It has only been natural to see Artificial Intelligence (AI) and machine learning (ML) carving their way into healthcare industry like they are in all other industrial sectors (Bilyk 2020). The efficiency and precision that AI brings in extracting features from medical images, which significantly helps disease diagnosis and patient outcome predictions, have attracted many researchers in the medical community. One such example is detecting cancerous tumors in images generated by chest computed tomography (CT) scans through AI (Rosso 2022).
A specialized AI convolutional neural networks called the Lung Cancer Prediction Convolutional Neural Network (LCP-CNN) was developed using CT scans. This Computer-aided diagnosis (CAD) model was then used in Roger Y. Kim, et.al.’s paper on detecting tumors and assess malignancy risk in indeterminate pulmonary nodules (IPNs) (Kim, Oke et al. 2022). This has significantly improved clinician’s assessment of malignancy risk (Rosso 2022).
Computer Vision (CV) as a branch of AI, which enables computer to “see” images and identify objects like humans, quickly prevails as deep convolutional neural network (CNN) becoming the main workhorse. In 2012, Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton made AI history with their revolutionary study where they trained a deep convolutional neural network (CNN) to classify 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into a thousand classes. It won the ImageNet Large-Scale Visual Recognition Challenge and ignited a gold rush in the AI computer vision industry (Rosso 2022).
In Healthcare, CV helps human experts in numerous aspects of disease diagnosis:
Detection of Catheters On Radiographs
Significant delays occur between the time of X-raying and the moment when the image ends up in a radiologist’s hand. In this situation, computer vision can be a significant help in detecting misplacements and prioritizing patients with such problems (Altexsoft 2021).
Brain Tumor Segmentation on MRIs
Magnetic resonance imaging or MRI produces the most detailed pictures of soft tissues like the brain and is widely used for diagnosis of brain tumors. The main challenge here is that such tumors can be of different shapes and sizes. The earlier the abnormality is detected, the better chance is for a positive treatment outcome. Machine learning techniques hold the potential of speeding up tumor localization dramatically (Altexsoft 2021).
Skin Cancer Classification on Photographs And Dermoscopic Images
Skin cancer such as melanoma is often detected visually. Yet, owing to the variability of skin lesions, its symptoms can be confusing. That’s where computer vision outperforms even living experts. Recent research demonstrates that machine learning models show better results in skin cancer classification than an average dermatologist (Altexsoft 2021).
Looking forward, it is not just about AI finding things in images, it is more about the how AI improves the holistic care pathways and patient treatment. large number of emphysema case histories allows for supervised learning of a neural network to identify problem areas, then advanced statistical models, now part of ML, even if not AI, to assist in the identification of lung segments and to measure and quantify identified features in those segments (Teich 2021). There’s the high level statistical analysis of which types of treatment are best for different patient demographics (Teich 2021). Significant movements are in high level identification, organization and prescriptive action for helping disease diagnosis and patient outcomes.
References:
Altexsoft, s. r. d. e. (2021, May 21st ). “Computer Vision in Healthcare: Creating an AI Diagnostic Tool for Medical Image Analysis.” Retrieved June 5th, 2022, from https://www.altexsoft.com/blog/computer-vision-healthcare/.
Bilyk, V. (2020, May 6th). “Computer vision opportunities in Medical Imaging Explained.” Retrieved June 1st, 2022, from https://volodymyrbilyk.medium.com/computer-vision-opportunities-in-medical-imaging-explained-9e046f9e2d88.
Kim, R. Y., J. L. Oke, L. C. Pickup, R. F. Munden, T. L. Dotson, C. R. Bellinger, A. Cohen, M. J. Simoff, P. P. Massion, C. Filippini, F. V. Gleeson and A. Vachani (2022). “Artificial Intelligence Tool for Assessment of Indeterminate Pulmonary Nodules Detected with CT.” Radiology: 212182.
Rosso, C. (2022, May 25th ). “AI Deep Learning Predicts Cancer From Scans of Lung Nodules.” Retrieved June 5th, 2022, from https://www.psychologytoday.com/au/blog/the-future-brain/202205/ai-deep-learning-predicts-cancer-scans-lung-nodules.
Teich, D. A. (2021, Feb 2nd). “Radiological Analysis Leveraging Artificial Intelligence Is Moving Past Pure Identification Of Tumors.” Retrieved June 5th, 2022, from https://www.forbes.com/sites/davidteich/2021/02/02/radiological-analysis-leveraging-artificial-intelligence-is-moving-past-pure-identification-of-tumors/?sh=3e00a2ce327a.
