講演抄録/キーワード |
講演名 |
2023-03-06 10:10
Radiogenomic signature based on CT images to predict HOPX gene expression and prognoses of patients with non-small cell lung cancer
-- Prognoses and prediction of HOPX expression for NSCLC based on CT image -- ○Yu Jin・Hidetaka Arimura・YunHao Cui・Takumi Kodama・Shinichi Mizuno(Kyushu Univ.)・Satoshi Ansai(Kyoto Univ.) MI2022-77 |
抄録 |
(和) |
(まだ登録されていません) |
(英) |
The homeodomain-only protein homeobox, HOPX, has recently been discovered that associated with the prognoses of non-small cell lung cancer (NSCLC) patients. This study aims to explore radiogenomic signature (RgS) based on computed tomography (CT) images connected with HOPX gene to predict prognosis for NSCLC patients. The best RgS, consisting of three image features, were used for building a stacking model, which exhibited the highest predictive power with an area under receiver operating characteristic curve of 0.705 (accuracy:0.750, specificity:0.765 and sensitivity:0.714) and showed the prognostic power in Kaplan-Meier curves (p-value=0.015) in a test dataset. The HOPX-expression status in lung cancer could be identified by using RgS in the stacking model, which suggested the potential of the RgS for the prediction of the gene expression and prognosis in NSCLC patients based on CT images. |
キーワード |
(和) |
/ / / / / / / |
(英) |
Radiogenomics / HOPX / CT image / machine learning / lung cancer / / / |
文献情報 |
信学技報, vol. 122, no. 417, MI2022-77, pp. 20-23, 2023年3月. |
資料番号 |
MI2022-77 |
発行日 |
2023-02-27 (MI) |
ISSN |
Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
PDFダウンロード |
MI2022-77 |