Introduction
On November 17th 2019, the first COVID-19 case was detected.1 Then, on the 16th March 2020, the UK entered its first lockdown.2 In 2021, the Lancet published a paper estimating that within the 4 most prevalent cancer types (breast, colorectal, lung and oesophageal) there were 3291 to 3621 avoidable deaths and an additional 59,204 to 63,229 years of life lost attributed to delays in cancer diagnosis as a result of COVID-19 lockdown in the UK alone.3 Countless lockdowns have resulted in pushbacks on many diagnoses and routine treatments, and this is no exception in oncology. Cancer is a leading cause of death globally,4 with the stage of diagnosis being the most important factor affecting survival.5 If diagnosed early, over 80% of patients survive 10 years or more.5 Therefore, it is no surprise that cancer screening and early detection has been one of the largest difficulties for oncologists, with more patients battling advanced stage cancer than in previous years.6
Artificial intelligence (AI) if applied well could help transform cancer detection and screening, making it more accessible, economically efficient and potentially even more accurate. These ideas are not just the musings of technology stars wanting to digitise everything, they have real scientific ground endorsed by the National Cancer Institute.7
COVID Inspired Evolution
The COVID-19 pandemic brought in a mind-set shift. It allowed us to challenge the status quo, to revaluate our priorities and how we live in a new world. COVID-19 accelerated the digital dawn, and allowed us to reimagine how we utilise technology in our daily lives. During the pandemic it allowed us to connect with loved ones, buy toilet roll at the click of a button, and work from the comfort of our own homes. Digitalisation has strengthened our global and socio-cultural structure, and broken down many traditional thought processes and hierarchies. However, what technology has really revolutionised is medicine. The past two years have seen humankind evolve drastically in medicine, from holding virtual clinics that utilise virtual reality to assess patients,8 to incorporating new vaccine technology to create a COVID-19 vaccination in record time.9
But the question remains, with all this incredible progress, can we harness the power of technology into something more meaningful than we already have?
Harnessing The Power of AI
Complexity shrouds cancer due to its thousands of genetic and molecular variations, and the need for early detection only escalates the need for a quick and effective detection mechanism. AI provides an innovative solution to this. "AI or artificial intelligence is defined as the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings"10 - with machine learning described as an application of AI, which provides systems with the ability to learn and improve from experience, without needing to be explicitly programmed.11 It uses computational methods based on mathematical algorithms and theories to detect images and patterns.11 The use of AI and more specifically machine learning algorithms in cancer detection not only could be highly economical, but the nature of AI would ensure personalised care, that under the right interface may one day allow detection from the comfort of our homes. AI excels at recognising patterns in large volumes of data, and has the ability to extract relationships between complex features in the data.7 It can identify characteristics in the data, such as images, that humans cannot perceive.7 AI has already been utilised in radiology to process images rapidly, enabling radiologists to use their time to focus on complex issues where their technical judgement is required.3 AI-assisted endoscopy has also shown huge potential for more accurate and sensitive early detection of gastric lesions, with a system demonstrating an AUC (an aggregate measure of performance, with the higher the AUC, the better the model)12 of 0.8684, and accuracy of 85.1%, equivalent or better than experts.13
However, the term AI is often met with hesitancy. We think of terminator style computers that plan to take over humanity. Sure, AI can be used to provide us tailored advertisements or to personalise your Instagram feed, but can we really trust AI with something as important as our own health? Multiple papers have discussed this at depth,14,15 with our lack of trust in AI systems being influenced by human factors such as user education, past experiences, user biases, along with the properties of the AI system itself, including controllability, transparency and complexity of the model to name a few.14 To add to the complexity surrounding this, anyone who has studied computational science will understand that AI is built on big data, and the robustness and granularity of each algorithm is the only way to ensure that AI is usable and functional. As someone who has dabbled in this during my time at university, having created my own machine learning algorithm to predict whether patients would develop coronary heart disease in 10 years, I can attest it is not easy, especially when applied to the healthcare industry, and will therefore require advocacy and policy change. However, the results could be transformative. The logistic regression algorithm (a specific form of machine learning algorithm) I created proved to have an AUC score of 0.78 when predicting whether a patient will develop coronary heart disease, in comparison to the AUC score of 0.68 for the Framingham method (clinicians’ current predictive method).16 Whilst I only applied my algorithm to a small dataset, this just shows the possibilities of AI when applied to a medical space.
Conclusion
If we are to beat cancer, early detection is arguably one of the most important factors. AI and machine learning tools could hold the key to this, but learning to trust AI and ensuring that the systems themselves are effective enough are clearly issues we must first address. However, if we have learnt anything from the past two years, anything could be possible with a little innovation. As AI inches closer to becoming a major tool in cancer treatment, robust medical communication campaigns will play a pivotal role in the success of this emerging landscape for oncology. As an award-winning medical communications agency, TVF specialises in educating and raising awareness of pharmaceutical advances, combining science, creativity, and strategy to deliver high quality communications to a range of audiences. To find out more about how our services could contribute to this exciting breakthrough within medicine, get in touch.
By Eleanor Barr
