Hanbaek Yi; Jeonghwan Youk; Yoojoo Lim; Hanseong Roh; Dongsoo Kyung; Hwang-Phill Kim; Duhee Bang; Bhumsuk Keam; Tae-Min Kim; Miso Kim; Dong-Wan Kim; Tae-You Kim

doi:10.4143/crt.2023.1294

Articles

Page Path: HOME > Cancer Res Treat > Volume 56(3); 2024 > Article

Original Article Lung and Thoracic cancer Analytical and Clinical Validation of a Highly Sensitive NGS-Based ctDNA Assay with Real-World Concordance in Non–Small Cell Lung Cancer: Hanbaek Yi¹, Jeonghwan Youk^1,2, Yoojoo Lim³, Hanseong Roh³, Dongsoo Kyung³, Hwang-Phill Kim³, Duhee Bang⁴, Bhumsuk Keam^1,2, Tae-Min Kim^1,2, Miso Kim^1,2,, Dong-Wan Kim^1,2,, Tae-You Kim^1,2,3; Cancer Research and Treatment : Official Journal of Korean Cancer Association 2024;56(3):765-773.
DOI: https://doi.org/10.4143/crt.2023.1294
Published online: January 8, 2024

¹Department of Internal Medicine, Seoul National University Hospital, Seoul National University College of Medicine, Seoul, Korea

²Cancer Research Institute, Seoul National University, Seoul, Korea

³IMBdx, Seoul, Korea

⁴Department of Chemistry, Yonsei University, Seoul, Korea

Correspondence: Miso Kim, Department of Internal Medicine, Seoul National University Hospital, 101 Daehak-ro, Jongno-gu, Seoul 03080, Korea
Tel: 82-2-2072-4035 Fax: 82-2-2072-7379 E-mail: misokim@snu.ac.kr

Co-correspondence: Dong-Wan Kim, Department of Internal Medicine, Seoul National University Hospital, 101 Daehak-ro, Jongno-gu, Seoul 03080, Korea
Tel: 82-2-2072-2995 Fax: 82-2-6072-5336 E-mail: kimdw@snu.ac.kr

• Received: December 7, 2023 • Accepted: January 7, 2024

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

5,405 Views
359 Download
2 Web of Science
3 Crossref
2 Scopus

prev next

Full Article

Download PDF

Abstract

Purpose
There have been needs to improve the sensitivity of liquid biopsy. This report aims to report the analytical and clinical validation of a next-generation sequencing (NGS)–based circulating tumor DNA (ctDNA) assay.
Materials and Methods
Analytical validation was conducted in vitro by evaluating the limit of detection (LOD), precision, and specificity for various genomic aberrations. The real-world performance in non–small cell lung cancer (NSCLC) was assessed by comparing the results of AlphaLiquid100 to the tissue-based results.
Results
The LODs with 30 ng input DNA were 0.11%, 0.11%, 0.06%, 0.21%, and 2.13 copies for detecting single nucleotide variants, insertions, deletions, fusions, and copy number alterations (CNA), respectively. Quantitatively, single nucleotide variants/insertions and deletions, fusions, and CNAs showed a good correlation (R²=0.91, 0.40, and 0.65; y=0.95, 1.06, and 1.19) to the manufacturer’s values, and per-base specificities for all types of variants were near 100%. In real-world NSCLC (n=122), key actionable mutations in NSCLC were detected in 60.7% (74/122) with the ctDNA assay. Comparative analysis against the NGS-based tissue results for all key mutations showed positive percent agreement (PPA) of 85.3%. For individual genes, the PPA was as high as 95.7% for epidermal growth factor receptor (EGFR) mutations and 83.3% for ALK translocations. AlphaLiquid100 detected drug-sensitive EGFR mutation at a variant allele frequency as low as 0.02% and also identified an EGFR mutation in a case where tissue sample missed. Blood samples collected post-targeted therapies revealed additional acquired mutations.
Conclusion
The AlphaLiquid100 ctDNA assay demonstrates robust analytical validity, offering clinically important information for NSCLC patients.
Key words: Liquid biopsy, Circulating tumor DNA, Analytical validity, Actionable mutations, Non–small cell lung carcinoma

Introduction

Liquid biopsy, which uses circulating tumor DNA (ctDNA) as a source, is increasingly becoming a practical option in obtaining tumor mutational profiles with recent advancements in related technologies [1,2]. Although tumor tissue remains the gold standard specimen for genomic analysis, it inherently presents several limitations. These include the costs and risks associated with the acquisition procedure, the potential of neglecting the spatial heterogeneity of the tumors, and the difficulty of repeating the analysis as needed. In contrast, liquid biopsy provides a minimally invasive alternative, facilitating the genomic profiling of tumors without the need for additional invasive procedures. Consequently, this approach can not only enable genomic analysis in patients for whom traditional tissue-based methods may have been unfeasible but also expedite the molecular assessment process for all patients, including those suitable for traditional methods [3,4]. Furthermore, it enables repeat testing throughout the disease course, and the analysis of ctDNA gathered from the bloodstream can potentially provide a more comprehensive genomic status in terms of spatial coverage compared to the analysis conducted using focal tissue biopsies. Current guidelines from groups such as the National Comprehensive Cancer Network (NCCN) [5] or the International Association for the Study of Lung Cancer (IASLC) [6] have started to encourage the use of liquid biopsy, especially to overcome the limitations of traditional tissue-based approaches.

As the understanding of cancer biology expands and targeted therapies continue to develop, the number of targets requiring evaluation is also increasing. The next-generation sequencing (NGS)–based approaches can be more suitable in genomic profiling as it can cover broad genomic variations and even potentially decrease the time and costs needed to understand a patient’s tumor status. Moreover, the decreasing cost and increasing accuracy of NGS-based sequencing suggest that it will become increasingly important in the future. However, ctDNA analysis using the NGS-based approaches requires highly precise techniques, as they have to derive accurate results from the limited amounts of DNA present in such samples. The adaptation of NGS-based liquid biopsy in real-world clinical settings has been limited due to concerns regarding the lack of sensitivity and sufficient validation data. Therefore, the accumulation of both analytical and clinical validation data is required.

The aim of this report is to assess the analytical validity of a ctDNA assay based on NGS technology, which has been developed for sensitive detection of trace amounts of ctDNA in blood. It also aims to demonstrate the assay’s performance in real-world samples of advanced non–small cell lung cancer (NSCLC).

Materials and Methods

1. AlphaLiquid100 design and workflow

AlphaLiquid100 is a comprehensive genomic profiling assay designed for advanced solid cancer patients, capable of detecting 118 genes using ctDNA isolated from plasma samples. The assay has been specifically designed to suppress errors by incorporating a proprietary High-Quality unique Sequence (HQS) technology, which has evolved the unique molecular identifier (UMI) method and context-based background error suppression. It detects single nucleotide variants (SNVs) and short insertions and deletions (INDELs) in full coding exons of 117 genes, spanning 321 kb of genome. In addition, it analyzes fusions in 10 genes by examining intronic regions, copy number alterations (CNA) in 42 genes, and estimates the microsatellite instability (MSI) status (S1 Table).

The workflow of AlphaLiquid100 is outlined in S2 Fig. cfDNA (cell-free DNA) is extracted from plasma, which has been separated from stabilized whole blood. UMI barcodes are added to both ends of cfDNA, to distinguish duplicate reads for sequencing error correction during the subsequent bioinformatics analysis steps. The barcoded DNA undergoes polymerase chain reaction (PCR) amplification and are enriched using hybridization capture, followed by sequencing at a depth of more than 50,000×. Using the proprietary HQS technology, the assay precisely detects SNVs and INDELs, while also identifying fusions, CNAs, and MSI.

2. Blood sample processing and sequencing

Each blood sample was collected using dedicated cfDNA bottles and was centrifuged with Ficoll solution at 1,500 ×g for 15 minutes. Plasma was separated by centrifugation at 16,000 ×g for 10 minutes to remove cell debris. cfDNA is isolated according to the manufacturer’s instructions from 2 to 4 mL plasma using a Maxwell RSC cfDNA Plasma Kit (Promega, Madison, WI) and quantified using 4200 TapeStation (Agilent Technologies, Santa Clara, CA). Peripheral blood mononuclear cells (PBMC) were separated following this protocol. Genomic DNA was isolated from PBMC using Maxwell RSC Blood DNA Kit (Promega). All experiments were processed in accordance with the guidelines meeting the pre-analytical conditions for analyzing cfDNA. Genomic DNA from tumor tissue samples was extracted using Maxwell RSC FFPE DNA Kit (Promega) for FFPE samples and Maxwell RSC Tissue DNA Kit (Promega) for fresh-frozen tissues. Following the extraction steps, NGS libraries were constructed using the IMBdx NGS DNA Library Prep Kit. Solution-based target enrichment was performed at IMBdx, Inc. (Seoul, Korea), using the AlphaLiquid100 target capture panel. Captured DNA libraries were sequenced using Illumina NovaSeq 6000 platform (Illumina, San Diego, CA) in 2×150 bp paired-end mode.

3. Bioinformatics processing

All sequencing reads were demultiplexed and converted into unmapped bam files that store extracted UMI sequences. Unmapped bam files were aligned to the reference genome (GRCh38) using Burrows-Wheeler Aligner MEM algorithm (0.7.17-r1188) [7]. Reads mapped on the same position and UMI that resulted in PCR duplicates were grouped as a family and were collapsed consensus sequence using in-house scripts.

The SNV and INDEL variants were detected using the deepblood (IMBdx proprietary software) that discriminates noise using the HQS technology. The HQS technology incorporates a pre-computed background error rate of every family size group for both SNVs and INDELs. Mutations occurring in genes frequently involved in clonal hematopoiesis of indeterminate potential (CHIP) (ATM, CBL, CHEK2, IDH2, JAK2, MPL, and U2AF1) were excluded if the variants detected were with variant allele fractions (VAFs) of below 0.1% or were classified as variants of uncertain significance, as these were considered to be associated with CHIP. Additionally, variants with VAF ranging between 40%-60% were considered germline and were consequently excluded.

Fusion variants were detected in three steps. First, the supporting reads were extracted from split read searching and k-mer mapping. Next, the supporting reads were mapped on the reference genome. Finally, high-confident fusion genes were selected if they have two or more reads with a mapping quality of 60, and if the predicted transcript is expected to be functional.

For the determination of CNA, the log2 ratios and p-values were calculated by comparing a sample to the panel of normal (PON) prepared independently and previous to this study. The log2 ratio was then converted to copy number (CN) values. CN values greater than 2.2 with p-values less than 0.001 were reported as CN amplification. Additionally, the allele frequency (AF) of single nucleotide polymorphisms was evaluated for any potential bias and utilized to confirm a CNA event.

To detect MSI, the proportion of deletion was calculated on the conventional MSI markers (BAT25, BAT26, D17S2720, D2S123, D5S346, NR21, NR24, and NR27) and 426 predicted unstable loci, and compared to the PON from healthy donors. If any markers were more than 3 standard deviations from the PON, it was determined to be unstable. More than two conventional MSI makers or more than 100 predicted unstable loci are considered MSI-high.

4. Preparation of standard reference materials for analytical validation

The analytical performance of SNV, INDEL, fusion, and CNA detection was evaluated using the Seraseq ctDNA Complete Mutation Mix samples with AF levels of 0.05%, 0.1%, 0.5%, and 1%. To prepare the sample with AF 0.05%, the material with AF 0.1% was diluted with the Seraseq ctDNA Complete Mutation Mix WT sample. The precision of SNV, INDEL, CNA, and fusion detection were evaluated in terms of coefficients of variability (CV) for 12 SNVs (ALK E17K, ALK G1202R, ALK F1174L, BRAF V600E, EGFR L858R, EGFR T790M, KIT D816V, KRAS G12C, KRAS G12D, KRAS Q61H, NRAS Q61R, PIK3CA H1047R), four deletions (BRCA2 R2645fs*3, EGFR L747_P753>S, EGFR S752_I759delSPKANKEI, EGFR E746_A750delELREA), two insertions (ERBB2 A775_G776insYVMA and PIK3CA N1068fs*4), three CNAs (ERBB2 amplification, MET amplification, and MYC amplification), and three fusions (NCOA4-RET, EML4-ALK, and CD74-ROS1). BRCA1 c.1961delA, which is included in the Seraseq ctDNA Complete Mutation Mix was excluded from the analysis as it is located on an 8-mer A homopolymer. For MSI detection, Seraseq MSI Reference Panel Mix samples with AF of 0.5%, 1%, and 5% were used. The samples with AF of 0.5% and 1% were prepared by diluting the AF 5% material using the Seraseq MSI Matched WT.

5. Estimation of analytical validity measures using the standard reference materials

To determine the sensitivity and the limit of detection (LOD) of SNVs, INDELs, fusions, and CNAs on both the DNA input amount and the AF, standard reference materials at varying AF were tested five times each, using input DNA amounts of 2 ng, 5 ng, 10 ng, 20 ng, and 30 ng. The LOD of MSI were tested five times per AF, using the MSI reference materials at input DNA amounts of 20 ng and 30 ng. The LOD was determined for each combination of DNA input amount and AF, as well as for each variant type. The 95% LOD was established empirically based on the sensitivity of each condition or through probit regression analysis.

Repeatability (intra-assay variation) and reproducibility (inter-assay variation) were evaluated by calculating CVs for short variants, fusions, and CNAs, and by calculating detection rate for MSI across different conditions. Seraseq ctDNA Complete Mutation Mix with an input DNA quantity of 20 ng at AF levels of 1%, 0.5%, and 0.1% was used to test SNVs, INDELs, fusions, and CNAs, and Seraseq MSI Reference Panel Mix with an input DNA quantity of 20 ng at AF levels of 5%, 1%, and 0.5% was used to test MSI. All testing results were compared to the manufacturer’s measured values to obtain the precision measures.

To validate the specificity level, blood samples from 50 healthy donors were tested, by comparing the sequencing results of cfDNA to PBMC. Additionally, triplicate tests were performed on NA12878, NA12891, and NA12892, respectively. Somatic mutations previously identified in these cell lines through another targeted sequencing panel were used as a comparator. The per-base specificity for SNV, deletion, and insertion were determined by dividing the total number of detected mutations by the cumulative target region of 18.96 million bp, which corresponds to 59 samples at coding sequence region (321,358 bp) per sample.

6. Patient samples and tumor tissue results

Blood samples were collected from 50 healthy donors and from patients with metastatic or locally-advanced unresectable NSCLC patients, who were planned for or are receiving systemic therapy at Seoul National University Hospital (SNUH, Seoul, Korea). A single collection of 20 mL of whole blood was collected in dedicated tube for cfDNA analysis. The genomic analysis results of tumor tissue samples from the corresponding patients, performed as a part of standard-of-care testing at SNUH, were collected as available regardless of method. The genomic analysis includes direct DNA sequencing of EGFR or immune histochemistry (IHC) or in situ hybridization for ALK or ROS1. Additionally, the results of targeted gene panel sequencing of tissue samples with NGS platform (FIRST-LCP) [8] were collected if performed. Key actionable mutations for this analysis were defined as alterations in genes for which corresponding drugs exist, either approved or under clinical trial. These included known mutations in EGFR, ALK, ROS1, BRAF V600E, RET, or NTRK, and additionally KRAS G12C, FGFR3 or HER2 mutations [9]. Comparative analysis between the tissue-based NGS findings and the AlphaLiquid100 assay results were performed based on the presence of these mutations.

Results

1. Sensitivity and LOD using standard materials

The LODs for identifying various genomic aberrations using AlphaLiquid100, were evaluated across a spectrum of input DNA quantities, ranging from 2 ng to 30 ng (Table 1). First, the LOD for SNV detection using AlphaLiquid100, was 0.11% at 30 ng of input DNA, and gradually increased with the decrease in the amount of input DNA, to 1.01% with the input DNA of 2 ng. When insertion, deletion, fusion, and CNA were tested in a similar manner, the LODs were observed to be 0.11%, 0.06%, 0.21%, and 2.13 CN respectively, at an input DNA level of 30 ng and increased with the change of input DNA in similar manner. The MSI test correctly detected MSIhigh status in 100% of samples under conditions of AF 1% and 5%. However, the detection rate decreased to 60% under AF 0.5% in 30 ng input DNA conditions.

2. Precision and specificity using standard materials

The average CVs obtained by comparing the sequencing results to the manufacturer’s values of testing materials, were 16.7%, 21.8%, and 37.9% for short variants, 24.9% 31.3%, and 31.4% for fusions, and 7.6%, 9.9% and 9.9% for CNAs, each at 1%, 0.5%, and 0.1% of AFs, respectively (S3 Table). Short variants, fusions, and CNAs showed good correlation (R²=0.91, 0.40, and 0.65; y=0.95, 1.06, and 1.19, respectively). The assay consistently detected MSI-high correctly at AF above 1% (S4 Table).

For specificity calculation, two SNVs from cell lines and eight SNVs and one deletion from healthy donors were discovered, which were not present in the respective comparator of each. As a result, per-base specificity was calculated at 99.999947%, 99.99999%, and 100% for detection of SNVs, deletions and other types of mutations, respectively.

3. Concordance analysis of AlphaLiquid100 compared to tissue NGS results in real-world NSCLC samples

Overall, results of AlphaLiquid100 from 122 advanced NSCLC patients were available for analysis. Blood samples were drawn before the initiation of any systemic therapies in 59 patients (48.4%), while the samples were obtained during or after any systemic therapies in 63 patients (51.6%). The patient characteristics are summarized in Table 2. By AlphaLiquid100, 74 of all 122 samples (60.7%) were found to harbor any of the key actionable mutations. The most frequent of these mutations was in the EGFR gene found in 57 patients (46.7%), followed by EML4-ALK translocations in seven patients (5.7%). Other less frequent mutations detected included MET exon 14 skipping mutations in two patients, ROS1 fusion in one patient, and KIF5B-RET fusion in one patient (Fig. 1). Notably, eight of the key actionable mutations identified by the ctDNA assay were detected at a VAF of 0.1% or less, with the lowest VAF at 0.02% for an EGFR exon19del.

The general concordance between ctDNA and tissue-based assessments was analyzed using data of the patients whose blood samples were collected before the initiation of systemic therapies and also had the NGS-based panel sequencing results for their tissue samples. Of the 59 patients having their blood samples drawn before systemic therapies, tissue NGS test results were available for comparative analysis in 50 patients. The tissue-based NGS testing was attempted in two additional patients but were unsuccessful due to insufficient tissue specimen volume. In 50 patients selected for concordance analysis, 36 key actionable mutations were detected overall, of which 29 were confirmed to show consistent results between tissue and blood. The blood analysis identified two additional mutations (FGFR3 R248C and BRAF V600E) that were not detected by the tissue testing. However, it did not identify five mutations that were detected from tissue testing. Consequently, the positive percent agreement (PPA) of AlphaLiquid100 in detecting the key actionable mutations from blood samples was 85.3% and the positive predictive value (PPV) was 93.5% (Table 3).

Comparative analysis between tissue and blood-based assay results were also carried out by individual genes (Table 4). The ctDNA assessment detected all EGFR mutations detected by the tissue NGS method with the exception of one case, yielding a PPA of 95.7% and a PPV of 100%. Notably, there was one case whose EGFR mutation that is concordant with the tissue result was detected from the blood sample with the VAF as low as 0.05% (Table 5). Both this case and the case that blood assay failed to detect the mutation had intrathoracic diseases only. When the comparison was extended to include six samples that were excluded from the general concordance analysis because it lacked tissue NGS test results but were collected before treatment initiation and had EGFR mutation status determined by PCR analysis, one additional EGFR mutation (L858R) was found from the blood of a patient whose tissue analysis had returned a negative result. This patient was subsequently treated with Osimertinib and remained under treatment for 10 months as of the most recent follow-up demonstrating the best response of stable disease.

Similar comparative analysis was conducted for ALK translocations. Overall, the ctDNA assay was able to detect three out of four cases identified by the tissue NGS method, yielding a PPA of 75.0%. When the analysis was extended to include patients having ALK assessments by methods other than NGS, two additional EML4-ALK translocations that were concordant between blood and tumor were added, to yield a PPA of 83.3%. The two added cases were confirmed from IHC, but tissue NGS analysis was not feasible due to small tumor volume.

4. Detection of resistance mutations using AlphaLiquid100

Among 40 patients exhibiting EGFR mutations as determined by tissue assays and having undergone treatment with EGFR-targeted therapies before the blood draw, 15 cases were found to have acquired mutations known to be related to treatment resistance by the ctDNA assay. These resistance mutations include T790M in 11 patients, C797X mutations in six, and other mutations (specifically in RET and PIK3CA genes) in two patients. Similarly, two among the five patients with known ALK translocation and treated with crizotinib and/or alectinib before the blood sampling exhibited mutations of ALK G1202R by the ctDNA assay, and one patient with CD74-ROS1 fusion who was treated with entrectinib before blood sampling exhibited additional V2089M mutation from the ctDNA assay. We successfully constructed Ba/F3 cells expressing the cDNA encoding CD74-ROS1 V2089M through targeted mutagenesis in CD74-ROS1–expressing Ba/F3 cells [11]. In vitro experiments demonstrated that these ROS1 V2089M-mutant Ba/F3 cells exhibited resistance to crizotinib, confirming the mutation’s role in drug resistance (S5 Fig.).

Discussion

In this study, we present the analytical validity of AlphaLiquid100, along with its performance compared to results obtained using tissue samples in the real-world advanced NSCLC samples. Our findings show that sensitive handling of cfDNA, which by nature can be minimal in amount, enables profiling of a patient’s tumor with good sensitivity and repeatability in vitro. Compared to tissue analysis results, the ctDNA assay was able to detect key druggable mutations in most cases. Notably, the ctDNA assay was able to detect important mutations in cases where tissue-based NGS analysis was not possible due to limited availability of tissue. In a case that showed heterogeneous EGFR mutation status between tissue and blood, the patient demonstrated a prolonged PFS after an EGFR TKI treatment.

The minimally invasive and convenient approach of liquid biopsy, combined with advancements in sensitive mutation detection technologies, is driving its increased adoption in clinical practice [1,5,6,12,13]. The clinical adoption is particularly pronounced in certain tumor types such as NSCLC, where the anatomical challenges complicate the acquisition of tissue samples, yet comprehensive molecular testing is becoming imperative for treatment decision making [5,6,14]. The United States Food and Drug Administration have started to approve liquid biopsies for single gene or multi-gene platforms, and the indications are expanding [15]. Current NGS methodologies enable the testing of a broader range of genes at once, thereby minimizing the volume of specimens required and potentially reducing the turnaround time needed from ordering of a test to receiving of the results [4]. Furthermore, the ease of sample acquisition and broad coverage make liquid biopsy an ideal method for monitoring the emergence of resistance mutations. Although concerns about clinical validation and sensitivity in low-burden disease patients persist, the expanding evidence increasingly supports the utility and reliability of liquid biopsy, with NGS-based approaches becoming more prominent [16].

By nature of ctDNA, successful liquid biopsy assay must be able to accurately discriminate true mutations from noise from a very small amount of DNA. In the process of sequencing low-concentration of DNA, elevating sensitivity inevitably leads to increased noise, leading to a higher chance of false positivity. Consequently, sophisticated noise reduction techniques are important in making liquid biopsy clinically viable. The method we used for ctDNA profiling was able to effectively reduce noise by building error profiles for contextual patterns for each set of consensus reads separated by family size. By leveraging HQS technology, we were able to efficiently distinguish authentic SNVs and INDELs from NGS artifacts, using a log-likelihood ratio score as a measure, rather than depending on read counts or VAFs. Such technology has enabled us to effectively eliminate noise and thus accurately identify true mutations down to 0.02% of VAF, thereby improving sensitivity.

The concordance and the sensitivity of ctDNA assays in NSCLC compared to the tissue assay have been reported in a wide range of variabilities [17-20]. For detecting EGFR mutations, generally higher sensitivity and concordances with tissue-based results have been observed, with a recent large-scale study reporting a PPA for EGFR of 78% [15]. However, structural aberrations such as ALK translocations or MET exon 14 skipping mutations have been more difficult to successfully detect using liquid biopsy. Our methodology of enhancing sensitivity has enhanced sensitivity to PPA of 95.7% for EGFR, by detecting variants at VAF as low as 0.02%. The sensitivity of the ctDNA assay in detecting important pathogenic SNVs was similar in other cancer types such as colorectal cancer and prostate cancer showing over 95% of sensitivity [21,22]. Although improved, the PPA for ALK translocations was still lower than EGFR, which is believed to be due to intrinsic characteristics of ctDNA and in trend with other reported findings. The improving sensitivity of liquid biopsy assays coupled with the minimally invasive nature of liquid biopsies, affirm the practical value for use of ctDNAs in clinics. Such advancements may broaden treatment options, especially for patients who were otherwise ineligible for molecular evaluation due to the unavailability of tissue samples for a variety of reasons such as poor general condition, anatomical challenges, or risk of invasive procedures. Strategically combining liquid biopsy with tissue testing for confirmation of structural variations could be an effective diagnostic approach in the future.

This study has limitations that only a limited number of real-world NSCLC samples were included for analysis. As the samples were collected from advanced NSCLC patients regardless of the treatment phase, even less number were available for adequate comparative analysis with tissue results. Further analysis in larger cohort may be necessary to assess the true performance in real-world and to find the practical challenges related to the assay.

In conclusion, the in vitro and real-world performance results presented in this study show that NGS-based ctDNA analysis using AlphaLiquid100 can be a reasonable option for profiling advanced NSCLC patients.

Electronic Supplementary Material

Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).

crt-2023-1294_S1_Table.xlsx

crt-2023-1294_S2_Fig.pdf

crt-2023-1294_S3_Table.xlsx

crt-2023-1294_S4_Table.xlsx

crt-2023-1294_S5_Fig.pdf

NOTES

Ethical Statement

All patients provided written informed consent before any study-specific procedures. The study protocol was approved by the Institutional Review Board of SNUH (IRB No. H-1805-049-944, H-2304-150-1416), and the study was conducted in accordance with the Declaration of Helsinki for biomedical research.

Author Contributions

Conceived and designed the analysis: Yi H, Lim Y, Bang D, Kim M, Kim DW, Kim TY.

Collected the data: Yi H, Youk J.

Contributed data or analysis tools: Roh H, Kyung D, Kim HP, Keam B, Kim TM, Kim M, Kim DW, Kim TY.

Performed the analysis: Yi H, Youk J, Lim Y, Kyung D.

Wrote the paper: Yi H, Youk J, Lim Y, Roh H, Kim M, Kim DW, Kim TY.

Conflicts of Interest

Y.L., H.R., D.K., H.-P.K. are employees of IMBdx. D.B. owns stocks of IMBdx and Celemics. T.-Y.K. is a founder of IMBdx. H.-P.K. and T.-Y. K. are inventors on patent application (no.10-2022-0038856, filed on 29 Mar 2022). T.M.K. performed consulting or advisory roles outside this work at Amgen, AstraZeneca/MedImmune, Boryung, Daiichi-Sankyo, HK inno.N, IMBdx, Janssen, Novartis, Regeneron, Roche/Genentech, Samsung Bioepis, Takeda, and Yuhan. M.K. performed consulting or advisory roles outside this work at MSD, BMS/Ono Pharmaceutical, Ipsen, Roche, Janssen, Merck, Astellas, Eisai, Bayer, Pfizer, Boehringer-Ingelheim, Boryung, and Yuhan. D.-W.K. received fundings or medical writing assistance outside this work from Alpha Biopharma, Amgen, Astrazeneca/Medimmune, Boehringer-Ingelheim, Bridge BioTherapeutics, Chong Keun Dang, Daiichi-Sankyo, GSK, Hanmi, InnoN, IQVIA, Janssen, Merck, Merus, Mirati Therapeutics, MSD, Novartis, ONO Pharmaceutical, Pfizer, Roche/Genentech, Takeda, TP Therapeutics, Xcovery and Yuhan. All other authors declare no competing interests.

Acknowledgements

This work was financially supported by a grant from the Seoul R&D Program (BT210184), and the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (HI14C1277). The ctDNA analysis for this study was conducted by IMBdx, Seoul, Korea.

Fig. 1.

Distribution of the key actionable mutations detected from ctDNA of NSCLC patients. ^a)Co-occurring resistance mutations found in 14 samples (T790M, C797S, C797Y), ^b)Co-occurring resistance mutations found in two samples (G1202R), ^c)Reference dataset from cBioPortal for Cancer Genomics. Metastatic Non–Small Cell Lung Cancer. Data originally published in Nature Medicine 2022 [10]. Retrieved on December 27, 2023, from https://www.cbioportal.org/study/summary?id=nsclc_ctdx_msk_2022.

Table 1.

Experimental LOD estimated across a spectrum of input DNA quantities

DNA amounts (ng)	95% LOD
DNA amounts (ng)	SNV (AF, %)	Insertion (AF, %)	Deletion (AF, %)	Fusion (AF, %)	CNA (CN)	MSI (AF, %)
30	0.11	0.11	0.06	0.21	2.13	0.65
20	0.15	0.19	0.05	0.32	2.45	0.62
10	0.19	0.21	0.16	0.53	2.58	-
5	0.50	0.52	0.19	0.64	2.63	-
2	1.01	2.11	0.91	1.42	2.47	-

AF, allele frequency; CN, copy number; CNA, copy number alteration; LOD, limit of detection; MSI, microsatellite instability; SNV, single nucleotide variant.

Table 2.

Patient characteristics

Characteristic	No. (%) (n=122)
Age (yr), median (range)	68 (40-89)
Sex
Male	66 (54.0)
Female	56 (46.0)
Histology
Adenocarcinoma	111 (91.0)
Squamous cell carcinoma	8 (6.6)
Others	3 (2.5)
Disease extent at the time of blood sample collection
Intrathoracic disease only	38 (31.1)
Involving extrathoracic metastasis or recurrent disease	84 (68.9)
Systemic treatment history before blood sample collection
Not treated for metastatic disease	59 (48.4)
Previously treated for metastatic disease	63 (51.6)

Table 3.

Concordance between AlphaLiquid100 ctDNA analysis and tissue-based NGS results in detecting key actionable mutations

	ctDNA (n=50)
	Mutation	No mutation
Tissue
Mutation	29	5^a)
No Mutation	2^b)	14

ctDNA, circulating tumor DNA; NGS, next-generation sequencing.

^a) Specific mutations in cases with mutations detected from tissue but not blood (number of cases): KIFB/RET (1), BRAF V600E (1), EML4/ALK (1), EGFR exon19del (1), ERBB2 exon20ins (1),

^b) Specific mutations in cases with mutations detected from blood but not tissue (number of cases): FGFR3 R248C (1), BRAF V600E (1).

Table 4.

Concordance between AlphaLiquid100 ctDNA analysis and tissue-based NGS results by individual genes

	ctDNA mutation (+)	ctDNA mutation (–)
50 Cases with matching NGS results
EGFR
Tissue mutation (+)	21	1
Tissue mutation (–)	0	28
ALK
Mutation (+)	3	1
Mutation (–)	0	45
All cases assessed for presence of genomic alterations (including PCR/IHC/ISH)
EGFR^a)
Tissue mutation (+)	21	1
Tissue mutation (–)	1	33
ALK
Mutation (+)	5	1
Mutation (–)	0	53

ctDNA, circulating tumor DNA; EGFR, epidermal growth factor receptor; IHC, immunohistochemistry; ISH, in situ hybridization; NGS, next-generation sequencing; PCR, polymerase chain reaction.

^a) Three patients did not have EGFR PCR results (ALK+ in 2, canceled by patient 1).

Table 5.

VAF of EGFR mutations detected by AlphaLiquid100, in cases included for concordance analysis compared to tissue NGS panel results

Case No.	ctDNA EGFR	VAF (%)	Tissue panel EGFR
3	Exon19del	0.57	Exon19del
6	L858R	1.47	L858R
10	Uncommon	15.92	Uncommon
11	L858R	10.49	L858R
13	L858R	19.37	L858R
14	Exon19del	0.92	Exon19del
15	Uncommon	0.1	Uncommon
16	Exon19del	1.91	Exon19del
19	L858R	3.72	L858R
23	Exon19del	19.09	Exon19del
26	Exon19del	17.66	Exon19del
28	Uncommon	22.37	Uncommon
33	Exon19del	9.93	Exon19del
34	L858R	17.44	L858R
37	Exon19del	0.05	Exon19del
41	Exon19del	3.3	Exon19del
45	L858R	0.21	L858R
47	L858R	0.54	L858R
48	Exon19del	0.32	Exon19del
49	Exon19del	2.89	Exon19del
51	Exon19del	4.97	Exon19del

ctDNA, circulating tumor DNA; EGFR, epidermal growth factor receptor; NGS, next-generation sequencing; VAF, variant allele fraction.

REFERENCES

1. Alix-Panabieres C, Pantel K. Liquid biopsy: from discovery to clinical application. Cancer Discov. 2021;11:858–73. Article PubMed PDF
2. Wan JCM, Massie C, Garcia-Corbacho J, Mouliere F, Brenton JD, Caldas C, et al. Liquid biopsies come of age: towards implementation of circulating tumour DNA. Nat Rev Cancer. 2017;17:223–38. Article PubMed PDF
3. Thompson JC, Aggarwal C, Wong J, Nimgaonkar V, Hwang WT, Andronov M, et al. Plasma genotyping at the time of diagnostic tissue biopsy decreases time-to-treatment in patients with advanced NSCLC: results from a prospective pilot study. JTO Clin Res Rep. 2022;3:100301.Article PubMed PMC
4. Nakamura Y, Taniguchi H, Ikeda M, Bando H, Kato K, Morizane C, et al. Clinical utility of circulating tumor DNA sequencing in advanced gastrointestinal cancer: SCRUM-Japan GI-SCREEN and GOZILA studies. Nat Med. 2020;26:1859–64. Article PubMed PDF
5. National Comprehensive Cancer Network. Non-small cell lung cancer (version 3.2023) [Internet]. Plymouth Meeting, PA: National Comprehensive Cancer Network; 2023 [cited 2023 Dec 5]. Available from: https://www.nccn.org/professionals/physician_gls/pdf/nscl.pdf
6. Rolfo C, Mack P, Scagliotti GV, Aggarwal C, Arcila ME, Barlesi F, et al. Liquid biopsy for advanced NSCLC: a consensus statement from the International Association for the Study of Lung Cancer. J Thorac Oncol. 2021;16:1647–62. Article PubMed
7. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009;25:1754–60. Article PubMed PMC PDF
8. Im SW, Chae J, Jang SS, Choi J, Yun J, Cha S, et al. A newly developed capture-based sequencing panel for genomic assay of lung cancer. Genes Genomics. 2020;42:751–9. Article PubMed PDF
9. Hendriks LE, Kerr KM, Menis J, Mok TS, Nestle U, Passaro A, et al. Oncogene-addicted metastatic non-small-cell lung cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann Oncol. 2023;34:339–57. Article PubMed
10. Jee J, Lebow ES, Yeh R, Das JP, Namakydoust A, Paik PK, et al. Overall survival with circulating tumor DNA-guided therapy in advanced non-small-cell lung cancer. Nat Med. 2022;28:2353–63. PubMed PMC
11. Song A, Kim TM, Kim DW, Kim S, Keam B, Lee SH, et al. Molecular changes associated with acquired resistance to crizotinib in ROS1-rearranged non-small cell lung cancer. Clin Cancer Res. 2015;21:2379–87. Article PubMed PDF
12. De Mattos-Arruda L, Siravegna G. How to use liquid biopsies to treat patients with cancer. ESMO Open. 2021;6:100060.Article PubMed PMC
13. Lone SN, Nisar S, Masoodi T, Singh M, Rizwan A, Hashem S, et al. Liquid biopsy: a step closer to transform diagnosis, prognosis and future of cancer treatments. Mol Cancer. 2022;21:79.Article PubMed PMC PDF
14. Rolfo C, Mack PC, Scagliotti GV, Baas P, Barlesi F, Bivona TG, et al. Liquid biopsy for sdvanced non-dmall vell lung cancer (NSCLC): a statement paper from the IASLC. J Thorac Oncol. 2018;13:1248–68. PubMed
15. U.S. Food and Drug Administration. List of cleared or approved companion diagnostic devices (in vitro and imaging tools) [Internet]. Silver Spring, MD: U.S. Food and Drug Administration; 2023 [cited 2023 Dec 5]. Available from: https://www.fda.gov/medical-devices/in-vitro-diagnostics/list-cleared-or-approved-companion-diagnostic-devices-in-vitro-and-imaging-tools
16. Sugimoto A, Matsumoto S, Udagawa H, Itotani R, Usui Y, Umemura S, et al. A large-scale prospective voncordance dtudy of plasma- and tissue-based next-generation targeted sequencing for advanced non-small cell lung cancer (LCSCRUM-Liquid). Clin Cancer Res. 2023;29:1506–14. PubMed
17. Desmeules P, Dusselier M, Bouffard C, Bafaro J, Fortin M, Labbe C, et al. Retrospective assessment of complementary liquid biopsy on tissue single-gene testing for tumor genotyping in advanced NSCLC. Curr Oncol. 2023;30:575–85. Article PubMed PMC
18. Woodhouse R, Li M, Hughes J, Delfosse D, Skoletsky J, Ma P, et al. Clinical and analytical validation of FoundationOne Liquid CDx, a novel 324-Gene cfDNA-based comprehensive genomic profiling assay for cancers of solid tumor origin. PLoS One. 2020;15:e0237802Article PubMed PMC
19. Yu C, Han Y, Wang M, Hua P, Zhang Y, Wang B. Concordance of ctDNA and tissue mutations in NSCLC: a meta-analysis. Cell Mol Biol (Noisy-le-grand). 2023;69:89–95. Article PDF
20. Mok T, Wu YL, Lee JS, Yu CJ, Sriuranpong V, Sandoval-Tan J, et al. Detection and dynamic changes of EGFR mutations from circulating tumor DNA as a predictor of survival outcomes in NSCLC patients treated with first-line intercalated erlotinib and chemotherapy. Clin Cancer Res. 2015;21:3196–203. Article PubMed PDF
21. Jeong SH, Kyung D, Yuk HD, Jeong CW, Lee W, Yoon JK, et al. Practical utility of liquid biopsies for evaluating genomic alterations in castration-resistant prostate cancer. Cancers (Basel). 2023;15:2847.Article PubMed PMC
22. Kang JK, Heo S, Kim HP, Song SH, Yun H, Han SW, et al. Liquid biopsy-based tumor profiling for metastatic colorectal cancer patients with ultra-deep targeted sequencing. PLoS One. 2020;15:e0232754Article PubMed PMC

Figure & Data

REFERENCES

Citations

Citations to this article as recorded by

Next-generation sequencing impact on cancer care: applications, challenges, and future directions
Mariano Zalis, Gilson Gabriel Viana Veloso, Pedro Nazareth Aguiar Jr., Nathalia Gimenes, Marina Xavier Reis, Silvio Matsas, Carlos Gil Ferreira
Frontiers in Genetics.2024;[Epub] CrossRef
Profiling Cell-Free DNA from Malignant Pleural Effusion for Oncogenic Driver Mutations in Patients with Treatment-Naive Stage IV Adenocarcinoma: A Multicenter Prospective Study
Shih-Chieh Chang, Yu-Feng Wei, Chung-Yu Chen, Yi-Chun Lai, Po-Wei Hu, Jui-Chi Hung, Cheng-Yu Chang
Molecular Diagnosis & Therapy.2024; 28(6): 803. CrossRef
Concordance of ctDNA and tissue genomic profiling in advanced biliary tract cancer
Sohyun Hwang, Seonjeong Woo, Beodeul Kang, Haeyoun Kang, Jung Sun Kim, Sung Hwan Lee, Chang Il Kwon, Dong Soo Kyung, Hwang-Phill Kim, Gwangil Kim, Chan Kim, Hong Jae Chon
Journal of Hepatology.2024;[Epub] CrossRef

PubReader
ePub Link

Cite

CITE: export Copy Download Format; Close

Download Citation

Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.

Format:

RIS — For EndNote, ProCite, RefWorks, and most other reference management software
BibTeX — For JabRef, BibDesk, and other BibTeX-specific software

Include:

Citation for the content below
Citation and abstract for the content below

Analytical and Clinical Validation of a Highly Sensitive NGS-Based ctDNA Assay with Real-World Concordance in Non–Small Cell Lung Cancer

Cancer Res Treat. 2024;56(3):765-773. Published online January 8, 2024

DOI: https://doi.org/10.4143/crt.2023.1294

XML Download

Abstract
Introduction
Materials and Methods
Results
Discussion
Electronic Supplementary Material
NOTES
REFERENCES

Clinical Validation of the Unparalleled Sensitivity of the Novel Allele-Discriminating Priming System Technology–Based EGFR Mutation Assay in Patients with Operable Non–Small Cell Lung Cancer

Analytical and Clinical Validation of a Highly Sensitive NGS-Based ctDNA Assay with Real-World Concordance in Non–Small Cell Lung Cancer

Fig. 1. Distribution of the key actionable mutations detected from ctDNA of NSCLC patients. a)Co-occurring resistance mutations found in 14 samples (T790M, C797S, C797Y), b)Co-occurring resistance mutations found in two samples (G1202R), c)Reference dataset from cBioPortal for Cancer Genomics. Metastatic Non–Small Cell Lung Cancer. Data originally published in Nature Medicine 2022 [10]. Retrieved on December 27, 2023, from https://www.cbioportal.org/study/summary?id=nsclc_ctdx_msk_2022.

Fig. 1.

Analytical and Clinical Validation of a Highly Sensitive NGS-Based ctDNA Assay with Real-World Concordance in Non–Small Cell Lung Cancer

DNA amounts (ng)	95% LOD
DNA amounts (ng)	SNV (AF, %)	Insertion (AF, %)	Deletion (AF, %)	Fusion (AF, %)	CNA (CN)	MSI (AF, %)
30	0.11	0.11	0.06	0.21	2.13	0.65
20	0.15	0.19	0.05	0.32	2.45	0.62
10	0.19	0.21	0.16	0.53	2.58	-
5	0.50	0.52	0.19	0.64	2.63	-
2	1.01	2.11	0.91	1.42	2.47	-

Characteristic	No. (%) (n=122)
Age (yr), median (range)	68 (40-89)
Sex
Male	66 (54.0)
Female	56 (46.0)
Histology
Adenocarcinoma	111 (91.0)
Squamous cell carcinoma	8 (6.6)
Others	3 (2.5)
Disease extent at the time of blood sample collection
Intrathoracic disease only	38 (31.1)
Involving extrathoracic metastasis or recurrent disease	84 (68.9)
Systemic treatment history before blood sample collection
Not treated for metastatic disease	59 (48.4)
Previously treated for metastatic disease	63 (51.6)

	ctDNA (n=50)
	Mutation	No mutation
Tissue
Mutation	29	5^a)
No Mutation	2^b)	14

	ctDNA mutation (+)	ctDNA mutation (–)
50 Cases with matching NGS results
EGFR
Tissue mutation (+)	21	1
Tissue mutation (–)	0	28
ALK
Mutation (+)	3	1
Mutation (–)	0	45
All cases assessed for presence of genomic alterations (including PCR/IHC/ISH)
EGFR^a)
Tissue mutation (+)	21	1
Tissue mutation (–)	1	33
ALK
Mutation (+)	5	1
Mutation (–)	0	53

Case No.	ctDNA EGFR	VAF (%)	Tissue panel EGFR
3	Exon19del	0.57	Exon19del
6	L858R	1.47	L858R
10	Uncommon	15.92	Uncommon
11	L858R	10.49	L858R
13	L858R	19.37	L858R
14	Exon19del	0.92	Exon19del
15	Uncommon	0.1	Uncommon
16	Exon19del	1.91	Exon19del
19	L858R	3.72	L858R
23	Exon19del	19.09	Exon19del
26	Exon19del	17.66	Exon19del
28	Uncommon	22.37	Uncommon
33	Exon19del	9.93	Exon19del
34	L858R	17.44	L858R
37	Exon19del	0.05	Exon19del
41	Exon19del	3.3	Exon19del
45	L858R	0.21	L858R
47	L858R	0.54	L858R
48	Exon19del	0.32	Exon19del
49	Exon19del	2.89	Exon19del
51	Exon19del	4.97	Exon19del

Table 1. Experimental LOD estimated across a spectrum of input DNA quantities

AF, allele frequency; CN, copy number; CNA, copy number alteration; LOD, limit of detection; MSI, microsatellite instability; SNV, single nucleotide variant.

Table 2. Patient characteristics

Table 3. Concordance between AlphaLiquid100 ctDNA analysis and tissue-based NGS results in detecting key actionable mutations

ctDNA, circulating tumor DNA; NGS, next-generation sequencing.

Specific mutations in cases with mutations detected from tissue but not blood (number of cases): KIFB/RET (1), BRAF V600E (1), EML4/ALK (1), EGFR exon19del (1), ERBB2 exon20ins (1),

Specific mutations in cases with mutations detected from blood but not tissue (number of cases): FGFR3 R248C (1), BRAF V600E (1).

Table 4. Concordance between AlphaLiquid100 ctDNA analysis and tissue-based NGS results by individual genes

ctDNA, circulating tumor DNA; EGFR, epidermal growth factor receptor; IHC, immunohistochemistry; ISH, in situ hybridization; NGS, next-generation sequencing; PCR, polymerase chain reaction.

Three patients did not have EGFR PCR results (ALK+ in 2, canceled by patient 1).

Table 5. VAF of EGFR mutations detected by AlphaLiquid100, in cases included for concordance analysis compared to tissue NGS panel results

ctDNA, circulating tumor DNA; EGFR, epidermal growth factor receptor; NGS, next-generation sequencing; VAF, variant allele fraction.