Research On DWI Monoexponential And Biexponential Signal Decay Analysis With Pathological Specimen Of Breast

Objective: Many studies have reported the importance of attenuation of Apparent DiffusionCoefficient （ADC） values of DWI signal to identify benign and malignant breast lesions.But tissues in vivo don’t display a linear mono-exponential decay of ADC values, but rathera bi-exponential one; and thus they could be divided into fast （ADCfast） and slow （ADCslow）diffusion component using a simple bi-exponential decay model. The purpose of this studywas to examine the non mono-exponential characteristics of the diffusion weightedmagnetic resonance imaging signal attenuations of breast, to estimate the fast and slowdiffusion components, and to compare them with pathological specimen. Fifty one casesof clinical breast lumps were collected who underwent a preoperative single-b and multi-bDWI. The DWI signal attenuations were analyzed using mono-exponential andbi-exponential decay function; and the ADC, ADCfast, ADCslow; and fraction of fast ADCwere calculated and compared.Materials and Methods: Patients with clinically and pathologically proven breast lumpswere collected and studied with conventional MRI scan, single-b DWI （with b=0,800s/mm²）, multi-b DWI （with11different b values up to3000s/mm²）, and dynamiccontrast-enhanced MRI scan. The multi-b DWI from the normal and abnormal portions ofthe breast tissue were analyzed and compared; and reconstruction of the ADC map, ADCfastmap, ADCslowmap and fraction of fast ADC were done using using AW4.4workstation.Results: The ADCmonoof various types of lesions were significantly different, in decreasingorder of: normal glands> benign lesions> malignant lesions; while the ADCmonoof cyst hada wide range of values. The ADCfast, ADCslowof cysts using bi-exponential signalattenuation model was greater than the parameters in the group of normal glandular tissue,and the difference was statistically significant. In parameters of benign breast lesion just ffasthad significant difference, whereas the ADCfastand ADCslowobviously follow the same decreasing order sequence: fibrous nodules <intraductal papilloma <fibrous nodules withmyxomatous change <chronic lymphocytic mastitis. Slow component （ADCslow=0.573±0.122mm²/s） of invasive ductal carcinoma II （n=21） was significantly lower than normalglands （ADCslow=0.974±0.342mm²/s）; their fast component (ADCfast=3.877±1.965×10^-3mm²/s) almost equal to the normal glandular tissue (ADCfast=3.613±0.670×10^-3mm²/s), but the fraction of the fast component （ffast） had significant difference （fraction offast component of cancer tissue ffast=0.289±0.131; fraction of fast component of normalglands ffast=0.534+0.198）. The benign and malignant lesions group also had a similarfinding.Conclusion: Analysis of DWI signal attenuation of breast tumours using mono and bi-exponential function shows that ADCmonousing mono-exponential signal attenuation modelhave statistically significant differences. The parameters of normal glands, benign lesionsand malignant lesions （ffast, fslow, ADCslow） using bi-exponential decay model also havesignificant differences, whereas ADCfasthas no statistically significant differences. Thisresult suggests that the main reason for the decrease in ADC values of a breast tumour isdecrease in fraction of the fast component and the increase in fraction of the slowcomponent having a low ADC values rather than the decrease in ADC values of the fastcomponent by the restricted water diffusion in the reduced extracellular spaces.