Search Results

Al-Rad Companion (Pulmonary) is image processing software that provides quantitative and qualitative analysis from previously acquired Computed Tomography DICOM images to support radiologists and physicians from emergency medicine, specialty care, urgent care, and general practice in the evaluation and assessment of disease of the lungs. It provides the following functionality:

· Segmentation and measurements of complete lung and lung lobes

· Identification of areas with lower Hounsfield values in comparison to a predetined threshold for complete lung and lung lobes

· Providing an interface to external Medical Device syngo.CT Lung CAD

· Segmentation and measurements of found lung lesions and dedication to corresponding lung lobe.

· Identification of areas with elevated Hounsfield values. where areas with elevated versus high opacities are distinguished.

The software has been validated for data from Siemens Healthineers (filtered backprojection and iterative reconstruction), GE Healthcare (filtered backprojection reconstruction), and Philips (filtered backprojection reconstruction).

Only DICOM images of adult patients are considered to be valid input.

Device Description

Al-Rad Companion is a software only medical system that investigates data from imaging systems. Al-Rad Companion receives these data and checks which post-processing algorithms may be applicable. Data that does not meet the Al-Rad Companion requirements are ignored while data that meets the requirements are sent for further processing. Applicable data are processed, and the results are provided to the user via their clinical workplace. The user has the option to accept, review or withdraw single results of Al-Rad Companion.

Al-Rad Companion includes a software operating platform (Al-Rad Companion (Engine)) and optional clinical extensions such as Al-Rad Companion (Pulmonary), Al-Rad Companion (Musculoskeletal) and Al-Rad Companion (Cardiovascular). The clinical extensions are post-processing applications that operate on the Al-Rad Companion (Engine) software platform and process CT datasets in specific regions of the thorax or use datasets from other modalities. The basic post-processing functions are landmark detection, segmentation, and classification. Al-Rad Companion uses Artificial Intelligence (Al)algorithms.

The Al-Rad Companion (Engine) platform is the interface for incoming and outgoing data for the complete Al-Rad Companion system that provides input data and collects results and status information from the extensions. Additionally, it is the interface for incoming and outgoing data for the complete Al-Rad Companion system.

The Al-Rad Companion extensions are optional post-processing applications that operate on the Al-Rad Companion (Engine) software platform. The platform and each of the extensions are distinct software components and thus separate medical devices.

The scope of this submission is the extension Al-Rad Companion (Pulmonary). It is an image postprocessing software that uses CT DICOM data to support clinicians in the evaluation and assessment of lung diseases. It utilizes machine-learning and deep-learning algorithms to provide quantitative and qualitative analysis from previously acquired Computed Tomography DICOM images to support qualified clinicians in the evaluation and assessment of disease of the major functionalities of Al-Rad Companion (Pulmonary) are as follows:

Segmentation and measurements of complete lung, lungs, and lung lobes.
ldentification of areas with lower Hounsfield values in comparison to a predefined threshold for . complete lung and lung lobes.
. Segmentation and measurements of found lung lesions
. Identification of areas with elevated Hounsfield values, where areas with elevated versus high opacities are distinguished

The results will be delivered in different image formats and, depending on the configuration, can be verified in the Results Preview and will be included in the overview with all findings. This will include DICOM Structured Report with measurements results

The software version VA13 of the Al-Rad Companion (Pulmonary) includes the following modifications:

Pulmonary Density: This feature provides the possibility to segment opacity regions inside the lung using an Al algorithm. Al-Rad Companion (Pulmonary) counts image voxels inside opacity regions and calculates the percentages of these voxels relative to the total number of voxels per lobe, lung and in total. Afterwards, the opacity results are assigned to a certain range as defined by Bernheim et al.
Bi-directional lesion diameter: This feature provides an additional measurement derived from the existing segmentation contour of a lung lesion. The existing list of measurements is extended with the maximum orthogonal diameter in 2D (short axis diameter) which is orthogonal to the lesion's maximum 2D diameter (2D diameter, long axis diameter).
Cloud and Edge Deployment: The system supports the existing cloud deployment as well as a new edge deployment. The system remains hosted in the teamplay digital health platform and remains driven by the Al-Rad Companion (Engine). Now the edge deployment allows the processing of clinical data and the generation of results on-premises within the customer network. The edge system is fully connected to the cloud for monitoring and maintenance of the system from remote.

AI/ML Overview

Here's a breakdown of the acceptance criteria and study details for Al-Rad Companion (Pulmonary) based on the provided document:

1. Table of Acceptance Criteria and Reported Device Performance

Feature/Metric	Acceptance Criteria (Implied)	Reported Device Performance
Lung Lobe Segmentation	Clinically acceptable segmentation accuracy and equivalence to predicate.	Average DICE coefficients ranged from 0.94 to 0.96 (for 250 datasets from US and Europe). "Demonstrated equivalent performance in comparison to the primary predicate device for segmentation." Consistency across population-specific subgroups and technical parameters.
Opacity Detection (PO values)	Clinically acceptable agreement with human reads during inter-reader variability assessment and equivalence to predicate.	95%-Limits of Agreement (LoA) were established against human reads. 93.0% of the PO (percentage of opacity) values were found within the LoA (for 150 datasets from US and Europe). "Demonstrated equivalent performance in comparison to the primary predicate device for lung parenchyma categorization." Consistency across population-specific subgroups and technical parameters.
Overall Software Performance	Device performs as intended, all software specifications met.	Non-clinical tests (integration and functional) were conducted, and the results "demonstrate that the subject device performs as intended." "The results of all conducted testing was found acceptable to support the claim of substantial equivalence." "The risk analysis was completed, and risk control implemented to mitigate identified hazards. The testing results demonstrate that all the software specifications have met the acceptance criteria. Testing for verification and validation of the device was found acceptable to support the claims of substantial equivalence."

Note: The document mainly focuses on proving substantial equivalence to a predicate device, thus the "acceptance criteria" are implied to be achieving performance comparable to, or improving upon, the predicate. Specific numerical thresholds for acceptance criteria are not explicitly stated, but are inferred from the reported performance which is deemed acceptable for substantial equivalence.

2. Sample Size for the Test Set and Data Provenance

Lung Lobe Segmentation: 250 datasets
Opacity Detection: 150 datasets
Data Provenance: Multiple sites across the US and Europe. The document states this was a "Clinical Data Based Software Validation." It does not explicitly state if it was retrospective or prospective, but clinical validation of existing images typically suggests a retrospective approach.

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

The document does not specify the number of experts or their qualifications for establishing ground truth for the test set. It mentions "human reads" for comparison in the opacity detection study, implying human experts were involved, but details are not provided.

4. Adjudication Method for the Test Set

The document does not explicitly state the adjudication method (e.g., 2+1, 3+1, none) used for the test set. For opacity detection, "Interreader-variability of the percentage of opacity (PO) was assessed" against "human reads," but the specific process of how those "human reads" were finalized as ground truth is not detailed.

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

No multi-reader multi-case (MRMC) comparative effectiveness study demonstrating human readers improve with AI vs. without AI assistance is explicitly described. The studies focus on the standalone performance of the AI algorithm against human reads or a "primary predicate device."

6. Standalone (Algorithm Only) Performance Study

Yes, a standalone performance study was done. The "Clinical Evaluation of the AI-based Algorithms" section details the validation of the lung lobe segmentation and opacity detection algorithms. The reported DICE coefficients and the LoA for PO values are measures of the algorithm's performance independent of human-in-the-loop interaction in the context of the reported studies.

7. Type of Ground Truth Used

The ground truth used for the opacity detection algorithm was based on "human reads" during an inter-reader variability assessment. For lung lobe segmentation, while not explicitly stated, it is commonly established through expert annotations. The phrasing "Description of ground truth / annotations generation" indicates that such ground truth was generated, likely by experts.

8. Sample Size for the Training Set

The document mentions "Training cohort: size and properties of data used for training" under the "Clinical Data Based Software Validation" section but does not provide the specific sample size for the training set. It states "Additional training data was added as compared to the primary predicate for the Pulmonary Density Feature."

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

The document states "Description of ground truth / annotations generation" for the training cohort, implying that ground truth was established through annotations, most likely by clinical experts. However, specific details about the process (e.g., number of annotators, their qualifications, adjudication) are not provided for the training set.

Ask a Question

Ask a specific question about this device

Page 1 of 1