21st century started with a bang for mankind with the Human Genome Project and the success story continues ever since. Integrating fundamental knowledge to devising feasible simplistic technology-based solutions to real-life problems, we have come a long way. As Isaac Newton famously quotes ‘If I have seen further than others, it is by standing upon the shoulders of the giants’, we are witnessing this transition from applying the knowledge gained earlier to developing useful techniques. This age aspires to plunge deeper into the fundamental understanding of nature and inventing translational technologies for the betterment of all, living and nonliving.
From reaching astronomical heights by exploring human habitable areas outside the earth to diving cellular depths investigating the mechanisms of our creation and successful existence, we continue to explore every possible way of increasing our presence. Central to this is the concept of surviving in a world brimming with potential hazards. These could be drastic climate changes to advanced weapons of mass destruction. But a silent assassin through all these, are the diseases that cut short billions of lives worldwide every day, COVID-19 being the latest. Bacterial, viral, developmental or regulatory: all pose a stern challenge for us. Diagnosis, prognosis and successful treatment are the critical steps. Due to a plethora of expensive diagnostic tests for different diseases, it is a challenge left to the human expertise of doctors to accurately pinpoint the disease, its stage and the course of treatment. The sheer ratio of patients to doctors makes it a challenge. The unavailability of these diagnostic facilities to people from varied economic backgrounds and geographical locations makes it an arduous task to provide quality healthcare services for one and all. A one-stop accurate solution to the increasing number of diagnostic tests is the need of the age.
The development of a Unified Diagnostic and Prognostic Assistant (UDPA) could serve the purpose. A machine that combines the powers of simple tests to determine the vital statistics of the patient as well as perform CT (Computed Tomography) or PET (Positron Emission Tomography) scan safely using image reconstruction techniques . The only problem is the occasional loss of important information. With the emergence of Artificial Intelligence, Machine Learning and increasingly human-friendly virtual assistants in the form of Alexa and Siri, we are not long from taking help in aspects of healthcare from them. Collected data obtained in the form of numbers or images from the UDPA should be categorized. The UDP device could be something similar in appearance to a CT scanner where the patient is completely enclosed in a tube. Vitals like temperature, blood pressure and pulse rate could be tested non-invasively while a blood sampling to determine different cell count and presence of any other abnormality could be done through a robotic arm piercing as tested by the Venipuncture robot at Rutgers University. A test for audible range and visual capacity could be included based on inbuilt facility of a speaker and a laser. High-quality image processing techniques would be required to obtain quick data from the blood sample . The blood sample should be undergoing segregation for use in different tests for the application of different chemicals like antibodies and fluorophores which could be added manually (costly!). Mechano-sensitive wearable sensors at joints of the patient could collect physical skeletal and muscular data at the joints . All the patient data should be stored virtually. On completion of all preliminary tests, should the assistant offer help. The assistant should be trained with a large number of artificial data sets and images to diagnose diseases using the data. Dengue, for example, could be diagnosed based on IgM and IgG (Simply proteins) values in the blood sample. The diagnosis for Malaria based on the 2-line test could be performed internally in the UDP.
The assistant becomes a powerful tool once it has attained the ability of ‘thinking’ that would enable further specific tests like ELISA to be done. Antibody check could be done by fluorescence sensors built within the UDP. Determining the presence of metastatic cancers, based on isolation of Circulating Tumor Cells (CTCs) and imaging could be possible . Genomic data could be compared with the isolated and sequenced DNA internally across online databases to suggest possible diseases. These tests might require more time based on the complexity, but the assistant should be trained to quantify and accurately determine the stage of the disease. For inadequate data, human expertise could be sought. Upon characterizing the disease, the assistant should be able to suggest a course of medication or treatment based on its previous experience on the artificial data sets and earlier patient data. It should be able to analyze thermal images, real biopsy data and quantitative data from the tests. Advanced machine learning algorithms would be needed for the UDPA to develop the experience to create a treatment regime. Human technical expertise in combination with the UDPA could do wonders. An increasing number of robot-assisted surgeries are a testament to this Machine-Man partnership. The Da Vinci Surgical robot, Robin, Suki medical record transcriptors, radiology assistant PowerScribe One and the virtual nurse Elizabeth are some of the technologies developed till now , , , , .
Although this looks farfetched, a closer look at the publication years of the references below give an idea of how close yet how far we are! Google’s acquisition of the healthcare arm of AI firm DeepMind is a step towards it. A machine such as a UDPA amalgamates distinct fields like biology, medicine, artificial intelligence, image analysis and big data analysis. Providing a solution under one roof, at least for largely investigated diseases reduces the strain on doctors and increases the access of health seekers. Ultimately, the science of today is the technology of tomorrow.
- H Shan, A Padole, F Homayounieh, U Kruger, RD Khera, C Nitiwarangkul, MK Kalra and G Wang. Competitive performance of a modularized deep neural network compared to commercial algorithms for low-dose CT image reconstruction.Nature Machine Intelligence. June 10, 2018.
- Balter, M. L. et al. Automated end-to-end blood testing at the point-of-care: Integration of robotic phlebotomy with downstream sample processing. TECHNOLOGY 06, 59–66 (2018).
- Barpanda, S.S. (2013). Use of Image Processing Techniques to Automatically Diagnose Sickle-Cell Anemia Present in Red Blood Cells Smear.
- Gurchiek RD, Cheney N, McGinnis RS. Estimating Biomechanical Time-Series with Wearable Sensors: A Systematic Review of Machine Learning Techniques. Sensors (Basel). 2019;19(23):5227. Published 2019 Nov 28. doi:10.3390/s19235227
- Shen, Z., Wu, A. & Chen, X. Current detection technologies for circulating tumor cells. Chemical Society Reviews 46 2038–2056 (2017).
By: Aditya Dixit, Department of Biological Sciences, IISER Bhopal
Other articles by the author: LIGER? Really!?
About the Author: Aditya is a 3rd year Integrated MS student pursuing a Biology major at IISER Bhopal. An ardent Indian Classical music fan and a trained Harmonium performer, he enjoys creating music, singing, learning instruments and trying anything and everything in it! A FIDE rated chess player, playing-watching-talking about any sport, penning down some original pieces in any language and w(a/o)ndering are his go-to activities.