Search Results

Lumipulse G CA15-3 is a Chemiluminescent Enzyme Immunoassay (CLEIA) for the quantitative determination of CA 15-3 in human serum or plasma (sodium heparin, lithium heparin, or dipotassium EDTA) on the LUMPULSE G System.

The assay is to be used as an aid in the management of patients previously diagnosed with stage II and III breast cancer. Serial testing for patient CA15-3 assay values should be used in conjunction with other clinical methods used for monitoring breast cancer.

WARNING: The concentration of CA 15-3 in a given specimen, as determined by assays from different manufacturers, can vary due to differences in assay methods and reagent specificity. The results reported by the laboratory to the physician must include the identity of the assay for CA 15-3 used. Values obtained with different assay methods cannot be used interchangeably. If, in the course of monitoring a patient, the assay method used of CA 15-3 is changed, the laboratory must perform additional serial testing to confirm baseline values. Prior to changing assays, the laboratory MUST confirm baseline values for patients being serially monitored. Lumipulse G CA 15-3 should not be used for cancer screening or diagnosis.

Lumipulse G CA15-3 is an assay system, including a set of immunoassay reagents, for the quantitative measurement of CA 15-3 in specimens based on CLEIA technology by a two-step sandwich immunoassay method on the LUMIPULSE G System.

Lumipulse G CA15-3 Immunoreaction Cartridges: REF 235102 The Lumipulse G CA15-3 Immunoreaction Cartridges consists of 3 x 14 tests. Each kit contains the following:

1.) Antibody-Coated Particle Solution (Liquid when used, 250 µL/Immunoreaction Cartridge) Contains 150 µg/mL anti-CA 15-3 monoclonal antibody (mouse)-coated particles, protein stabilizers (bovine and mouse) and chemical stabilizers in 0.15 M sodium chloride/Tris buffer. This solution contains gelatin and turns into gel at 15°C or lower. Preservative: sodium azide.

2.) Enzyme-Labeled Antibody Solution (Liquid, 350 µL/Immunoreaction Cartridge) Contains 0.2 µg/mL alkaline phosphatase (ALP: calf)-labeled anti-CA 15-3 monoclonal antibody (mouse), protein stabilizers (bovine and calf) and chemical stabilizers in 0.1 M sodium chloride/MES buffer. Preservative: sodium azide.

Here's an analysis of the acceptance criteria and study detailed in the provided document, restructured to answer your specific questions:

1. A table of acceptance criteria and the reported device performance

The document primarily focuses on validating the Lumipulse G CA15-3 assay as "substantially equivalent" to a predicate device (ARCHITECT CA 15-3) and demonstrating its analytical and clinical performance. The acceptance criteria are largely implicit, based on meeting established guidelines (CLSI) and demonstrating satisfactory performance within defined ranges or against a predicate.

2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)

• Precision (ARUP 20-day): n=80 for each of 6 panels.
• Precision (Lot-to-Lot Reproducibility): n=90 for each of 6 samples (combined for Lots A, B, and C).
• Precision (Site-to-Site Reproducibility): n=102 for each of 6 panels (for Lot A).
• Linearity: Patient serum samples containing naturally expressed CA15-3. (Specific number not provided, but implies multiple dilutions for a range of samples).
• Detection Limit: Not explicitly detailed on sample size but derived from technical measurements.
• Interfering Substances: Human serum specimen pools with CA 15-3 concentrations ranging from approximately 30-35 U/mL, 80-100 U/mL, and 300-350 U/mL. (Number of pools or individual samples not provided, but implies multiple concentrations for each interferent).
• Method Comparison: n=117 samples.
• Matrix Comparison: Not explicitly detailed sample size, but involved various tube types (SST, K2EDTA, Lithium Heparin, and Sodium Heparin) versus control samples (Red top serum).
• Monitoring of Disease State in Patients Diagnosed with Breast Cancer:
  • Patients: 112 patients previously diagnosed with stage II and III breast cancer.
  • Observations: 566 pairs of observations, with an average of 6.1 observations per patient.
  • Data Provenance: Not explicitly stated as retrospective or prospective, nor country of origin, but described as patients "previously diagnosed with stage II and III breast cancer" and "monitoring recurrence or progressive disease," suggesting real-world patient data likely collected over time.
• Expected values/Reference range (Apparently Healthy): N=356 (120 males, 236 females).
• Expected values/Reference range (Apparently Healthy and Benign Subjects): N=591 (Includes the 356 healthy, plus additional benign groups: Benign Breast N=75, Benign Ovarian N=40, Urogenital N=40, Pregnant N=40, Hypertension N=40).
• Expected values/Reference range (Subjects with Cancer): N=368 (Treatment Naïve Breast N=130, Uterine/Endometrial N=40, Ovarian N=40, Lung N=40, Colorectal N=38, Pancreatic N=40, Liver N=40).

The provenance details (country, retrospective/prospective) are not explicitly mentioned for most studies, although the patient cohorts for disease monitoring and reference ranges imply real clinical data collection.

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)

This device is an in vitro diagnostic (IVD) assay for a biomarker (CA15-3).

• For the analytical performance studies (precision, linearity, detection limits, interference, cross-reactivity), the "ground truth" is established through highly controlled laboratory procedures, reference materials, and defined statistical methods (e.g., CLSI guidelines). There is no "expert" per se establishing ground truth in the way a radiologist does for an imaging study. The accuracy is relative to reference measurements or established methods.
• For the method comparison study (against ARCHITECT CA 15-3), the predicate device serves as the comparative reference, which itself would have been validated.
• For the clinical performance study (Monitoring of Disease State), the ground truth for "disease progression" or "no progression" was not established by expert consensus of imaging or pathology specific to this study. Instead, disease status changes were assessed by "other clinical methods used for monitoring breast cancer" which likely includes physician assessment, imaging, and possibly pathology reports. The study assessed changes in CA15-3 levels against these "changes in disease status," which is the outcome used as the ground truth. No specific number or qualification of experts defining the clinical ground truth within this study is provided. However, the initial "staging of the patients was done according to AJCC 7th edition," which implies expert clinical assessment in their initial diagnosis.

4. Adjudication method (e.g. 2+1, 3+1, none) for the test set

For an IVD assay measuring a biomarker, adjudication methods like 2+1 or 3+1 (common in image-based AI studies where human readers interpret images) are generally not applicable. The "ground truth" for quantitative assays is based on a reference standard or, in clinical studies, the established clinical diagnosis/status.

In the clinical monitoring study, the "change in disease status" serves as the clinical outcome (ground truth), but the method for determining this status retrospectively or prospectively, and if it involved adjudication, is not described.

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, a multi-reader multi-case (MRMC) comparative effectiveness study was not done. This is an in vitro diagnostic (IVD) assay for a blood biomarker (CA15-3), not an AI-assisted diagnostic imaging device meant to assist human readers. Therefore, the concept of human readers improving with AI assistance is not applicable here.

6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done

Yes, the analytical and clinical performance studies are standalone performance evaluations of the Lumipulse G CA15-3 assay. The device provides a quantitative measurement of CA15-3 in a sample. The performance characteristics (precision, linearity, detection limit, interference, method comparison) and the clinical utility for monitoring disease progression are all evaluated based on the assay's output directly, without a human in the loop interpreting the assay's numerical result in a subjective way. The clinical monitoring study, for instance, directly correlates changes in the assay's output with disease progression, demonstrating its standalone utility for this purpose.

7. The type of ground truth used (expert consensus, pathology, outcomes data, etc)

• Analytical Studies: For analytical performance characteristics (precision, linearity, detection limit, interference, cross-reactivity), the ground truth is established by carefully prepared reference materials, spiked samples, and highly controlled laboratory procedures consistent with CLSI guidelines, aiming for accurate measurement against these known values.
• Method Comparison: The ground truth is the measurement obtained by the legally marketed predicate device (ARCHITECT CA 15-3).
• Clinical Monitoring Study: The ground truth for disease progression/no progression was based on outcomes data defined as "changes in disease status" over time, using "other clinical methods used for monitoring breast cancer." This likely refers to a combination of clinical assessments, imaging, and possibly pathology results, but no specific details on the exact methods or their adjudication are provided within the document.

8. The sample size for the training set

The terms "training set" and "test set" are typically used for machine learning or AI models. This document describes the validation of an IVD assay, not an AI system. Therefore, there isn't a traditional "training set" in the machine learning sense. The bulk of the studies (analytical and clinical performance) serve as the validation/testing to demonstrate the device's performance characteristics. If we loosely interpret "training set" as data used during the development phase to optimize the assay before formal validation, this information is not provided in the 510(k) summary, as it focuses on the performance of the finalized device.

9. How the ground truth for the training set was established

As explained above, there isn't a "training set" in the context of an AI model for this IVD device. The assay development would involve rigorous chemical and biological optimization, and calibration against reference materials, but this specific "ground truth establishment" for a training set is not applicable or detailed here.

Ask a specific question about this device