Is #AI a cancer game-changer?

brAIn gAIn: Episode 005

Your doctor says, “you have cancer.”?Three devastating words. Time stops.

Regardless of the location, language, or utmost concern and care with which this diagnosis is delivered … time stops.

Regardless of what your doctor or oncologist says after about the type of cancer, if it’s curable and can be put into remission, the type of surgery and/or therapy(ies) to be tried, or even your five-year survival odds (good or horrible) … time stops. And your life changes.

The Big Picture

According to the World Cancer Research Fund (WCRF), 18.1 million new cancer diagnoses were made in 2020. Meanwhile, the World Health Organization (WHO) estimates almost 10 million people died from cancer – the leading cause of death on earth – in the same year. And the Union for International Cancer Control (UICC) puts the challenge of cancer control into stark perspective given factors like:

• The ageing global population;
• Climate change;
• Sedentary lifestyles;
• Poor diet;
• Tobacco and alcohol use;
• Avoidable radiation exposure;
• Immunization avoidance in other diseases that can cause cancer, and
• Lack of access to proper healthcare – education, screening, diagnosis, and treatment – resources (especially in low- and middle-income countries).

Today, time will STOP for approximately 50,000 of our fellow travellers here on this third rock from the sun. Every minute in the next 24 hours, 35 human beings (one every two seconds) will receive a cancer diagnosis … and tomorrow? Well, a case of rinse, lather, repeat. And on and on it goes. One more sobering fact, cancer is not one disease, depending on how it is classified, it is some 200 separate diseases according to the Canadian Cancer Society.

Turning Hurt into Hope

Despite this global challenge, and not to discount* the impact of any individual cancer diagnosis, the 21st century, right now, is the most advantageous time in human history to be diagnosed with cancer. Today, advancements in prediction of predisposition to the disease, early detection, diagnosis down to the molecular level, precision tailoring of technology (drugs and other interventions) treatments, and our understanding of environmental and psychosocial factors that can cause or even help in healing cancer, is astounding.

* Author Note: I lost my father to a glioblastoma multiforme stage 4 brain tumour in 1996 (66 days from diagnosis to tumour resection to post-surgery radiation to death), I live with a pre-cancerous digestive tract condition, and have had several pre-cancerous skin lesions ‘burned’ off my skin by my dermatologist.

With medical knowledge increasing – by one measure, every 73 days (an assertion, not a fact) – there is justifiable hope that some of this knowledge will be usable in the battle against cancer and lead to more cures, safer treatments with fewer side effects, and turn life-threatening cancers into more manageable chronic disease.

Is #AI a cancer game-changer?

Last month, The Lancet-Oncology offered a nuanced answer around the theme of this question. But I will be more bullish, unequivocally, YES! In so many areas, the deployment of #artificialintelligence is making those time stopping three words a mere moment in time, not a countdown clock as they were often marked just two generations, or even one generation ago. Whether alone, or in concert with other treatment regimens and technologies, #AI is helping to make cancer a journey to be navigated toward a cure (full remission), more encouraging five-year survival rates, and sometimes, still, and sadly, better predicting the time a person living with a terminal cancer diagnosis has left before death.

The question of where to begin is not an easy one given the range of #AIinHealthcare interventions that have arisen in the treatment of cancer by oncology (with its fascinating history) researchers, healthcare professionals, allied health professionals, and non-traditional approaches to care. So here goes …

Medical Imaging

Perhaps the most well-known and older (like new teenager old ??) use of #AIinHealthcare is in the medical imaging space including X-rays, CT scans, MRIs, PET scans, and ultrasounds. Algorithms, trained on hundreds of thousands (if not millions) of images are now being used to detect and classify heart disease, stroke cancer, brain lesions that can indicate forms of dementia, tuberculosis, and retinal scans in vision care and other conditions.

The degree of sophistication and granularity of these approaches can detect abnormalities that may be missed by even the most trained and experienced human eyes. When combined with other information such as patient history, genomic profiles, recent symptoms and the latest lab results, this richness of data points yields patterns of information that help clinicians plan further steps in care for their patients.

A 2022 post from the US National Cancer Institute provides more detail on how AI-assisted imaging technology can help identify cancers faster, determine courses of treatment (chemotherapy, radiation, resection, pharmacotherapy) and be used to train younger doctors and technicians. And the range of cancers where AI-enabled computer vision, algorithms, and combinations with other diagnostic tools and therapy regimens is about as broad as the subtypes of cancer itself. This includes:

• A blood test that can detect 50 different types of cancer;
• Predicting the possible tissue of origin (breast, colorectum, liver, lung, lymphoma, esophagus, ovary, pancreas, and stomach) in individuals where a blood test has indicated the presence of cancer;
• Using nanotechnology (quantum sensors) to bounce light off of a blood sample to see cancer biomarkers and use #MachineLearning to determine the type of cancer;
• The range of imaging tools and algorithms to assess various forms of skin cancer;
• #MachineLearning combined with high-resolution ultrasound images to better detect if thyroid growths are cancerous; and
• A sketch-based tool that AI can then use to search imaging files for aggressive tumours.

This list is not exhaustive and if an imaging and computer vision “rabbit hole” is your thing, I recommend a dive into this Future Science article and/or learning more about CVPR 2023 (the annual IEEE Computer Vision and Pattern Recognition Conference) which was held this past June in Vancouver, Canada.

Cancer Treatment

Over the past 25 years, significant strides have been made in cancer care. This includes wider adoption of targeted therapies (drugs that target specific proteins that drive how cancer grows, divides, and spreads), increased use of immunotherapy, and the development of personalized medicine (using your genetic information and your specific cancer type to combat the disease). These treatments have led to better survival rates and quality of life for people living with cancer.

Moreover, broader and more targeted deployment of radiation therapy, the advent of radiopharmaceutical therapy, improved chemotherapy regimens, and innovations in tumour resection (aka: surgery) have further transformed cancer care.

Other AI-enabled (or embedded) treatment innovations include:

• Determining the molecular profile of a patient’s brain tumor during surgery which can help surgeons know how much tissue to remove or whether to place anti-cancer drugs directly in the brain with the patient still ‘on the table’;
• Using #DeepLearning to target drug resistance of adaptive cancer cells (acquired through genetic mutations, activation of alternative signaling pathways, or the presence of cancer stem cells;
• Predicting the side-effects of immunotherapy;
• Identifying predictive biomarkers to assess if a patient’s prostate cancer could spread and determining if a patient could benefit from hormone therapy and radiotherapy; and
• Better segmenting healthy and disease groups in clinical trials of new medications.

Drug Discovery and Development

One of the most promising areas of #AIinHealthcare is the potential to accelerate drug development through the discovery of new molecules, repurposing older approved treatments for other diseases into cancer medicines, or shortening the lengthy (and costly) timeline to research, develop, run clinical trials, and bring drugs forward for regulatory approval. In this realm, the promise for new cancer therapies is mind-blowing. Consider the following examples:

• The first AI-generated drug, from Insilico Medicine, advancing through clinical trials offering hope for idiopathic-pulmonary fibrosis (IPF) patients who have five times the risk of developing lung cancer compared to the general population;
• Developing a potential liver cancer treatment in just 30 days;
• Collaborations around the world (Europe being just one example) to develop new precision medicines; and
• mRNA vaccine pioneers Moderna (second story bonus) and BioNTech – alongside other biopharmaceutical companies – are leading AI-assisted efforts to develop a cancer vaccine(s).

If you fancy another “rabbit hole”, of the AI and drug discovery variety, Nature and Techopedia have you covered. ??

?

Infinite Promise

The examples provided so far barely scratch the surface of the scratching the surface of the surface of the potential of #ArtificialIntelligence to help eradicate cancer from our planet. This potential includes:

• Conversational chatbots for patients (and caregivers) during their cancer journey;
• Navigation options and plans for care teams; and
• Use of #GenerativeAI across a range of cancer research and treatment areas.

?

?#Hope, #HealthEquity, and #HealthForAll

To recap, deploying #AI in cancer research, prediction, drug discovery, treatment, and patient care provides immense hope for current and future patients. However, we must:

• Protect these innovations and provide further incentives for public- and private-funded efforts;
• Integrate validated #AIinCancer applications/algorithms/medicines into clinical practice with appropriate caution(s) respecting transparency, bias, legal, and ethical considerations; and
• Address health equity issues including the social determinants of health that can result in a lack of access to education, screening, treatment options, and important non-healthcare supports (both in developed nations and across lower- and middle-income countries).

To borrow the tagline line of biopharma company BeiGene: Cancer Has No Borders.

And neither does #AI: It is a game-changer in cancer!

= = = = =

brAIn gAIn 005 -- Resources

• @erictopol on Twitter … Author of Deep Medicine (see brAIn gAIn 001 – Reading recommendation) and a consistent reader of JAMA, Nature, and the Lancet, he often tweets about articles on #AIinHealthcare
• https://www.uicc.org/ … The Union for International Cancer Control – multilateral and mission-focused
• https://www.newsweek.com/worlds-best-specialized-hospitals-2022/oncology ... 2022 ranking of the top 250 cancer research and treatment facilities across the planet!

brAIn gAIn 005 -- Reading

• Artificial Intelligence and Precision Oncology: Bridging Cancer Research and Clinical Decision Support edited by Zodwa Dlamini: https://www.amazon.com/Artificial-Intelligence-Precision-Oncology-Bridging/dp/3031215052/ … be warned, this is more of a textbook which is written for a knowledgeable audience (not your Barnes & Noble / Chapters-Indigo / WH Smith general purpose read)

?

brAIn gAIn 006 -- Reveal

• Imagining Your #AIinHealthcare Navigator

?

brAIn gAIn -- Request

Future newsletters will cover topics that include AI and future pandemics, AI regulation and legal issues in medicine, AI to combat the overdose crisis, and AI to transform clinical workflow. But this is what I want to write about.

However, you invest your time to read this newsletter, for which I am grateful and humbled, what #AIinHealthcare topics interest you?

Please let me know with a comment post or LinkedIn message …

your feedback is always a cherished gift.

Looking for other #AIinHealthcare content?

Please follow @AI_4_Healthcare on Twitter.

Be well!