An Application Study Of Multiple Myeloma With Diffusion-weighted MR Imaging And Quantitative Parameters Of Spectral CT

Part ⅠWhole-body diffusion-weighted MRI for evaluation of response inmultiple myeloma patients following bortezomib-based therapyObjective:To determine the feasibility of whole-body diffusion-weighted imaging(WB-DWI)MRI for evaluation of response in patients with multiple myeloma(MM)following bortezomib-based therapy and to explore the direction of apparent diffusion coefficient(ADC)changes upon treatment.Materials and Methods:Seventy-two MM patients who underwent WB-DWI MRI before and after bortezomib-based chemotherapy(21 weeks)were evaluated retrospectively.According to the pattern of bone marrow infiltration based on MRI,patients were divided into three groups as follows:focal,diffuse and mixed.According to the consensus of International Myeloma Working Group(IMWG),patients were classified as deep responders and non-deep responders.In the patient-based analysis,the changes of the ADCmean values of the lesions at baseline and follow-up in the three groups of patients were evaluated respectively.In the lesion-based analysis,the changes of the ADCmean values of the lesions at baseline and follow-up between the diffuse lesions and focal lesions grouped by responds were evaluated.The estimated tumor volume(eTV)and ADCmean values before and after chemotherapy were calculated and compared between deep and non-deep responders.Paired t-test was used to assess the samples with normal distribution and Wilcoxon signed rank test was use to assess the samples with non-normal distribution.The significance of differences of three gourps was evaluated by one-way analysis of variance(ANOVA).Predictive value of baseline ADCmean was calculated to predict the trend of ADCmean change following treatment.Results:A total of 846 diffuse lesions and 327 focal lesions were evaluated in this study.Out of the 72 patients,55 patients were classified as deep responders and 17 patients were classified as non-deep responders.For 846 representative diffuse lesions,the ADCmean value was significantly decreased from baseline to posttreatment.However,the ADCmean value was significantly increased following treatment in 327 focal lesions(FLs).Diffuse lesions showed a significantly decreased ADCmean value in deep responders,while significant increasing was found in FLs and mixed lesions.No significant variation in ADCmean value in diffuse and mixed lesions was found between deep and non-deep responders,whereas signicantly increased in FLs.ROC analysis of percentage change in ADCmean of diffuse lesions indicated that a 33.07%decrease in ADCmean correctly identified deep response to treatment with a sensitivity of 64.5%and specificity of 77%.Baseline ADCmean at a specific value(0.808×10-3 mm 2/s)yielded a maximum specificity(68.05%)and sensitivity(54.09%)in predicting increase of post-treatment ADCmean.Conclusions:1、The ADCmean value was significantly decreased in MM patients with diffuse pattern,while it was significantly increased in those with focal and mixed patterns following bortezomib-based treatment.2、WB-DW MRI could be used to discriminate deep response to induction treatment in MM patients with diffuse infiltration pattern.3、Baseline ADCmean value might have a potential to predict the trend of ADCmean change following treatment.Part ⅡA study of bone marrow infiltration pattern and tumor burden based on diffusion-weighted MR imagingObjective:The purpose of our study is to explore and compare the tumor load of different bone marrow(BM)infiltration patterns and evaluate the feasibility of ADC value to discriminate infiltration patterns of BM in multiple myeloma.Materials and Methods:93 patients with newly proven multiple myeloma and 23 controls had undergone routine MRI and DWI on a 3.0-T magnetic resonance imaging(MRI)unit.All the 93 patients had exhaustive laboratory data.According to the International Myeloma Working Group(IMWG)guidelines and MRI findings,five BM infiltration patterns were detected:normal bone marrow(N),pure diffuse infiltration(D),focal involvement(F),combined diffuse/focal infiltration(M/F+D)and a“salt-and pepper”pattern(SP).The values of hemoglobin,β2-MG,creatinine levels,the number of plasma cells,M protein and the ADC values of five patterns were analyzed.The Kruskal-Wallis H test was adopted to compare the laboratory data and the ADCs of different patterns.The chi-square test was adopted to compare the percentages of ISS staging,R-ISS staging and high-risk MM among different patterns.Spearman correlation coefficient was used to assess the correlation between the ADC values of diffuse involvement and the numbers of plasma cells.ROC analysis was used to establish the best diagnostic ADC value for identifying five patterns and distinguishing normal pattern patients from controls.Results:A total of 440 ROIs were collected from 93 MM patients,of which 113 ROIs were collected from 23 patients with diffuse pattern,76 ROIs from 17 patients with focal pattern,100 ROIs from 20 patients with normal pattern,and 83 ROIs from 19 patients with mixed pattern and 68 ROIs from 14 patients with“salt and pepper”pattern.111 ROIs were collected from 23 normal controls.The values of hemoglobin,β2-MG,plasma cell,M protein,the percentages of ISS staging and R-ISS staging and high-risk FISH,and ADC values showed significantly different among five BM infiltration patterns.However,there was no statistical difference in creatinine and lactate dehydrogenase among five patterns.ADCmean at a specific value(368.5 × 10-6 mm 2/s)yielded a maximum specificity(95.5%)and sensitivity(92.0%)in diagnosing MM BM.ADCmean at a specific value(335.5 × 10-6mm 2/s)yielded a maximum specificity(84.7%)and sensitivity(88.0%)in discriminating normal involvement from controls.There was a moderate positive correlation between the values of plasma cells and ADCs of diffuse involvement(r=0.648,P<0.05).Conclusions:1、The BM infiltration pattern in MM patients can indicate the tumor load correspondingly.2、ADC value has the ability to discriminate the five kinds of BM infiltration in MM patients.3、There is a moderate positive correlation between the ADC values of diffuse involvement and the numbers of plasma cells in MM patients.Part ⅢQuantitative evaluation of Bone Marrow infiltration in patients with multiple myeloma Using fast kVp switching Dual-Energy Spectral CTObjective:To explore the potential value of quantitative parameters derived from fast kVp switching dual-energy spectral computed tomography(DESCT)in detecting bone marrow(BM)infiltration and distinguishing different patterns of BM infiltration in patients with multiple myeloma(MM).Materials and Methods:This study included 35 patients with newly diagnosed,untreated MM and 15 healthy control subjects who had undergone spinal fast kVp switching DESCT and MRI.Pattern assignment was based on visual assessment of conventional magnetic resonance images,and the regions of interest were defined on both computed tomography and apparent diffusion coefficient maps.The values of DESCT parameters were measured and compared between infiltrated and healthy bone marrow.Receiver operating characteristic analysis was performed to determine the potential usefulness of these quantitative parameters in identifying BM infiltration and different patterns defined by MRI.The sensitivity and specificity under the optimal thresholds determined by the Youden Index were also calculated.Results:Statistical differences were observed between the fast kVp switching DESCT quantitative parameters including Ca(Water),Water(Ca),HAP(Fat),Fat(HAP)and effective atomic number(Eff-Z)but not for the 70 keV CT value between the infiltrated and healthy BM(all P<0.001).The 70keV CT value and Ca(Water),HAP(Fat)and Eff-Z values were also found to be statistically different in a comparison of different infiltration patterns(all P<0.05).The performance of the model-based parameter Ca/Water was shown to be superior in its capacity to differentiate between infiltrated and healthy BM(AUC=0.856[95%CI,81.4-89.1%],sensitivity=0.841 and specificity=0.768),as well as between MM patients and control subjects(AUC=0.910[95%CI,79.5-97.3%),sensitivity=0.829 and specificity=1.000).Conclusion:1、The quantitative analysis of fast kVp switching DESCT offers potential as a quantitative method for detecting infiltrated BM.2、The quantitative analysis of fast kVp switching DESCT offers potential as a quantitative method for evaluating infiltration patterns of BM in patients with MM.