Premium
Assessment of histological differentiation in gastric cancers using whole‐volume histogram analysis of apparent diffusion coefficient maps
Author(s) -
Zhang Yujuan,
Chen Jun,
Liu Song,
Shi Hua,
Guan Wenxian,
Ji Changfeng,
Guo Tingting,
Zheng Huanhuan,
Guan Yue,
Ge Yun,
He Jian,
Zhou Zhengyang,
Yang Xiaofeng,
Liu Tian
Publication year - 2017
Publication title -
journal of magnetic resonance imaging
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.563
H-Index - 160
eISSN - 1522-2586
pISSN - 1053-1807
DOI - 10.1002/jmri.25360
Subject(s) - kurtosis , histogram , effective diffusion coefficient , receiver operating characteristic , percentile , nuclear medicine , magnetic resonance imaging , cancer , medicine , correlation coefficient , skewness , diffusion mri , radiology , pathology , mathematics , statistics , computer science , artificial intelligence , image (mathematics)
Purpose To investigate the efficacy of histogram analysis of the entire tumor volume in apparent diffusion coefficient (ADC) maps for differentiating between histological grades in gastric cancer. Materials and Methods Seventy‐eight patients with gastric cancer were enrolled in a retrospective 3.0T magnetic resonance imaging (MRI) study. ADC maps were obtained at two different b values (0 and 1000 sec/mm 2 ) for each patient. Tumors were delineated on each slice of the ADC maps, and a histogram for the entire tumor volume was subsequently generated. A series of histogram parameters (eg, skew and kurtosis) were calculated and correlated with the histological grade of the surgical specimen. The diagnostic performance of each parameter for distinguishing poorly from moderately well‐differentiated gastric cancers was assessed by using the area under the receiver operating characteristic curve (AUC). Results There were significant differences in the 5 th , 10 th , 25 th , and 50 th percentiles, skew, and kurtosis between poorly and well‐differentiated gastric cancers ( P < 0.05). There were correlations between the degrees of differentiation and histogram parameters, including the 10 th percentile, skew, kurtosis, and max frequency; the correlation coefficients were 0.273, –0.361, –0.339, and –0.370, respectively. Among all the histogram parameters, the max frequency had the largest AUC value, which was 0.675. Conclusion Histogram analysis of the ADC maps on the basis of the entire tumor volume can be useful in differentiating between histological grades for gastric cancer. Level of Evidence: 4 J. Magn. Reson. Imaging 2017;45:440–449.