Application Of 3.0T MRI Imaging In The Pathologic Features Of Cervical Cancer And Preoperative Staging

Objective:To investigate the relationship between multi-b-value DWI based on bi-exponential decay model and stretched-exponential model and clinical stage and pathological features of cervical cancer. And compare these two models with mono-exponential decay model.Material and Methods:74consecutive patients with pathologically confirmed cervical cancer by surgery or biopsy were collected, including59cases of squamous cell carcinoma (high, medium and low differentiation were13cases,35cases and11cases separately) and15cases of adenocarcinoma. All patients underwent3.0T MRI scan with conventional MRI, DWI (b=0,800s/mm2) and multi-b-value DWI (b=0,10,20,50,100,200,500,800,1200,1500,2000s/mm2).58patients had radical resection of the uterus or cervix within two weeks after MRI scanning. Single-exponential decay model was used to calculate the average of conventional apparent diffusion coefficient (ADC). Bi-exponential decay model was used to calculate slow apparent diffusion coefficient (ADCslow), fast apparent diffusion coefficient (ADCfast) and the fraction of fast diffusion component (Ffast), and stretched-exponential model was used to calculate distributed diffusion coefficient (DDC) and a. Then compare these parameters in different clinical stage and pathological features of cervical cancer. Receiver operating characteristics (ROC) analysis was performed in order to evaluate the diagnostic performance of mono-, bi-exponential decay model and stretched-exponential model in differentiating pathological features of cervical cancer. According to ROC curves, the optimal cut off value was extracted. Relevant factors of mono-, bi-exponential decay model and stretched-exponential model parameters were analyzed by multiple stepwise regression.Results:The values of ADC, ADCslow, DDC were lower in cervical squamous cell carcinoma than those in adenocarcinoma, while the value of ADCfast and Ffast in cervical squamous cell carcinoma were higher with statistical significance (P<0.05). ADCslow for differentiating squamous cell carcinoma from adenocarcinoma had a largest Az (0.979)(P<0.05). Using ADCsiow=0.38×10-3mm2/s as the threshold of identifying squamous cell carcinoma and adenocarcinoma, the sensitivity and specificity were91.5%and93.3%. The values of ADC, ADCslow, DDC and a were statistically different in different differentiation subtypes of cervical squamous cell carcinoma (P<0.05). α for differentiating high/medium from poorly differentiated squamous cell carcinoma had a largest Az (0.958)(P<0.05). Using α=0.60as the threshold of identifying high/medium from poorly differentiated subtypes, the sensitivity and specificity were89.1%and92.3%. The diffusion index ADC, ADCslow and DDC had a significantly positive correlation between each two of them (P<0.05). The index ADCfast and Ffast which reflect perfusion showed a moderate positive correlation (P<0.05). There was no statistical difference in these model parameters for different FIGO staging, size, depth of cervical infiltration, with/without lymphovascular invasion, with/without lymphnode metastasis groups of cervical cancer(P>0.05). Only pathological type and degree of differentiation were selected into stepwise multiple regression eqution.Conclusion:ADC, ADCslow, DDC, ADCfast and Ffast were helpful in identifying squamous cell carcinoma from adenocarcinoma, and the ability of ADCslow and DDC was superior to that of ADC. ADC, ADCslow, DDC and α all contributed to identify different differentiated subtypes of cervical squamous carcinoma. There is a correlation between mono-, bi-exponential decay model and stretched-exponential model parameres and pathological type and degree of differentiation. Objective:To evaluate the diagnostic performance of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in different clinical stage and pathological features of cervical cancer and to investigate the correlation between quantitative and semi-quantitative parameters of DCE-MRI and perfusion indexes ADCfast and Ffast of DWI based on bi-exponential decay model.Materials and Methods:74consecutive patients with pathologically confirmed cervical cancer by surgery or by biopsy were recruited, including59cases of squamous cell carcinoma (high, medium and low differentiation were13cases,35cases and11cases separately) and15cases of adenocarcinoma. All patients underwent3.0T MRI scan with conventional sequence, multi-b-value DWI (b=0,10,20,50,100,200,500,800,1200,1500,2000s/mm2) and DCE-MRI.58patients had radical resection of the uterus or cervix within two weeks after MRI scanning. Analyze the DCE derived parameters including time-signal intensity curve (TIC), semi-quantitative and quantitative parameters, that is SER, MSI, Tpeak, Kep, Ktrans and Ve. Then compare these parameters in different clinical stage and pathological features of cervical cancer. Receiver operating characteristics (ROC) analysis was performed in order to evaluate the diagnositic performance of DCE parameters in differentiating pathological features of cervical cancer. According to ROC curves, the optimal cut off value was extracted. Relevant factors of DCE parameters were analyzed by multiple stepwise regression.Bi-exponential signal decay model was used to calculate fast apparent diffusion coefficient (ADCfast) and the fraction of fast diffusion component (Ffast). Analyze the correlation between quantitative and semi-quantitative parameters of DCE-MRI and ADCfast and Ffast of DWI.Results:TIC type distribution was significantly different in well and moderately differentiated squamous cell carcinomas (P<0.05). There was no difference in TIC type distribution in different pathological types, FIGO staging, size, depth of cervical infiltration, with/without lymphovascular invasion, with/without lymphnode metastasis groups of cervical cancer (P>0.05). The values of SER, MSI, Ktrans, Kep, and Ve were higher in cervical squamous cell carcinoma than those in adenocarcinoma, while the value of Tpeak in cervical squamous cell carcinoma was lower than that in adenocarcinoma with statistical significance (P<0.05). Ktrans for differentiating squamous cell carcinoma from adenocarcinoma had a largest Az (0.864)(P<0.05). Using Ktrans=0.26min-1as the threshold of identifying squamous cell carcinoma and adenocarcinoma, the sensitivity and specificity were96.6%and86.7%. Values of SER, MSI, Ktrans, Kep, and Ve were statistically different in different differentiation subtypes of cervical squamous cell carcinoma (P<0.05). Ktrans for differentiating high/medium from poorly differentiated squamous cell carcinoma had a largest Az (0.864)(P<0.05). Using Ktrans=0.32min-1as the threshold of identifying high/medium from poorly differentiated subtypes, the sensitivity and specificity were82.6%and92.3%. There was no statistical difference in DCE parameters for different FIGO staging, size, depth of cervical infiltration, with/without lymphovascular invasion, with/without lymphnode metastasis groups of cervical cancer(P>0.05). Only pathological type and degree of differentiation were selected into stepwise multiple regression eqution. SER, MSI, Ktrans, Kep, and Ve showed a weak to moderate positive correlation with ADCfast and Ffast; Tpeak had a weak negative correlation with ADCfast and Ffast (P<0.05).Conclusion:SER, MSI, Ktrans, Kep, Ve and Tpeak were helpful in identifying cervical squamous cell carcinoma from adenocarcinoma. TIC type, SER, MSI, Ktrans, Kep and Ve all contributed to identify different differentiated subtypes of cervical squamous cell carcinoma. There was a correlation between DCE parameres and pathological type and degree of differentiation. There was a certain correlation between quantitative and semi-quantitative parameters of DCE-MRI and ADCfast and Ffast of DWI. Objective:To evaluate the diagnostic performance of T2WI, DWI, small FOV eDWI and DCE-MRI in parametrial invasion and vaginal involvement of cervical cancer.Materials and Methods:58consecutive patients with pathologically confirmed cervical cancer by surgery were collected, with an average of43.12±10.80years, range31to58years. All patients underwent routine MRI, DWI, eDWI(26×10.4cm) and DCE-MRI examination two weeks before surgery. Multiple sequences of MRI were divided into six groups:T2WI, T2WI+DWI, T2WI+eDWI, T2WI+DCE, T2WI+DWI+DCE and T2WI+eDWI+DCE. Compared with postoperative pathological findings, calculate the sensitivity, specificity and consistent rate of six groups in the diagnosis of parametrial invasion and vaginal involvement. The consistency of six MRI groups and pathological findings in diagnosis of parametrial invasion and vaginal involvement was calculated by Kappa test.Results:Patients diagnosed with parametrial invasion were7cases,4cases,3cases,1case, lease and1case in T2WI, T2WI+DWI, T2WI+eDWI, T2WI+DCE, T2WI+DWI+DCE and T2WI+eDWI+DCE separately. The pathologically diagnosed were2cases. The sensitivity, specificity and consistent rate of T2WI in diagnosis of parametrial invasion were100.0%,8.9%and91.4%; The sensitivity, specificity and consistent rate of T2WI+DWI were50.0%,96.4%and94.8%; The sensitivity, specificity and consistent rate of T2WI+eDWI were50.0%,98.2%and96.6%; The sensitivity, specificity and consistent rate of T2WI+DCE, T2WI+DWI+DCE, T2WI+eDWI+DCE were same, that’s50.0%,100.0%and98.3%. Kappa values were0.413,0.374,0.482,0.659,0.659and0.659for each group above (P<0.05).Patients diagnosed with vaginal involvement were12cases,10cases,9cases,8case,8case and8case in T2WI, T2WI+DWI, T2WI+eDWI, T2WI+DCE, T2WI+DWI+DCE and T2WI+eDWI+DCE separately. The pathologically diagnosed were10cases. The sensitivity, specificity and consistent rate of T2WI in diagnosis of vaginal involvement were60.0%,89.6%and84.5%; The sensitivity, specificity and consistent rate of T2WI+DWI were70.0%,95.8%and89.7%; The sensitivity, specificity and consistent rate of T2WI+eDWI were80.0%,95.8%and93.1%; The sensitivity, specificity and consistent rate of T2WI+DCE, T2WI+DWI+DCE, T2WI+eDWI+DCE were the same, that’s80.0%,100.0%and96.6%. Kappa values were0.477,0.685,0.758,0.869,0.869and0.869for each group above (P<0.05).Conclusion:T2WI alone has limited value in diagnosis of parametrial invasion. T2WI+DWI, T2WI+eDWI sequences improve the diagnositic accuracy of parametrial invasion and vaginal involvement. T2WI+DCE is the optimal sequence combination in diagnosis of parametrial invasion and vaginal involvement, while DCE combined with DWI and eDWI could not improve the diagnositic accuracy.