By: Thalia Kaylyn Averil
Artificial Intelligence (AI) is experiencing rapid growth recently. Its impact is not only limited to specific fields, but has also begun to contribute to the healthcare world. One example is the use of AI in Magnetic Resonance Imaging (MRI). AI can significantly improve MRI capabilities in several ways, such as improving image quality, enabling faster and more accurate interpretation, detecting certain diseases or health conditions, and assisting in more personalized treatment planning for each patient.
In recent years, one of the advancements in MRI is the increasingly sophisticated techniques of data acquisition and reconstruction. MRI scans can be performed more quickly with an undersampling approach, which reduces the amount of data required for image acquisition. The quality of the generated images remains preserved with advanced reconstruction methods, even when performed more quickly. The reconstruction methods are designed meticulously to minimize image domain artifacts, resulting in high-quality images. Deep learning, a part of AI, plays a role in advancing these reconstruction methods used in MRI. Deep learning can learn and make decisions from large amounts of medical data using algorithms inspired by the structure and function of the human brain to process data and recognize decision-making patterns. This approach not only speeds up the MRI scanning process but also enhances image clarity and accuracy.
AI can also optimize MRI workflows by automating various stages of the imaging process, such as planning, data collection, reconstruction, parameter mapping and segmentation, thereby increasing overall MRI efficiency. AI can overcome limitations commonly encountered in MRI, such as improving patient motion tolerance. AI can also enhance the quality of MRI images to detect diseases and plan further medical actions more accurately. Additionally, the use of AI in MRI enables automated diagnosis and prognosis determination, aiding in the detection of various diseases. Advances in AI within MRI can further impact the introduction of new applications to guide procedures, such as low-field MRI and real-time MRI, thereby expanding the scope and effectiveness of MRI technology.
MRI image analysis can be automated with AI to enhance efficiency and accuracy of results. AI simplifies and improves various aspects of MRI, such as accurately segmenting structures in images for diagnosis. AI can help doctors identify potential problems by automatically detecting and classifying lesions, tumors, etc. Tissue characteristics can be precisely measured using AI algorithms for quantitative analysis (e.g. measuring tumor volume and detecting changes that occur over time).
AI plays a crucial role in determining diagnoses based on MRI images, providing doctors with a second opinion that can reduce the likelihood of human error. Additionally, AI can integrate MRI data with patients’ Electronic Health Records (EHR), enabling doctors to have a comprehensive view of patients’ health histories for better decision-making. It can also improve the accuracy of diagnosis and treatment plans. Prognosis determination and risk assessment are examples of AI’s role in healthcare for predictive analysis. By analyzing MRI images and patient data, AI can estimate disease progression, evaluate risks, and assist in tailoring personalized treatments. AI can also detect early signs of disease that are often missed, allowing for timely intervention.
AI also plays a role in advancing research in the fields of radiomics and radiogenomics. Radiomics involves the extraction and analysis of a large number of quantitative features (e.g., size, shape, intensity, texture) from medical images to aid decision making. On the other hand, radiogenomics integrates radiomics with genomic data to understand the relationship between imaging features and genetic information, which can potentially predict outcomes and guide treatments. Additionally, AI can enhance participant selection in clinical trials by analyzing MRI scans and other data, thereby enabling more efficient identification of suitable participants.
One of the newest technologies at Abdi Waluyo Hospital is an AI software called AIR Recon DL that has been used on MRI machines. AIR Recon DL allows doctors to reconstruct high quality images in less time, thereby increasing diagnostic accuracy. This technology utilizes a deep learning approach to process MRI images to produce clear and accurate images. In addition, automatic analysis is one of the features of AIR Recon DL that can help differentiate tumors, infections and soft tissue in order to determine the most accurate and effective treatment plan for each patient. With AIR Recon DL, doctors at Abdi Waluyo Hospital can optimize the diagnostic process and provide the best service to patients.
Resources
- Shimron E, Perlman O. AI in MRI: Computational Frameworks for a Faster, Optimized, and Automated Imaging Workflow. Bioengineering (Basel). 2023 Apr 20;10(4):492. doi: 10.3390/bioengineering10040492. PMID: 37106679; PMCID: PMC10135995. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10135995/
- Kang J, Nam Y. Applications of Artificial Intelligence in MR Image Acquisition and Reconstruction. J Korean Soc Radiol. 2022 Nov;83(6):1229-1239. Korean. doi: 10.3348/jksr.2022.0156. Epub 2022 Nov 30. PMID: 36545429; PMCID: PMC9748458. Available from: https://pubmed.ncbi.nlm.nih.gov/36545429/
- Gaube S, Suresh H, Raue M, Merritt A, Berkowitz SJ, Lermer E, Coughlin JF, Guttag JV, Colak E, Ghassemi M. Do as AI say: susceptibility in deployment of clinical decision-aids. NPJ Digit Med. 2021 Feb 19;4(1):31. doi: 10.1038/s41746-021-00385-9. PMID: 33608629; PMCID: PMC7896064. Available from: https://pubmed.ncbi.nlm.nih.gov/33608629/
- Fan H, Luo Y, Gu F, Tian B, Xiong Y, Wu G, Nie X, Yu J, Tong J, Liao X. Artificial intelligence-based MRI radiomics and radiogenomics in glioma. Cancer Imaging. 2024 Mar 14;24(1):36. doi: 10.1186/s40644-024-00682-y. PMID: 38486342; PMCID: PMC10938723. Available from: https://pubmed.ncbi.nlm.nih.gov/38486342/
