Development of screening test using texture analysis in CT colonography

2023 Fiscal Year Final Research Report

• Project/Area Number: 21K15778
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 52040:Radiological sciences-related
• Research Institution: Chiba University
• Principal Investigator: Tochigi Toru 千葉大学, 医学部附属病院, 助教 (00645567)
• Project Period (FY): 2021-04-01 – 2024-03-31
• Keywords: テクスチャ解析 / CTコロノグラフィー / 大腸癌 / GLCM

Outline of Final Research Achievements

We investigated whether the texture analysis using images obtained by CTC could be used as a biomarker for colorectal tumors. We performed texture analysis using CTC images of 263 patients. We used the software developed with Pixspace Inc. for texture analysis, which is possible to identify the tumor site from the virtual endoscopic image and to obtain the ROI of the maximum section of tumor in the MPR image. As a result, GLCM correlation and homogeneity were significantly associated with several pathological factors such as depth, lymphovascular invasion, and lymph node metastasis. These parameters were also significantly associated with overall survival and the high value groups of each of these parameters had significantly poor prognosis. Furthermore, multivariate analysis showed that the GLCM correlation was an independent factor for survival. This study revealed that texture analysis using CTC image is useful for the practice of colorectal cancer.

Free Research Field

大腸外科

Academic Significance and Societal Importance of the Research Achievements

CTC画像を用いて大腸癌のテクスチャ解析を行い、GLCMといった新たなパラメーターがバイオマーカーとして悪性度診断に有用であることを見出したのは、本研究が初であると考えており、学術的な意義は大きい。CTCは比較的簡便な検査手法で大腸癌の存在診断から手術に対する精密検査まで可能である。さらに本研究による解析方法を用いることで初期段階から大腸癌の悪性度診断まで可能となり、CTC画像を用いたテクスチャ解析が大腸癌個別化診療の一助となると考えられる。