(70 days)
The SYNCHRON LX20 Clinical Chemistry System is a fully automated, computer controlled, clinical chemistry analyzer intended for the in vitro quantitative measurement of a variety of analytes of clinical interest in biological fluids, such as, serum, plasma, urine, and cerebral spinal fluid (sample type is chemistry dependent).
The LX20 Clinical Chemistry System is a fully automated, computer controlled, clinical chemistry analyzer intended for the in vitro determination of a variety of general chemistries, therapeutic drugs, and other chemistries of clinical interest in biological fluids such as serum, plasma, urine, and cerebral spinal fluid (sample type is chemistry dependent). The analyzer operates in conjunction with reagents, calibrators, and controls designed for use with the system. The instrument features bar code identification of samples and reagents. It automatically dilutes samples and delivers them to the reaction cuvette along with readents and reaction constituents. The system analyzes up to 100 samples per run with up to 41 analytes per sample. Major hardware components include a reagent compartment, sample and reagent cranes, cartinge chemistry section, modular chemistry section, sample carousel and crane, hydropneumatics, electronics, and power supplies.
Here's a breakdown of the acceptance criteria and study information for the SYNCHRON LX™20 Clinical Chemistry System, based on the provided text:
Acceptance Criteria and Device Performance
The acceptance criteria for the SYNCHRON LX™20 Clinical Chemistry System appear to be defined by its substantial equivalence to existing predicate devices. This is demonstrated through method comparison and imprecision experiments, where the device's results are compared to those obtained from selected predicate methods. The performance is considered acceptable if it shows good correlation (high 'r' value, close to 1) and a slope close to 1 and intercept close to 0 in the method comparison studies, along with acceptable levels of imprecision (low %C.V.).
Table of Acceptance Criteria and Reported Device Performance:
Since specific acceptance thresholds for slope, intercept, r, SD, and %C.V. are not explicitly stated as numerical criteria in the document, I will infer them based on what would typically be considered good performance for substantial equivalence in clinical chemistry, and present the reported performance against general expectations.
| Analyte | Acceptance Criteria (General Expectation) | Reported Device Performance (Method Comparison: Slope, Intercept, r) | Reported Device Performance (Within-run Imprecision: Mean, SD, %C.V., N) |
|---|---|---|---|
| Method Comparison | Slope close to 1.0 (e.g., 0.95-1.05) | ||
| Intercept close to 0.0 (e.g., within ± a small absolute value) | |||
| Correlation coefficient (r) > 0.97 (ideally > 0.99) | - | - | |
| Albumin | - | 0.970, 0.070, 0.997 | Level 1: 2.21 g/dL, 0.04, 1.9%, 80 |
| Level 2: 4.81 g/dL, 0.03, 0.6%, 80 | |||
| Creatinine | - | 1.010, 0.04, 0.999 | Level 1: 0.59 mg/dL, 0.07, 12.2%, 80 |
| Level 2: 8.24 mg/dL, 0.18, 2.1%, 80 | |||
| Glucose | - | 0.991, -0.250, 0.996 | Level 1: 43.7 mg/dL, 1.3, 2.9%, 80 |
| Level 2: 397.1 mg/dL, 1.7, 0.4%, 80 | |||
| Phosphorus | - | 1.023, 0.27, 0.996 | Level 1: 1.80 mg/dL, 0.04, 2.5%, 80 |
| Level 2: 7.04 mg/dL, 0.05, 0.7%, 80 | |||
| Total Protein | - | 0.962, -0.01, 0.993 | Level 1: 3.54 g/dL, 0.06, 1.8%, 80 |
| Level 2: 7.52 g/dL, 0.05, 0.7%, 80 | |||
| Urea Nitrogen | - | 0.949, 1.13, 0.999 | Level 1: 7.7 mg/dL, 0.4, 5.0%, 80 |
| Level 2: 58.7 mg/dL, 0.5, 0.9%, 80 | |||
| Carbon Dioxide | - | 1.015, -0.70, 0.992 | Level 1: 10.58 mmol/L, 0.26, 2.5%, 80 |
| Level 2: 30.16 mmol/L, 0.29, 1.0%, 80 | |||
| Calcium | - | 0.990, -0.92, 0.998 | Level 1: 7.5 mg/dL, 0.08, 1.1%, 80 |
| Level 2: 13.5 mg/dL, 0.11, 0.8%, 80 | |||
| Chloride | - | 0.974, 1.04, 0.997 | Level 1: 81.11 mmol/L, 0.47, 0.6%, 80 |
| Level 2: 120.53 mmol/L, 0.84, 0.7%, 80 | |||
| Potassium | - | 1.008, -0.01, 0.999 | Level 1: 2.4 mmol/L, 0.02, 0.9%, 80 |
| Level 2: 7.29 mmol/L, 0.04, 0.6%, 80 | |||
| Sodium | - | 1.028, -4.22, 0.993 | Level 1: 111.1 mmol/L, 0.52, 0.5%, 80 |
| Level 2: 170.2 mmol/L, 1.01, 0.6%, 80 | |||
| Benzodiazepine | Concordance = 100% agreement | Concordance = 100% agreement | Cutoff Calibrator: 293 mA/min, 2.4, 0.8%, 20 |
| High Control: 362 mA/min, 3.0, 0.8%, 20 | |||
| Iron | - | 0.976, 9.03, 0.997 | Level 1: 51.2 ug/dL, 1.2, 2.4%, 80 |
| Level 2: 262.6, 2.2, 0.9%, 80 | |||
| Magnesium | - | 0.971, 0.09, 0.997 | Level 1: 1.12 mg/dL, 0.03, 3.0%, 80 |
| Level 2: 3.63 mg/dL, 0.06, 1.5%, 80 | |||
| Phenobarbital | - | 0.976, 0.050, 0.992 | Level 1: 9.32 ug/mL, 0.22, 2.3%, 80 |
| Level 2: 65.86 ug/mL, 1.42, 2.2%, 80 | |||
| Uric Acid | - | 0.977, -0.02, 0.999 | Level 1: 2.42 mg/dL, 0.03, 1.1%, 80 |
| Level 2: 10.48 mg/dL, 0.05, 0.5%, 80 | |||
| Lactate Dehydrogenase | - | 1.018, 0.20, 0.999 | Level 1: 49.9 IU/L, 1.6, 3.2%, 80 |
| Level 2: 363.5 IU/L, 3.4, 1.0%, 80 | |||
| Imprecision | %C.V. generally below 5%, with some exceptions for very low concentrations or specific analytes | - | - |
Study Details
2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)
- Test Set Sample Size: The sample sizes are explicitly stated for the "Estimated Serum/Plasma Within-run Imprecision" for each analyte and level. For most analytes, the imprecision study used N=80 samples (likely runs or replicates) for each of two levels (Level 1 and Level 2). For Benzodiazepines, it was N=20 for each of the Cutoff Calibrator and High Control.
- Data Provenance: The document does not specify the country of origin of the data. The studies described are performance validation studies for a new medical device, implying them to be prospective in nature, conducted to evaluate the device against established predicate methods. The samples were "Serum/Plasma" samples.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g. radiologist with 10 years of experience)
- The concept of "ground truth" established by experts (like radiologists) is typically relevant for interpretative diagnostic tests, not for quantitative clinical chemistry analyzers. For this type of device, the "ground truth" or reference standard for comparison is the results obtained from the predicate device itself. Therefore, specific human experts were not explicitly used to establish a subjective "ground truth" for the test set in the way they would be for image interpretation. The predicate device's results are taken as the validated reference.
4. Adjudication method (e.g. 2+1, 3+1, none) for the test set
- This concept of "adjudication" is also generally applied to subjective diagnostic evaluations where multiple human readers interpret data, and discrepancies need to be resolved. For a quantitative clinical chemistry analyzer, adjudication between human readers is not applicable. The comparison is objective, between the results of the new device and the predicate device.
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, an MRMC comparative effectiveness study was not done. This type of study assesses improvements in human performance with AI assistance and is irrelevant for a highly automated quantitative clinical chemistry analyzer. The device performs the measurements automatically without direct human interpretation of raw data in a way that would be "assisted" by AI.
6. If a standalone (i.e. algorithm only without human-in-the loop performance) was done
- Yes, this entire study represents standalone performance. The SYNCHRON LX™20 Clinical Chemistry System is an automated analyzer, and the reported performance data (method comparison, imprecision) reflects the algorithm and hardware's ability to measure analytes independently. Human involvement is in operating the instrument and interpreting the final quantitative results, but the measurement itself is automated.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc)
- For the method comparison studies, the "ground truth" as a reference for comparison was the results obtained from the specified predicate devices. These predicate devices are already cleared for commercial distribution and their performance is established. For imprecision studies, the ground truth is statistical reliability of the device's own measurements.
8. The sample size for the training set
- Not Applicable / Not Provided. Clinical chemistry analyzers like the SYNCHRON LX™20 typically do not rely on "training sets" in the way machine learning algorithms do. Their performance is based on established chemical reactions, photometric principles, and instrument design. Calibration is performed using specific calibrators (e.g., SYNCHRON LX™ AQUA CAL 1, 2, 3) which are part of routine operation, but this is distinct from an AI "training set." The document does not mention any machine learning or AI components that would require a dedicated training set.
9. How the ground truth for the training set was established
- Not Applicable. As explained above, for this type of device, there is no "training set" in the context of AI/machine learning, and therefore no ground truth established for such a set.
§ 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.