Preliminary Study Of Multiple B Value Diffusion Weighted Imaging In Cervical Cancer Using 1.5T MR

Part I:Difference analysis of multiple bvalue diffusion weighted imaging parameter values(ADC、D、f、D*) in junctional zone and muscular layer in normal cervixObjective: To investigate the parameters(ADC、D、f、D*)difference between junctional zone and muscle layer of normal cervix of uterus in healthy adult women using multiple b value diffusion weighted imaging.Materials and Methods: The recruitment of 30 healthy volunteers as research objects(mean47±10 years old, range 28~65 years old). All healthy volunteerswere informed consent before examination. 30 healthy female volunteer’s inclusion and exclusion criteria were as follows:Inclusion criteria:(1) there were no known history gynecological diseases, no pelvic diseases and no treatment history;(2) there was no MRI contraindication, and claustrophobia;(3) all subjects examined by conventional MRI and multiple b value diffusion weighted imaging, and meeting the research requirements;(4) all subjects examined by MRI after the menstrual period within one week.Exclusion criteria:(1) MRI image quality was so poor that can not analysis image well;(2) ROI can not be set truely on junctional zone and muscular layer of normal cervix, Or the ROI’s max area was less than 20mm2;(3) MRI image showed cervical or pelvic lesions.All subjects underwent 1.5-T MRI system with conventional pelvic MRI(achieva,Philips) and multi-b-value(b=0,10,20,50,100,200,400,600,800,1000s/mm2) DWI. DWI images generated were post-processed by using Firevoxel software respectively. Combining T2 MI image, on the corresponding DWI image to identify normal junctional zone and muscle layer, draw the ROI respectively, and through the software selection calculation and measurement for mono- exponential and bi-exponential model respectively corresponding parameter values. Mono-exponential model selected value b= 0, 600 s/mm2 to obtain apparent diffusion coefficient(ADC); Bi-exponential model selected all b value(b=0,10,20,50,100,200,400,600,800,1000s/mm2), respectively performed to derive the perfusion parameters f(perfusion fraction) and D*(pseudodiffusion coefficient) as well as the diffusion parameter D(true molecular diffusion coefficient) in normal junctional zone and muscle layer. ROI setting: ROI area >20mm2, ROI setting by mono-exponential was correspondence with by bi-exponential model. When measuring junctional zone and muscle layer of normal cervix, the researcher choosed a continuous two levels of junctional zone and muscle layer of normal cervix, drawing the ROI respectively for three times to obtain the average parameter value.Multiple parameter values of junctional zone and muscle layer in normal cervix were analyzed by SPSS 17.0, using paired sample t test and when p<0.05, the difference was statistically significant.All MRI images in this part study were evaluated by two high qualification physicians, and reach agreement.Results: All MRI images in 30 cases of normal healthy volunteers can clearly display the normal cervical junctional zone and muscle layer. The normal cervical junctional zone’s average ADC value was(1.56±0.45) x 10-3mm/s, the average f value was(40.87±9.84)%, the average D*value was(17.64±4.80) x 10-3mm2/s, the average D value was(1.00±0.28) x 10-3 mm2/s; the normal cervical muscle layer’s average ADC value was(1.83±0.54) x 10-3mm2/s, the average f value was(36.00±5.99)%, the average D* value was(17.69±4.37) x 10-3mm2/s, the average D value was(1.24±0.20) x 10-3 mm2/s. By SPSS 17.0 statistical software analysis, the normal cervical junctional zone’s average ADC value was below the normal cervical muscle layer, P< 0.05, the difference was statistically significant; the normal cervical junctional zone’s average f value was above the normal cervical muscle layer, P< 0.05, the difference was statistically significant; the normal cervical junctional zone’s average D value was below the normal cervical muscle layer, P< 0.05, the difference was statistically significant. D* value had no statistically difference between normal cervical junctional zone and muscle layer, P> 0.05.Conclusion: This part study results indicate that ADC, f, D values between the normal cervical junctional zone and muscle layer were statistically significant by 1.5T MRI multiple b value DWI. But D * value had no obvious difference between junctional zone and muscle layer. So ADC, f, D values can effectively distinguish from normal cervical junctional zone and muscle layer, and having the potential value of evaluating cervical cancer invasion depth.Multiple b value DWI parameter(ADC、D、f、D*)average values form the fist part study were used as the second part study’s contrast group.Part II:Preliminary study ofmultiple b value diffusion weighted imaging parameter values(ADC、D、f、D*) in cervical cancerObjective: To investigate the parameters(ADC、D、f、D*) difference between cervical cancer and normal cervix. To analysis the resolving ability of the parameters(ADC、D、f、D*) betweenin cervical cancer and non cancerous area.Materials and Methods: From March 2015 to April 2016, we collected patients with surgery or biopsy pathologically proven cervical cancer in our hospital. According to the inclusion criteria and exclusion criteria, cervical cancer patients who conformed to the standard included were included in the cervical cancer group in this study. All subjects were informed consent before inspection and signed informed consent. Inclusion criteria and exclusion criteria were as follows:Inclusion criteria:(1) cervical cancer patients were proven with surgery or biopsy pathologically;(2) there was no MRI contraindication or claustrophobia;(3) all cervical cancer patients were examined by MRI after the menstrual period within one week;(4) all cervical cancer patients always had not any cervical cancer treatment;(5) time interval between MRI examination and pathological materials within one week.Exclusion criteria:(1) MRI image quality was so poor that can not analysis image well;(2) cervical cancer was so small that can not set ROI correctly.All subjects underwent 1.5-T MRI system with conventional pelvic MRI(achieva,Philips) and multi-b-value(b=0,10,20,50,100,200,400,600,800,1000s/mm2) DWI. DWI images generated were post-processed by using Firevoxel software respectively. Combining T2 MI, DWI, ADC image, the normal cervical zonal structure was damaged, cancerous tissue slightly high signal in T2 WI, high signal in DWI, slow signal in ADC image, and can have parametrium tissue infiltration. On corresponding DWI image to identify cervical cancer, draw the ROI, and through the software selection calculation and measurement for mono-exponential and bi-exponential model respectively corresponding parameter values. Mono-exponential model selected value b= 0, 600 s/mm2 to obtain apparent diffusion coefficient(ADC); Bi-exponential model selected all b value(b=0,10,20,50,100,200,400,600,800,1000s/mm2), respectively performed to derive f, D*, D values.ROI setting: ROI setting by mono-exponential was correspondence with by bi-exponential model. When measuring cervical cancer, the researcher choosed all levels of cervical cancer, drawing the ROI for three times to obtain the average parameter value.Multiple b value DWI parameter(ADC、D、f、D*)average values form the fist part study were used as the second part study’s contrast group.Multiple parameter values form cervical cance were analyzed by SPSS 17.0 and Medcalc15.2. Independent sample t test was used and when P<0.05, the difference was statistically significant. Receiver operating characteristics(ROC) curve determined the optimal cut-off values for significant parameters in achieving the best diagnostic accuracy in differentiating cervical cancer from non cancerous area, P<0.05, the difference were considered statistically significant.Results: 24 cervical cancer subjects were included in this part study, mean age 50±11years old, range 37~66 years, The cervical cancer’s average ADC value was(1.05±0.17)x 10-3mm/s, the average f value was(25.83±6.12)%, the average D*values was(16.83±4.02)x 10-3mm2/s, the average D values was(0.95±0.15) x 10-3 mm2/s. The normal cervical junctional zone’s average ADC value was(1.56±0.45) x 10-3mm/s, the average f value was(40.87±9.84)%, the average D*value was(17.64±4.80) x 10-3mm2/s, the average D value was(1.00±0.28) x 10-3 mm2/s; The normal cervical muscle layer’s average ADC value was(1.83±0.54) x 10-3mm2/s, the average f value was(36.00±5.99)%, the average D* value was(17.69±4.37) x 10-3mm2/s, the average D value was(1.24±0.20) x 10-3 mm2/s.By SPSS 17.0 and Medcalc15.2 statistical software analysis, the cervical cancer’s average ADC value was below the normal cervical junctional zone, the difference was statistically significant(P< 0.05); D and D* values had no statistically difference between and normal cervical junctional zone(P> 0.05). The cervical cancer’s average(ADC, f and D) values were below the normal cervical muscle layer, the difference was statistically significant(P< 0.05); D* values had no statistically difference between and normal cervical muscle layer(P> 0.05). Between cervical cancer and normal junctional zone, ADC value and f value were evaluated by ROC curve analysis, f value’s Az(0.923) was larger than ADC value’s(0.912), and no statistically difference between them(P> 0.05).When f value was 32% as the best threshold, sensitivity and specific degree were 83.3%, 86.7% respectively; when ADC value was 1.13×10-3mm2/s as the best threshold, sensitivity and specific degree were 79.2%、90% respectively. Between cervical cancer and normal cervical muscle layer, ADC, f and D value were evaluated by ROC curve analysis, ADC value’s Az(0.974) was the largest one, then D value(0.887) and f value(0.881), the difference between ADC’s Az and f’s Az was statistically significant(P< 0.05); when ADC value was 1.45×10-3mm2/s as the best threshold, sensitivity and specific degree were 100%、83.3% respectively; when D value was 1.04×10-3mm2/s as the best threshold, sensitivity and specific degree were 79.2%、90% respectively; when f value was 30% as the best threshold, sensitivity and specific degree were 79.2%、86.7% respectively.Conclusion: This part study results indicate that ADC, f, values can effectively distinguish from cervical cancer and normal cervical junctional zone by 1.5 MRI multiple b value DWI, both the identification ability was higher. ADC, f, D values can effectively distinguish from cervical cancer and normal cervical muscle layer, ADC values’ s differential capacity was the highest, D value and f value’s differential ability were medium.