(88 days)
The Hp Detect™ Stool Antigen ELISA is an in vitro diagnostic qualitative enzyme immunoassay for the detection of Helicobacter pylori (H. pylori) antigens in human stool or feces. The Hp Detect™ Stool Antigen ELISA is intended to aid in the initial diagnosis and post-therapy diagnosis of H. pylori infection. Additionally, the test may be used to assess H. pylori infection status after treatment. Retesting at a minimum of 4 weeks after the completion of treatment may be done to assess H. pylori status. Test results should always be taken into consideration by the physician in conjunction with patient's clinical information (history and symptoms). For Prescription Use Only.
The Hp Detect Stool Antigen ELISA is an enzyme immunoassay which detects the H. pylori antigen in human fecal samples. The Hp Detect Stool Antigen ELISA comes in a kit that contains materials to assay a total of 92 samples. The device consists of a 96-well clear flat bottom polystyrene high bind microplate coated with affinity purified rabbit anti-human H. pvlori polyclonal antibody. The device is provided with detection antibody which is a purified mouse monoclonal antibody specific for H. pylori antigen and has been conjugated to horseradish peroxidase (HRP). The device kit is also provided with sample diluent buffer, wash buffer, substrate solution, stop solution along with negative and positive controls. Negative control is a phosphate buffered protein solution and positive control is composed of purified H. pylori antigen (ATCC strain 43504) from cell lysate. Polyclonal anti-H. pylori captures antibodies that are immobilized on microwells. Patient samples prepared in sample diluent are added to the microwells and incubated for one hour at 37 ± 2℃. If the H. pvlori antigen is present in the sample, it will bind to the immobilized antibody on the plate. Following this incubation, the plate is washed thoroughly. A peroxidase conjugated anti-H. pylori monoclonal antibody is then added to the microwells and incubated for 30 minutes at 37 ± 2℃. If H. pylori antigen is bound to the microwells in the first step, the detection antibody would now bind in this step to form a sandwich complex. Following this incubation, a thorough wash step is performed to remove non-specific and non-binding materials. Substrate is then added and incubated for 10 minutes at 37±2℃ to generate a color in the presence of the enzyme complex. Stop solution is then added to end the reaction. The results are read spectrophotometrically at the following wavelengths: 1. Single Wavelength Measurement at 450 nm 2. Dual Wavelength Measurement 450/620 nm or 450/630 nm
The provided document is an FDA 510(k) Pre-Market Notification for the Biomerica, Inc. Hp Detect Stool Antigen ELISA for the detection of Helicobacter pylori (H. pylori) antigens in human stool or feces. It is a qualitative immunoassay intended to aid in the initial diagnosis and post-therapy diagnosis of H. pylori infection.
Based on the provided text, the device itself has acceptance criteria for its analytical performance (e.g., reproducibility, LoD, specificity, inclusivity) and clinical performance (Positive Percent Agreement - PPA, Negative Percent Agreement - NPA). The study detailed in the document serves to prove that the device meets these internal acceptance criteria set by the manufacturer for FDA clearance.
Here's a breakdown of the requested information based on the document:
1. A table of acceptance criteria and the reported device performance
The document does not explicitly state pre-defined acceptance criteria values in a formal table for PPA and NPA. However, it presents the achieved performance and concludes that the results are "acceptable." We can infer the implicit acceptance criteria from these reported values. For analytical performance, criteria are implied by the "acceptable" statement for reproducibility, precision, LoD, cross-reactivity, interference, and inclusivity.
| Performance Metric | Acceptance Criteria (Inferred from "Acceptable") | Reported Device Performance (Dual Wavelength) | Reported Device Performance (Single Wavelength) |
|---|---|---|---|
| Analytical Performance | |||
| Reproducibility (Detection Rate) | High Negative: Low % detection | High Negative (0.42xLoD): 3.9% (7/180) | Equivalent to Dual Wavelength |
| Low Positive (1.60xLoD): 95%+ detection | Low Positive (1.60xLoD): 100% (180/180) | Equivalent to Dual Wavelength | |
| Low Positive (2.43xLoD): 95%+ detection | Low Positive (2.43xLoD): 100% (180/180) | Equivalent to Dual Wavelength | |
| Moderate Positive: 95%+ detection | Moderate Positive (3.93xLoD): 100% (180/180) | Equivalent to Dual Wavelength | |
| Within-Lab Precision (Detection Rate) | High Negative: Low % detection | High Negative (0.42xLoD): 5% (13/288) | Equivalent to Dual Wavelength |
| Low Positive (1.60xLoD): 95%+ detection | Low Positive (1.60xLoD): 97% (278/288) | Equivalent to Dual Wavelength | |
| Low Positive (2.43xLoD): 95%+ detection | Low Positive (2.43xLoD): 100% (288/288) | Equivalent to Dual Wavelength | |
| Moderate Positive: 95%+ detection | Moderate Positive (3.93xLoD): 100% (288/288) | Equivalent to Dual Wavelength | |
| Limit of Detection (LoD) | Quantified LoD | Strain ATCC 43504: 2.53 ng/mL (0.38 ng/test) or 1.69 x 10^3 CFU/mL | |
| Strain ATCC 49503: 5.86 ng/mL (0.88 ng/test) | Equivalent to Dual Wavelength | ||
| Cross-Reactivity & Microbial Interference | No interference expected | No cross-reactivity or microbial interference observed with listed microorganisms | Equivalent to Dual Wavelength |
| Interfering Substances | No interference expected | No interference observed with listed substances | Equivalent to Dual Wavelength |
| Inclusivity (Detection Rate) | 100% detection of tested strains | 100% detection for all 6 H. pylori strains (whole cells) and 1 purified H. pylori antigen tested | Equivalent to Dual Wavelength |
| Prozone / Hook Effect | No hook effect up to high antigen concentration | No high-dose hook effect observed up to 20,000 ng/mL | Equivalent to Dual Wavelength |
| Clinical Performance | |||
| Frozen Specimen PPA | High PPA | 99.11% (111/112) | 99.11% (111/112) |
| Frozen Specimen NPA | High NPA | 98.13% (315/321) | 95.95% (308/321) |
| Fresh Stool Specimen PPA | High PPA | 100.00% (20/20) | 100.00% (20/20) |
| Fresh Stool Specimen NPA | High NPA | 98.36% (120/122) | 98.36% (120/122) |
| Post-Therapy Sensitivity | High Sensitivity | 100% (10/10) | Not explicitly stated whether single wavelength was assessed for post-therapy; likely same as dual for qualitative results. |
| Post-Therapy Specificity | High Specificity | 100% (4/4) | Not explicitly stated whether single wavelength was assessed for post-therapy; likely same as dual for qualitative results. |
2. Sample sizes used for the test set and the data provenance (e.g., country of origin of the data, retrospective or prospective)
- Clinical Study - Frozen Specimen:
- Sample Size: 433 frozen and de-identified fecal samples.
- Provenance:
- 355 specimens from Italy.
- 78 specimens from three geographically different regions of the USA (west, southwest, and southeast).
- Nature: Retrospective (frozen, de-identified samples). Patients were presenting with dyspepsia, undergoing endoscopy/biopsy, not on certain medications, and no H. pylori treatment within 6 months.
- Clinical Study - Fresh Stool Specimen:
- Sample Size: 142 fresh, de-identified fecal specimens.
- Provenance: Collected through multiple biospecimen vendors and clinical laboratories. Locations not specified beyond "collection centers" and "Biomerica (internal site)."
- Nature: Likely prospective (freshly collected and then immediately tested/shipped). Patients had symptoms of H. pylori infection.
- Post-Therapy Diagnosis:
- Sample Size: 14 paired (pre- and post-therapy) frozen retrospective specimens.
- Provenance: Italy.
- Nature: Retrospective (frozen, paired samples). All subjects initially positive by CRM and completed eradication therapy. Post-therapy samples collected minimum 4 weeks after treatment.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g., radiologist with 10 years of experience)
The document states that the ground truth for the clinical studies (frozen and fresh specimens) was established by comparison with an "FDA cleared device" (predicate device). For the frozen specimen study, discrepant results were further analyzed by "chart review and determined to have a RUT or history result."
For the post-therapy study, the ground truth was a "composite reference method (CRM) consisting of histology and Rapid Urease Test."
The document does not specify the number of experts or their qualifications (e.g., pathologists, gastroenterologists) involved in establishing the ground truth via histology, RUT, or chart review. It implies laboratory testing capabilities for the predicate device.
4. Adjudication method (e.g., 2+1, 3+1, none) for the test set
The document describes adjudication for discrepant results in the frozen specimen clinical study:
- Discrepant results were "further analyzed by chart review and determined to have a RUT or history result."
This suggests a form of adjudication after the initial comparison, using additional clinical information or laboratory results (RUT or histology) as a reference. This is not a multi-reader adjudication method but rather a method for determining the "true" status of discrepant samples.
There is no mention of multi-reader adjudication for the Hp Detect Stool Antigen ELISA results themselves. The device's results are read spectrophotometrically based on defined cut-offs.
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, a multi-reader multi-case (MRMC) comparative effectiveness study was not conducted. This device is an in-vitro diagnostic (ELISA), not an AI-powered diagnostic imaging tool that would typically involve human readers and MRMC studies. The device provides a quantitative or qualitative output (positive/negative) that is read spectrophotometrically, not interpreted by human readers in the same way as medical images.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
This refers to the performance of the device itself (the "algorithm" in this context being the ELISA assay) without human interpretation variables beyond the initial lab procedure. Yes, the performance characteristics (analytical and clinical studies) directly report the standalone performance of the Hp Detect Stool Antigen ELISA device against comparator methods or reference standards. The results (PPA, NPA, sensitivity, specificity) represent the device's accuracy in detecting H. pylori antigens.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc)
The ground truth for the test sets was primarily established through:
- Comparison to an FDA cleared predicate device for the frozen and fresh specimen clinical studies.
- Composite Reference Method (CRM) consisting of histology and Rapid Urease Test (RUT) for the initial H. pylori status and post-therapy evaluation.
- Chart review combined with RUT or histology results for resolving discrepant cases in the frozen specimen study.
This is a form of "laboratory reference standard" and "clinical/pathology confirmation."
8. The sample size for the training set
The document describes the analytical and clinical performance studies, which are validation studies or test sets. It does not provide information on the sample size used for the training set of the Hp Detect Stool Antigen ELISA device. As an ELISA assay, it is a biochemical test, not a machine learning model that typically undergoes data-driven training. The "training" for such a device involves assay development, optimization, and establishment of reagents and procedures, rather than a machine learning training set of data.
9. How the ground truth for the training set was established
As the Hp Detect Stool Antigen ELISA is a biochemical immunoassay (not an AI/ML device), the concept of a "training set" and establishing ground truth for it in the machine learning sense is not applicable. The "ground truth" for developing and optimizing such a device would implicitly rely on well-characterized H. pylori positive and negative samples, purified antigens, and potentially clinical samples with known infection status (established through methods like culture, PCR, histology, RUT) to guide the reagent selection, assay design, and cut-off determination. The document details the methodology for setting assay cut-off based on a panel of 227 specimens (83 positive, 144 negative by predicate device), which could be considered part of the "development" or "optimization" phase for establishing the device's operational parameters, but not a "training set" in the context of deep learning.
§ 866.3110
Campylobacter fetus serological reagents.(a)
Identification. Campylobacter fetus serological reagents are devices that consist of antisera conjugated with a fluorescent dye used to identifyCampylobacter fetus from clinical specimens or cultured isolates derived from clinical specimens. The identification aids in the diagnosis of diseases caused by this bacterium and provides epidemiological information on these diseases.Campylobacter fetus is a frequent cause of abortion in sheep and cattle and is sometimes responsible for endocarditis (inflammation of certain membranes of the heart) and enteritis (inflammation of the intestines) in humans.(b)
Classification. Class I (general controls).