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

The Platelia™ Aspergillus EIA is an immunoenzymatic sandwich microplate assay for the detection of Aspergillus galactomannan antigen in adult and pediatric serum samples.

The Platelia™ Aspergillus EIA is a test which, when used in conjunction with other diagnostic procedures such as microbiological culture, histological examination of biopsy samples and radiographic evidence, can be used as an aid in the diagnosis of Invasive Aspergillosis.

Device Description

The Platelia™ Aspergillus EIA is a one-stage immunoenzymatic sandwich microplate assay which detects galactomannan in human serum. The assay uses rat EBA-2 monoclonal antibodies, which are directed against Aspergillus galactomannan, and have been characterized in previous studies 19,16. The monoclonal antibodies are used to coat the wells of the microplate and bind the antigen, and to detect the antigen bound to the sensitized microplate (conjugate reagent: peroxidase-linked monoclonal antibodies).

Serum samples are heat-treated in the presence of EDTA in order to dissociate immune complexes and to precipitate serum proteins that could possibly interfere with the test . The treated serum samples and conjugate are added to the wells coated with monoclonal antibodies, and incubated. A monoclonal antibody - galactomannan - monoclonal antibody / peroxidase complex is formed in the presence of galactomannan antigen. The strips are washed to remove any unbound material. Next, the substrate solution is added, which will react with the complexes bound to the well to form a blue color reaction. The enzyme reaction is stopped by the addition of acid, which changes the blue color to yellow. The absorbance (optical density) of specimens and controls is determined with a spectrophotometer set at 450 and 620/630 nm wavelength.

AI/ML Overview

Here's a summary of the acceptance criteria and study details for the Bio-Rad Platelia™ Aspergillus EIA, extracted from the provided 510(k) summary:

Acceptance Criteria and Device Performance:

Performance Metric	Acceptance Criteria (Implied)	Reported Device Performance (Pediatric)	Reported Device Performance (Adult)
Sensitivity (Patients)	Sufficiently high to aid in diagnosis of Invasive Aspergillosis.	Combined Proven & Probable IA:	Combined Proven & Probable IA:
		52.9% (9/17) [95% CI: 31.0-73.8%]	79.3% (23/29) [95% CI: 61.6-90.2%]
		Proven IA: 44.4% (4/9)	Proven IA: 81.8% (9/11)
		Probable IA: 62.5% (5/8)	Probable IA: 77.8% (14/18)
Specificity (Patients)	Sufficiently high to minimize false positives, when used in conjunction with other diagnostic procedures.	87.0% (94/108) [95% CI: 79.4-92.1%]	88.8% (127/143) [95% CI: 82.6-93.0%]
Specificity (Samples)	Sufficiently high to minimize false positives at the sample level.	98.5% (1600/1625) [95% CI: 97.7-99.0%]	98.5% (1243/1262) [95% CI: 97.7-99.0%]
Positive Predictive Value (PPV) - Study Prevalence	(No explicit numerical criteria stated, expected to be clinically useful)	39.1% [95% CI: 22.2-59.2%]	59.0% [95% CI: 43.4-72.9%]
Negative Predictive Value (NPV) - Study Prevalence	(No explicit numerical criteria stated, expected to be clinically useful)	92.2% [95% CI: 85.3-96.0%]	95.5% [95% CI: 90.5-97.9%]
Positive Predictive Value (PPV) - 5% Prevalence	(No explicit numerical criteria stated, expected to be clinically useful)	17.6% [95% CI: 6.5-39.8%]	27.2% [95% CI: 13.7-46.7%]
Negative Predictive Value (NPV) - 5% Prevalence	(No explicit numerical criteria stated, expected to be clinically useful)	97.2% [95% CI: 92.1-99.1%]	98.8% [95% CI: 95.4-99.7%]
Reproducibility (Inter-assay & Intra-assay %CV)	(No explicit numerical criteria stated, expected to be acceptable for lab diagnostics)	Generally below 20-30% for most panels	(Assumed similar, studies were prior)
Cross-Reactivity	No (or minimal) positive results with common interfering conditions.	0 positives across 15 interfering conditions (10 samples/condition)	(Assumed similar, studies were prior)

Note: The document implies acceptance criteria by presenting performance data within clinical contexts and stating that the device is "substantially equivalent" to a predicate device, which inherently means it meets similar performance standards.

Study Details:

Sample Sizes and Data Provenance (Test Set):
- Pediatric (new study for this submission):
  - Total Patients: 129 immunocompromised pediatric patients.
  - Total Samples: 1954 serum samples.
  - Invasive Aspergillosis (IA) Patients: 17 (9 Proven, 8 Probable).
  - Control/Non-IA Patients: 108 (from whom 1625 samples were tested).
  - Data Provenance: United States (three testing centers).
  - Retrospective/Prospective: Not explicitly stated, but the collection of patients diagnosed with IA and controls suggests it could be a mix or primarily retrospective given the diagnostic criteria often established over time.
- Adult (previously conducted study for K023857):
  - Total Patients: 172 bone marrow transplant (BMT) and leukemic patients.
  - Total Samples: 1724 serum samples.
  - Invasive Aspergillosis (IA) Patients: 29 (11 Proven, 18 Probable).
  - Control/Non-IA Patients: 143 (from whom 1262 samples were tested).
  - Data Provenance: North America (three testing centers).
  - Retrospective/Prospective: Not explicitly stated, likely retrospective as it refers to a "previously conducted" study with diagnosed patients.
Number of Experts and Qualifications (Ground Truth for Test Set):
- The ground truth for Invasive Aspergillosis (IA) was established based on EORTC/NIAID definitions. These definitions are rigorous and rely on a combination of:
  - Proven IA: Positive microbiological culture from a sterile site AND histopathological demonstration of fungal forms.
  - Probable IA: At least one microbiological criterion AND one major or two minor clinical criteria.
- The document does not specify the number or qualifications of individual experts who applied these definitions to each patient/sample. It relies on the widely accepted, expert-consensus-derived EORTC/NIAID criteria.
Adjudication Method (Test Set):
- The document does not explicitly state an adjudication method (e.g., 2+1, 3+1) for establishing the ground truth of IA. The reliance on EORTC/NIAID definitions suggests a standardized, objective application of these criteria, which may involve review by treating clinicians or infectious disease specialists as part of the diagnostic process.
Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:
- No, an MRMC comparative effectiveness study was not done. This device is an in-vitro diagnostic (IVD) assay designed to detect a biomarker (galactomannan antigen). Its performance is evaluated biochemically (sensitivity, specificity, reproducibility) rather than by human readers interpreting outputs, so a study comparing human reader performance with and without AI assistance is not applicable.
Standalone Performance Study (Algorithm Only):
- Yes, this is a standalone performance study of the diagnostic assay. The "Performance Evaluation Studies" sections detail the assay's sensitivity and specificity based on its direct measurements of galactomannan antigen in serum, independent of human interpretation or a human-in-the-loop scenario.
- The reproducibility and cross-reactivity studies also evaluate the algorithm's intrinsic performance.
Type of Ground Truth Used:
- The ground truth used was "Proven or Probable Invasive Aspergillosis" as defined by EORTC/NIAID definitions. This is a robust clinical standard for diagnosing IA, incorporating:
  - Pathology: Histological examination of biopsy samples.
  - Microbiological culture: Positive culture from sterile sites.
  - Clinical evidence/outcomes data: Major/minor clinical criteria defined by the EORTC/NIAID.
Sample Size for the Training Set:
- The document does not explicitly describe a separate "training set" in the context of an algorithm that learns from data.
- This is an enzymatic immunoassay (EIA) kit, a traditional IVD device. The development and calibration of such a kit typically involve internal studies and optimization (analogous to "training" in AI), but these details are not provided as a distinct "training set" and "validation set" in the way they would be for a software algorithm. The performance evaluation discussed in the 510(k) is essentially the validation of the finalized assay itself.
How Ground Truth for the Training Set Was Established:
- As there isn't a traditional "training set" for an AI algorithm in this context, the question of its ground truth establishment is not directly applicable.
- For the initial development and optimization of the Platelia™ Aspergillus EIA (which would be analogous to "training"), the ground truth for positive and negative controls would have been established through well-characterized samples (e.g., purified galactomannan antigens, sera from confirmed IA patients, sera from healthy controls) using established reference methods and standards. However, these specific details for the "training" phase are not presented in this 510(k) summary.

Summary

Regulation Number and Section

§ 866.3040

Aspergillus spp. serological reagents.(a)
Identification. Aspergillus spp. serological reagents are devices that consist of antigens and antisera used in various serological tests to identify antibodies toAspergillus spp. in serum. The identification aids in the diagnosis of aspergillosis caused by fungi belonging to the genusAspergillus. Aspergillosis is a disease marked by inflammatory granulomatous (tumor-like) lessions in the skin, ear, eyeball cavity, nasal sinuses, lungs, and occasionally the bones.(b)
Classification. Class I (general controls). The device is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to § 866.9.