2023 Fiscal Year Final Research Report
Construction of Deep Learning Model for Determining Histological Treatment Effect after Neoadjuvant Therapy of Lung Cancer
Project/Area Number |
21K06923
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Multi-year Fund |
Section | 一般 |
Review Section |
Basic Section 49020:Human pathology-related
|
Research Institution | Nara Medical University (2023) Kyoto University (2021-2022) |
Principal Investigator |
|
Project Period (FY) |
2021-04-01 – 2024-03-31
|
Keywords | 肺癌 / ネオアジュバント化学療法 / 主病理学的奏功率 / バイオマーカー / デジタルパソロジー / 人工知能 / 深層学習 |
Outline of Final Research Achievements |
Pathologists must histologically evaluate the effect of Neoadjuvant therapies (NAT) with resected specimens. Major pathological response (MPR) has recently been proposed for the evaluation; however, poor reproducibility is often noted. The aim of this study was to develop a deep learning (DL) model to predict MPR from digital images and to validate its utility for clinical use. We collected data on 125 non-small cell lung carcinomas resected after NAT and estimated MPR using an original DL model which we previously developed. In cross-validation, accuracy and mean F1 score were over 0.800. During testing, accuracy and mean F1 score were over 0.943. The areas under the receiver operating characteristic curve were over 0.978. The disease-free survival based on MPR predicted by the DL-based model was almost identical to that by pathologists. The DL model may support pathologist evaluations and can offer accurate determinations of MPR in patients.
|
Free Research Field |
診断病理
|
Academic Significance and Societal Importance of the Research Achievements |
近年、非小細胞肺癌は従来の白金製剤を主体とした化学療法に加え,分子標的治療薬(TC)や免疫チェックポイント阻害剤(ICI)など治療オプションが増えた癌腫でもある。その評価は治療後の切除病理標本における残存腫瘍の病理組織学的評価となりつつある。一方で病理医の負担は増し,また標準化も進んでいるとはいえない。今回我々が開発したDLモデルは独自性が高く,それを解消するツールとして実臨床において大きな成果をもたらす可能性があり,学術的意義は高いものと考える。また,TCやICIは多くの癌腫で利用され始めており,開発したDLモデルは癌腫を超え展開することのできる可能性を秘め、社会的意義は高いものと思われる。
|