Comparative study of diffusion kurtosis imaging model and diffusion weighted imaging model in diagnosisofbreastcancer / 中华放射学杂志

ABSTRACT

Objective To compare the value of diffusion kurtosis imaging (DKI) model with single-index DWI model parameters in the differential diagnosis of benign and malignant breast lesions,and to explore the correlation between the parameters and molecular subtypes and prognostic factors of breast cancer.Methods A retrospective analysis was performed with inclusion of 64 cases of breast diseases from January 2016 to May 2017 in Shanghai First People's Hospital.The patients were pathologically confirmed and typed, 30 cases are malignant tumors and 34 cases are benign lesions. DKI and DWI were performed within 2 weeks before the pathological examination. Invasive ductal carcinoma of grade Ⅰ, Ⅱ and Ⅲ were revealed in 1, 7 and 13 cases respectively. Luminal A breast cancer was found in 10 cases, Luminal B breast cancer was diagnosed in 11 cases, HER-2 positive breast cancer was 4 cases and triple negative breast cancer was 5 cases. The expressions of estrogen receptor (ER), Progesterone receptor (PR), and HER-2 positive were found in 20, 14 and 15 cases respectively. Ki67 was highly expressed in 24 cases and low expression in 6 cases. All patients underwent both plain and enhanced mammography scanning. The kurtosis (MK), mean diffusivity (MD) and ADC value were measured. Prognosis analysis was performed according to the maximum diameter (>2 cm, ≤2 cm), vascular or neurological invasion (positive, negative), lymph node metastasis (positive, negative), ER (positive, negative), PR (positive, negative), HER-2 (positive, negative),Ki67 (positive, negative), pathological grade (grade Ⅰ+Ⅱ,Ⅲ). Two independent samples t test was used to compare DKI and DWI parameters between benign and malignant lesions. ROC analysis was performed for assessing the values of parameters in discriminating benign and malignant breast lesions. Mann-Whitney U and Kruskal-Wallis H tests were used for the comparison of various prognostic factors or molecular subtypes.Spearman rank correlation analysis was used to explore the correlation of different prognostic factors and DKI and DWI parameters. Results The MK value of malignant group was higher than that of benign group,and the MD value and ADC value were lower than that of benign group (P<0.05). The area under the ROC for MK, MD and ADC were 0.897, 0.827 and 0.776, respectively. The area under the ROC was improved to 0.935 when three parameters were combined. The MK of ER positive group was higher than that of negative group (P<0.05). There was no significant difference of parameters among the other prognostic groups (all P>0.05). There was a low positive correlation between ER and MK (r= 0.417, P= 0.022). There was no correlation between the other prognostic factors and parameters (r=－0.086 to 0.313, all P>0.05). There was no significant difference in the MD, MK and ADC values among the four different subtypes of breast cancer (all P>0.05). Conclusions MK, MD and ADC values can be used to discriminate benign and malignant breast tumors, among which MK value has the best diagnostic performance. There is a certain correlation between DKI model parameters and prognostic factors.