Increasingly, artificial intelligence is being used to perform specialised human processes – only more efficiently and at a lower cost. Especially in the field of oncology and pathology, AI can be used for diagnosis, prognosis, and treatment.
Artificial intelligence (AI) and robotics will throughout the coming years become an integrated part of Denmark’s healthcare system. By assisting healthcare professionals in their daily routines and clinical decision making – such as scanning microscope slides with stained tissue to a digital format allowing for advanced image analysis and decision support – AI is providing faster time to diagnosis and improved certainty for patients waiting for answers in the most critical phase of their life.
For their part, Agilent Technologies are using algorithms to increase precision and diagnostic certainty. As the Associate Vice President, Morten Frost Norgreen explains:
“Our aim is to provide an end-to-end laboratory workflow incorporating AI and machine learning tools into clinical pathology and oncology. We focus on how to automate processes, pool data, and make more accurate and faster decisions using advanced algorithms as a complement to our existing foundational diagnostic solutions.”
Marking a major leap in the field of pathology, this process not only enhances patient outcomes. It also enables the ability to cross reference tissue samples, blood tests and genetic information with patient records and medical information to create consensus and validate the diagnostic input along the patient journey.
Using AI to prioritise urgent cases
Sarah Lidé, Senior Strategy and Project Manager at Medicon Valley Alliance, sees a rise in potential for assisted intelligence to aid with faster and better diagnoses. For example, DeepMind and Google Health recently developed a new AI system to help doctors detect breast cancer early, which performed better than human radiologists.
Another example is an algorithm developed by a UK tech start-up, which is capable of identifying lung cancer. Delivering an ‘instant triage’ of chest X-rays in a few seconds, it helps researchers and clinicians to determine which ones need a closer look to confirm or rule out the presence of cancer.
According to Lidé:
“Such algorithms can ease backlogs and enable medical staff to prioritise the most urgent cases first to ensure these patients are treated on a timely basis and have the best possible outcomes.”
Norgreen and Lidé agree that further down the road, we may see a rise in tele-robotics. Using remote-operated robotics, physicians will be able to treat patients in different locations, facilitated by ultra-fast, high-bandwidth network infrastructure, such as 5G. This would help to bridge the urban-rural divide and foster health equity for specialised care.
Remote diagnostics is opening up for full leverage of the workforce and addressing some of the challenges we see with the lack of trained healthcare professionals across critical specialties such as Pathology, where Pathologists are becoming a scarcity.
Identifying at-risk populations for early interventions
According to Lidé, AI can help healthcare administrators to address three priority areas: diagnosis, human error, and healthcare inequities:
“AI-enabled solutions can reduce cost pressures through optimising healthcare operations, such as patient flow and logistics, and on a larger scale, supporting better population management. This second point, which touches on predictive and preventive aspects of healthcare, is particularly important, where it is about identifying at-risk population groups for early interventions. Machine learning could thus enable us to identify at-risk patients earlier on, who could benefit from preventive treatment.”
More broadly, the increased focus on AI and machine learning can improve the patient experience generally. Paraphrasing Lidé, patients and stakeholders want to be engaged in healthcare decision-making. Facilitating this, AI applications – such as chatbots – can not only empower individuals to engage in their health but also provide them with the information they need in a personalised and conversational way.
Weathering the perfect storm
For over 200 years, pathologists were reliant on microscopes. Today, Agilent Technologies delivers laboratory solutions covering automated instrument platforms, antibody reagents and software which provides an essential backbone of all cancer diagnostics around the world. While the technologies are improving, Norgreen sees a significant challenge with the lack of trained healthcare professionals across critical specialties such as pathology, where pathologists are becoming a scarcity. This challenge is amplified by the increasing number of people being diagnosed with cancer every year.
While this has planted the seeds for a perfect storm in the healthcare system, artificial intelligence promises to be an effective tool to weather it.
“AI and digitalised processes allow full leverage of the workforce no matter where people are located, and it allows even graduate pathologists with limited experience to provide high quality and consistent diagnosis. Finally, the technologic advancements allow us to save time and resources to ensure cost optimal diagnosis and treatment”, Norgreen explains.
To meet the rise in AI, healthcare and life sciences personnel will require new competencies. Lidé refers to the book “The Most Human Human”, authored by Brian Christian as required reading in this area. Essentially, it describes the rise of AI as a type of maggot therapy – one that consumes only those portions of the physician’s work that are no longer human.
“This means that healthcare administrators need to think critically about where human touch and wisdom – seeing beyond the data – make the most sense and are irreplaceable, and where we can benefit from AI as a source of support.”
Thus, when it comes to technology, healthcare professionals require a nuanced understanding of what can and cannot be delivered. On a deeper level, they must reflect on how the role of healthcare professionals may be changing – from information providers to sense-makers, who have empathy that simply cannot be replicated by AI.