Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Multi-Center Evaluation of the Fully Automated PCR-Based Idylla™ KRAS Mutation Assay for Rapid KRAS Mutation Status Determination on Formalin-Fixed Paraffin-Embedded Tissue of Human Colorectal Cancer

Abstract

Since the advent of monoclonal antibodies against epidermal growth factor receptor (EGFR) in colorectal cancer therapy, the determination of RAS mutational status is needed for therapeutic decision-making. Most prevalent in colorectal cancer are KRAS exon 2 mutations (40% prevalence); lower prevalence is observed for KRAS exon 3 and 4 mutations (6%) and NRAS exon 2, 3, and 4 mutations (5%). The Idylla™ KRAS Mutation Test on the molecular diagnostics Idylla™ platform is a simple (<2 minutes hands-on time), highly reliable, and rapid (approximately 2 hours turnaround time) in vitro diagnostic sample-to-result solution. This test enables qualitative detection of 21 mutations in codons 12, 13, 59, 61, 117, and 146 of the KRAS oncogene being clinically relevant according to the latest clinical guidelines. Here, the performance of the Idylla™ KRAS Mutation Assay, for Research Use Only, was assessed on archived formalin-fixed paraffin-embedded (FFPE) tissue sections by comparing its results with the results previously obtained by routine reference approaches for KRAS genotyping. In case of discordance, samples were assessed further by additional methods. Among the 374 colorectal cancer FFPE samples tested, the overall concordance between the Idylla™ KRAS Mutation Assay and the confirmed reference routine test results was found to be 98.9%. The Idylla™ KRAS Mutation Assay enabled detection of 5 additional KRAS-mutated samples not detected previously with reference methods. As conclusion the Idylla™ KRAS Mutation Test can be applied as routine tool in any clinical setting, without needing molecular infrastructure or expertise, to guide the personalized treatment of colorectal cancer patients.

Citation: Solassol J, Vendrell J, Märkl B, Haas C, Bellosillo B, Montagut C, et al. (2016) Multi-Center Evaluation of the Fully Automated PCR-Based Idylla™ KRAS Mutation Assay for Rapid KRAS Mutation Status Determination on Formalin-Fixed Paraffin-Embedded Tissue of Human Colorectal Cancer. PLoS ONE 11(9): e0163444. https://doi.org/10.1371/journal.pone.0163444

Editor: John Souglakos, University of Crete, GREECE

Received: June 13, 2016; Accepted: September 7, 2016; Published: September 29, 2016

Copyright: © 2016 Solassol et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper and its Supporting Information files.

Funding: The authors received no specific funding for this work.

Competing interests: Peter Sillekens is an employee of Biocartis. He is senior scientist at the Research & Development department. Biocartis has no financial contribution to this study, besides free of charge Idylla™ KRAS Mutations Assay cartridges.This does not alter our adherence to PLOS ONE policies on sharing data and materials.

Introduction

The Kirsten rat sarcoma viral oncogene (KRAS) belongs to a family of related RAS genes, comprising three known human isoforms, i.e., KRAS, NRAS, and HRAS [1, 2]. KRAS encodes the KRAS protein, consisting of 188 or 189 amino acids (depending on exon 4 utilization), which is a small membrane-bound GTPase that plays a pivotal role in cell signal transduction. Normally, KRAS exists in an inactive state. Ligand binding of a nearby transmembrane tyrosine kinase receptor, like the epidermal growth factor receptor (EGFR), leads to activation of KRAS, which is directly downstream of this receptor. Once in its active state, KRAS in turn activates a wide variety of downstream effectors, hence influencing cell proliferation and cell survival. Mutated KRAS remains in the active state, leading to a loss of its regulatory function on downstream effectors and eventually to cancer cell survival.

There are several etiological pathways leading to colorectal cancer, the traditional biomarkers being microsatellite instability, the CpG island methylator phenotype, and somatic mutations in BRAF and KRAS [3–5]. In a recent analysis of randomized controlled trials in metastatic colorectal cancer patients, RAS mutation prevalence was found to be 55.9%, with KRAS exon 2 mutations being most common (42.6% prevalence), followed by KRAS exon 4 (6.2%), NRAS exon 3 (4.2%), KRAS exon 3 (3.8%), NRAS exon 2 (2.9%), and NRAS exon 4 (0.3%) mutations [6]. The corresponding KRAS single amino acid missense mutations are located at codons 12 or 13 for exon 2, codons 59 or 61 for exon 3, and codons 117 or 146 for exon 4, with G12D (13.1% prevalence), G12V (11.6%), and G13D (8.1%) being most prevalent in colorectal cancer.

Colorectal cancer is the second most common cause of cancer death in Europe, the third most common cause in the US, and the fourth most common cause worldwide [5, 7, 8]. For many years, intravenous 5-fluorouracil/leucovorin or oral capecitabine were the backbone of first-line palliative chemotherapy for colorectal cancer [5, 9]. Since 2000, the addition to 5-fluorouracil/leucovorin of oxaliplatin (FOLFOX) or irinotecan (FOLFIRI), or the combination of capecitabine with oxaliplatin (CAPOX), led to increased response rates and survival. More recently, the advent of targeted therapies including human vascular endothelial growth factor (VEGF) and EGFR monoclonal antibodies, expanded treatment options and further increased treatment response. Thanks to the improvements in treatment and detection, 5-year colorectal cancer survival rates increased significantly from about 50% in the 1970s to about 65% currently [8], while median overall survival of patients with metastatic disease increased from 8–12 months to 21–24 months [10].

Addition of the anti-EGFR monoclonal antibody cetuximab or panitumumab to chemotherapy regimens improves outcome for metastatic colorectal cancer [9]. It was found that presence of mutations in KRAS exon 2 (codon 12/13) considerably reduced the efficacy of these EGFR inhibitors [11, 12]. Exon 2 mutations have, therefore, been used as routine biomarkers for predicting lack of response to cetuximab and panitumumab, thereby protecting metastatic colorectal cancer patients from the undesirable side effects and the considerable cost of ineffective therapy. As the presence of exon 2 mutations did not fully explain poor response to anti-EGFR therapy, further analysis identified additional mutations predicting resistance, i.e., mutations in KRAS exon 3 (codons 59 and 61) and 4 (codons 117 and 146), and in NRAS exon 2 (codons 12 and 13), exon 3 (codons 59 and 61), and exon 4 (codons 117 and 146) [13–15]. Hence, extended RAS analysis, beyond KRAS exon 2, is necessary to identify patients eligible for EGFR-targeted therapy. This is reflected by the latest clinical practice guidelines of the European Society for Medical Oncology (ESMO) and the National Comprehensive Cancer Network (NCCN) recommending genotyping of tumor tissue (primary or metastatic) for the presence of KRAS exon 2 mutations as well as non-exon 2 mutations and NRAS mutations [9, 16].

Various laboratory-based or commercial KRAS mutation assays are used in routine practice and the majority of those have been optimized to be compatible with DNA extracted from formalin-fixed paraffin-embedded (FFPE) samples, which represent the most common form of tumor tissue specimens [17]. However, these assays are characterized by differences in the range of mutations covered, sensitivity, labor, cost, level of automation and multiplexing, and in the need for specialized equipment and highly skilled staff. Importantly, it was recently shown that KRAS genotyping results were only made available within 15 days for 82%, 51%, and 98% of tested patients in Europe, Latin America, and Asia, respectively [18]. Since a timeframe of 15 working days is undesirable, in particular for the therapeutic management of rapidly progressing metastatic patients, shorter timelines are necessary. In some countries, guidelines even require results to be available within 8 to 10 days [19].

The Idylla™ BRAF Mutation Test, CE IVD, on the Idylla™ platform (Biocartis, Mechelen, Belgium) was successfully launched for the fast and accurate detection of BRAF V600 mutations in melanoma patients, directly on FFPE samples [20, 21]. More recently, the Idylla™ KRAS Mutation Test, CE IVD, which enables detection of 21 clinically relevant KRAS mutations, was launched, and the Idylla™ NRAS Mutation Assay was made available for Research Use Only. Unlike conventional methods that use time-consuming FFPE processing steps, Idylla™ enables integration of pre-analytical and analytical processes in a single cartridge, thus eliminating the need of manually and time-consuming successive deparaffinization, tissue digestion, DNA extraction procedures, and PCR set-up steps.

Here, the results of a multi-center study testing archived FFPE tumor samples from colorectal cancer patients for their KRAS mutational status with the Idylla™ KRAS Mutation Assay (RUO) are reported. To establish test performance, these results were compared to the results previously obtained for these samples using routine reference methods present in the participating laboratories.

Materials and Methods

Tissue Sample Collection

Archived clinical FFPE materials of 374 colorectal cancer patients were selected for this study. Samples were obtained from 12 European centers: Klinikum Augsburg (Augsburg, Germany; n = 29), Hospital del Mar (Barcelona, Spain; n = 31), Queen Elizabeth Hospital (Birmingham, UK; n = 32), Hôpital Erasme (Brussels, Belgium; n = 30), Rigshospitalet Copenhagen (Copenhagen, Denmark; n = 29), Royal London Hospital (London, UK; n = 31), Laboratoire National de Santé (Dudelange, Luxembourg; n = 32), Niguarda Cancer Center (Milano, Italy; n = 34), Institut du Cancer de Montpellier (Montpellier, France; n = 32), Oxford University Hospitals NHS Foundation Trust (Oxford, UK; n = 30), Surgical Pathology Unit (Padua, Italy; n = 30), Centre Hospitalier Régional Universitaire de Tours (Tours, France; n = 34).

FFPE tissue sections (mostly 1 or 2, up to 12) of 5 μm to 25 μm thick were sampled as close as possible (within the same FFPE block) to the sections used before to generate the reference result (Fig 1). FFPE tissue sections were placed directly into the Idylla™ cartridge following the assay instructions. Tumor content and area were determined on hematoxylin-eosin-stained slides by a Pathologist and, if needed, macro-dissection was performed to achieve a tumor cell content of at least 25%.

thumbnail
Download:
Fig 1. Decision tree of pre-analytical FFPE sample preparation prior to Idylla™ KRAS Mutation Assay.

https://doi.org/10.1371/journal.pone.0163444.g001

Artificial FFPE samples (i.e., FFPE Reference Standards from Horizon Discovery, Cambridge, UK) were included as external controls: 5 sites tested KRAS G12D at 5% allelic frequency as well as KRAS G13D at 50% allelic frequency, 2 sites only tested KRAS G12D at 5%, and 1 site only tested KRAS G13D at 50%.

In-House KRAS Mutation Tests Used as Reference Method

To guide cancer therapy, the KRAS mutational status had been assessed previously on FFPE material using routine reference methods. Depending on each center, different protocols were used.

Sanger sequencing was used as a routine reference method by adopting different protocols and equipment at each site.

Digital droplet PCR was performed on a QX100™ system (Bio-Rad Laboratories, Hercules, CA) according to the instructions of the manufacturer.

High-resolution melting (HRM) analysis was performed following three different protocols: (i) fragments from KRAS exon 2, 3, and 4 were PCR amplified using a Rotor-Gene 6000 instrument (Qiagen) in combination with the LightCycler 480 High-Resolution Melting Master Reaction Mix (Roche Diagnostics, Meylan, France); (ii) using an automated ABI Prism 7900HT Sequence Detection System (Life Technologies) according to the manufacturer’s recommendations; or (iii) using a LightCycler® 480 II (Roche) using the LightCycler® 480 High Resolution Melting Master Kit (Roche).

Pyrosequencing was performed on a PyroMark Q24 MDx (Qiagen) following the manufacturer’s instructions.

Next-generation-sequencing (NGS) was performed as described [22], using the Ampliseq™ Colon and Lung Cancer panel (Life Technologies, Carlsbad, CA) for amplification and sequencing of hotspot mutations in 22 lung and colon cancer relevant genes, and massively parallel sequencing was carried out on an Ion Torrent Personal Genome Machine sequencer (Life Technologies) following the manufacturer’s instructions. Alternatively, the Multiplicom (Niel, Belgium) Somatic 1Mastr v2 Colon Cancer Panel was used for amplification and sequencing of the complete coding regions of the BRAF, KRAS, and NRAS genes, and massively parallel sequencing was carried out on a MiSeq Sequencer (Illumina, San Diego, CA) following the manufacturer’s instructions.

Commercial Routine KRAS Mutation Tests Used as Reference Method

At several centers to guide cancer therapy, the KRAS mutational status had been assessed previously on FFPE material using commercial routine tests.

The cobas® KRAS Mutation Test (Roche), the Ion Torrent AmpliSeq™ Colon and Lung Cancer Research Panel (Thermo Fisher Scientific), therascreen® KRAS Pyro Kit (Qiagen), therascreen® RAS Extension Pyro Kit (Qiagen), the KRAS, BRAF, PIK3CA* Array (Randox Molecular, Crumlin, UK), and the QClamp™ KRAS Codon Specific Mutation Detection Kit (Exon 2, 3, 4) (DiaCarta, Hayward, CA), were performed according to the manufacturer’s instructions.

Idylla™ KRAS Mutation Assay

The Idylla™ KRAS Mutation Assay enables mutation detection in exons 2, 3, and 4 of the KRAS oncogene. This assay consists of five allele-specific multiplex PCR reactions, designed for the specific amplification of KRAS gene sequences with a mutation in codons 12, 13, 59, 61, 117, or 146. As such, the test enables detection of 21 KRAS mutations, i.e., seven mutations in exon 2 (codons 12 and 13), nine mutations in exon 3 (codons 59 and 61), and five mutations in exon 4 (codons 117 and 146) of the KRAS oncogene. All possible Idylla™ KRAS Mutation Assay genetic call values are shown in Table 1. In case of multiple mutations, only the dominantly detected mutation (lowest ΔCq value) is currently reported.

thumbnail
Download:
Table 1. KRAS mutations detected by the Idylla™ KRAS Mutation Test.

https://doi.org/10.1371/journal.pone.0163444.t001

FFPE tissue sections were placed directly into the cartridge of the fully automated Idylla™ platform (Biocartis, Mechelen, Belgium) following the manufacturer’s instructions, without requiring prior manual deparaffinization or FFPE pre-processing. Briefly, after insertion of the FFPE tissue section into the cartridge and insertion of the cartridge into the Idylla™ instrument, a combination of reagents, enzymes, heat, and high intensity focused ultrasound (HIFU) induces deparaffinization, disruption of the tissue, and lysis of the cells. The nucleic acids are liberated for subsequent real-time PCR amplification. Allele-specific primers and fluorescent probes are present in a stable formulation. A sample processing control, involving the simultaneous amplification of a conserved fragment in the intron 4 / exon 5 junctional region of the KRAS gene, was performed in each of the five multiplex PCR reactions to check for adequate execution of the complete sample-to-result process and as a measure for the amount of amplifiable DNA in the sample.

For every valid PCR curve, the Idylla™ software calculates a quantification cycle (Cq) value. If the difference between the measured Cq for a KRAS mutant PCR signal and the KRAS wild-type Cq value, i.e. the ΔCq value, is within a validated range, the sample is characterized as KRAS mutation positive, and the specific mutation or mutation group is indicated. Samples having a valid KRAS wild-type signal but a ΔCq value outside the validated range are reported as being KRAS mutation negative (wild-type). Results expressed as invalid are due to the presence of inhibitors in the sample, insufficient amplifiable DNA present in the sample, incorrect placement of the sample in the cartridge, or a sample volume that is out of range. In addition, incorrectly stored cartridges, use of cartridges that exceeded their in-use period after removal from the pouch, or cartridge malfunctioning can cause invalid results.

The results obtained with the Idylla™ KRAS Mutation Assay on archival material were not used for diagnostic purposes of any kind.

Analytical Sensitivity and Specificity of Idylla™ KRAS Mutation Test

Both analytical sensitivity and specificity have been established for the Idylla™ KRAS Mutation Test (CE IVD). Idylla™ KRAS Mutation Test (CE IVD) and Idylla™ KRAS Mutation Assay (RUO) both consist of the same design and decision tree.

The test enables identifying the presence of a KRAS mutation with a limit of detection (LOD) of ≤5% allelic frequency in a standard Horizon FFPE tissue section for the vast majority of KRAS mutations, while G12A, G13D, and A146P/T/V show LOD values of 9%, 10%, and 16%, respectively (Table 1).

The LOD is defined as the lowest KRAS mutation copy number consistently detected in ≥95% of cases (with 95% confidence), at an allelic frequency of 5%. For KRAS mutations G12A, G12C, G12D, G12R, G12S, G12V, G13D, Q61H, Q61L, and A146T, dilution series were prepared by blending liquefied KRAS mutant FFPE material with liquefied KRAS wild-type FFPE material to the desired allelic frequency, and for mutations Q61K (c.180_181TC>AA), K117N (c.351A>T), and A59E (c.176C>A), synthetic target DNA was spiked to the desired allelic frequency into liquefied KRAS wild-type FFPE material. Two-fold dilution series were prepared in 12-fold and tested with cartridges from two Idylla™ KRAS Mutation Test lots. Mutation call rates were used to determine the LOD by logistic regression, and the corresponding tissue area equivalent of an artificial FFPE specimen with the allelic frequency as indicated (Table 1) was deduced from these copy numbers. LOD values were verified by testing 20 replicates at or below the upper limit of the 95% confidence interval.

In silico analysis of the human genome sequence did not identify reactivity for any of the oligonucleotide primers outside the KRAS gene that could possibly result in non-specific amplicon formation and/or detection, thereby excluding cross-reactivity of Idylla™ KRAS Mutation Test primers with sequence homologues in the NRAS gene or the KRASP1 pseudogene. Screening of single point mutations (SNPs) reported for the human KRAS gene by in silico analysis, did not reveal any mutations outside the codons targeted by the Idylla™ KRAS Mutation Test that could lead to false-positive results for KRAS. Known variants for the KRAS mutations targeted by the Idylla™ KRAS Mutation Test were identified in this in silico SNP analysis to be detected and correctly reported by the assay. The only exceptions that will not be detected are two rarely occurring variants for G12V and Q61R, respectively. Of the known mutations within the codons covered by the Idylla™ KRAS Mutation Test but for which the primers were not primarily designed, mutation G12W will be detected and reported as G12C, and mutations G13N and G13E will result in a G13D call.

Statistical Analysis

For method correlation, the one-sided 95% confidence intervals were calculated using the Wilson method. Percentage agreement was calculated for the Idylla™ KRAS Mutation Assay against routine reference methods. When comparing two methods, overall percent agreement was calculated from the number of specimens tested positive and negative by both methods, and the total number of specimens. Positive percent agreement was calculated from the number of specimens tested positive by both methods, and the total numbers of specimens tested positive for the reference method. Negative percent agreement was calculated from the number of specimens tested negative by both methods, and the total numbers of specimens tested negative for the reference method.

Results

Inter-Laboratory Reproducibility Idylla™ KRAS Mutation Assay

Artificial FFPE samples (i.e., FFPE Reference Standards from Horizon Discovery) were included as external controls: 5 sites tested KRAS G12D at 5% allelic frequency and KRAS G13D at 50% allelic frequency, 2 sites only tested KRAS G12D at 5%, and 1 site only tested KRAS G13D at 50%.

The assay identified each of these mutations correctly (testing 1 replicate per mutation per site). Hence, inter-laboratory reproducibility was 100% (based on 6 out of 6 G13D, and 7 out of 7 G12D).

Idylla™ KRAS Mutation Assay

Using the Idylla™ KRAS Mutation Assay, the KRAS mutational status of 374 archived clinical colorectal cancer FFPE samples was tested at 12 centers. The Idylla™ results were compared with the original assessments made by routine reference methods (Tables 2 and 3, and S1 Table).

thumbnail
Download:
Table 2. Comparison between results of the Idylla™ KRAS Mutation Assay and of routine reference methods.

https://doi.org/10.1371/journal.pone.0163444.t002

thumbnail
Download:
Table 3. Comparison between results of the Idylla™ KRAS Mutation Assay and of routine reference methods (codon level).

https://doi.org/10.1371/journal.pone.0163444.t003

Several routine reference methods were used: cobas® KRAS Mutation Test (Roche; n = 119), Ion Torrent AmpliSeq™ Colon and Lung Cancer Research Panel (Life Technologies, n = 52), therascreen® KRAS Pyro Kit (Qiagen, n = 85), therascreen® RAS Extension Pyro Kit (Qiagen, n = 57), HRM screening (combined with the previous 2 therascreen® kits, n = 34), HRM screening and pyrosequencing (n = 32), HRM screening and Sanger (n = 34), Sanger sequencing (as only method n = 59). Macro-dissection was performed in 141 cases to enrich tumor area, in order to maximize sensitivity.

Of the 374 FFPE samples analyzed by the Idylla™ KRAS Mutation Assay, 231 tested positive for a mutation in KRAS codon 12, 13, 59, 61, 117, or 146 (Table 4A): 13 samples with mutation G12A, 18 with G12C, 57 with G12D, 5 with G12R, 10 with G12S, 39 with G12V, 38 with G13D, 5 with A59T/E/G, 2 with Q61K, 8 with Q61L/R, 11 with Q61H, 4 with K117N, and 21 with A146P/T/V. In 141 cases no mutation was found, and an “invalid” call was reported in 2 cases with the Idylla™ KRAS Mutation Assay and in 1 case with the therascreen® KRAS Pyro Kit (Qiagen) (Table 4A). Overall, in this first assessment, agreement between the Idylla™ results and the results of the routine reference methods was observed in 347 cases. These 347 cases include 3 cases in which a different KRAS mutation was identified by the Idylla™ KRAS Mutation Assay as compared to the routine reference method (Table 2). However, as treatment decisions are driven by the mere absence or presence of a mutation in one of the relevant codons and are not influenced by the exact nature of such a mutation, these 3 Idylla™ results were classified as being concordant.

thumbnail
Download:
Table 4. Comparison between results of the Idylla™ KRAS Mutation Assay and (A) results of routine reference methods, or (B) results of routine reference methods including further analysis.

https://doi.org/10.1371/journal.pone.0163444.t004

In 12 cases, no KRAS mutation was detected by the Idylla™ KRAS Mutation Assay while the reference method had identified a mutation. In another 12 cases, the Idylla™ KRAS Mutation Assay identified a mutation that had not been detected by the routine reference method. These 24 discordant samples were further analyzed (Table 5).

thumbnail
Download:
Table 5. Discordant results between the Idylla™ KRAS Mutation Assay and routine reference methods.

https://doi.org/10.1371/journal.pone.0163444.t005

Discordant Results: Mutation Detected by Routine Reference Method and Not By Idylla™ KRAS Mutation Assay

In FFPE material of 12 colorectal cancer patients, a mutation in the KRAS oncogene previously detected by a routine reference method was not detected when reanalyzing the clinical archival material with the Idylla™ KRAS Mutation Assay (Table 5).

As the Idylla™ KRAS Mutation Assay is not intended to identify mutations G12F, G13C, or G13R, the “no mutation” call for the 6 samples (Bi_01, Bi_07, Co_13, Mo_31, Pa_28, and To_19) harboring one of these mutations was expected. Hence, no further analysis was performed on 5 of these samples. For sample Co_13, the presence of a mutation in codon 13 but with identity G13C was confirmed by NGS.

For sample Au_06, the presence of the A146T mutation detected earlier with Sanger sequencing was confirmed by ddPCR, which detected A146T with 29.7% allelic frequency. Therefore, slices from the same FFPE block were retested three times with the Idylla™ KRAS Mutation Assay using one tenth of the original sample, and each time an A146P/T/V call was obtained. The exact reason why the Idylla™ KRAS Mutation Assay originally missed this mutation was not further investigated.

ddPCR confirmed the presence of mutation G12D in sample Au_18, although at an allelic frequency (0.87%) below the LOD of the Idylla™ KRAS Mutation Assay for this mutation, which may have been the reason for not detecting it with Idylla™. The allelic frequency might have decreased when cutting further into the block.

In sample Lo_23, ddPCR confirmed the presence of G13D at low allelic frequency (0.45%), which is below the LOD of the Idylla™ KRAS Mutation Assay for this mutation (10%).

Mutation G12A in sample Du_26 was not detected by the Idylla™ KRAS Mutation Assay, while its presence was confirmed by both Illumina NGS (7.3% allelic frequency) and ddPCR (4.7% allelic frequency). Because the allelic frequency value of G12A was around the LOD of the Idylla™ KRAS Mutation Assay for this mutation (9%), insufficient sample input might be the cause for not detecting the mutation at this level of allelic frequency. Retesting FFPE slices of this sample with the Idylla™ KRAS Mutation Assay, while using three times more input material, indeed resulted in successful detection of G12A.

In sample Mo_14, ddPCR confirmed the presence of mutation G13D, although at a low allelic frequency (0.11%). Hence, the presence of this mutation in the sample was well below the Idylla™ LOD for this mutation. Additionally, in this sample mutation V14I was identified by HRM and pyrosequencing; the presence of this double mutation (G13D+V14I) in close vicinity of each other most likely has affected detection by the Idylla™ KRAS Mutation Assay as well.

The Idylla™ KRAS Mutation Assay normally detects mutation G12A but not G13R, the presence of both mutations in close vicinity of each other in sample Mo_32 might have hindered the detection of G12A. Moreover ddPCR did not confirm the presence of mutation G12A in sample Mo_32, but this ddPCR result might also have been affected by the presence of the double mutation.

Overall, in 4 samples (Au_18, Lo_23, Mo_14, and Mo_32), a mutation detected by a routine reference method was not detected by the Idylla™ KRAS Mutation Assay. Of note, in three of these cases, the mutation was present at an allelic frequency below the Idylla™ LOD, possibly due to decreasing allelic frequencies when cutting further into the block.

Discordant Results: Mutation Detected by Idylla™ KRAS Mutation Assay and Not By Routine Reference Method

In 12 of the archival clinical FFPE samples tested in this study, the Idylla™ KRAS Mutation Assay identified mutations that were not identified by routine reference methods before (Table 5).

ddPCR analysis of the samples confirmed the mutations found by the Idylla™ KRAS Mutation Assay in 4 samples (Lo_12, Mi_13, Ox_13, and Pa_12). Although A146T was present in an allelic frequency of 25.5% in sample Lo_12, the cobas® KRAS Mutation Test (Roche) was not able to detect this mutation, which is expected as this test is not designed for the identification of mutations in codons 59, 117, and 146 of the KRAS oncogene.

For sample Bi_15, the mutation identified by the Idylla™ KRAS Mutation Assay was confirmed by the KRAS, BRAF, PIK3CA* Array (Randox Molecular) and by the QClamp™ KRAS Codon Specific Mutation Detection Kit (Exon 2, 3, 4) (DiaCarta).

In 3 samples (Mi_03, Mi_07, and Mi_31), ddPCR did not detect the Q61H mutation identified by the Idylla™ KRAS Mutation Assay. Moreover, a repeat test of these samples did not identify this mutation anymore either. Hence, the original discordance between the Idylla™ KRAS Mutation Assay and the routine reference method might have been due to a contamination issue in the laboratory. The results for sample Mi_10 might have been affected by the same issue, however there was no material left for further analysis and possible confirmation of this hypothesis.

For sample Du_05 no further analysis was performed as the “no mutation” call of the cobas® KRAS Mutation Test (Roche) for KRAS exon 4 mutations was expected. The Idylla™ KRAS Mutation Assay identified mutation A146P/T/V.

Concerning sample Bi_04, the Idylla™ KRAS Mutation Assay initially found mutation G12C while the therascreen® KRAS Pyro Kit (Qiagen) and therascreen® RAS Extension Pyro Kit (Qiagen) did not find this mutation. The “no mutation” finding was confirmed by the KRAS, BRAF, PIK3CA* Array (Randox Molecular) and QClamp™ KRAS Codon Specific Mutation Detection Kit (Exon 2, 3, 4) (DiaCarta). Retesting with the Idylla™ KRAS Mutation Assay did no longer find the G12C mutation.

In sample Ox_28, the initial cobas® KRAS Mutation Test result was confirmed by ddPCR.

In total, in 5 samples (Bi_15, Lo_12, Mi_13, Ox_13, and Pa_12), the Idylla™ KRAS Mutation Assay was able to identify a mutation that was not identified by the first-line routine reference method but that was confirmed during further analysis.

Performance of the Idylla™ KRAS Mutation Assay

Further analysis enabled the identification of 8 samples (Bi_15, Mi_03, Mi_07, Mi_13, Mi_31, Mo_32, Ox_13, and Pa_12) for which the Idylla™ KRAS Mutation Assay result was confirmed by an independent reference method, and in 3 samples (Au_06, Bi_04, and Du_26) Idylla™ KRAS Mutation Assay retesting results were in agreement with the reference method result. Therefore, the number of concordant samples was increased from 347 to 358 upon discordance analysis.

The Idylla™ KRAS Mutation Assay is not intended to identify the rare mutations G12F, G13C, or G13R (<0.5% prevalence in colorectal cancer), which were previously detected in 6 archival FFPE samples by routine reference methods. Despite the correct “no mutation detected” Idylla™ call, these 6 samples (Bi_01, Bi_07, Co_13, Mo_31, Pa_28, and To_19) were excluded from the dataset. Likewise, samples Du_05 and Lo_12 were excluded from the dataset as the reference method used is not designed to detect the KRAS exon 4 mutation identified by the Idylla™ KRAS Mutation Assay. Sample Mi_10 is removed from the dataset as there was no material left to confirm the mutation found by the Idylla™ KRAS Mutation Assay, while a contamination issue was suspected. Also the samples with invalid results for the Idylla™ KRAS Mutation Assay (2 samples) or the reference method (1 sample) were removed from the dataset. As a result, the total number of samples was reduced to 362.

Taking into account the additional concordant samples and the decreased number of samples in the dataset, the results obtained by the Idylla™ KRAS Mutation Assay were in agreement with the confirmed reference method results in 358 out of 362 samples, resulting in an overall concordance of 98.89% (95% confidence interval: 97.56% - 100%) (Table 4B). The negative percent agreement was 99.25% (95% confidence interval: 96.75% - 100%) and positive percent agreement was 98.68% (95% confidence interval: 97.74% - 100%).

Discussion

The innovative Idylla™ KRAS Mutation Test, performed on the molecular diagnostic Idylla™ platform (Biocartis, Mechelen, Belgium), enables the qualitative detection in human colorectal cancer FFPE material of 21 KRAS mutations being clinically relevant according to the latest CAP/AMP/ASCO, ESMO, and NCCN guidelines [23]. These mutations encompass codons 12, 13, 59, 61, 117, and 146 of the KRAS oncogene.

Here, an analysis comparing the Idylla™ KRAS Mutation Assay (RUO) results with the original assessments made by routine reference methods for determining the KRAS mutational status on FFPE samples is reported. Twelve clinical centers contributed to this study, and a total of 374 archival clinical colorectal cancer FFPE samples were tested. The samples were specifically selected for this study; hence the mutational rates in the current sample set are not representative for the incidence of KRAS mutations in the general colorectal cancer patient population. The results showed an overall presence of KRAS mutations in the samples of 62.0% when measured with routine reference methods and of 61.8% measured with the Idylla™ KRAS Mutation Assay, demonstrating that the sensitivity of the Idylla™ KRAS Mutation Assay is comparable to the sensitivity currently used routine reference methods. The overall concordance between the Idylla™ KRAS Mutation Assay and the confirmed reference routine test results for colorectal cancer samples was found to be 98.9%, with a negative percent agreement of 99.25% and a positive percent agreement of 98.68%.

The initial results showed discordances between results of the Idylla™ KRAS Mutation Assay and of the original routine reference method for 24 of the 374 samples. In 12 of these 24 cases, the Idylla™ KRAS Mutation Assay did not identify the KRAS mutation that was previously detected by the reference method, while in another 12 cases the Idylla™ KRAS Mutation Assay identified a mutation not detected by the routine reference method. These 24 discordant samples were analyzed further to establish the overall concordance. Compared to pyrosequencing (including in-house methods, and the therascreen® KRAS Pyro Kit and RAS Extension Pyro Kit [Qiagen]), the cobas® KRAS Mutation Test (Roche), and NGS (Ion Torrent AmpliSeq™ Colon and Lung Cancer Research Panel [Life Technologies]), the observed concordances varied between 91.5% and 100% for these initial results.

Several reasons were found why the Idylla™ KRAS Mutation Assay did not identify KRAS mutations detected earlier by routine reference methods. Six of the 12 specimens displayed G13C, G12F, or G13R KRAS exon 2 mutations, which are not included in the Idylla™ KRAS Mutation Assay because they are rarely observed (<0.5%) in the large colorectal cancer cohorts published [24–26]. Four other samples exhibited G12A, G12D, or G13D mutations according to the reference results, but the observed allelic frequencies were below the LOD values of the Idylla™ KRAS Mutation Assay (i.e., 9% for G12A, 5% for G12D, and 10% for G13D; Table 1); in one of these samples, retesting with three times more input material indeed detected the mutation. A decrease in allelic frequency when cutting further into the block might be an issue in particular when using archival material. In one sample, the presence of an additional mutation may have hindered the detection of the G12A mutation by the Idylla™ KRAS Mutation Assay; interestingly, ddPCR could not confirm the presence of mutation G12A in this sample either, demonstrating that the presence of a double mutation located in two consecutive codons can affect allele-specific-based techniques. In addition, co-mutations for G12D and G13D have exceptionally been described in colorectal cancer.

The other way around, the Idylla™ KRAS Mutation Assay identified mutations that were not identified by routine reference methods in 12 archival clinical FFPE samples. Three of these cases exhibited mutations with too low allelic mutation frequencies to be detected by the used reference method (HRM, Sanger sequencing, or cobas® KRAS Mutation Test), and two other cases displayed mutations not intended to be detected by the routine approach (i.e., KRAS exon 4 mutations by the cobas® KRAS Mutation Test). When considering all KRAS mutations, there was a strong agreement in detection rates between the Idylla™ KRAS Mutation Assay and the routine reference methods. Overall, in 5 archival clinical samples, the Idylla™ KRAS Mutation Assay detected KRAS mutations (confirmed by further analysis) which were not detected by the routine reference methods previously used to guide therapy.

Of note, the Idylla™ KRAS Mutation Assay identified different KRAS mutations in three cases as compared to routine reference methods, but since treatment decisions are driven by the presence of a mutation in the relevant codons and not by its exact nature, these results were considered concordant.

According to manufacturer’s instructions, the minimal tumor cell percentage in samples for the Idylla™ KRAS Mutation Assay should be at least 25% to guarantee reliable results. In case this lower limit is not met, macro-dissection should be performed in order to enrich tumor area (and thus tumor cells) and to maximize sensitivity. In the current study, macro-dissection was performed in 141 cases.

It is now well established that the frequency of detected KRAS mutations in colorectal cancer is influenced by the analytical sensitivity of the method applied for their detection. The sensitivity of the Idylla™ KRAS Mutation Assay is comparable to most of the routine approaches used in laboratories such as competitive allele-specific Taqman, amplification refractory mutation system-PCR, HRM, and pyrosequencing. These reference methods require process steps including deparaffinization, DNA extraction, PCR amplification, and mutation detection, which does not allow on demand testing and therefore leads to batch testing. Hence turnaround times are more than a day, and even several days in the case of NGS. In addition, they are labor-intensive, and require a specialized staff and complex infrastructure. The Idylla™ KRAS Mutation Assay is a simple and fast procedure that does not entail any special additional equipment, has a very short hands-on time (i.e., less than 2 minutes) and has an estimated time of approximately 2 hours from FFPE sample to the final reporting. Since the response time for analyzing KRAS status might be a critical factor, in particular for patients with rapid disease progression, the Idylla™ KRAS Mutation Assay could offer new prospects by significantly decreasing turnaround time. Given the fact that the Idylla™ KRAS Mutation Assay can be easily implemented in any pathology laboratory setting, even those without prior molecular diagnostics experience, and needs a very short turnaround time to obtain test results, incorporation of the KRAS mutational status in the initial histological report is possible, enabling immediate initiation of targeted therapy. Although the Idylla™ KRAS Mutation Assay is not intended to determine all possible KRAS mutations or to quantify allelic frequency, the test is capable of detecting the great majority of the relevant and most prevalent KRAS mutations. However, other techniques should be utilized to obtain quantitative information or complete overviews of mutations present, or to discover new useful tumor biomarkers.

The Idylla™ KRAS Mutation Assay is a fully-automated and integrated real-time-PCR-based test with high sensitivity, offering a “one-step” solution for routine application. Up-front sample preparation, apart from possible macro-dissection, is not necessary as FFPE tissue sections are directly loaded into the single-use cartridge. As well, the interpretation of the results is fully automated. Due to the platform characteristics, there is no need for additional molecular infrastructure or a highly skilled staff to perform the test on site. Interestingly, the closure of the leak-proof cartridge after loading the sample limits the risk of cross contamination between different samples. In large molecular laboratories, the Idylla™ KRAS Mutation Assay is considered a useful addition to other technologies including NGS, in particular for urgent requests where a very quick turnaround time is needed for occasional patients.

Conclusion

In conclusion, the Idylla™ KRAS Mutation Test on the Idylla™ platform is to our knowledge the only fully integrated and automated test for detection of 21 clinically relevant mutations included in current guidelines starting from FFPE tissue samples and represents a relevant solution for simple, highly reliable, and rapid routine determination of the KRAS mutational status needed to guide colorectal cancer therapy.

Supporting Information

S1 Table. Overview of all samples analyzed, with results of Idylla™ and of routine reference methods.

https://doi.org/10.1371/journal.pone.0163444.s001

(DOCX)

Acknowledgments

The authors wish to thank Hélène Frugier and Frédéric Bibeau (Laboratoire d’Anatomo-Pathologie, Institut du Cancer de Montpellier, Montpellier, France), Phillipe Taniere (Queen Elizabeth Hospital, Birmingham, UK), Erica Torres and Ester Moragón (Pathology Department, Hospital del Mar, Barcelona, Spain), Lai Mun Wang (Department of Cellular Pathology, Oxford University Hospitals NHS Foundation Trust, Oxford, UK), Luc Geeraert (Bench to Pen Scientific Writing, Louvain, Belgium; for his help in the preparation of this manuscript), and the Biocartis team (for their support and reviewing of the manuscript). Work at the Hospital del Mar Medical Research Institute (Barcelona, Spain) was partially funded by a Grant for Oncology Innovation (GOI), which is an initiative funded by EMD Serono (Merck KGaA, Darmstadt, Germany).

Author Contributions

  1. Conceptualization: JS PS.
  2. Data curation: JS JV BM CH BB C. Montagut MS BO ND MLM MG LCM EB FC DS FDLS VM CL RC ES MF C. Mescoli CC JCP.
  3. Formal analysis: JS PS.
  4. Investigation: JS JV BM CH BB C. Montagut MS BO ND MLM MG LCM EB FC DS FDLS VM CL RC ES MF C. Mescoli CC JCP.
  5. Methodology: JS JV BM CH BB C. Montagut MS BO ND MLM MG LCM EB FC DS FDLS VM CL RC ES MF CM CC JCP.
  6. Project administration: PS.
  7. Resources: JS JV BM CH BB C. Montagut MS BO ND MLM MG LCM EB FC DS FDLS VM CL RC ES MF C. Mescoli CC JCP PS.
  8. Supervision: JS PS.
  9. Validation: JS JV BM CH BB C. Montagut MS BO ND MLM MG LCM EB FC DS FDLS VM CL RC ES MF CM CC JCP.
  10. Visualization: JS.
  11. Writing – original draft: JS.
  12. Writing – review & editing: JS JV BM CH BB C. Montagut MS BO ND MLM MG LCM EB FC DS FDLS VM CL RC ES MF C. Mescoli CC JCP PS.

References

  1. 1. Malumbres M, Barbacid M. RAS oncogenes: the first 30 years. Nat Rev Cancer. 2003 Jun;3(6):459–65. pmid:12778136
  2. 2. Baines AT, Xu D, Der CJ. Inhibition of Ras for cancer treatment: the search continues. Future Med Chem. 2011 Oct;3(14):1787–808. pmid:22004085
  3. 3. Colussi D, Brandi G, Bazzoli F, Ricciardiello L. Molecular pathways involved in colorectal cancer: implications for disease behavior and prevention. Int J Mol Sci. 2013 Aug 7;14(8):16365–85. pmid:23965959
  4. 4. Kocarnik JM, Shiovitz S, Phipps AI. Molecular phenotypes of colorectal cancer and potential clinical applications. Gastroenterol Rep (Oxf). 2015 Nov;3(4):269–76. pmid:26337942
  5. 5. Kuipers EJ, Grady WM, Lieberman D, Seufferlein T, Sung JJ, Boelens PG, et al. Colorectal cancer. Nat Rev Dis Primers. 2015 Nov 5;1:15065. pmid:27189416
  6. 6. Peeters M, Kafatos G, Taylor A, Gastanaga VM, Oliner KS, Hechmati G, et al. Prevalence of RAS mutations and individual variation patterns among patients with metastatic colorectal cancer: A pooled analysis of randomised controlled trials. Eur J Cancer. 2015 Sep;51(13):1704–13. pmid:26049686
  7. 7. Ferlay J, Steliarova-Foucher E, Lortet-Tieulent J, Rosso S, Coebergh JW, Comber H, et al. Cancer incidence and mortality patterns in Europe: estimates for 40 countries in 2012. Eur J Cancer. 2013 Apr;49(6):1374–403. pmid:23485231
  8. 8. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2015. CA Cancer J Clin. 2015 Jan-Feb;65(1):5–29. pmid:25559415
  9. 9. Van Cutsem E, Cervantes A, Nordlinger B, Arnold D; ESMO Guidelines Working Group. Metastatic colorectal cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2014 Sep;25 Suppl 3:iii1–9. pmid:25190710
  10. 10. Hocking CM, Price TJ. Panitumumab in the management of patients with KRAS wild-type metastatic colorectal cancer. Therap Adv Gastroenterol. 2014 Jan;7(1):20–37. pmid:24381645
  11. 11. Amado RG, Wolf M, Peeters M, Van Cutsem E, Siena S, Freeman DJ, et al. Wild-type KRAS is required for panitumumab efficacy in patients with metastatic colorectal cancer. J Clin Oncol. 2008 Apr 1;26(10):1626–34. pmid:18316791
  12. 12. Karapetis CS, Khambata-Ford S, Jonker DJ, O'Callaghan CJ, Tu D, Tebbutt NC, et al. K-ras mutations and benefit from cetuximab in advanced colorectal cancer. N Engl J Med. 2008 Oct 23;359(17):1757–65. pmid:18946061
  13. 13. Douillard JY, Oliner KS, Siena S, Tabernero J, Burkes R, Barugel M, et al. Panitumumab-FOLFOX4 treatment and RAS mutations in colorectal cancer. N Engl J Med. 2013 Sep 12;369(11):1023–34. pmid:24024839
  14. 14. Hecht JR, Douillard JY, Schwartzberg L, Grothey A, Kopetz S, Rong A, et al. Extended RAS analysis for anti-epidermal growth factor therapy in patients with metastatic colorectal cancer. Cancer Treat Rev. 2015 Sep;41(8):653–9. pmid:26220150
  15. 15. Sorich MJ, Wiese MD, Rowland A, Kichenadasse G, McKinnon RA, Karapetis CS. Extended RAS mutations and anti-EGFR monoclonal antibody survival benefit in metastatic colorectal cancer: a meta-analysis of randomized, controlled trials. Ann Oncol. 2015 Jan;26(1):13–21. pmid:25115304
  16. 16. NCCN clinical practice guidelines in oncology: colon cancer v.3.2015. National Comprehensive Cancer Network. Available: http://www.nccn.org/professionals/physician_gls/pdf/colon.pdf; 2015.
    • 17. Westwood M, van Asselt T, Ramaekers B, Whiting P, Joore M, Armstrong N, et al. KRAS mutation testing of tumours in adults with metastatic colorectal cancer: a systematic review and cost-effectiveness analysis. Health Technol Assess. 2014 Oct;18(62):1–132. pmid:25314637
    • 18. Ciardiello F, Tejpar S, Normanno N, Mercadante D, Teague T, Wohlschlegel B, et al. Uptake of KRAS mutation testing in patients with metastatic colorectal cancer in Europe, Latin America and Asia. Target Oncol. 2011 Sep;6(3):133–45. pmid:21630045
    • 19. Jouret-Mourin A, Cuvelier C, Demetter P, D’Haene N, Driessen A, Hoorens A, et al. RAS-testing in colorectal cancer: Belgian guidelines. Belg J Med Oncol. 2015 9(5):183–90.
    • 20. Melchior L, Grauslund M, Bellosillo B, Montagut C, Torres E, Moragón E, et al. Multi-center evaluation of the novel fully-automated PCR-based Idylla™ BRAF Mutation Test on formalin-fixed paraffin-embedded tissue of malignant melanoma. Exp Mol Pathol. 2015 Dec;99(3):485–91. pmid:26407762
    • 21. Schiefer AI, Parlow L, Gabler L, Mesteri I, Koperek O, von Deimling A, et al. Multicenter Evaluation of a Novel, Automated Rapid Detection System of BRAF Status in Formalin-Fixed, Paraffin-Embedded Tissues. J Mol Diagn. 2016 May;18(3):370–7. pmid:26921540
    • 22. D'Haene N, Le Mercier M, De Nève N, Blanchard O, Delaunoy M, El Housni H, et al. Clinical Validation of Targeted Next Generation Sequencing for Colon and Lung Cancers. PLoS One. 2015 Sep 14;10(9):e0138245. pmid:26366557
    • 23. Association for Molecular Pathology. ASCP/CAP/AMP/ASCO Evaluation of Molecular Markers for Colorectal Cancer Draft Recommendations. Available: http://www.amp.org/committees/clinical_practice/CRCOpenComment.cfm; 2015. Refers to Guideline on the Evaluation of Molecular Markers for Colorectal Cancer Expert Panel Draft Recommendations Summary for Open Comment Period.
      • 24. Tong JH, Lung RW, Sin FM, Law PP, Kang W, Chan AW, et al. Characterization of rare transforming KRAS mutations in sporadic colorectal cancer. Cancer Biol Ther. 2014 Jun 1;15(6):768–76. pmid:24642870
      • 25. Cox AD, Fesik SW, Kimmelman AC, Luo J, Der CJ. Drugging the undruggable RAS: Mission possible? Nat Rev Drug Discov. 2014 Nov;13(11):828–51. pmid:25323927
      • 26. Jiang L, Huang J, Morehouse C, Zhu W, Korolevich S, Sui D, et al. Low frequency KRAS mutations in colorectal cancer patients and the presence of multiple mutations in oncogenic drivers in non-small cell lung cancer patients. Cancer Genet. 2013 Sep-Oct;206(9–10):330–9. pmid:24200637