| Objective:To explore the clinical application value of multiple b value diffusion weighted image based on IVIM mode in node lymphoma and to compare with monoexponential mode. To Explore the repeatability of D, f, D*and ADC values, and evaluate the feasibility of its use in clinical practice. D value derived from IVIM model was compared with ADC value derived from single index model.To explore the clinical value of D, f, D*in monitoring and prognosis of chemotherapy response.Material and methods:1. The inclusion and exclusion criteriaInclusion criteria:Age>18years, lymphoma confirmed by pathology and patients accept chemotherapy in our hospital.Exclusion criteria:(1) patients with internal metal implants, such as patiens with coronary stent, pathents with internal fixation after fracture;(2) patients with claustrophobia;(3) patients cannot lie down for a long time, such as some patients with a lot of pleural effusion.This prospective, single-center study was approved by the hospital ethics committee,and be carried out under their supervision. Informed consent was obtained.2. Magnetic resonance imaging (MRI) scan methodPhilips Archieve1.5T (Philips Healthcare, Best, Netherlands) superconducting MRI scanner and its corresponding software were used in this study. The whole body lymph node was divided into four regions:neck, chest, abdomen and pelvic respectively. The neck region was scanned with head and neck combined coil.The chest, abdomen and pelvic region were scanned with sensitivity encoding technique (sensitivity encoding SENSE) of abdominal coil. In addition to the neck region scanned by free breathe, the rest regions were scanned by respiratory gating. All patients did MR scan before chemotherapy, after1cycle of chemotherapy and2cycle of chemotherapy. Scanning sequences included routine MRI non-enhancing scan and multiple b value of DWI. Conventional MRI sequences included axial T1WI, axial T2WI, axial T2WI fat supression, coronal T2WI fat supression (the neck region with STIR, the abdomen region with SPIAR). Multiple b value of DWI examination adopted single plane echo (SS-EPI) sequence, parallel sampling space sensitive coding technology (ASSET).Axial DWI images(TR6000ms, TE64ms, excitation frequency (NEX)2, matrix176x176, field of view (FOV)28～35cm, slice thickness/gap,5/1mm) were collected with spectral presaturation inversion recovery and diffusion sensitive gradient pulse in X, Y, Z axis direction applied with weighting factors of b0,10,20,30,40,50,75,100,200,300,500,800,1000s/mm2.3. Image analysis and data processingThe bi-exponential model from an IVIM sequence was expressed by the equation Sb/So=(1-f)exp(-bD)+fexp[-b(D*+D)].The mono-exponential model was expressed by the equation Sb=So exp (-bADC). Diffusion-weighted imaging data were transferred to a computer equipped with a manufacturer-supplied Software PRIDE DWI Tool, version1.5, Philips Healthcare) for IVIM and mono-exponential analysis. ROI measurements of IVIM parametric maps were obtained with Image J software(National institute of Health, Bethesda, MD). ROI was manually placed on each lesion at the level of maxium transverse diameter of lesions on D map and were chosen to be as large as possible,consistent with minimal contamination from surrounding unintended tissues.Large cystic, necrotic or heamorrage areas by visual inspection were excluded from ROI in order to focus upon viable tumor tissue in deriving IVIM parameters and ADC values.ROI was copied on f map and D*map to derive f value and D*value. ADC was measured at the same slice with D map, and ROI was manually drawn on the same area as possible.1～5largest nodules or masses (the largest diameter>1.0cm) were chosen for analysis, and different regional lymph nodes were chosen as possible. A total of74swollen lymph nodes, of which52in neck,2in mediastinum,20in abdomen were chosen for analysis.The comparison of lymphoma before and after treatment was evaluated on the axial T2WI based on calculating the reduction rate of each lymphoma lesion before and after treatment in the largest diameter, the calculation formula as follows:the largest diameter smaller ratio=(the maximum diameter pretreatment-the maximum diameter posttreatment)/the maximum diameter pretreatment x100%. According to RECIST criteria, it is divided into four groups:(1) complete remission (complete response, CR):tumor completely disappeared and sustain more than4weeks, a total of49lesions.(2) the partial response (partial response, PR):largest tumor diameter reduced30%or more, a total of17lesions;(3) stability (stable diseases, SD):tumor maximum diameter changed between PR and PD, a total of eight lesions;(4) progress (progressive diseases, PD):tumor maximum diameter increased20%or more, no lesions.4. statistical analysisThe intraclass correction coefficient(ICC) was calculated to derive the data varability for the2different obserbers. D value was compared with the ADC values with paired sample t test. Pretreatment D,ADC, f value were compared between PR group, CR group and SD group with Kruskal Wallis H test. Pretreatment D*value were compared between PR group, CR group and SD group with one-way ANOVA. Pretreatment D,ADC value were compared between two of the CR,PR,SD group by Nemenyi test. D, f, D*,ADC value of PR group were compare between pretreatment and posttreatment with paired samples t test. D, f, D*,ADC value were compared between solid group and necrosis group with two independent samples t test. SPSS statistical software19.0was used for above statistical analysis. P<0.05was thought to be statistically significant. Pretreatment D,ADC value for predicting reponse was evaluated with ROC curve. ROC curve with done by Medcal(Medcalc software, Acacialaan22,B-8400Ostend, Belgium).ResultsA total of22cases of nodal lymphoma patients did MRI routine scan before chemotherapy and multiple b value DWI scan, only12patients did MRI scan after one cycle of chemotherapy,21patients did MRI scan after two cycle of chemotherapy.14patients were male, and8were female.Age range from24to78, with an average age of49.6years. Pathological types:15patients of diffuse large B cell lymphoma,1of Gray zone lymphoma with features of Burkitt lymphoma and diffuse large B cell lymphoma,1of follicular lymphoma,1of mantle cell lymphoma,1of small lymphocytic lymphoma/chronic lymphocytic leukemia,1of anaplastic large cell lymphoma,1of lymphoblastic lymphoma,1of typical Hodgkin’s lymphoma, nodular sclerosis type.1. The repeatability of parameter between different observersD value between two observers of the ICC was0.997(95%CI:0.996～0.996). F value between two observers of the ICC was0.959(95%CI:0.940～0.940). D*value between two observers of the ICC was0.936(95%CI:0.909～0.909). Two observers of ADC values between the ICC was0.992(95%CI:0.986～0.986). The repeatability of D, f, D*,ADC between two observers was good, the best of which is D value, followed by the ADC, f, D*.2.The comparison of D value and ADC value74lymph nodes were measured before treatment with D value of (0.59±0.22) x10"3mm2/s, ADC of (0.65±0.27) x10-3mm2/s. ADC value was higher than D value (t=6.360, P=0.000).3. The comparison of pretreatment D, f, D*, ADC between different response groupPretreatment D value of CR group was (0.63±0.26) X10-3mm2/s. Pretreatment f value of CR group was (7.26±5.48)%. Pretreatment D*value of CR group was (101.09±59.18) X10-3mm2/s. Pretreatment ADC value of CR group was (0.69±0.31) X10-3mm2/s. Pretreatment D value of PR group was (0.57±0.10) X10-3mm2/s. Pretreatment f value of PR group was (6.44±3.25)%. Pretreatment D*value of PR group was(90.37±45.33) X10-3mm2/s. Pretreatment ADC value of PR group was (0.63±0.11) X10-3mm2/s. Pretreatment D value of SD group was (0.42±0.04) X10-3mm2/s. Pretreatment f value of SD group was (4.51±2.31)%. Pretreatment D*value of SD group was (82.74±37.66) X10-3mm2/s. Pretreatment ADC value of SD group was (0.46±0.02) X10-3mm2/s.D value was different between CR, PR and SD group (χ2=12.944, P=0.002). Pretreatment D value with no significant difference between CR and PR group (χ2=0.072, P=0.965), while the pretreatment D value of SD group was obviously lower than that of CR(χ2=12.090, P=0.002) and PR group (χ2=10.684, P=0.005). ADC value was different between CR, PR and SD group (x2=13.164, P=0.001). Pretreatment ADC value with no significant difference between CR and PR group (χ2=0.092, P=0.955), while the pretreatment D value of SD group was obviously lower than that of CR group (χ2=12.223, P=0.002) and PR group (χ2=10.945, P=0.004). Pretreatment f value with no significant difference between CR,PR and SD group (χ2=2.312, P= 0.315). Pretreatment D*value with no significant difference CR,PR and SD group (F=0.535, P=0.588). D≤0.48×10-3mm2/s indicated poor chemotherapy response, and the area under the curve (AUC) was0.890, sensitivity and speciality rate were100%and75.76%, respectively. ADC≤0.49×10-3mm2/s indicated poor chemotherapy response, and the area under the curve (AUC) was0.893, sensitivity and speciality rate were100%and84.85%, respectively. There is no statistical difference of predictive effect between pretreatment ADC and D (Z=0.086, P=0.931).4. The change of D、f、D*after chemotherapy in different response groupPosttreatment D value of PR group was (1.03±0.37)×10-3mm2/s. Posttreatment f value of PR group was (9.39±4.52)%. Posttreatment D*value of PR group was (99.72±42.12)×10-3mm2/s. Posttreatment ADC value of PR group was (1.12±0.37)×10-3mm2/s.D value increased significantly after treatment in PR group (t=-4.781, P=0.000). ADC value increased significantly after treatment in PR group (t=-4.728,P=0.000). f value increased significantly after treatment in PR group (t=-2.294,P=0.036), D*value slightly increased after treatment in PR group, but there was no statistically significant difference (t=-0.579,P=0.570).5. The comparison between necrotic group and solid groupThe D value of solid part in necrotic group was (1.09±0.12)×10-3mm2/s. The f value of solid part in necrotic group was (16.40±4.06)%.The D*of solid part in necrotic group was (95.81±44.47)×10-3mm2/s. The ADC value of solid part in necrotic group was (1.19±0.20)×10-3mm2/s.The D value of solid group was (0.54±0.10)×10-3mm2/s. The f value of solid group was (5.90±3.11)%, The D*value of solid group was (160.28±72.55)×10-3mm2/s. The ADC value of solid group was (0.60±0.12)×10-3mm/s. The D value of solid part in necrosis group was significantly higher than that of solid group (t=11.292, P=0.000). The f value of solid part in necrosis group was significantly higher than that of solid group (t=6.484, P=0.000). The D*value of solid part in necrosis group was significantly lower than that of solid group (t=-2.396, P=0.028). The ADC value of solid part in necrosis group was significantly higher than that of solid group (t=8.188, P=0.000).Conclusions1. D derived by IVIM model is significantly lower than ADC derived by mono-exponential model, and the repeatability of D is better than that of ADC.The repeatability of D value is the best,followed by f, D*.2. Pretreatment D,ADC value can predict lymphoma chemotherapy response. Pretreatment D≤0.48X10-3mm2/s indicate poor chemotherapy response in lymphoma. There is no statistical difference of predictive effect between pretreatment ADC and D value. f, D*with no help in predicting of lymphoma chemotherapy response.3. D, ADC and f increased significantly after chemotherapy in an early stage.4. IVIM DWI can be utilized as a noninvasive assessment of lymphoma chemotherapy response and predict reponse. |
PDF Full Text Request