![AI IN RADIOLOGY MODALITIES
• AI shows promising results in detection of additional lung nodules in
patients with complex lung diseases [1]
• It is also an important tool in detecting additional Intracranial lesions [2],
increased sensitivity and accuracy in diagnosing cerebral aneurysm [3]
and also reduces time of aneurysm detection [4]
• Overall increase in sensitivity, accuracy and reduced workflow is recorded
in AI-Based diagnosis in computed tomography [5,6,7,8]](https://image.slidesharecdn.com/arn2024-250809150742-e0c03339/85/ARTIFICIAL-INTELLIGENCE-AN-EMERGING-TECHNOLOGY-IN-IMAGING-9-320.jpg)
![AI IN RADIOLOGY MODALITIES
• There are many AI applications in the interpretive domains of lesion
detection and diagnosis, triage and density assessment and also
noninterpretive domains of risk assessment, image quality control, image
acquisition and dose reductions [9]
• In Mammography, AI proved to be an aiding tool by serving as the first
reader [10] and as such reduced the overall workload of breast
radiologists with increased diagnostic accuracy [5,11]](https://image.slidesharecdn.com/arn2024-250809150742-e0c03339/85/ARTIFICIAL-INTELLIGENCE-AN-EMERGING-TECHNOLOGY-IN-IMAGING-10-320.jpg)
![AI IN RADIOLOGY MODALITIES
• Researches on AI in MRI shows increased detection with Functional
connectivity MRI (FCM) algorithm [12], AI is also useful by accurately
locating and segmenting acute ischemic stroke lesions and the effects of
early rehabilitation [13]. By apply Faster R-CNN in evaluating
Circumferential Resection Margin status of rectal cancer patients in MRI,
it shows increased accuracy and sensitivity [14,15].](https://image.slidesharecdn.com/arn2024-250809150742-e0c03339/85/ARTIFICIAL-INTELLIGENCE-AN-EMERGING-TECHNOLOGY-IN-IMAGING-11-320.jpg)
![AI IN RADIOLOGY MODALITIES
• Deep transfer learning also shows accuracy of up to 98.01% in
differentiation of breast lesions on dynamic contrast enhanced MRI (DCE-
MRI) [16,17]](https://image.slidesharecdn.com/arn2024-250809150742-e0c03339/85/ARTIFICIAL-INTELLIGENCE-AN-EMERGING-TECHNOLOGY-IN-IMAGING-12-320.jpg)
![AI IN RADIOLOGY MODALITIES
• Increasing throughput while maintain excellence is one of the main goals
of every department and studies in nuclear medicine shows increased
throughput, AI-driven theranostic drug discovery and labelling, dose
precision and as such increased accuracy and dose reduction [18]
• AI also helps in assessing image quality while using FDG-18 in whole
body imaging [19]
• AI-radiomics is also an important tool in research [20]](https://image.slidesharecdn.com/arn2024-250809150742-e0c03339/85/ARTIFICIAL-INTELLIGENCE-AN-EMERGING-TECHNOLOGY-IN-IMAGING-13-320.jpg)
![LIMITATIONS
• AI works based on data it was trained with, so when integrated in a
different population it might cause inequities [5]
• Risk of leakage of patient data privacy which can be as a result of too
much access of AI models to patient data [5,6,7]
• Integration of AI to existing institution requires harmony between AI and
medical experts to avoid undermining patient care [6]
• The black box Enigma [18]](https://image.slidesharecdn.com/arn2024-250809150742-e0c03339/85/ARTIFICIAL-INTELLIGENCE-AN-EMERGING-TECHNOLOGY-IN-IMAGING-14-320.jpg)
![18. Saboury, B., Bradshaw, T., Boellaard, R., Buvat, I., Dutta, J., Hatt, M., Jha, A. K., Li, Q., Liu, C.,
McMeekin, H., Morris, M. A., Scott, P. J. H., Siegel, E., Sunderland, J. J., Pandit-Taskar, N., Wahl, R. L.,
Zuehlsdorff, S., & Rahmim, A. (2022). Artificial intelligence in nuclear medicine: Opportunities,
challenges, and responsibilities toward a trustworthy ecosystem. Journal of Nuclear Medicine, 64(2),
188-196.
19. Qi, C., Wang, S., Yu, H., Zhang, Y., Hu, P., Tan, H., Shi, Y., & Shi, H. (2023). An artificial intelligence-
driven image quality assessment system for whole-body [(18) F] FDG PET/CT. European Journal of
Nuclear Medicine and Molecular Imaging, 50(5), 1318-1328.
20. Tomaszewski, M. R., Fan, S., Garcia, A., Qi, J., Kim, Y., Gatenby, R. A., Schabath, M. B., Tap, W. D.,
Reinke, D. K., Makanji, R. J., Reed, D. R., & Gillies, R. J. (2022). AI-radiomics can improve inclusion
criteria and clinical trial performance. Tomography, 8(1), 341-355.](https://image.slidesharecdn.com/arn2024-250809150742-e0c03339/85/ARTIFICIAL-INTELLIGENCE-AN-EMERGING-TECHNOLOGY-IN-IMAGING-21-320.jpg)
ARTIFICIAL INTELLIGENCE: AN EMERGING TECHNOLOGY IN IMAGING
- 1. ARTIFICIAL INTELLIGENCE: AN EMERGING TECHNOLOGY IN IMAGING BELLO ISAH UMAR DEPARTMENT OF RADIOLOGY FEDERAL TEACHING HOSPITAL BIRNIN KEBBI, KEBBI STATE NOVEMBER, 2024.
- 2. TABLE OF CONTENTS • INTRODUCTION • OBJECTIVES • BENEFITS OF AI IN RADIOLOGY • AI IN RADIOLOGY MODALITIES • CONCLUSION • LIMITATIONS • RECOMMENDATIONS • REFERENCES
- 3. INTRODUCTION • Radiology, a central medical discipline that over the years witnessed rapid evolution from x-rays to magnetic fields, film to function and now into the new era of artificial intelligence, particularly with the advent of deep learning approaches especially the convolutional neural networks. • Radiology which serve as an eye of other departments is tasked with fast patient throughput, reliability and accuracy of results that will aid in patients diagnosis, decision making and treatment.
- 4. INTRODUCTION CONT’D • A review of literatures in AI and radiology was performed. 109 papers were searched for across PubMed, Google Scholar, Elsevier and ScienceDirect. Careful review was made of 83 papers, 37 were excluded due to limitations in accessing the Main PDF, 19 were also excluded due date of publication before 2019 and 7 due to duplications. A total of 20 papers were included in this study.
- 5. • Video presentations on Youtube were also utilized from renowned sites like RSNA, Mayo Clinic, ECR, AI NeuroCare, GE healthcare and Radiopedia
- 6. OBJECTIVES • Clarify the importance of AI in radiology as an aiding tool rather than a threat • Discuss the success of AI in detection, characterization and image interpretation in multidimensional imaging and mammography • Discuss the best way to utilize AI in multidimensional imaging and mammography screening • Discuss the future and challenges faced by AI especially in developing countries and the best ways to overcome those challenges
- 7. AI, MACHINE LEARNING AND DEEP LEARNING
- 8. BENEFITS OF AI IN RADIOLOGY • Early detection of lesions • Enhanced accuracy • Increased efficiency and patient throughput • Research and innovation • Reduced patient expenditure • Personalized medicine
- 9. AI IN RADIOLOGY MODALITIES • AI shows promising results in detection of additional lung nodules in patients with complex lung diseases [1] • It is also an important tool in detecting additional Intracranial lesions [2], increased sensitivity and accuracy in diagnosing cerebral aneurysm [3] and also reduces time of aneurysm detection [4] • Overall increase in sensitivity, accuracy and reduced workflow is recorded in AI-Based diagnosis in computed tomography [5,6,7,8]
- 10. AI IN RADIOLOGY MODALITIES • There are many AI applications in the interpretive domains of lesion detection and diagnosis, triage and density assessment and also noninterpretive domains of risk assessment, image quality control, image acquisition and dose reductions [9] • In Mammography, AI proved to be an aiding tool by serving as the first reader [10] and as such reduced the overall workload of breast radiologists with increased diagnostic accuracy [5,11]
- 11. AI IN RADIOLOGY MODALITIES • Researches on AI in MRI shows increased detection with Functional connectivity MRI (FCM) algorithm [12], AI is also useful by accurately locating and segmenting acute ischemic stroke lesions and the effects of early rehabilitation [13]. By apply Faster R-CNN in evaluating Circumferential Resection Margin status of rectal cancer patients in MRI, it shows increased accuracy and sensitivity [14,15].
- 12. AI IN RADIOLOGY MODALITIES • Deep transfer learning also shows accuracy of up to 98.01% in differentiation of breast lesions on dynamic contrast enhanced MRI (DCE- MRI) [16,17]
- 13. AI IN RADIOLOGY MODALITIES • Increasing throughput while maintain excellence is one of the main goals of every department and studies in nuclear medicine shows increased throughput, AI-driven theranostic drug discovery and labelling, dose precision and as such increased accuracy and dose reduction [18] • AI also helps in assessing image quality while using FDG-18 in whole body imaging [19] • AI-radiomics is also an important tool in research [20]
- 14. LIMITATIONS • AI works based on data it was trained with, so when integrated in a different population it might cause inequities [5] • Risk of leakage of patient data privacy which can be as a result of too much access of AI models to patient data [5,6,7] • Integration of AI to existing institution requires harmony between AI and medical experts to avoid undermining patient care [6] • The black box Enigma [18]
- 15. RECOMMENDATIONS • Continued investment in AI research development • Establishment of clear ethical guidelines and robust privacy frameworks • Adequate training of healthcare personnel • Collaborative approach including AI developers, healthcare personnel and regulatory bodies • Continues monitoring and evaluation of AI in clinical practice. • Addressing data privacy concerns
- 16. CONCLUSION • Artificial intelligence is a great tool for radiology departments especially when utilized alongside radiologist and radiographers as it aids in increased workflow, accuracy, sensitivity and decision making. • Despite the enormous successes of AI in radiology, there are many limitations related to integration into existing departments, upgrading AI models, ethical considerations and data privacy concerns • Increased investment in AI and training staff is of utmost importance for proper utilization
- 17. REFERENCES 1. Abadia, A. F., Yacoub, B., Stringer, N., Snoddy, M., Kocher, M., Schoepf, U. J., Aquino, G. J., Kabakus, I., Dargis, D., Hoelzer, P., Sperl, J. I., Sahbaee, P., Vingiani, V., Mercer, M., & Burt, J. R. (2022). Diagnostic accuracy and performance of artificial intelligence in detecting lung nodules in patients with complex lung disease: A noninferiority study. Journal of Thoracic Imaging, 37(3), 154-161. 2. Kundisch, A., Hönning, A., Mutze, S., Kreissl, L., Spohn, F., Lemcke, J., Sitz, M., Sparenberg, P., & Goelz, L. (2021). Deep learning algorithm in detecting intracranial hemorrhages on emergency computed tomographies. PLoS ONE, 16(11), e0260560. 3. Park, A., Chute, C., Rajpurkar, P., Lou, J., Ball, R. L., Shpanskaya, K., Jabarkheel, R., Kim, L. H., McKenna, E., Tseng, J., Ni, J., Wishah, F., Wittber, F., Hong, D. S., Wilson, T. J., Halabi, S., Basu, S., Patel, B. N., Lungren, M. P., Ng, A. Y., & Yeom, K. W. (2019). Deep learning-assisted diagnosis of cerebral aneurysms using the HeadXNet model. JAMA Network Open, 2(6), e195600. 4. Martinez-Gutierrez, J. C., Kim, Y., Salazar-Marioni, S., Tariq, M. B., Abdelkhaleq, R., Niktabe, A., Ballekere, A. N., Iyyangar, A. S., Le, M., Azeem, H., Miller, C. C., Tyson, J. E., Shaw, S., Smith, P., Cowan, M., Gonzales, I., McCullough, L. D., Barreto, A. D., Giancardo, L., & Sheth, S. A. (2023). Automated large vessel occlusion detection software and thrombectomy treatment times: A cluster randomized clinical trial. JAMA Neurology, 80(11), 1182-1190.
- 18. 5. Khalifa, M., & Albadawy, M. (2024). AI in diagnostic imaging: Revolutionising accuracy and efficiency. Computer Methods and Programs in Biomedicine Update, 5, 100146. 6. Najjar, R., Smith, J., Lee, K., & Patel, A. (2023). Redefining radiology: A review of artificial intelligence integration in medical imaging. Diagnostics, 13(17), 2760. 7. Fan, X., Zhang, H., Li, X., Liu, J., & Chen, Q. (2022). Artificial intelligence-based CT imaging on diagnosis of patients with lumbar disc herniation by scalpel treatment. Computational Intelligence and Neuroscience, 2022, Article 3688630. 8. Alowais, S. A., Smith, J., Lee, K., & Patel, A. (2023). Revolutionizing healthcare: The role of artificial intelligence in clinical practice. BMC Medical Education, 23 9. Brancho, P. E., Franco, A. H. S., Oliveira, A. P., Carneiro, I. M. C., Carvalho, L. M. C., Souza, J. I. N., Leandro, D. R., & Cândido, E. B. (2024). Artificial intelligence in mammography: A systematic review of the external validation. Revista Brasileira de Ginecologia e Obstetrícia, 4, e-rbgo71.
- 19. 10. Lamb, L. R. (2022). Artificial intelligence (AI) for screening mammography, from the AJR special series on AI applications. AJR American Journal of Roentgenology, 219(3), 369-381. 11. Lauritzen, A. D., Lillholm, M., Lynge, E., Nielsen, M., Karssemeijer, N., & Vejborg, I. (2024). Early indicators of the impact of using artificial intelligence in mammography screening for breast cancer. Radiology, 311(3), e232479. 12. Du, Q., Zhang, H., Li, Y., Wang, X., Liu, J., & Chen, L. (2022). Evaluation of functional magnetic resonance imaging under artificial intelligence algorithm on plan-do-check-action home nursing for patients with diabetic nephropathy. Contrast Media & Molecular Imaging, 2022, Article 9882532. 13. Sun, L., Zhang, Y., Wang, H., Li, X., Liu, J., & Chen, Q. (2023). Exploration of the influence of early rehabilitation training on circulating endothelial progenitor cell mobilization in patients with acute ischemic stroke and its related mechanism under a lightweight artificial intelligence algorithm. European Review for Medical and Pharmacological Sciences, 27(12), 5338-5355.
- 20. 14. Xu, J. H., Wang, Y., Li, X., Zhang, H., Liu, J., & Chen, Q. (2020). Application of convolutional neural network to risk evaluation of positive circumferential resection margin of rectal cancer by magnetic resonance imaging. Zhonghua Wei Chang Wai Ke Za Zhi, 23(6), 572-577. 15. Wang, D., Zhang, H., Li, X., Liu, J., & Chen, Q. (2020). Evaluation of rectal cancer circumferential resection margin using faster region-based convolutional neural network in high-resolution magnetic resonance images. Diseases of the Colon & Rectum, 63(2), 143-151. 16. Meng, M., Zhang, H., Li, X., Liu, J., & Chen, Q. (2022). Differentiation of breast lesions on dynamic contrast- enhanced magnetic resonance imaging (DCE-MRI) using deep transfer learning based on DenseNet201. Medicine, 101(45), e31214. 17. Yin, H. L., Jiang, Y., Xu, Z., Jia, H. H., & Lin, G. W. (2023). Combined diagnosis of multiparametric MRI-based deep learning models facilitates differentiating triple-negative breast cancer from fibroadenoma magnetic resonance BI-RADS 4 lesions. Journal of Cancer Research and Clinical Oncology, 149(6), 2575-2584.
- 21. 18. Saboury, B., Bradshaw, T., Boellaard, R., Buvat, I., Dutta, J., Hatt, M., Jha, A. K., Li, Q., Liu, C., McMeekin, H., Morris, M. A., Scott, P. J. H., Siegel, E., Sunderland, J. J., Pandit-Taskar, N., Wahl, R. L., Zuehlsdorff, S., & Rahmim, A. (2022). Artificial intelligence in nuclear medicine: Opportunities, challenges, and responsibilities toward a trustworthy ecosystem. Journal of Nuclear Medicine, 64(2), 188-196. 19. Qi, C., Wang, S., Yu, H., Zhang, Y., Hu, P., Tan, H., Shi, Y., & Shi, H. (2023). An artificial intelligence- driven image quality assessment system for whole-body [(18) F] FDG PET/CT. European Journal of Nuclear Medicine and Molecular Imaging, 50(5), 1318-1328. 20. Tomaszewski, M. R., Fan, S., Garcia, A., Qi, J., Kim, Y., Gatenby, R. A., Schabath, M. B., Tap, W. D., Reinke, D. K., Makanji, R. J., Reed, D. R., & Gillies, R. J. (2022). AI-radiomics can improve inclusion criteria and clinical trial performance. Tomography, 8(1), 341-355.