LogoFDA 510(k) Search
Search Filters

Search Results

Found 1 results

510(k) Data Aggregation

    K Number
    K243348

    Validate with FDA (Live)

    Device Name
    Athelas Home
    Manufacturer
    Athelas, Inc.
    Date Cleared
    2026-02-06

    (466 days)

    Product Code
    GKZ
    Regulation Number
    864.5220
    Type
    Traditional
    Panel
    Hematology
    Age Range
    21 - 120
    Reference & Predicate Devices
    K200828,K181288
    Predicate For
    N/A
    AI/MLSaMDIVD (In Vitro Diagnostic)TherapeuticPediatricDiagnosticis PCCP AuthorizedThirdpartyExpeditedreview
    Intended Use

    The Athelas Home is indicated for the quantitative determination of white blood cells (WBC) and Neutrophil percentages (NEUT%) in capillary whole blood from fingerstick for patient self-testing with results viewable by healthcare professionals, and in capillary or K2EDTA venous whole blood for multiple-patient use in point-of-care settings. For self-testing, the Athelas Home is intended to be used by a single person and should not be shared. The Athelas Home is only to be used with Athelas Test Strips. The Athelas Home is intended for adult patients (aged 21 and older) at risk of neutropenia. For self-testing patients with psychiatric conditions, clinical judgment should be exercised when deciding the end-user and based on the instructions for use (IFU), the treating physician should determine which patients are competent to perform the test by themselves. Results obtained with the Athelas Home should not be the sole basis for patient diagnosis, treatment, or management of leukopenia and neutropenia, and all results should be evaluated by a healthcare provider. Prescription Use Only.

    Device Description

    The Athelas Home is identical to the Athelas Home (K200828) and Athelas One (K181288) in its technology, core software, the analytes measured, all components, and usage process.

    The Athelas Home is an automated diagnostic device intended to perform tests on whole blood samples collected in K2EDTA or capillary finger stick samples collected directly into the Athelas test strip. Athelas Home returns quantitative WBC and Neut% metrics from the blood sample. The Athelas Home system is for In Vitro Diagnostic use and for prescription use only.

    The Athelas Home is only to be used with Athelas Test Strips. The Athelas Home is indicated for use by patients and caregivers within home settings and at Point-of-Care settings with results viewable by health care professionals. The Athelas Home is indicated for adult populations only (aged 21 and older).

    The Athelas Home is not indicated to completely replace a patient's monitoring regimen on reference lab systems. Prescribers should continue monitoring patients using traditional lab-based means as deemed necessary and possible.

    The Athelas Home is not indicated to directly diagnose or assist in disease management. Rather, the Athelas Home should be used to quantify WBC and Neut% only, with all decision making owned by the health care professional as per standard of care.

    The Athelas Home uses image processing and microfluidics in order to measure WBC and NEUT% values from whole blood in point-of-care and home settings. A drop of blood is placed on the Athelas test strip either directly from the finger or via pipette from K2EDTA whole blood tube. The device uses single-use, disposable test strips, ensuring there is no direct patient contact or contamination with the device itself.

    The Athelas Home test strip stains and creates a monolayer of the blood sample within the chamber. The strip is inserted into the test strip slot of the Athelas Home device where a built-in camera captures images of the sample. The images are automatically transmitted to the cloud server where they are processed and analyzed by an image processing algorithm. The quantitative results of WBC, NEUT% results are based on the concentrations and types of cells present in the sample.

    AI/ML Overview

    Here's a breakdown of the acceptance criteria and study information for the Athelas Home device, based on the provided FDA 510(k) clearance letter:

    1. Acceptance Criteria and Reported Device Performance

    Acceptance CriteriaReported Device Performance (WBC)Reported Device Performance (Neut%)
    WBC Precision:
    - 7.5% CV above 2K/µL WBCWithin-run Precision/Reproducibility: 2.20 K/µL (5.87% CV), 3.75 K/µL (5.46% CV), 4.12 K/µL (5.09% CV), 5.11 K/µL (5.96% CV), 7.89 K/µL (4.82% CV), 10.01 K/µL (5.45% CV), 14.64 K/µL (4.69% CV), 17.52 K/µL (4.34% CV), 23.33 K/µL (5.81% CV). All met.
    - 0.25 K/µL SD below 2K/µL WBCNot explicitly detailed for each sample below 2K/µL, but overall "All results met the predefined specifications (CV%) for precision." "Overall reproducibility levels were found to meet the acceptance criteria of 7.5% CV for WBC." (Implicitly including SD for lower range). Low-level QC (2.746 K/µL) showed a SD of 0.153 K/µL.
    WBC Bias/Error:
    - ±7.5% error above 2K/µL WBCMethod Comparison: -2.31% Mean % Bias (1.1 - 23 K/µL range). Satisfied.
    - ±0.25 K/µL error below 2K/µL WBCNot explicitly detailed for each sample below 2K/µL.
    Neutrophil % Precision:
    - 5% SD OR 15% CVWithin-run Precision/Reproducibility: Not explicitly detailed for each sample, but overall "All results met the predefined specifications...for precision." (Implicitly including Neut%).
    Reproducibility (Overall Summary Table - Low): 2.680 SD (5.278% CV) Within-run. Total SD 3.443 (6.780% CV). Satisfied.
    Reproducibility (Overall Summary Table - Medium): 3.015 SD (6.030% CV) Within-run. Total SD 3.344 (6.689% CV). Satisfied.
    Reproducibility (Overall Summary Table - High): 2.967 SD (5.838% CV) Within-run. Total SD 3.316 (6.525% CV). Satisfied.
    Neutrophil % Bias:
    - ±10% bias or ±5% Neut% total error (whichever larger)Method Comparison: 1.18% Mean % Bias (8 - 92.89% range). Satisfied.
    At-Home Precision (WBC)At-Home Precision Repeatability: 5.4% CV (combined patient blood and QC material). Pass (Evaluation Criteria: 7.5% CV).
    At-Home Precision (Neut%)At-Home Precision Repeatability: 1.2% SD (combined patient blood and QC material). Pass (Evaluation Criteria: 5% SD or 15% CV).
    Linearity (WBC)R² = 0.997, Slope = 1.013, Intercept = 0.0449, CVr = 5.08% as per OLS regression. Demonstrated to be linear from lower limit to upper limit and within measured allowable max % diff for each interval.
    Flagging Accuracy% Positive Agreement (Sensitivity): 90.91%. % Negative Agreement (Specificity): 96.71%. % Overall Agreement: 94.87%. Met the specification of >= 90%.
    Matrix Comparability (WBC - venous vs capillary)WBC: Slope 1.026 (1.000, 1.055), Intercept -0.145 (-0.319, 0.018), Mean Bias 0.056 K/µL, Mean % Bias -0.588. Met all evaluation criteria.
    Matrix Comparability (Neut% - venous vs capillary)Neut%: Slope 0.999 (0.938, 1.058), Intercept 0.457 (-2.502, 3.614), Mean Bias 0.162 Percentage Points, Mean % Bias -0.333. Met all evaluation criteria.
    At-Home Method Comparison (WBC - Athelas One vs Athelas Home)WBC: Slope 1.0 (0.95, 1.03), Intercept 0 (-0.16, 0.31), Mean Bias 0, Mean % Bias 0.0. Met all evaluation criteria.
    At-Home Method Comparison (Neut% - Athelas One vs Athelas Home)Neut%: Slope 1.0 (0.98, 1.09), Intercept -1.0 (-6.09, 0.0), Mean Bias -0.5, Mean % Bias 1.3. Met all evaluation criteria.
    At-Home Error RateOver 94% of tests run successfully with no errors. Errors that were found were resolved after running a single additional test-strip.
    At-Home UsabilityLarge majority of patients found the Athelas Home easy to use and had very little difficulty running the test end to end. (e.g., Lancet use: 32/32 Yes; Test-strip fill: 30/32 Yes; No errors (non-code): 31/32 Yes; App use: 32/32 Yes).
    Fill-Volume TestingRobust to variable volumes of blood. Errors returned for 0µL. Accurate results within bias criteria for 1.4µL, 3.4µL, and 5.4µL fill volumes across various sample concentrations.Robust to variable volumes of blood. Errors returned for 0µL. Accurate results within bias criteria for 1.4µL, 3.4µL, and 5.4µL fill volumes across various sample concentrations.
    Device Tilt TestingNot affected by tilted placement. No significant bias across tilt orientations and samples. All WBC criteria met.Not affected by tilted placement. No significant bias across tilt orientations and samples. All Neut% criteria met.
    Device Drop TestingRobust to drop and vibration schedules. All 12 devices passed functional tests after testing, with no critical physical damage or functional defects, and WBC results remained within reference range.Robust to drop and vibration schedules. All 12 devices passed functional tests after testing, with no critical physical damage or functional defects, and Neut% results remained within reference range.
    Device/Test-Strip Temperature TestingBoth device and test-strip continued to meet performance specifications after exposure to extreme temperatures. All WBC bias criteria met.Both device and test-strip continued to meet performance specifications after exposure to extreme temperatures. All Neut% bias criteria met.
    Humidity Flex Testing (Test-Strips)Test-strips continued to function within specifications after exposure to extreme humidity conditions. No evidence of condensation, fill-time within criteria (< 25 seconds), and no color change present for all 8 test strips across schedules.
    Vibration TestingSystem successfully met evaluation criteria post vibration test.
    Cleaning and DisinfectionNo change in performance or external materials after 10,000 cleaning and disinfection cycles using Caviwipe.

    2. Sample Sizes Used for the Test Set and Data Provenance

    • Bench Testing (Within-run Precision):
      • Sample Size: 9 whole blood samples. 10 replicates per sample, per operator, per test strip lot. Total of 90 tests per sample (90 tests x 9 samples = 810 tests).
      • Data Provenance: Not specified, but generally from a laboratory setting.
    • Bench Testing (Between-run Reproducibility):
      • Sample Size: Three levels of ATH-check quality control material (Low, Normal, High). 240 readings per level (2 runs x 20 days x 2 replicates x 3 sites). Total of 720 readings across all levels.
      • Data Provenance: Not specified, but implied to be from laboratory/controlled settings across three sites.
    • At-Home Precision Study (Patient In-home Whole Blood):
      • Sample Size: 24 patients, 2 replicates per patient. Total of 48 individual tests.
      • Data Provenance: From at-home users in home environments in the US (implied by "at-home settings in the US" in subsequent sections and the study name).
    • At-Home Precision Study (Patient In-home QC Material):
      • Sample Size: 10 patients, 3 replicates per patient. Total of 30 individual tests.
      • Data Provenance: From at-home users in home environments in the US.
    • Linearity:
      • Sample Size: 10 samples run in 4 replicates on 4 devices. Total of 160 tests.
      • Data Provenance: Samples obtained by pooling "one low WBC concentration fresh whole blood sample one high WBC concentration sample." Generally laboratory prepared samples.
    • Interfering Substances:
      • Sample Size: Not explicitly stated how many individual samples were spiked and tested for each substance, but a list of 16 substances were tested.
      • Data Provenance: Prepared whole blood samples, some naturally occurring, some spiked.
    • Reference Intervals:
      • Sample Size: 120 healthy donors.
      • Data Provenance: Not specified, implied to be based on healthy donors.
    • Limit of Blank (LoB):
      • Sample Size: 120 total repeated measurements of blank samples.
    • Limit of Detection (LoD):
      • Sample Size: 2 test strip lots (60 strips each), 5 low-level samples, 2 replicates per sample. Total of 20 low-level tests for LoD calculation (across each lot, 5 samples x 2 replicates).
    • Limit of Quantification (LoQ):
      • Sample Size: 2 test-strip lots, 1 instrument, 3 days, 3 replicates per sample, 4 independent low-level whole blood samples. Total of 36 low-level tests (4 samples x 3 replicates x 3 days?) + more for the two lots.
      • Data Provenance: Distinct whole blood samples acquired from patients in K2EDTA tubes, presumably from a laboratory or clinical setting. Diluted to target concentrations.
    • Specimen Stability (Whole Blood):
      • Sample Size: 9 different venous blood samples (low, normal, high WBC levels).
      • Data Provenance: Venous blood samples.
    • Specimen Stability (Test Strip):
      • Sample Size: Two test strip lots compared over 26 months, 3 replicates, multiple QC lots. Three replicates per two lots of strips across 3 levels of control fluid run every 60 days.
    • Method Comparison:
      • Sample Size: 312 patient samples.
      • Data Provenance: Taken at 3 point-of-care sites in the US.
    • Flagging Comparison:
      • Sample Size: 312 samples (same as Method Comparison).
      • Data Provenance: Same as Method Comparison (3 point-of-care sites in the US).
    • Matrix Comparison (Venous vs. Capillary):
      • Sample Size: 59 patients.
      • Data Provenance: Capillary finger-prick blood samples and K2EDTA venous whole blood samples from the same patients.
    • At-Home Method Comparison:
      • Sample Size: 165 samples.
      • Data Provenance: Taken in at-home settings in the US.
    • At-Home Error Analysis:
      • Sample Size: 81 tests (derived from percentages: 76 successful, 3+2+1+1=7 errors, 76/0.94 = 80.85; 7/0.06 = 116.6; approximation derived from given percentages. The actual number is not precisely 81 based on the sum, but implies around this amount.)
      • Data Provenance: At-home testing.
    • At-Home Usability Analysis:
      • Sample Size: 32 users.
      • Data Provenance: At-home users during at-home visits.
    • Fill-Volume Testing:
      • Sample Size: 3 samples tested across 4 fill volumes (0µL, 1.4µL, 3.4µL, 5.4µL). Number of replicates not explicitly stated, but common for such studies.
      • Data Provenance: K2EDTA whole blood samples for predicate comparison.
    • Device Tilt Testing:
      • Sample Size: 3 samples tested across 4 tilt conditions. Number of replicates not explicitly stated.
      • Data Provenance: K2EDTA whole blood samples for predicate comparison.
    • Device Drop Testing:
      • Sample Size: 12 devices.
      • Data Provenance: Device testing.
    • Device and Test-Strip Temperature Testing:
      • Sample Size: Device testing: 3 devices, 9 replicates each, across 3 environmental conditions. Test-strip testing: 3 test-strip lots, 9 replicates each, across 3 environmental conditions.
      • Data Provenance: Device and test strips exposed to controlled environmental conditions.
    • Vibration Testing:
      • Sample Size: Not specified, but refers to "The system".
      • Data Provenance: Third party ISTA certified test lab Quanta Testing.
    • Humidity Flex Testing:
      • Sample Size: 8 test-strips, subject to specified schedules.
      • Data Provenance: Test-strips exposed to controlled humidity chambers at Quanta Laboratories.

    3. Number of Experts Used to Establish Ground Truth and Qualifications

    The document does not explicitly state the number or qualifications of experts used to establish ground truth.

    However, it implicitly relies on:

    • Predicate devices: The Sysmex XE-5000 and Athelas One are frequently used as the "ground truth" or reference method for comparison. These are legally marketed, validated automated hematology analyzers.
    • Clinical Laboratory Standards Institute (CLSI) recommendations: The studies (precision, linearity, LoQ, stability) frequently refer to CLSI guidelines (e.g., CLSI EP05-A3, EP06-A, EP17-A2, EP25-A) as the basis for their methodologies and acceptance criteria, implying adherence to established best practices in laboratory medicine.
    • Medical decision levels: The testing was conducted around "medical decision levels," which are established by clinical experts.
    • Morphological flagging: This implies comparison against expert (e.g., pathologist, trained medical technologist) review of blood smears, though not directly stated.

    Without explicit mention, it's not possible to state "X number of radiologists with Y years of experience," as the device is for hematology, not radiology.

    4. Adjudication Method for the Test Set

    The document does not explicitly describe an adjudication method for the test set (e.g., 2+1, 3+1, none).

    • For method comparison studies, the predicate device (Sysmex XE-5000 or Athelas One) serves as the reference, which performs automated analysis. The "ground truth" is effectively the results from this established, high-accuracy device.
    • For flagging comparison, the Sysmex XE-5000's flagging is used as the reference, which would typically be based on its internal algorithms and potentially validated against expert manual differential counts in its own clearance studies.

    5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

    No Multi-Reader Multi-Case (MRMC) comparative effectiveness study was mentioned. The studies focus on the device's analytical performance against predicate devices or its own established metrics, not on how human readers improve with or without AI assistance. The Athelas Home is an automated system providing quantitative results, not an AI-assisted diagnostic aid for human interpretation.

    6. Standalone (Algorithm Only) Performance

    Yes, the studies primarily describe standalone performance of the algorithm.

    • The device captures images, transmits them to a cloud server, where they are "processed and analyzed by an image processing algorithm."
    • The results for WBC and NEUT% are "based on the concentrations and types of cells present in the sample" as determined by this algorithm.
    • All the performance metrics (precision, bias, linearity, LoD, LoQ, method comparison, flagging) describe the output of the device (i.e., the algorithm and hardware combined) compared to a reference method, without direct human-in-the-loop assistance for interpretation of individual results. Human healthcare professionals evaluate the overall results and provide diagnosis/treatment.

    7. Type of Ground Truth Used

    • Predicate device measurements: The primary ground truth for quantitative measurements (WBC, NEUT%) was the Sysmex XE-5000 analyzer (a widely accepted automated hematology analyzer) and the Athelas One (the previous generation predicate Athelas device).
    • Quality Control (QC) material values: For precision and linearity studies, established values of QC materials (ATH-CHECK) were also used.
    • Blank samples: For Limit of Blank.
    • Diluted patient samples: For Limit of Quantification.
    • Expert consensus/pathology (implied): For flagging comparison and verification of morphological flags, the reference Sysmex XE-5000's flagging implies an underlying ground truth (likely derived from expert manual differentials or pathology review during its own validation), which the Athelas One then compares to. This is not directly stated for Athelas Home validation, however.
    • Pre-established reference intervals: For verifying reference intervals against Sysmex XE-5000.
    • Outcomes data (No): Not mentioned as a ground truth.

    8. Sample Size for the Training Set

    The document does not specify the sample size for the training set. It focuses entirely on the validation/test set performance. It mentions that the "images are automatically transmitted to the cloud server where they are processed and analyzed by an image processing algorithm," implying a machine learning component, but provides no details on its development or training data.

    9. How the Ground Truth for the Training Set Was Established

    As the training set sample size is not specified, neither is the method for establishing its ground truth. However, given the nature of the device (automated cell counter), it would typically involve:

    • Expert manual differential counts: Morphological classification and counting of various WBC types on blood smears by highly trained medical technologists or pathologists.
    • Reference laboratory analyzer results: Pairing manual counts with results from established automated hematology analyzers for quantitative parameters.
    • Annotated images: Images used for training would require meticulous annotation by experts to identify and classify different cell types.
    Ask a Question

    Ask a specific question about this device

    Page 1 of 1