Computational optimisation of targeted DNA sequencing for cancer detection

Area of Science:

• Oncology
• Genomics
• Molecular Diagnostics

Background:

• Personalized cancer medicine faces challenges from intra-tumor heterogeneity and drug resistance.
• Early diagnosis is crucial for improving survival rates in advanced metastatic cancer patients.
• Circulating tumor DNA (ctDNA) analysis offers a potential non-invasive method for early cancer detection.

Purpose of the Study:

• To define reduced gene panels for assessing and optimizing targeted ctDNA scans for early tumor detection.
• To evaluate the feasibility of using a limited gene set for broad cancer mutation screening.

Main Methods:

• Utilized publicly available datasets comprising 4,467 samples.
• Divided samples into discovery and two independent validation cohorts.
• Analyzed mutation presence within a 25-gene panel across various cancer types.

Main Results:

• Identified at least one mutation in the 25-gene panel in up to 76% of 10 cancer types.
• Demonstrated high sensitivity of the panel across most tumor types.
• Found that targeting "hotspot" regions introduces biases and compromises reproducibility.

Conclusions:

• A targeted 25-gene panel can effectively detect mutations for early cancer detection via ctDNA analysis.
• This approach shows potential for improving early diagnosis and personalized medicine.
• Avoiding "hotspot" biases is critical for reproducible ctDNA-based tumor detection.