(45 days)
The Dimension® EXL™ with LM system is an in vitro diagnostic device that is intended to measure a variety of analytes in human body fluids. The system utilizes photometric, turbidimetric, chemiluminescence and integrated ion selective multisensor technology for chemical and immunochemical applications for clinical use.
The FT4L method is an in vitro diagnostic test for the quantitative measurement of Free Thyroxine in human serum and plasma on the Dimension® EXL™ with LM system. Measurements of free thyroxine are used in the diagnosis and monitoring of thyroid disease.
The Dimension EXL with LM system is a floor model, fully automated, microprocessorcontrolled, integrated instrument which uses prepackaged Siemens Dimension Flex® reagent cartridges to measure a variety of analytes in human body fluids. The system can process samples in random access, batch or STAT modes. The instrument has a heterogeneous module (HM) for processing chromium-based heterogeneous immunoassays and a LOCI® module for chemiluminescent immunoassays. The instrument can also perform photometric, turbidimetric and ACMIA tests. This 510(k) is being submitted because a new photomultiplier (from a different vendor) will be used to count the signal for the chemiluminescent methods (LOCI).
The Dimension® FT4L Flex® regent cartridge consists of prepackaged liquid reagents containing two synthetic beads, and a biotinvlated anti-T4 mouse monoclonal antibody in a plastic eight-well cartridge.
The FT4L method is a homogeneous, sequential, chemiluminescent immunoassay based on. LOCI® technology. The LOCI® reagents include two synthetic bead reagents and a biotinylated anti-T4 mouse monoclonal antibody. The first bead reagent (Chemibeads) is coated with trijodothyronine (T3), a naturally occurring, weaker binding analog of T4, and contains chemiluminescent dye. The second bead reagent (Sensibeads) is coated with streptavidin and contains a photosensitizer dye. In a first step, sample is incubated with biotinylated antibody which allows T4 from the sample to saturate a fraction of the biotinylated antibody that is directly related to the free thvroxine (FT4) concentration. In a second step, T3 Chemibeads are added and form bead/biotinylated antibody immunocomplexes with the nonsaturated fraction of the biotinylated antibody. Sensibeads are then added and bind to the biotin to form bead pair immunocomplexes. Illumination of the complex at 680 nm generates singlet . oxygen from Sensibeads which diffuses into the Chemibeads, triggering a chemiluminescent reaction. The resulting signal is measured at 612 nm and is an inverse function of the FT4 concentration in the sample.
Here's an analysis of the acceptance criteria and study detailed in the provided 510(k) summary for the Siemens Dimension® EXL™ with LM system and FT4L Flex® reagent cartridge.
Note: This submission focuses on a change in a photomultiplier tube (PMT) vendor for an existing device. Therefore, the "acceptance criteria" discussed are primarily related to demonstrating that the new PMT does not negatively impact the performance of a representative assay (FT4L) compared to the previously cleared system. This is a comparative effectiveness study rather than assessing absolute performance against disease diagnosis benchmarks.
Acceptance Criteria and Reported Device Performance
1. Table of Acceptance Criteria and the Reported Device Performance
| Performance Metric | Acceptance Criteria | Reported Device Performance (New PMT vs. Current PMT on FT4L Assay) | Meets Criteria? |
|---|---|---|---|
| Method Comparison (FT4L Assay) | |||
| Correlation Coefficient (Linear Regression) | Not explicitly stated as a numerical threshold, but implied to be "good agreement" and "substantially equivalent." Typically, a value close to 1.0 indicates strong correlation. | 0.998 | Yes (Excellent) |
| Slope (Passing Bablok) | Expected to be close to 1.0 (indicating proportional agreement). | 0.99 (95% CI: 0.96 – 1.01) | Yes |
| Intercept (Passing Bablok) | Expected to be close to 0.0 (indicating no constant bias). | -0.03 ng/dL (95% CI: -0.07 – 0.00) | Yes |
| Precision (FT4L Assay) - Repeatability (%CV) | FT4L Target: $0.6 \pm 0.2$ ng/dL (0.4-0.8) | ||
| $\leq 5.0%$ | |||
| FT4L Target: $1.5 \pm 0.4$ ng/dL (1.1-1.9) | |||
| $\leq 3.0%$ | |||
| FT4L Target: $4.0 \pm 2.0$ ng/dL (2.0-6.0) | |||
| $\leq 3.0%$ | BioRad QC Level 1 (Mean 0.89 ng/dL): 2.0% | ||
| BioRad QC Level 2 (Mean 2.44 ng/dL): 1.9% | |||
| BioRad QC Level 3 (Mean 6.85 ng/dL): 1.5% | |||
| Patient Pool 1 (Mean 1.17 ng/dL): 1.7% | |||
| Patient Pool 2 (Mean 3.93 ng/dL): 1.6% | Yes (All values are well below the specified CV % limits) | ||
| Precision (FT4L Assay) - Within-Lab (%CV) | FT4L Target: $0.6 \pm 0.2$ ng/dL (0.4-0.8) | ||
| $\leq 7.0%$ | |||
| FT4L Target: $1.5 \pm 0.4$ ng/dL (1.1-1.9) | |||
| $\leq 5.0%$ | |||
| FT4L Target: $4.0 \pm 2.0$ ng/dL (2.0-6.0) | |||
| $\leq 5.0%$ | BioRad QC Level 1 (Mean 0.89 ng/dL): 3.4% | ||
| BioRad QC Level 2 (Mean 2.44 ng/dL): 2.8% | |||
| BioRad QC Level 3 (Mean 6.85 ng/dL): 2.4% | |||
| Patient Pool 1 (Mean 1.17 ng/dL): 2.7% | |||
| Patient Pool 2 (Mean 3.93 ng/dL): 1.7% | Yes (All values are well below the specified CV % limits) |
Notes on Acceptance Criteria: The document primarily states that the acceptance criteria for precision are either "as follows" (referring to the table provided) or "must be determined to be substantially equivalent to the predicate device." Since the new PMT system demonstrated precision values well within the specified limits, and also showed very similar performance to the current PMT system (data for current PMT also provided in the document, which are very close), it successfully met these criteria. The method comparison's strong correlation, slope near 1, and intercept near 0 directly support "substantially equivalent" performance.
Study Details
2. Sample size used for the test set and the data provenance
- Method Comparison Test Set Sample Size: 45 patient samples
- Precision Test Set Sample Size: 3 commercial quality controls (BioRad Liquichek Immunoassay QC - 3 levels) and 2 patient serum pools.
- Data Provenance:
- Country of Origin: Not explicitly stated, but the studies were conducted "internally by Siemens Healthcare Diagnostic Inc. R & D organization personnel," suggesting a US or European location where Siemens R&D is prominent. Given the FDA submission, the data is likely intended to be representative for the US market.
- Retrospective or Prospective:
- Method Comparison: Retrospective. "Remnant de-identified human serum samples were tested." This implies samples were collected previously for other purposes.
- Precision: Prospective. Testing was performed "over twenty (20) days, one (1) run per day for each test material." This indicates systematic, planned testing for this study.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts
- Not Applicable. This is an in vitro diagnostic (IVD) device measuring a quantitative biomarker (Free Thyroxine). The ground truth for such devices is established through laboratory methods and reference materials, not expert human interpretation (like in imaging or clinical diagnosis). The "test set" and "ground truth" here refer to the accurate measurement of FT4L concentrations.
- The "ground truth" for the comparative study is the measurement obtained from the predicate device (Dimension® EXL™ with LM system with the current PMT), which is the standard of comparison for demonstrating substantial equivalence. The study aims to show that the new device yields the same results as the predicate.
4. Adjudication method (e.g., 2+1, 3+1, none) for the test set
- None. Adjudication methods like 2+1 or 3+1 are typical for studies involving human interpretation or subjective assessments (e.g., in radiology studies where multiple readers interpret images, and discrepancies need resolution). For quantitative IVD devices, analytical measurements are directly compared.
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. This is not a multi-reader multi-case (MRMC) comparative effectiveness study. This type of study is relevant for AI-powered diagnostic tools that assist human readers in interpreting complex data (e.g., medical images).
- This submission is for a change in a component of an automated laboratory analyzer that measures a quantitative biomarker. There is no human "reader" involved in the direct output of the Free Thyroxine measurement, nor is there AI assistance to a human for interpretation in the context of this device's function.
6. If a standalone (i.e., algorithm only without human-in-the-loop performance) was done
- Yes, effectively (for the measurement component). The device itself is an automated analyzer. The "algorithm only" performance is the measurement output by the instrument. The study directly compares the numerical results from the device with the new PMT to the device with the current PMT. There is no "human-in-the-loop" once the sample is loaded; the instrument performs the analysis automatically.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.)
- Reference Method/Predicate Device Performance: The primary "ground truth" for demonstrating substantial equivalence in this context is the performance of the previously cleared predicate device (Dimension® EXL™ with LM system with the original PMT). The study aims to show that the new device produces results that are statistically indistinguishable from those produced by the predicate device.
- For the precision study, commercial quality controls and patient pools with known, stable concentrations serve as the reference materials monitored for consistency.
- For traceability, "USP-grade thyroxine spiked into stripped human serum" and a "Master Pool" calibrated against an "Anchor Pool" are used for defining the FT4L values.
8. The sample size for the training set
- Not Applicable. This is not an AI/machine learning device that requires a training set in the conventional sense. The device is a traditional in vitro diagnostic analyzer. While the instrument and assay are developed and optimized (which involves internal testing and calibration), the concept of a "training set" as understood in AI studies does not apply here.
9. How the ground truth for the training set was established
- Not Applicable. As explained above, there is no "training set" in the context of AI/ML. The device's underlying principles are based on established chemical and immunoassay reactions, not trained algorithms. The "ground truth" for calibrating the system and ensuring its accuracy is established through stringent analytical chemistry practices, including the use of reference materials, calibrators, and validation against established methods (as mentioned in the traceability section).
§ 862.2160 Discrete photometric chemistry analyzer for clinical use.
(a)
Identification. A discrete photometric chemistry analyzer for clinical use is a device intended to duplicate manual analytical procedures by performing automatically various steps such as pipetting, preparing filtrates, heating, and measuring color intensity. This device is intended for use in conjunction with certain materials to measure a variety of analytes. Different models of the device incorporate various instrumentation such as micro analysis apparatus, double beam, single, or dual channel photometers, and bichromatic 2-wavelength photometers. Some models of the device may include reagent-containing components that may also serve as reaction units.(b)
Classification. Class I (general controls). The device is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to § 862.9.