Quantification of CT perfusion curve shape and it's diagnosis of breast cancer: Computer simulation and breast perfusion

袁小东,张静,敖国昆,王培军,田嫒
DOI: https://doi.org/10.13929/j.1003-3289.2010.05.046
2010-01-01
Abstract:Objective To assess the influence of hemodynamic parameters and parameter combinations on shapes of CT perfusion time density curve (TDC),and to seek a method of quantitative evaluation on TDC shapes,as well as to use it in prospective research of CT perfusion in breast cancer. Methods Programming Excel 2003 to automate convolution and deconvolution calculations,Weibull function was used as impulse residue function (R(t)). When fixing the other parameters,progressive value of blood flow (BF),mean transit time (MTT),BF to MTT ratio (BF/MTT),parameters were taken respectively to obtain a series of virtual TDC,by which influence of the above parameters on TDC shape was verified and parameters that could describe TDC shape quantitatively were screened out. Twenty-eight patients with breast tumors underwent CT perfusion scan,and the presentations were classified as washout,persistent and plateau types according to TDC shape. Taking postoperative pathologic diagnosis as the golden standard,the value of BF,MTT,BV,BF/MTT and TDC shape type in the diagnosis of breast cancer was analyzed with receiver operating characteristic curve (ROC). Results ①The shape parameter of Weibull function had little influence on TDC. BF affected the height of TDC:the larger the BF was,the higher the TDC was. MTT affected the height and peak time of TDC:the larger the MTT was,the higher the TDC and at the same time peak time was delayed. BF/MTT determined TDC shape,TDC presenting as a washout type when relatively large values were taken,and as a persistent type when relatively small values were taken. ②BF/MTT of breast tumor was correlated with the TDC shape type. ③The area under ROC of all parameters of breast perfusion decreased in sequence as:BF/MTT,BF,TDC shape type,blood volume (BV) and MTT. The three former were larger than the area under the diagonal line (P<0.05),and had significance in differential diagnosis of malignant and benign breast tumors. Conclusion BF/MTT can describe TDC shape of CT perfusion quantitatively. It integrates the diagnostic information of BF and MTT in a non-linear way,and is therefore of significant reference value in the diagnosis of breast cancer.
What problem does this paper attempt to address?