Search Results

HemosIL Chromogenic Factor IX is an automated assay for the photometric, quantitative determination of factor IX activity in 3.2% citrated plasma on the ACL TOP® Family and ACL TOP Family 50 Series in the laboratory setting by a healthcare professional. HemosIL Chromogenic Factor IX is indicated for use on patients when identifying factor IX deficiency or measuring factor IX activity from patients on replacement therapy. For adult population only. For prescription use only.

Factor IX activity in a patient's plasma is determined using a chromogenic method, in which human factor IX is activated by human factor XIa, and, when formed, factor IXa activates human factor X in the presence of human factor VIII, calcium and phospholipid. The amount of factor Xa generated is proportionate to the factor IX activity and is determined from the hydrolysis of a chromogenic factor Xa substrate. Results are determined by comparing a chromogenic signal to a calibration curve.

Here's a breakdown of the acceptance criteria and the studies that demonstrate the device's performance, based on the provided FDA 510(k) summary for the HemosIL Chromogenic Factor IX:

1. Table of Acceptance Criteria and Reported Device Performance

The document doesn't explicitly list "acceptance criteria" in a separate table. However, it presents performance characteristics that implicitly serve as success metrics for the device's substantial equivalence. I've extrapolated these based on the study findings.

2. Sample Sizes Used for the Test Set and Data Provenance

• Precision Studies (Separate for ACL TOP Family and ACL TOP Family 50 Series):
  • Sample Size: For each sample level (three patient sample pools, two control levels), there were 80 replicates (20 days * 2 runs/day * 2 replicates/run). So, 5 levels * 80 replicates = 400 measurements per instrument type for the representative reagent lot. For the aggregated data, this would be 3 reagent lots * 5 levels * 80 replicates = 1200 measurements.
  • Data Provenance: Retrospective (clinical samples are often already collected). The documentation does not specify the country of origin, but it implies a laboratory setting.
• Reproducibility Study:
  • Sample Size: 90 replicates per level (pooled three-site data). There were 10 levels (2 controls, 5 patient samples, 3 concentrate-spiked samples). So, 10 levels * 90 replicates = 900 measurements.
  • Data Provenance: Retrospective (clinical samples) and prospective (concentrate-spiked samples). Conducted at two external sites and one internal site. Country of origin not specified, but likely US-based given FDA submission.
• LoB, LoD, LoQ Studies:
  • Sample Size: Not explicitly stated as a number of distinct samples or runs, but performed using "three reagent lots" on "representative members" of both ACL TOP families.
  • Data Provenance: Not specified, but generally involves laboratory-prepared samples.
• Linearity Study:
  • Sample Size: Not explicitly stated as number of samples or points, but performed using "three reagent lots" on ACL TOP Family and "one reagent lot" on ACL TOP Family 50 Series. Typically, linearity studies involve multiple dilutions of a high concentration sample.
  • Data Provenance: Not specified, likely laboratory-prepared samples.
• Interference Studies:
  • Sample Size: Not explicitly stated as a number of distinct samples or runs, but used a "minimum of one reagent lot." Interference studies usually involve adding varying concentrations of potential interferents to samples with known analyte concentrations.
  • Data Provenance: Not specified, likely laboratory-prepared or spiked samples.
• Normal Reference Interval Study:
  • Sample Size: 120 plasma samples from "apparently healthy individuals."
  • Data Provenance: Retrospective (plasma samples). Country of origin not specified.
• Recovery of Factor IX Replacement Therapies Study:
  • Sample Size: Immunodepleted FIX deficient plasma was spiked with four factor IX replacement therapies at "seven to fourteen concentrations." So, 4 therapies * (7 to 14 concentrations) = between 28 and 56 samples tested.
  • Data Provenance: Laboratory-prepared spiked samples using immunodepleted plasma.
• Multicenter Method Comparison Study:
  • Sample Size: 344 samples in total (N for each site: 105, 105, 104, 30). These samples were from "patients with von Willebrand disease, patients with hemophilia A and B and patients on factor IX replacement therapy."
  • Data Provenance: Retrospective patient samples. Conducted at four sites (three external and one internal). Country of origin not specified, but multi-center suggests potentially diverse data.
• Internal Method Comparison Study (ACL TOP Family 50 Series vs. ACL TOP Family):
  • Sample Size: 126 samples.
  • Data Provenance: Not explicitly stated, but likely patient samples or a mix. Internal study suggests data from the manufacturer's own facilities.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts

The document does not mention the use of experts to establish a "ground truth" for the test sets in the typical sense of subjective interpretation (e.g., radiologists, pathologists). This device is an in vitro diagnostic (IVD) assay designed for quantitative determination of Factor IX activity.

For IVDs, the "ground truth" is generally established by:

• Reference Methods: Such as the predicate device (HemosIL Factor IX deficient plasma) in the method comparison study, or established laboratory methods for determining analyte concentrations.
• Known Concentrations: In studies like linearity, LoB/LoD/LoQ, interference, and recovery, samples are often prepared with known concentrations of the analyte or interferent.
• Clinical Diagnosis: For studies involving patient samples, the "ground truth" for the patient's condition (e.g., Factor IX deficiency, hemophilia A/B) would be based on established clinical diagnostic criteria, which involve a healthcare professional's assessment, not a single 'expert' review of the test results themselves.

Therefore, the concept of "number of experts" and their "qualifications" for ground truth establishment, as often seen in imaging or pathology AI, is not directly applicable here.

4. Adjudication Method for the Test Set

Since this is an IVD device for quantitative measurement, there is no mention of an adjudication method (like 2+1 or 3+1 consensus) for establishing ground truth. The "ground truth" for quantitative assays is typically based on:

• Reference measurements: The results from the predicate device or a recognized reference method.
• Prepared concentrations: Samples with precisely known concentrations.
• Statistical analysis: Agreement between the device and the reference, or the consistency of results over time/across sites.

5. 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, an MRMC comparative effectiveness study was not done. This device is an in vitro diagnostic (IVD) assay, not an imaging or diagnostic AI tool that assists human readers (e.g., radiologists, pathologists). It provides a quantitative measurement of Factor IX activity directly. It does not involve human interpretation that could be "assisted" by AI in the way an imaging algorithm would.

6. If a Standalone (i.e. algorithm only without human-in-the-loop performance) was done

Yes, the performance studies described are essentially "standalone" algorithm performance. The HemosIL Chromogenic Factor IX assay is an automated photometric assay. The precision, reproducibility, linearity, LoB/LoD/LoQ, interference, normal reference interval, and recovery studies assess the device's analytical performance on its own without human interpretation in the results generation. While a healthcare professional operates the instrument and interprets the final numerical result in the context of a patient's condition, the device's output itself is a direct quantitative value. The method comparison studies also evaluate the standalone output of the new device against a predicate device.

7. The Type of Ground Truth Used

The ground truth for the various studies primarily involved:

• Reference Method Comparison: The predicate device, HemosIL Factor IX deficient plasma (K031829), which uses an activated partial thromboplastin time (APTT) assay. This serves as the comparative "ground truth" in the method comparison studies.
• Known Concentrations/Spiked Samples: For studies like LoB/LoD/LoQ, linearity, interference, and recovery of Factor IX replacement therapies, ground truth was established by preparing samples with precisely known concentrations of Factor IX or interferents, or by spiking immunodepleted plasma with known amounts of Factor IX concentrates.
• Clinically Characterized Samples: For studies involving patient samples (e.g., method comparison, normal reference interval), the samples were sourced from "patients with von Willebrand disease, patients with hemophilia A and B and patients on factor IX replacement therapy" or "apparently healthy individuals." The "ground truth" for their clinical status would have been based on established clinical diagnostic criteria and patient outcomes/history, not specifically pathology or ad-hoc expert consensus on the test results themselves.

8. The Sample Size for the Training Set

The document does not explicitly describe a "training set" in the context of machine learning or AI algorithm development. This is a traditional IVD assay, not an AI/ML-based device that would typically undergo a distinct training phase with a dedicated dataset. The development and optimization of such assays rely on chemical/biological principles, reagent formulations, and analytical validation rather than iterative learning from data.

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

As there's no explicit "training set" described for an AI/ML algorithm, this question is not directly applicable. The "ground truth" for the various performance studies (as described in point 7) was established through reference methods, known concentrations, and clinical characterization of patient samples for validation, not for training a model.

Ask a specific question about this device