Predicting the nodal status in gastric cancers: The role of apparent diffusion coefficient histogram characteristic analysis

Song Liu,Yujuan Zhang,Jie Xia,Ling Chen,Wenxian Guan,Yue Guan,Yun Ge,Jian He,Zhengyang Zhou

DOI: https://doi.org/10.1016/j.mri.2017.07.013

IF: 3.13

2017-01-01

Magnetic Resonance Imaging

Abstract:Purpose To explore the application of histogram analysis in preoperative T and N staging of gastric cancers, with a focus on characteristic parameters of apparent diffusion coefficient (ADC) maps. Materials and methods Eighty-seven patients with gastric cancers underwent diffusion weighted magnetic resonance imaging (b=0, 1000s/mm2), which generated ADC maps. Whole-volume histogram analysis was performed on ADC maps and 7 characteristic parameters were obtained. All those patients underwent surgery and postoperative pathologic T and N stages were determined. Results Four parameters, including skew, kurtosis, s-sDav and sample number, showed significant differences among gastric cancers at different T and N stages. Most parameters correlated with T and N stages significantly and worked in differentiating gastric cancers at different T or N stages. Especially skew yielded a sensitivity of 0.758, a specificity of 0.810, and an area under the curve (AUC) of 0.802 for differentiating gastric cancers with and without lymph node metastasis (P<0.001). All the parameters, except AUClow, showed good or excellent inter-observer agreement with intra-class correlation coefficients ranging from 0.710 to 0.991. Conclusion Characteristic parameters derived from whole-volume ADC histogram analysis could help assessing preoperative T and N stages of gastric cancers.