Filigree pattern - A new concept for lung carcinoma pathology

2023 Fiscal Year Final Research Report

• Project/Area Number: 20K16180
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 49020:Human pathology-related
• Research Institution: National Hospital Organization Saitama National Hospital (2023); Keio University (2020-2022)
• Principal Investigator: Emoto Katsura 独立行政法人国立病院機構埼玉病院(臨床研究部), 診療部, 病理診断医師 (40570859)
• Project Period (FY): 2020-04-01 – 2024-03-31
• Keywords: Adenocarcinoma / Lung cancer / Filigree pattern / Histological subtype

Outline of Final Research Achievements

In this study, we have established artificial intelligence (AI) for seven subtypes of lung adenocarcinoma (LUAD), including the concept of filigree pattern. In the procedure, we have revealed the importance of color standardization and picture size for AI. LUAD histological grades by WHO 4th and 5th editions indicated the similar tendency to the previous studies, however, the power of WHO 4th grading system was a little higher than WHO 5th in the present study. Focusing on tumor with 5-20% of high-grade component, tumors with centrally located high-grade components had a higher malignant potential. Our result suggests that histological subtype including the concept of filigree pattern is important, of not only its proportion but also its distribution.

Free Research Field

Thoracic pathology

Academic Significance and Societal Importance of the Research Achievements

Filigree patternは肺癌WHO分類第5版から採用された新しい概念であるが、その概念を含む肺腺癌組織亜型を分類可能なAIを設計した。肺腺癌亜型分類は複雑で、本研究でも7種類を対象としたが、病理医に近い分類が可能であった。組織亜型はその割合のみが予後予測に用いられているが、本研究ではAIを補助的に使用することでヒトが作業・分類するのが難しい組織亜型の分布を可視化することに成功し、組織亜型の分布も予後に影響し得ることを示した。本研究により、filigree patternを含む組織亜型の臨床的・研究的意義が広がったと考えられる。