Search Results
Found 1 results
510(k) Data Aggregation
(148 days)
NinesMeasure is a semi-automatic tool indicated for use by trained radiologists to aid in the analysis and review of adult thoracic CT images. NinesMeasure provides quantitative information about pulmonary nodule size on a single study or over the time course of several thoracic studies by providing long and short axis diameter measurements in the axial plane.
Based on analysis of DICOM images and provided input from a radiologist, indicating the location of the pulmonary nodule, the device uses artificial intelligence algorithms to automatically perform the measurements, and allows the axial measurements to be displayed and reviewed. NinesMeasure is limited for use on solid pulmonary nodules.
The device is intended to be used as a measurement tool by a trained radiologist and is limited to analysis of imaging data and should not be used in-lieu of full patient evaluation or relied upon to make or confirm a diagnosis. The device does not alter the original medical image.
NinesMeasure is a semi-automatic, diagnostic patient imaging tool used to measure the size of selected pulmonary nodules in a radiological image. The software system is comprised of a set of software modules for performing image analysis at a specified image location to calculate measurements of pulmonary nodules on adult thoracic CT images. The system operates over a standard network interface and receives the DICOM images and coordinates of the pulmonary nodule to measure. The system then returns the measurements for the long and short axis diameters for review by a trained radiologist.
NinesMeasure is designed to be used with a standard PACS, where the user can indicate a location of the pulmonary nodule to measure, and then review and edit the measurements on the DICOM image.
The image analysis uses Artificial Intelligence (Al) technology to analyze chest CT images for computing the measurements. Specifically, the device utilizes a machine learning (ML) algorithm to compute segmentations of nodules, from which the long and short axis measurements are then calculated.
The provided document describes the NinesMeasure device, a semi-automatic tool for measuring pulmonary nodule size on adult thoracic CT images. The document includes information about the device's performance testing.
Here's a breakdown of the acceptance criteria and the study that proves the device meets them:
1. Table of Acceptance Criteria and Reported Device Performance
The document presents primary endpoints related to the normalized error in measuring nodule diameters. While explicit "acceptance criteria" (thresholds the device must meet to be considered effective) are not stated numerically, the narrative confirms that "The performance goals for the primary endpoints were met." This implies internal acceptance criteria for these normalized error values.
| Primary Endpoint - All Nodules | Reported Device Performance [95% CI] | Implied Acceptance Criterion (Met) |
|---|---|---|
| Normalized error on long axis diameter | 0.113 [Upper Bound 0.124] | < 0.124 (or similar, met) |
| Normalized error on short axis diameter | 0.131 [Upper Bound 0.143] | < 0.143 (or similar, met) |
Stratified by Nodule Size:
| Nodule Size | Number of Nodules | Reported Normalized error on Long Axis Diameter [95% CI] | Reported Normalized error on Short Axis Diameter [95% CI] | Implied Acceptance Criterion (Met) for Long Axis | Implied Acceptance Criterion (Met) for Short Axis |
|---|---|---|---|---|---|
| 3-6 mm | 100 | 0.104 [Upper Bound: 0.119] | 0.123 [Upper Bound: 0.139] | < 0.119 (or similar, met) | < 0.139 (or similar, met) |
| 6-8 mm | 63 | 0.119 [Upper Bound: 0.138] | 0.143 [Upper Bound: 0.161] | < 0.138 (or similar, met) | < 0.161 (or similar, met) |
| 8-10 mm | 46 | 0.13 [Upper Bound: 0.156] | 0.133 [Upper Bound: 0.166] | < 0.156 (or similar, met) | < 0.166 (or similar, met) |
2. Sample Size and Data Provenance for the Test Set
- Sample Size for Test Set:
- All Nodules: The total number of nodules can be inferred by summing the stratified categories: 100 (3-6mm) + 63 (6-8mm) + 46 (8-10mm) = 209 nodules.
- The study used cases that were "diverse, and included 3 different major scanner manufacturers, 7 different scanner models, 11 different clinical sites."
- Data Provenance: The study was a retrospective, multi-center image comparison study. The document does not specify the country of origin for the data.
3. Number of Experts and Qualifications for Ground Truth
The document does not explicitly state the "number of experts used to establish the ground truth" or their specific qualifications (e.g., "Radiologist with 10 years of experience"). However, the device "is limited for use by trained radiologists" and relies on "provided input from a radiologist, indicating the location of the pulmonary nodule." This strongly suggests that ground truth measurements would likely be derived from expert radiologist annotations or measurements.
4. Adjudication Method for the Test Set
The document does not describe the specific adjudication method (e.g., 2+1, 3+1) used to establish the ground truth measurements for the test set.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
An MRMC study was not explicitly mentioned or detailed. The study described ("retrospective, multi-center image comparison study") evaluates the device's performance in measuring nodules, but it doesn't compare human readers with and without AI assistance to quantify improvement. The device is a "semi-automatic tool" and requires "provided input from a radiologist," implying a human-in-the-loop design, but the study focuses on the device's measurement accuracy after radiologist input, not the radiologist's performance change.
6. Standalone Performance (Algorithm Only)
The study describes the performance of the NinesMeasure device, which "uses artificial intelligence algorithms to automatically perform the measurements, and allows the axial measurements to be displayed and reviewed."
Given the description "semi-automatic tool" and "provided input from a radiologist, indicating the location of the pulmonary nodule," the reported performance metrics likely represent the algorithm's performance after the initial human input (locating the nodule). It does not appear to be a purely standalone (algorithm-only, end-to-end detection and measurement) performance evaluation. The algorithm performs the measurement after the radiologist pinpoints the nodule.
7. Type of Ground Truth Used
Based on the context of measuring pulmonary nodule size, the ground truth would most likely be expert consensus measurements (manual measurements performed by radiologists or other imaging experts) of the nodule diameters. Pathology or outcomes data are not typically used for direct measurement accuracy ground truth in this context.
8. Sample Size for the Training Set
The document does not specify the sample size used for the training set. It only describes the validation (test) set.
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. Given the nature of a machine learning (ML) algorithm for segmentation and measurement, it would typically involve expertly annotated images (manual segmentations or measurements) on a separate dataset from the test set.
Ask a specific question about this device
Page 1 of 1