Key messages

  • Roche Molecular Systems recently launched a new real-time polymerase chain reaction test, the cobas® Factor II and Factor V Test, to detect common mutations associated with thrombophilia [13]

  • The cobas F2F5 test allows for multiplex testing with flexible reporting and a user-selected sample DNA extraction method

  • Technical performance and clinical validation studies found 100% agreement between genotypes reported by the cobas F2F5 test and reference Sanger sequencing

  • Compared with the LightCycler® 1.2 platform–based testing method, the cobas F2F5 test also reduces the total processing time and number of steps required in amplification and detection

  • These results show that the new cobas F2F5 test is both time saving and cost-efficient and provides a high level of accuracy in Factor II and Factor V genotype identification

Introduction

Thrombophilia, characterized by a predisposition to the development of thrombi in veins, arteries, or both, is the result of either inherited or acquired defects (or an interaction between the 2) in the coagulation system [1, 2]. The most commonly associated genetic mutations for inherited thrombophilia are gain-of-function mutations in the genes for Factor V and Factor II (prothrombin). The American College of Medical Genetics and recent guidelines from other professional societies recommend genetic testing only in certain circumstances for patients with suspected hereditary venous thrombosis [3, 4]. Patients who test positive for Factor V Leiden should be considered for Factor II G20210A testing, particularly pregnant women with a previous history of venous thromboembolism and recurrent pregnancy loss [3,4,5].

Several different laboratory testing methods are currently available for Factor II and Factor V mutation detection [6]. Based on internal data from market assessments and sales, we estimate that approximately 2.2 million Factor II and Factor V genotyping tests are performed annually worldwide (excluding Japan) [7]. The laboratory costs of thrombophilia testing are estimated to exceed $650 million (USD) annually [8]. Therefore, there is an urgent need to optimize screening and diagnosis algorithms, make testing more cost-effective for laboratories, and reduce the frequency of inappropriate testing [9, 10].

In 2003, Roche Molecular Diagnostics launched 2 qualitative real-time polymerase chain reaction (PCR)-based diagnostic tests for use with the LightCycler® 1.2 Instrument—the Factor II G20210A Kit and the Factor V Leiden Kit—each allowing for detection and genotyping of a single point mutation from whole-blood genomic DNA (gDNA) [11, 12]. In 2017, keeping with evolving laboratory practices and technological advances, the cobas® Factor II and Factor V Test (cobas F2F5 test) was launched for use with the cobas z 480 analyzer. This test not only allows for multiplex testing but also flexible reporting—either both mutations together or individually—depending on testing requirements. The testing algorithm was optimized to distinguish between wild-type and mutant alleles and designed to prevent the occurrence of incorrect results. For greater flexibility prior to amplification and detection, DNA extraction from whole blood can be performed offline using a user-selected manual or automated method. Multiplex testing also improves cost-efficiency by reducing the testing time and minimizing the risk of error. Figure 1 summarizes the differences between the earlier LightCycler and the new cobas z 480 platforms.

Fig. 1
figure 1

Comparison of the 2 Roche Molecular Systems qualitative polymerase chain reaction (qPCR) methods for Factor II and Factor V detection using the LightCycler and cobas z 480 instruments. FII Factor II, FV Factor V, K2EDTA dipotassium ethylenediaminetetraacetic acid, LIS laboratory information system

Full size image

Here, we report the results from technical performance verification, clinical validation, and external laboratory performance studies for the cobas F2F5 test.

Methods

The overall workflow for Factor II and Factor V testing using the cobas F2F5 test platform is summarized in Fig. 2.

Fig. 2
figure 2

Overall workflow for Factor II and Factor V identification using the cobas z 480 platform. K2EDTA dipotassium ethylenediaminetetraacetic acid

Full size image

Technical performance verification

As part of the development of the cobas F2F5 test, studies were designed to investigate the performance characteristics of the test, such as the limit of detection, analytical specificity, stability (including samples and reagents), and substances that may interfere with the performance of the test. Further details about the studies are provided in Supplementary Table 1.

Clinical validation studies

Method comparison

Frozen whole-blood and DNA specimens (n = 300) were obtained from laboratories that performed Factor II and/or Factor V testing per routine medical care. Genotypes tested included Factor II or Factor V homozygous mutations, single or compound heterozygous mutations, and wild-type samples. Factor II homozygous mutant samples (n = 16) were available as gDNA only. Manual gDNA isolation using the High Pure (HP) PCR Template Preparation Kit (Roche Molecular Systems Inc, Pleasanton, CA) and cobas F2F5 testing were carried out at an external site. Bidirectional Sanger DNA sequencing was carried out at a commercial laboratory.

Clinical reproducibility

Nine panel members were tested at three sites, with each site using a different commercially available manual DNA extraction kit: HP kit (Roche Molecular Systems Inc, Pleasanton, CA), ReliaPrep™ Blood gDNA Miniprep System (Promega, Madison, WI), and QIAamp® DSP DNA Blood Mini Kit (Qiagen, Venlo, the Netherlands). The nine panel members tested consisted of four dipotassium ethylenediaminetetraacetic acid (K2EDTA) whole-blood samples, three contrived whole-blood samples, and two gDNA samples (diluted to 0.2 ng/µL). Each testing site had two operators and one instrument. Each operator performed 1 run per day over 5 non-consecutive days. A total of 18 cobas F2F5 tests were carried out for each panel member (in duplicate) for each reagent lot.

External laboratory performance testing

Performance of the cobas F2F5 test in combination with automated DNA extraction methods was assessed using frozen and fresh clinical samples. Samples were de-identified prior to processing and testing. Banked (frozen) whole-blood specimens (n = 200)—a subpopulation of samples used for the method comparison clinical validation study (using manual DNA isolation; outlined above)—were provided by Roche Molecular Systems. gDNA was isolated using both the MagNA Pure 24 (MP 24) and MagNA Pure 96 (MP 96) automated DNA isolation methods. Overall percentage agreement (OPA) was assessed against archived Sanger sequencing results and between DNA isolation methods. Remnant clinical samples (fresh, never frozen; n = 200) from routine testing were provided by an external study site and DNA was isolated using the MP 96 isolation method. The cobas F2F5 test results were compared with results from routine Factor II and Factor V testing methods used at the external site (Roche Factor II prothrombin G20210A kit and Roche Factor V Leiden Kit coupled to the LightCycler® 1.2 instrument).

Workflow comparison

The workflow for cobas Factor II and Factor V testing using the cobas z 480 platform was compared with the LightCycler® 1.2 platform. The hands-on and total processing times in addition to the number of steps from specimen receipt to release of results were observed and documented for different batch sizes. DNA isolation was carried out using the MP 96 instrument prior to testing with both platforms. Figure 3 shows representative steps in the workflow comparison study using the example of the MP 96 instrument paired with the cobas z 480 for the new cobas F2F5 test (MP 96 paired with LightCycler steps not shown).

Fig. 3
figure 3

Overview of the workflow using the MP 96 and cobas z 480 platform for the cobas® Factor II and Factor V Test. MP 96, MagNA Pure System

Full size image

Study ethics

All studies were conducted in compliance with their protocols. Clinical studies were conducted with the International Conference on Harmonisation Good Clinical Practice Guidelines and regulations of the US Food and Drug Administration (FDA). Where appropriate, study protocols were submitted to an institutional review board in accordance with FDA and local regulatory requirements prior to the start of the study.

Results

Technical performance verification studies

All studies met predefined (prior to test development) acceptance criteria. The aims and key results from each technical performance study are outlined in Table 1.

Table 1 Summary of results from technical performance verification studies for the cobas F2F5 test
Full size table

Clinical validation studies

Method comparison

All runs and sample results were valid (no repeat tests were needed). A genotype result was classified as correct if the same genotype was detected by both the cobas F2F5 test and the reference method, bidirectional Sanger sequencing.

For Factor II testing (n = 300), the OPA with bidirectional Sanger sequencing was 100%, with a 2-sided 95% lower confidence boundary (LCB; exact method) of 98.78%. The negative percentage agreement (NPA; for wild-type samples) was 100% (95% LCB, 97.55%; n = 149). The positive percentage agreement (PPA; for Factor II mutation-positive samples) was also 100% (95% LCB, 97.59%; n = 151). Of the 151 positive samples, the percentage agreement for both heterozygous and homozygous mutant samples was 100% each (95% LCB, 97.20%; n = 130 and 95% LCB, 83.89%; n = 21, respectively).

Similarly, for Factor V testing, the OPA was 100% (95% LCB, 98.78%; n = 300), NPA was 100% (95% LCB, 97.52%; n = 147), and PPA was 100% (95% LCB, 97.62%; n = 153). For the positive samples, the percentage agreement for both heterozygous and homozygous mutant samples was 100% each (95% LCB, 97.20%; n = 130 and 95% LCB, 85.18%; n = 23, respectively).

Clinical reproducibility

Four hundred twenty whole-blood samples (240 clinical and 180 contrived) and 120 gDNA samples were tested. A total of 540 tests obtained from 30 valid runs yielded 539 valid results. One invalid result was due to a lack of agreement in the data parameter checks for software result interpretation, designed to prevent inaccurate result reporting. The invalid rate for whole-blood samples in the study was 0.24% (1 of 420).

The cobas F2F5 test demonstrated 100% correct call rates for each Factor II and Factor V genotype tested (Table 2).

Table 2 Summary of results from the reproducibility study by site, lot, sample preparation method, and genotype for Factor II and Factor V testing using the cobas® Factor II and Factor V Test
Full size table

External laboratory performance testing

For banked (frozen) specimens, the cobas F2F5 test results agreed with the genotypes determined by Sanger sequencing for 100% of all genotypes tested regardless of the DNA isolation method used, manual or automated (n = 200; 95% LCB, 98.2%). For Factor II genotypes, the percentage agreement was 100% each for wild-type (n = 103), heterozygous mutant (n = 94), and homozygous mutant (n = 3) samples. For Factor V genotypes, the percentage agreement was also 100% each for wild-type (n = 87), heterozygous mutant (n = 96), and homozygous mutant (n = 17) samples.

Results from remnant whole-blood specimens tested with the MP 96 platform showed that the cobas F2F5 test generated 100% correct calls and 100% OPA compared with the LightCycler method for all Factor II and Factor V genotypes tested (95% LCB, 98.17%; n = 200). For Factor II genotypes, the percentage agreement was 100% for both wild-type (n = 196) and heterozygous mutant (n = 4) samples (no homozygous mutant samples were tested). Similarly, for Factor V genotypes, the percentage agreement was 100% each for wild-type (n = 178), heterozygous mutant (n = 21), and homozygous mutant (n = 1) samples.

Workflow overview study

Results for the hands-on and total processing times in the pre-analytic (gDNA isolation) phase are provided in Supplementary Table 2.

Compared with the LightCycler® 1.2–based method, the cobas F2F5 test required less hands-on time (141.2 vs 8.6 min), less total processing time (420.2 vs 98.6 min), and fewer steps for amplification and detection (1091 vs 303; n = 96 samples). The processing times for 24 and 48 samples are provided in Supplementary Table 2.

Discussion

In this paper, we report results from several studies evaluating the performance characteristics of the new cobas F2F5 test for thrombophilia. In all, these studies show that the cobas F2F5 test is robust and has a very high degree of accuracy in determining Factor II and Factor V genotypes from frozen and fresh whole-blood and gDNA samples.

The minimum recommended gDNA input for the cobas F2F5 test is 0.1 ng/μL, which is close to that observed for > 99% of gDNA samples used in our testing (≥ 0.2 ng/μL). However, correct results were generated at a concentration that is 10 times lower than recommended (Table 1). The maximum recommended gDNA concentration is 150 ng/μL, which is a conservative recommendation since correct results were generated using a concentration up to 300 ng/μL (Table 1). No incorrect genotypes were reported in these studies.

One gDNA plasma sample with a rust color, most likely due to hemoglobin (a polymerase chain reaction inhibitor), yielded an invalid result in the DNA extraction method study (Table 1). It is recommended, therefore, that the appearance of gDNA isolated from whole blood should be clear and colorless; samples that contain a red-colored tint may yield invalid or incorrect results. Extraction buffers and ethanol in commercial DNA isolation kits can also cause invalid results (Table 1). These buffers typically contain chaotropic salts (eg, guanidine hydrochloride), which are potent denaturants and inhibit DNA polymerase. However, any residual extraction buffer is typically removed during the wash steps between lysis and elution in commercial DNA isolation kits. Ethanol is a common ingredient in the final wash buffer, and it is recommended that both the lysis and wash buffers are completely removed from the sample before running the cobas F2F5 test.

Eight known single-nucleotide polymorphisms close to the Factor II or Factor V locus were tested with the cobas F2F5 test and no false positives were detected. In the rare event that a Factor V 1689 G > A mutation is present on both Factor V alleles, the cobas F2F5 test will yield an invalid result. All the aforementioned considerations are included in the cobas® Factor II and Factor V Test kit Instructions for Use.

In these studies, a variety of commercially available manual and automated DNA isolation methods were used, and the cobas F2F5 test identified Factor II and Factor V genotypes with a high degree of accuracy. The reproducibility study found 100% agreement between the cobas F2F5 test and Sanger sequencing for all genotypes and manual DNA isolation methods tested. Similarly, external laboratory performance testing found 100% agreement for all frozen and whole-blood specimens isolated using both manual and automated DNA extraction methods. These results highlight the “open front-end” capability of the cobas F2F5 test, which allows the user to select any gDNA isolation method, provided it meets the recommended minimum gDNA quantity [13].

Compared with the preceding LightCycler® 1.2 platform, the cobas F2F5 test and cobas z 480 platform reduced the hands-on and total processing times for 96 samples by 16- and fourfold, respectively. The total number of steps was also reduced by approximately fourfold, decreasing the likelihood of user processing errors.

Factor II and Factor V mutation testing is expensive, time consuming [7, 10], and associated with errors that can lead to misdiagnosis or overtreatment [14]. The Italian Committee for the Standardization of Laboratory Tests reported a 4.9–17.0% error rate for Factor V Leiden and 4.9–19.5% error rate for Factor II G20210 identification [15]. The Royal College of Pathologists of Australasia reported a 98.63% success rate for Factor V and Factor II mutation identification but found large variations between laboratories: 51% made ≥ 1 error and 3 of 39 laboratories were responsible for nearly half the reported errors [16]. Results of genetic analyses can impact clinical decision making and patient levels of concern, thus testing laboratories should adhere to a high level of internal quality control and participate in external quality assurance programs [16].

In conclusion, the new cobas F2F5 test can improve the efficacy of thrombophilia testing through its established accuracy in genotype identification. In addition, the multiplex method and short processing time can reduce the cost and time associated with Factor II and Factor V testing. Lastly, the flexible reporting and “open front-end” design can enable easy integration into established testing procedures in laboratories.