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

Wondfo One Step Fecal Occult Blood Test is a rapid test for the qualitative detection of human occult blood in feces. It is used as an aid in the diagnosis of gastrointestinal(GI) bleeding. The device is suitable for use in laboratories and physician's offices as well as for over the counter use.

For in vitro diagnostic use only. For prescription use and over the counter use.

Device Description

The Wondfo One Step Fecal Occult Blood Test utilizes double antibodies sandwich immunoassay for the detection of hemoglobin in test samples. The test kit consists of:

Test devices, one test in one pouch. One pouch contains a test cassette and a desiccant. The desiccant is for storage purposes only and is not used in the test procedures.
Collection tubes with 1.5mL extraction buffer solution.
Clean collection papers.
Instructions for use.

AI/ML Overview

Here's a breakdown of the acceptance criteria and study information for the Wondfo One Step Fecal Occult Blood Test, based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance:

The document doesn't explicitly state "acceptance criteria" for each performance characteristic. Instead, it presents results that were deemed "acceptable" or demonstrated "substantial equivalence" to the predicate device. Therefore, the table below reflects the reported performance which implies it met the internal acceptance benchmarks.

Performance Characteristic	Acceptance Criteria (Implied)	Reported Device Performance
Precision	High overall, positive, and negative percent agreement.	Repeatability: Overall 99.3%, Positive 100.0%, Negative 98.1%
Lot-to-Lot Reproducibility: Overall 99.2%, Positive 99.2%, Negative 99.2%
Between-run Reproducibility: Overall 98.9%, Positive 99.2%, Negative 98.5%
Between-Device Reproducibility: Overall 99.3%, Positive 99.4%, Negative 99.2%
Between-site Reproducibility: Overall 98.9%, Positive 99.2%, Negative 98.5%
Combined Reproducibility: Overall 99.1%, Positive 99.2%, Negative 99.0%
Linearity (Prozone)	No false negatives at high hemoglobin concentrations.	Not susceptible to prozone/hook effect up to 200,000 ng/mL.
Sample Stability	Maintain performance over specified storage periods.	- Room Temp: Up to 15 days

2-8°C: Up to 30 days
-20°C: Up to 3 months |
| In-use Stability | Maintain performance for a reasonable in-use period. | Up to 1 hour under various temperature/humidity conditions. |
| Test Kit Stability | Maintain performance over specified storage periods. | - Accelerated: Estimated 24 months at 4-30°C
Real Time: Stable for at least 27 months at 4, 10, 20, 30°C. |
| Transport Stability | Maintain performance after exposure to transport conditions. | Stable up to 35 transport days when stored at -20°C and 40°C. |
| Interference | No significant interference from common substances. | No significant interference from vegetable extracts, Vitamin C, iron, or horseradish peroxidase. False negatives with 1% toilet cleaner. |
| Specificity (Hb Variants) | Detect common human hemoglobin variants. | Equivalently recognized HbA, HbS, and HbC. |
| Specificity (Cross-Reactivity) | No significant cross-reactivity with non-human heme or meat extracts. | No significant interference from non-human hemoglobin or animal meat extracts. |
| Cut-off | Consistent performance at the established cut-off. | Determined to be 1.35 µg hemoglobin/g stool or 45 ng/mL. Overall Agreement of 99.3% in cut-off study. |
| Method Comparison | Acceptable overall, positive, and negative percent agreement with predicate. | Overall Combined Sites: Overall 99.6%, Positive 99.2%, Negative 99.7% |
| Lay-user Study (Own Samples) | High agreement between lay-user and professional results. | Own Samples: 100% agreement between lay-users and professional testers. |
| Lay-user Study (Spiked Samples) | High agreement between lay-user and professional results for spiked samples. | Spiked Samples: Overall 98% agreement. (2 discrepant results at 37.5 ng/mL). |
| Lay-user Study (Instructions) | Instructions are easily understood. | All lay users indicated instructions were easily followed. |
| Reaction Time | Optimal reaction time for accurate results. | Demonstrated as 10 minutes (100% positivity at 50 ng/mL at 10 minutes). |
| Operator Intensity Reading | Consistency in reading test line intensity across operators. | No statistical significance in analyses by operators for intensity reading. |
| Specimen Collection Verification | Consistent delivery of specified stool amount by applicator. | Average collected volume of 0.0517g with standard deviation of 0.0027, demonstrating consistency. |

2. Sample Sizes and Data Provenance:

Precision/Reproducibility:
- Repeatability: 21 replicates for each of 7 concentrations (total 147 samples). Data provenance is "in house".
- Reproducibility: 21 replicates for each of 7 concentrations, tested at 3 POC sites, with 3 kit lots, 3 operators, and 5 non-consecutive days each. This would be 21 * 7 * 3 * 3 * 5 = 6615 tests in total across sites for the primary experiment, with reported summarized results from a subset of this (e.g., 735 for lot-to-lot, between-run, between-device, between-site, and 2205 for combined reproducibility). Data provenance: 3 POC sites in the U.S.
Linearity (Prozone Hook Effect): 5 aliquots for each of 7 concentrations (total 35 samples). Data provenance is "in house".
Sample Stability: For each condition (room temp, 2-8°C, -20°C): 7 different hemoglobin concentrations, tested repeatedly over time. The exact total sample count for stability studies (especially across all time points and conditions) is not explicitly totaled but would be substantial.
Interference: 21 aliquots of hemoglobin-negative stool spiked with 7 different Hb concentrations (total 147 samples per interfering substance tested). Tested against various vegetable extracts, vitamin C, iron, horseradish peroxidase, toilet cleaners.
Specificity (Hb Variants): 21 aliquots for each of 7 concentrations of each Hb variant (HbA, HbS, HbC). (Thus, 21 * 7 * 3 = 441 samples).
Cross-Reactivity: 21 aliquots of hemoglobin-negative stool spiked with 7 different Hb concentrations (total 147 samples per non-human Hb/meat extract). Tested against 8 non-human hemoglobins and 8 animal meat extracts.
Cut-off Study: 21 aliquots of each of 7 concentrations (total 147 samples). Tested side-by-side with predicate using the same sample set. Data provenance is "in house".
Comparison Studies (Method):
- 407 patient samples.
- 18 stool samples around the cut-off (purchased).
- 100 spiked samples.
- Total 525 samples analyzed in the comparison study tables.
- Data provenance: Performed at three POC testing sites.
Lay-user Study:
- Own stool samples: 100 lay users each testing their own stool sample (100 samples).
- Spiked samples: 20 spiked samples at each of 5 concentrations (0, 37.5, 50, 62.5, 2000 ng/mL) resulting in a total of 100 spiked samples. Each lay user tested one spiked sample.
- Data provenance: Performed at three intended user sites.
Test Kit Reaction Time: 21 replicates of each of 7 concentrations (total 147 samples).
Operator Intensity Reading: 25 replicates for each of 7 concentrations (total 175 samples).
Specimen Collection Verification: 5 positive and 5 negative clinical samples per lay user (5 * 10 = 50 samples for the primary part of the study). Stool weight measurements were taken from these.

The overall provenance of the data appears to be a mix of "in-house" (likely in China, given the manufacturer's location) and "US POC sites" for reproducibility and method comparison studies. All studies appear to be prospective for the purpose of device validation.

3. Number of Experts and Qualifications:

Adjudication and Ground Truth for Test Sets:
- In the precision, stability, linearity, interference, specificity, and cut-off studies, the "ground truth" was established by the known concentrations of human hemoglobin or other substances spiked into Hb-free stool samples. This is an analytical ground truth, not requiring expert review of the results in the same way clinical data would.
- For the Method Comparison Study, the predicate device (Orient Gene Biotech One Step Rapid FOB Test) served as the comparator, with its results being the "reference standard" against which the new device was compared. This implies the predicate device's results were considered the "ground truth" for comparison.
- For the Lay-user study: "Professional testing" results were used as the ground truth for comparison. The document does not specify the qualifications of these "professionals," but it implies they are trained laboratory or clinical staff.
- Specimen Collection Verification: A "professional operator" was involved in weighing samples. Qualifications are not specified.

4. Adjudication Method:

There is no mention of an adjudication method (like 2+1 or 3+1 consensus) for establishing the ground truth of the test sets. The studies rely on:
- Analytical ground truth: Spiked samples with known concentrations.
- Comparative ground truth: Comparison against a legally marketed predicate device or "professional testing" results.

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

No, a multi-reader multi-case (MRMC) comparative effectiveness study focusing on how much human readers improve with AI vs. without AI assistance was reported. This device is a rapid diagnostic test (FOBT) for the qualitative detection of human occult blood. It is not an AI-powered image analysis device, nor does it appear to involve human "readers" interpreting complex medical images in the way an MRMC study would typically evaluate. The "operators" or "lay users" in this context are performing the test and reading a simple positive/negative result, not interpreting complex data.
The "Operator Intensity Reading" study involved multiple readers (five) but it was to assess consistency of subjective intensity grading, not comparative effectiveness with/without AI assistance.

6. Standalone Performance:

Yes, extensive standalone performance was done. The entire "Analytical Performance" section (Precision/Reproducibility, Linearity, Stability, Interference, Specificity, Cut-off) details the algorithm's (device's) performance characteristics independently of human-in-the-loop clinical decision making, by testing samples with known analytical parameters.
The "Method Comparison Studies" also describe the standalone performance of the device against a predicate device.
The "Lay-user study" also tested the standalone performance of the device when used by lay users, comparing their results to professional results.

7. Type of Ground Truth Used:

Primarily analytical ground truth (known concentrations of human hemoglobin or other substances in spiked stool samples) for most of the analytical performance studies (precision, stability, linearity, interference, specificity, cut-off).
Comparative ground truth using a legally marketed predicate device (Orient Gene Biotech One Step Rapid FOB Test) for the method comparison study.
Comparison to "professional testing" results for the lay-user study.

8. Sample Size for the Training Set:

The document does not mention or specify a training set in the context of an algorithm or AI. This device is a lateral flow immunochromatographic assay, which is a chemical and biological test, not an AI/machine learning model that typically requires a large training dataset. The development and optimization of such a test would involve internal R&D experiments to determine optimal reagent concentrations, membrane properties, etc., but these are not referred to as a "training set" in the context of AI/ML.

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

As there is no mention of a training set for an AI/ML algorithm, this question is not applicable to the submitted document.

Summary

Regulation Number and Section

§ 864.6550 Occult blood test.

(a)
Identification. An occult blood test is a device used to detect occult blood in urine or feces. (Occult blood is blood present in such small quantities that it can be detected only by chemical tests of suspected material, or by microscopic or spectroscopic examination.)(b)
Classification. Class II (special controls). A control intended for use with an occult blood test is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 864.9.