Clarification of sequencing errors in all patients who underwent cancer gene panel testing in Japan

2023 Fiscal Year Final Research Report

• Project/Area Number: 22K15571
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 50020:Tumor diagnostics and therapeutics-related
• Research Institution: Tokyo Metropolitan Hospital Organization Tokyo Metropolitan Komagome Hospital
• Principal Investigator: Ikegami Masachika 地方独立行政法人東京都立病院機構東京都立駒込病院(臨床研究室), 骨軟部腫瘍科, 医長 (90830521)
• Project Period (FY): 2022-04-01 – 2024-03-31
• Keywords: がんゲノムプロファイリング / FFPEアーチファクト

Outline of Final Research Achievements

We analyzed sequencing data from the development phase of the NCC OncoPanel and showed that 40% of the mutations called by the pipeline are FFPE artifacts, and that a small number of FFPE artifacts are annotated with pathological mutations. Furthermore, we created a machine learning algorithm to predict FFPE artifacts for the large number of mutations reported in the AACR project GENIE, and confirmed that the error frequency differs among gene panels. Although the study is not yet complete due to the delay in the release of raw data from C-CAT, the implementation system is in place, and we aim to clarify the necessity and significance of FFPE artifact removal and implement it in clinical practice in the future.

Free Research Field

がんゲノム医療

Academic Significance and Societal Importance of the Research Achievements

がんゲノム医療において、検出された変異が真の変異なのかエラーなのか不確実な状態では、仮に遺伝子変異にマッチした治療が存在したとしても実施に踏み切ることに心理的抵抗が生じ、また無意味な治療が行われるリスクもある。本研究はNCCオンコパネルの開発時点のデータではあるが治療選択に影響を与えうるエラーがあることを明らかとし、エラー予測アルゴリズムの臨床実装の必要があることが確認できた。