(61 days)
MDA® D-Dimer is a homogeneous latex particle based immunoassay for the quantitative determination of crosslinked fibrin degradation products containing the D-dimer domain in citrated human plasma. MDA D-Dimer is an assay for use in conjunction with a clinical Pre-test Probability Assessment (PTP) model in excluding deep vein thrombosis (DVT) in outpatients suspected of a first episode of DVT. MDA D-Dimer can also be used as and in the assessment and evaluation of patients suspected of pulmonary embolism (PE). The assay is designed for use on the MDA automated coagulation analyzers.
bioMerieux Inc. MDA® D-Dimer is a homogeneous latex particle based immunoassay for the quatitative determination of cross-linked fibrin degradation products containing the D-dimer domain in citrated human plasma. These latex particles aggregate in the presence of fibrin derivatives containing the D-dimer domain. The rate of latex microparticle aggregation is proportional to the concentration of D-dimer in the sample. D-dimer concentration may be interpolated from a reference curve.
This summary describes the acceptance criteria and the study proving the MDA® D-Dimer device meets these criteria.
Acceptance Criteria and Device Performance
| Parameter | Acceptance Criteria | Reported Device Performance |
|---|---|---|
| Specificity | MDA D-Dimer Latex Reagent aggregates in the presence of cross-linked fibrin degradation products D-dimer and D-dimer E. | MDA D-Dimer Latex Reagent aggregates in the presence of cross-linked fibrin degradation products D-dimer and D-dimer E. |
| Accuracy (Correlation) | High correlation (r) to a commercially available assay. (Implicitly, the predicate device K000942's performance) | r = 0.91 (compared to Fibrinostika® FbDP EIA) |
| Accuracy (Slope) | Slope close to 1.0 (Implicitly, the predicate device K000942's performance) | Slope = 1.005 (compared to Fibrinostika® FbDP EIA) |
| Accuracy (Intercept) | Intercept close to 0.0 (Implicitly, the predicate device K000942's performance) | Intercept = 0.293 (compared to Fibrinostika® FbDP EIA) |
| Precision (Positive Control CV) | Low CV (total) for positive control. (Implicitly, the predicate device K000942's performance) | Positive Control: CV (total) = 6.67% |
| Precision (Normal Control CV) | Low CV (total) for normal control. (Implicitly, the predicate device K000942's performance) | MDA Verify 1 (Normal Control): CV (total) = 12.65% |
| Clinical Sensitivity | High clinical sensitivity for excluding DVT. (No explicit threshold, but clinical utility required) | 98.2% (95% CI: 90-100%) for suspected DVT |
| Clinical Specificity | Reasonable clinical specificity for excluding DVT. (No explicit threshold, but clinical utility required) | 60.4% (95% CI: 56-65%) for suspected DVT |
| Negative Predictive Value (NPV) | Very high NPV for excluding DVT. (No explicit threshold, but clinical utility required) | 99.7% (95% CI: 98-100%) for suspected DVT |
Study Details for Clinical Performance (DVT Exclusion)
-
Sample Size used for the test set and the data provenance:
- Sample Size: 556 consecutive eligible outpatients.
- Data Provenance: Multi-center, prospective cohort study conducted at three hospitals. The country of origin is not explicitly stated but is implied to be the USA given the submission to the FDA by a US-based company.
-
Number of experts used to establish the ground truth for the test set and the qualifications of those experts:
- The document does not explicitly state the number or qualifications of experts used to establish the ground truth. It mentions that an MDA D-Dimer test was performed on presentation without knowledge of the Pre-test Probability Assessment (PTP) results. The "ground truth" for DVT diagnosis in patients with a positive D-Dimer and/or high PTP was established through serial compression ultrasound (CUS). For patients with a negative D-Dimer and low/moderate PTP, they underwent no further diagnostic testing but were "followed up for 3 months for development of DVT," which implies clinical outcomes as part of the ground truth.
-
Adjudication method for the test set:
- The document does not explicitly describe a formal adjudication method (e.g., 2+1, 3+1 consensus) for the diagnosis of DVT. The ground truth appears to be based on a combination of diagnostic imaging (serial CUS for some) and clinical follow-up for outcomes.
-
If a multi-reader multi-case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance:
- No, a multi-reader multi-case (MRMC) comparative effectiveness study was not done. This device is an in-vitro diagnostic (IVD) assay, not an AI-assisted diagnostic imaging tool that would typically involve human readers.
-
If a standalone (i.e. algorithm only without human-in-the loop performance) was done:
- Yes, the clinical performance described ("The sensitivity, specificity and negative predictive value of the MDA D-Dimer assay for a clinical cut off of 0.50 µg FEU/n1") represents the standalone performance of the assay when used in conjunction with a clinical PTP model. The assay itself (
MDA D-Dimer) generates a quantitative output. The clinical "human-in-the-loop" component is the interpretation of that output in the context of the PTP model, but the assay's performance metrics are standalone for its measurement capabilities.
- Yes, the clinical performance described ("The sensitivity, specificity and negative predictive value of the MDA D-Dimer assay for a clinical cut off of 0.50 µg FEU/n1") represents the standalone performance of the assay when used in conjunction with a clinical PTP model. The assay itself (
-
The type of ground truth used:
- The ground truth for DVT diagnosis was a combination of:
- Diagnostic Imaging: Serial compression ultrasound (CUS) for patients with a positive MDA D-Dimer and/or high Pre-test Probability (PTP).
- Clinical Outcomes Data: 3-month follow-up for development of DVT in patients with a negative MDA D-Dimer test result and a low or moderate PTP.
- The ground truth for DVT diagnosis was a combination of:
-
The sample size for the training set:
- The document does not specify a training set for algorithm development for the MDA D-Dimer assay. As an immunoassay, its "training" pertains more to establishing its analytical performance parameters (linearity, precision, etc.) and defining a clinical cutoff value. It states that a "clinical cut off value of 0.50 ug FEU/ml previously validated in an accuracy study" was used, suggesting prior studies informed this threshold, but no sample size for that "accuracy study" is provided here.
-
How the ground truth for the training set was established:
- As noted above, a formal "training set" in the AI/machine learning sense is not applicable here for this immunoassay. However, the clinical cutoff value of 0.50 µg FEU/ml was "previously validated in an accuracy study." The mechanism for establishing ground truth within that prior accuracy study is not detailed in this document.
§ 864.7320 Fibrinogen/fibrin degradation products assay.
(a)
Identification. A fibrinogen/fibrin degradation products assay is a device used to detect and measure fibrinogen degradation products and fibrin degradation products (protein fragments produced by the enzymatic action of plasmin on fibrinogen and fibrin) as an aid in detecting the presence and degree of intravascular coagulation and fibrinolysis (the dissolution of the fibrin in a blood clot) and in monitoring therapy for disseminated intravascular coagulation (nonlocalized clotting in the blood vessels).(b)
Classification. Class II (performance standards).