(77 days)
The ACL TOP is a bench top, fully automated, random access analyzer 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 system provides results for both direct hemostasis measurements and calculated parameters.
The ACL TOP is a bench top, fully automated, random access analyzer 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 system provides results for both direct hemostasis measurements and calculated parameters.
This document describes the ACL TOP, an automated coagulation analyzer, and presents data to support its substantial equivalence to a predicate device, the ACL Advance.
1. Acceptance Criteria and Reported Device Performance:
The primary acceptance criteria for the ACL TOP are based on demonstrating substantial equivalence to the predicate device, the ACL Advance, through method comparison. This is achieved by showing statistical similarity between the measurements obtained by both devices for various coagulation parameters.
| Reagent Type | Performance Metric (ACL TOP vs. ACL Advance) | Reported Device Performance | Acceptance Criteria (Implied by Predicate Equivalence) | |
|---|---|---|---|---|
| Antithrombin (%) | Slope | 1.03 | Slope close to 1.0 | |
| Intercept | -1.418 | Intercept close to 0.0 | ||
| Correlation Coefficient (r) | 0.9660 | Acceptably high correlation (e.g., >0.95 or similar to predicate's known performance) | ||
| APTT (Seconds) | Slope | 1.076 | Slope close to 1.0 | |
| Intercept | -0.380 | Intercept close to 0.0 | ||
| Correlation Coefficient (r) | 0.9943 | Acceptably high correlation | ||
| D-Dimer (ng/mL) | Slope | 1.12 | Slope close to 1.0 | |
| Intercept | -16.0 | Intercept close to 0.0 | ||
| Correlation Coefficient (r) | 0.993 | Acceptably high correlation | ||
| Factor II (%) | Slope | 0.95 | Slope close to 1.0 | |
| Intercept | -0.551 | Intercept close to 0.0 | ||
| Correlation Coefficient (r) | 0.9753 | Acceptably high correlation | ||
| Factor V (%) | Slope | 0.81 | Slope close to 1.0 | |
| Intercept | 4.742 | Intercept close to 0.0 | ||
| Correlation Coefficient (r) | 0.9822 | Acceptably high correlation | ||
| Factor VII (%) | Slope | 0.88 | Slope close to 1.0 | |
| Intercept | 3.153 | Intercept close to 0.0 | ||
| Correlation Coefficient (r) | 0.9922 | Acceptably high correlation | ||
| Factor X (%) | Slope | 0.97 | Slope close to 1.0 | |
| Intercept | 2.995 | Intercept close to 0.0 | ||
| Correlation Coefficient (r) | 0.9954 | Acceptably high correlation | ||
| Fibrinogen-C (mg/dL) | Slope | 1.00 | Slope close to 1.0 | |
| Intercept | -8.740 | Intercept close to 0.0 | ||
| Correlation Coefficient (r) | 0.9759 | Acceptably high correlation | ||
| Protein C (%) | Slope | 1.15 | Slope close to 1.0 | |
| Intercept | -0.323 | Intercept close to 0.0 | ||
| Correlation Coefficient (r) | 0.9902 | Acceptably high correlation | ||
| Prothrombin Time (PT) (Seconds) | Slope | 0.990 | Slope close to 1.0 | |
| Intercept | 1.46 | Intercept close to 0.0 | ||
| Correlation Coefficient (r) | 0.9987 | Acceptably high correlation | ||
| PT-Based Fibrinogen (mg/dL) | Slope | 1.084 | Slope close to 1.0 | |
| Intercept | -9.93 | Intercept close to 0.0 | ||
| Correlation Coefficient (r) | 0.9587 | Acceptably high correlation | ||
| Precision | Percentage Coefficient of Variation (%CV) within acceptable clinical limits, and comparable to the predicate device's known precision. (No explicit numerical acceptance criteria are stated, but the reported values show good precision). | |||
| Antithrombin (%) | Within Run %CV | Normal: 5.7, Low Abnormal: 5.6, High Abnormal: 6.8 | ||
| Total %CV | Normal: 5.8, Low Abnormal: 6.8, High Abnormal: 9.1 | |||
| APTT (Seconds) | Within Run %CV | Normal: 1.2, Low Abnormal: 0.9, High Abnormal: 0.9 | ||
| Total %CV | Normal: 1.6, Low Abnormal: 2.1, High Abnormal: 1.4 | |||
| D-Dimer (ng/mL) | Within Run %CV | Low Control: 4.6, High Control: 2.5 | ||
| Total %CV | Low Control: 7.7, High Control: 4.5 | |||
| ... (and so on for all listed reagents, similar precision metrics apply) |
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:
- Antithrombin: n=123 (method comparison), n=80 (precision)
- APTT: n=205 (method comparison), n=80 (precision)
- D-Dimer: n=120 (method comparison), n=80 (precision)
- Factor II: n=101 (method comparison), n=80 (precision)
- Factor V: n=93 (method comparison), n=80 (precision)
- Factor VII: n=96 (method comparison), n=80 (precision)
- Factor X: n=110 (method comparison), n=80 (precision)
- Fibrinogen-C: n=98 (method comparison), n=80 (precision)
- Protein C: n=123 (method comparison), n=80 (precision)
- Prothrombin Time (PT): n=150 (method comparison), n=80 (precision)
- PT-Based Fibrinogen: n=93 (method comparison), n=80 (precision)
- Data Provenance: The document states "in-house performance data" and "method comparison studies evaluating citrated plasma samples." It does not specify the country of origin of the data or whether the study was retrospective or prospective. However, given the context of a 510(k) summary for an in vitro diagnostic device, it is highly likely that these were prospective studies conducted in a laboratory setting. The samples were "citrated plasma samples," which are common for coagulation testing.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts
The concept of "ground truth" as typically used for AI/ML devices (e.g., expert consensus, pathology, outcomes data) does not directly apply here. For an in vitro diagnostic device like the ACL TOP, the "ground truth" for the method comparison study is implicitly established by the measurements obtained from the predicate device (ACL Advance). The predicate device itself has been cleared by the FDA and its performance characteristics are accepted as a standard against which the new device is compared. Similarly, for precision studies, the "ground truth" is the true analytical variability inherent to the control materials or samples used.
There is no mention of external human experts establishing ground truth for these types of analytical performance studies; the focus is on the analytical agreement between the new device and the predicate device.
4. Adjudication method (e.g. 2+1, 3+1, none) for the test set
Not applicable. Adjudication methods are typically employed in studies where human interpretation of medical images or data is being evaluated, often to resolve discrepancies between readers or between human readers and an AI algorithm. For an in vitro diagnostic instrument like the ACL TOP, the performance is assessed through quantitative measurements and statistical comparison against a predicate device, rather than subjective interpretations requiring adjudication.
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
Not applicable. An MRMC study is relevant for diagnostic imaging AI/CAD systems that assist human readers in tasks like lesion detection or diagnosis. The ACL TOP is an automated laboratory instrument measuring coagulation parameters; it does not involve human readers interpreting AI output.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
Yes, the studies presented are effectively standalone performance evaluations. The ACL TOP is a "fully automated, random access analyzer" and its performance data (precision and method comparison) reflects the device operating independently to produce results. There is no human-in-the-loop component described for its basic operation or for these performance studies, other than potentially loading samples and controls. The method comparison directly compares the ACL TOP's measurements to those of the predicate device (ACL Advance), with both devices acting in a standalone capacity.
7. The type of ground truth used (expert concensus, pathology, outcomes data, etc)
The "ground truth" for the method comparison study is the analytical results generated by the predicate device, ACL Advance. For precision data, the ground truth is the inherent variability of the controls and samples, which is quantified by statistical measures like %CV. This is an analytical rather than a clinical ground truth.
8. The sample size for the training set
Not applicable. The ACL TOP is an automated analyzer, not an AI/ML device that requires a "training set" in the conventional sense of machine learning. Its operation is based on established analytical principles for coagulation testing, calibrated using standard laboratory calibration materials, not trained on a dataset.
9. How the ground truth for the training set was established
Not applicable, as there is no "training set" for an AI/ML algorithm. Calibration and quality control for such instruments typically involve using reference materials with known concentrations or activities, established by manufacturers or regulatory bodies, to ensure accurate measurement.
§ 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.