Whole-volume histogram derived from apparent diffusion coefficient maps for assessment of gastric cancer: the intra-and inter-observer reproducibility study

Tingting GUO,Song LIU,Jian HE,Changfeng JI,Nan ZHOU,Zhengyang ZHOU
2017-01-01
Journal of Medical Imaging
Abstract:Objective To evaluate the intra-and inter-observer reproducibility using whole-volume histogram derived from apparent diffusion coefficient (ADC) maps for assessment of gastric cancer.Methods Seventy patients with gastric carcinoma proved by biopsy pathology underwent diffusion weighted (DW) magnetic resonance imaging preoperatively.Tumors were delineated on each slice of the ADC maps, and a histogram from the entire tumor volume was subsequently generated.A series of histogram parameters (including ADCmean, Standard deviation, max frequency, mode, ADCmin, ADCmax, the 5 th, 10 th, 25 th, 50 th, 75 th, 90 th percentiles of the cumulative ADC histograms etc.) were calculated.Intra-and inter-observer agreement of each histogram parameter was analyzed by calculating the intraclass correlation coefficient (ICC).Results The intra-observer agreement of both reviewers and inter-observer agreement between them were generally good.For observer 1, the intra-observer agreement was perfect in 5 parameters, excellent in 8 parameters, very good in 2 and good in 1 parameters.For observer 2, the intra-observer agreement was perfect in 1 parameter, excellent in 8 parameters, very good in 4 and good in 3 parameters.And the inter-observer agreement was perfect in 5 parameters, excellent in 8 parameters and very good in 3 parameter.Conclusion The intra-and inter-observer reproducibility was good enough for the assessment of ADC histogram parameters in gastric cancers, especially for (10 th, 25 th, 50 th) percentile ADC, sample number and ADCmean.
What problem does this paper attempt to address?