Abstract
Background/Aim: Comprehensive genomic profiling (CGP) with tissue- and blood-based next-generation sequencing (NGS) is integral to the delivery of personalized medicine for targeted cancer therapy. This study aimed to evaluate the variant concordance for somatic variants using two clinical NGS systems for conducting both tissue- and blood-based analyses: Genexus-OCA v3 (OCA) vs. FoundationOne CDx (F1) for tissues and Genexus OPA (OPA) vs. FoundationOne CDx Liquid (F1L) for blood.
Patients and Methods: The concordance of genomic alterations between the two NGS analyses was compared in six patients with breast, head, and neck cancers using tissue and circulating tumor DNA biopsies.
Results: A total of 130 genes were common between F1 and OCA, and 41 between F1L and OPA. When comparing FoundationOne to Genexus for common genes, the sensitivity and specificity of OCA and OPA were 55% and 99%, respectively. Nine single-nucleotide variants (SNVs), one copy number alteration (CNA), and one fusion were detected by both Genexus and FoundationOne. However, one SNV (MAP2K1 F53V), two CNAs (AKT3 and MYC), and one fusion (ESR-CCDC170) were detected only in Genexus, whereas two SNVs (TP53 Q331* and KRAS G12V) were detected only in FoundationOne.
Conclusion: The two cancer genome panels were equivalent but not perfect in terms of the detection of variants using tissue and blood, indicating that different assays and analytical methods may have influenced the results. When performing CGPs, it is important to consider the characteristics of each NGS-based CGP test and the genetic variants associated with each disease.
Introduction
Comprehensive genomic profiling (CGP) testing by next-generation sequencing (NGS), leading to precision medicine, has become a clinical practice to select effective treatments based on druggable genomic alterations. Among several CGP tests, FoundationOne CDx (Foundation Medicine) is the first U.S. Food and Drug Administration (FDA)-approved tissue-based broad companion diagnostic (CDx) test, and was approved in 2019 in Japan as a medical device for both CGP and CDx tests. FoundationOne CDx can detect substitutions, insertions, deletions, and copy number alterations in 324 genes and selected gene rearrangements (1, 2). However, the indications in Japan are limited to patients with advanced solid tumors who likely completed or have completed standard treatment, or rare cancers that have no standard treatment. Considering the indications in Japan, patients with cancer may miss the opportunity to receive molecularly matched therapies (3). Implementing readily available, optional, private CGP testing is of great benefit to patients (4, 5). To meet the needs of patients with cancer, CGP testing has been implemented at the patients’ own expense.
The Ion Torrent™ Genexus™ Sequencer (Genexus) launched by Thermo Fisher Scientific is a highly integrated instrument that can automate library construction, templating, and sequencing in a single-instrument run (6). Unlike FoundationOne, Genexus has not yet been approved as a clinical medical device by the Ministry of Health, Labor, and Welfare in Japan. However, Genexus requires only a few steps: setting up a run with a few iterations of pipetting, and then loading all the prefilled reagents and consumables according to the instructions on the sequencer’s touchscreen. Genexus has the potential to establish NGS services within a routine clinical diagnostic laboratory, even in a regional hospital with no previous experience with NGS (7).
Therefore, this study aimed to compare two CGP tests, Genexus and FoundationOne. The concordance of somatic variants using two CGP tests with tissue- and highly sensitive blood-based analyses were evaluated. In addition, the differences in the detailed test results were examined and the clinical situations in which these tests should be performed were clarified.
Patients and Methods
Patients. Six patients with cancer were enrolled in this study. Three primary tumor tissues from two patients with breast cancer and one patient with parotid gland cancer, and three peripheral blood samples from three patients with breast cancer were analyzed. Detailed patient information is shown in Table I. All samples were analyzed using Genexus Oncomine Comprehensive Assay v3 (OCAv3) and FoundationOne for tissue, and Genexus Oncomine Precision Assay (OPA) and FoundationOne Liquid for blood. The study was approved by Kurume University Hospital’s Institutional Review Board (numbers: 2022008) and Japan Registry of Clinical Trials (number: 072230003). All methods were carried out in accordance with relevant guidelines and regulations. Participants of this study were fully informed of the purpose and procedures of the study and had adequate time to ask questions and ponder about their voluntary participation. A written informed consent was obtained from all patients before enrollment.
Patient characteristics.
Sample, DNA extraction, and quantification. DNA and RNA from formalin-fixed paraffin-embedded (FFPE) tissue specimens were extracted using the Maxwell RSC Instrument (Promega, Fitchburg, WI, USA, catalog #AS4500) with a Maxwell RSC FFPE Plus DNA kit (Promega, catalog #AS1720) and Maxwell RSC RNA FFPE kit (Promega, catalog #AS1440), respectively. To obtain blood plasma, 14 ml of whole blood with EDTA-2Na was cooled and centrifuged (4°C, 2,000 × g, 10 min) twice. Cell-free total nucleic acid (cfTNA), which included both DNA and RNA, was extracted using a Maxwell RSC Instrument with a Maxwell RSC miRNA Plasma and Serum Kit (Promega, catalog #AS1680). DNA and RNA concentrations were measured using the QuantiFluor ONE dsDNA System (Promega, catalog #E4871) and QuantiFluor RNA System (Promega, catalog # E3310), respectively. The concentration of nucleic acid >1.1 ng/μl for DNA and >0.95 ng/μl for RNA extracted from tissues, and >1.33 ng/μl from blood were recommended to proceed in the next sequencing steps. The fragment lengths of DNA and RNA were evaluated using an Agilent 4200 TapeStation system (Agilent, Santa Clara, CA, USA).
Genexus sequencing. The Ion Torrent™ Genexus™ Integrated Sequencer (Thermo Fisher Scientific, Waltham, MA, USA) is a fully automated NGS system that integrates library preparation, including cDNA synthesis, template preparation, sequencing, and data analysis from purified and quantified nucleic acids (DNA, RNA, and cfTNA). In this study, DNA and RNA derived from FFPE were used as samples for OCAv3 and cfTNA for OPA. Twenty-five microliters of DNA (>1.1 ng/μl) and RNA (>0.95 ng/μl) for OCAv3 and 20 μl of cfTNA (>1.33 ng/μl) for OPA were recommended to use for library preparation. After assigning the assay to the Genexus Integrative System, the sample type and sequence run settings were specified and loaded onto sample plates. The Multiplex I cfDNA Reference Standard Set (Horizon Diagnostics HD780) was used to evaluate the performances of OCAv3 and OPA (6).
Genexus variant analyses. The sequencing data were mapped to the standard reference genome, the human genome assembly 19, using on-instrument Genexus software, and aligned using the torrent mapping alignment program. After the initial mapping, a variant call was performed using the Torrent Variant Caller. The variants were selected using a built-in variant filter and subsequently called by the system The four major classes of mutations evaluated included single nucleotide variants (SNV), insertions and deletions (INDEL), copy number amplifications (CNA), and gene fusions. Variants with a low frequency of <0.3%, an allele count of ≤5, and a C to T or G to A change were excluded from the results as potential deaminations for the OPA of the liquid samples. The results were compared with those of the same samples from F1 or F1L. In cases where the variants did not match with F1 and F1L, the unfiltered analysis data were examined (6).
FoundationOne® and FoundationOne® Liquid CDx. FoundationOne® CDx (F1, Foundation Medicine, Inc., Cambridge, MA, USA) has been previously described and validated (8, 9). Briefly, FFPE samples retrieved from surgical specimens were obtained and 10 unstained sections and one hematoxylin and eosin stained (H&E) section (thickness: 5 μm, area: >25 mm2) were delivered to Foundation Medicine, Inc. All samples were confirmed to be carcinomas on routine H&E-stained slides and contained a minimum of 20% tumor cells. F1 applies NGS across 324 genes known to be drivers of solid tumors with high accuracy by sequencing the coding regions of 309 cancer-related genes and the introns of 36 genes. Sequencing was performed using the Illumina HiSeq® 4000 (Illumina, San Diego, CA, USA) to identify base substitutions, INDELS, CNA, and rearrangements. In addition, the tumor mutational burden (TMB) and microsatellite instability (MSI) were assessed. MSI was classified as: stable (MSS), intermediate (MSI-I), high (MSI-H), and ‘cannot be determined’ (MSI status could not be accurately determined). TMB was classified into four categories: low [≤5 mutations per megabase (mut/Mb)], intermediate (6-19 mut/Mb), high (≥20 mut/Mb), and ‘cannot be determined’ which meant that the TMB score could not be accurately calculated.
FoundationOne® CDx Liquid (F1L, Foundation Medicine) has been previously described and validated (2, 8, 9). F1L is an NGS-based in vitro diagnostic tool that analyzes a panel of 324 genes using circulating cell-free DNA isolated from plasma-derived anticoagulated peripheral whole blood. Using a novel hybrid capture approach, a subset of targeted regions in 75 genes was baited for greater sensitivity through ultra-deep sequencing coverage. Blood samples were collected in two tubes of whole blood (8.5 ml per tube). The samples were shipped to Foundation Medicine, Inc. at ambient temperature. The F1L assay employs a single DNA extraction method to obtain cfDNA from the whole blood plasma (10).
Statistical analyses. Genexus and F1 were compared by calculating the percentage match, sensitivity, specificity, positive predictive value, and negative predictive value. The number of genes in F1 was 324 (https://www.foundationmedicine.com/test/foundationone-cdx), whereas in Genexus-OCAv3 it was 161 (https://www.thermofisher.com/jp/ja/home/clinical/preclinical-companion-diagnostic-development/oncomine-oncology/oncomine-cancer-research-panel-workflow/oncomine-comprehensive-assay.html), and the number of matched genes was 130. The number of F1L genes limited to the high-sensitivity regions was 75, the number of Genexus-OPA genes was 50 (https://www.thermofisher.com/jp/ja/home/clinical/preclinical-companion-diagnostic-development/oncomine-oncology/oncomine-precision-assay.html), and the number of matched genes was 41.
The concordance rate refers to the percentage of genes that are consistent between the two tests, F1 and OCAv3 and between F1L and OPA, regardless of whether the result is positive or negative. Sensitivity refers to the percentage of genes that tested positive in both F1 and OCAv3, or in F1L and OPA, relative to the total number of genes that were positive in F1 or F1L. Specificity is the percentage of genes for which both F1 and OCAv3 or F1L and OPA were negative, out of the total number of genes for which F1 or F1L was negative. The positive predictive value was calculated as the ratio of the number of genes positive for OCAv3 or OPA to the number of genes positive for F1 or F1L for the genetic variants detected in the F1 or F1L. Similarly, the negative predictive value was calculated as the ratio of the number of genes negative for OCAv3 or OPA to the number of genes negative for F1 or F1L for genetic variants not detected in the F1 or F1L.
Results
DNA and RNA concentration of five cases (case# 1, 2, 3, 4, 6) were passed recommended criteria, but DNA concentration of case #5 was 0.83 ng/μl which was not cleared according to recommended criteria. However, all six samples including case #5 passed the Genexus quality control criteria and the mapped reads were highly consistent, demonstrating the accuracy of calls among the Ion Torrent platforms, informatic pipelines, and panels. Compared to FoundationOne, the sensitivity, specificity, positive predictive value, and negative predictive value of all six cases using OCA and OPA were 77%, 99%, 66%, and 99%, respectively. Those of the three cases using OCA with tissue samples were 89%, 99%, 80%, and 99%, respectively, and those of the other three cases using OPA with blood samples were 50%, 97%, 40%, and 98%, respectively (Table II).
Sensitivity, specificity, and positive predictive value of Genexus vs. Foundation one.
A list of somatic variants compared between Genexus and FoundationOne is shown in Table III. Among the six cases, 12 SNVs, three CNAs, and two fusions were detected. Nine SNVs, one CNA, and one fusion were detected using both Genexus and FoundationOne. However, one SNV (MAP2K1 F53V), two CNAs (AKT3 and MYC), and one fusion (ESR-CCDC170) were detected only using Genexus, whereas two SNVs (TP53 Q331* and KRAS G12V) were detected only using FoundationOne. Using the default Genexus analysis system, nine variants were not concordant between Genexus and FoundationOne. However, two SNVs (PIK3CA P449_L452del and TP53 L257_E258>Q) and one CNS (MYC) were identified in the original VCF files.
List of somatic variants comparing Genexus vs. FoundationOne.
Discussion
Our findings demonstrated that the sensitivity and specificity of Genexus were 55% and 99%, respectively, compared to those of FoundationOne, which is one of the standard CGP certified by Japanese National Insurance. Among the thirteen variants that were detected using FoundationOne, only two were not detected using Genexus. Four variants were detected using Genexus, but not using FoundationOne. Our findings suggest that Genexus is comparable to FoundationOne, which has been approved in Japan, and that different assays and analytical methods can influence its results.
Several studies have compared Genexus with standard-of-care genomic testing methods. A study of 119 non-small cell lung cancer cases using cell-free total nucleic acids indicated that Genexus showed high sensitivity and specificity compared with standard-of-care assays, such as a PCR-based assay (Cobas EGFR test and Oncomine Dx) and anaplastic lymphoma kinase-immunohistochemistry (12). This study compared genetic variant concordance using two clinical NGS systems, Genexus OPA and GeneStrat, run as the Ion GeneStudio S5 PRIME system for blood-based analyses. The results indicated 100% concordance in the variant calls for SNVs, INDELs, and fusions between the two platforms. However, the Genexus workflow failed to call one CNA observed in S5, resulting in an overall concordance of 97% (32 of 33) (13). In line with these studies comparing Genexus with other standards or NGS testing methods, our study also showed a high concordance rate between Genexus and FoundationOne. Moreover, this result suggests that Genexus and F1, which use different analytical methods, can detect the same genetic variants. However, the default variant-calling settings in Genexus failed to detect three variants identified by F1 or F1L, suggesting the need to develop user-customized variant-calling parameters to enhance detection sensitivity. Genexus, which can analyze original data in detail, may expand the possibility of drug selection for treatment and participation in clinical trials.
The Genexus system is an automated NGS tool that can reduce turnaround time. It comprises two software-linked instruments: the Ion Torrent Genexus Purification System and Ion Torrent Genexus Integrated Sequencer. The system detected genetic variations consistently with a sensitivity and positive predictive value of ≥95%, as compared to the Ion GeneStudio S5 System, an orthogonal method (14). Another study conducted at a Canadian academic center analyzed 578 solid tumor FFPE samples using Genexus. Four hundred and eighty-one (83%) of the cases were completed in less than five business days (15). Additionally, a study using cell-free DNA from non-small cell lung cancer demonstrated that Genexus is a fully automated and highly accurate NGS system with a short turnaround time (16). This suggests that the Genexus system can provide accurate identification of tumor markers for oncology research with a fast turnaround, be implemented in a hospital laboratory, and help clinicians make timely decisions for the oncologist (7).
Study limitations. The study was limited by a small sample size and a narrow range of tumor types, consisting of six cases involving breast and parotid gland cancers. Moreover, our analytical workflow was not customized using the Genexus accompanying workflow. Because this was a preliminary study, we are starting to construct a custom workflow to expand patient treatment selection. Several aspects of NGS analysis should be noted. First, it is important to determine the quantity and quality of nucleic acids present. CGP testing using tissue samples remains the standard approach due to the stability and adequacy of nucleic acids extracted from FFPE blocks. These allow adjustment of sample volume based on tumor content in surgical or biopsy specimens. In contrast, blood-derived samples often yield lower concentrations of nucleic acids, which can result in less reliable assay performance. Second, ensuring the accuracy of NGS analysis is critical. Accuracy is maintained through strict quality control parameters at each analytical stage, and expert validation is recommended, especially for blood samples where nucleic acid content may be suboptimal. Engaging molecular biology specialists is advisable when assay parameters are unclear to confirm the reliability of genomic findings. Finally, it is necessary to understand the characteristics and principles of the panels. While F1 and F1L were analyzed using an Illumina sequencer, Genexus was analyzed using an Ion Torrent sequencer. A significant difference between these sequencers is the method of library construction: hybridization capture-based approaches for Illumina vs. amplicon-based approaches for Ion Torrent by Genexus. It is necessary to understand that while there is no basic difference between their results, differences in characteristics of methods lead to different results (17). A comparative study of these two approaches using whole-exome sequencing showed that amplicon methods have higher on-target rates, whereas hybridization capture-based approaches demonstrate better uniformity (17). This study emphasizes the importance of selecting an appropriate methodology and offers guidance for laboratories in selecting their preferred approach.
Conclusion
Our results demonstrated high sensitivity and specificity of Genexus compared to the standard CGP testing FoundationOne, suggesting that Genexus is comparable to FoundationOne, which was approved in Japan, and that different assays and analytical methods may influence the results. Given the Japanese national insurance rules that limit CGP to only once to a single patient, the use of Genexus as a private CGP, which can analyze original data in detail, may expand the possibility of drug selection for treatment and participation in clinical trials.
Acknowledgements
The Authors express their gratitude to the Institutional Review Board of Kurume University Hospital, Professor Fumihiko Fujita of the Department of Surgery at Kurume University, and the patients who participated in this study. Portions of this research have been previously published as a preprint: “Comparison of Comprehensive Genomic Profiling Testing ‘Ion Torrent Genexus Sequencer’ with FoundationOne” (18).
Footnotes
Authors’ Contributions
YN and KF designed the study and wrote the protocol. KF, YW, TS, SN, KY, HU, MN, and YN contributed to the design. NM, JA, HA, YI, KI, and FY extracted DNA and implemented NGS sequencing. KF and YN managed the literature search and analyses. RS, SO, YO, TO, KO, MK, RK, HK, AK, RT, RU, UT, KH, TH, SH, MM, and DM enrolled the patients and performed the clinical assessment. KF undertook the statistical analysis. KF and NM wrote the first draft of the manuscript. YN and TS supervised the whole process and critically reviewed the article. All Authors reviewed and approved the final version of the manuscript.
Conflicts of Interest
The Authors have no conflicts of interest to disclose in relation to this study.
Funding
Not applicable.
Artificial Intelligence (AI) Disclosure
No artificial intelligence (AI) tools, including large language models or machine learning software, were used in the preparation, analysis, or presentation of this manuscript.
- Received May 23, 2025.
- Revision received June 13, 2025.
- Accepted June 19, 2025.
- Copyright © 2025 The Author(s). Published by the International Institute of Anticancer Research.
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.