Source: Artificial Intelligence in Biomedical Science · Zenodo Authors: Mitra, Manu Licence: CC-BY-4.0 — https://creativecommons.org/licenses/by/4.0/
ISSN: 2640-4133
Short CommunicationsAdvances in Bioengineering & Biomedical Science Research
Artificial Intelligence in Biomedical Science
Manu Mitra
*
Department of Alumnus with Electrical Engineering, University of Bridgeport, Bridgeport, United States
Introduction
Biomedical Sciences has very broad range and deals with various disciplines of medical research such as genetics epidemiology, clinical epidemiology, clinical virology, medical microbiology. It also includes science disciplines whose fundamental aspect is biology of human health and diseases. It is also aims on relevant sciences that includes but not limited to anatomy, cell biology, biochemistry, microbiology, genetics, molecular biology, immunology, mathematics, statistics and bioinformatics. Biomedical sciences have wider range of research, academic and economic significance than that defined by hospital laboratory sciences [1].
Artificial Intelligence (AI) in biomedical is usage of software and complex structure of algorithms to mirror human intelligence in the analysis of composite medical data. Specifically, Artificial Intelligence is the capability for computer algorithms to estimate results without direct human interaction [2]. Some key features interest include but not limited to clinical text mining, patient centric information retrieval, biomedical text evaluation, diagnostic assistance, clinical event forecasting, data-driven prognostics, precision medicine, human computation [3].
Corresponding author
Manu Mitra, Department of Alumnus with Electrical Engineering, University of Bridgeport, Bridgeport, United States, E-mail: mmitra@my.bridgeport.edu
Submitted: 01 Nov 2019; Accepted: 12 Nov 2019; Published: 16 Nov 2019
Biomedical Imaging through Artificial Intelligence
Academics at University of Zurich used various methods in machine learning to improve optoacoustic imaging. This medical technique can be used for studying brain activity, visualizing blood vessels, characterizing skin lesions and diagnosing cancer. Nevertheless, quality of rendered images are dependent on number of sensors distribution used by the apparatus. This new technique developed by the scientists allows for significant reduction of number of sensors without reducing quality of image. This makes it possible to reduce the cost of apparatus and therefore increasing imaging speed and improve diagnosis.
To accomplish this task researchers used self-developed high-end optoacoustic scanner which has around 512 sensors which conveyed superior quality images. Next scientists discarded majority of the sensors and around 128 to 32 sensors remained; with a detrimental effect on the quality of image. Due to insufficient data, various distortions appears in the images. However, previously trained neural network was able to correct for these distortions and conveying the quality of image closer to the measurements obtained with 512 sensors [5, 6].
Figure 1: Illustrates a new MIT-developed model automates a
critical step in using AI for medical decision making, where experts usually identify important features in massive patient datasets by hand.
The model was able to automatically identify voicing patterns of people with vocal cord nodules (shown here) and, in turn, use those features to predict which people do and don’t have the disorder [4].
Figure 2: Illustrates Scientists use optoacoustic tomography to
create cross-sectional images of a mouse. Using machine learning, they were able to largely restore quality of images recorded with fewer sensors. Image Credit: Davoudi N et al. Nature Machine Intelligence 2019 [7].
Artificial Intelligence to Detect Cancer Tumors
Scientists at University of Central Florida Computer Vision Center developed and taught a computer on how to detect tiny particles of lung cancer in CT scans, where radiologists cannot identify it
Adv Bioeng Biomed Sci Res, 2019 www.opastonline.com Volume 2 | Issue 4 | 1 of 2
accurately. The accuracy of artificial system is around 95 percent compared to 65 percent when done by human eyes.
This approach is related to the algorithms that facial recognition software uses. It scans around thousands of faces looking for a particular pattern to match. Group delivered more than 1000 CT scans which were supported by the National Institutes of Health through collaboration with Mayo Clinic. They were able to develop software to identify cancer tumors. They used Machine Learning to ignore other tissue, nerves and other masses it encountered in the CT scans and analyze lung tissues.
Figure 3: Illustrates Fast Fiducial Tracking Algorithm [10].
New Advances in Plastic Surgery with Machine Learning
With ever increasing of electronic data collected in the health care, scientists are considering the use of a sub field of artificial intelligence
- Machine learning to improve medical care and patient outcomes. An analysis of machine learning can contribute to advancements in plastic surgery. Machine learning analyzes historical data to evolve algorithms capable of knowledge acquisition. Projects with healthcare applications that includes IBM Watson Health cognitive computing system. Authors have five areas where machine learning can improve efficiency and clinical outcomes – Burn surgery, Micro surgery, craniofacial surgery, hand and peripheral nerve surgery, aesthetic surgery. Authors also expect useful applications of machine learning to improve plastic surgery training. Nevertheless, they concentrate the need for measures to make sure the safety and clinical relevance of the results collected by machine learning and also remember at the same time that computer generated algorithms cannot replace the trained human eye yet [11, 12].
Artificial Intelligence Improves Dementia Diagnosis
Machine Learning has identified one of the common causes of dementia and stroke in most widely used form of brain scan more accurately than current methods. Advanced software developed by experts at Imperial College London and the University of Edinburgh were able to detect and measure the sternness of small vessel disease one of the common cause of stroke and dementia.
Scientists asserts that this technology can help physicians to carry out the best treatment to patients more swiftly in emergency settings and can predict person’s likelihood of developing dementia. This development also makes way for personalized medicine. [13, 14].
Artificial Intelligence Predicts Alzheimer’s Before Diagnosis
Timely diagnosis Alzheimer’s disease is very critical because treatments and interventions are more effective at early in the course of the disease. Yet, early diagnosis has proven to be very challenging.
A research group, multidisciplinary team of physicians and clinicians focusing on radiological data science. Dr. Franc was involved in applying deep learning, a type of AI in which machines learn like humans do, to find changes in brain metabolism of Alzheimer’s disease. They trained the algorithm on special imaging technology known as 18-F-fluorodeoxyglucose positron emission tomography (FDG-PET). In this, radioactive glucose compound is injected into the blood, then PET scans and takes measure of metabolic activity [15, 16].
References
- Biomedical sciences (2019) https://en.wikipedia.org/wiki/ Biomedical_sciences.
- Artificial intelligence in healthcare (2016) https://en.wikipedia. org/wiki/Artificial_intelligence_in_healthcare.
- Brown University (2019) http://brown.edu/Research/AI/index. html.
- Matheson, R (2019) Automating artificial intelligence for medical decision-making http://news.mit.edu/2019/automating- ai-medical-decisions-0806.
- Neda Davoudi, Xosé Luís Deán-Ben, Daniel Razansky (2019) Deep learning optoacoustic tomography with sparse data. Nature Machine Intelligence.
- ETH Zurich (2019) Artificial intelligence improves biomedical imaging. ScienceDaily. www.sciencedaily.com/ releases/2019/09/190930101259.
- Bergamin, F, ETH Zurich (2019) Artificial intelligence improves biomedical imaging.
- Naji Khosravan, Ulas Bagci (2018) S4ND: Single-Shot Single-Scale Lung Nodule Detection.
- University of Central Florida (2018) Engineers develop artificial intelligence system to detect often-missed cancer tumors.
- Stanford Medicine (2019) Image-Guided Radiation Therapy.
- Jonathan Kanevsky, Jason Corban, Richard Gaster, Ari Kanevsky, Samuel Lin, et al. (2016) Big Data and Machine Learning in Plastic Surgery. Plastic and Reconstructive Surgery137: 890-897.
- Wolters Kluwer Health (2016) ‘Machine learning’ may contribute to new advances in plastic surgery.
- Liang Chen, Anoma Lalani Carlton Jones, Grant Mair, Rajiv Patel, Anastasia Gontsarova, et al. (2018) Rapid Automated Quantification of Cerebral Leukoaraiosis on CT Images: A Multicenter Validation Study 288: 573-581.
- Imperial College London (2018) Artificial Intelligence improves stroke and dementia diagnosis in most common brain scan.
- Yiming Ding, Jae Ho Sohn, Michael G Kawczynski, Hari Trivedi, Roy Harnish, et al. (2018) A Deep Learning Model to Predict a Diagnosis of Alzheimer Disease by Using 18F-FDG PET of the Brain 290: 456-464.
- Radiological Society of North America (2018) Artificial intelligence predicts Alzheimer’s years before diagnosis. Copyright: ©2019 Manu Mitra. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Adv Bioeng Biomed Sci Res, 2019 www.opastonline.com Volume 2 | Issue 4 | 2 of 2