large coverage scan, even whole body scan, while free breathing. STIR is employed in this technique to better suppress the signal of normal organ, favoring a better view of pathologic tissues.A large number of oncologic patients was enrolled in this study to underwent WB-DWIBS to assess its ablity to detect the metastatic lesion for patients with primary tumors and the primary tumor for patients with metastatic tumors and unknown origin and to evaluate its value in evaluating the cell types of lung cancer and in detecting the hilar and/or mediastinal lymph node metastasis and further in differential diagnosis with benign lymph nodes which may also be visualized on diffusion weighted images.1. Purpose:(1) to summarize the characteristics of WB-DWIBS images of healthy volunteers, the ADC, EADC and SI value ranges of normal organs and quantitatively evaluate the influence of different b values on the WB-DWIBS imaging.(2) to assess the ability of WB-DWIBS in detecting metastatic lesions, the consistency in bone metastasis detection with SPECT.(3) to assess the clinical utility of WB-DWIBS in detecting primary tumors for patients with metastatic tumor and unknown origin.(4) to assess the ability of WB-DWIBS in identifying different cell types and metastatic hilar or/and mediastinal lymph node metastassi of lung cancer.2. Material and methods:(1) Study objects①From May 2008 to February 2009,60 cases of healthy volunteers were enrolled, including 38 males and 22 females. The age range was 30-70 years, with the average age of 44.6±6.9 years.②From May 2009 to October 2009,66 patients, including 44 males and 22 females, who were diagnosed as primary malignant tumor. The age range was 21-79 years, with the average age of 56.2±12.9 years.③From February 2009 to June 2009,29 patients, including 19 males and 10 females, who were diagnosed as metatatic tumor of unknown origin. The age range was 30-79 years, with the average age of 57.1±11.1 years.④From March 2009 to February 2010,34 patients, including 26 males and 8 females, who were diagnosed as lung mass. The age range was 43-78 years, with the average age of 59.0±9.3 years.(2) MR scanThe extending board was used and the study objects were asked to take a supine and feet-first position and keep quiet breath. The scan range was from head to knees. For section 1, two groups of b values (0s/mm2,600s/mm2, and 0s/mm2,1000 s/mm2) were used, while for the following sections, one group of b values(0s/mm2,600s/mm2) were used.(3) Image post-processingWB-DWIBS images were obtained with a multi-stack scanning, and the whole body images were obtained with 'Mobiview' function. ADC map and EADC map were obtained by using 'diffusion' function.(4) Image analysis①By drawing and copying ROIs, the ADC value, EADC value and SI value was measured.②Direct visual assessment of the metastatic tumors was used. The number of different types of metastatic tumors was recorded. The number of bone metastasis was compared with the finding of SPECT to assess the coherence between the two modalities. The characteristics of the ADC value, EADC value and SI value of the bone metastasis and liver metastasis was analysed and compared with common benign lesions.③Direct visual assessment was used to detect the primary tumors on the basis of the SI of lesions according to a 5 point confidence scale.④The ADC value, EADC value and SI value of lung masses and lymph nodes were measured and compared with the pathologic findings⑤Statistical analysisThe data was calculated by using the statistical software SPSS11.0. The results of measured values were recorded in the format of 'mean±standard deviation'. Two-tailed student's T test or non-parametric test were used to compared means of different groups according to whether the data was in accordance with normal distribution. The coherence between two modalities in identifying lesions was evaluated by usingxanalysis. The correlation and regression analysis were performed and the regression equation was established. The significance of the coefficient of regression was verified by using t test. ROC analysis was used to analyse the diagnostic value of this technique in detecting primary tumors and in detecting lymph node metastasis for lung cancer patient. The optimal cut-off point of ADC, EADC and SI values, and the correspongding sensitivity and specificity of the three parameters in diagnosing was determined. 3. Results:(1) The signal of most of the solid organs were attenuated on WB-DWIBS images, except for spleen, gallbladder, testis, seminal vesicle, prostate and the cerebrospinal fluid in the spinal canal due to T2 shine-through effect. Vessles in the inguinal, axillary and abdominal regions may be visualized as high signal, which can be differenciated with pathologic lesions by its symmetrical distribution and beam like appearance. The value of ADC, EADC, SI and SNR of the group with b value of 600s/mm2 was greater than that of the group with b value of 1000s/mm2, and for most organs (or parts) of body, the difference was significant. While the background noise intensity of the former group was weaker smaller than the latter group.(2) There was good coherence between WB-DWIBS and SPECT in detection of bone metastasis(κ=0.705). No statistical difference was found in the measured diameter of metastatic lymph nodes between WB-DWIBS and CT/MR. The ADC, EADC and SI value of bone metastasis were statistically different with those of normal lumbar vertebrea. No statistical difference was found in ADC and EADC value between metastatic lymph nodes and muscle (Z=-1.31,P=0.19; Z=-1.503, P=0.133) and between metastatic lymph node and liver(Z=-1.844, P=0.065; Z=-7.80, P=0.435). There was statistical difference in SI value between metastatic lymph node and muscle and between metastatic lymph node and liver(Z=-7.756, P=0.000 and Z=-7.662, P=0.000).The SI value of metastatic lymph nodes was significantly higher than that of muscle and liver. There was statistical difference in the ADC, EADC and SI value between hepatic metastasis and hepatic cyst(Z=-4.797,P=0.000; Z=-4.798, P=0.000 and Z=-3.094,P=0.001). There was statistical difference in the ADC and EADC value between hepatic metastasis and hepatic hemangioma(Z=-3.178,P=0.001 and Z=-2.963,P=0.002), but no statistical difference in the SI value(Z=-0.593, P=0.581). There was statistical difference in the ADC, EADC and SI value between hepatic cyst and hepatic hemangioma(Z=-3.092, P=0.001; Z=-3.377, P=0.000和Z=-2.718,P=0.005). Compared with normal liver, liver metastasis has statistical difference in ADC value and SI value (Z=-2.836,P=0.005 and Z=-5.403, P=0.000) and no difference in EADC value (Z=-1.669, P=0.095), liver cyst has statistical difference in ADC value, EADC value and SI value (Z=-5.012,P=0.000; Z=-5.013, P=0.000 and Z=-3.077, P=0.002), and liver hemangioma has statistical difference in ADC value, EADC value and SI value (Z=-2.844, P=0.004; Z=-2.590, P=0.010和 Z=-3.821,P=0.000)(3) For 82.8%(24/29) of the patients, primary malignancies were correctly identified with this whole-body DW MR imaging, and subsequently confirmed by biopsy and/or pathology. While for 17.2%(5/29) of patients, no definite primary malignancy was identified. ROC analysis yielded Az values of 0.907,0.914, and 0.910 for the three reviewers respectively. The sensitivity and specificity was 82.8% (24/29) and 100%(25/25) for the three viewers. The inter-observer agreement between each pair of reviewer was 0.935,0.804, and 0.873, respectively.(4) There was statistical difference in ADC, EADC and SI value between benign and malignant lung lesions(t=-9.932, P=0.000; t=6.567, P=0.000 and t=5.383, P=0.000). No statistical difference was found in ADC, EADC and SI value between each pair of the three cell types of squamous cell carcinoma, adenocarcinoma and small cell lung carcinoma. There was no statistical difference in the diameter of lung cancer measured by WB-DWIBS and CT/MR(t=-0.020, P=0.984). There was a significant correlation between them(r=0.984, P=0.000). WB-DWIBS vusualized 89.6%(26/29) of the pathologically proven metastatic lymph node. The metastatic lymph nodes had significantly lower ADC value(Z=-4.420, P=0.000), higher EADC(Z=-4.287,P=0.000) value and SI value(Z=-3.360,P=0.000) than normal lymph node. ROC analysis showed that the optimal cut-off point of ADC, EADC and SI value in diagnosing metastatic lymph node was 2.42×10-3mm2/s,0.245 and 227.5, respectively.4. Conclusions:(1) WB-DWIBS proved to yield images of good suppression of background signal. The ADC, EADC, SI values of organs and the image SNR was affect by selection of different b values.(2) WB-DWIBS proved to be able to detect metastatic lesions in different part of body. It showed good coherence with SPECT in detecting bone metastasis and proved to be as accurate in metastatic lymph node diameter measurement as CT and MR.(3) For patients originally presented with metastases, WB-DWIBS imaging might have potential utility in detecting primary malignancies.(4) WB-DWIBS proved to be able to differenciate between benign and malignant lesions in lung, but had a limited value in identifying different cell types of lung cancer. WB-DWIBS could detect hilar or mediastinal lymph node metastasis. ADC, EADC and SI values were usefull in differenciate metastatic and benign lymph nodes.