Search Results

Philips Magnetic Resonance (MR) systems are Medical Electrical Systems indicated for use as a diagnostic device.

This MR system enables trained physicians to obtain cross-sectional images, spectroscopic images and/or spectra of the internal structure of the head, body or extremities, in any orientation, representing the spatial distribution of protons or other nuclei with spin.

Image appearance is determined by many different physical properties of the tissue and the MR scan technique applied, and presence of contrast agents. The use of contrast imaging applications should be performed consistent with the approved labeling for the contrast agent.

The trained clinical user can adjust the MR scan parameters to customize image appearance, accelerate image acquisition, and synchronize with the patient's breathing or cardiac cycle.

The systems can use combinations of images to produce physical parameters, and related derived images, spectra, and measurements of physical parameters, when interpreted by a trained physician, provide information that may assust diagnosis, and therapy planning. The accuracy of determined physical parameters depends on system and scan parameters and must be controlled and validated by the clinical user.

In addition, the Philips MR systems provide imaging capabilities, such as MR fluoroscopy, to guide and evaluate interventional and minimally invasive procedures in the head, body and extremities. MR Interventional procedures, performed inside or adjacent to the Philips MR system, must be performed with MR Conditional or MR Safe instrumentation as selected and evaluated by the clinical user for use with the specific MR system configuration in the hospital. The appropriateness and use of information from a Philips MR system for a specific interventional procedure and specific MR system configuration must be validated by the clinical user.

The proposed Ingenia, Ingenia CX, Ingenia Elition, Ingenia Ambition, MR 5300 and MR 7700 MR Systems R12 are 60 cm and 70 cm bore 1.5 and 3.0 Tesla (1.5T and 3.0T) Magnetic Resonance Diagnostic Devices, hereafter to be known as Ingenia, Ingenia CX, Ingenia Elition, Ingenia Ambition, MR 5300 and MR 7700 MR Systems.

This bundled abbreviated 510(k) submission, is prompted by the introduction of a new optional software feature called Precise Image contained in software R12 for the proposed Ingenia, Ingenia CX, Ingenia Elition, Ingenia Ambition, MR 5300 and MR 7700 MR Systems, as compared to our legally marketed primary predicate device Achieva, Ingenia, Ingenia CX, Ingenia Elition, and Ingenia Ambition MR Systems (R11) (K213583) and the secondary predicate device MR 5300 and MR 7700 R11 MR Systems (K223442).

Precise Image is a deep learning based reconstruction technique designed to increase signal-to-noise ratio (SNR), increase sharpness and decrease ringing artefacts from MR images.

The introduction of Precise Image only required updates to the MR System Software.

The proposed Ingenia, Ingenia CX, Ingenia Elition, Ingenia Ambition, MR 5300 and MR 7700 MR Systems are intended to be marketed with the following pulse sequences and coils that are previously cleared by FDA:

1. mDIXON (K102344)
2. SWIp (K131241)
3. mDIXON-Quant (K133526)
4. MRE (K140666)
5. mDIXON XD (K143128)
6. O-MAR (K143253)
7. 3D APT (K172920)
8. Compatible System Coils

The accessories to be used with the proposed device Ingenia, Ingenia CX, Ingenia Elition, Ingenia Ambition, MR 5300 and MR 7700 MR Systems have not changed compared to the primary predicate device and secondary predicate device and can be found in the Instructions for Use accompanying the device:
System coils PPU Sensor for wireless physiology Pediatric PPU Sensor FlexTrak trolleys (FlexTrak / HA FlexTrak II) Acoustic Hood MR Elastography

When Philips MRI system is used in combination with the Philips MR-RT or MR-OR solutions, the user is referred to the dedicated MR-RT and MR-OR Instructions for Use for information on additional accessories that may apply:
Flextrak OR MR-RT CouchTop RT CouchTop XD

The provided text describes the acceptance criteria and the study conducted for the Philips Ingenia, Ingenia CX, Ingenia Elition, Ingenia Ambition, MR 5300, and MR 7700 MR Systems, specifically focusing on the new "Precise Image" software feature.

Here's a breakdown of the requested information:

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

2. Sample size used for the test set and the data provenance

The document mentions "a reader evaluation by ABR board certified radiologists was performed" and refers to "MR images," but does not explicitly state the sample size (number of images or cases) used for the test set. It also does not specify the country of origin of the data or whether it was retrospective or prospective.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts

The document states "ABR board certified radiologists" were used for the reader evaluation. It does not specify the exact number of experts or their years of experience.

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

The document does not specify the adjudication method used for the reader evaluation. It only mentions that radiologists performed the evaluation.

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

The study was a reader evaluation comparing MR images reconstructed with the new "Precise Image" feature (deep learning-based) to images from the predicate device reconstruction technology. While it involved multiple readers (radiologists) and multiple cases (implied by "MR images" and "abnormalities and pathologies"), it was primarily an evaluation of the image quality and diagnostic equivalence of the algorithm's output rather than a direct MRMC comparative effectiveness study measuring human reader improvement with AI assistance versus without.

Therefore, the document does not provide an effect size of how much human readers improve with AI vs without AI assistance. The focus was on the performance of the algorithm's output itself in comparison to existing technology.

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

Yes, a standalone evaluation was implicitly done through the "reader evaluation by ABR board certified radiologists." The radiologists assessed the properties of the images generated by the new Precise Image technique against images from the predicate reconstruction technology. This assesses the algorithm's output quality and diagnostic value in a standalone capacity, even though humans are interpreting its output.

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

The ground truth for the reader evaluation seems to be based on expert assessment/consensus by the ABR board-certified radiologists, who evaluated image properties (SNR, artifact level, sharpness, CNR) and the visualization of abnormalities/pathologies. The document doesn't explicitly mention external pathology or outcomes data as the primary ground truth for this specific evaluation, rather the expert interpretation of the images.

8. The sample size for the training set

The document does not mention the sample size used for the training set of the deep learning-based "Precise Image" reconstruction technique.

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

The document does not provide information on how the ground truth for the training set was established for the deep learning-based "Precise Image" technique. It only describes the nature of the technique as "deep learning based reconstruction."

Ask a specific question about this device

The SIGNA Prime is a whole body magnetic resonance scanner designed to support high resolution, high signal-to-noise ratio, and short scan times. It is indicated for use as a diagnostic imaging device to produce axial, sagittal, coronal, and oblique images, spectroscopic images, parametric maps, and/or spectra, dynamic images of the structures and/or functions of the entire body, including, but not limited to, head, neck, TMI, spine, abdomen, pelvis, joints, prostate, blood vessels, and musculoskeletal regions of the body.
Depending on the region of interest being imaged, contrast agents may be used. The images produced by SIGNA Prime reflect the spatial distribution or molecular environment of nuclei exhibiting magnetic resonance.
These images and/or spectra when interpreted by a trained physician yield information that may assist in diagnosis.

SIGNA™ Prime is a whole body magnetic resonance scanner designed to support high resolution, high signal to-noise ratio, and short scan times. The systems use a combination of time-varying magnet fields (Gradients) and RF transmissions to obtain information regarding the density and position of elements exhibiting magnetic resonance. The system can image in the sagittal, coronal, axial, oblique, and oblique planes, using various pulse sequences, imaging techniques and reconstruction algorithms. The system features a 1.5T superconducting magnet with 60cm bore size. The system is designed to conform to NEMA DICOM standards (Digital Imaging and Communications in Medicine).

The provided text is a 510(k) summary for the GE Healthcare SIGNA Prime magnetic resonance diagnostic device. The summary states that the device did not require clinical studies to support substantial equivalence and instead relied on non-clinical tests and sample clinical images to demonstrate acceptable diagnostic image performance.

Therefore, the study details requested cannot be fully provided as a formal comparative effectiveness study or standalone performance study as typically understood for AI/CADe devices was not conducted with predefined acceptance criteria for diagnostic metrics.

Here's a breakdown of the available information:

1. Table of Acceptance Criteria and Reported Device Performance:

No specific numerical acceptance criteria for diagnostic performance (e.g., sensitivity, specificity, AUC) are provided in the document. The general acceptance criterion was that the image quality of the SIGNA Prime is "substantially equivalent" to that of the predicate device (SIGNA Creator, K143251).

2. Sample size used for the test set and the data provenance:

• Sample Size: The document mentions "Sample clinical images have been included in this submission" but does not specify the number of images or cases used as a "test set."
• Data Provenance: Not specified. It's unclear if these were retrospective or prospective, or the country of origin.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

• The document states that "These images and/or spectra when interpreted by a trained physician yield information that may assist in diagnosis." However, it does not specify the number or qualifications of experts who interpreted the "sample clinical images" for the purpose of demonstrating substantial equivalence. The mechanism for establishing ground truth for these sample images is not detailed.

4. Adjudication method for the test set:

• Not specified. Given that a formal clinical study with a detailed ground truth process is not described, an adjudication method for a test set is not mentioned.

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. The document explicitly states: "The subject of this premarket submission, the SIGNA™ Prime, did not require clinical studies to support substantial equivalence." Therefore, an MRMC comparative effectiveness study involving human readers with and without AI assistance was not performed or reported.

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

• The SIGNA Prime is a magnetic resonance scanner, not an AI algorithm or CADe device. Therefore, the concept of "standalone (algorithm only without human-in-the-loop performance)" is not applicable in this context. The device itself produces images for human interpretation.

7. The type of ground truth used:

• For the "sample clinical images" used to demonstrate acceptable diagnostic image performance, the ground truth is implicitly expert interpretation by a "trained physician" as stated in the Indications for Use. However, the exact methodology for establishing this ground truth for the purpose of the submission is not detailed (e.g., expert consensus vs. pathology vs. outcomes data).

8. The sample size for the training set:

• Not applicable. The SIGNA Prime is a hardware device (MRI scanner) with associated software, not an AI/ML algorithm that is "trained" on a dataset in the typical sense. While the software platform and reconstruction algorithms were likely developed and refined, the document does not describe a "training set" in the context of an AI algorithm evaluation.

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

• Not applicable. (See point 8).

Summary of the Study:

The K211980 submission for the SIGNA Prime focused on demonstrating substantial equivalence to a predicate device (SIGNA Creator, K143251). This was primarily achieved through:

• Non-clinical tests: Compliance with various electrical, safety, software, and biocompatibility standards (e.g., ANSI/AAMI ES60601-1, IEC 60601-1-2, IEC 60601-2-33, IEC 62304, IEC 60601-1-6, IEC 62366-1, ISO 10993-1, NEMA MS, NEMA PS3 DICOM).
• Risk management activities: Including risk analysis, design reviews, and various levels of testing (unit, integration, performance, simulated use).
• Sample clinical images: These images were provided to demonstrate acceptable diagnostic performance and visual equivalence to the predicate device. However, the specific methodology for selecting, evaluating, or establishing ground truth for these sample images is not detailed, nor are any quantitative metrics provided for their performance.

The submission explicitly states that clinical studies were not required to support substantial equivalence. Therefore, the detailed information typically sought for the evaluation of AI/CADe devices (such as sample sizes for test/training, expert qualifications, adjudication methods, or MRMC study results) is not present in this document.

Ask a specific question about this device

The SIGNA 7.0T system is a whole-body magnetic resonance scanner designed to support high signal-tonoise ratio, and short scan times. It is indicated for use as a diagnostic imaging device to produce axial, sagittal, coronal, and oblique images, spectroscopic images, parametric maps, and/or spectra, dynamic images of the structures and/or functions of the head and extremities.

The images produced by the SIGNA 7.0T system reflects the spatial distribution or molecular environment of nuclei exhibiting magnetic resonance. These images and/or spectra when interpreted by a trained physician yield information that may assist in diagnosis.

The device is intended for patients > 20 kg / 44 lb.

SIGNA 7.0T is a high performance magnetic resonance imaging system designed to support high resolution imaging at 7.0T in particular anatomical regions determined by the available RF coils. The system includes a 7.0T superconducting magnet and an ultra-high performance gradient coil with a 60 cm patient bore, supporting scanning in axial, coronal, sagittal, oblique, and double oblique planes using a variety of pulse sequences, imaging techniques, acceleration methods, and reconstruction algorithms.

This 510(k) submission is for the SIGNA 7.0T MR System, and has been triggered by the addition of the AIR Recon DL software feature and inclusion of installed base magnet system upgrades.

The provided text describes a 510(k) submission for the GE SIGNA 7.0T MRI system, specifically focusing on the addition of the AIR Recon DL software feature. While it discusses performance testing and a reader study, it does not explicitly define specific numerical acceptance criteria for the device's performance in a table format, nor does it provide detailed quantitative results against such criteria. The document states that "The nonclinical testing passed the defined acceptance criteria," but these criteria are not enumerated.

However, based on the provided text, we can infer and report on the study details as much as possible:

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

As stated above, no explicit numerical acceptance criteria table is provided in the document. The text broadly states:

• "The nonclinical testing passed the defined acceptance criteria, and did not identify any adverse impacts to image quality or other concerns related to safety and performance."
• "The nonclinical testing demonstrated that AIR Recon DL does improve SNR and image sharpness while maintaining low contrast detectability."
• "Objective measures of in vivo images were analyzed to confirm that AIR Recon DL improves SNR and image sharpness for typical clinical use cases."
• "This study showed that AIR Recon DL feature provides images with better SNR and equivalent or better sharpness."
• "The radiologists uniformly preferred the AIR Recon DL images for clinical evaluation."

2. Sample size used for the test set and the data provenance

• Test Set Sample Size: Not explicitly stated for either the non-clinical or clinical tests. The text mentions "in vivo images" for clinical analysis and "typical clinical use cases," implying a dataset was used, but the size is absent.
• Data Provenance: Not explicitly stated (e.g., country of origin, retrospective/prospective).

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts

• Number of Experts: "Radiologists" (plural), but the exact number is not specified.
• Qualifications of Experts: Only stated as "trained physician" for general image interpretation within the Indications for Use, and "Radiologists" for the reader study. No specific experience levels (e.g., years of experience) are provided for the radiologists in the study.

4. Adjudication method for the test set

• Adjudication Method: Not explicitly stated. The text says "Radiologists were asked to rate the images, and to comment on any notable aspects related to image quality," and that they "uniformly preferred" the AIR Recon DL images. This implies a consensus or preference-based evaluation rather than a formal adjudicated ground truth establishment process for specific findings.

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

• MRMC Study: Yes, a reader study was performed comparing images "with and without AIR Recon DL feature."
• Effect Size: No quantitative effect size (e.g., magnitude of improvement in diagnostic accuracy, AUC, or other metrics) is provided for how much human readers improved. The improvement is described qualitatively: "The radiologists uniformly preferred the AIR Recon DL images for clinical evaluation." and that images had "better SNR and equivalent or better sharpness."

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

• Standalone Performance: The document describes "nonclinical testing" and "objective measures of in vivo images" where AIR Recon DL's impact on SNR and sharpness was evaluated. This implies an algorithmic evaluation independent of human readers, where the algorithm processes images and its output is analyzed for objective metrics (SNR, sharpness, low contrast detectability). However, specific performance metrics for this standalone performance (e.g., a specific SNR improvement percentage, or sharpness metric) are not quantified in the text.

7. The type of ground truth used

• Clinical Study: For the reader study, the ground truth appears to be based on the radiologists' preference and qualitative assessment of image quality (SNR, sharpness, low contrast detectability) for clinical evaluation. It's not explicitly stated as a definitive "diagnosis" ground truth derived from pathology or outcomes data.
• Non-Clinical/Objective Measures: For objective measures, the "ground truth" aligns with quantifiable physical properties like Signal-to-Noise Ratio (SNR) and image sharpness, which are inherent characteristics of the image output itself. Low contrast detectability was also assessed.

8. The sample size for the training set

• Training Set Sample Size: Not specified in the provided text. The text mentions that "AIR Recon DL has been previously cleared for use with GE Healthcare's 3T SIGNA Premier system through K193282" and that "Due to the technical similarities, SIGNA Premier (K193282) is used as a reference device for this submission." This implies that the algorithm was likely trained on a separate dataset prior to this specific submission, but details of that training set are not in this document.

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

• Training Set Ground Truth Establishment: Not specified in the provided text.

Ask a specific question about this device

The SIGNA 7.0T System is a whole-body magnetic resonance scanner designed to support high signal-tonoise ratio, and short scan times. It is indicated for use as a diagnostic imaging device to produce axial, sagittal, coronal, and oblique images, spectroscopic images, parametric maps, and/or spectra, dynamic images of the structures and/or functions of the head and extremities.

The images produced by the SIGNA 7.0T System reflect the spatial distribution or molecular environment of nuclei exhibiting magnetic resonance. These images and/or spectra when interpreted by a trained physician yield information that may assist in diagnosis.

The device is intended for patients > 20 kg / 44 lbs.

SIGNA 7.0T is a high performance magnetic resonance imaging system designed to support high resolution imaging at 7.0T in particular anatomical regions determined by the available RF coils. The system includes a 7.0T superconducting magnet and an ultra-high performance gradient coil with a 60 cm patient bore, supporting scanning in axial, coronal, sagittal, oblique, and double oblique planes using a variety of pulse sequences, imaging techniques, acceleration methods, and reconstruction algorithms.

Here's an analysis of the acceptance criteria and the study conducted for the GE SIGNA 7.0T Magnetic Resonance Diagnostic Device, based on the provided text:

Important Note: The provided text is a 510(k) summary, which focuses on demonstrating substantial equivalence to a predicate device rather than presenting a detailed clinical study for novel claims or specific performance metrics against pre-defined acceptance criteria for a new AI-powered diagnostic. Therefore, some of the requested information (like specific quantitative acceptance criteria for image quality, expert qualifications beyond "radiologist," or a detailed MRMC study effect size) is not explicitly present in the given document. The analysis below extracts what is available and makes inferences where appropriate based on the nature of a 510(k) submission for an imaging device.

1. Table of Acceptance Criteria and Reported Device Performance

Explanation of Implied Acceptance Criteria: The document describes the study's purpose as evaluating image diagnostic quality and usability against the predicate device. For a 510(k), the acceptance criterion for image quality is generally that the new device's images are diagnostically equivalent or non-inferior to the predicate device's images for the stated indications for use. "Acceptance" here means that the FDA agreed that the device met the requirements for substantial equivalence.

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

• Sample Size for Image Quality Study (Test Set):
  • Subjects: Not explicitly stated, but images were acquired from "normal subjects and subjects with self-reported common neuropathology" for brain scans, and individuals over a "broad age range representative of a range of knee health" for knee images. The specific number of subjects/cases is not provided.
  • Image Series: Included brain MR scans and knee images.
• Sample Size for PNS Limits Study: Not explicitly stated, but involved "adult human volunteers without reported pathology."
• Data Provenance: Not explicitly stated, but given it's a US submission and refers to "U.S. board-certified radiologists," it is highly probable the data was collected in the United States. The study was prospective in the sense that the images were acquired specifically for this evaluation using both the proposed and predicate devices.

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

• Number of Experts: 4
• Qualifications: "U.S. board-certified radiologists." No specific years of experience are provided, but board certification implies a certain level of expertise.

4. Adjudication Method for the Test Set

The document states, "The study involved 4 U.S. board-certified radiologists evaluating proposed device's image diagnostic quality and usability, personal preferences, and general commentary using radiology terms against same subject images scanned on the predicate device."

This description suggests that the radiologists independently evaluated images from both devices. It does not explicitly describe a formal adjudication method (like 2+1 or 3+1 consensus). It's more likely that their individual assessments were collected and analyzed, potentially looking for consistency or a majority opinion on diagnostic quality comparison. Without further detail, we cannot definitively state a formal adjudication method was used, beyond individual evaluations.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, and the Effect Size of How Much Human Readers Improve with AI vs. Without AI Assistance

• MRMC Study: Yes, a reader evaluation study was performed. It involved multiple readers (4 radiologists) evaluating multiple cases (brain and knee images).
• Effect Size (AI Assistance): This device is an MR scanner, not an AI diagnostic algorithm. Therefore, the study did not involve human readers improving with AI assistance. The comparison was between the image quality from the new 7.0T MR scanner versus the predicate 3.0T MR scanner.

6. If a Standalone (Algorithm Only Without Human-in-the-Loop Performance) Was Done

• No, this question is not applicable. The SIGNA 7.0T is an imaging device, not an AI algorithm. No standalone algorithm performance was evaluated.

7. The Type of Ground Truth Used

For the image quality evaluation, the "ground truth" was established by expert consensus/opinion (or individual expert assessments, given the lack of explicit adjudication) regarding the diagnostic quality and usability of the images. They compared images from the proposed 7.0T system against the predicate 3.0T system. There is no mention of pathology, clinical outcomes, or other objective ground truth for the diagnostic findings themselves, but rather whether the images produced were suitable for diagnosis.

For the PNS limits, the ground truth was direct determination in human volunteers.

8. The Sample Size for the Training Set

• This device is an MR scanner, not an AI algorithm that requires a "training set" in the machine learning sense. Therefore, this question is not applicable. The device's "training" would refer to its engineering development and calibration, which is a different concept.

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

• This question is not applicable, as there is no "training set" in the context of an AI algorithm.

Ask a specific question about this device

The SIGNA Artist system is a whole body magnetic resonance scanner designed to support high signalto-noise ratio, and short scan times. It is indicated for use as a diagnostic imaging device to produce axial, sagittal, coronal. and oblique images, spectroscopic images, parametric maps, and/or spectra, dynamic images of the structures and/or functions of the entire body, including, but not limited to, head, neck, heart, abdomen, pelvis, joints, prostate, blood vessels, and musculoskeletal regions of the body. Depending on the region of interest being imaged, contrast agents may be used.

The images produced by the SIGNA Artist system reflect the spatial distribution or molecular environment of nuclei exhibiting magnetic resonance. These images and/or spectra when interpreted by a trained that may assist in diagnosis.

The SIGNA Artist system is a whole body magnetic resonance scanner designed to support high resolution, high signal-to-noise ratio, and short scan times. The system features a superconducting magnet. The data acquisition system accommodates up to 128 independent receive channels in various increments and multiple independent coil elements per channel during a single acquisition series. The system uses a combination of time varying magnetic fields (gradients) and RF transmissions to obtain information regarding the density and position of elements exhibiting magnetic resonance. The system can image in the sagittal, coronal, axial, oblique, and double oblique planes, using various pulse sequences and reconstruction algorithms.
This 510(k) submission is for the SIGNA Artist 1.5T MR system, and has been triggered by the addition of the AIR Recon DL software feature.
The AIR Recon DL feature has been previously cleared for use on the SIGNA Premier 3T system through K193282, which is used as a reference device for this submission.

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

Acceptance Criteria and Device Performance for GE Healthcare SIGNA Artist with AIR Recon DL

The provided document describes the 510(k) submission for the GE Healthcare SIGNA Artist system with the added AIR Recon DL software feature. The study focuses on evaluating the impact of this new feature on image quality.

1. Table of Acceptance Criteria and Reported Device Performance

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

The document does not explicitly state the specific sample size for the test set used in the clinical evaluation. It mentions "sample images from clinically indicated scans."

The data provenance for the clinical evaluation is implied to be retrospective as it involves "sample images from clinically indicated scans" that were then evaluated with and without the AIR Recon DL feature. The country of origin of the data is not specified.

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

The document states that "Radiologists were asked to rate the images, and to comment on any notable aspects related to image quality." It does not specify the exact number of experts or their qualifications (e.g., years of experience, subspecialty).

4. Adjudication Method for the Test Set

The adjudication method is not explicitly stated. The text only mentions that "Radiologists were asked to rate the images, and to comment on any notable aspects related to image quality." This suggests an individual review process, but it doesn't detail how discrepancies or consensus building was handled if multiple radiologists reviewed the same case. It doesn't mention methods like 2+1, 3+1, or majority vote.

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

A multi-reader multi-case (MRMC) comparative effectiveness study was not explicitly stated as the primary methodology. The clinical evaluation described involves radiologists rating images "both with and without the AIR Recon DL feature" and stating their preference. While this provides comparative feedback, it does not quantify human reader improvement in terms of diagnostic accuracy or a specific effect size. The study concludes that radiologists "preferred the AIR Recon DL images for clinical use" and that lesion conspicuity was maintained, indicating a subjective improvement, but not a measurable effect size of diagnostic performance.

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

Yes, a standalone performance evaluation was done as part of the "Nonclinical Tests." These tests were designed to evaluate the AIR Recon DL feature "and its impact on image quality, including SNR, sharpness, low contrast detectability, and noise spectral content. Analysis was performed to confirm that the feature does not introduce significant bias that might impact quantitative measurements based on signal intensity. The influence of motion during image acquisition on the performance of AIR Recon DL was also evaluated." These are objective, quantitative measurements of the algorithm's output without human interpretation being the primary endpoint.

7. The Type of Ground Truth Used

For the nonclinical tests, the ground truth appears to be based on objective image quality metrics, physical phantoms, and simulated conditions. For instance, evaluating SNR, sharpness, noise spectral content, a lack of signal intensity bias, and motion artifact influence against established benchmarks or predefined ideal conditions.

For the clinical tests, the ground truth for "equivalent or better image quality" and "maintained lesion conspicuity" was established by expert consensus/opinion from radiologists.

8. The Sample Size for the Training Set

The document does not specify the sample size for the training set used for the AIR Recon DL algorithm. It only mentions that the AIR Recon DL feature "has been previously cleared for use on the SIGNA Premier 3T system through K193282, which is used as a reference device for this submission." This implies the training was done prior to this specific submission for the SIGNA Artist.

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

The document does not provide details on how the ground truth for the training set was established for the AIR Recon DL algorithm. While it mentions the algorithm was previously cleared for another device, it does not elaborate on the specific data used for its initial training and ground truth annotation.

Ask a specific question about this device