(157 days)
Meuro Quant is intended for automatic labeling, visualization and volumetric quantification of segmentable brain structures and lesions from a set of MR images. Volumetric measurements may be compared to reference percentile data.
NeuroQuant is a fully automated MR imaging post-processing medical device software that provides automatic labeling, visualization and volumetric quantification of brain structures and lesions from a set of MR images and returns segmented images and morphometric reports. The resulting output is provided in a standard DICOM format as additional MR series with segmented color overlays and morphometric reports that can be displayed on third-party DICOM workstations and Picture Archive and Communications Systems (PACS). The high throughput capability makes the software suitable for use in both clinical trial research and routine patient care as a support tool for clinicians in assessment of structural MRIs.
NeuroQuant provides morphometric measurements based on 3D T1 MRI series. The output of the software includes volumes that have been annotated with color overlays, with each color representing a particular segmented region, and morphometric reports that provide comparison of measured volumes to age and gender-matched reference percentile data. In addition, the adjunctive use of the T2 FLAIR MR series allows for improved identification of some brain abnormalities such as lesions, which are often associated with T2 FLAIR hyperintensities.
The NeuroQuant processing architecture includes a proprietary automated internal pipeline that performs artifact correction, atlas-based segmentation, volume calculation and report generation.
Additionally, automated safety measures include automated quality control functions, such as tissue contrast check, atlas alignment check and scan protocol verification, which validate that the imaging protocols adhere to system requirements.
From a workflow perspective, NeuroQuant is packaged as a computing appliance that is capable of supporting DICOM file transfer for input and output of results.
The provided text describes the 510(k) summary for the NeuroQuant device (K170981). Here's a breakdown of the acceptance criteria and the study proving the device meets them, based on the provided document:
1. Table of Acceptance Criteria and Reported Device Performance
The acceptance criteria are implied by the performance statistics reported. While explicit acceptance thresholds are not given in a "PASS/FAIL" format, the document presents quantitative results from the performance testing.
| Acceptance Criteria (Implied) | Reported Device Performance |
|---|---|
| Segmentation Accuracy (Dice's Coefficient): | |
| - Major Subcortical Structures (compared to expert manual) | In the range of 80-90% |
| - Major Cortical Regions (compared to expert manual) | In the range of 75-85% |
| - Brain Lesions (T1 and T2 FLAIR, compared to expert manual) | Exceeds 80% |
| Segmentation Reproducibility (Percentage Absolute Volume Differences): | |
| - Major Subcortical Structures (repeated T1 MRI scans) | Mean percentage absolute volume differences were in the range of 1-5% |
| - Brain Lesions (repeated T1 and T2 FLAIR MRI scans) | Mean absolute lesion volume difference was less than 0.25cc, while the mean percentage lesion absolute volume difference was less than 2.5%. |
2. Sample Size Used for the Test Set and Data Provenance
The document does not explicitly state the sample size used for the test set. It mentions "3D T1 MRI scans" and "3D T1 and T2 FLAIR MRI scan pairs of subjects with brain lesions" were used for evaluation.
The document 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 Qualifications of Those Experts
The document states that segmentation accuracy was evaluated by "comparing segmentation accuracy with expert manual segmentations." However, it does not specify the number of experts used or their qualifications (e.g., radiologist with 10 years of experience).
4. Adjudication Method for the Test Set
The document mentions "expert manual segmentations" as the ground truth, but it does not describe any adjudication method (e.g., 2+1, 3+1, none) used to establish this ground truth among multiple experts if more than one was involved.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done
The document does not mention a multi-reader multi-case (MRMC) comparative effectiveness study or any effect size of how human readers improve with or without AI assistance. The performance testing focuses solely on the device's accuracy and reproducibility against manual segmentation and repeated scans.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done
Yes, a standalone performance evaluation was done. The "Performance Testing" section describes how "NeuroQuant performance was evaluated by comparing segmentation accuracy with expert manual segmentations and by measuring segmentation reproducibility between same subject scans." This refers to the algorithm's performance directly, independent of a human reader's interaction with the output for primary diagnosis.
7. The Type of Ground Truth Used
The ground truth used for the segmentation accuracy evaluation was "expert manual segmentations." For reproducibility, the ground truth was the measurements from repeated scans of the same subjects, with the expectation that the device produces consistent results on these repeated scans.
8. The Sample Size for the Training Set
The document does not specify the sample size used for the training set. It describes the device's "proprietary automated internal pipeline that performs... atlas-based segmentation," and "dynamic probabilistic neuroanatomical atlas, with age and gender specificity." This implies a trained model, but the size of the dataset used for this training is not disclosed.
9. How the Ground Truth for the Training Set Was Established
The document states the device uses "atlas-based segmentation" and a "dynamic probabilistic neuroanatomical atlas, with age and gender specificity." This suggests the training involves the creation or utilization of an anatomical atlas, which typically involves expert anatomical labeling and segmentation of a representative set of MR images to build probabilities for different brain regions. However, the specific methodology for establishing this ground truth for the training set (e.g., number of experts, their qualifications, adjudication) is not detailed in this summary.
§ 892.2050 Medical image management and processing system.
(a)
Identification. A medical image management and processing system is a device that provides one or more capabilities relating to the review and digital processing of medical images for the purposes of interpretation by a trained practitioner of disease detection, diagnosis, or patient management. The software components may provide advanced or complex image processing functions for image manipulation, enhancement, or quantification that are intended for use in the interpretation and analysis of medical images. Advanced image manipulation functions may include image segmentation, multimodality image registration, or 3D visualization. Complex quantitative functions may include semi-automated measurements or time-series measurements.(b)
Classification. Class II (special controls; voluntary standards—Digital Imaging and Communications in Medicine (DICOM) Std., Joint Photographic Experts Group (JPEG) Std., Society of Motion Picture and Television Engineers (SMPTE) Test Pattern).