(256 days)
The ACL TOP Family 50 Series (ACL TOP 750; ACL TOP 750 CTS; ACL TOP 750 LAS; ACL TOP 550 CTS; ACL TOP 350 CTS) are bench top, fully automated, random access analyzers designed specifically for in vitro diagnostic clinical use in the hemostasis laboratory for coagulation and/or fibrinolysis testing in the assessment of thrombosis and/or hemostasis. The systems provide results for both direct hemostasis measurements and calculated parameters.
The ACL TOP Family 50 Series are fully automated coagulation analyzers that utilize the same intuitive software, the same consumables, reagents, calibrators and provide the same analytical methodology for routine and specialty assay result reporting as the predicate ACL TOP Family. The ACL TOP Family 50 Series instrument performs the following types of tests, using the same optical measuring wavelengths and test parameters as the predicate ACL TOP Family: Coagulometric (Turbidimetric) Measurements, Chromogenic (Absorbance) Measurements, Immunological Measurements. The ACL TOP Family 50 Series also offers new pre-analytical features not available on the current ACL TOP Family: Pre-Analytical HIL Check, Pre-Analytical Tube Fill Height (THF) Check, and Pre-Analytical Clog Detection.
Acceptance Criteria and Device Performance for ACL TOP Family 50 Series
This document outlines the acceptance criteria and the results of the studies demonstrating that the ACL TOP Family 50 Series coagulation analyzers meet these criteria. The device is intended for in vitro diagnostic clinical use for hemostasis and fibrinolysis testing.
1. Table of Acceptance Criteria and Reported Device Performance
The acceptance criteria for this device are established through internal and external precision, linearity, and method comparison studies, specifically aiming for performance comparable to the predicate device (ACL TOP Family) and within CLSI guidelines. While explicit numeric acceptance criteria are not presented as a single consolidated table, the pass/fail status indicated in the reproducibility study and the strong correlation coefficients/slopes in the method comparison studies serve as de-facto acceptance criteria. The device consistently passed these criteria across various assays and models.
| Study Type | Assay/Parameter | Acceptance Criteria (Implied) | Reported Device Performance |
|---|---|---|---|
| Precision | Within-run %CV, Total %CV (various assays and levels) | Within CLSI EP05-A2 guidelines (Specific %CVs noted in tables) | All reported %CVs (within-run and total) passed acceptance criteria. |
| Linearity | Slope, Y-Intercept, r² (various assays) | r² close to 1, slope close to 1, intercept close to 0 (equivalent linearity as predicate) | All r² values were >0.9781, slopes were generally close to 1, and intercepts were low, demonstrating strong linearity. |
| Method Comparison | Slope, Intercept, r (against predicate ACL TOP 500) | r > 0.94, slope close to 1, intercept close to 0 (equivalent performance as predicate) | All r values were >0.94, slopes were generally close to 1, and intercepts were small across all sites and assays. |
| Pre-Analytical HIL Check (Hemoglobin, Bilirubin, Lipemia) | Strong correlation (r) to reference methods | Hemoglobin r: >0.944, Bilirubin r: >0.952, Lipemia: >91% Overall Matching | |
| Functional Check | Pre-Analytical Clog Detection | Correct detection of occluded sample probe | All instrument models correctly detected an occluded sample probe. |
| Pre-Analytical Tube Fill Height Check | Correct detection of underfilled tubes | All instrument models correctly detected when tubes were underfilled. |
2. Sample Sizes Used for the Test Set and the Data Provenance
-
Precision Study (Internal):
- Sample Size: 80 replicates per level for each of 12 assays on each instrument model (ACL TOP 350 CTS, ACL TOP 550 CTS, ACL TOP 750 CTS, ACL TOP 750). This involved commercially available assays and their assayed control materials, as well as a prepared patient plasma pool.
- Data Provenance: Internal, not specified country of origin, retrospective or prospective not explicitly stated but implies prospective as part of a 20-day study.
-
Precision Study (External):
- Sample Size: 80 replicates per level for each of 11 assays (two levels of control materials) on one ACL TOP 550 CTS model at three external US sites.
- Data Provenance: Three US external sites, implies prospective as part of a 20-day study.
-
Reproducibility Study (External):
- Sample Size: 30 replicates per level for each of 12 assays (two levels of control materials) on one ACL TOP 550 CTS model at three external US sites.
- Data Provenance: Three US external sites, implies prospective as part of a 5-day study.
-
Linearity Study (Internal):
- Sample Size: Minimum of 9 levels tested in quadruplicate for each of 10 assays on each instrument model (ACL TOP 350 CTS, ACL TOP 550 CTS, ACL TOP 750 CTS, ACL TOP 750). This involved prepared plasma pool panels.
- Data Provenance: Internal, not specified country of origin, implies prospective.
-
Method Comparison Study (External):
- Sample Size (Assays): Ranged from 61 to 146 patient samples per assay at each of the three US external sites.
- Sample Size (HIL Check): Hemoglobin (244-269 samples), Bilirubin (249-267 samples), Lipemia (257-272 samples) at each of the three US external sites.
- Data Provenance: Three US external sites, patient samples, implies prospective.
-
Pre-Analytical Clog Detection Testing:
- Sample Size: Not explicitly stated beyond "All instrument models" (2 ACL TOP 350 CTS; 2 ACL TOP 550 CTS; 2 ACL TOP 750; 2 ACL TOP 750 CTS; 1 ACL TOP 750 LAS).
- Data Provenance: Internal, likely simulated occluded samples.
-
Pre-Analytical Tube Fill Height Check Testing:
- Sample Size: Not explicitly stated beyond "All instrument models" (2 ACL TOP 350 CTS; 2 ACL TOP 550 CTS; 2 ACL TOP 750; 2 ACL TOP 750 CTS).
- Data Provenance: Internal, likely simulated underfilled tubes.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and the Qualifications of Those Experts
- For clinical assays (Precision, Linearity, Method Comparison): The ground truth is established by the calibrated reference methods and control materials themselves, which are well-established in clinical laboratory standards. No human "experts" are explicitly mentioned whose adjudication was required to establish ground truth for these quantitative measurements, as the values are determined by the analytical process and certified controls.
- For Pre-Analytical HIL Check (Lipemia): "Visual Matching" was used as a reference for Lipemia. While not explicitly stated, this would typically involve trained laboratory personnel or experts in visual assessment of lipemia. The number and qualifications of these individuals are not specified in the provided text.
- For Pre-Analytical Clog Detection and Tube Fill Height Check: The ground truth for these functional tests was whether the instrument correctly detected the intentionally created conditions (occluded probe, underfilled tube). This doesn't rely on expert human judgment for each instance, but rather on the designed functionality of the system.
4. Adjudication Method for the Test Set
- For the quantitative clinical assays, there is no indication of an adjudication method involving multiple human readers as the output is a numerical measurement. The CLSI guidelines followed (EP05-A2, EP09-A3) dictate statistical analysis of quantitative results.
- For the Lipemia "Visual Matching" reference method, no specific adjudication method (e.g., 2+1, 3+1) is mentioned.
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 MRMC comparative effectiveness study was done, as this device (ACL TOP Family 50 Series) is a fully automated laboratory instrument for coagulation testing, not an AI-assisted diagnostic imaging or interpretation tool that would involve human readers. The new pre-analytical features (HIL Check, Tube Fill Height Check, Clog Detection) are described as alerts to the instrument operator, but there is no mention of an AI component or a study on human reader performance with or without such assistance. The device's primary function is quantitative measurement.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
Yes, the studies presented are all standalone validations of the instrument's performance. The "ACL TOP Family 50 Series" is a fully automated analyzer. The precision, linearity, and method comparison studies directly evaluate the algorithm's (instrument's) ability to accurately and precisely measure coagulation parameters and detect pre-analytical issues without human intervention in the measurement process itself. The "Pre-Analytical HIL Check," "Pre-Analytical Clog Detection," and "Pre-Analytical Tube Fill Height Check" are new algorithmic features designed to flag samples automatically. The performance of these flags is evaluated in a standalone manner against a reference method (for HIL) or by simply checking if the instrument's detection mechanism worked correctly (for clog and tube fill).
7. The type of ground truth used
- For clinical assays (Precision, Linearity, Method Comparison): The ground truth is primarily based on reference methods and assayed control materials with known concentrations/values, as per established clinical laboratory standards (e.g., CLSI guidelines). For method comparison, the predicate device (ACL TOP 500) served as the reference for patient samples.
- For Pre-Analytical HIL Check:
- For Pre-Analytical Clog Detection and Tube Fill Height Check: The ground truth was known induced conditions (occluded probe, underfilled tubes) that the instrument was designed to detect.
8. The sample size for the training set
The provided document describes validation studies (precision, linearity, method comparison) for the ACL TOP Family 50 Series. It does not explicitly mention a "training set" or "test set" in the context of machine learning, implying that the device's algorithms for analytical measurements are based on established physicochemical principles and validated through these performance studies rather than being developed through a machine learning training paradigm. The studies described are for validation of the device's performance, not for training a model.
9. How the ground truth for the training set was established
As there is no explicit mention of a "training set" in the context of machine learning for generating the core analytical measurements or the pre-analytical flagging algorithms, the question of how ground truth was established for a training set is not applicable based on the provided information. The device's foundational analytical capabilities would have been developed using known chemical/biological principles and standard calibration processes. The new pre-analytical features are described as programmed detections (e.g., measuring at 535 nm for HIL, pressure transducer for clogs, optical detection for tube fill) and their performance validated against reference methods or known conditions.
§ 864.5400 Coagulation instrument.
(a)
Identification. A coagulation instrument is an automated or semiautomated device used to determine the onset of clot formation for in vitro coagulation studies.(b)
Classification. Class II (special controls). A fibrometer or coagulation timer intended for use with a coagulation instrument is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 864.9.