Search Results

The 23andMe PGS Carrier Screening Test for Bloom Syndrome is indicated for the detection of the BLMAsh variant in the BLM gene from saliva collected using an FDA cleared collection device (Oragene DX model OGD-500.001). This test can be used to determine carrier status for Bloom syndrome in adults of reproductive age, but cannot determine if a person has two copies of the BLM140 variant. The test is most relevant for people of Ashkenazi Jewish descent.

The 23andMe Personal Genome Service (PGS) Carrier Screening Test for Bloom Syndrome (hereafter the "PGS") is a non-invasive genetic information service that combines qualitative genotyping data for an individual. The PGS is indicated for use for the detection of the BLM4sh variant in the BLM gene from saliva collected using the Oragene•Dx Saliva Collection Device (Oragene Dx model OGD-500.01). The core components of the PGS consist of the saliva collection kit: custom genotyping chip: laboratory procedures, equipment and analysis; and result reporting software.

Here's a breakdown of the acceptance criteria and the study details for the 23andMe Personal Genome Service Carrier Screening Test for Bloom Syndrome, based on the provided document:

1. Table of Acceptance Criteria and Reported Device Performance

Note: The document sometimes mixes "acceptance criteria" for the studies themselves with broader regulatory criteria. This table focuses on the performance metrics explicitly stated as needing to be met within the studies.

Study Details

2. Sample Size and Data Provenance for Test Set

• Human Cell Line Precision Study (Test Set):

  • Sample Size: 6 DNA samples from cell lines (4 BLM homozygous common ("DD"), 1 BLMAsh heterozygous ("DI"), 1 BLMAsh homozygous rare ("II")). These samples were replicated extensively: "DD" samples had 360 replicates each (total 1440), "DI" and "II" samples had 720 replicates each (total 1440). Grand total of 2,880 replicates.
  • Data Provenance: Not explicitly stated, but implies laboratory-prepared cell lines, not patient-derived samples from a specific country. This is retrospective in the sense of using pre-existing cell lines.

• Saliva Sample Reproducibility Study (Test Set):

  • Sample Size: 105 BLMAsh homozygous common ("DD") saliva samples.
  • Data Provenance: Obtained from individuals using the 23andMe Saliva Collection kit (Oragene-DX, OGD500.001). Implies prospective collection by users for 23andMe's biobank but used retrospectively for this study. Country of origin not specified.

• Limit of Detection Study (LoD) (Test Set):

  • Sample Size: DNA samples from cell lines: BLMAsh (homozygous wild type, DD) - 4 samples/4 replicates per sample; BLMAsh (heterozygous variant, DI) - 1 sample/8 replicates per sample; BLMAsh (homozygous variant, II) - 1 sample/8 replicates per sample. Each DNA sample tested at 3 concentrations (5, 15, 50 ng/uL).
  • Data Provenance: Cell lines (retrospective, laboratory-prepared).

• Interference Studies (Test Set):

  • Endogenous Interference: 10 individuals (homozygous common "DD") with saliva samples.
  • Exogenous Interference: 5 individuals with saliva samples, tested under 5 conditions (eating beef, eating non-beef, drinking, chewing gum, mouthwash) at 3 time points, in triplicate. Total 225 replicates.
  • Smoking Interference: 5 donors with saliva samples, tested under 1 condition (smoking) at 3 time points, in triplicate. Total 45 replicates.
  • Microbial Interference: 6 DNA samples from cell lines (4 "DD", 1 "DI", 1 "II") with 3 replicates each, spiked with 5 microbial DNAs. Total 90 replicates (6 samples x 5 microbes x 3 reps; plus controls).
  • Data Provenance: Saliva from individuals for endogenous/exogenous/smoking (likely retrospective from 23andMe's biobank, country unspecified); cell lines for microbial interference (retrospective, laboratory-prepared).

• Accuracy/Method Comparison Study (Test Set):

  • Sample Size: 65 saliva samples (25 DD and 22 DI at Site 1; 18 DI at Site 2). Also, 6 human cell line samples (4 DD, 1 DI, 1 II) at both sites. Totaling 71 unique samples, but with some samples counted across sites (e.g., 22 DI at Site 1 + 18 DI at Site 2 = 40 DI unique saliva samples tested across sites).
  • Data Provenance: Saliva samples randomly selected from the 23andMe Biobank (retrospective, country unspecified). Human cell line samples (retrospective, laboratory-prepared).

• User Comprehension Study (Test Set):

  • Sample Size: A total of 11 of 678 (1.6%) participants were excluded, meaning approximately 667 participants were included. Quota-based sampling targeting at least 100 subjects per 5 different representative test reports.
  • Data Provenance: Prospective collection for the study, conducted at 5 locations across the U.S.

3. Number of Experts and Qualifications for Ground Truth for the Test Set

• Analytical Validation Studies (Precision, LoD, Interference, Accuracy):

  • Ground Truth Providers: Performed by bi-directional Sanger Sequencing. The document does not specify a "number of experts" or their qualifications. Bidirectional Sanger sequencing is considered a gold standard for genetic variant confirmation, and its results are interpreted by trained laboratory personnel. The "experts" in this context are the established and validated sequencing protocol itself and the standard bioinformatics analysis of the sequencing data, rather than individual human experts adjudicating cases for complex image interpretation.

• User Comprehension Study:

  • Ground Truth Providers: The ground truth is the "correct answer" to comprehension questions, which would have been established by the study designers (likely geneticists, educators, regulatory affairs specialists) based on the intended meaning of the labeling. No external "experts" were used to establish ground truth for individual participant responses; the correctness of responses was evaluated against pre-defined correct answers.

4. Adjudication Method for the Test Set

• Analytical Validation Studies:

  • Adjudication Method: None, in the typical sense of expert review for ambiguous cases. The ground truth (bi-directional Sanger sequencing) is considered definitive. Discordant results between the device and sequencing would be thoroughly investigated, but not "adjudicated" by multiple human reviewers. The document states that "BeadChip genotypes were compared with sequenced genotypes to determine the rates of correct BeadChip genotype calls." For FQC (failed quality control) replicates, they were re-run as per laboratory SOPs.

• User Comprehension Study:

  • Adjudication Method: None. Participant responses were assessed against pre-defined correct answers established by the study design. There was no mention of multiple reviewers adjudicating ambiguous participant answers.

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

• No MRMC comparative effectiveness study was done. This device is a standalone genetic test for detecting a specific variant, not an AI-assisted diagnostic tool that aids human readers (e.g., radiologists interpreting images). Therefore, there is no "human readers improve with AI vs without AI assistance" effect size reported or relevant for this type of device.

6. Standalone Performance (Algorithm Only without Human-in-the-Loop)

• Yes, a standalone performance study was done. All the analytical performance studies (Precision/Reproducibility, LoD, Interference, Accuracy) directly assess the performance of the 23andMe PGS Carrier Screening Test (BeadChip, iScan, GenomeStudio, and Coregen software) in detecting the BLMAsh variant without human geneticists or counselors "in the loop" for the initial genotype calling. The test produces a genotype call ("DD", "DI", "II", or "no-call") directly. Human interaction comes in downstream for result interpretation and counseling (which the user comprehension study addresses).

7. Type of Ground Truth Used

• Analytical Performance Studies (Precision, LoD, Interference, Accuracy):

  • Ground Truth Type: Bi-directional Sanger Sequencing. This is a highly accurate molecular method considered the gold standard for confirming specific DNA sequences and variants.

• User Comprehension Study:

  • Ground Truth Type: Expert consensus for the correct answers to comprehension questions about the test report and genetic concepts. This was established by the designers of the study, based on scientific and educational principles.

8. Sample Size for the Training Set

• The document does not explicitly describe a training set sample size for the device's core genotyping algorithm. The methods described (Infinium BeadChip, Illumina iScan System, GenomeStudio, and Coregen software) are based on established genotyping technologies that typically involve pre-defined probe sets and analytical algorithms. While these systems require initial development and calibration (analogous to 'training'), the document focuses on the validation of the specific assay and workflow for the BLMAsh variant.
• However, the GenomeStudio software is mentioned to perform "primary data analyses, such as raw data normalization, clustering, and genotype calling." These steps often involve some form of algorithmic learning or optimization that could be considered "training" on reference data sets, but specific sizes or details are not provided here for the 23andMe application.

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

• Since a specific "training set" with independent ground truth is not explicitly described for the underlying genotyping algorithm (see point 8), the method for establishing its ground truth is also not detailed.
• Implied ground truth for algorithm development/calibration: For systems like GenomeStudio's genotype calling, ground truth datasets for clustering and calling would typically be laboratory-prepared reference samples with known genotypes (confirmed by orthogonal methods like Sanger sequencing) or large population datasets where genotypes are well-established for common variants. This process is part of the general development and validation of the Illumina platform itself, rather than a specific 'training set' for this particular Bloom Syndrome test beyond the analytical validation datasets.

Ask a specific question about this device