FDA 510(k) Search - By Innolitics (SaMD and AI Experts)

HemosIL D-Dimer HS is an automated latex enhanced immunoassay for the quantitative determination of D-Dimer in human citrated plasma on the ACL TOP as an aid in the diagnosis of venous thromboembolism (VTE) [deep venous thrombosis (DVT) and pulmonary embolism (PE)]. The D-Dimer HS Latex Reagent is a suspension of polystyrene latex particles of uniform size coated with the F(ab')2 fragment of a monoclonal antibody highly specific for the D-Dimer domain included in fibrin soluble derivatives. The Use of the F(ab)2 fragment allows a more specific D-Dimer detection avoiding the interference of some endogenous factors like the Rheumatoid Factor. When a plasma containing D-Dimer is mixed with the Latcx Reagent and the Reaction Buffer included in the D-Dimer HS kit, the coated latex particles agglutinate. The degree of agglutination is directly proportional to the concentration of D-Dimer in the sample and is determined by measuring the decrease of the transmitted light caused by the aggregates (turbidimetric immunoassay).

AI/ML Overview

Here's an analysis of the acceptance criteria and study detailed in the provided K050544 submission for the HemosIL D-Dimer HS device:

1. Table of Acceptance Criteria and Reported Device Performance

The submission does not explicitly state pre-defined acceptance criteria for the "Management Study" in terms of specific sensitivity, specificity, and NPV values. Rather, it presents the results of the study. However, the intent is clearly to demonstrate the device's utility in a clinical context, and thus, the reported performance itself serves as the de facto "met" criteria for this submission. For precision and method comparison, the results are presented, implying they met internal or regulatory expectations for substantial equivalence.

Performance Metric	Acceptance Criteria (Implied / Contextual)	Reported Device Performance (HemosIL D-Dimer HS on ACL TOP)
Precision	Low CV% values (indicating good precision)
- D-Dimer Plasma Pool		Within run CV%: 8.3%, Total CV%: 11.0%
- HemosIL D-Dimer Low Control		Within run CV%: 3.7%, Total CV%: 7.0%
- HemosIL D-Dimer High Control		Within run CV%: 2.0%, Total CV%: 7.0%
Method Comparison	Strong correlation with predicate (R value close to 1)	Slope: 1.0407, Intercept: -50.298, R: 0.973
Management Study for VTE Diagnosis (Cut-off 230 ng/mL)	High Sensitivity, high Negative Predictive Value (NPV) are critical for VTE rule-out devices. Specificity is typically lower.
- Sensitivity	(Implied: High, ideally near 100%)	100% (95% CI: 95.4% to 100%)
- Specificity	(Implied: Acceptable for a rule-out test)	47% (95% CI: 40.1% to 53.6%)
- Negative Predictive Value (NPV)	(Implied: High, ideally near 100%)	100% (95% CI: 96.5% to 100%)

2. Sample Size and Data Provenance for the Test Set

Precision Test Set: Not explicitly stated as a separate "test set" in the context of the management study. For precision, three levels of control plasma were used over multiple runs.
Method Comparison Test Set:
- Sample Size: n=229
- Data Provenance: Not explicitly stated, but given it's a 510(k) submission from a US-based company, it's highly probable the samples were from the US. The samples were "citrated plasma samples ranging in D-Dimer concentration from 87 to 20869 ng/mL." Retrospective or prospective is not explicitly stated, but the nature of comparing archived samples suggests it could be retrospective.
Management Study Test Set:
- Sample Size: 300 frozen samples.
- Data Provenance: The samples were from "patients admitted consecutively to an emergency unit with suspected PE or DVT." The country of origin is not specified, but the submission is for the US market. The samples were "frozen," suggesting a retrospective study design where samples were collected and then later analyzed.

3. Number of Experts and Qualifications for Ground Truth (Test Set)

Precision and Method Comparison: No external experts were used to establish ground truth for these studies, as they are analytical performance evaluations against instrument controls or a predicate device.
Management Study: The ground truth for VTE positive and negative status was established by "standard objective tests." The specific number or qualifications of experts (e.g., radiologists interpreting imaging, clinicians making a final diagnosis based on multiple tests) used to interpret these objective tests are not specified in the document.

4. Adjudication Method (Test Set)

Precision and Method Comparison: Not applicable, as these are analytical performance studies.
Management Study: An adjudication method for determining the "standard objective tests" outcome (VTE positive/negative) is not described in the document. It simply states "standard objective tests" were used.

5. Multi Reader Multi Case (MRMC) Comparative Effectiveness Study

No, a multi-reader multi-case (MRMC) comparative effectiveness study of human readers with vs. without AI assistance was not done. This device is an automated in vitro diagnostic immunoassay, not an imaging analysis or decision support AI tool designed to assist human readers directly.

6. Standalone (Algorithm Only) Performance Study

Yes, the "Management Study" represents a standalone performance evaluation. The HemosIL D-Dimer HS is an automated immunoassay device. The reported sensitivity, specificity, and NPV were generated by the algorithm (the automated immunoassay system) interpreting the D-Dimer levels in the plasma samples, independent of human interpretation of the assay results, other than reading the final numerical output. The performance metrics reflect the device's ability to classify patients based on a single D-Dimer concentration cut-off.

7. Type of Ground Truth Used (Test Set)

Precision and Method Comparison: Not applicable in the same way as a diagnostic study. Ground truth for precision is based on statistical measures of repeatability using known controls. Ground truth for method comparison is the measurement obtained from the predicate device.
Management Study: The ground truth for VTE (DVT/PE) diagnosis was established using "standard objective tests." These typically include imaging modalities like ultrasound (for DVT) and CT pulmonary angiography (CTPA) or ventilation/perfusion (V/Q) scans (for PE), potentially combined with clinical assessment. This is a form of clinical outcome/reference standard ground truth.

8. Sample Size for the Training Set

The document does not provide information about a separate "training set" for the HemosIL D-Dimer HS device. Automated immunoassays like this are typically developed and validated using a structured process that involves reagent optimization, calibration, and then performance verification studies as described. They are not "trained" in the machine learning sense with large datasets to learn features. The "training" in this context refers to the initial development and optimization of the assay's chemical and optical properties, and calibration procedures.

9. How the Ground Truth for the Training Set Was Established

As no "training set" in the machine learning sense is described, this question is not applicable. The "ground truth" for developing such an assay would be precisely measured D-Dimer concentrations in various samples to ensure the assay accurately quantifies the analyte.

Summary

Regulation Number and Section

§ 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).