AI-Rad Companion - Brain MR - Data sheet

Data sheet AI-Rad Companion Brain MR Version VA51 Valid from January 2025 siemens-healthineers.com/ai-rad SIEMENS Healthineers AI-Rad Companion Brain MR VA51 · Data sheet The AI-Rad Companion product family provides automated post-processing of medical images by integrating Artificial Intelligence (AI) algorithms into the clinicians workflow. The results generated by AI-Rad Companion can then be sent to predefined systems that can receive DICOM objects. AI-Rad Companion supports cloud-based and on-edge deployment options. Thus enabling the institution to select the option that better suits their requirements. In many countries, the number of imaging Advanced image processing tools in AI-Rad Companion examinations is steadily increasing, but the number of Brain MR support clinicians in analyzing and processing experts is not growing proportionally. This results in a medical images. dramatic increase in the workload per radiologist. The effect of this is that shorter reading times can have a The AI-Rad Companion Brain MR automatically prepares negative impact on the quality of the reports 1 valuable information for clinicians in order to facilitate the . subsequent image interpretation workflow. AI powered Al-Rad Companion Brain MR on teamplay Cloud Platform or dge deployment MRI Input Confirmation UI Image Series MRI Scanner Functionalities Automated Image Aquisition results in PACS to Quantification Labels and assist clinicians Map DICOM SR The AI-Rad Companion Brain MR is a post-processing The Results Preview function displays a preview of image analysis software that assists clinicians in automatically generated results and allows basic viewing, analyzing, and evaluating MR brain images viewing tasks, such as windowing, zooming, and for the assessment of the volumetric properties of a panning. variety of brain structures and white matter hyperintensities (WMH). All displayed results can be confirmed or rejected before they are passed on to a picture archiving The longitudinal assessment of Brain MR images system or an image reading system. from two timepoints provides the rate of change of volumes of brain structures and the count and Optionally, the clinical user can be informed about volume of new or enlarged WMH. the availability of results by using the AI-Rad Companion Notifier. The pop-up message from the Clinicians can configure the workflow within AI-Rad notifier indicates whether the processing of cases is Companion Brain MR. Clinicians can define whether completed, processing is in progress, or processing the results require manual confirmation or if the has failed. images will be sent to the Picture Archiving and Communication System (PACS) automatically. AI-Rad Companion Brain MR supports cloud-based and on-edge deployment options. The user interface can only be accessed with a web browser and optionally, from the AI-Rad Companion Notifier component. 1 Berlin L: Faster Reporting Speed and Interpretation Errors: Conjecture, Evidence, and Malpractice Implications. J Am Coll Radiol. 2015 2 Sep;12(9):894-6. doi: 10.1016/j.jacr.2015.06.010 Data sheet · AI-Rad Companion Brain MR VA51 General features of AI-Rad Companion Brain MR althineers Al-Rad Companion R DS ? Brain MR O Al Results 81 / 160 Deviations Patient Information Examination Details Deviations Normalized %@ Percentile 7th - 93th percentiles ] Cerebellum 104.8 6.68 [6.91 - 8.85] Ventricles 158.9 >99 10.13 [1.64 - 5.17] Lateral Ventricle Left 81.8 >99 5.21 [0.67 - 2.44] Lateral Ventricle Right 66.2 >99 4.21 [0.66 - 2.35] Ln 3rd ventricle 52 98 0.33 [0.12-0.28] Ith ventricle 5.8 >99 0.37 [0.12 - 0.27] E Temporal left 24.6 1.57 [1.59 - 2.11] 14.0 0.89 [0.96 - 1.30] AI Results E'I Morphometry to PACS Deviation Map Label Map Legend = Deviations General Info 7th and 93th percentiles of healthy age-/gender-matched population Decline All Send to PACS Show All Contours Figure 2: Morphometry Deviation Map SIEMENS Healthineers Al-Rad Companion DS ? Brain MR 4 Al Results 89 / 176 Report Patient Information All Examination Details Global Values Volume Percentile : Normalized % Rate of Change [ml] [ 11th - 89th percentiles ] [%/yT] ---- Total Intracranial 1,424.2 [0.00 - 0.00] -1.21 Volume Total Brain Parenchyma 1,169.6 97 82.10 [71.90 - 79.80] -1.03 E Total Gray Matter 682.7 84 47.90 [42.10 - 48.60] -0.83 E'I Total White Matter 486.9 95 34.20 [27.80 - 33.20] -1.31 AI Results Morphometry TO PACS Cerebellum 149.3 > 99 10.49 [7.49 - 9.20] 0.26 Deviation Map Label Map Cerebral Spinal Fluid 254.6 <1 17.90 [20.40 - 26.70] -2.01 Legend = Deviations General Info Decline All Send to PACS Show All Contours 11th and 89th percentiles of healthy age-/gender-matched population Figure 2: Morphometry Follow-Up 3 AI-Rad Companion Brain MR VA51· Data sheet SIEMENS. Healthineers Al-Rad Companion DS ? 0 ₽ Brain MR O 89 / 176 Quality Information @ Comments Values Acceptable Range @ Patient Information Current ExamInation Detalls Image Quality Not Assessed [n.a.J -- Segmentation Quality Acceptable Contours Structures Lobar Gray Matter Lobar White Matter Cerebral Spinal Fluid AIB AI Results White Matter Abnormalities (WMAb) Morphometry TO PACS Deviation Map Thalamus - Left + Right Label Map Lateral Ventricle - Left White Matter Hyperintensities Lateral Ventricle - Right VMH Map 3rd Ventricle 4th Ventricle Caudate - Left + Right Putamen - Left + Right Pallidum - Left + Right Hippocampus - Left + Right Insula - Left + Right Decline Al Send to PACS Show All Contours Pons Figure 3: Morphometry Label Map SIEMENS . Al-Rad Companion DS ? " Healthineers . Brain MR 1 129 / 256 WMH Map Patient Information Spatial Distribution Count Volume (ml) Periventricular 28.09 Juxtacortical 6 0.43 Infratentorial 0.13 Deep white matter 26 0.24 Total 28.88 AI Results Morphometry To PACS Deviation Map Label Map White Matter Hyperintensities WMH Map Caution Input data not consistent with the recommended guideline. Please check the validity of WMH results. Decline All Send to PACS show All Contours Figure 4: White Matter Hyperintensities Map 4 Data sheet · AI-Rad Companion Brain MR VA51 SIEMENS ... Healthineers Al-Rad Companion DS ? Brain MR 138 / 256 Progression Map Patient Information Examination Details Progression Count Volume ( mal ) New or Enlarged 10.69 AI Results Morphometry TO PACS Deviation Map bel Map White Matter Hyperintensities WMH Map Progression Map Decline All Send to PACS Show All Contours Figure 5: Progression Map AI-Rad Companion Brain MR provides automated • Longitudinal Volumetry report: Includes Cross- segmentation and quantitative analysis of individual sectional volumetry reports and rate of change of brain structures and White Matter Hyperintensities volumes for prior and current timepoints. (WMH). • Label Map: Displays each region of the brain in a AI-Rad Companion Brain MR Morphometry feature different color provides automatic volumetric quantification of different brain structures and compares these to • Label Map fused with MPRAGE data image data normative data from a healthy population. (color-coded and partially blended) AI-Rad Companion Brain MR WMH feature includes • Deviation Map: Displays the deviation for each segmentation and quantification of White Matter brain structure from an average age- and gender- Hyperintensities. matched normative volume All generated data can be confirmed from the Results • Deviation Map fused with MPRAGE data (color- Preview and sent to PACS, or rejected. coded and partially blended) AI-Rad Companion Brain MR includes the following • DICOM Structured Reporting output format functions. • WMH map visualizing the white matter Automatic segmentation and quantitative analysis hyperintensities fused with the registered T2 • of individual brain structures FLAIR image (color-coded and partially blended) • Progression map visualizing new or enlarged • Quantitative comparison of each brain structure with normative data from a healthy population white matter hyperintensities Presentation of results for reporting that includes • Cross-sectional WMH Report: Includes table • all numerical values as well as visualization of these summarizing all numeric results, indicating the results total count and volume of the WMH, and the WMH count and volume for each of the four • Cross-sectional Volumetry report: Includes the table subregions periventricular, juxtacortical, of all relevant volumetric tissue class and brain infratentorial, and deep white matter structure values together with the segmentation quality metrics and volume estimates for the • Longitudinal WMH Report: Cross-sectional different brain regions. volumetry reports for prior and current timepoints and the count and volume of new or enlarged WMH regions are provided together. 5 AI-Rad Companion Brain MR VA51 · Data sheet • Volumetric Report: In this report, the numeric results are Brain Morphometry feature supports the images that are listed in tabular format, which includes the volume of acquired from different vendors and does not depend on each brain structure and comparison to the normative any specific vendor DICOM tag 2 . Brain Morphometry range. The report is generated in Secondary Capture and feature can detect >30 different brain structures. Encapsulated PDF Report formats. WMH feature only supports data acquired on • Reporting WMH count and volume distribution as per Siemens Healthineers MR scanners the McDonald criteria over four brain regions juxtacortical, periventricular, infratentorial, and deep white matter to aid Multiple Sclerosis evaluation. • The report is also generated in Secondary Capture and Encapsulated PDF Report formats. . 6 2 AI-Rad Companion Brain MR has been tested and validated on multi-vendor datasets including Siemens Healthineers, GE, Philips. Data sheet · AI-Rad Companion Brain MR VA51 Examination Date Image Quality Segmentation Quality SIEMENS . Current Exam 34 25.06.2014 Not assessed: n.a. - Range: [n.a.] Acceptable: 0.74 - Range: [0.70 - 1.00] Healthineers Morphometry - Al-Rad Companion Brain MR Brain Structures Volume[ml] Percentile Normalised %[10th-90th Percentile] Hippocampus 7.3 96 0.53 [0.40 - 0.51] Temporal lobe-Left 62.1 58 4.46 [4.01 - 4.83] Temporal lobe-Right 62.4 24 4.49 [4.29 - 5.23] Cortical Gray Matter 480.1 15 34.50 [34.00 - 40.20] Total Brain Parenchyma 989.9 <1 71.10 [75.20 - 83.90] Ventricles 55.3 99 3.98 [1.00 - 2.72] Hippocampus Temporal lobe-Left Temporal lobe-Right Volume in & of The Volume in & of TIV Volume in & of THE 30 15 Agetyr) Agetting Cortical Gray Matter Total Brain Parenchyma Ventricles Volume in'S of TIV Ageral Agebra Agetting · - Deviations - 10th / 90th percentile. Rate of Change - Annualized change calculated using current and prior values Current Timepoint - 25.06.2014 A Cautions Input data of current study not consistent with the morphometry protocol guideline. Please check validity of the current study's volumetric results. Figure 6: Morphometry PDF Report 7 AI-Rad Companion Brain MR VA51· Data sheet Age Examination Date Input Data Input Data SIEMENS Current Exam 34 25.06.2014 FLAIR MPRAGE Healthineers White Matter Hyperintensities - Al-Rad Companion Brain MR Spatial Distribution Area Count Volume[ml] Juxtacortical 0.4 Periventricular 6 26.1 Deep white matter 19 0.3 Infratentorial 3 0.1 Total 35 26.9 General Info Progression could not be calculated because no valid prior was found. Cautions 1/1 Input data not consistent with the recommended guideline. Please check the validity of WMH results. FIgure 7: White Matter Hyperintensities PDF 8 Data sheet · AI-Rad Companion Brain MR VA51 Preconditions and requirements The following preconditions apply for AI-Rad Companion The following preconditions apply for AI-Rad Companion Brain MR Morphometry feature. Brain MR WMH feature. • Patient age must be ≥2 years • Patient age must be > 18 years • For Deviation Map only, patient age must be ≥ 16 • T1-weighted MPRAGE and T2-weighted 3D FLAIR image years and <=98 years series • 3D T1-weighted image series • MPRAGE series must have been acquired before contrast Series must have been acquired before contrast agent agent application • application • It is expected to use Pre-scan Normalization Input T1-weighted protocol with scan parameters • MR Image or Enhanced MR Image • according to the Alzheimer's Disease Neuroimaging • Initiative (ADNI) standard Image series acquired from Siemens Healthineers MR scanners • No significant asymmetry between left and right • hemispheres For MR images from Siemens scanners, if the input data was not already 3D distortion corrected, it will be • MR Image or Enhanced MR Image distortion corrected and the result series is used for • It is expected to use Pre-scan Normalization further processing. In case of 3D distortion algorithm failure, system will process the input data with an • Image data from third party vendors needs to be 3D information message. distortion corrected The following preconditions additionally apply for AI-Rad • For MR images from Siemens scanners, if the input data Companion Brain MR WMH follow up feature. was not already 3D distortion corrected, it will be The calculation of the progression (New or Enlarged) for distortion corrected and the result series is used for further processing. For third party vendors, the brain regions is based on the consistency check between prior and current studies and the follwoing DICOM tags distortion correction type of third party data may not be should be condistent between prior and current study correctly identified by the system. Hence, for optimal result quality, it is recommended to only use third party • Magnetic Field Strength (0018, 0087) input data that was already distortion corrected • Slice Thickness (0018,0050) The following preconditions additionally apply for AI-Rad • Companion Brain MR Morphometry follow up feature. Pixel Spacing (0028, 0030) The follow-up atrophy calculations will be performed only The following common requirements additionally apply: when all the following parameters are consistent between • Ensure that keep patient age is enabled in case of any the prior and current study data privacy profile is selected. This setting ensures association of the correct patient age with the reference • In-plane resolution population for the calculation of the Deviation Map • Size of the image (number of rows and number of • The mix up of different SOP class UID in a single series columns) during acquisition of the images is not supported by the Slice thickness Application. For example RAW Data (Raw Data Storage • 1.2.840.10008.5.1.4.1.1.66) along with MR Image (MR • TI (Inversion Time) Image Storage1.2.840.10008.5.1.4.1.1.4) • TR (Repetition Time) • Internet browser: Google Chrome, Microsoft Edge • Magnetic field strength • It is recommended to use Google Chrome as a • Auto Align preferred web browser for use with AI-Rad Companion Brain MR. 9 AI-Rad Companion Brain MR VA51 · Data sheet Factors influencing accuracy The following factors influence measurement accuracy of AI-Rad Companion Brain MR. • Inhomogeneities of the magnetic field or Distortion correction. For example, if scanning off-center or Coil selection mode is On/Automatic. T1-weighted MPRAGE and T2-weighted 3D FLAIR image series • Image reconstruction settings, such as matrix size, slice thickness, slice gap, bandwidth, and Field of View (FoV) • Patient compliance during image acquisition, for example, breathhold and motion • Artefacts present in the input images for example, susceptibility or motion artefacts. For example, in case of off-center scanning, or activation of spine coil elements while using automatic coil selection. Minimum hardware requirements for on-edge deployment AI-Rad Companion Brain MR supports cloud-based and on-edge deployment options. For AI-Rad Companion edge deployment, data processing is performed within the institution. This enables the institution to select the option that suits their requirementS. For on-edge deployment, the following table lists the minimum hardware requirements. These requirements correspond to a small-grade server. Processor Type CPU: 1 x 10 Cores (1 x 20 threads) Intel ® Xeon® (4th generation) 2.7GHz base frequency with Intel ® AVX2 support, for example, Intel ® Xeon® Silver 4410T or - - CPU: 1 x 12 Cores (1 x 24 threads) Intel ® Xeon® (4th generation) 2.0 GHz base frequency with Intel ® AVX2 support, for example, Intel ® Xeon® Silver 4410Y Memory RAM: 96GB (DDR5) Disk Disk: Solid State Drive (SSD) with >2.5TB; at least 500GB free disk space on Windows system drive Operating System Windows Server 2019 / Windows Server 2022 • Optionally, if you want to use AI-Rad Companion Notifier, see Interfaces, Options, and Accessories in the AI-Rad Companion Engine Instructions for Use. • Depending on the expected frequency of use of the follow-up feature, it is strongly recommended to extend the HDD capacity to a minimum of 8TB of HDD for storing the prior details in the long term storage (LTS). • In the case of cloud deployment, the computer hosting the teamplay Receiver requires access to the hospital Intranet and the Internet with at least 100Mbit/s bandwidth to the Internet. • For AI-Rad Companion Notifier, it is recommended to use Microsoft Windows 10. For more information on medium-grade and large -grade server configurations, see AI-Rad Companion Engine Datasheet. 10 AI-Rad Companion Brain MR VA51 · Data sheet Licensing Data privacy and security AI-Rad Companion Brain MR is available as a subscription As processor of your data, we are highly driven by the aim with different tiers, based on the number of cases processed. to support you in your compliance with the US standards of The services are provisioned per institution. The users HIPAA (Health Insurance Portability and Accountability Act) affiliated with a given institution can access the application, and the privacy laws applicable throughout the world, and if authorized by the institution's administrator. in particular the European GDPR (General Data Protection Regulation). Additional requirements AI-Rad Companion Brain MR supports the following privacy levels. AI-Rad Companion Brain MR requires teamplay digital health platform. • No data minimization • Standard privacy level and High privacy level, with keep patient age enabled Data minimization is applicable for cloud-based service only For more information, see AI-Rad Companion Data Privacy and Security Whitepaper. 11 AI-Rad Companion Brain MR VA51· Data sheet The information in this document contains general CE technical descriptions of specifications and options as well as standard and optional features that do not always 0123 have to be present in individual cases. Some products and services mentioned in this data sheet are not Caution: Federal law restricts this device to sale by or on commercially available in all countries. Due to certain the order of a physician (21 CFR 801.109(b)(1)). regional limitations of sales rights, we cannot guarantee the future availability of the products and services The original language of this document is English. described in this datasheet. Please contact your local Made in Germany Siemens Healthineers organization for more information. Siemens Healthineers Headquarters Siemens Healthcare GmbH Siemens Healthineers AG Henkestr. 127 Siemensstr. 3 91052 Erlangen 91301 Forchheim, Germany Germany Phone: +49 9191 18-0 siemens-healthineers.com Published by Siemens Healthcare AG · Print no. P10-BR-MR-DS.627.02.01.02 · ©Siemens Healthineers AG, 2024 12