Search Results

The Eonis™ SCID-SMA kit is intended for the qualitative detection of the SMN1 gene exon 7 as an aid in screening newborns for Spinal Muscular Atrophy (SMA). The test is intended for DNA from blood specimens dried on a filter paper and for use on the QuantStudio™ Dx Real-Time PCR instrument.

This test is only intended for use for screening of SMA that bear the homozygous deletion of SMN1 exon 7.

This test is not intended for use as a diagnostic test and a positive screening result should be followed by confirmatory testing.

The Eonis SCID-SMA kit contains reagents to detect three biomarkers: TREC, KREC and exon 7 in the SMN1 gene. Detection of TREC and KREC was cleared in K203035.

The newborn screening workflow for the Eonis SCID-SMA kit includes:

• Two liquid handling platforms (one for DNA extraction and one for PCR master mix . setup)
• QuantStudio Dx Real-Time PCR instrument .
• . Eonis Analysis Software

Each Eonis SCID-SMA kit contains reagents for up to 384 reactions or 1152 reactions including kit controls. The kit contents are listed in Table 1. Materials required but not provided include the Eonis DNA Extraction Kit, Eonis Analysis Software and consumables (Table 2).

Here's a breakdown of the acceptance criteria and the study proving the device meets them, based on the provided text:

Acceptance Criteria and Device Performance

Study Details

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

• Precision Study (Test Set 1):
  • Sample Size: 13 representative DBS samples (SMA positive, carrier, and normal), tested in 108 replicates (106 for some) per sample over 54 runs. Total measurements: 13 samples * 108 measurements = 1404 measurements.
  • Data Provenance: Analytical performance studies conducted using contrived samples (cord blood or adult whole blood with hematocrit adjusted to neonate levels). SMA positive sample created by spiking SMN1 negative Coriell cells into leukocyte-depleted blood.
• Reproducibility Study (Test Set 2):
  • Sample Size: 13 samples (same as precision study), tested in 150 replicates per sample across 3 study sites over 5 operating days. Total measurements: 13 samples * 150 measurements = 1950 measurements.
  • Data Provenance: Contrived samples (cord blood or adult whole blood with hematocrit adjusted to neonate levels).
• Filter Paper Reproducibility Study (Test Set 3):
  • Sample Size: 6 samples (from precision study set) prepared on 3 lots of 2 brands of filter paper each (total 36 conditions). 5 replicates per condition. Total 900 results.
  • Data Provenance: Contrived samples.
• Limit of Blank Study (Test Set 4):
  • Sample Size: 150 replicates of contrived analyte-negative samples per kit lot (total 300 replicates across 2 kit lots).
  • Data Provenance: Contrived samples (SMN1-negative cells from Coriell Institute into leukocyte-depleted human blood).
• Interference Study (Test Set 5):
  • Sample Size: 7 interfering substances, 2 interferent levels, 3 target DNA levels, 13 replicates per level. Total 544 sample results.
  • Data Provenance: Contrived samples (SMN1 presumptive normal).
• qPCR Method Equivalency Study (Test Set 6):
  • Sample Size: 13 samples (from precision study set), 5 replicates per sample, test for 2 PCR methods. Total 1560 results.
  • Data Provenance: Contrived samples.
• DNA Extraction Equivalency Study (Test Set 7):
  • Sample Size: 7 samples (from precision study set), 5 replicates per sample, test for 3 extraction/PCR methods. Total 1050 results.
  • Data Provenance: Contrived samples.
• Clinical Screening Study (Test Set 8):
  • Sample Size: 3069 DBS specimens. This included 51 retrospective archived DBS specimens from subjects confirmed positive for SMA and 3018 routine newborn screening specimens.
  • Data Provenance: Retrospective archived DBS specimens from the US and Denmark. Routine newborn screening specimens obtained from the Danish Newborn Screening Biobank (NBS-Biobank).
• Accuracy Study (Test Set 9):
  • Sample Size: 51 confirmed positive SMA samples and 55 presumed negative DBS samples. Total 106 samples.
  • Data Provenance: Confirmed positive SMA samples (molecular genetic testing result showing homozygous deletion of SMN1 exon 7) and presumed negative DBS samples matched by storage time.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications

The document mentions that the clinical status of the routine subjects in the Clinical Screening Study was determined through a "retrospective review by clinical experts." However, it does not specify the number of experts or their specific qualifications (e.g., "radiologist with 10 years of experience"). For the confirmed SMA cases, "confirmatory test results" (molecular genetic testing) were used as the comparator, which is a definitive method rather than expert consensus on imaging.

4. Adjudication Method for the Test Set

The document does not describe any specific adjudication method (like 2+1 or 3+1) for the interpretation of results in the test sets. For the Eonis SCID-SMA kit, the interpretation of results appears to be largely automated by the Eonis Analysis Software based on pre-set Ct cut-off values. For the clinical screening study, samples with values above the cut-off were re-tested in duplicate to obtain the final result.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done

No, an MRMC comparative effectiveness study was not performed. This device is a quantitative PCR-based assay with automated interpretation software, not an imaging-based AI system that assists 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 performance of the Eonis SCID-SMA Kit is essentially standalone. The Eonis Analysis Software "automatically flags quality control (QC) violations and interprets results according to the cut-offs," presenting results as "Presumptive Positive" or "Presumptive Normal." While human operators perform the lab procedures (DNA extraction, PCR setup), the final interpretation of the test result itself is automated by the algorithm based on the measured Ct values against a pre-set cut-off.

7. The Type of Ground Truth Used

• Analytical Studies (Precision, Reproducibility, Limit of Blank, Interference, Method/Extraction Equivalency, Carry-over): Ground truth was based on the contrived nature of the samples. For example, SMA positive samples were created by spiking specific cells, and analyte-negative samples were prepared to contain no target analyte.
• Clinical Screening Study:
  • For SMA positive cases (51 samples): Ground truth was established by "confirmatory test results" (molecular genetic testing showing homozygous deletion of SMN1 exon 7).
  • For routine newborn screening specimens (3018 samples): Ground truth was established by "retrospective review by clinical experts to confirm the routine subject cohort samples were from unaffected individuals." This suggests a form of clinical outcome/diagnosis as ground truth, likely based on further clinical evaluations, not just genetic testing for SMN1 deletion.
• Accuracy Study:
  • For confirmed SMA samples (51 samples): Ground truth was "molecular genetic testing result showing homogenous deletion of SMN1 gene exon 7."
  • For presumed negative samples (55 samples): Ground truth was confirmed by "molecular genetic testing for SMN1" using a CE-IVD labeled assay. This is molecular genetic testing/pathology ground truth.

8. The Sample Size for the Training Set

The document does not explicitly state a separate "training set" for the Eonis SCID-SMA Kit. As a PCR-based assay, its "learning" primarily involves setting the appropriate Ct cut-off value (31.24). This cut-off is pre-set in the Eonis Analysis Software. The document does not describe how this specific cut-off was initially determined (e.g., through a separate study for calibration or training). The studies described here are verification and validation studies to demonstrate the performance with that pre-set cut-off.

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

Since a distinct "training set" is not explicitly mentioned for algorithmic development in a machine learning sense, the establishment of ground truth for training is not detailed. The inherent "training" of such a system would involve optimizing the Ct cut-offs based on a set of known positive and negative samples to achieve desired diagnostic sensitivity and specificity. However, the provided text focuses on the validation of the device's performance given its pre-determined operational parameters (like the 31.24 Ct cut-off).

Ask a specific question about this device

The Eonis™ SCID-SMA kit is intended for the semi-quantitative determination of TREC (T-cell receptor excision circle) as an aid in screening newborns for Severe Combined Immunodeficiency (SCID) and for the semi-quantitative determination of KREC (Kappa-deleting recombination excision circle) as an aid in screening newborns for X-linked agammaglobulinemia (XLA). The test is intended for DNA from blood specimens dried on a filter paper and for use on the QuantStudio™ Dx Real-Time PCR instrument.

This test is not intended for screening of SCID-like Syndromes, such as DiGeorge Syndrome, or Omenn Syndrome. lt is also not intended to screen for less acute SCID syndromes such as leaky-SCID or variant SCID. The test is not indicated for screening B-cell deficiency disorders other than XLA, such as atypical XLA, or for screening of XLA carriers.

This test is not intended for use as a diagnostic test and a positive screening result should be followed by confirmatory testing.

The Eonis SCID-SMA kit is a multiplex real-time PCR-based assay. It uses target sequence-specific primers and TaqMan™ probes to amplify and detect three targets: TREC, and RPP30, in the DNA extracted from newborn dried blood spot (DBS) using Eonis DNA Extraction kit in a single PCR reaction.

Each Eonis SCID-SMA kit contains reagents for up to 384 reactions (for 3241-001U) or 1152 reactions (for 3242-001U) including kit controls.

The document describes the Eonis SCID-SMA kit, a real-time PCR-based assay for newborn screening of Severe Combined Immunodeficiency (SCID) and X-linked agammaglobulinemia (XLA). The study provided demonstrates the device's analytical and screening performance to support its substantial equivalency to a predicate device.

Here's a breakdown of the requested information:

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly state "acceptance criteria" as a separate table. However, it presents Sensitivity and Specificity for both TREC and KREC analytes, which serve as key performance metrics. These values are compared to the predicate device.

Reported Device Performance of Eonis SCID-SMA Kit:

Comparison to Predicate Device (PerkinElmer EnLite Neonatal TREC Kit) for TREC:

The reported performance clearly aims to meet or exceed the performance of the predicate device, demonstrating 100% sensitivity and high specificity for both analytes.

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

• Sample Size for Screening Performance Study (Test Set):
  • Total DBS specimens: 3090
  • Confirmed SCID positive: 17
  • Confirmed XLA positive: 6
  • Normal newborn screening specimens: 3018 (retrospective archived)
• Data Provenance: Retrospective archived dried blood spot specimens.
  • Country of Origin: US and Denmark.
  • Study conducted in Denmark.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts

• Number of Experts: Not explicitly stated as a specific number. The document mentions "clinical experts" were used.
• Qualifications of Experts: The document states "clinical experts" retrospectively reviewed the clinical status of routine subjects to confirm they were from unaffected individuals. Further specific qualifications (e.g., specific medical specialty, years of experience) are not provided in this document.

4. Adjudication Method for the Test Set

• Adjudication Method: The document describes a retesting protocol for initial "screen positive" results.
  • "The specimens having TREC and KREC levels below the cut-off values in the initial round of testing were re-tested in duplicate."
  • "The final results (presumptive positive, invalid result) were classified after the second round of testing."
  • This implies a form of internal re-adjudication based on duplicate retesting for samples near the cut-off. There is no mention of external expert adjudication for discordant results or a specific "X+Y" type of adjudication.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was done

• MRMC Study: No, a Multi-Reader Multi-Case (MRMC) comparative effectiveness study was not done. This device is an in-vitro diagnostic (IVD) kit for semi-quantitative determination of biomarkers, not an AI assisting human readers of medical images. Therefore, the concept of human readers improving with AI assistance is not applicable to this type of device.

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

• Standalone Performance: Yes, the entire performance data regarding sensitivity, specificity, reproducibility, precision, limit of detection, and linearity are based on the standalone performance of the Eonis SCID-SMA kit (the algorithm of the kit combined with the instrument) on dried blood spot samples. This device operates as an automated assay, therefore, its performance is inherently "standalone" in terms of its analytic and clinical validity.

7. The Type of Ground Truth Used

• Ground Truth Type:
  • Confirmatory testing: For SCID and XLA positive cases, "Confirmatory test results were used as the comparator." This implies clinical diagnosis or gold standard laboratory tests.
  • Clinical expert retrospective review: For normal newborn screening specimens, "The clinical status of the routine subjects was determined through a retrospective review by clinical experts to confirm the routine subject cohort samples were from unaffected individuals." This indicates clinical outcomes or medical records adjudicated by experts.

8. The Sample Size for the Training Set

• Training Set Sample Size: The document does not explicitly state the sample size of a separate "training set" for the assay. The study described is a clinical validation (test set). For assay development (which would include "training" for establishing parameters like cut-offs), the document mentions:
  • Cut-off values were established using "an independent dataset." The size of this independent dataset is not specified.
  • Reproducibility and precision studies used panels of dried blood spots at different TREC/KREC levels, but these are for analytical validation rather than establishing classification criteria.

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

• Training Set Ground Truth Establishment: As no specific "training set" is detailed, the method for establishing ground truth for any data used during the assay's development or cut-off determination (the "independent dataset" mentioned for cut-off study) is not explicitly described. However, it's reasonable to infer that similar methods to the test set ground truth (confirmatory testing for affected individuals and clinical review for unaffected individuals) would have been applied during the development phase.

Ask a specific question about this device

The NeoBase™ 2 Non-derivatized MSMS kit is intended for the measurement and evaluation of amino acid, succinylacetone, free carnitine, acylcarnitine, nucleoside and lysophospholipid concentrations (Table 1) with a tandem mass spectrometer from newborn heel prick blood specimens dried on filter paper. Quantitative analytes and their relationship with each other is intended to provide analyte concentration profiles that may aid in screening newborns for metabolic disorders.

Each NeoBase 2 Non-derivatized MSMS kit contains reagents for 960 assays. The kit is designed to be used with NeoBase 2 Non-derivatized Assay Solutions consisting of Neo MSMS Flow Solvent and NeoBase 2 Extraction Solution and NeoBase 2 Succinylacetone Assay Solution.

• NeoBase 2 Internal Standards - 1 vial
• NeoBase 2 Controls Low, High - 3 filter paper cassettes (Whatman, no. 903) containing 3 spots of each level per cassette
• Microplate, U-bottomed - 20 plates
• Adhesive microplate covers - 20 sheets
• Barcode labels for the plates - 30 pcs (10 different barcodes, 3 pcs of each)
• Lot-specific quality control certificate
  This kit contains components manufactured from human blood. The source materials have been tested by FDA-approved methods for hepatitis B surface antigen, anti-hepatitis C and anti-HIV 1 and 2 antibodies and found to be negative.
  Instruments used:
• . QSight® 210 MD Screening System is comprised of:
  • QSight® 210 MD Mass Spectrometer
  • QSight® HC Autosampler MD
  • QSight® Binary Pump MD ●
  • Simplicity™ 3Q MD Software
• PerkinElmer MSMS Workstation software ●

Here's a breakdown of the acceptance criteria and the study that proves the device meets them, based on the provided text.

Note: This document describes a medical device for newborn screening. Performance is demonstrated through analytical studies (precision, sensitivity, linearity, interference) and screening performance data comparing it to a predicate device, rather than human-in-the-loop studies common for AI/imaging devices. Thus, several sections typically relevant to AI-based devices (e.g., number of experts, adjudication methods, MRMC studies) are not applicable here.

Acceptance Criteria and Device Performance

The acceptance criteria for the NeoBase 2 Non-derivatized MSMS kit are implied by the comprehensive analytical and screening performance evaluation studies conducted. The goal is to demonstrate that the device performs equivalently to the legally marketed predicate device (NeoBase 2 Non-derivatized MSMS kit, K173568) and provides reliable results for newborn screening.

The reported device performance below highlights key analytical and screening metrics. Specific numerical acceptance criteria are not explicitly stated as hard thresholds (e.g., "CV must be < X%"), but rather the results of the studies are presented to demonstrate acceptable performance comparable to a cleared predicate.

Table of Acceptance Criteria (Implied) and Reported Device Performance

Study Details

1. Sample Size and Data Provenance:

   • Test Set (Screening Performance): 2530 routine newborn screening specimens.
     • Country of Origin: United States (one routine screening laboratory in the US).
     • Retrospective/Prospective: The nature of "routine screening specimens" usually implies retrospective use of collected samples for evaluating a new method, but explicit detail (e.g., prospective collection specifically for this study) is not given. However, the study focuses on comparing the new device (QSight) to the predicate (TQD) on these samples, suggesting an evaluation of existing data or new data collected in a routine setting.
   • Analytical Studies (Precision, Linearity, Interference):
     • Precision: 80 determinations (40 plates x 2 replicates) for repeatability and within-laboratory; 75 determinations (15 plates x 5 replicates x 3 lots) for between-lot; 50 determinations (10 plates x 5 replicates x 2 QSights) for between-instrument.
     • Reproducibility: 75 determinations (5 plates x 5 working days x 5 replicates) across 3 labs (1 internal, 2 external).
     • Linearity, Sensitivity, Interference: Not explicitly quantified by sample size in the same way as precision. These studies typically use a controlled set of samples spiked with known concentrations or interfering substances.

2. Number of Experts and Qualifications for Ground Truth:

   • This device is a quantitative assay for measuring analyte concentrations, not an AI/imaging device requiring expert interpretation of images.
   • The "ground truth" for the screening performance study are confirmed positive specimens for various metabolic disorders (Amino Acid, Fatty Acid Oxidation, Organic Acid, ADA-SCID, X-ALD). These confirmations are typically established through gold standard clinical diagnostic methods (e.g., genetic testing, enzyme assays, or further biochemical analysis), not through expert panel review of the screening test results themselves. The document states "confirmed positive specimens," implying these are definitive diagnoses.
   • Therefore, the concept of "experts" establishing ground truth for the test set, in the way a radiologist reads an image, does not directly apply here.

3. Adjudication Method for the Test Set:

   • Not applicable as this is a quantitative diagnostic assay. "Adjudication" typically refers to resolving disagreements among human readers/interpreters, which is not the primary function of this device or study design. The study compares the new device's quantitative results and detection rates against those of a predicate device and known confirmed positive cases.

4. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:

   • Not applicable. This is not a study comparing human readers with and without AI assistance. It is a comparison of two analytical instruments/kits (the new NeoBase 2 on QSight vs. predicate NeoBase 2 on TQD).

5. Standalone Performance (Algorithm-only without human-in-the-loop):

   • Yes, this study is a standalone performance evaluation. The device (NeoBase 2 kit run on QSight MD Screening System) provides quantitative measurements of analytes. The performance tables directly show the device's analytical capabilities (precision, sensitivity, linearity, reproducibility) and its ability to detect known positive cases in a patient population without human interpretation as part of the core measurement. Human expert judgment comes in for confirming the disease in patient specimens (the ground truth), but not in interpreting the raw output of the device itself.

6. Type of Ground Truth Used:

   • Confirmed positive specimens: For the screening performance study, the ground truth was based on specimens from individuals clinically confirmed to have the targeted metabolic disorders (Amino Acid Disorders, Fatty Acid Oxidation, Organic Acid Conditions, ADA-SCID, X-ALD). This likely involved definitive diagnostic tests beyond initial screening, such as genetic testing, enzyme activity assays, or more specific quantitative metabolite analysis. This is a form of outcomes data/definitive diagnosis.

7. Sample Size for the Training Set:

   • This is a medical device for quantitative measurement, not a machine learning or AI model that requires a "training set" in the conventional sense (i.e., data used to teach an algorithm to learn patterns). The device's performance validation relies on analytical studies to demonstrate its accuracy, precision, and consistency, and then clinical performance by comparing its results to a predicate device and confirmed clinical cases. Therefore, the concept of a training set as understood in AI/ML is not relevant here.

8. How Ground Truth for the Training Set Was Established:

   • Not applicable, as there is no "training set" in the context of an AI/ML model for this type of device. The device's performance is intrinsically linked to its chemical reagents and mass spectrometry technology, which are developed and calibrated through standard analytical chemistry procedures.

Ask a specific question about this device

The GSP Neonatal Creatine Kinase-MM kit, is intended for the quantitative in vitro determination of creatine kinase MM-isoform (CK-MM) concentration in blood specimens dried on filter paper as an aid in screening newborns for Duchenne Muscular Dystrophy (DMD) using the GSP instrument.

The GSP Neonatal Creatine Kinase-MM assay is a solid phase, two-site fluoroimmunometric assay based on the direct sandwich technique and utilizes standard PerkinElmer DELFIA chemistry with the GSP instrument. The kit contains:

• . The CK-MM Calibrators (containing 0, 30, 120, 500, 2000 and 8000 ng/mL of creatine kinase) consisting of 7 cassettes each containing 1 set of dried blood spots.
• The CK-MM Controls (containing 130, 500 and 2000 ng/mL of creatine kinase) consisting of 5 cassettes each containing 2 set of dried blood spots.
• Anti-CK-MM-Eu Tracer ●
• CK-MM Assay Buffer ●
• Anti-CK-MM Microtitration strips ●
• Extra barcodes for the plates ●

Here's an analysis of the GSP Neonatal Creatine Kinase-MM kit's acceptance criteria and the study proving it meets them, based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance

The document doesn't explicitly list "acceptance criteria" for the overall device performance in a summary table. Instead, it describes analytical performance studies (precision, linearity, detection limits, specificity) with implied acceptance based on CLSI guidelines. For clinical performance, the results are presented with different cut-off values, and the benefit-risk assessment provides the overall conclusion regarding the device's benefit.

However, we can infer some key performance metrics from the clinical study results and regulatory context. The primary clinical acceptance is tied to its ability to aid in screening for DMD.

Inferred Acceptance Criteria and Reported Device Performance:

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

• Test Set Sample Size:

  • Routine Newborn Screening Samples: 3,041
  • Retrospectively Confirmed DMD Positive Samples: 30
  • Total in Clinical Study: 3,071 (for screening performance analysis using both routine and confirmed positive samples)

• Data Provenance:

  • Routine Samples: Prospective clinical study of routine newborn screening samples. Most (97.3%) were from newborns ≤ 72 hours old. Stored for (b)(4) days prior to testing.
  • DMD Confirmed Positive Samples: Retrospective. Ranged from 1 to 12 years of storage.
  • Country of Origin: Not explicitly stated but implied to be from the US, given "state U.S. laboratory" mentioned in the "Expected Values" section for a different study (though this doesn't directly refer to the provenance of the clinical study samples).

3. Number of Experts and Qualifications for Ground Truth

• Number of Experts: Not explicitly stated for establishing ground truth for the test set. However, "FDA consulted with experts in the field of DMD diagnosis" regarding the acceptance of DMD gene sequencing as a diagnostic method.
• Qualifications of Experts: "Experts in the field of DMD diagnosis." More specific qualifications (e.g., years of experience, subspecialty) are not provided.

4. Adjudication Method for the Test Set

• Initial Testing: Samples were initially tested in singlicate.
• Retest for High Concentration: If the CK-MM concentration was greater than the cut-off, new dried blood spot punches were re-tested in duplicate to confirm high concentration.
• Final Categorization for Retested Specimens: Final screening categorization was based on the mean value of the replicate retest results.
• Ground Truth Determination:
  • DMD Positive Confirmation: Next-generation sequencing of the DMD gene was used for routine samples with CK-MM concentrations above 1250 ng/mL (initially tested) and between 984 and 1210 ng/mL (initially tested).
  • Confirmed DMD Positive Samples: Clinical diagnosis of the 30 retrospective positive samples was known.
  • This is a form of hierarchical or tiered testing with a retest for "above cut-off" and then a definitive diagnostic test (NGS) for those flagging for further investigation. There is no mention of multiple human readers or a consensus process for the screening test results themselves. The NGS results served as the ground truth.

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

• No MRMC study was done. This device is an automated in vitro diagnostic (IVD) assay quantifying CK-MM levels, not an imaging or interpretive AI device where human readers interact with AI. Therefore, the concept of "human readers improve with AI vs without AI assistance" is not applicable here.

6. Standalone (Algorithm Only) Performance

• Yes, a standalone performance study was done. The entire clinical study described is a standalone performance evaluation of the GSP Neonatal Creatine Kinase-MM kit. The device operates automatically, performing quantitative analysis of CK-MM. Its output (CK-MM concentration) is then compared against established cut-offs and ground truth (NGS or clinical diagnosis) to determine screening performance metrics like false positive and false negative rates. There is no human-in-the-loop interacting with an AI algorithm for interpretation in the context of this device.

7. Type of Ground Truth Used

• Expert Consensus: Not explicitly stated as the primary method for all samples, but "experts in the field of DMD diagnosis" confirmed the validity of DMD gene sequencing.
• Pathology: Not applicable in this context as it's not a tissue-based diagnosis.
• Outcomes Data: Not primary. The ground truth refers to a definitive diagnosis rather than long-term outcomes.
• Genetic Testing / Clinical Diagnosis: This is the primary method of ground truth.
  • For the investigated routine samples (those above initial cut-off or in a specific lower range), Next-Generation Sequencing (NGS) of the DMD gene was used as the ground truth.
  • For the 30 retrospectively confirmed positive DMD samples, their clinical diagnosis was the ground truth.

8. Sample Size for the Training Set

• The document does not provide details on a specific "training set" for an AI algorithm. This device is a fluoroimmunometric assay, not an AI/ML-based diagnostic system that would typically undergo a training phase with a distinct dataset.
• The "calibration curve" mentioned in calibration (Section M.5) is part of the analytical process to quantify CK-MM concentration, not training in the AI/ML sense. Similarly, "expected values" are derived from a separate study (2019 routine newborn screening specimens) but are not a "training set" for an algorithm.

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

• Since there's no mention of a "training set" in the context of an AI/ML algorithm, this question is not applicable. The device's calibration and analytical performance were established through standard laboratory practices and CLSI guidelines for IVDs, which do not involve AI/ML training concepts or associated ground truth establishment.

Ask a specific question about this device

The GSP Neonatal Total Galactose kit is intended for the quantitative determination of total galactose and galactose-1-phosphate) concentrations in blood specimens dried on filter paper as an aid in screening newborns for galactosemia using the GSP® instrument.

The GSP Neonatal Total Galactose test system measures total galactose, i.e. both galactose and galactose-1-phosphate, using a fluorescent galactose oxidase method. The fluorescence is measured using an excitation wavelength of 505 nm and an emission wavelength of 580 nm. The GSP Neonatal Total Galactose kit contains sufficient reagents to perform 1152 assays. The kit contains Calibrators, Controls, Neonatal Total Galactose Assay Reagent 1, Neonatal Total Galactose Assay Reagent 2, Neonatal Total Galactose Assay Buffer, Neonatal Total Galactose Assay Reconstitution Solution, and Neonatal Extraction Solution.

The provided document describes the K190335 submission for the GSP Neonatal Total Galactose kit, a device used for screening newborns for galactosemia. It primarily focuses on demonstrating the substantial equivalence of the new device (3309-002U) to a previously cleared predicate device (3309-001U).

Here's an analysis of the acceptance criteria and the study proving the device meets them, based on the provided text, with the understanding that this is a medical device clearance document, not an AI/ML model acceptance study. Therefore, some of the requested information (like number of experts for AI ground truth, MRMC study, training set details) are not directly applicable to this type of device and submission.

Acceptance Criteria and Reported Device Performance

The acceptance criteria are implicitly derived from the comparison to the predicate device and standard analytical performance metrics for in vitro diagnostic tests. The goal is to show that the new device performs equivalently to the predicate.

1. Table of Acceptance Criteria and Reported Device Performance:

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

• Analytical Validation (Precision, LoB, LoD, LoQ, Linearity, Recovery, Interference, Hook Effect): Various sample sizes specific to each experiment (e.g., 150 for LoB, 60 for LoD, 40 plates/80 results for repeatability, 15 plates/75 results for between-instrument, 15 plates/75 results for between-lot variation). Samples were human red blood cell enriched with galactose, human blood enriched with galactose and galactose-1-phosphate, whole blood with added substances, or contrived dried blood spot samples.
• Method Comparison:
  • Comparison with 3029-0010 Neonatal Total Galactose kit (different method): n=139 samples (routine newborn screening dried blood spot samples and dried adult human whole blood samples spiked with galactose and galactose-1-phosphate).
  • Comparison with 3309-001U GSP Neonatal Total Galactose kit (predicate): n=545 routine newborn screening dried blood spot samples.
• Screening Performance Study:
  • Test Set Size: 2161 samples.
  • Data Provenance: Conducted at one newborn screening laboratory in the United States.
  • Retrospective/Prospective: The samples included "routine newborn screening dried blood spot samples" (implying prospective collection in a screening program context) and "retrospective galactosemia diagnosed screening samples" (implying retrospective identification of confirmed positive cases). Specifically, the 95th percentile analysis included 5 retrospective galactosemia diagnosed screening samples and 1 retrospective galactosemia screening sample collected 22 hours after birth. The 99th percentile analysis included 4 retrospective galactosemia diagnosed screening samples, 1 retrospective galactosemia diagnosed screening sample collected 22 hours after birth, and 1 retrospective galactosemia diagnosed screening sample collected 16 hours after birth.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts:

• This is an in vitro diagnostic (IVD) device, not an AI/ML model for image interpretation. The "ground truth" for the screening performance study is clinical diagnosis of galactosemia, or the results from the predicate device using routine newborn screening samples.
• The document does not specify the number or qualifications of experts (e.g., radiologists) in the context of establishing ground truth, as this is not a study involving subjective interpretation like medical imaging by human experts. The 'truth' is derived from the biochemical measurements and clinical outcomes associated with galactosemia screening performed by a validated screening program.

4. Adjudication Method for the Test Set:

• Not applicable in the context of an IVD device. The 'comparison' and 'screening performance' results are based on quantitative measurements by both the new device and the predicate device compared to each other, and against known clinical results (for retrospective samples). There is no "adjudication" between human readers or AI outputs in the way it's understood for image interpretation or diagnosis.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was done:

• No. An MRMC study is relevant for perception tasks like image interpretation where human readers' performance is evaluated. This is an in vitro diagnostic assay that produces quantitative results. The comparison is between the new device's quantitative output and the predicate device's quantitative output, as well as their agreement on screening classification (positive/negative) against the established screening program's outcomes.

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

• Yes, in a sense. The GSP Neonatal Total Galactose kit is a standalone in vitro diagnostic system. Its performance (accuracy, precision, linearity, etc.) is measured intrinsically and compared to the predicate, independent of human operators' subjective interpretation. The "screening performance study" evaluates the device's ability to classify samples as screen positive or negative based on its quantitative output, without direct human "interpretation" of the assay result itself. Human decision-making uses this quantitative result but isn't part of the direct device performance.

7. The Type of Ground Truth Used:

• For analytical performance (Precision, LoB, LoD, LoQ, Linearity, Recovery, Interference, Hook Effect): The ground truth is established by known concentrations of analytes (galactose, galactose-1-phosphate) in spiked samples, or by established measurement principles for blank and low-level samples.
• For Method Comparison: The ground truth is the measurement obtained from the predicate device and the 3029-0010 Neonatal Total Galactose kit. This establishes agreement.
• For Screening Performance Study: The ground truth is a combination of:
  • Predicate Device Results: For routine samples, the predicate device's classification (screen positive/negative) serves as the comparator.
  • Clinical Diagnosis/Outcomes: For "retrospective galactosemia diagnosed screening samples," the confirmed clinical diagnosis of galactosemia (presumably through follow-up testing and clinical presentation) serves as the ultimate truth for these specific cases.

8. The Sample Size for the Training Set:

• This device is an IVD kit, not an AI/ML algorithm. Therefore, there is no "training set" in the context of machine learning. The device's calibration curve is established using known calibrators provided in the kit. The "training" of an IVD like this involves chemical formulation, assay optimization, and manufacturing process control, not data-driven learning from a "training set" like an AI model.

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

• Not applicable, as there is no training set for this type of device. The calibrators have been prepared from human red blood cells enriched with galactose and calibrated against primary calibrators gravimetrically prepared using a U.S. Pharmacopeia Reference Standard Preparation for galactose. This establishes the "truth" for calibration.

Ask a specific question about this device

The NeoLSD™ MSMS kit is intended for the quantitative measurement of the activity of the enzymes acid-9glucocerebrosidase (ABG), acid-sphingomyelinase (GAA), 8 galactocerebrosidase (GALC), agalactosidase A (GLA) and a-L-iduronidase (IDUA) in dried blood spots (DBS) from newborn babies. The analysis of the enzymatic activity is intended as an aid in screening newborns for the following lysosomal storage disorders (LSD) respectively; Gaucher Disease, Niemann-Pick A/B Disease, Krabbe Disease, Fabry Disease, and Mucopolyaccharidosis Type I (MPS I) Disease.

The NeoLSD MSMS test system uses mass spectrometry to quantitatively measure the activity of six lysosomal enzymes simultaneously from a dried blood spot sample. The NeoLSD MSMS test system is comprised of:

1. NeoLSD MSMS kit, including substrates, internal standards, solutions and controls
2. The QSight Instrument is comprised of:
   QSight® 210 MD Mass Spectrometer O
   QSight HC Autosampler MD Instrument Software O
   QSight Binary Pump MD O
   Simplicity Instrument control software: O
   Simplicity Data Processing software (by sample): O
   O PerkinElmer MSMS Workstation Data Processing Software

The NeoLSD MSMS kit evaluates enzyme activities by measuring the product generated when an enzyme reacts with a synthesized substrate to create a specific end product. The activities of the six lysosomal enzymes present in a 3.2 mm punch from a dried blood spot (DBS) are simultaneously measured by the NeoLSD MSMS kit. The punches are incubated with the assay reagent mixture which contains;

• six substrates, one corresponding to each lysosomal enzyme
• six stable-isotope mass-labeled internal standards (IS) each designed to chemically resemble each product generated
• . a buffer to maintain the reaction pH, and to carry inhibitors to limit activity from competing enzymes if present and additives to enhance the targeted enzyme reactions.

Here's a breakdown of the acceptance criteria and study information for the NeoLSD MSMS Kit, based on the provided document:

1. Table of Acceptance Criteria and Reported Device Performance

The document doesn't explicitly state acceptance criteria in a dedicated table for screening performance per se, but it details the analytical performance and implies that meeting the predicate device's performance characteristics for screening, along with established analytical limits, constitutes acceptance. The "Comparison Chart" on page 6 includes some comparable metrics between the proposed and predicate device.

Below is a table summarizing the reported analytical performance, with implied acceptance criteria that the device's performance should be within acceptable clinical/analytical ranges and comparable to the predicate device.

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

• Sample Size for Screening Performance Study: 2489 routine newborn samples + 12 archived confirmed LSD positive newborn Dried Blood Spot (DBS) specimens = 2501 samples.
• Data Provenance:
  • Country of Origin: The study was conducted at a "US newborn screening laboratory (Site A)".
  • Retrospective/Prospective: Primarily retrospective for the routine samples as they were "routine newborn samples". The 12 positive samples were "archived confirmed LSD positive newborn DBS specimens", making them retrospective as well. The device also mentions testing samples from newborns ≤ 48 hours old, common for newborn screening.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications

The document does not explicitly state the number of experts or their qualifications. It refers to the 12 archived confirmed LSD positive newborn DBS specimens as having a "confirmed" diagnosis. This implies clinical confirmation, likely by medical specialists and further diagnostic testing, but the specific process for establishing this ground truth (e.g., expert panel review of clinical, biochemical, and genetic data) is not detailed.

4. Adjudication Method for the Test Set

The document does not describe a formal adjudication method for the test set in terms of expert review for AI output. Instead, the study compares the screening results of the new device (QSight) against a predicate device (Waters Acquity TQD) and against retrospectively confirmed positive samples. The "ground truth" for the 12 positive samples was their established "confirmed" LSD status, not an adjudication process by experts specifically for this study.

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

This is not applicable. The device is an in-vitro diagnostic kit for quantitative measurement of enzyme activity using mass spectrometry (MSMS), not an image-based AI system that assists human readers. The study involves comparing the performance of the new kit on one MSMS instrument (QSight) against its predicate on a different MSMS instrument (Waters Acquity TQD).

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

Yes, this study represents a standalone performance evaluation of the "NeoLSD MSMS Kit" on the "QSight Instrument" system. The kit/instrument system performs the quantitative measurement of enzyme activity. While a human laboratory technician operates the instrument and interprets the numerical results based on cut-off values, the core diagnostic output (enzyme activity) is determined by the automated system without subjective human interpretation of raw data fields that would typically be associated with "standalone AI" vs "human-in-the-loop" in other diagnostic contexts (e.g., radiology). The comparison is between two automated systems (QSight vs. TQD).

7. The Type of Ground Truth Used

The ground truth for the 12 positive samples was confirmed LSD diagnoses, which implies a combination of:

• Clinical diagnosis: Based on patient symptoms.
• Biochemical confirmation: Enzyme assays, metabolite analysis.
• Genetic confirmation: DNA sequencing to identify pathogenic variants.

For the 2489 routine newborn samples, the ground truth is implied to be their "routine" healthy status, where no LSD was suspected or later confirmed, and their performance was used to establish median values and cut-offs.

8. The Sample Size for the Training Set

The document does not explicitly mention a "training set" in the context of machine learning. The term "training set" is typically used for developing AI models. This device is a diagnostic kit based on mass spectrometry, not a machine learning algorithm that is "trained." Performance characteristics like expected values (means, medians, percentiles) were derived from a large population of 2488 (or 2489) routine newborn samples, which could be considered a reference population for establishing normal ranges and cut-offs.

• Expected Values (Reference Population): 2488 or 5041 (depending on the enzyme) de-identified residual DBS samples from routine newborn screening. (Page 7, Table "Expected Values")

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

As noted above, there's no explicit "training set" in the AI sense. However, the "Expected Values" table on page 7 lists statistics derived from "N=2488" and "N=5041" samples, which are described as de-identified residual DBS samples from routine newborn screening. These would represent a population assumed to be largely unaffected (normal controls) in terms of LSDs, allowing for the establishment of normal enzyme activity ranges, medians, and percentiles. The ground truth for these samples would be their implicit "normal" or "no confirmed LSD" status based on routine screening outcomes or lack of follow-up diagnoses.

Ask a specific question about this device