(255 days)
The eSensor® Warfarin Sensitivity Saliva Test is an in vitro diagnostic for the detection and genotyping of the *2 and *3 alleles of the cytochrome P450 (CYP450) 2C9 gene locus and the Vitamin K epoxide reductase C1 (VKORC1) gene promoter polymorphism (-1639G>A) from genomic DNA extracted from human saliva samples collected using the Oragene® Dx and ORAcollect® Dx devices, as an aid in the identification of patients at risk for increased warfarin sensitivity.
The eSensor® XT-8 instrument is an in vitro diagnostic device intended for genotyping multiple mutations or polymorphisms in an amplified DNA sample utilizing electrochemical detection technology.
The kit consists of the eSensor® Warfarin Sensitivity Saliva Test cartridge, the eSensor® Warfarin Sensitivity Saliva Test amplification reagents (including PCR mix and DNA polymerase), the eSensor® Warfarin Sensitivity Saliva Test detection reagents (including exonuclease, probes and hybridization buffer ingredients) and the eSensor® XT-8 System. One eSensor® Warfarin Sensitivity Saliva Test Kit has sufficient materials for 24 tests.
The provided document, a 510(k) summary for the eSensor® Warfarin Sensitivity Saliva Test, details performance data primarily focused on a new specimen collection kit (ORAcollect®·Dx Device) and its impact on the existing device's performance. The study aims to demonstrate that incorporating this new collection kit does not compromise the accuracy or reliability of the Warfarin Sensitivity Saliva Test.
Here's an analysis of the acceptance criteria and study findings:
1. Table of Acceptance Criteria and Reported Device Performance
The document doesn't explicitly state 'acceptance criteria' in terms of specific numerical thresholds for accuracy, reproducibility, or call rates. However, the studies consistently aim for 100% agreement with the ground truth (bi-directional sequencing), and 100% correct calls with 0% incorrect calls. The "Method Comparison" section indicates an aspiration for high concordance and call rate.
| Metric (Implicit Acceptance Criterion) | Reported Device Performance |
|---|---|
| Reproducibility | |
| Sample-to-Sample, Lot-to-Lot, Day-to-Day, Operator-to-Operator Agreement with Sequencing (CYP2C92, CYP2C93, VKOR) | 100% Agreement (60/60 correct calls across 3 operators for each SNP, totaling 180 successful calls out of 180 attempts for each SNP, or 120/120 for each donor when considering total calls). |
| Multi-center Reproducibility Agreement with Sequencing | 100% Agreement (30/30 for Site 1, 30/30 for Site 2, 29/29 for Site 3, after final pass, excluding one sample that did not meet input criteria). |
| Method Comparison | |
| Overall Concordance with DNA Sequencing | 99.4% concordance for all polymorphisms (after retests). |
| First-pass Call Rate | 98.1% |
| Final Pass Call Rate | 99.4% |
| % Agreement per Genotype (post-retest) | 2C9*2: 99.0% (wt/wt), 100.0% (wt/*2), 100.0% (*2/*2) |
| 2C9*3: 100.0% (wt/wt), 95.0% (wt/*3), 100.0% (*3/*3) | |
| VKORC1: 98.4% (G/G), 100.0% (G/A), 100.0% (A/A) | |
| (Note: Lower 95% LCBs for less common genotypes due to smaller sample sizes). | |
| Interference Studies | |
| Endogenous Interfering Substances Agreement with Sequencing | 100% Agreement (for Amylase, Hemoglobin, IgA, Total Protein - all 14 samples tested per substance). |
| Exogenous Interfering Substances Agreement with Sequencing | 100% Agreement (for Eating, Drinking, Chewing Gum, Smoking, Mouthwash, Brushing Teeth at immediate and 30-minute time points - sample sizes ranging from 5 to 9 per group). |
2. Sample Sizes and Data Provenance
-
Reproducibility (Sample-to-Sample, Lot-to-Lot, Day-to-Day, Operator-to-Operator):
- Test set sample size: 10 donors, with 3 samples each (total 30 samples collected). These were processed by 3 operators, with genotyping data evaluated for each operator for each of the 3 SNPs. This implies 20 tests per SNP per operator (based on the "20" under "Samples Tested" for each SNP for each operator), totaling 60 tests per SNP across all operators, or 120 calls per SNP when considering all samples and operators for the "Summary of Results by Sample and Genotype" table.
- Data Provenance: Not explicitly stated, but likely prospective and from a controlled lab environment.
-
Multi-center Reproducibility:
- Test set sample size: 30 donors. Multiple saliva samples collected from each.
- Data Provenance: Saliva samples were collected from 3 sites. Two sites were described as "professional setting" with "supervised collections," and the third site had "unsupervised collections." One sample from each donor was transported to three independent sites for extraction. All eSensor Warfarin Sensitivity Saliva testing was conducted at Site 1. This suggests a prospective data collection design, with samples from multiple (unspecified) locations. No country of origin is mentioned.
-
Method Comparison:
- Test set sample size: 156 saliva samples.
- Data Provenance: Not explicitly stated, but implies a prospective or retrospectively collected set of human saliva samples. No country of origin is mentioned.
-
Interfering Substances (Endogenous):
- Test set sample size: 14 donors, each providing 4 saliva samples.
- Data Provenance: Not explicitly stated, but likely prospective, controlled lab study.
-
Interfering Substances (Exogenous):
- Test set sample size: Varied per activity, ranging from 5 to 9 donors per activity group. Each donor provided samples at two time-points (immediate and 30 minutes post-activity).
- Data Provenance: Not explicitly stated, but likely prospective, controlled lab study.
3. Number of Experts and Qualifications for Ground Truth
The ground truth for all studies was established by bi-directional DNA sequencing. This is a laboratory-based method, not typically requiring "experts" in the sense of clinical specialists. The interpretation of sequencing results is a standard molecular biology technique. No specific number of experts or their qualifications (e.g., geneticists, molecular biologists) are mentioned for establishing the ground truth, as it's assumed to be a direct, objective measurement.
4. Adjudication Method for the Test Set
The studies used bi-directional DNA sequencing as the gold standard, not a human expert adjudication process. The agreement was calculated directly between the device's genotypes and the sequencing results. If there were discrepancies in the initial device results (e.g., initial miscalls or no-calls in the method comparison), retests were performed. For instance, in the method comparison, two miscalls were attributed to "operator error (sample mix-up) occurring during the first XT8 testing" and one remaining no-call to "an operator error at the purification step." This suggests an internal investigation and re-testing process rather than an external adjudication panel.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
No Multi-Reader Multi-Case (MRMC) comparative effectiveness study was done. This device is a diagnostic test for genotyping, not an imaging device or a system requiring human interpretation with or without AI assistance. Therefore, the concept of "readers improving with AI vs. without AI" is not applicable here. The device itself performs the genotyping.
6. Standalone Performance Study
Yes, a standalone performance study was done. All the performance data described (reproducibility across operators/sites, method comparison with sequencing, and interfering substances studies) directly assesses the performance of the eSensor® Warfarin Sensitivity Saliva Test (algorithm/device only, without human-in-the-loop performance influencing the genotyping call itself, though human operators are involved in sample preparation and running the test). The percentage agreement and call rates reflect the direct output of the device compared to the ground truth.
7. Type of Ground Truth Used
The primary ground truth used for all studies was bi-directional DNA sequencing. This is a highly accurate molecular method for determining genetic sequences and polymorphisms (genotypes).
8. Sample Size for the Training Set
The document is a 510(k) summary for a labeling modification (addition of a new specimen collection kit) to an already cleared device. It primarily presents validation data for the continued performance of the device with the new collection kit. It does not mention a "training set" for the eSensor® Warfarin Sensitivity Saliva Test itself, as this device likely relies on established molecular biology principles and probes designed against known genetic targets, rather than machine learning algorithms that require extensive training data. It's a deterministic diagnostic test, not an AI model that learns from data in the way a computer vision algorithm would.
9. How Ground Truth for the Training Set Was Established
As noted above, the concept of a "training set" doesn't directly apply in the context of this device. The device's design is based on known genetic sequences and electrochemical detection rather than a machine learning model. The validation studies verify its performance against the established ground truth of bi-directional DNA sequencing.
§ 862.3360 Drug metabolizing enzyme genotyping system.
(a)
Identification. A drug metabolizing enzyme genotyping system is a device intended for use in testing deoxyribonucleic acid (DNA) extracted from clinical samples to identify the presence or absence of human genotypic markers encoding a drug metabolizing enzyme. This device is used as an aid in determining treatment choice and individualizing treatment dose for therapeutics that are metabolized primarily by the specific enzyme about which the system provides genotypic information.(b)
Classification. Class II (special controls). The special control is FDA's guidance document entitled “Class II Special Controls Guidance Document: Drug Metabolizing Enzyme Genotyping Test System.” See § 862.1(d) for the availability of this guidance document.