ThrombUS+ D6.1: Product Design Specifications

D6.1. Product Design Specifications V. Marozas [KTU], N. Portokallidis [ATHENA], R. Jurkonis [KTU], M. Jucevičius [KTU], R. Eitminavičius [KTU], A. Lukoševičius [KTU], S. Daukantas [KTU], A. Rapalis [KTU], A. Sonda [KTU], A. Moutafidou [ATHENA], J. Cesana [ComfTech], P. Moustakidis [EchoNous], N. Lange [FRAUNHOFER IPMS], S. Balling [MEDIS], J. Querengässer [MEDIS], M. Legros [VERMON], D. Novikov [TELEMED], S. Maja [TAU], T. Prinz [VDE], E. Kaldoudi [ATHENA] Due Date: 28 February 2025 Delivery Date: 28 February 2025 Revision Date: 12 November 2025 Horizon Innovation Action | Agreement No. 101137227 HORIZON-HLTH-2023-TOOL-05-05 Co-funded by the European Union D6.1| Specifications v2.0 | 12 Nov 2025 ii ThrombUS+ Consortium ATHENA Research and Innovation Center in Information, Communication and Knowledge Technologies, Greece Eleni Kaldoudi [email protected] KTU Kaunas University of Technology Lithuania Vaidotas Marozas [email protected] VERMON Vermon SA France Mathieu Legros [email protected] FRAUNHOFER Institute for Photonic Microsystems IPMS, Fraunhofer Germany Nicolas Lange [email protected] TELEMED Telemed Ultrasound Medical Systems Lithuania Dmitry Novikov [email protected] EchoNous EchoNous Inc USA Pavlos Moustakidis [email protected] MEDIS medis Medizinische Messtechnik GmbH Germany Susann Balling [email protected] ComfTech ComfTech SLR Italy Lara Alessia Moltani [email protected] TAU Faculty of Medicine and Health Technology Tampere University, Finland Antti Vehkaoja [email protected] LMSU Lithuanian University of Health Science Lithuania Andrius Macas [email protected] GNP Papageorgiou General Hospital Greece Maria Bigaki [email protected] CSS-IRCCS Home Relief of Suffering Hospital Italy Elvira Grandone e.grand[email protected] HSV Simon Veil Hospital France Maxime Gautier [email protected] VDE Association for Electrical, Electronic & Information Technologies, Germany Thorsten Prinz thorsten.pri[email protected] MEDEA MEDEA SRL Italy Pietro Dionisio [email protected] PHAZE Clinical Research and Pharma Consulting SA Greece Spiros Anagnostopoulos [email protected] PBY PredictBy Research and Consulting SL Spain Frans Folkvord [email protected] SciGen SciGen Technologies SA Greece Katerina Pavlidi [email protected] Disclaimer This document contains description of the ThrombUS+ project work, findings, and products. The authors of this document have taken any available measure for its content to be accurate, consistent and lawful. However, neither the project consortium as a whole nor the individual partners that implicitly or explicitly participated in the creation and publication of this document hold any sort of responsibility that might occur as a result of using its content. Views and opinions expressed are those of the author(s) only and do not necessarily reflect those of the European Union or HADEA. Neither the European Union nor the granting authority HADEA can be held responsible for them. In case you believe that this document harms in any way intellectual property held by you as a person or as a representative of an entity, please do notify us immediately. ThrombUS+ is an Innovation Action Project co-funded by the European Union, under HORIZON-HLTH-2023-TOOL05-05 “Harnessing the potential of real-time data analysis and secure Point-of-Care computing for the benefit of person-centred health and care delivery”. D6.1| Specifications v2.0 | 12 Nov 2025 iii Document Control Page Project Grant Agreement: 101137227 Acronym: ThrombUS+ Title: Wearable Continuous Point-of-Care Monitoring, Risk Estimation and Prevention for Deep Vein Thrombosis Type: Innovation Action Start: 1 January 2024 End: 30 June 2027 Programme: Horizon Europe Call Identifier: HORIZON-HLTH-2023-TOOL-05-05 Call Topic Harnessing the potential of real-time data analysis and secure Point-of-Care computing for the benefit of person-centred health and care delivery Website: http://thrombus.eu/ Deliverable # 22 No: D6.1 Deliverable Title: Product design specifications Deliverable Type: R Classification: PU Task: 6.1. Integrated product technical specifications [M09 - M14] Task Leader: KTU [V. Marozas] Work Package: Work package WP6 - Integration and validation [M09-M32] Work Package Leader: KTU [V. Marozas] Responsible Partner: KTU Authors: V. Marozas [KTU], N. Portokallidis [ATHENA], R. Jurkonis [KTU], M. Jucevičius [KTU], R. Eitminavičius [KTU], A. Lukoševičius [KTU], S. Daukantas [KTU], A. Rapalis [KTU], A. Sonda [KTU], A. Moutafidou [ATHENA], J. Cesana [ComfTech], P. Moustakidis [EchoNous], N. Lange [FRAUNHOFER], S. Balling [MEDIS], J. Querengässer [MEDIS], M. Legros [VERMON], D. Novikov [TELEMED], S. Maja [TAU], T. Prinz [VDE], E. Kaldoudi [ATHENA] Input from: All consortium partners Peer Reviewers: Z. Narten [VDE], S. Didaskalou [ATHENA] Due Date: 28 February 2025 [M14] Delivery Date: 28 February 2025 Revision Date: 12 November 2025 Document Status Version: 2.0 Status: Draft Consortium reviewed WP leader endorsed Coordinator endorsed D6.1| Specifications v2.0 | 12 Nov 2025 iv Revision History Version Date Modification Contributors 0.1 06 Oct 2024 New content: outline V. Marozas, A. Lukosevicius [KTU] 0.1 12 Dec 2024 New content: outline V. Marozas, A. Lukosevicius [KTU] 0.1 09 Jan2025 New content: outline V. Marozas [KTU] 0.1 19 Jan 2025 New content: outline V. Marozas [KTU] 0.3 06 Feb 2025 New content: outline R. Eitminavičius, V. Marozas, R. Jurkonis, M. Jucevičius, A. Rapalis, S. Daukantas, A. Sonda [KTU] 0.4 07 Feb 2025 New content: outline R. Jurkonis [KTU] 0.4 09 Feb 2025 New content: outline, Ch4.2 V. Marozas [KTU] 0.4 10 Feb 2025 New content: outline, Ch4.2, V. Marozas [KTU] 0.4 13 Feb 2025 New content Ch. 5.4.1, Ch. 5.5 S. Maja [TAU], S. Balling [medis] 0.4 14 Feb 2025 New content Ch 5.1.2, Ch. 5.5 N. Portokallidis [ATHENA], S. Balling [medis] 0.4 16 Feb 2025 New content: outline, Ch2 V. Marozas [KTU] 0.4 17 Feb 2025 New content: outline, fixing Ch3 V. Marozas [KTU] 0.5 18 Feb 2025 New content: Ch 3.1, V. Marozas, R. Eitminavičius [KTU] 0.5 18 Feb 2025 New content: Ch 3.1.2, Ch 4.3, Ch 4.4 N. Portokallidis [ATHENA], E. Kaldoudi [ATHENA] R. Eitminavičius [KTU] 0.5 19 Feb 2025 Ch. 5.4; Ch 3.2 A. Rapalis [KTU], D. Novikov [Telemed] 0.5 20 Feb 2025 Fixed Ch 3.3; modifications: Ch 2.3, 2.4 R. Eitminavičius, V. Marozas [KTU] 0.5 21 Feb 2025 Fixed Ch. 3.4.1; modifications Ch 3.3; S. Maja [TAU], R. Eitminavičius [KTU] 0.5 22 Feb 2025 Added block diagram in Ch 3.2.2 A. Sonda [KTU] 0.6 22 Feb 2025 New content in Ch. 3 V. Marozas [KTU] 0.6 23 Feb 2025 New content in Ch. 3, 4, and 5. V. Marozas [KTU] 0.6 23 Feb 2025 New content in Ch. 3, 4, and 5. V. Marozas [KTU] 0.6 24 Feb 2025 New content in Ch. 3, 4, and 5. V. Marozas [KTU] 0.6 24 Feb 2025 New content in Ch 3 N. Moustakidis (Echonous) 0.6 24 Feb 2025 Update content in Ch 5 N. Portokallidis, A. Moutafidou [ATHENA] 0.6 25 Feb 2025 New content in Ch 3.3.1.; Ch.4, Conclusions, Summary R. Eitminavičius, V. Marozas [KTU] 0.6 26 Feb 2025 Document Control Page A. Rapalis [KTU] 0.6 26 Feb 2025 New content in Ch 3.3.1. and 3.3.2. R. Eitminavičius, S. Daukantas [KTU], J. Querengässer [MEDIS] 0.7 28 Feb 2025 Final editing A. Rapalis [KTU] 1.0 28 Feb 2025 Final editing for conformity E. Kaldoudi [ATHENA] 2.0 12 Nov 2025 Revised based on PR1 review comments: A clarification of the extended reality functionality and the distinction between the serious gaming was added in Sections 4.4 and 5.1.3. E. Kaldoudi [ATHENA] D6.1| Specifications v2.0 | 12 Nov 2025 v Contents About ThrombUS+ ...................................................................................................................................................... 1 Deliverable D6.1 Description ...................................................................................................................................... 1 Cite this Document as ................................................................................................................................................. 1 Terms and Definitions ................................................................................................................................................ 2 Executive Summary .................................................................................................................................................... 4 1. Introduction ....................................................................................................................................................... 5 2. Background ........................................................................................................................................................ 5 2.1. Intended purpose of ThrombUS+ device ........................................................................................................... 5 2.2. Use cases and main user requirements ............................................................................................................. 5 2.3. General architecture of ThrombUS+ system ..................................................................................................... 6 2.4. Compliance and standards ................................................................................................................................ 7 3. Technical specifications of the ThrombUS+ modules .......................................................................................... 8 3.1. Central Hub & Intelligence (CHI) ........................................................................................................................ 8 3.1.1. Hardware .................................................................................................................................................. 8 3.1.2. Software ................................................................................................................................................... 9 3.2. Ultrasound Monitoring Module (USM) ........................................................................................................... 15 3.2.1. Hardware ................................................................................................................................................ 15 3.2.2. Software ................................................................................................................................................. 22 3.3. Wearable Actuator Module (WAM) ................................................................................................................ 25 3.3.1. Hardware ................................................................................................................................................ 25 3.3.2. Software ................................................................................................................................................. 27 3.4. Electrical Impedance Module (EIM) ................................................................................................................ 29 3.4.1. Hardware ................................................................................................................................................ 29 3.4.2. Software ................................................................................................................................................. 30 3.5. Light Rheography Module (LRM)..................................................................................................................... 32 3.5.1. Hardware ................................................................................................................................................ 32 3.5.2. Software ................................................................................................................................................. 34 3.6. Limb Activity Module (LAM) ............................................................................................................................ 36 3.6.1. Hardware ................................................................................................................................................ 36 3.6.2. Software ................................................................................................................................................. 38 3.7. Extended Reality & Serious Gaming Module (XGM) ........................................................................................ 40 3.7.1. Software ................................................................................................................................................. 40 4. DVT monitoring methods and working modes ..................................................................................................43 4.1. Operator-independent Compression Duplex Ultrasonography for DVT monitoring ....................................... 43 4.2. Operator-independent Venous Occlusion Plethysmography for DVT monitoring ........................................... 44 4.3. Operator-Independent Light Reflection Rheography-Based Monitoring of DVT ............................................. 45 4.4. Serious Gaming and Extended Reality for DVT prevention and rehabilitation ................................................ 46 5. Critical specifications of the ThrombUS+ integrated product ............................................................................47 D6.1| Specifications v2.0 | 12 Nov 2025 vi 5.1. Core functions.................................................................................................................................................. 47 5.1.1. Ultrasound imaging for DVT detection in the lower limbs ..................................................................... 47 5.1.2. Plethysmography measurements for DVT detection in the lower limbs ............................................... 48 5.1.3. Extended reality environment and serious games for DVT prevention ................................................. 48 5.1.4. Integrated data processing and visualization ......................................................................................... 48 5.2. User flow ......................................................................................................................................................... 49 5.3. User interfaces ................................................................................................................................................ 50 5.4. System roles and user interface features ........................................................................................................ 50 5.5. Functional Overview ........................................................................................................................................ 51 5.6. Testing, verification, and validation ................................................................................................................ 51 5.7. Potential ThrombUS+ system improvements and extensions ......................................................................... 51 D6.1| Specifications v2.0 | 12 Nov 2025 1 About ThrombUS+ Deep vein thrombosis (DVT) is the formation of a blood clot within the deep veins, most commonly those of the lower limbs, causing obstruction of blood flow. In 50% of people with DVT, the clot eventually breaks off and travels to the lung to cause pulmonary embolism. Clinical assessment of DVT is notoriously unreliable because up to 2/3 of DVT episodes are clinically silent and patients are symptom free even when pulmonary embolism has developed. Early diagnosis of DVT is crucial and despite the progress made in ultrasound imaging and plethysmography techniques, there is a need for new methods to enable continuous monitoring DVT diagnosis at the point of care. ThrombUS+ brings together an interdisciplinary team of industrial, technology, regulatory, social science and clinical trial experts to develop a novel wearable diagnostic device for point-of-care, operator free, continuous monitoring in patients with high DVT risk. The device will combine autonomous, AI driven DVT detection based on a novel wearable ultrasound hardware, impedance plethysmography and light reflection rheography for immediate detection of blood clot formation in the lower limb. Activity and other physiological measurements will be used to provide a continuous assessment of DVT risk and support DVT prevention via serious gaming. The aggregated data will drive an intelligence decision support unit that will provide accurate monitoring and alerts. Extended reality will be used to guide experts to design exercises and patients to use the device optimally. ThrombUS+ is intended for use by postoperative patients in the ward, during long surgical operations, cancer patients or otherwise bedridden patients at home or in care units, and women during pregnancy and postpartum. ThrombUS+ will use big data sets for AI training collected in the project via 3 large scale clinical studies and will validate the outcome in the clinical setting via 1 early feasibility study and 1 multi-center clinical trial. Deliverable D6.1 Description Task 6.1 aims at preparing a thorough review of relevant standards, quality guidelines of medical devices, and statutory regulations that ultimately need to be met by the finished product. These will be used to establish critical product specifications, which in turn, will be used to devise a testing strategy to ascertain compliance with the predefined standards and specifications. Key end-users requirements will be prioritized through a Quality-Function Deployment tool, and they will be translated to specifications, considering outcomes of WP2. The task is led by KTU, an expert in sensor and sensor technology hardware and software, thus ensuring an extended expertise to overview the technical integration of the ThrombUS+ solution. The regulatory partner VDE has a crucial role in ensuring that compliance by design aspects are defined. All partners participate to ensure continuity with the rest of the project activities. Cite this Document as Marozas V, Portokallidis N, Jurkonis R, Jucevičius M, Eitminavičius R, Lukoševičius A, Daukantas S, Rapalis A, Sonda A, Moutafidou A, Cesana J, Moustakidis P, Lange N, Balling S, Querengässer J, Legros M, Novikov D, Maja S, Prinz T, Kaldoudi E. Product Design Specifications, Deliverable 6.1, ThrombUS+ Horizon Europe Innovation Action, EC Grant Agreement No. 101137227, 28 February 2025. Revised 12 November 2025. https://doi.org/10.5281/zenodo.17642925 D6.1| Specifications v2.0 | 12 Nov 2025 2 Terms and Definitions Abbreviation Definition AI Artificial Intelligence AIA Artificial Intelligence Act API Application Programming Interface AUC Area Under Curve CDUS Compression Duplex Ultrasound CHI Identifier for ThrombUS+ Central Hub & Intelligence Module CNN Convolutional Neural Network CSW C-shell type Semirigid Wearable CUS Compression Ultrasound Imaging DICOM Digital Imaging and Communications in Medicine DoA Description of Action DVT Deep Vein Thrombosis EAC Identifier for ThrombUS+ Electronic Actuator Controller EC European Commission EIM Identifier for ThrombUS+ Electrical Impedance Module EIP Electrical Impedance Plethysmography EU European Union GDPR General Data Protection Regulation HADEA European Health and Digital Executive Agency IMU Inertial Measurement Unit IoU Intersection over Union IEC International Electrotechnical Commission ISO International Organization for Standardization LAM Identifier for ThrombUS+ Limb Activity Module LRM Identifier for ThrombUS+ Light Rheography Module LRR Light Reflection Rheography MEMS Micro-Electro-Mechanical System MDR Medical Device Regulation ML Machine Learning npm Node Package Manager PPG Photoplethysmography PZT Pb (ZrTi) - shorten form of Lead zirconate titanate D6.1| Specifications v2.0 | 12 Nov 2025 3 ReLU Rectified Linear Unit ROC Receiver-Operating Characteristic RUCs Reference Use Cases SDK Software Development Kit UI User Interface US Ultrasound USM Identifier for ThrombUS+ Ultrasound Monitoring Module VOP Venous Occlusion Plethysmography WAM Identifier for ThrombUS+ Wearable Actuator Module WI-Fi Wireless Fidelity WP Work Package WSN Identifier for ThrombUS+ Wearable Sensors Network WTC Identifier for ThrombUS+ Wearable Thigh Cuff WTE Identifier for ThrombUS+ Wearable Tetrapolar Electrode WUSD Identifier for ThrombUS+ Wearable Ultrasound Device XGM Identifier for ThrombUS+ Extended Reality and Serious Gaming Module XR Extended Reality D6.1| Specifications v2.0 | 12 Nov 2025 10 Functional scope Data Handling Import and store patient data: demographics, medical records, sensor streams. Manage local records until explicit clearance. Perform ML-based DVT risk analysis with TensorFlow.NET. User Interfaces Clinician Mode: Advanced control panels, device configuration, manual override. Patient Mode: Guided exercises, restricted controls, PIN-locked interface. Security & Compliance Mandatory PIN protection for lock/unlock states. Automatic logging of user actions (audit trail). Local data storage (SQLite) with no external transmission unless authorized. Integration REST endpoints for control/configuration. WebSocket for real-time communication (alerts, updates). UDP for high-frequency sensor data (IMU, external devices). Optional Unity-based extended reality features. CHI software architecture Frontend Windows Forms for core UI elements. Embedded WebView2 (Vue 3) for modern web-based components. Unity integration for XR scenarios (if required by the clinical setting). Backend C# (.NET 8) runtime managing data flow and core logic. SQLite database for persistent local storage. TensorFlow.NET library for DVT risk detection. Communication protocols REST API for device and application-level commands. WebSocket for low-latency streaming of alerts/ data visualization. UDP for raw sensor packet reception. IPC on Windows for local inter-process messaging. Figure 2. The Central Hub & Intelligence software architecture. The main application consists of a Service Manager, which runs on the system’s tray and a graphical user interface (GUI) for interacting with the application’s functionality. D6.1| Specifications v2.0 | 12 Nov 2025 11 The various components of the database, their description and their relations are presented in Table 3 and a graphical representation of the database schema is presented in Figure 3. The fields for each database table, the data type of each field and the relationships between the different database tables are also represented in Figure 3. Table 3. Database schema. Table Description Relations Games Stores game configurations linked to exercises (title, difficulty, description, data) FK ExerciseId → Exercises(Id) UNIQUE on (ExerciseId) SensorData Captures sensor outputs for each patient (type, timestamp, binary/string/JSON data) FK PatientId → Patients(Id) UNIQUE on (PatientId) RiskProfiles Holds DVT-related risk factors and details for each patient FK PatientId → Patients(Id) UNIQUE on (PatientId) MaintenanceLog Logs maintenance or calibration events (sensor type, version, notes). None Sessions Session records (start/end times, type, data) per patient, specifying a source (exercise, game, exam). FK PatientId → Patients(Id) UNIQUE on (PatientId) Column Source references domain (e.g., Exercise, Game, Exam) Settings System-wide or device-specific configuration entries (key-value pairs). None LogFiles Logs user actions (timestamp, action, details) for auditing and compliance. None Notifications Issues alerts per patient (severity, message, timestamps), referencing other entities. FK PatientId (linked to Patients(Id) though it’s also marked as a primary key in the schema) Patients Core patient records (names, DOB, health record). FK to Notifications(PatientId) as per schema definition PatientPlan Encapsulates a patient’s therapy plan, linking them to exercises and games (data, timestamp). FK PatientId → Patients(Id) FK Exercises → Exercises(Id) FK Games → Games(Id) UNIQUE on (PatientId) Exercises Definitions of activities (description, media data, metadata). Referenced by Games(ExerciseId) and PatientPlan(Exercises) D6.1| Specifications v2.0 | 12 Nov 2025 12 Figure 3. The ThrombUS+ database tables, with their respective fields and their data types, used within the Central Hub & Intelligence module. Table 4 presents the CHI software frontend Node Package Manager (npm) 4 dependencies. Node Package Manager is a dependency management tool for JavaScript, simplifying the installation, updating, and distribution of libraries to ensure version control and maintainability. It supports both open-source and private packages, contributed by a vast developer community. Table 4. Software Node Package Manager dependencies of the Central Hub & Intelligence module. npm Package Name Version Description @quasar/extras ^1.16.4 Provides icon libraries and additional styles used by Quasar. axios ^1.2.1 A promise-based HTTP client for making requests from browsers and Node.js. d3 ^7.9.0 A JavaScript library for producing dynamic, interactive data visualizations in web browsers. 4 Node Package Manager www.npmjs.com D6.1| Specifications v2.0 | 12 Nov 2025 13 d3-sankey ^0.12.3 A plugin for D3 that simplifies the creation of Sankey diagrams. pinia ^2.0.11 An intuitive, lightweight store library for Vue 3, serving as an alternative to Vuex. quasar ^2.16.0 A high-performance Vue.js framework for building rich, responsive UIs with Material Design. vue ^3.4.18 The core progressive JavaScript framework for building user interfaces. vue-i18n ^9.0.0 Internationalization plugin for Vue, enabling multi-language support. vue-router ^4.0.0 The official router for Vue, managing navigation and route matching in single-page apps. Table 5 presents the list of NuGet packages 5 and their corresponding versions. NuGet Package Manager is a dependency management tool for .NET, streamlining the installation, updating, and distribution of libraries to ensure version control and maintainability. All referenced packages are open source with many contributors. To connect, communicate, control, and collect data from the various sensors of the ThrombUS+ system, custom SDK libraries will be utilized. Table 6 presents the list of SDK embedded libraries under development for the ThrombUS+ system. For the assessment of the ultrasound images two convolutional neural network (CNN) models tailored for real-time ultrasound image analysis will be developed and deployed within the Central Hub & Intelligence module. The first model, based on MobileNet architecture, is designed for multi-class classification, grading target veins using clinically relevant ultrasound DICOM frames as input. The second model, leveraging a UNet architecture, performs pixel-wise classification to segment and label vascular structures. Both models receive preprocessed B-mode ultrasound images, where cropping, pixel normalization, resizing, and any additional adjustments are applied beforehand to ensure optimal input quality and alignment with the project's goals and objectives. The MobileNet-based classifier employs Leaky ReLU activation, optimizing performance with Cross Entropy or Focal Loss, while the U-Net segmentation model uses Dice Loss, JS divergence (Jensen-Shannon divergence), or KL divergence (Kullback–Leible divergence) for accurate mask generation. Evaluation metrics include Top-K and Balanced Accuracy for classification and Intersection over Union (IoU) and receiver-operating characteristic curve (ROC) and area under the curve (AUC) for segmentation. The models are trained in PyTorch and exported in .tflite format, ensuring compatibility with a real-time deployment pipeline in C# via TensorFlow.NET. Targeting an inference latency of under 30ms, this implementation supports efficient and accurate real-time vascular analysis for clinical applications. The specification of the SDK for DVT detections based on ultrasound images are listed below in Table 7. Table 5. The list of NuGet packages and the version used within the ThrombUS+ Central Hub & Intelligence module. NuGet Package Name Version TensorFlow.NET 0.150.0 AForge.Video 2.2.5 AForge.Video.DirectShow 2.2.5 Microsoft.EntityFrameworkCore 9.0.0 Microsoft.EntityFrameworkCore.Design 9.0.0 Microsoft.EntityFrameworkCore.Sqlite 9.0.0 5 NuGet packages www.nuget.org D6.1| Specifications v2.0 | 12 Nov 2025 14 Microsoft.Extensions.Configuration 9.0.0 Microsoft.Web.WebView2 1.0.2903.40 Swashbuckle.AspNetCore 7.2.0 System.Management 9.0.1 System.Net.Security 4.3.2 System.Net.WebSockets.Client 4.3.2 Microsoft.Extensions.Hosting 9.0.0 Zeroconf 3.7.16 Table 6. The list of the ThrombUS+ SDK embedded libraries. Module SDK embedded libraries Ultrasound Module Telemed_SDK Libs\Interop.Usgfw2Lib.dll Limb Activity Module LAM_SDK Libs\LAM_SDK.dll Light Reflection Module LRM_SDK Libs\LRM_SDK.dll Electrical Impedance Module EIM_SDK Libs\EIM_SDK.dll Wearable Actuator Module WAM_SDK Libs\WAM_SDK.dll Table 7. The specification of SDK for DVT AI detection module. General information Model Type CNN MobileNet Based (Real-time Image Grading of Target Veins) CNN U-Net Based (Real-time Labelling of Vascular Structures) Model Format .tflite Trained On Ultrasound DICOM frames with clinically relevant information (.dcm files) and corresponding labels/annotations saved in .json format Technical characteristics Input shape/format [1, 1, 128, 128] Model output Multi-Class Classification (5 classes) - Probability (0-1) (Real-time Image Grading of Target Veins) Segmentation Masks – Pixel level classification with probability (0-1) (Real-time Labelling of Vascular Structures) Activation Function Leaky ReLU (Real-time Image Grading of Target Veins) Leaky ReLU (Real-time Labelling of Vascular Structures) Loss Function Cross Entropy Loss OR Focal Loss (Real-time Image Grading of Target Veins) Dice Loss OR JS div loss OR KL div loss (Real-time Labelling of Vascular Structures) Evaluation Metrics Top-K accuracy and balanced accuracy (real-time image grading of target veins) Preprocessing B-mode US area cropping, pixel array scaling [0,1], color-space correction if needed, input frame resizing to input shape, augmentations if needed, etc.) Deployment C# via TensorFlow.NET Latency Target < 30ms per frame for real-time analysis D6.1| Specifications v2.0 | 12 Nov 2025 15 3.2. Ultrasound Monitoring Module (USM) 3.2.1. Hardware Table 8, Table 9, Table 10, and Table 11 present the specifications of the USM integrating together ultrasound beamformer and ultrasound transducers developed by TELEMED, VERMON, and FRAUNHOFER and software developed by ATHENA, respectively. Table 8. Specifications of the ultrasound beamformer by TELEMED. General information Generic name Ultrasound Imaging System (ArtUs Compact Beamformer) Type Hardware and software Version Latest released version Author TELEMED Technical characteristics Imaging modes B-mode (bidimensional, simultaneous) M-mode Color Doppler Pulsed Doppler Harmonic Imaging Depth of imaging 40 - 80 mm Width of imaging 50 - 100 % Receiver’s gain range 0-100% Emission power range -20 to 0 dB Dynamic range 36 - 102 dB Frequency range 1-15 MHz Focal depth Up to 50 mm Time - gain control Adjustable in range 0-100% individually in each of five depth zones Image memory Frame by frame or cine-loop Streaming data to CHI USB 3.2, physical layer USB C. Triggering Electronic Actuation Controller (EAC) BNC connection, two ports IN and OUT Memory capacity Recording up to 1024 frames ( PC memory dependent) Power supply External AC~220 V ±10% 50 Hz or DC +5V 500mA power from USB-C interface from the CHI module. Environmental requirements Transport and storage conditions Temperature: 10 to 40 °C Relative humidity: 15 to 85 % (non-condensing) Atmospheric pressure: 700 hPa to 1060 hPa Temperature: 10 to 40 °C D6.1| Specifications v2.0 | 12 Nov 2025 16 Operating and storage conditions Relative humidity: 15 to 85 % (non-condensing) Atmospheric pressure: 700 hPa to 1060 hPa Physical and chemical characteristics Physical components Beamformer with socket connector for coaxial cables (multi-line) to imaging transducer and two SMA Jack sockets for IN and OUT triggers; USB cable with both endings USB-C; external module for power supply +5V. USB cable with both endings USB-C external module for power supply +5V. Raw materials N/A Accessories, consumables, spare parts, and other components Accessories USB-C cable 0,5 m Sterilization / cleaning of accessories List of recommended disinfectants for transducer soaking or wiping Consumables / reagents N/A Spare parts Spare parts to be replaced on a regular basis Other components Medical-grade power supply ACM18US05 Maintenance Maintenance tasks The measurement system must always be kept in a safe and reliable working order and should be checked regularly, at least once a year, by the manufacturer or authorized persons. The check covers: − Inspection for potential mechanical and functional damage. − Readability and integrity of safety labels. − Function test of the device. Table 9. Specifications of the Linear Array ultrasound transducer by VERMON. General information Generic name Linear array transducer Type Hardware Version Latest released version Author VERMON Mechanical structure Technology Piezocomposite, PZT ceramic Transducer Type: Linear Array - LA 6.0/128-2886 Number of elements 128 Element pitch 0.3 mm Elevation aperture 6 mm Curvature radius (convex, mm) N/A Acoustic lens radius (mm) N/A Angle element pitch (convex, deg) N/A D6.1| Specifications v2.0 | 12 Nov 2025 17 Elevation focus 30 mm Housing shape rectangular parallelepiped (see D 3.1) External dimensions of housing (LxWxH) 90x25x20 mm Handling by-hand Detachable ergonomic handle. Electrical specifications Central Frequency at -6dB 5.4 – 6.6 MHz Fractional Bandwidth at -6dB ≥60 min Pulse Width at -20dB (us MAX): ≤900 ns Cable specifications Cable Type: AWG size 44 Outer Diameter 6.2±0.3 mm Number of coax wires: 135 Length 2 m Capacitance 60±7(pF/m) Connector specifications Connector Type ArtUs-Compact (Telemed specific) Connector Pinout ArtUs-Compact (Telemed specific) Operating environment Operating temperature range +10 to +40 °C Operating relative humidity range 10 % to 80 % Operating atmospheric pressure range 700 hPa to 1060 hPa Storage Temperature Range -20 to +55 °C Storage relative Humidity Range 10% to 85 % RH Storage Atmospheric Pressure Range 500 hPa to 1060 hPa Safety and reliability tests Salt Immersion Test Salt solution (saline) immersion Dielectric Strength (HIPOT) Test HIPOT soaking limit (HIPOT specification: 1500 V AC) Leakage Current Test* ≤ 50 μA at 264 V AC Disinfection and sterilization Commercially available compatible disinfectants Provided in the following list Compatible product Alkazyme Korsolex Basic Bodedex Forte Korsolex Extra Bomix Plus Mikrozid D6.1| Specifications v2.0 | 12 Nov 2025 18 Cavi Wipes XL Mikrozid AF Cidex 2% Prolystica 2X Cidex OPA Protex Wipes Cidex Plus Metrizyme Cidezyme Salvanios PH10 Gigasept AF Steranios Gigasept FF Klenzyme Incidin Foam S Biocompatibility test reports Patient contact material Reports provided, tests passed Acoustic lens Silicone Seam line acoustic lens housing Silicone Housing Plastic Seam line housing (2 halves) Glue Table 10. Specifications of the Raw-Column Array ultrasound transducer by VERMON. General information Generic name Row-Column array transducer Type Hardware Author VERMON Mechanical structure Technology Piezocomposite, PZT ceramic Transducer Type Row-Column Array - RCA 6.0/96+32-2887 Number of Elements 128 (Azimuth: 96/ Elevation: 32) Element Pitch (mm) Azimuth: 0.3 / Elevation:0.2 Elevation Aperture (mm) Azimuth: 6.4 / Elevation: 28.8 Electrical specifications Central Frequency at -6dB 5.4 – 6.6 MHz Fractional Bandwidth at -6dB ≥ 60 min Cable specifications Cable Type AWG size 44 Outer Diameter 6.2±0.3(mm) Number of coaxial wires 135 Length 2 m Capacitance 60±7 pF/m Connector specifications Connector Type ArtUs-Compact (Telemed specific) D6.1| Specifications v2.0 | 12 Nov 2025 19 Connector Pinout ArtUs-Compact (Telemed specific) Operating environment Operating Temperature Range +10 to +40 °C Operating Relative Humidity Range 10 % to 80 % Operating Atmospheric Pressure Range 700 hPa to 1060 hPa Storage Temperature Range -20 to +55 °C Storage relative Humidity Range 10% to 85 % RH Storage Atmospheric Pressure Range 500 hPa to 1060 hPa Safety and reliability tests available Tests description Salt solution (saline) immersion Dielectric Strength (Hipot) Test HIPOT soaking limit (HIPOT specification : 1500 V AC) Leakage Current Test* ≤ 50 μA at 264 V AC Disinfection and sterilization Commercially available compatible disinfectants Provided in following list Compatible product Alkazyme Korsolex Basic Bodedex Forte Korsolex Extra Bomix Plus Mikrozid Cavi Wipes XL Mikrozid AF Cidex 2% Prolystica 2X Cidex OPA; Protex Wipes Cidex Plus Metrizyme Cidezyme Salvanios PH10 Gigasept AF Steranios Gigasept FF Klenzyme Incidin Foam S Biocompatibility test reports Patient contact material Reports provided, tests passed Acoustic lens Silicone Seam line accoustic lenshousing Silicone Housing Plastic Seam line Housing (2 halves) Glue D6.1| Specifications v2.0 | 12 Nov 2025 26 Table 13. Wearable actuator module (WAM) specifications. General information Generic name Wearable Actuator Module Type Hardware and firmware Version Latest released version Author KTU Technical characteristics Inflation patterns Fully controllable from CHI Communication (input) Via USB-C from CHI Pump technology Pneumatic piezoelectric (ultrasonic) micropump Exhaust valves Normally open quick exhaust valves for safety for each bladder Number of bladders 2 Number of pressure sensors 2 (monitoring each bladder separately) Number of pumps 2 Input (air) filtration < 3μm membrane filter (for each pump) Operational noise < 40 dB (20 Hz – 20kHz) Power supply Via a USB-C from the CHI device Response time Ultrafast, millisecond response time Pressure range 0 – 150 mmHg Free flow rate 2.2 L/min (continuous) and 2.7 L/min (intermittent) (combined) Maximum pressure 150 mmHg (continuous) and 188 mmHg (intermittent) Displayed parameters Returns pressure readings, time stamps, and data on WAM module status User adjustable The Velcro straps and tension-mounting system, wearable on the thigh, are adjustable for the patient Mechanical characteristics of WAM for CDUS (C-shell type Semirigid Wearable (CSW) option) The inner circumference of the adjustable WAM frames Approximately 38-63 cm The total quantity of C-shaped frames per WAM 2 Size of the insert for the US transducer in one of the WAM frames Rectangular slot, with internal dimensions of 90x25x15 mm Envelope of the frame Fabric: textile combination of rigid Polyamide and soft fabric composed of Polyamide and Elastane Inflatable bladder size Approximately: 110 x 60 mm (2 units) or 220 x 100 mm (1 unit). Electronic Actuator Controller (EAC) dimensions. Approximately 90 x 90 x 40 mm Mechanical characteristics of WAM for VOP (Wearable Thigh Cuff (WTC) option) D6.1| Specifications v2.0 | 12 Nov 2025 27 Number of bladders 1 Inflatable bladder size Approximately 280 x 110 mm Circumference of the envelope for mounting on the thigh Regulated by Velcro straps (≤ 50 cm) Envelope material Fabric: textile combination of rigid Polyamide and soft fabric composed of Polyamide and Elastane Physical and chemical characteristics Components Two bladers with two rigid C-shaped frames and associated textiles, wearable actuator electronics module, and USB type-C cable. Raw materials Enclosure – PLA (biocompatible), C-shaped frames – PLA (biocompatible). Textile materials in contact with the skin are biocompatible and provide reliable skin contact with comfort Utility and environmental requirements Power USB-C 5V @ 1.5A max Context-dependent requirements Capable of being stored continuously in ambient temperature of 0 to 50 deg C and relative humidity of 15 to 90% Capable of operating continuously in the ambient temperature of 10 to 40 deg C and relative humidity of 15 to 90% Accessories, consumables, spare parts, and other components Accessories USB-C cable 1,5 m Sterilization / cleaning of accessories The textiles of the actuator can be washed Consumables / reagents - Spare parts Spare parts are to be replaced regularly Other components - Maintenance Maintenance tasks Verification of the integrity and connection of air tubes, connectors, and external housing before each use Upgrade tasks Firmware upgrade through a dedicated GUI application 3.3.2. Software Table 14 presents the specification the WAM related software. Table 14. Specifications of the WAM related software. General information Generic name WAM_SDK Type Software and firmware Operating systems Windows 11 or higher (64-bit) The latest .NET runtime Programming languages C and C++ (for ESP32 microcontroller) D6.1| Specifications v2.0 | 12 Nov 2025 28 Firmware EAC Version Latest released version Author KTU Software requirements Programming language C Communication protocol Communication with the WAM module is achieved via USB-C type cable and USB communications device class (CDC). Serial communication is 115200 8-N-1. Functional scope Data streaming Status and Pressure sensor reading responses are constructed on the device, depending on the actual pump module and valve states. Application programming interface (API) Text-based commands should be sent to the WAM module. All commands are terminated by carriage return symbol ‘\r’ (ASCII CR). Every command received is responded to by echoing the command back or “ERR\r” if the wrong command is received. The full list of the commands is presented in the deliverable D3.3. GUI application A dedicated GUI application for the testing and maintenance of the WAM component. Technical characteristics Supported algorithms Closed loop pressure control by pressure sensor readings for each pump Core functionalities Collect and transmit device Status and pressure sensor readings Receive control commands for pump and valve control Configuration Predefined settings Approximately 62Hz sampling rate for pressure sensors Pump power 1000mW Slow-release valves open Exhaust valves open Adjustable settings 1-62Hz sampling rate Enable/Disable pumps Set desired inflation pressure Close/Open slow-release valves Close/Open exhaust valves Compatibility & Integration APIs WAM_SDK.dll Protocols Virtual Com Port over USB connection Baudrate 115200 bps Data mode 8-N-1 Command termination Carriage return symbol (‘\r’) Maintenance Update mechanism Via dedicated GUI Error handling Error codes in Status Report Communication & Connectivity D6.1| Specifications v2.0 | 12 Nov 2025 29 Data transmission Wired, by USB cable Data formats Text-based (ASCII) input for commands and readings as output Data storage No internal storage present 3.4. Electrical Impedance Module (EIM) 3.4.1. Hardware Table 15 presents the electrical impedance module (EIM) specifications. Table 15. Electrical impedance module (EIM) specifications. General information Generic name Electrical Impedance Module Type Hardware and firmware Version Latest released version Author TAU Technical characteristics Measuring principle Singleand multifrequency impedance plethysmography Electrode configuration Bipolar, tetrapolar Excitation frequency Up to 20 adjustable frequencies in the range of 10-500 kHz Excitation current Adjustable: up to 2.9 mA Impedance measurement range 0.1-400 Ω Sampling rate Adjustable: 25, 50, 100 or 200 Hz Communication Wi-Fi Power Rechargeable lithium polymer battery Battery voltage 3.7 V Battery capacity 4500 mAh Battery charging Via a USB Type-C connector during non-use periods Displayed parameters Impedance plethysmography signal in the time domain, time stamps, EIPextracted parameters (at least venous capacity and outflow values), data on EIM status (charge level, measurement status, data recording status) User adjustable settings Selection between single-frequency and multi-frequency measurements, the number of excitation frequencies measured simultaneously, Excitation frequency, selection between bipolar and tetrapolar configurations, excitation current, sampling rate, electrode tension Physical and chemical characteristics Components EIM controller: Two controllers for bilateral measurement. EIM controller dimensions: 157 x 89 x 26 mm EIM controller weight: 257 g D6.1| Specifications v2.0 | 12 Nov 2025 30 Textile electrodes for both legs: Four electrodes for both legs: two outer electrodes for current injection and two inner for voltage measurement. Electrodes are adjustable and washable and integrated into a wearable system. Sterilization / cleaning of components Textile electrodes can be washed with mild detergent in common washing machines at 30°C. Raw materials For EIM controller: ABS plastic housing, lithium polymer battery (LP123497), PVC insulated electrode wires For the textile electrodes: Silver-plated fabric OEKO-TEX® STANDARD 100 certified Environmental conditions Transport and storage conditions Temperature: 0 to 50 °C Humidity: 15 to 95 % (non-condensing) Altitude: 0 to 3000 m Atmospheric pressure: 400 hPa to 1100 hPa Operating conditions Temperature: 10 to 30 °C Humidity: 30 to 75 % (non-condensing) Atmospheric pressure: 700 hPa to 1050 hPa Altitude: 0 to 3000 m Accessories, consumables, spare parts, and other components Accessories USB-C cable 1.5 m Electrode connection wires with safety connector and snap: Four wires for both legs Consumables / reagents - Spare parts Spare parts are to be replaced regularly Electroconductive textile-based electrodes keep their functions after several washing cycles if washing instructions are followed Other components - Maintenance Maintenance tasks Checking the integrity of wires, connectors, and external housing before each use. Washing textile electrodes regularly and between patients 3.4.2. Software Table 16 presents the EIM-related software specifications. Table 16. Electrical Impedance Module-related software specifications. General information Generic name EIM Software Type Software and firmware SDK and related firmware EIM Version Latest released version D6.1| Specifications v2.0 | 12 Nov 2025 31 Author TAU Dependencies Operating System Windows 11 or higher (64-bit) Runtime The latest .NET runtime Physical communication medium Wi-Fi, 802.11b/g/n Firmware programming language C (for ATSAM4S8B microcontroller) Software programming language C# (for PC) SDK functional scope Discovery mode mDNS (automatic IP address configuration) Device status IP, connected state Device configuration Selection between single-frequency and multi-frequency measurements The number of excitation frequencies measured simultaneously Excitation frequency Selection between bipolar and tetrapolar configurations Excitation current Sampling rate Real-time data streaming TCP protocol using a fixed binary structure that includes data from EIM sensor Heartbeat mechanism Maintains and detects connectivity, otherwise reverts to discovery mode Specialised algorithms Self-calibration Excitation waveform calculation Presentation of the measurement data as impedances Feature extraction Diagnostic classifiers GUI application A dedicated GUI application for the testing and maintenance of the EIM component Technical characteristics Functionalities Start measurement Stop measurement Record data Adjust measurement settings Non-Functionalities N/A User Interface (UI) UI Components Start button Stop button Device setting section D6.1| Specifications v2.0 | 12 Nov 2025 32 Results section Data Display Formats Real-time impedance signal EIP-extracted parameters Accessibility According to the medical device requirements Configuration Predefined Settings N/A Adjustable Settings Selection between single-frequency and multi-frequency measurements The number of excitation frequencies measured simultaneously: Up to 20 Excitation frequencies: 10-500 kHz Selection between bipolar and tetrapolar configurations Excitation current: Maximum peak amplitude 2.9 mA Sampling rate: 25, 50, 100 or 200 Hz Compatibility & Integration APIs EIM_SDK.dll Protocols Communication protocol: TCP Third-Party Integrations N/A Security Requirements Authentication N/A Data Privacy N/A (Data anonymity) Audit Logging Logging for user actions, system errors, and security events Maintenance Update Mechanism Manual installations for EIM device Error handling Exporting error stack trace and related metrics Communication & Connectivity Data Transmission TCP for communication Data Formats Sequences of vector data Data Storage Binary file, MATLAB (.mat file) 3.5. Light Rheography Module (LRM) 3.5.1. Hardware Table 17 presents the main hardware specifications of the light reflection module (LRM). Table 17. Light rheography module (LRM) specifications. General information Generic name Light rheography module Type Hardware and firmware D6.1| Specifications v2.0 | 12 Nov 2025 33 Version Latest released version Author medis Technical characteristics Main characteristics An optical rheography measuring channel is implemented in the device. The technical parameters can be found in detail in the table below. Displayed parameters Venous blood displacement volume, refilling time. User adjustable settings None Electrical and mechanical specification Power Rechargeable battery Charging method Wired charging using a magnetic force connector Charging power Max. 0,5 W (100 mA @ 5 V using an USB-A connector) Wavelength 950 nm Light source Infrared light, pulsed light source Frequency range 0 to 3 Hz Status display RGB LED Dimensions Almost round case, diameter: 45 mm; height: 15 mm Weight Approx. 20 g Length of the charging cable Approx. 50 cm Environmental Conditions Transport and storage conditions Temperature: 0 to 50 °C Humidity: 15 to 95 % (non-condensing) Altitude: 0 to 3000 m Atmospheric pressure: 400 hPa to 1100 hPa Operating conditions Temperature: 10 to 30 °C Humidity: 30 to 75 % (non-condensing) Atmospheric pressure: 650 hPa to 1050 hPa Altitude: 0 to 3000 m Safety tests IEC 60601-1 Not yet carried out IEC 60601-1-2 Not yet carried out IP Protection Not yet carried out Disinfection and sterilization Cleaning and Disinfection: The sensors must be carefully cleaned and disinfected after each skin contact Clean and disinfect the sensors by wiping all contact surfaces Follow the hygienic procedures of the users hospital/facility Compatible product: Cloth damped with a 70% isopropyl alcohol solution, e.g., Prolystica® / STERIS plc., terralin® protect / Schülke & Mayr GmbH D6.1| Specifications v2.0 | 12 Nov 2025 34 Biocompatibility test report Part of contact: Housing LRR Sensor (Not intended, but possible) Material identification and evaluation available User contact material: Thermoplastic polymer; Polypropylene copolymer; biocompatible Contact duration and classification acc. to EN ISO 10993-1 < 30 min / A - short time (<= 24 h) Accessories, consumables, spare parts, and other components Accessories Associated textile wearable Sterilization / cleaning of accessories Associated textile wearable can be washed in a common washing machine at 30°C, with all electronics removed. Consumables / reagents Associated textile wearable Spare parts None Maintenance Maintenance tasks The measurement system must always be kept in safe and reliable working order and should be checked regularly, at least once a year, by the manufacturer or authorized persons. The check covers: − Inspection for potential mechanical and functional damage − Readability and integrity of safety labels − Function test of the device. 3.5.2. Software Table 18 presents the LRM-related software specifications: Table 18. Light reflection module-related software specifications. General information Generic name LRM Software Type Software and firmware SDK and related firmware LRM Version Latest released version Author medis Dependencies Operating system Windows 11 or higher (64-bit) Runtime The latest .NET runtime Physical communication medium Wi-Fi, 802.11ac Firmware programming language C (for ESP32 microcontroller) Software programming language C# (for PC) SDK functional scope D6.1| Specifications v2.0 | 12 Nov 2025 35 Discovery mode mDNS (automatic IP address configuration) Device status and configuration IP, connected state, set appropriate LED current Real-time data streaming UDP protocol using a fixed binary structure that includes data from LRM sensor. Heartbeat mechanism Maintains and detects connectivity, otherwise reverts to discovery mode Specialised algorithms Calculation of venous blood displacement volume and refilling time GUI application A dedicated GUI application for the testing and maintenance of the LRM component. Technical characteristics Functional requirements Start measurement, stop measurement, record data, calculate parameters Non-functional requirements N/A Dependencies N/A User Interface (UI) UI components Start button, stop button, guidance pictures show feet positions, metronome sound Data display formats Real time curve of venous emptying and refilling phase, calculated parameters Accessibility According to the medical device requirements Configuration Predefined settings N/A Adjustable settings N/A Compatibility & Integration APIs LRM_SDK.dll Protocols Communication protocol: TCP Third-Party Integrations N/A Security Requirements Authentication N/A Data privacy N/A Audit logging Logging for user actions, system errors Maintenance Update mechanism Manual updates Error handling Export of system errors Communication & Connectivity Data transmission TCP for communication Data formats ASCII data Data storage Text file D6.1| Specifications v2.0 | 12 Nov 2025 42 − Scalable UI elements to accommodate various screen resolutions and highcontrast design for visibility. − Consideration for assistive technologies (e.g., screen readers) to enhance usability. Configuration Predefined Settings − WebSocket Port − IMU data header and format Adjustable Settings − Customizable exercise parameters (number of repetitions, sets, sequence order) for individual patient needs. − Options to adjust UI preferences and sensor data visualization settings (e.g., refresh rate, precision). Compatibility & Integration APIs − REST API endpoints for external control commands and status queries. − WebSocket API for real-time streaming of IMU sensor data. − Unity API for embedding the XR module within the CHI application. Protocols − TCP for REST API communications. − WebSocket protocol for live sensor data transmission. Third-Party Integrations Integration with the CHI WinForm C# application via process embedding of the Unity runtime. Security Requirements Authentication Handled by CHI Data Privacy Sensor data is handled in-memory Audit Logging Handled by CHI Communication & Connectivity Data Transmission − Real-time streaming via Websockets for IMU sensor data. − REST API communications over TCP for control and status updates. Data Formats − IMU sensor data transmitted as binary or text, depending on implementation requirements. − 3D models provided in .fbx format. Data Storage - D6.1| Specifications v2.0 | 12 Nov 2025 43 4. DVT monitoring methods and working modes The following chapter describes how the modules described in the previous sections will be used in different configurations to implement DVT monitoring methods. 4.1. Operator-independent Compression Duplex Ultrasonography for DVT monitoring The Compression Duplex Ultrasonography (CDUS) method is the main method used for DVT detection 6 . It combines B-mode ultrasound imaging with Doppler flow analysis to assess venous compressibility and blood flow. In this method, a transducer is placed over the vein while gentle pressure is applied. A healthy vein compresses completely under pressure, whereas a thrombus prevents full compression, indicating deep vein thrombosis (DVT). Doppler imaging further evaluates blood flow characteristics, detecting abnormalities such as reduced or absent flow, which can suggest the presence of a clot. This technique is widely used due to its high sensitivity and specificity for DVT diagnosis. However, this technique requires highly skilled medical professional or operator. To implement operator-independent post-surgical monitoring for DVT in a hospital setting, the ThrombUS+ system will combine an ultrasound module (USM) for imaging, a wearable actuator module (WAM) for controlled compression, and a central control and processing unit (CHI) to automate the procedure (see Figure 6). Figure 6. The ThrombUS+ subsystem for operator-independent Compression Ultrasonography: only the highlighted modules are used in this scenario. The system will operate in three modes: 6 A. W. Lensing et al., “Detection of deep-vein thrombosis by real-time B-mode ultrasonography,” N Engl J Med, vol. 320, no. 6, pp. 342–345, Feb. 1989 D6.1| Specifications v2.0 | 12 Nov 2025 44 Imaging: In this mode, the system is controlled by a medical specialist, i.e., operator. Upon activation, the WAM applies controlled, sequential compression to the patient’s thigh, mimicking the manual compression performed by a clinician. Simultaneously, the USM captures real-time ultrasound images of the deep veins in the compressed region. Advanced image processing algorithms automatically segment the veins, track their dimensions, and assess their response to compression. The primary objective of this mode is to enable the operator to evaluate the situation, calibrate the system, and personalize the detection thresholds for DVT. Monitoring: After ThrombUS+ CDUS subsystem is calibrated, it can be switched to autonomous working mode for DVT monitoring. In this mode, CHI will periodically initiate WAM-based compression and USMbased imaging procedure. If a thrombus is present, the system detects abnormal vein behavior, such as incomplete compression or reduced vein collapsibility, which are key indicators of DVT. CHI processes this data in real time and provides an assessment of DVT risk, displaying results for clinical review and initiating a visual and audible alert. Analysis: When the monitoring stage is finished, the analysis mode can be initiated manually by the operator. During periodic measurements in the monitoring mode, a dataset is generated, consisting of raw ultrasound data and time series of monitored parameters. The analysis mode enables deeper processing, analysis, and visualization of the collected data, identification of trends in extracted features, and potentially machine learning-based anomaly detection. Any working mode could be initiated and stopped on demand, i.e., at any time by the operator. By enabling operator-independent DVT monitoring, the system reduces reliance on specialized personnel, enhances accessibility to routine screenings, and ensures standardized, objective evaluations of venous thrombosis risk. 4.2. Operator-independent Venous Occlusion Plethysmography for DVT monitoring Venous Occlusion Plethysmography (VOP) is a non-invasive technique used to assess venous function and detect deep vein thrombosis (DVT) by measuring changes in calf volume during controlled venous occlusion. In this method, a cuff is inflated around the thigh to temporarily obstruct venous outflow while maintaining arterial inflow, leading to a gradual increase in calf volume. The rate of volume change is recorded using plethysmography sensors, such as electrical bioimpedance sensors. In healthy individuals, venous outflow resumes rapidly upon cuff deflation, whereas in DVT, the presence of a thrombus impairs venous drainage, leading to abnormal volume changes. By analyzing these hemodynamic patterns, VOP provides valuable insights into venous obstruction and serves as a diagnostic tool for DVT assessment. To enable an operator-independent VOP method in a hospital setting, the ThrombUS+ system will integrate an electrical impedance module (EIM) for imaging, a wearable actuator module (WAM) for controlled compression, and a central control and processing unit (CHI) to automate the procedure (see Figure 7). The WAM will repurpose the electronic actuator controller (EAC) from the CDUS subsystem, utilizing a wearable thigh cuff (WTC) instead of the C-shell-type semirigid wearable (CSW) used in CDUS. The VOP subsystem, similar to the CDUS, will operate in three modes: imaging, monitoring, and analysis. In imaging mode, the system is operated manually. Upon activation, the wearable actuator module (WAM) applies controlled compression to the patient’s thigh via the wearable thigh cuff (WTC), just enough to close the veins. As the WTC deflates, the electrical impedance module (EIM) simultaneously registers volume changes. The primary objective of this mode is to allow the operator to evaluate the situation, calibrate the system, and personalize the detection thresholds for deep vein thrombosis (DVT). Once the ThrombUS+ VOP subsystem is calibrated, it can be switched to monitoring mode for autonomous DVT detection. In this mode, the central control and processing unit (CHI) periodically initiates VOP sessions. If a thrombus forms, the system detects an abnormal relationship between venous filling and outflow. The CHI processes this data in real-time, assesses the DVT risk, and triggers an alert if the risk exceeds a predefined threshold. The results are displayed for clinical review, allowing healthcare professionals to take appropriate action. D6.1| Specifications v2.0 | 12 Nov 2025 45 Figure 7. The ThrombUS+ subsystem for operator-independent Venous Occlusion Plethysmography: only the highlighted modules are used in this scenario. After the monitoring phase, the analysis mode can be manually initiated by the operator. During periodic measurements in monitoring mode, the system generates a dataset consisting of raw bioimpedance data and time-series recordings of monitored parameters. The analysis mode enables in-depth data processing, visualization, and trend identification in extracted features. Additionally, this mode can support machine learning-based anomaly detection, further enhancing the accuracy of DVT assessment. Any working mode could be initiated and stopped on demand, i.e., at any time by the operator. 4.3. Operator-Independent Light Reflection Rheography-Based Monitoring of DVT Light Reflection Rheography (LRR) is an optical technique used for monitoring venous function (valve insufficiency) and detecting DVT by assessing changes in blood volume dynamics in the skin microcirculation. It relies on the optical reflectance properties of the skin, which vary based on blood volume fluctuations in the underlying venous network. An infrared light source illuminates the skin, typically over a superficial vein, while a photodiode measures the intensity of the reflected light. When venous occlusion occurs—either through controlled compression (e.g., with a thigh cuff) or natural changes in circulation—variations in blood volume alter the reflectance properties. A healthy venous system exhibits a rapid decrease in reflectance upon venous filling and a quick recovery after occlusion is released. However, in cases of DVT, the presence of a thrombus disrupts normal venous drainage, leading to delayed or reduced reflectance changes. Foot dorsiflexion or the tiptoe test in a sitting position is a standard procedure used to induce variations in blood volume in the lower limb. During this test, the contraction of calf muscles pushes blood out of the limb. If venous occlusion occurs, potentially due to a thrombus, light reflectance changes will appear, indicating a risk of DVT. However, the standard tiptoe test has an inherent repeatability issue, as it is patient dependent. D6.1| Specifications v2.0 | 12 Nov 2025 46 The repeatability of the test could be improved through synchronized movement guidance and patient response monitoring (see Figure 8). To achieve this, an optical light reflection module (LRM) will monitor reflectance changes, an extended reality and serious gaming module (XGM) will provide movement guidance, and a limb activity module (LAM) will track the patient's response. Additionally, the central control and processing unit (CHI) will automate the procedure. By integrating these components, LRR method enables not only standardized assessment but also intermittent monitoring of DVT, allowing for early detection of hemodynamic changes and increased DVT risk over time. Figure 8. The ThrombUS+ subsystem for LRM-based DVT monitoring: only the highlighted modules are used in this scenario. 4.4. Serious Gaming and Extended Reality for DVT prevention and rehabilitation DVT prevention can be enhanced through the integration of serious gaming and limb activity monitoring, promoting patient engagement and adherence to movement protocols. Serious games, designed with interactive and motivational elements, encourage regular physical activity, reducing the risk of venous stasis—a primary contributor to DVT. When combined with real-time limb activity monitoring, these systems provide objective feedback on movement patterns, also alerting users and healthcare providers to periods of inactivity. This dual approach not only improves patient compliance but also enables personalized interventions, ultimately reducing the incidence of DVT in at-risk populations, such as post-surgical patients or individuals with limited mobility. Additionally, patients of most of the ThrombUS+ RUCs (i.e. surgery, peripartum, obesity) are normally prescribed physical therapy exercises to promote mobilization, prevent thrombosis, and allow for proper patient physical rehabilitation. In the pilot use case chosen at proposal stage, namely, patients undergoing orthopedic surgery for the lower limb, the physical rehabilitation process includes lower limb exercises prescribed by a physiotherapist. The extended reality environment will provide the platform for these D6.1| Specifications v2.0 | 12 Nov 2025 47 exercises to be selected by the physiotherapist so that a dynamic physical therapy can be constructed. The extended reality will provide the patient with a virtual physiotherapy coach. The patient will have the opportunity to replay individual exercises and the entire physical therapy prescription and assess their own performance against the model exercises. Real-time visual guides will be provided to help the patient perform the prescribed exercise correctly. A summary of patient activity will be available for the physiotherapist to assess progress and compliance. Operator-independent serious gaming and the extended reality virtual physical exercise coach for DVT prevention and rehabilitation will be achieved by integrating the XGM module, which provides movement guidance; the LAM module, which tracks the patient’s response to guidance and monitors limb activity over time; and the CHI module, which automates the procedure (Figure 9). Figure 9. The ThrombUS+ subsystem for DVT prevention: only the highlighted modules are used in this scenario. In addition, incorporating an LRM could offer the advantage of monitoring the efficiency of prevention exercises by assessing subtle variations in limb movement and muscle engagement. This real-time feedback could further optimize exercise regimens, ensuring their effectiveness in promoting venous circulation and reducing DVT risk. 5. Critical specifications of the ThrombUS+ integrated product 5.1. Core functions This section outlines the core functional capabilities of the integrated ThrombUS+ system. 5.1.1. Ultrasound imaging for DVT detection in the lower limbs The novel ultrasound probes, in conjunction with the compact and portable beamformer provide real-time imaging of deep veins in the lower limbs. To achieve operator-independent tissue and therefore, vein D6.1| Specifications v2.0 | 12 Nov 2025 48 compression, the probe is embedded in a dedicated cuff that also contains two airbladders, connected to the wearable actuator module (WAM). The cuff also provides stability over prolonged use times. The integrated convolution neural network (CNN) models analyze the ultrasound images in real-time, giving feedback to the beamformer and the electronic actuator controller respectively, via the central hub & intelligence module. The key functionalities of the ThrombUS+ device, with respect to the ultrasonography are: 1. Real-time image acquisition and processing. 2. Automated image interpretation and DVT detection using AI models. 3. Intermittent or on-demand monitoring modes. 5.1.2. Plethysmography measurements for DVT detection in the lower limbs The ThrombUS+ device incorporates electrical impedance and light reflection sensors to assess blood flow within the lower limbs. The electrical impedance module in conjugation with the wearable actuator module works synergically to implement the operator independent venous occlusion method, which measures changes in electrical resistance due to blood pooling, indicative of thrombosis presence. The key functionalities of the ThrombUS+ device, with respect to the plethysmography sensors are: 1. Real-time data acquisition and analysis. 2. Automated data interpretation and DVT detection using AI models. 3. Intermittent or on-demand monitoring modes. 5.1.3. Extended reality environment and serious games for DVT prevention The ThrombUS+ device also integrates software for patient guidance and motivation on performing prescribed physiotherapy exercises for DVT prevention. The extended reality (XGM) leverages the limb activity module, embedded in the ThrombUS+ wearable, for real-time activity recognition and visualization with the aim of providing a virtual physiotherapy coach. The serious game module leverages the limb activity module as input to drive the game, thus providing an alternative means of engagement with physical activity and, ultimately, supporting DVT prevention. The key functionalities of the ThrombUS+ device, with respect to DVT prevention are: 1. Real-time visualization of lower limbs and guidance against model physical therapy exercises. 2. Patient motivation via serious gaming. 3. Automated interpretation of patient’s progress. 5.1.4. Integrated data processing and visualization The central hub and intelligence (CHI) module serve as the central node, for synchronizing and controlling the several services required by each measurement mode. In addition, it provides a user friend graphical interface for patients, clinicians and technicians. The key functionalities of the ThrombUS+ device, with respect to central hub are: 1. Real-time feedback on device’s status and sensor reading. 2. Efficient data processing when required. 3. User friendly interface with data visualization capabilities. D6.1| Specifications v2.0 | 12 Nov 2025 49 5.2. User flow This section introduces a user flow diagram that maps the entire lifecycle of the ThrombUS+ device. The diagram details how the device transitions from its initial configuration—where the healthcare professional sets up the device and adds a new patient—to a locked state that ensures secure patient usage. Figure 10 illustrates the user flow of the medical device, detailing the transition between different operational states: 1. Device starts in a configuration state. 2. The clinician (Healthcare Professional - HP) configures the device, such as adding a new patient. 3. After patient selection, the clinician sets up the patient specific device configuration, such as monitoring plan, physiotherapy prescription, and locks the device. If the device is not locked the clinician can have access to manual imaging modes, or to the analysis mode if patient data are available. 4. When the device is locked, patient monitoring mode is activated. 5. After the patient session ends, the device is unlocked by the clinician and data is reviewed. 6. The clinician can either repeat for another patient or conclude the session flow. Figure 10. ThrombUS+ device user flow diagram. The process begins in a configuration state, where the Healthcare Professional (HP) sets up the device and prepares it for patient monitoring mode. Once a patient is selected and their profile is configured, the device is locked, ensuring restricted access during the session. D6.1| Specifications v2.0 | 12 Nov 2025 50 During the session, the patient interacts with the device while it continuously monitors progress and collects sensor data. At the end of the session, the HP unlocks the device, reviews the recorded data, and analyzes patient progress. The HP can then either prepare the device for another patient or conclude the session flow, completing the operational cycle. This structured flow ensures that the device operates securely, efficiently, and in compliance with medical protocols, facilitating both patient treatment and professional oversight. 5.3. User interfaces This section outlines the essential interface design for the ThrombUS+ device, highlighting the clear delineation of user roles and access levels within the device’s workflow. Figure 11 presents the workflow of system usage, helping users understand how the device is configured, locked in patient monitoring mode, and reviewed post-session. It ensures that clinicians control device access, manage patient interactions, and oversee data collection, while patients can only access relevant functions in View Mode. Figure 11. System roles and views. 5.4. System roles and user interface features This section highlights how the user interface (UI) segregates access by defining distinct roles, ensuring that each user, whether a technician, clinician, or patient, interacts only with the functions pertinent to their responsibilities, thereby enhancing both security and usability. D6.1| Specifications v2.0 | 12 Nov 2025 51 At the top of Figure 11, the diagram illustrates three primary system roles: − Clinician (Healthcare Professional - HP): Assist patients on wearing and using the device managing patient data, configuring exercises, and reviewing treatment progress. − Patient: Interacts with the system in a controlled environment where functionalities are restricted to their prescribed exercises and health data. − Technician: Responsible for setting up the device (e.g. assembling the sensors after disinfection), for maintaining the device (e.g. software or sensor update) as well as troubleshooting (e.g. calibration and communication of sensors). 5.5. Functional Overview The interface is structured into three primary views that dictate available features: Maintenance, Management & Patient view: − Maintenance View (Technician Only): Provides access to device settings, maintenance, calibration, and game management. Technicians ensure optimal system performance, perform necessary calibrations, and manage game content. − Management View (Clinician Only): Enables patient management, exercise studio, and data management. Clinicians configure patient data, customize exercise plans, and handle medical records while ensuring compliance with treatment protocols. − Patient View (Clinician & Patient with distinct access levels): Grants patients access to their prescribed exercises, therapeutic games, health status, and dashboard. Simultaneously, clinicians can review prescription plans, patient history, manual exams, and customizations. This structured access ensures that patients interact only with their assigned treatment programs, while clinicians retain control over medical decisions and session reviews, maintaining data security and integrity. This structured approach ensures that clinicians retain full control over patient management and patients engage only with their prescribed therapy, enhancing security, usability, and compliance within a clinical environment. Moreover, technicians can maintain and fix any potential issues without interfering with the clinician and patient usage. 5.6. Testing, verification, and validation This document establishes critical product specifications, subsystems, use case scenarios, and sub scenarios. This information will be used to design a testing strategy to ensure compliance with predefined standards and specifications. Based on this document, a testing methodology and test plan will be developed in task T6.2 to demonstrate that the ThrombUS+ system and its subsystems meet the requirements specified in T2.3 and T6.1, as well as the relevant regulations and standards analyzed in T2.2. Task T6.2 will define a general process for testing, documenting test evidence, and managing test failure processes, including feedback on the overall design of the ThrombUS+ system. Testing will focus on the integrated beta prototype while also addressing individual modules of the alpha prototype. The testing plan will outline the types of tests, descriptions of specific test environments, responsibilities for each test, required equipment or instruments, ergonomic assessments, and other organizational requirements. 5.7. Potential ThrombUS+ system improvements and extensions There are several avenues for improving the ThrombUS+ system development process, mitigating risks, and exploring potential research directions: Model-based methodology (MBM) for embedded systems design. MBM enhances abstraction and system understanding by allowing designers to create high-level models, making it easier to visualize and verify