CXOToday has engaged in an exclusive interview with Dr. Ravi Gupta, Vice President – Bioinformatics
- Role of big data in diagnostics. How is it transforming the medical diagnostics sector in India?
By supplying healthcare providers with a wealth of patient data that can be evaluated and utilised to guide diagnostic and treatment decisions, big data plays a key role in diagnostics. By providing deeper insights and better patient management and treatment with lower costs and improved quality, this has the potential to increase the accuracy of diagnoses as well as the efficiency of the diagnostic process. The healthcare industry is changing because of the usage of big data in medical diagnostics since it allows for a more thorough comprehension of patient data. To tailor personalized treatment regimens, healthcare providers can, for instance, employ big data analytics to find patterns and trends in patient information.
Furthermore, the improvement of communication networks has allowed medical professionals to share and access information more effectively, potentially enhancing the speed and precision of diagnosis. The COVID-19 testing data management platform from the ICMR has been essential in helping the government to respond to the pandemic quickly and efficiently, and it emphasises the significance of big data in the field of medical diagnostics.
- What are the trends today that encourage the healthcare industry to embrace big data?
Healthcare practitioners are increasingly gathering enormous volumes of patient data thanks to the quick development of electronic health records (EHRs), wearables, and other digital health technology, which can be examined to guide treatment and diagnosis choices. Healthcare professionals are finding it simpler to build stronger relationships with their patients thanks to the development of advanced analytics and machine learning algorithms as well as the expanding availability of cloud computing resources. Many hospitals and healthcare organisations are currently looking at AI and big data solutions in fields including radiography, genomics, and telemedicine. These developments are pushing the healthcare sector to embrace big data and take use of its potential to raise the standard of treatment and enhance patient outcomes.
- Discuss the various applications and challenges.
The use of AI in radiology has the potential to significantly improve the accuracy of medical image interpretation, which can lead to earlier and more accurate diagnoses, better treatment plans, and improved patient outcomes. With the increasing availability of large datasets, AI techniques are becoming more sophisticated and effective in detecting diseases from medical images and are likely to play an increasingly important role in healthcare in the future. One of the most used medical scans for the identification of numerous thoracic disorders is the chest X-ray, which receives approximately 2 billion scans annually worldwide. A deep learning program called CheXNet, created by Rajpurkar et al. in 2017, uses a 121-layer Convolutional Neural Network (CNN) trained on a large publicly available dataset of chest X-ray records with over 100,000 records from 14 diseases, and is better than trained radiologists at detecting pneumonia from chest X-rays. According to tumor size and cohort, the clinical community’s manual sensitivity for identifying malignant pulmonary nodules ranges from 36% to 84%, which is not sufficient. Deep neural networks (DNNs) have recently been used to detect malignant lung nodules from chest X-rays (Nam et al., 2018). This technique performed better overall in 16 out of 18 doctors and produced significantly superior outcomes as compared to manual identification methods. The doctors that outperformed the AI technique had more than 13 years of expertise. Identifying bone fractures from photos as opposed to human interpretation is another intriguing application of AI. For example, utilising a DNN method on wrist fractures detection raised accuracy from 81% to 92% and lowered false alarm rates by 47% (Lindsey et al., 2018).
Big data analysis is also extensively used in genetics in addition to radiology. MedGenome applies data analysis approach to detect the causative variations quickly and accurately from patient genetic data. Identifying causal variants from a large genomics data is a needle in a haystack problem. Building a machine learning model that has been trained on a big dataset allows us to identify causal variants reliably and quickly. To handle the astronomical genomics data that conventional procedures are unable to handle, big data approaches are also applied. Deep learning algorithms are now also employed to detect early cancer symptoms in otherwise healthy persons.
Most of the models in use were developed using data from western populations. This can result in skewed findings. Population-specific computer models must be developed to provide more accurate and fair predictions.
- How can big data enable diagnostics players to expand their services?
The myths around big data are fundamental to the industry’s transformation in the field of medical diagnosis. The preventative healthcare sector may proactively identify and address potential health risks by utilising the power of predictive analytics, which lowers treatment costs and enhances patient outcomes. Mobile health applications and continuous wearable device data monitoring can assist diagnostics players in expanding their services and offering in-the-moment monitoring and analysis of patient health metrics. Early diagnosis of potential health problems and rapid intervention as a result can improve patient outcomes and lessen the strain on the healthcare system. Additionally, resource usage can be optimised with the aid of data-driven forecasting tools, resulting in more effective and efficient patient care.
- What are the latest technologies that are being used in big data for the diagnostics sector?
Big data technologies are classified into four categories: data storage, data mining, data analytics, and data visualisation. Hadoop is the most widely used framework in medical imaging. The Hadoop framework aids in the efficient processing and storage of large amounts of various types of data. The Hadoop framework is made up of three distinct parts: the Hadoop distributed file system (HDFS), MapReduce, and yet another resource negotiator (YARN). HDFS is used for efficient data storage, MapReduce for faster data processing, and YARN for efficient resource allocation required to process and manage data clusters. To address some of the shortcomings of the Hadoop framework, the Apache Spark framework was developed at the University of California, Berkeley’s AMP Lab.
- How can big data give more scope to precision medicine?
A complete picture of a patient’s medical history, including diagnoses, treatments, and outcomes, can be obtained from the study of their electronic health records (EHRs). Precision medicine can be made more effective by using this information to guide treatment choices. The EHR system will unavoidably be a part of the clinical decision support system. The usage of this method will increase clinician productivity and is also adopted in several Western countries for different therapeutic judgements. By supplying the information and insights required to create better-informed and more efficient treatment decisions, big data has the potential to revolutionise precision medicine. Big data is being used by precision medicine, particularly pharmacogenetics, to enhance patient outcomes and lower the possibility of negative side effects. Healthcare professionals can better meet the needs of each patient and increase the efficacy of treatment by understanding a patient’s genetic composition and how it affects how they react to various medications. Prior to prescribing a prescription or making a therapeutic judgement, patients are already being screened for specific gene mutations as part of precision medicine. For example, CYP2D6 polymorphism has been associated with Tamoxifen response, BRAF mutations (Y472C) have been associated with Dasatinib response in non-small cell lung cancer, and numerous other of these genes have recently been associated with the response of rectal cancer to chemotherapy and radiotherapy.
