A Standardized Technique for Ultrasound–MRI Fusion Imaging of the Carotid Artery for Vascular Assessment

Abstract

Purpose:

To describe a reproducible ultrasound–magnetic resonance imaging(MRI) fusion imaging technique developed for quantitative assessment of carotid wall strain and plaque morphology.

Methods:

MRI volumes (Oblique T2 Cube) were imported into the Sequoia ultrasound system equipped with an electromagnetic tracking module. The fusion workflow aligned ultrasound and MRI coordinate spaces through anatomical landmark registration of the internal and external carotid arteries. Real-time B-mode and strain imaging were performed with a 10L4 transducer. The fused MRI volume provided spatial context for localization of plaque and intima-media regions. Following alignment, additional B-mode, color Doppler, and radiofrequency (RF) data were acquired on a Siemens S2000/S3000 system to match the imaging plane used during fusion.

Results:

This workflow enabled consistent cross-modality localization of vascular features between MRI and ultrasound, improving repeatability of acquisitions. The S2000 acquisition step ensured that color Doppler and RF data corresponded to the same anatomical orientation defined during fusion.

Conclusion:

This technique offers a practical framework for ultrasound–MRI fusion in carotid imaging and facilitates direct localization of similar structures across two imaging modalities. This approach also enables visualization across different system with manual visual corroborations, that in this case enabled standardized RF data collection for quantitative elastography research.

1. Introduction

Ultrasound-MRI fusion imaging offers a powerful method for improving the spatial consistency of vascular ultrasound studies.(1) In carotid imaging, the reproducibility of cross-sectional views is highly sensitive to small variations in transducer orientation, imaging plane selection, and patient anatomy. These factors introduce significant challenges for studies requiring session-to-session consistency or cross-platform data acquisition.(2) Fusion imaging addresses this by aligning real-time ultrasound with high-resolution MRI volumes, providing a consistent anatomical reference that can guide probe positioning and image interpretation.

MRI is widely recognized for its ability to visualize soft tissue morphology, vessel wall structure, and plaque components with high spatial resolution. However, its cost, acquisition time, and lack of real-time capability make it impractical for routine use in mechanical or functional imaging tasks.(3) Ultrasound, in contrast, enables dynamic, bedside evaluation but suffers from limited reproducibility due to its operator-dependent nature(2). By integrating these modalities through image fusion, it becomes possible to combine the anatomical precision of MRI with the real-time feedback of ultrasound.

Recent advancements in ultrasound systems equipped with electromagnetic tracking have made real-time Ultrasound-MRI fusion increasingly feasible. These systems can register live ultrasound images to pre-acquired MRI datasets, allowing users to navigate in the MRI-defined coordinate space.(4) Despite the growing availability of such platforms, detailed implementation protocols for vascular applications—particularly those requiring high spatial reproducibility—are still limited.(1)

This Technical Note presents a standardized, step-by-step fusion imaging workflow developed for carotid artery studies, using the Siemens Sequoia platform for fusion alignment and Siemens S2000/S3000 systems for radiofrequency (RF) data acquisition. The protocol is optimized for research scenarios where precise plane matching across systems is essential, such as in ultrasound-based strain imaging. Special emphasis is placed on the landmark registration and plane matching steps that enable reproducible alignment of ultrasound images to MRI anatomy, forming the basis for consistent data acquisition and longitudinal analysis.

2. Materials and Equipment

• Ultrasound Systems: Siemens Acuson Sequoia (with Fusion module and electromagnetic tracking); Siemens Acuson S2000/S3000 for RF data collection

• Transducers: 5V1, 10L4, 18L6 (10L4 used primarily for fusion)

• MRI Protocol: Oblique T2 Cube sequence of the carotid bifurcation

• Tracking Hardware: Electromagnetic transmitter and probe-mounted sensors

• Data Transfer: Encrypted USB for secure MRI import

• Software Presets: Sequoia Carotid Strain preset for real-time elastography

• Personnel: Experienced sonographer and second operator for registration and workflow management

3. Fusion Imaging Protocol

3.1. MRI Data Import

MRI volumes were imported into the Sequoia system via USB using the Review USB > Import Data function. A new study is created under the Local Database, and the 10L4 probe is selected with the Carotid Strain preset.

3.2. Tracking Initialization

The electromagnetic transmitter was connected to the Sequoia’s transmitter port; the probe’s tracking sensor was connected to ports 1–3 on the left aligning on the port number of the transducer. The system will confirm successful initialization with an on-screen message (“Tracking Initialized”), indicating real-time probe localization within the electromagnetic field.

3.3. Fusion Workflow Setup

The MRI volume can be loaded by selecting Workflow > Fusion. Brightness and contrast adjustments can be made via the touchscreen to enhance vessel visibility, particularly at the carotid bifurcation.

3.4. Landmark Registration and Alignment

Accurate registration is essential for aligning ultrasound and MRI datasets. This process begins with coarse alignment: the sonographer positions the 10L4 probe in a transverse orientation. A sweep is performed from the distal common carotid artery (CCA) through the proximal bifurcation of the internal (ICA) and external (ECA) branches.

The MRI volume image is then rotated on the Z axis to match the orientation of the ICA and ECA on the ultrasound image (Figure 1). Once alignment is achieved, the operator selects points to initiate landmark-based registration of the ICA and ECA.

Figure 1:

Example placement of points used for alignment in the axial plane for a representative subject. Points are placed at the wall regions of the ICA and ECA as well as the bifurcation region for both the MRI image (A) and Ultrasound image (B).

• Anatomical landmarks are manually selected on both modalities, typically the origins of the ICA and ECA.

• To minimize misalignment due to patient motion or drift, 2–3 corresponding points should be placed within a short time frame (ideally within 15 seconds).

Upon placing the final point, the operator will initiate and complete a rigid registration of the ultrasound and MRI coordinate systems. An example of accurate point placement and registration is shown in Figure 1.

The system displays live ultrasound beside the registered MRI slice, updating in real time with probe motion. If misalignment is detected, previously marked points can be deleted, and the registration repeated. This iterative capability ensures precise overlap of vascular structures. The accuracy of landmark registration directly affects downstream strain reproducibility. Close attention to vascular bifurcation geometry in both image modalities and prompt initialization minimize registration drift.

3.5. Plane Matching and Confirmation

Following successful registration, the goal is to fine-tune the probe orientation so that the live ultrasound cross-section precisely matches the MRI geometry in the longitudinal plane.

The sonographer incrementally adjusts probe position while visually comparing the live ultrasound image to the registered MRI plane until the vessel can be visualized in the longitudinal plane as shown in Figure 2.

Figure 2:

Longitudinal view obtained from the same registration example shown in Figure 1. Two of the three registration points remain visible in this imaging plane; the third point (MK03), placed in the ECA during axial registration, is outside the longitudinal imaging plane and therefore not visible. MRI image (A) and Ultrasound image (B).

Alignment is considered optimal when:

• The ICA and ECA branches show consistent branching angles and wall curvature

• Morphological features—such as eccentric plaque, wall thickening, or bifurcation contour—are correctly aligned between modalities

Once a high-confidence match is reached, the view is frozen to record the transducer’s spatial orientation. This defines the fusion-matched plane, serving as the anatomical reference for RF data acquisition. (Figure 3)

Figure 3:

Independent example of longitudinal registration from a different subject, with labeled vessel and plaque regions. Landmark identifiers shown in this figure are specific to this registration instance. MRI image (A) and Ultrasound image (B).

4. Cross-System Data Acquisition

With the fusion-matched plane established, imaging transitions to the Siemens S2000 or S3000 system for additional B-mode, color Doppler, and RF data collection. The sonographer uses the frozen Sequoia image as a reference to manually replicate the same imaging plane.

Special attention is given to reproducing bifurcation geometry, wall contours, and transducer orientation (including tilt and rotation).

RF data is acquired without beam steering and named using a structured convention: patXXXX_a2_r_fusion_match_[timestamp].

5. Practical Considerations

The full fusion and RF acquisition process typically takes 10 to 15 minutes per subject. Consistent patient positioning, particularly neck position and transducer pressure, is essential to reduce anatomical distortion between modalities. To prevent interference with electromagnetic tracking, metallic objects near the scanning field should be removed in advance.

Although in-phase/quadrature (IQ) data is available on the Sequoia for research purposes, the system exits fusion mode when setup for IQ data acquisition. Similar visual registration as described for RF data acquisition would have to be implemented for IQ data acquisition.

Limitations of this protocol include operator-dependent landmark placement, potential registration drift due to patient motion, and susceptibility of electromagnetic tracking to metallic interference. In accordance with manufacturer recommendations for the tracking system used in this study, patients with implanted cardiac pacemakers should not undergo fusion imaging with active electromagnetic tracking.

6. Discussion

This Technical Note outlines a reproducible Ultrasound-MRI fusion imaging technique specifically tailored for vascular applications, with a focus on spatial alignment and cross-system consistency. By leveraging rigid registration and precise plane matching on the Siemens Sequoia platform, the workflow enables high-fidelity localization of carotid anatomy that can be replicated on other ultrasound systems for advanced data acquisition.

The protocol provides a generalizable framework for any vascular study requiring multimodal correlation or consistent imaging across sessions. The ability to anchor real-time ultrasound to MRI-derived anatomy enhances anatomical fidelity, improves reproducibility, and supports integration with downstream techniques such as elastography, Doppler analysis, or contrast-enhanced imaging.

The novelty of the report lies in the application-specific protocol, including identification of carotid bifurcation landmarks, strategies for registration of the ICA and ECA, longitudinal plane matching, and the subsequent transfer of the fusion-defined imaging plane to a second ultrasound platform for RF-based strain imaging. While individual components of the workflow are available within commercial fusion systems, to our knowledge a detailed protocol describing their use for carotid artery fusion imaging and cross-system vascular data acquisition has not been previously reported. The protocol can also be readily adapted to computed tomography (CT) datasets on fusion platforms that support CT image import and registration.

By standardizing the fusion process and highlighting the critical steps of landmark registration and plane matching, this workflow addresses a key gap in vascular imaging methodology. It offers a practical foundation for future studies that aim to combine the complementary strengths of MRI and ultrasound in the investigation of vascular structure, function, and disease progression.

7. Conclusions

This technique provides a reproducible method for aligning ultrasound and MRI data in carotid ultrasound and using visual registration between systems for elastography studies. By emphasizing landmark registration and plane matching, the workflow ensures anatomical consistency. The Acuson Sequoia system serves as a fusion workstation, leveraging MRI guidance for precise anatomical localization, while the Acuson S2000/S3000 platforms are used to collect high-fidelity RF data for strain analysis. While the current process relies on manual landmark placement, it provides sufficient accuracy for rigid vascular targets like the carotid bifurcation.