Mind Over Chatter
Cancer and artificial intelligence
Episode guests
Podcast summary created with Snipd AI
Quick takeaways
- Artificial intelligence enhances cancer detection by analyzing images to identify smaller tumors that traditional methods often miss.
- The WIRE clinical trial employs innovative drug treatment and imaging techniques during the critical period between cancer diagnosis and surgery.
- Ethical considerations surrounding data privacy and consent are crucial as AI becomes integrated into healthcare for improved patient care.
Deep dives
Understanding Cancer
Cancer is characterized as an uncontrolled overproduction of cells that can arise from any type of organ or tissue. It stems from genetic abnormalities that can be inherited or induced by environmental factors, with smoking cited as a major contributor, accounting for nearly a third of all cancer cases. There are over 200 distinct types of cancer correlated to the more than 200 types of cells in the human body, with the most common types being lung, breast, bowel, and prostate cancer. As cancer progresses, it can destroy the local organ and spread to other parts of the body, making early detection crucial for successful treatment.
Challenges in Cancer Detection
Current imaging technology struggles to detect cancers smaller than half a centimeter in size, leading to late diagnoses in many cases. Techniques such as mammography and CT scans can identify tumors once they reach a certain size, but they are not sensitive enough to catch smaller, potentially curable cancers at their early stages. Efforts are underway to improve detection methods, including the development of AI-based virtual biopsy processes, which aim to analyze imaging data thoroughly enough to eliminate the need for physical biopsies in the future. This would allow for quicker assessments without invasive procedures, enhancing patient care.
The Role of Artificial Intelligence in Healthcare
Artificial intelligence (AI) refers to systems capable of learning and reasoning to perform tasks that mimic human intelligence, and it is increasingly applied in healthcare, particularly in imaging. AI can assist radiologists by tagging and annotating images, accelerating the detection of lesions and abnormalities. Furthermore, machine learning enables algorithms to identify patterns within large data sets that the human eye may miss, enhancing the accuracy of diagnoses. As these systems are integrated into medical practices, they can support rather than replace healthcare professionals, freeing them up to focus on more complex clinical judgments.
Integrating Data for Personalized Medicine
Current advancements aim to consolidate multiple streams of patient data—ranging from clinical information to imaging and genetic profiles—into comprehensive analyses for more informed treatment decisions. This holistic approach stands in contrast to traditional methods, which often isolate individual aspects of a patient's health. The integration of AI into this process can help healthcare providers synthesize information more efficiently, allowing for tailored treatments that consider the unique characteristics of each patient's cancer. This initiative relies on patient involvement and consent, as ethical data handling is critical in utilizing AI technology securely and responsibly.
Ethical Considerations and Future Directions
As AI technologies evolve, ethical considerations regarding data privacy and consent become paramount, especially in the medical field where algorithms influence patient care decisions. Regulations such as the General Data Protection Regulation (GDPR) in Europe are essential to ensure transparency in how AI systems make decisions and utilize patient data. There is a growing emphasis on understanding the rationale behind AI algorithms, ensuring that they are developed and maintained without biases that could affect treatment outcomes. The ultimate goal remains to enhance clinical decision-making, improve patient experiences, and drive better health outcomes through the responsible use of AI in medicine.
What’s cancer got to do with crabs, artist Jackson Pollock, and artificial intelligence? It’s not a riddle; these are some of the things we’ll explore with surgeon Grant Stewart, computer scientist Mateja Jamnik and radiologist Evis Sala from the Mark Foundation Institute for Integrated Cancer Medicine.
In this episode, we’ll discover how artificial intelligence is making it easier for doctors to diagnose and treat cancer and we’ll share some cancer facts that are both amazing and disturbing. We also learn about the WIRE clinical trial for kidney cancer. WIRE evaluates the effectiveness of giving a short course of drug treatment to patients in the one-month “window of opportunity” between diagnosis and surgery. Patients on the WIRE trial also undergo a suite of new imaging techniques that have been brought together for the first time globally in this clinical trial.
This episode was produced by Nick Saffell, James Dolan, Naomi Clements-Brod and Annie Thwaite.
How did you find us? What do you like about Mind Over Chatter? We want to know. So we put together this survey https://forms.gle/r9CfHpJVUEWrxoyx9. If you could please take a few minutes to fill it out, it would be a big help.
Timestamps:
[00:00] - Introductions
[01:15] - A bit about the guests’ research
[02:45] - The origins of cancer. Why Hippocrates is known as the father of medicine.
[04:00] - How cancer starts
[05:05] - How many types of cancer are there? What are the most common types of cancer?
[06:10] - How do cancers develop? The lifecycle of cancer.
[09:00] - How early in the lifecycle can we see or detect cancer? What size does the cancer cell need to be for us to see it?
[10:15] - What improved machines and AI help with detection and characterisation?
[11:00] - Can we turn imaging into a virtual biopsy?
[12:20] - Defining Artificial Integiilence (AI)
[14:20] - AI and machine learning and how they interlink.
[15:10] - Deep learning and statistical learning.
[16:10] - Origins of AI in medicine and healthcare.
[17:15] - Intro to AI and cancer imagery
[18:30] - How the AI algorithm assists the radiologist
[20:10] - AI and prepared models. How the data is trained to understand what cancer looks like.
[21:20] - The importance of sharing the data set.
[22:00] - Time for a recap!
[28:40] - Ai and surgical robots
[29:30] - AI and screening kidney cancer. Grant’s and Evis’s work using models, imagery, automation to screen for kidney cancer
[31:50] - Explaining the types of imaging in oncology
[33:10] - How Evis uses AI in her imagery
[34:10] - How to scan for ovarian cancer
[35:20] - Comparing images of tumours to paintings. Comparing Jackson to a Mark Rothko painting. Homogeneous or heterogeneous
[37:40] - Describing what the images actually look like from a non-radiologist perspective. Grades of grey. What CT scans and MRI scans look like.
[41:10] - How AI is used throughout the imagery process, not just for clarification.
[42:30] - Comparing the AI in oncology imagery to an Instagram filter. Do we lose any information when we use AI?
[43:15] - Time for another recap!
[48:15] - How do we create and ensure a high quality of data in a healthcare context?
[50:50] - Is there any governance for introducing AI into clinical practice. GDPR and how it impacts AI decisions around the care of a human being. A huge area of research around explainability.
[53:20] - The typical process (modality of data) What Evis, Grant and Matejia are doing with Integrated Cancer Medicine. The techniques
[56:30] - The time has come for integrated care and shared streams of data. Increase the involvement of the patient in their care.
[58:15] - Grant explains the WIRE trial (WIndow-of-opportunity clinical trial platform for evaluation of novel treatment strategies in renal cell cancer).
[1:02:40] - Is it possible to do a holistic analysis? The goal of AI is to help clinicians with personalised medicine.
[1:03:20] - Why it is so important for patients to be involved in oncology AI-based studies.
[1:04:10] - Let's break this episode down and close this thing out.
Guests
Professor of Artificial Intelligence in the Department of Computer Science and Technology (Computer Laboratory) at the University of Cambridge, UK. I am also an associate fellow at the Leverhulme Centre for the Future of Intelligence. Recently I served as Specialist Adviser to the House of Lords Select Committee on Artificial Intelligence. I founded the women@CL initiative.
I am interested in human intuitive reasoning and want to make computers think intuitively too. I build computational models that capture human informal reasoning - I am essentially trying to humanise computer thinking. I combine AI reasoning with machine learning techniques, and apply them to personalise medicine and tutoring systems.
Broadly, my research is in the areas of artificial intelligence, human-like computation, machine learning, automated reasoning, diagrammatic reasoning, knowledge representation, theorem proving, cognitive science, human-computer interaction.
Dr Sala is an academic radiologist with a special interest in Cancer Imaging. She is the Professor of Oncological Imaging at the University of Cambridge, UK. She leads the Radiogenomics and Quantitative Imaging Group in the Department of Radiology. Her current research focuses on integrated diagnostics, through the clinical development and validation of functional imaging biomarkers to rapidly evaluate treatment response using physiologic and metabolic tumour habitat imaging. Her research in the new field of “radio genomics” has focused on understanding the molecular basis of cancer by demonstrating the phenotypic patterns which occur as a result of multiple genetic alterations that interact with the tumour microenvironment to drive the disease in several tumours types. Her work integrates quantitative imaging methods for evaluation of spatial and temporal tumour heterogeneity with genomics, proteomics and metabolomics. She is also working on development and clinical translation of several novel PET tracers.
Professor Grant Stewart @grantissimus
As Professor Surgical Oncology, Professor Stewart aims to leverage the strengths of academic surgical oncology (clinical surgery, trials and biosampling/translational research) to enhance the cure rate following surgical treatment of cancer. He has a specific interest in optimising management of patients with initially localised renal cancer, an area of great need within the disease. Clinically, Professor Stewart undertakes a full range of treatments for renal cancer at Addenbrooke’s Hospital, from robotic partial nephrectomy for small renal cancers to multi-speciality surgery for locally advanced disease. In order to make practice changing developments in this area, Professor Stewart has developed a range of interlinked clinical trials and translational research, which are all underpinned by clinical excellence in managing renal cancer at Addenbrooke’s Hospital, Cambridge. To deliver on the above goals, he coordinates the Cambridge Renal Cancer Collaboration (CamRenCan) a group of over 40 clinicians, translational researchers and basic scientists across the Cambridge Biomedical Campus with a shared interest in renal cancer research. His research is focused on the key clinical questions in initially localised RCC, namely: early detection/screening, early diagnosis approaches, optimal follow-up strategies, neoadjuvant and adjuvant therapies.
Remember Everything You Learn from Podcasts
