Search Results

VitruvianScan is indicated for use as a magnetic resonance diagnostic device software application for non-invasive fat and muscle evaluation that enables the generation, display and review of magnetic resonance medical image data.

VitruvianScan produces quantified metrics and composite images from magnetic resonance medical image data which when interpreted by a trained healthcare professional, yield information that may assist in clinical decisions.

VitruvianScan is a standalone, post processing software medical device. VitruvianScan enables the generation, display and review of magnetic resonance (MR) medical image data from a single timepoint (one patient visit).

When a referring healthcare professional requests quantitative analysis using VitruvianScan, relevant images are acquired from patients at MRI scanning clinics and are transferred to the Perspectum portal through established secure gateways. Perspectum trained analysts use the VitruvianScan software medical device to process the MRI images and produce the quantitative metrics and composite images. The device output information is then sent to the healthcare professionals for their clinical use.

The metrics produced by VitruvianScan are intended to provide insight into the composition of muscle and fat of a patient. The device is intended to be used as part of an overall assessment of a patient's health and wellness and should be interpreted whilst considering the device's limitations when reviewing or interpreting images.

Here's an analysis of the acceptance criteria and the study provided in the document for the VitruvianScan (v1.0) device:

1. Table of Acceptance Criteria and Reported Device Performance

The provided document unfortunately does not explicitly state the specific quantitative acceptance criteria for each performance aspect (e.g., a specific percentage for repeatability, a particular correlation coefficient). Instead, it states that "All aspects of the performance tests met the defined acceptance criteria." and that the device "successfully passed the acceptance criteria with no residual anomalies."

However, based on the described performance tests, we can infer the types of acceptance criteria that would have been defined. The document also lacks specific numerical results for the device performance that directly map to these criteria.

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

The document does not specify the sample size used for the test set.

The data provenance (country of origin, retrospective/prospective) is not explicitly stated. However, the context of an FDA submission for a device used in clinical settings suggests the data would likely be from a clinical or research environment, potentially multi-center, but this is an inference.

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

• Number of Experts: 3 radiologists
• Qualifications of Experts: The document states "3 radiologists results" but does not provide specific qualifications (e.g., years of experience, subspecialty).

4. Adjudication Method for the Test Set

The adjudication method used for the test set is implicitly a "mean (average) of 3 radiologists results" for establishing the gold standard. This suggests a form of consensus, where the average of their interpretations serves as the reference. It's not a typical "X+1" method (like 2 out of 3, or 3 out of 1 for disagreement), but rather a central tendency measure of their assessments.

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

A MRMC comparative effectiveness study was NOT explicitly done in the sense of comparing human readers with AI assistance versus without AI assistance.

The study did involve "Comparative testing between the operators' results and the gold standard (mean of 3 radiologists results)," which evaluates the device's output and how operators use it against a human expert consensus. However, it doesn't describe a scenario where human readers improve with AI assistance in their own diagnostic performance compared to their performance without the AI. The stated use case is that "Perspectum trained analysts use the VitruvianScan software medical device to process the MRI images and produce the quantitative metrics and composite images," and then these are sent to "trained Healthcare Professionals who then utilize these to make clinical decisions." This suggests the device provides quantitative data to healthcare professionals, rather than directly assisting their image interpretation to improve their diagnostic accuracy from images alone.

6. Standalone (Algorithm Only) Performance Study

Yes, a standalone performance assessment was conducted. The "Comparative testing between the operators' results and the gold standard (mean of 3 radiologists results)" and the "benchmarking performance" against OSIRIX MD inherently assess the algorithm's output (via the trained analysts) against a reference, which signifies a standalone evaluation of the device's quantitative capabilities.

7. Type of Ground Truth Used

The primary type of ground truth used for the comparative testing was expert consensus (mean of 3 radiologists results).

8. Sample Size for the Training Set

The document does not provide any information regarding the sample size used for the training set for the VitruvianScan algorithm. This information is typically crucial for understanding the generalizability and robustness of an AI/ML device.

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

The document does not provide any information on how the ground truth for the training set was established.

Ask a specific question about this device

CoverScan is a medical image management and processing software package that allows the display, analysis and postprocessing of DICOM compliant medical images and MR data.

CoverScan provides both viewing and analysis capabilities to ascertain quantified metrics of multiple organs such as the heart, lungs, liver, spleen, pancreas and kidney.

CoverScan provides measurements in different organs to be used for the assessment of longitudinal and transversal relaxation time and rate (T1, SR-T1, cT1, T2), fat content (proton density fat fraction or PDFF) and metrics of organ function (e.g., left ventricular ejection fraction and lung fractional area change on deep inspiration).

These metrics derived from the interpreted by a licensed physician, vield information that may assist in diagnosis, clinical management and monitoring of patients.

CoverScan is not intended for asymptomatic screening. This device is intended for use with Siemens 1.5T MRI scanners.

CoverScan is a post-processing software system comprised of several software modules. It uses acquired MR data to produce metrics of quantified tissue characteristics of the heart, lungs, liver, kidneys, pancreas and spleen.

Metrics produced by CoverScan can be used by licensed physicians in a clinical setting for the purposes of assessing multiple organs.

The provided documentation describes the CoverScan v1 device, which is a medical image management and processing software. While it mentions internal verification and validation testing, and that product specifications were met, it does not explicitly state specific quantitative acceptance criteria or detailed results of a study proving the device meets those criteria, especially in a comparative effectiveness context (MRMC).

The document primarily focuses on establishing substantial equivalence to predicate devices through a qualitative comparison of intended use, technological characteristics, and performance features. It indicates that "bench testing included functional verification to ensure software installation, licensing, labeling, and feature functionality all met design requirements" and that "The accuracy and precision of device measurements was assessed using purpose-built phantoms and in-vivo acquired data from volunteers." However, it does not provide the specific quantitative results of these assessments against defined acceptance criteria.

Therefore, much of the requested information cannot be directly extracted from the provided text. I will explain what information is available and highlight what is missing.

Here's an attempt to structure the information based on the provided text, with clear indications where the information is not present:

Acceptance Criteria and Device Performance Study (CoverScan v1)

The provided document indicates that CoverScan v1 underwent internal verification and validation testing to confirm it met product specifications and its overall ability to meet user needs was validated. However, specific, quantitative acceptance criteria for metrics like accuracy, sensitivity, specificity, etc., are not explicitly defined in the provided text. Similarly, the reported numerical device performance against such criteria is not detailed. The document broadly states that "The accuracy and precision of device measurements was assessed using purpose-built phantoms and in-vivo acquired data from volunteers."

Missing Information: A detailed table of acceptance criteria with numerical targets and the corresponding reported device performance values.

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

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

• Test Set Sample Size: The document mentions "in-vivo acquired data from volunteers" and that "Volunteers participating in the performance testing were representative of the intended patient population." However, the specific number of cases or volunteers used in the test set is not provided.
• Data Provenance: The document does not explicitly state the country of origin of the data or whether it was retrospective or prospective. It only mentions "in-vivo acquired data from volunteers," implying prospectively collected data for assessment, but this is not explicitly stated.

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

This information is not provided in the document. The text indicates that the device produces quantitative metrics that are "interpreted by a licensed physician" to "assist in diagnosis, clinical management and monitoring of patients." However, it does not describe how ground truth was established for the performance testing, nor the number or qualifications of experts involved in that process.

4. Adjudication method for the test set

This information is not provided in the document. There is no mention of an adjudication process (e.g., 2+1, 3+1, none) for the test set's ground truth.

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 document does not indicate that a MRMC comparative effectiveness study was performed. The device is described as a "post-processing software system" that provides "quantified metrics" and does not describe AI assistance for human readers in a diagnostic workflow. The primary method of performance assessment mentioned is the accuracy and precision of the measurements themselves using phantoms and volunteer data, not reader performance.

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

The performance testing described involves the device generating quantitative metrics. The phrase "The accuracy and precision of device measurements was assessed" suggests a standalone performance assessment of the algorithm's output (measurements) against some reference (phantoms, in-vivo data). While the final interpretation is by a physician, the core performance reported relates to the device's ability to produce these measurements consistently and accurately, which aligns with a standalone assessment of the algorithms.

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

The ground truth for the "accuracy and precision of device measurements" was established using:

• Purpose-built phantoms: These phantoms "containing vials with different relaxation times corresponding to the physiological ranges of tissue values." This provides a known, controlled physical ground truth.
• In-vivo acquired data from volunteers: For this type of data, the text does not specify how ground truth was established (e.g., through other validated methods, clinical outcomes, or expert consensus on a final diagnosis based on all available information). It only mentions that the studies "covering a range of physiological values for cT1, T1 and PDFF."

8. The sample size for the training set

This information is not provided in the document. The document describes CoverScan v1 as software that takes acquired MR data and processes it; it does not detail any machine learning training processes or associated datasets.

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

As there is no mention of a specific training set in the provided text, the method for establishing its ground truth is also not described. The document implies that the device is a measurement and processing tool rather than a machine learning model that requires a dedicated training set.

Ask a specific question about this device

Hepatica (Hepatica v1) is a post-processing medical device software that presents quantified metrics which may contribute to the assessment of a patient's liver health.

Hepatica (Hepatica v1) uses image visualisation and analysis tools to process DICOM 3.0 compliant magnetic resonance image datasets to produce semi-automatic segmented 3D models of the work of Couinaud and the Brisbane 2000 terminology. For each identified Couinaud segment, volumetric data is determined and reported.

Hepatica v1) may also report iron corrected-T1 (cT1) and PDFF calculated using the IDEAL method from multi-slice acquisitions, on a per segment basis, over the whole liver. Both metrical values of different fundamental liver tissue characteristics that can be used as measures of liver tissue health.

Hepatica (Hepatica v1) provides trained clinicians with additional information to evaluate the volume and health of a patient's liver on a segmental basis. It is not intended to replace the established procedures for the assessment of a patient's liver health. However, information gathered through existing diagnostic tests, clinical evaluation of the patient, as well Hepatica (Hepatica v1), may support surgical decision making.

Hepatica v1 is a standalone software device that imports MR datasets encompassing the abdomen, including the liver. Visualisation and display of T1-weighted MR data which can be analysed, and quantitative metrics of tissue characteristics and liver volume are then reported. Datasets imported into Hepatica are DICOM 3.0 compliant and reported metrics are independent of the MRI equipment vendor. It allows for the 3D visualisation of the liver and quantification of metrics (cT1, PDFF and volumetry) from liver tissue and exportation of results and images to a deliverable report. Hepatica v1 supports semi-automatic liver segmentation of T1-weighted volumetric data. Liver segmentation in Hepatica v1 requires the placement of anatomical landmarks to define the outer contours of the liver and can be adjusted by the operator, where necessary. Where available, whole liver and segmental cT1 and PDFF quantitative metrics derived from the predicate device may be presented in the final report. Hepatica uses volumetric datasets to create 2D anatomical views from all supported scanners. Where available, cT1 and PDFF parametric maps are derived from the predicate device. Quantified metrics and images derived from the analysis of liver volume and tissue characteristics are collated into a report for evaluation and interpretation by a clinician.

Here's a breakdown of the acceptance criteria and study information for the Hepatica (Hepatica v1) device, based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly state acceptance criteria in a quantitative format (e.g., "accuracy must be >X%"). Instead, it states that the performance testing "demonstrates that Hepatica v1 is at least as safe and effective as the predicate device and does not introduce any new risks."

However, it provides "Upper and Lower Limits of Agreement" from a Bland-Altman analysis, which can be interpreted as the range within which the device's measurements agree with the gold standard. A tighter range indicates higher accuracy. The precision results are also presented as "Upper and Lower limits of Agreement" for repeatability and reproducibility.

The table below summarizes the reported device performance, which implies that these values met the internal acceptance parameters for demonstrating substantial equivalence:

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

• Sample Size for Test Set: The document states that the performance testing used "previously acquired in-vivo data from healthy and non-healthy volunteers." It also mentions "Volunteers participating in the performance testing were representative of the intended patient population." However, a specific number for the sample size (N) of these volunteers or images in the test set is not provided.
• Data Provenance: "Previously acquired in-vivo data." The country of origin is not explicitly stated. It can be inferred that the data is likely from the UK, given the submitter's address (Oxford, UK). The data is retrospective, as it was "previously acquired."

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

• Number of Experts: The document refers to the gold standard as "radiologists." It does not specify the number of radiologists involved in establishing the ground truth.
• Qualifications of Experts: The qualification is "radiologists." No further details on their years of experience or sub-specialty are provided.

4. Adjudication Method for the Test Set

The document states that the "gold standard" is "radiologists." It does not describe any specific adjudication method (e.g., 2+1, 3+1 consensus) used to establish this ground truth. It implies that the radiologists' readings were considered the definitive truth.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, and the Effect Size

No, a Multi-Reader Multi-Case (MRMC) comparative effectiveness study was not done. The study compared the device's measurements to a "gold standard" (radiologists' assessments), not the performance of human readers with vs. without AI assistance. Therefore, there is no effect size reported for human readers improving with AI assistance.

6. If a Standalone Study Was Done

Yes, a standalone (algorithm only without human-in-the-loop performance) study was done. The performance testing section directly reports on the "Accuracy" and "Precision" of Hepatica v1's measurements (cT1, PDFF, and volumetry) when compared to the gold standard. The device operators are "trained Perspectum internal operators," but the reported metrics are explicitly from the device's output. The statement "The variation introduced by operator measurements are well within the acceptance criteria" also suggests an understanding of the device's standalone performance separate from human interpretation of the reports.

7. The Type of Ground Truth Used

The primary ground truth used for accuracy comparison is expert consensus/interpretation, specifically "radiologists." For cT1 and PDFF, these are quantitative measurements derived from imaging, which radiologists would interpret or measure. For volumetry, it's also based on radiological assessment.

8. The Sample Size for the Training Set

The document does not provide the sample size for the training set. It only mentions the "previously acquired in-vivo data from healthy and non-healthy volunteers" used for performance testing.

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

Since the training set sample size and its specifics are not mentioned, how its ground truth was established is not described in the provided document.

Ask a specific question about this device

LiverMultiScan (LMSv4) is indicated for use as a magnetic device software application for noninvasive liver evaluation that enables the generation, display and review of 2D magnetic resonance medical image data and pixel maps for MR relaxation times.

LiverMultiScan (LMSv4) is designed to utilize DICOM 3.0 compliant magnetic resonance image datasets, acquired from compatible MR systems, to display the internal structure of the abdomen including the liver. Other physical parameters derived from the images may also be produced.

LiverMultiScan (LMSv4) provides a number of tools, such as automated liver segmentation and region of interest (ROI) placements, to be used for the assessment of selected regions of an image. Quantitative assessments of selected regions include the determination of triglyceride fat fraction in the liver (PDFF), T2* and iron-corrected T1 (cT1) measurements. T2* may be optionally computed using the DIXON or LMS MOST methods.

These images and the physical parameters derived from the images, when interpreted by a trained clinician, yield information that may assist in diagnosis.

LiverMultiScan is a standalone software device. The purpose of the LiverMultiScan device is to assist a trained operator with the evaluation of information from Magnetic Resonance (MR) images from a single time-point (a patient visit). LiverMultiScan is a post-processing software device, a trained operator uses tools such as automatic liver segmentation and region of interest placement upon previously acquired MR images, from which a summary report is generated. The summary report is subsequently sent to an interpreting clinician at the acquiring site.

LiverMultiScan is not intended to replace the established procedures for the assessment of a patient's liver health by an interpreting clinician, providing many opportunities for competent human intervention in the interpretation of images and information displayed.

The metrics are intended to be used as an additional diagnostic input to provide information to clinicians as part of a wider diagnostic process. It is expected that in the normal course of liver disease diagnosis, patients with clinical symptoms or risk factors which may indicate liver disease. The interpreting clinician needs to take into consideration the device's limitations and accuracy during clinical interpretation.

Liver function tests, blood tests, ultrasound scanning as well as liver biopsy are all expected to be used at the discretion of a qualified clinician in addition to information obtained from the use of LiverMultiScan metrics. The purpose of LiverMultiScan metrics is to provide imaging information to assist in characterizing tissue in the liver, in addition to existing methods for obtaining information relating to the liver. LiverMultiScan metrics are not intended to replace any existing diagnostic source of information but can be used to identify patients who may benefit most from further evaluation, including biopsy.

Information gathered through existing diagnostic tests and clinical evaluation of the patient, as well as information obtained from LiverMultiScan metrics, may contribute to a diagnostic decision.

LiverMultiScan is not a computer-aided diagnostic device and can only present imaging information which must be interpreted by a qualified clinician. LiverMultiScan is an aid to diagnosis and treatment decisions remains the responsibility of the clinician.

In consequence, the product is considered to have no adverse effect on health since the results represent only a part of the information that the user will utilize for final interpretation. In this regard, LiverMultiScan presents a moderate level of concern with respect to patient safety.

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

1. Table of Acceptance Criteria and Reported Device Performance

The acceptance criteria are implicitly defined by the performance test results demonstrating acceptable levels of accuracy, repeatability, and reproducibility. The document states that the observed variations were "well within the prescribed acceptance criteria," but the numerical values for the criteria themselves are not explicitly listed. Instead, the reported performance values are provided.

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

• Sample Size for Test Set: The document does not explicitly state the numerical sample size for the test set. For the bench testing, it mentions "purpose-built phantoms" and for clinical testing, "in-vivo acquired volunteer data." The number of phantoms or volunteers is not specified.
• Data Provenance:
  • Bench Testing: Data was acquired from "purpose-built phantoms." The country of origin is not specified, but the applicant (Perspectum Ltd) is based in the UK.
  • Clinical Testing: "in-vivo acquired volunteer data" from "volunteers participating in the performance testing were representative of the intended patient population." The country of origin for these volunteers is not specified. The study appears to be prospective for the gathered "volunteer data" in the sense that data was acquired for the purpose of testing, but the details are sparse.

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

The document does not mention the use of experts to establish ground truth for the test set.

• Bench Testing: Ground truth for phantom accuracy was based on the "gold standard" to which the LMSv4 values were compared. This "gold standard" is implied to be the known physical properties or reference measurements of the phantoms.
• Clinical Testing: The clinical performance relates to repeatability and reproducibility of the device's measurements themselves, not against a human-established ground truth. Operator variability was assessed, implying trained operators were involved, but their qualifications and number are not detailed beyond "trained internal Perspectum operators."

4. Adjudication Method for the Test Set

No adjudication method for the test set is mentioned. The performance testing focuses on the device's intrinsic accuracy, repeatability, and reproducibility, rather than agreement with human interpretation that would typically require adjudication.

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

No MRMC comparative effectiveness study was done. The document states: "No clinical investigations or studies were conducted during performance testing of LMSv4." The device is positioned as an "aid to diagnosis" and "not intended to replace any existing diagnostic source of information," suggesting its role is as a tool rather than a standalone diagnostic.

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

Yes, a standalone performance study was clearly done. The "Performance Testing - Bench" section directly assesses the "accuracy and precision of device measurements" using phantoms. The "Performance Testing - Clinical" section assesses the "precision of LMSv4 measurements" (repeatability and reproducibility) using volunteer data. While operator variability was assessed, the core measures are of the algorithm's output. The device itself is described as a "post-processing software device" where an "operator uses tools such as automatic liver segmentation and region of interest placement," implying the algorithm performs the quantification and the operator interfaces with it.

7. Type of Ground Truth Used

• Bench Testing: The ground truth for accuracy was established using a "gold standard" derived from the "purpose-built phantoms" that contained vials with known relaxation times.
• Clinical Testing: For repeatability and reproducibility, the "ground truth" is essentially the device's own consistent measurement across repeated scans, different scanners/field strengths, and different operators. It is not externally validated against pathology or clinical outcomes in this document.

8. Sample Size for the Training Set

The document does not provide any information regarding the sample size for the training set. It focuses solely on the performance testing of the device.

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

The document does not provide any information regarding the training set or how its ground truth was established, as it does not describe the development or training of the LMSv4 algorithm. The emphasis is on proving substantial equivalence to a predicate device through performance benchmarking.

Ask a specific question about this device