Escribir un comentario

PREreview of “High-Concordance Validation of Droplet Digital PCR and Next-Generation Sequencing for EGFR Mutation Detection Across Diverse Biospecimens in a Large-scale NSCLC Cohort Study”

Short summary of the research and contribution to the field

This preprint evaluates the concordance between droplet digital PCR (ddPCR) and next-generation sequencing (NGS) for detecting clinically important EGFR mutations in non-small cell lung cancer (NSCLC), including L858R, exon 19 deletions, and T790M. The authors report a large retrospective comparison across approximately 1,000 EGFR-positive samples from diverse biospecimen types, including cfDNA, FFPE-derived DNA, fresh tumor tissue DNA, pleural effusion DNA, and corresponding pre-capture NGS libraries.

The main findings show very high overall mutation detection concordance between ddPCR and NGS, strong VAF correlation, and strong agreement between original DNA samples and their paired pre-capture NGS libraries. The study is clinically relevant because NSCLC molecular testing often depends on limited tissue, liquid biopsy, or residual library material. If validated carefully, the finding that pre-capture NGS libraries can serve as substitute material may reduce repeat biopsies, preserve limited samples, and support additional testing when original DNA is insufficient.

Overall, this work contributes to the field by providing practical evidence for integrating NGS multiplexing and ddPCR sensitivity in EGFR testing workflows and by exploring the potential utility of pre-capture NGS libraries as an alternative analyte source in clinical molecular diagnostics.

Positive feedback / strengths

1. Clinically important question. The study addresses a real challenge in NSCLC testing: how to reconcile EGFR mutation detection across tissue, liquid biopsy, pleural effusion, and limited residual sample material.

2. Large sample size. A retrospective cohort of approximately 1,000 EGFR-positive samples is a major strength and provides a broader evidence base than many smaller concordance studies.

3. Multiple biospecimen types. Including cfDNA, FFPE DNA, fresh tumor tissue DNA, pleural effusion DNA, and pre-capture libraries makes the study highly relevant to real-world diagnostic workflows.

4. Focus on both detection and VAF quantification. Evaluating both mutation concordance and VAF correlation is useful because clinical laboratories often need to understand not only whether a mutation is detected, but also how quantitative results compare between platforms.

5. Practical value of pre-capture library testing. The investigation of pre-capture NGS libraries as substitute material is especially valuable for cases with limited FFPE tissue or insufficient remaining DNA.

Major issues

1. The study appears enriched for EGFR-positive samples, so specificity and false-positive performance need clarification

The abstract states that the study used approximately 1,000 EGFR-positive samples. This is useful for concordance among positive cases, but it may not fully evaluate specificity, false-positive rate, or clinical performance in EGFR-negative samples.

Suggested improvement: The authors should include or clearly describe an EGFR-negative control cohort across relevant biospecimen types. Reporting positive percent agreement, negative percent agreement, overall percent agreement, sensitivity, specificity, and false-positive/false-negative cases would make the clinical interpretation stronger.

2. Sample pairing and specimen distribution need clearer explanation

Because the study includes many biospecimen types, it is important to know how many samples were available in each category and how many were truly paired.

Suggested improvement: The authors should provide a clear sample-flow diagram and table showing:

• number of samples per specimen type

• number of paired original DNA and pre-capture library samples

• number of cfDNA/tissue-matched pairs

• number of samples per mutation subtype

• DNA input amounts and quality metrics

• clinical stage/treatment context where available

This would help readers understand whether the reported concordance reflects broad sample-type performance or is driven mainly by one or two dominant specimen categories.

3. ddPCR limit-of-detection claims require more validation detail

The reported ddPCR LOD of 0.01% at 100 ng input DNA is very sensitive and clinically meaningful, but it requires careful explanation.

Suggested improvement: The authors should describe:

• number of replicates tested at each allele fraction

• DNA input copy number assumptions

• false-positive droplet thresholds

• limit of blank and limit of detection calculations

• acceptance criteria for positive calls

• performance near the LOD

• whether LOD was separately validated for L858R, Ex19del, and T790M

This is especially important for very low-VAF clinical interpretation.

4. Exon 19 deletion assay design should be described in more detail

EGFR exon 19 deletions are heterogeneous, with multiple deletion breakpoints and sequence contexts. A single ddPCR assay may not detect all clinically relevant exon 19 deletion variants equally.

Suggested improvement: The authors should clarify which Ex19del variants are detected by the ddPCR assay, whether the assay covers common and uncommon deletion types, and how discordant Ex19del calls were handled. A table of Ex19del subtypes detected by NGS versus ddPCR would be very helpful.

5. NGS assay details and variant-calling thresholds need more transparency

The validity of concordance depends heavily on NGS assay design and bioinformatics thresholds.

Suggested improvement: The manuscript should provide details on:

• NGS panel content and EGFR coverage

• sequencing depth across sample types

• molecular barcode/UMI use, if applicable

• minimum VAF threshold for calling variants

• minimum unique read depth

• variant-calling pipeline

• quality-control metrics

• criteria for reportable positive results

Without this information, it is difficult to judge whether discordance reflects platform differences, bioinformatics thresholds, or sample limitations.

6. VAF correlation should not be interpreted as full quantitative equivalence without bias analysis

The reported Pearson correlation is strong, but correlation alone does not prove quantitative agreement. Two methods can correlate strongly while still having systematic bias.

Suggested improvement: The authors should include additional agreement analyses, such as:

• Bland–Altman plots

• Passing–Bablok or Deming regression

• mean bias and limits of agreement

• VAF comparison by low, medium, and high allele fraction groups

• separate VAF analysis by specimen type and mutation subtype

This would make the quantitative claims more robust.

7. Pre-capture library substitution is promising but requires more workflow-specific validation

The finding that pre-capture NGS libraries correlate strongly with source DNA is important, but pre-capture libraries may introduce bias through PCR amplification, library complexity differences, fragmentation patterns, and storage conditions.

Suggested improvement: The authors should clarify:

• library preparation method

• number of PCR cycles before capture

• library input amount used for ddPCR or re-testing

• storage duration and freeze-thaw history

• whether library concentration/quality affected concordance

• whether results remained accurate at low VAF

• whether this approach works across all specimen types or only high-quality libraries

The authors should also clearly state whether pre-capture libraries are being proposed as a routine clinical substitute or as a rescue strategy when source DNA is exhausted.

8. Discordant cases should be deeply analyzed

With high concordance, discordant cases become especially informative.

Suggested improvement: The authors should include a dedicated discordance analysis showing:

• mutation type

• specimen type

• VAF by each method

• DNA input and QC metrics

• NGS read depth

• ddPCR droplet counts

• whether orthogonal testing resolved the discrepancy

• likely explanation: low VAF, sample degradation, tumor heterogeneity, ctDNA shedding, technical threshold, or biological discordance

This would strengthen confidence in the conclusions.

Minor issues

1. Clarify terminology. Terms such as cfDNA-prePCR, ffpeDNA-prePCR, ttDNA-prePCR, and peDNA-prePCR should be clearly defined early and used consistently.

2. Avoid overgeneralizing “clinical testing” claims. The manuscript should distinguish analytical concordance from clinical utility, especially if patient outcome data are not included.

3. Add a workflow diagram. A figure showing original sample → DNA extraction → ddPCR / NGS → pre-capture library → repeat testing would improve readability.

4. Report mutation subtype-specific sample numbers. Concordance percentages should be accompanied by numerator/denominator values for L858R, Ex19del, and T790M.

5. Clarify pleural effusion sample handling. Pleural effusion supernatant DNA can vary substantially in tumor fraction and background DNA. Extraction and QC details should be included.

6. Discuss tissue–liquid discordance as both technical and biological. Differences between tissue and cfDNA may reflect tumor heterogeneity, disease burden, treatment timing, and ctDNA shedding—not only assay performance.

7. Include confidence intervals. Concordance rates, correlation coefficients, and LOD estimates should include confidence intervals where possible.

8. Clarify retrospective cohort limitations. The authors should discuss selection bias, sample availability bias, and whether only successful clinical samples were included.

9. Add practical implementation guidance. If pre-capture libraries are recommended as substitute samples, the manuscript should specify minimum library concentration, quality, storage conditions, and acceptable use cases.

10. Clarify regulatory status. If the assay is intended for research or clinical validation only, the manuscript should clearly state this and avoid implying broad clinical adoption without further validation.

Overall assessment

This is a clinically relevant and practically useful study addressing an important challenge in NSCLC molecular diagnostics: how to ensure reliable EGFR mutation detection and VAF interpretation across NGS, ddPCR, and diverse biospecimen types. The large sample size, inclusion of multiple mutation types, and evaluation of pre-capture NGS libraries are major strengths.

The most important areas for improvement are clarification of cohort structure, inclusion of EGFR-negative controls, deeper LOD validation, more transparent NGS methods, bias analysis beyond Pearson correlation, and detailed review of discordant cases. The pre-capture library findings are particularly promising, but they should be framed carefully as a validated rescue or supplemental strategy unless broader prospective clinical implementation data are available.

With these additions, the study would provide a stronger and more actionable framework for integrated EGFR testing using ddPCR, NGS, and residual pre-capture library material in NSCLC diagnostics.

Competing interests

The author declares that they have no competing interests.

Use of Artificial Intelligence (AI)

The author declares that they used generative AI to come up with new ideas for their review.