(119 days)
The Epstein Barr Virus Early Diffuse Antigen (EBV EA-D) IgG ELISA Test is for the qualitative detection of IgG antibodies to EBV EA-D as an aid in the diagnosis of EBV infection in patients with clinical symptoms of Infectious Mononucleosis (IM). The PANBIO EBV EA-D IgG ELISA should be used in conjunction with other EBV serologies.
The EBV EA-D IgG ELISA Kit is an enzyme-linked immunosorbent assay (ELISA) for the detection of IgG antibodies to EBV EA-D antigen in human serum.
The provided text describes the performance of the PANBIO EBV EA-D IgG ELISA Kit. However, it does not explicitly state "acceptance criteria" for the device. Instead, it presents various performance characteristics (sensitivity, specificity, agreement) of the PANBIO ELISA and compares them to an equivalent device (Trinity Biotech Captia™ EBV EA-D IgG ELISA) and to the EBV serological status determined by other serologies. The reproducibility and cross-reactivity studies also demonstrate the device's analytical performance.
Since no explicit numerical acceptance criteria are given, I will present the reported performance as if it were the criteria and the device meets it.
Here's a breakdown of the requested information based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance
As no specific acceptance criteria were explicitly stated, the reported performance characteristics from the primary study sites (Site 2 and Site 3) are presented below. It is implied that these values were deemed acceptable for the device's clearance.
Study Site 2 (Retrospective Sera - Maryland, USA)
| Metric | Acceptance Criteria (Implied) | Reported Device Performance (PANBIO EBV EA-D IgG ELISA) |
|---|---|---|
| Relative Sensitivity (Acute EBV) | Not explicitly stated | 31.4% (95% CI: 19.1 – 45.9 %) |
| Relative Sensitivity (Past EBV) | Not explicitly stated | 24.7% (95% CI: 19.3 – 30.1%) |
| Relative Specificity (Past EBV) | Not explicitly stated | 66.3% (95% CI: 60.3 – 72.2 %) |
| Relative Specificity (Seronegative) | Not explicitly stated | 98.1% (95% CI: 89.7 – 100.0 %) |
| Relative Agreement (Overall) | Not explicitly stated | 65.9% (95% CI: 60.9 – 70.9 %) |
| Relative Sensitivity (vs Trinity Biotech) | Not explicitly stated | 57.6% (95% CI: 48.9 – 66.3%) |
| Relative Specificity (vs Trinity Biotech) | Not explicitly stated | 94.3 % (95% CI: 90.0 – 97.1%) |
| Relative Agreement (vs Trinity Biotech) | Not explicitly stated | 73.1% (95% CI: 68.5 – 77.8%) |
Study Site 3 (Prospective Sera - Queensland, Australia)
| Metric | Acceptance Criteria (Implied) | Reported Device Performance (PANBIO EBV EA-D IgG ELISA) |
|---|---|---|
| Relative Sensitivity (Acute EBV) | Not explicitly stated | 37.1% (95% CI: 21.5 - 55.1%) |
| Relative Sensitivity (Past EBV) | Not explicitly stated | 13.3% (95% CI: 9.1 - 17.5%) |
| Relative Specificity (Past EBV) | Not explicitly stated | 86.3% (95% CI: 82.0 - 90.6%) |
| Relative Specificity (Seronegative) | Not explicitly stated | 95.7% (95% CI: 85.5 - 99.5%) |
| Relative Agreement (Overall) | Not explicitly stated | 82.4% (95% CI: 78.3 - 86.5%) |
| Relative Sensitivity (vs Trinity Biotech) | Not explicitly stated | 58.3% (95% CI: 44.9 - 70.9%) |
| Relative Specificity (vs Trinity Biotech) | Not explicitly stated | 95.2% (95% CI: 91.9 - 97.4%) |
| Relative Agreement (vs Trinity Biotech) | Not explicitly stated | 88.2% (95% CI: 84.7 - 91.7%) |
Reproducibility (Across 3 Australian Sites)
| Sample Group | Within-run CV | Between Day CV | Between Site CV | Total CV |
|---|---|---|---|---|
| Positive | 20.3% | 0.0% | 4.3% | 20.1% |
| Cut-off | 10.9% | 0.0% | 0.0% | 10.1% |
| Negative | 18.9% | 0.0% | 5.8% | 19.4% |
| Sample 1 | 9.0% | 0.0% | 4.1% | 9.4% |
| Sample 2 | 11.7% | 2.4% | 4.7% | 12.5% |
| Sample 3 | 10.8% | 0.0% | 0.0% | 10.4% |
| Sample 4 | 10.9% | 0.0% | 0.0% | 10.2% |
| Sample 5 | 14.0% | 0.0% | 3.3% | 13.9% |
| Sample 6 | 12.2% | 2.7% | 0.0% | 12.2% |
| Sample 7 | 9.5% | 9.1% | 5.5% | 13.0% |
| Sample 8 | 14.0% | 9.9% | 3.7% | 16.5% |
Cross-Reactivity (Study Site 5)
| Disease (IgG Antibodies) | Total Specimens | Positive Result (PANBIO EBV EA-D IgG ELISA) |
|---|---|---|
| Cytomegalovirus | 10 | 0/10 |
| Varicella zoster | 13 | 0/13 |
| Herpes simplex virus 1 | 8 | 0/8 |
| Herpes simplex virus 2 | 2 | 0/2 |
| Anti-Nuclear Antibody | 8 | 0/8 |
| Rheumatoid Factor | 9 | 0/9 |
| Total | 50 | 0/50 |
2. Sample sizes for the test set and data provenance:
- Study Site 2:
- Sample Size: 346 frozen retrospective sera.
- Data Provenance: Maryland, USA. The samples were submitted to a state health laboratory for EBV testing.
- Study Site 3:
- Sample Size: 330 prospective sera.
- Data Provenance: Queensland, Australia. The samples were tested at a private pathology laboratory for EBV testing.
- Reproducibility Study (Sites 3, 4 & 5):
- Sample Size: 8 sera, each tested 3 times on 3 different days (total 27 tests per sample for analysis).
- Data Provenance: Three Australian study sites (two private pathology laboratories and PANBIO).
- Cross-Reactivity Study (Site 5):
- Sample Size: 50 specimens.
- Data Provenance: Not explicitly stated beyond "Study Site 5," but conducted in relation to the Australian reproducibility sites.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:
The text does not specify the number or qualifications of experts used to establish the ground truth for the test sets. The ground truth ("EBV Status") was determined based on "EBV serological status" defined by a combination of other EBV serologies (VCA IgM, VCA IgG, EBNA IgG), which are established diagnostic markers. This suggests the ground truth was derived from established clinical laboratory testing rather than individual expert adjudication of images or complex clinical cases.
4. Adjudication method for the test set:
The term "adjudication method" as typically used for expert review (e.g., 2+1, 3+1) is not applicable here. The ground truth was established by interpreting results from a panel of other EBV serological tests (VCA IgG, VCA IgM, EBNA IgG), indicating a "rule-based" or "diagnostic criteria" approach rather than expert consensus on individual case interpretations.
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 MRMC comparative effectiveness study was conducted as this device is an ELISA assay, not an AI-powered diagnostic tool requiring human interpretation improvement. The device itself is designed to provide a result directly.
6. If a standalone (i.e., algorithm only without human-in-the-loop performance) was done:
Yes, the studies presented are essentially "standalone" performance evaluations. The EBV EA-D IgG ELISA Kit is an in vitro diagnostic assay. Its performance (sensitivity, specificity, agreement) is assessed based on its ability to detect antibodies in patient samples, independent of real-time human interpretation during the testing process. The "PANBIO results were compared to the EBV status of the sera," meaning the algorithm (the test kit's biochemical reaction and result interpretation) operated independently to produce a classification, which was then compared against the established reference.
7. The type of ground truth used:
The ground truth used was based on EBV serological status determined by a combination of other established EBV serologies:
- Seronegative: VCA IgG (-), VCA IgM (-), EBNA IgG (-)
- Acute Infectious Mononucleosis: VCA IgM (+), EBNA IgG (-)
- Past Infection: VCA IgG (+), VCA IgM (-), EBNA IgG (+)
This can be categorized as diagnostic criteria or reference method comparison using other established laboratory tests.
8. The sample size for the training set:
The document does not explicitly mention a "training set" for the ELISA kit. ELISA kits are generally developed and validated using biochemical principles and often use internal reference panels for calibration and quality control. The samples described in the performance studies are test sets used to evaluate the final device's performance, not to "train" an algorithm in the machine learning sense.
9. How the ground truth for the training set was established:
Since no explicit training set for an algorithm is mentioned, this question is not applicable to the description provided. The "training" of an ELISA kit primarily involves optimizing reagent concentrations, reaction conditions, and cutoff values during its development phase, typically using characterized control samples and internal validation studies, not a large patient-derived training dataset in the AI sense.
§ 866.3235 Epstein-Barr virus serological reagents.
(a)
Identification. Epstein-Barr virus serological reagents are devices that consist of antigens and antisera used in serological tests to identify antibodies to Epstein-Barr virus in serum. The identification aids in the diagnosis of Epstein-Barr virus infections and provides epidemiological information on diseases caused by these viruses. Epstein-Barr viruses are thought to cause infectious mononucleosis and have been associated with Burkitt's lymphoma (a tumor of the jaw in African children and young adults) and postnasal carcinoma (cancer).(b)
Classification. Class I (general controls).