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

The FlexSure ® OBT (Occult Blood Test) is a rapid, visually read, qualitative immunochemical chromatographic method for detection of human hemoglobin from blood in fecal samples. Fecal occult blood tests are useful screening aids for detecting primarily lower gastrointestinal (g.i.) disorders that may be related to iron deficiency anemia, diverticulitis, ulcerative colitis, polyps, adenomas, colorectal cancers or other g.i. lesions that can bleed. FlexSure ® OBT is recommended for use by health professionals as part of routine physical examinations or when lower g.i. disorders are suspected.

Device Description

Not Found

AI/ML Overview

Here's an analysis of the acceptance criteria and the studies performed for the FlexSure OBT device, based on the provided text:

Acceptance Criteria and Device Performance for FlexSure® OBT

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly state pre-defined "acceptance criteria" with specific numerical targets. Instead, it describes performance characteristics and compares them to the predicate device or demonstrates the device's capability. I've extrapolated the implied acceptance criteria from the reported performance results, assuming that the observed performance was deemed acceptable for market clearance.

Characteristic	Implied Acceptance Criteria (Based on Study Results)	Reported Device Performance (FlexSure® OBT)	Study Section
Analytical Sensitivity	Reliably detect ≥ 0.2 mL blood/100g feces (approx. 2-3 mL daily GI bleeding).	Reliably detected 0.2 mL or more of added blood per 100 g of feces (0.3 mg hemoglobin per g of stool). This level is approximately equal to 2 to 3 mL of daily in vivo lower g.i. bleeding.	Analytical Performance - Sensitivity
Prozone Effect	Reliably detect blood levels up to where blood is generally visible in stool (~17mL/100g).	Reliably detected up to 17 mL of added blood per 100 g of feces (25 mg hemoglobin per g of stool).	Analytical Performance - Sensitivity
Cross-Reactivity	No cross-reactivity with common dietary animal hemoglobins or horse myoglobin.	Results were negative for horse myoglobin and hemoglobin from beef, chicken, fish, horse, pig, rabbit, goat, sheep, and turkey, even at levels exceeding normal dietary intake.	Analytical Performance - Cross Reactivity
Interference	100% accurate results with known interfering substances (plant peroxidases, ferrous iron, Vitamin C + hemoglobin).	Yielded 100% accurate test results with FlexSure® OBT when fecal samples spiked with excessive quantities of plant peroxidases, ferrous iron (to check for false-positives), and mixture of vitamin C and human hemoglobin (to check for false-negatives).	Analytical Performance - Interference
Reproducibility (Within-site)	High agreement (e.g., ≥95-100%) for negative, borderline, and positive results.	100% (90/90)	Analytical Performance - Reproducibility
Reproducibility (Between-site)	High agreement (e.g., ≥95%) for negative, borderline, and positive results.	97% (30/30 for negative, 30/30 for positive, and 27/30 for borderline positive sample).	Analytical Performance - Reproducibility
Readability / Repeatability (Agreement b/w readers)	Greater than 95% agreement between experienced and inexperienced readers, with expected variation at analytical cutoff.	Greater than 95% agreement among the two reader groups (experienced vs. inexperienced) except at the analytical cutoff of the test.	Analytical Performance - Readability/Repeatability
Apparent Specificity (Average Risk - Colorectal Neoplasia)	Comparable to or better than predicate device (Hemoccult®).	98.4% (1,702/1,729) for FlexSure® OBT vs. 97.2% (1,680/1,729) for Hemoccult®. (FlexSure® OBT performed better)	Clinical Performance - Average Risk Screening
Apparent Specificity (Average Risk - Any lower G.I. Pathology)	Comparable to or better than predicate device (Hemoccult®).	99.0% (1,666/1,682) for FlexSure® OBT vs. 98.8% (1,661/1,682) for Hemoccult®. (FlexSure® OBT performed slightly better)	Clinical Performance - Average Risk Screening
Positive Predictive Value (Average Risk - Colorectal Neoplasia)	Comparable to or better than predicate device (Hemoccult®).	15.6% (5/32) for FlexSure® OBT vs. 3.9% (2/51) for Hemoccult®. (FlexSure® OBT performed better)	Clinical Performance - Average Risk Screening
Relative Sensitivity (High Risk - Colorectal Neoplasia)	Comparable to or better than predicate device (Hemoccult®).	55.6% (25/45) for FlexSure® OBT vs. 51.1% (23/45) for Hemoccult®. (FlexSure® OBT performed better)	Clinical Performance - High Risk Population
Lower G.I. Specificity (Presumed Normal)	High specificity (e.g., ≥95%).	97.7% (86/88)	Clinical Performance - Studies with Presumed Normal Subjects
Effect of Diet (Specificity)	100% specificity regardless of restricted, unrestricted, or rare red meat diets.	100% specificity regardless of diet (restricted, unrestricted, and rare red meat).	Clinical Performance - Studies with Presumed Normal Subjects
Simulated Upper G.I. Bleeding (Specificity)	100% specificity for lower G.I. bleeding.	100% specificity in the four individuals who participated (ingested own blood to simulate upper g.i. bleeding).	Clinical Performance - Studies with Presumed Normal Subjects

2. Sample Size and Data Provenance (Test Set)

Average Risk Screening: 1,734 individuals.
- Data Provenance: Not explicitly stated, but clinical studies are generally assumed to be prospective in nature unless otherwise specified. The setting "worldwide" and mention of "U.S., Europe, Australia and Canada" for reader locations suggests potential multi-national data, or at least a multi-national perspective in study design, but the actual data collection for this cohort is not precisely identified by country. The context of a 510(k) submission to the FDA implies a focus on U.S. regulatory requirements, but doesn't preclude international data.
High Risk Population: 283 individuals.
- Data Provenance: Same as above.
Presumed Normal Subjects (Lower G.I. Specificity): 88 individuals.
- Data Provenance: Same as above.
Presumed Normal Subjects (Effect of Diet): 19 individuals.
- Data Provenance: Same as above.
Presumed Normal Subjects (Simulated Upper G.I. Bleeding): 4 individuals.
- Data Provenance: Same as above.
Analytical studies (cross-reactivity, interference, reproducibility, readability): Number of individual fecal samples or spiked samples varied; specific numbers are mentioned in some sections (e.g., 90 for within-site reproducibility, 30 for between-site, 16 for inexperienced readers). These are laboratory-based studies.

3. Number of Experts and Qualifications for Ground Truth

Analytical Studies (Readability/Repeatability):
- Number of experts: 3 "experienced" SKD technicians.
- Qualifications: "one with a doctorate degree and two with BA/BS degrees," who had "used the test extensively."
- An "inexperienced" group of 16 individuals with varied educational backgrounds (High School to M.D.) also participated for comparison, but were not establishing ground truth.
Clinical Studies (Colorectal Neoplasia, Lower G.I. Pathology):
- The document does not explicitly state the number or qualifications of experts used to establish the ground truth (e.g., diagnosis of colorectal neoplasia or other lower G.I. pathology). However, for these types of conditions, the ground truth would typically be established by clinical diagnoses from specialists (e.g., gastroenterologists, pathologists) based on colonoscopy, biopsy, and other clinical findings. The text states "apparent specificity for colorectal neoplasia" and "for any lower g.i. pathology," implying that a clinical follow-up or diagnostic gold standard was used, but the details of the specialists involved are absent.

4. Adjudication Method (Test Set)

The document does not specify an adjudication method (e.g., 2+1, 3+1, none) for the clinical ground truth. The clinical diagnoses were presumably based on standard medical practice for confirming conditions like colorectal neoplasia or other lower G.I. pathologies.

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

No, a typical MRMC comparative effectiveness study was not performed as commonly understood for AI/CAD devices.
The study did involve:
- Comparison of device performance to a predicate device (Hemoccult®) in clinical settings. This is a comparative effectiveness study in a broad sense, but not specifically an MRMC study comparing human readers with and without AI assistance.
- A "readability/repeatability" study comparing experienced vs. inexperienced device users (readers) using the device. This is a type of reader variability study, but not an MRMC study assessing AI's impact on diagnostic accuracy. The device itself is a visually-read, qualitative test; it's not an AI system assisting interpretation of complex images.

6. Standalone (Algorithm Only) Performance Study

Yes, the FlexSure® OBT demonstrably performs as a standalone device. The entire document details the performance of the OBT device itself, which is a rapid, visually-read immunochemical chromatographic method for detecting hemoglobin. There is no "human-in-the-loop" component in the sense of an AI interpreting images for a human. The human reader simply observes the qualitative result on the test strip. The "readability" study confirms that the device's output is consistently interpreted.

7. Type of Ground Truth Used

Analytical Studies:
- Spiked Samples: For sensitivity, cross-reactivity, interference, and reproducibility, the ground truth was established by precisely known quantities of human hemoglobin or other substances added to fecal samples in vitro.
- Expected Results: For interference, knowledge of substances' effects was used (e.g., plant peroxidases should not yield a positive result with FlexSure® OBT).
Clinical Studies:
- Clinical Diagnosis / Pathology: For colorectal neoplasia, lower G.I. pathology, and presumed normal status, the ground truth would have been established through a medical gold standard, likely involving colonoscopy, biopsy, and pathological examination for positive cases, and clinical assessment for presumed normal subjects. The text refers to "colorectal neoplasia" and "lower g.i. pathology," which are clinically diagnosed conditions.
- Simulated in vivo (for upper G.I. bleeding): For the simulated upper G.I. bleeding study, the ground truth was the knowledge that individuals ingested their own blood, and thus should not show lower G.I. bleeding as detected by the device.

8. Sample Size for the Training Set

The document does not mention a training set in the context of an AI/ML algorithm. This device is a manual, visually-read immunochemical test, not an AI-powered diagnostic system. Its development would involve chemical and biological optimization, not algorithm training on data like an AI.

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

Not applicable, as there is no mention or indication of a training set or AI algorithm for this device.

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.