So-called “labs-on-a-chip” are leading a testing revolution.
It’s well known that the world’s healthcare needs are increasing as the population is aging. The proportion of the world’s population over 60 years old is projected to rise from around 10% today to 16% by 2040. With this aging, the types of required treatments are also expected to change. Instances of cancer, for example, are expected to increase 47% by 2040.
We would all like to see treatment for serious illnesses like cancer become less invasive and deliver better patient outcomes. In a previous column, we discussed the opportunity for machines like ORNL’s Frontier to perform supercomputing to help get ahead of diseases and delay their onset. We are also seeing great advancements in diagnostic research using lab-on-a-chip biosensors by the University of Bath’s Centre for Biosensors, Bioelectronics and Biodevices in the UK.
These sensors shrink all the processes needed to analyze a disease onto a single chip, so instead of sending samples away to the lab, diagnoses can happen in the presence of the patient. It’s not only faster, but also more efficient, more affordable and easier to deliver in developed and developing countries. There is also no chance for samples to become mixed up or lost in transit.
In a conventional lab, where analysts must process slide after slide, the repetitive and laborious nature of the work increases risks of misdiagnosis. In contrast, the new biosensors eliminate errors by digitalizing and processing the information directly on the chip. The technology is already enabling the world’s first personal saliva diabetes tests, which make routine self-testing easier, as well as less invasive and more comfortable than at any time before.
Whereas traditional medicine has been performed “in vivo” on patients or “in vitro” on a glass slide or culture dish, we can now say that advanced healthcare is going “in silico” with these lab-on-a-chip sensors.
Among the projects ongoing at Bath, researchers are working with the University of São Paulo to develop biosensors suited to point-of-care testing for diseases such as prostate and breast cancer. As lab-on-chip sensors continue demonstrating great results, the research is beginning to focus on ways to make the sensors manufacturable in high volume, at low cost. Printed circuit board technologies have a critical role in this and, indeed, Dr. Despina Moschou – a microelectronics Ph.D. leading the lab-on-chip research at Bath – is closely connected with the EIPC and has presented her department’s work at several events.
It's great that our industry can contribute to this work, leveraging, for example, the proven processes that companies already have in place for scaling up production and minimizing costs. There will also be issues to tackle such as biocompatibility of the chosen materials, and further development of materials and fabrication techniques cannot fail to deliver improvements such as increased performance, greater repeatability, and yield growth leading to lower costs.
And then, of course, there are the opportunities to leverage AI to assist with analysis and diagnostics. After digitizing the sensed data – be they blood or saliva analysis from a lab-on-chip sensor, cell biopsies, or radiographic image data – AI can help in several ways, including screening the data to filter items of concern and highlighting them for the attention of specialists. This can greatly accelerate analysis and drive out human error, as well as reducing the number of specialists needed to serve the aging population.
Reducing our reliance on human analytical skills is important. Although there is an increasing demand for pathologists to perform these analyses, the number of trained practitioners is an acknowledged shortfall. AIs can be trained relatively quickly, which can help to meet demands for healthcare services in the future. Trials of Deep Learning-based Automatic Detection (DLAD) algorithms are already showing that these can help reduce instances of overlooked tumors when assisting specialists to analyze patient radiographs. By integrating the use of these algorithms within standard practice, healthcare providers can enable more patients to begin treatment at an early stage, leading to better outcomes.
Moreover, AI’s ability to analyze data collected from large numbers of patients can improve the study of disease progression throughout a specific population or throughout the world. With these machines, we can identify trends hidden in vast quantities of data a human observer could never detect, and thus gain insights that help to plan future services and drug development. It’s also worth mentioning they can be programmed to operate without bias and can therefore contribute toward democratizing access to high-quality healthcare.
One major issue to overcome is public trust in such powerful technology. When clinical errors are caused by humans, the reasons are examined, any negligence discovered is punishable, and practices are modified if necessary to prevent recurrence, but people continue visiting their doctors. On the other hand, misdiagnoses due to machine error – though far less common – will not be tolerated so easily.
Trust is usually aided and enhanced by transparency, which could be achieved if software providers were open about their algorithms. Naturally, they are likely to be guarded about their intellectual property, but we have already seen how the open-source software model has become commercially successful, benefiting both from community engagement in reviewing and refining published code while also effectively protecting creators’ rights and revenues.
It will probably just take time for us to adjust, and for machine-based healthcare to establish its own track record. In a couple of generations’ time, many people will have no experience of receiving healthcare any other way. Being diagnosed by machines, and even describing our own symptoms to machines, will become the norm quite quickly. We know human beings are highly adaptable creatures, and many of us will soon appreciate that machines can take better care of us than we could ourselves. •
Alun Morgan is technology ambassador at Ventec International Group (ventec-group.com); This email address is being protected from spambots. You need JavaScript enabled to view it..
