Tumor-naïve ctDNA detection with deep learning-enhanced error suppression for sensitive mutation calling.

AI / informaticsGenome medicine· Jun 2026

Tumor-naïve ctDNA detection with deep learning-enhanced error suppression for sensitive mutation calling.

Shaya Akbarinejad, Sarah Doppler, Jos de Graaf … David Weber

Read paper

Molecular Genetic PathologyAI / informaticsDiagnosisQ1IF 10.6

New tumor-naïve ctDNA pipeline achieved 100% sensitivity detecting pre-surgical colorectal cancer

Tumor-naïve ctDNA detection with deep learning-enhanced error suppression for sensitive mutation calling.

Shaya Akbarinejad … David Weber

Genome medicine · Jun 2026 · Q1 in Genetics

Background

Circulating tumor DNA (ctDNA) detection offers minimally invasive monitoring of cancer from blood samples.

Methods

Here, we developed DEEPctMUT, a computational pipeline for tumor-naïve ctDNA detection that integrates unique molecular identifiers (unique molecular identifiers) with three complementary error-polishing strategies: (1) machine learning-based sequencing error suppression, (2) deep learning-based background noise filtering (DeepES), and (3) removal of clonal hematopoietic and germline variants using patient matched PBMC.

Results

the pipeline caught every pre-surgical CRC case, missing none

DEEPctMUT

100%

Roche Avenio Surveilla

50%

More results

DEEPctMUT accurately detected mutations down to 0.03% variant allele frequency (variant allele frequency), outperforming other tumor-naïve methods.

Furthermore, a panel-independent version of DEEPctMUT could improve the performance of Roche Avenio data without additional assay-specific training.

More results

We provide an extensive ground-truth dataset and make DEEPctMUT, including all trained models, available as an easy-to-use Nextflow pipeline.

“

Conclusion

By using advanced computational error-polishing techniques, DEEPctMUT can substantially reduce technical artifacts, allowing our tumor-naïve method to achieve a sensitivity comparable to tumor-informed approaches.

Tumor-naïve ctDNA detection with deep learning-enhanced error suppression for sensitive mutation calling.

Tumor-naïve ctDNA detection with deep learning-enhanced error suppression for sensitive mutation calling.

Related papers