Font Size: a A A

The Study Of Diffusion Kurtosis MR Imaging For Assessing The Diagnosis And Therapy Of Breast Diasease

Posted on:2017-05-06Degree:DoctorType:Dissertation
Country:ChinaCandidate:K SunFull Text:PDF
GTID:1484305906961519Subject:Medical imaging and nuclear medicine
Abstract/Summary:PDF Full Text Request
OBJECTIVES To compare the differences of Kurtosis(K)and Diffusivity(D)among various DKI schemas,and search for the optimal combination of b values for characterising breast cancer.And to evaluate the diagnostic value of K and D in characterizing benign and malignant breast lesions,and also compare the diagnostic accuracy between DKI and DWI derived parameters.And to evaluate the potential association between DKI-derived parameters and ER,PR,Her2?Ki67 expression,histological grade,tumor size and lymph nodes metastasis of invasive breast cancer,and also evaluate the combined diagnostic value of DCE and DKI in invasive breast cancer.And to assess the value of DKI in monitoring therapy effect of neoadjuvant chemotherapy(NAC)in patients with locally advanced breast cancer(LABC).MATERIALS AND METHODS Patients from January 2014 and Novermber 2014,who were treated in our hospital were retrospectively analyzed.20 patients with proven breast cancer were selcted to compare the differences between various b values groups.Nighty-seven patients who undergone surgery were used for the comparison of K and D between benign and malignant breast lesions and the diagnostic accuracy of DWI and DKI were compared.Fifty-two invasive carcinoma were used for the correlation analysis of K,D and different expression of ER,PR,cerbB2,Ki67,lesions ' size and lymph nodes metastastis.Sixteen patients with LABC undergoing NAC were used for the evaluation the accuracy of K and D.MRI scanning with DKI sequence was examined before NAC,the second point and after all cycles.All these patients underwent conventional MRI,DKI(0,700,1400,2100,2800)and DWI(50,1000).ROls were drawn on ADC maps by manul,then copy to K and D maps.DCE features were analyzed by using the BI-RADS MRI lexicon.One calssification was based on DCE only;the other was based on DCE and DKI.We classify all these data into 5b,4b and 3b values groups.Friedman test were used to compare differences between various b values groups.Student t test were used for the comparision between benign and malignant lesions of ADC,K and D.ROC curves were used to find the best calssification threshold of benign and malignant lesions,and calculating its sensitivity and specificity.Wilcoxon-signed-rank and Jonckheere-Terpstra test were used for the comparision between different expression of ER,PR,cerbB2,Ki67,different sizes,lymph nodes metastastis and histological grades of K,D and ADC.Spearman correlation was used to evaluate the correlation between diffusion parameters and all prognostic factors.Paired-Sampled T test were used in all comparasion of K and D between the response and non-response groups.ROC curve was drawn to compare the diagnostic accuracy of K and D in evaluating different pathological response.RESULTS 1)There were significant differences among different b value groups of both kurtosis and diffusivity(P<0.001).The 4b values group of 0,700,1400,2100 had the highest K value(1.109±±0.120),while the 5 b values group of 0,700,1400,2100,2800 had the lowest D value(0.968±0.128).The 5 b values group can maximize the average diagnosis of K and D.2)K were significantly higher in the malignant lesions than in the benign lesions(1.05±0.22 vs 0.65±0.11,respectively;P<0.0001).D and ADC in the malignant lesions were significantly lower than those in the benign lesions(1.13±0.27vs1.97±0.33 and 1.02±0.18vs1.48±0.33,respectively;P<0.0001).Significantly higher specificity for differentiation of malignant from benign lesions was shown with the use of kurtosis and diffusivity coefficients than with the use of ADC(83%[34/41]and 83%[34/41]vs76%[31/41],respectively;P<0.0001)with equal sensitivity(95%[54/57]).3)Patients with grade 3 breast cancer or tumors with high expression of Ki67 were associated with higher kurtosis and lower diffusivity coefficients.The combined use of DCE and DKI showed higher specificity(90%[37/41])than did the only use of DCE(37%[15/41],P<0.0001),with equal sensitivity(97%[55/57]).4)There were 11 lesions in response group and 5 lesions in non-response group.There was no signifiant difference of both K and D in response group and non-response group of baseline(1.083±0.128 vs 0.984±0.090;0.970±0.078 vs 1.103±0.162;P>0.05).There was no signifiant difference of both K and D in response group and non-response group of the second point(0.835±0.220 vs 0.840±0.142;1.342±0.305 vs 1.242±0.30;P>0.05).There were significant differences of both K and D in response and non-response group In the last point(0.581±0.048 vs 0.791±0.111;1.858±0.146 vs 1.375±0.209;P<0.001).The change rates of K and D between the last point and the baseline had the highest area under the ROC curves(0.950).CONCLUSIONS Equally spaced b-values schema was recommended and the 5b schema based on its estimation accuracy of both K and D in characterizing breast cancers.K and D can provide valuable information in the diagnosis of benign and malignant breast lesions.DKI showed higher specificity than did DWI for the assessment of benign and malignant breast lesions.The kurtosis and diffusivity values derived from the DKI model may be helpful for the preoperative differentiation of histologic grade and proliferative activity of breast cancer.The combined use of DCE and DKI had a higher diagnostic specificity and accuracy.DKI model may be helpful for monitoring therapy effect of NAC in patients with LABC.
Keywords/Search Tags:Breast cancer, Magnetic resonance imaging, Diffusion kurtosis imaging, Kurtosis coefficient, Diffusivity coefficient, Dynamic contrast enhancement
PDF Full Text Request
Related items