Historic Review of MRI Acceleration Technology Development
The journey of MRI acceleration technology has been marked by several significant advancements. Initially, MRI scans were time-consuming, which limited their clinical utility. The introduction of parallel imaging (PI) in the 1990s was a watershed moment. PI techniques, such as Sensitivity Encoding (SENSE) and Generalized Autocalibrating Partially Parallel Acquisitions (GRAPPA), leveraged the redundancy in multi-coil data to accelerate image acquisition. These methods significantly reduced scan times while maintaining image quality1.
Another major development was Compressed Sensing (CS), which utilized the sparsity of MR images to reconstruct high-quality images from undersampled data. CS techniques allowed for even faster scans by acquiring fewer data points in k-space, the frequency domain where MRI data is initially captured. However, these methods had limitations in robustness and required complex reconstruction algorithms2.
Need for Even Higher Speed Acceleration
In recent years, the healthcare system in Vietnam has faced significant challenges due to high patient loads. Most hospitals are outdated and face chronic overcrowding. Hospitals in Hanoi and Ho Chi Minh City receive up to 60% of the country’s patients and operate at 200% of the capacity4. This situation has created an urgent need for faster and more efficient diagnostic tools to manage the high patient volumes.
Deep Resolve Boost addresses this need by providing even higher speed acceleration for MRI scans. The ability to perform faster scans without compromising image quality is crucial in managing the high patient loads. By reducing scan times, we can accommodate more patients, alleviate overcrowding, and ensure timely diagnoses and treatments. This technology has been instrumental in helping us cope with the increasing demand for medical imaging services3.
Introduction to Deep Learning Reconstruction
Deep Learning Reconstruction represents a groundbreaking advancement in MRI scan acceleration. This technology leverages the power of artificial intelligence and deep learning algorithms to reconstruct high-quality images from raw MRI data. Unlike traditional methods, Deep Learning Reconstruction can handle complex data patterns and noise, resulting in clearer and more accurate images5.
Deep Resolve Boost, a part of the Deep Resolve family, applies AI to image reconstruction and uses deep learning for noise reduction. This approach not only accelerates MRI scans but also enhances the Signal-to-Noise Ratio (SNR) of MR images, ensuring high-resolution images even at faster scan speeds. The combination of deep learning with parallel imaging (PAT) and Simultaneous-Multi-Slice (SMS) has revolutionized the speed and efficiency of MRI scans, making it a crucial tool in modern medical imaging6.
Faster Scan Times, Better Patient Experience
One of the most significant benefits we've experienced with Deep Resolve Boost is the drastic reduction in scan times. By combining deep learning with parallel imaging (PAT) and Simultaneous-Multi-Slice (SMS), we have been able to cut down our MRI scan times by up to 73 percent. This has allowed us to accommodate more patients in a day, reducing wait times and enhancing the overall patient experience.
Enhanced Image Quality
Despite the faster scan times, the image quality has not been compromised. Deep Resolve Boost has significantly improved the Signal-to-Noise Ratio (SNR) of our MR images, providing us with high-resolution images that are crucial for accurate diagnoses. Our radiologists have been impressed with the clarity and detail of the images, which have made their job easier and more efficient.
Streamlined Workflow
The integration of Deep Resolve Boost into our existing MR XA50 software has been seamless. The technology works in combination with GRAPPA in the Resolution, Acceleration tab card within a Turbo Spin Echo sequence, simplifying the workflow for our radiologists and technologists. This has allowed our team to focus more on patient care and less on adjusting technical parameters.
Increased Clinical Capacity
With the ability to perform faster scans and obtain high-quality images, we have been able to handle a higher volume of patients. This has not only boosted our clinical productivity but also improved patient satisfaction. The positive feedback from our patients has been overwhelming, and we are proud to offer them the best possible care with the help of Deep Resolve Boost.
Case Sharing: Success Stories with Deep Resolve Boost
Case Study 1: Enhancing Diagnostic Accuracy in MSK Imaging
At UMC, we implemented Deep Resolve Boost to improve our diagnostic accuracy in musculoskeletal (MSK) imaging. One notable case involved a patient with a complex joint condition. The enhanced SNR provided by Deep Resolve Boost allowed our radiologists to detect subtle abnormalities that were previously challenging to identify. This led to a more accurate diagnosis and timely intervention, significantly improving the patient's outcome.
Image courtesy of University Medical Center (UMC), Ho Chi Minh City, Vietnam
Figure 1. Image comparison between MAGNETOM Altea 1.5T with Deep Resolve Boost (left) and MAGNETOM Verio 3T (right). The scan time for the left images was 49.57 seconds, while for the right images it was 1 minute and 50 seconds. The matrix size for the left image was 0.23 x 0.23 x 3 mm, and for the right image it was 0.50 x 0.50 x 3.5 mm. The Deep Resolve Boost image demonstrated higher resolution and more anatomical details, achieving a 54% reduction in scan time. This indicates that the MAGNETOM Altea 1.5T, enhanced by Deep Resolve Boost, outperforms the 3T model.
Case Study 2: Improving Workflow Efficiency in MSK Imaging
At UMC, we implemented Deep Resolve Boost to improve our diagnostic accuracy in musculoskeletal (MSK) imaging. One notable case involved a patient with a complex joint condition. The enhanced SNR provided by Deep Resolve Boost allowed our radiologists to detect subtle abnormalities that were previously challenging to identify. This led to a more accurate diagnosis and timely intervention, significantly improving the patient's outcome.
Image courtesy of University Medical Center (UMC), Ho Chi Minh City, Vietnam
Figure 2. Image comparison between MAGNETOM Altea1.5T with Deep Resolve Boost (left) and MAGNETOM Avanto 1.5T (right). Scan time for the left images was 28.25 seconds, and for the right images was 49.27 seconds. The matrix size for the left image was 0.37 x 0.37 x 4 mm, and for the right image was 0.63 x 0.63 x 4.5 mm. Higher SNR and more anatomical details were observed from the Deep Resolve Boost image with a 43% scan time reduction.
Conclusion
Deep Resolve Boost has truly transformed our MSK MRI imaging processes at UMC. The combination of faster scan times, enhanced image quality, and streamlined workflow has made a significant impact on our clinical productivity and patient care. I highly recommend Deep Resolve Boost to any healthcare facility looking to enhance their imaging capabilities and improve patient outcomes.
About the University Medical Center (UMC)
The University Medical Center (UMC) in Ho Chi Minh City is one of the largest public hospitals in Vietnam. Established in 1994, UMC is the first combined hospital-university model in the country. The hospital has achieved significant milestones in patient care, training, and research, establishing itself as a leading university hospital. UMC is committed to providing optimal healthcare solutions through the integration of treatment, research, and training, focusing on pioneering various areas of expertise, including diagnostic imaging.
Grand Opening of MAGNETOM Altea at UMC
Recently, we celebrated the grand opening of the MAGNETOM Altea 1.5T MRI system at UMC, which features Deep Resolve Boost technology. This event marked a significant milestone in advancing medical imaging technology. The MAGNETOM Altea MRI system embodies the fusion of technology and compassion, heralding a future where constraints are shattered and barriers are transcended.
Innovative Features
The MAGNETOM Altea MRI system at UMC is equipped with several groundbreaking features:
- Compressed Sensing GRASP VIBE Technology: This innovation eliminates motion picture noise, transcending physical barriers, and ensuring access to care for a broader patient demographic, including geriatric and pediatric patients.
- BioMatrix Technology: This technology adapts to the uniqueness of each patient's body, allowing personalized examinations and expanding the frontiers of precision medicine.
- Deep Resolve Boost: AI-powered image reconstruction technology, coupled with Compressed Sensing and Simultaneous Multi-Slice capabilities, revolutionizes the speed and efficiency of patient scans.
Võ Tấn Đức is a distinguished Associate Professor and the Head of the Department of Diagnostic Imaging at the University Medical Center Ho Chi Minh City. His career has been dedicated to the advancement of medical imaging and surgery. Dr. Đức achieved his PhD in Digestive Surgery from the University of Medicine & Pharmacy, HCMC, in 2015, and was honored with the title of Associate Professor in 2023.
With an impressive academic foundation, including undergraduate and master's degrees in General Surgery from the same institution, Dr. Đức has held various significant positions throughout his professional journey. He has served as a resident doctor and lecturer in both General Surgery and Diagnostic Imaging, with international residencies in France at the University of Marseille II and the University of Clermont-Ferrand.
Dr. Đức has made notable contributions to medical research, with numerous publications in esteemed journals such as Scientific Reports, Frontiers in Oncology, and European Radiology. His research primarily focuses on the applications of magnetic resonance imaging and deep learning models in medical diagnostics, particularly in the evaluation and treatment of complex conditions like hepatocellular carcinoma and thymic epithelial tumors.
Residing in Ho Chi Minh City, Dr. Đức continues to lead and innovate in the field of diagnostic imaging, significantly contributing to both academic and clinical advancements.
