Texture analysis of CT colonography to develop a novel imaging biomarker for the management of colorectal cancer

• DOI: 10.1002/ags3.12852
• Publication Date: 2024-08-26T08:05:20Z
• Authors: Hisashi Mamiya; Toru Tochigi; Koichi Hayano; Gaku Ohira; Shunsuke Imanishi; Tetsuro Maruyama; Yoshihiro Kurata; Yumiko Takahashi; Atsushi Hirata; Hisahiro Matsubara
• Keywords: Kurtosis; Imaging biomarker

Abstract Background Recent studies have focused on evaluating the biomarker value of textural features in radiological images. Our study investigated whether or not a texture analysis computed tomographic colonography (CTC) images could be novel for colorectal cancer (CRC). Methods This retrospective 263 patients with CRC who underwent contrast‐enhanced CTC (CE‐CTC) before curative surgery between January 2014 and December 2017. Multiple analyses (fractal, histogram, gray‐level co‐occurrence matrix [GLCM] analyses) were applied to CE‐CTC (portal‐venous phase), fractal dimension (FD), skewness, kurtosis, entropy, GLCM parameters, including GLCM‐correlation, GLCM‐autocorrelation, GLCM‐entropy, GLCM‐homogeneity, tumor calculated. These parameters compared pathological factors (tumor depth, lymph node metastasis, vascular invasion, lymphatic invasion) overall survival (OS). Results Tumor depth was significantly associated FD, GLCM‐homogeneity ( p = 0.001, 0.018, 0.008, respectively); metastasis 0.004); invasion GLCM‐correlation 0.001 0.012, venous GLCM‐entropy 0.033, 0.021, 0.046, respectively). In Kaplan–Meier analysis, high tumors showed worse OS than others 0.04, Multivariate that correlation an independent prognostic factor 0.021). Conclusion CE‐CTC‐derived may clinically useful biomarkers managing patients.