(90 days)
The BÜHLMANN fCAL turbo is an in vitro diagnostic assay intended for the quantitative measurement of fecal calprotectin, a neutrophilic protein that is a marker of inflammation, in human stool. The BÜHLMANN fCAL turbo aids in the diagnosis of inflammatory bowel disease (IBD), specifically Crohn's disease (CD) and ulcerative colitis (UC) and aids in the differentiation of IBD from irritable bowel syndrome (IBS) in conjunction with other laboratory and clinical findings.
The BÜHLMANN CALEX Cap is a single use tube intended for the preparation of human stool samples to be used with the BÜHLMANN fCAL turbo.
The BÜHLMANN fCAL® turbo, a particle-enhanced turbidimetric immunoassay (PETIA), is performed using patient stool extracts collected without preservatives. Calprotectin within the sample extract mediates immunoparticle agglutination; sample turbidity is proportional to calprotectin concentration. The detected light absorbance allows quantification of calprotectin concentration via interpolation of an established calibration curve. The assay is validated for use on clinical chemistry analyzers such as the Roche cobas® c501/c502 platforms.
The BÜHLMANN fCAL® turbo Reagent Kit is to be used in conjunction with the BÜHLMANN fCAL® turbo Calibrator Kit and BÜHLMANN fCAL® turbo Control Kit, which are available separately.
Sample extracts may be prepared using manual weighing extraction methods or the CALEX® Cap.
The CALEX® Cap is a single use tilled with extraction buffer. The sampling pin houses a dosing tip which is used to obtain sufficient stool sample for the extraction process. The extraction method leads to stool specimen extracts which can be measured directly using the BÜHLMANN fCAL® turbo assay.
This document describes the validation of the BÜHLMANN fCAL® turbo and CALEX® Cap device, an in vitro diagnostic assay for measuring fecal calprotectin.
1. Table of Acceptance Criteria (Implied) and Reported Device Performance
The document doesn't explicitly list "acceptance criteria" in a separate table, but the performance data sections demonstrate that the device meets various analytical and clinical performance standards. The key performance metrics are presented as results from studies.
| Performance Metric | Reported Device Performance |
|---|---|
| Precision | Single-Site Repeatability (Manual Weighing): %CVs ranged from 0.7% to 8.3%. |
| Multi-Site Reproducibility (Manual Weighing): Total Precision %CVs ranged from 3.2% to 9.1%. | |
| Lot-to-Lot Precision (Manual Weighing): Total Precision %CVs ranged from 3.6% to 11.3%. | |
| Extraction Reproducibility (Manual Weighing): Total Precision %CVs ranged from 4.5% to 13.0%. | |
| Extraction Reproducibility (CALEX® Cap): Total Precision %CVs ranged from 8.1% to 19.7%. | |
| Linearity | R^2 values: 0.9983 and 0.9984 for two dilution series. |
| Measuring Range: Direct 30 - 2000 µg/g; with automatic dilution 30 - 10,000 µg/g. | |
| High Dose Hook Effect | No high dose hook effect observed up to 45,715 µg/g. |
| Accuracy/Recovery | Total recovery ranged from 93.6% to 102.0% across various spiked samples. |
| Analytical Sensitivity | LoB: 16.7 µg/g |
| LoD: 23.7 µg/g | |
| LoQ: 30 µg/g | |
| Interfering Substances | No interference observed from various common pharmaceuticals, nutritional supplements, and enteropathological microorganisms at tested concentrations. |
| Method Comparison (vs. predicate ELISA) | Correlation (r): 0.972 |
| PPA (lower cutoff): 93.6% [88.5%, 96.9%] | |
| NPA (lower cutoff): 91.3% [83.6%, 96.2%] | |
| PPA (upper cutoff): 93.9% [87.1%, 97.7%] | |
| NPA (upper cutoff): 95.3% [90.6%, 98.1%] | |
| Extraction Method Comparison (CALEX® Cap vs. Manual) | Correlation (r): 0.921 |
| PPA (lower cutoff): 98.1% [94.7%, 99.6%] | |
| NPA (lower cutoff): 89.9% [81.0%, 95.5%] | |
| PPA (upper cutoff): 97.6% [93.1%, 99.5%] | |
| NPA (upper cutoff): 96.6% [91.4%, 99.1%] | |
| Clinical Sensitivity/Specificity (Manual Weighing) | IBD vs. IBS (Borderline Considered Positive): Sensitivity 91.1%, Specificity 76.2% |
| IBD vs. IBS (Borderline Considered Negative): Sensitivity 80.0%, Specificity 87.7% | |
| IBD vs. non-IBD (Borderline Considered Positive): Sensitivity 91.1%, Specificity 74.3% | |
| IBD vs. non-IBD (Borderline Considered Negative): Sensitivity 80.0%, Specificity 85.1% | |
| Expected Values/Reference Range | 75.2% of apparently healthy adults (160 µg/g). |
2. Sample Size and Data Provenance
-
Test Set Sample Sizes:
- Precision Studies: 80 replicates per sample for Single-Site and Extraction Reproducibility, 75 replicates per sample for Multi-Site and Lot-to-Lot precision.
- Linearity: Not explicitly stated as a number of unique patient samples, but involves two dilution series with "various mixing ratios" and 4 replicates per dilution.
- Accuracy/Recovery: 7 samples.
- Interfering Substances: Stool specimen extracts with 4 approximate calprotectin concentrations; specific number of runs not explicitly stated.
- Method Comparison (vs. predicate ELISA): 248 clinical study samples.
- Extraction Method Comparison (CALEX® Cap vs. Manual): 241 clinical study samples.
- Clinical Sensitivity/Specificity: 265 samples for IBD vs. IBS analysis, 337 samples for IBD vs. non-IBD analysis.
- Expected Values/Reference Range: 141 apparently healthy normal adults.
-
Data Provenance: The document does not explicitly state the country of origin for the clinical study data or whether it was retrospective or prospective. Given the submitter's location (Switzerland), it's possible some or all studies were conducted there or internationally. The term "clinical study samples" usually implies prospective collection, but it's not explicitly stated.
3. Number of Experts and Qualifications for Ground Truth
- This document is for an in vitro diagnostic (IVD) assay, specifically an immunological test system for fecal calprotectin. The "ground truth" for such devices often relies on established laboratory methods, clinical diagnoses based on a constellation of evidence (e.g., endoscopy, histology, imaging, other lab findings), or consensus among clinicians.
- No "experts" in the sense of human readers for images are mentioned or used for establishing ground truth. Instead, the device's performance is compared against:
- Itself (for precision, linearity, recovery, sensitivity).
- A legally marketed predicate device (for method comparison).
- External clinical diagnoses (for clinical sensitivity/specificity).
- For clinical diagnosis (e.g., IBD, IBS, non-IBD), the ground truth is implicitly established by clinical findings, which would typically involve qualified physicians (e.g., gastroenterologists) and pathologists, but their specific number or qualifications are not detailed as this is standard for IVD test validation.
4. Adjudication Method for the Test Set
- Adjudication methods (like 2+1, 3+1) are typically used for studies involving human interpretation of complex data (e.g., medical images) where a consensus is needed to define the ground truth for equivocal cases.
- This is an IVD device, not an image-based AI system. Therefore, there is no "adjudication method" as described in the context of human reader studies. The "ground truth" for clinical sensitivity/specificity would be the established clinical diagnosis for each patient, determined by standard medical practice, not by expert adjudication of the test results themselves.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
- No MRMC study was done. MRMC studies are designed to compare the diagnostic performance of human readers, often with and without AI assistance, on a set of medical cases.
- This document describes the validation of an in vitro diagnostic (IVD) test kit, which is a laboratory assay. It does not involve human readers interpreting images or any AI assistance for human interpretation.
6. Standalone (Algorithm Only Without Human-in-the-Loop Performance) Study
- Yes, this entire submission effectively represents a standalone performance study of the BÜHLMANN fCAL® turbo and CALEX® Cap device.
- As an IVD assay, its core function is to produce quantitative results based on the chemical reaction. The performance data (precision, linearity, accuracy, analytical sensitivity, method comparison) demonstrate its intrinsic performance independent of human interpretation of its raw output. The subsequent clinical sensitivity and specificity studies evaluate the diagnostic accuracy of the device's numerical output against clinical diagnoses. There is no "human-in-the-loop" in the direct operation or result generation of the assay itself, apart from standard laboratory procedures and subsequent clinical interpretation of the numerical calprotectin values.
7. Type of Ground Truth Used
- Analytical Ground Truths:
- Known concentrations: For linearity, analytical sensitivity (LoB, LoD, LoQ), accuracy/recovery, and interfering substances, the ground truth involves samples with known or spiked concentrations of calprotectin or interferents.
- Reference Methods/Predicate Device: For method comparison, the predicate BÜHLMANN fCAL® ELISA assay serves as a comparative ground truth.
- Clinical Ground Truths:
- Clinical Diagnosis: For clinical sensitivity and specificity, the ground truth for patients was their established clinical diagnosis (e.g., IBD, IBS, non-IBD). This diagnosis is typically based on an aggregation of clinical findings, endoscopic results, histological reports, and other laboratory tests, representing the best available clinical determination.
- Healthy Normal Cohort: For expected values/reference range, the ground truth was a cohort of "apparently healthy normal adults" based on lack of symptoms or signs of gastrointestinal disease, implying a clinical assessment.
8. Sample Size for the Training Set
- This document describes the validation studies for a pre-market submission (510(k)) for an IVD device. For such devices, particularly those based on immunoassay technology, the concept of a "training set" like that used in machine learning (ML) is generally not applicable in the same way. The device's underlying immunoassay methodology is laboratory-based and its parameters are established through extensive analytical development rather than data-driven "training" in the ML sense.
- The document presents performance data from verification and validation studies, which are essentially "test sets" for the final device. There is no mention of a separate "training set" of patient data for model development as one would see with a deep learning algorithm. The analytical parameters and performance characteristics of the immunoassay are developed and optimized by the manufacturer.
9. How the Ground Truth for the Training Set Was Established
- As explained above, the concept of a distinct "training set" and associated ground truth establishment in the context of an ML algorithm is not directly applicable to this immunoassay device's validation as described in the 510(k) summary. The development of such assays relies on biochemical principles, calibration curves established with known standards, and analytical validation experiments performed on various matrices and samples. The "ground truth" for developing the assay itself would be the precisely measured values of known calprotectin standards.
§ 866.5180 Fecal calprotectin immunological test system.
(a)
Identification. A fecal calprotectin immunological test system is anin vitro diagnostic device that consists of reagents used to quantitatively measure, by immunochemical techniques, fecal calprotectin in human stool specimens. The device is intended forin vitro diagnostic use as an aid in the diagnosis of inflammatory bowel diseases (IBD), specifically Crohn's disease and ulcerative colitis, and as an aid in differentiation of IBD from irritable bowel syndrome.(b)
Classification. Class II (special controls). The special control for these devices is FDA's guidance document entitled “Class II Special Controls Guidance Document: Fecal Calprotectin Immunological Test Systems.” For the availability of this guidance document, see § 866.1(e).