CAN ARTIFICIAL INTELLIGENCE HELP MANAGE OR PREVENT CANCER? Today, we focus on how researchers are cracking the code: How is AI unveiling cancer’s secrets?
I recently thought about my day as an oncologist and was surprised at how artificial intelligence (AI) infiltrates every corner of my specialty, breast cancer.
Screening mammograms? Check. Helping read the pathology report? Check again. Helping predict what drug is best for breast cancer treatment? Yup, AI can help.
First, a sobering observation from the late Stephen Hawking:
“The development of full artificial intelligence could spell the end of the human race….It would take off on its own, and re-design itself at an ever increasing rate. Humans, who are limited by slow biological evolution, couldn’t compete, and would be superseded.”
AI and Breast Cancer
But what about artificial intelligence in healthcare? AI is involved in the care of the vast majority of my patients.
Today, I want to share how AI is revolutionizing the practice of cancer medicine. We’ll look at the following examples:
- Breast imaging and artificial intelligence
- Breast pathology and artificial intelligence
- Breast cancer risk prediction
Artificial Intelligence (AI) has tremendous potential to advance clinical practice and patient care delivery.
Cancer remains among our most challenging health issues, affecting millions worldwide.
Despite significant advances in medical science and technology, the fight against cancer remains an uphill battle. However, in recent years, a formidable ally has emerged in this battle:
Artificial Intelligence (AI). AI is rapidly becoming a game-changer in cancer prediction, diagnosis, and treatment.
In this essay, we will explore the profound impact of AI in unraveling the secrets of cancer, revolutionizing our approach to its detection and management.
Medical Information Volume is Exploding
Artificial intelligence to improve healthcare delivery has arrived. AI is involved in various aspects of my patients’ care.
And just in time. There has been an exponential increase in the medical knowledge needed to guide management for my patients.
The numbers for medical information are breathtaking. Dr. Peter Densen offered these staggering estimates:
- In 1980, medical knowledge doubled every seven years.
- In 2010, the doubling period was 3.5 years.
- In 2020, medical knowledge doubled every 73 days.
Real-World Implications of the Knowledge Explosion
Let’s translate that knowledge doubling time into the life of a medical trainee. What medical students learn in their first three years nowadays would be only six percent of known medical information at graduation.
Sisyphus (or Sisyphos) is a figure from Greek mythology. This Corinthian king gained infamy for his trickery and fierce intelligence.
But his greatest feat? Cheating death and Hades. Not once, but twice. Sisyphus lived up to Homer’s description of him as “the most cunning of men.” (Iliad, 6:153).
However, Corinth’s first king ultimately got his comeuppance when Zeus dealt him the eternal punishment of forever rolling a boulder up a hill in the depths of Hades.
If you practice medicine, you may relate to Sisyphus’ challenge.
The Cancer Conundrum
Cancer is a complex group of diseases characterized by uncontrolled cell growth and the potential to spread to other parts of the body.
Its complexity stems from the fact that there are over 100 different types of cancer, each with its unique behavior, genetic mutations, and treatment challenges.
Traditional cancer detection and treatment approaches have relied heavily on human expertise, limited by human error, subjectivity, and the inability to process vast amounts of data quickly.
AI, on the other hand, offers a new paradigm for understanding and combating cancer. It excels at handling vast datasets, identifying subtle patterns, and making predictions based on data-driven insights.
By leveraging AI, we can delve deeper into the mysteries of cancer, uncovering hidden patterns and gaining a more comprehensive understanding of its underlying mechanisms.
AI in Early Detection
Early detection is often the key to successful cancer treatment. The earlier the cancer is diagnosed, the higher the chances of effective intervention and survival.
AI-powered tools have proven to be invaluable in this regard.
For instance, machine learning algorithms can analyze medical images such as X-rays, MRIs, and CT scans with remarkable precision, detecting abnormalities that might elude even the most skilled human radiologists.
One noteworthy example is the application of deep learning algorithms in breast cancer detection. This is the AI application I encounter most frequently.
Mammography is a common screening tool for breast cancer, but it can produce false positives and negatives.
AI models trained on vast datasets can improve mammogram interpretation accuracy, reducing unnecessary biopsies and ensuring that potential cancers are not missed.
If you read your mammogram report, there is a good chance that the second reader was a robot.
AI and Breast Cancer Screening
Retrospective studies have shown promising results using artificial intelligence to improve mammography screening accuracy and reduce screen-reading workload.
But what about higher-level evidence, such as a randomized trial? An August 2023 research study offers proof of AI’s value.
Swedish researchers assessed the clinical safety of an AI-supported screen-mammogram reading protocol compared with standard screen reading by radiologists following mammography.
As reported in the Lancet Oncology, here are the results:
AI-supported mammography screening resulted in a similar cancer detection rate compared with standard double reading, with a substantially lower screen-reading workload.
The study authors concluded that AI for mammography screening is safe.
Moreover, AI can process a patient’s medical history, genetic information, and other clinical data to assess their risk of developing cancer.
By identifying high-risk individuals, AI can enable healthcare providers to implement proactive screening and preventive measures, potentially saving lives.
AI to Analyze Cancer Cell Signatures
Understanding a breast cancer’s personality or signature is important before recommending treatment.
The first thing I do when I see a new patient with breast cancer is to look at the pathology report, a description of the cancer cells, and some features driving their behavior.
For example, most breast cancers are driven by estrogen and progesterone. We call them estrogen- and progesterone-receptor-positive (ER/PR +).
When these hormones fuel breast cancer cells, we can dramatically drop the available hormone levels. Alternatively, we might use a drug that clogs up the cancer cell receptors that receive the hormone growth signal.
HER2-positive breast cancer is a breast cancer that tests positive for a protein called human epidermal growth factor receptor 2 (HER2). This protein promotes cancer cell growth.
In approximately one of every five breast cancers, the cancer cells have extra copies of the gene that makes the HER2 protein. HER2-positive breast cancers tend to be more aggressive than other breast cancer types.
Human epidermal growth factor receptor 2 (HER2) protein expression is critical for therapeutic decision-making in breast cancer.
Treatments that specifically target HER2 are very effective. These treatments are so effective that the prognosis for HER2-positive breast cancer can be remarkably good.
(On a side note, two weeks ago, at a breast cancer thought leader conference on the Big Island of Hawaii, I sat next to one of the primary creators of the primary drug that targets Her-2. A fan-boy moment for this clinician.)
AI to Determine HER-2 Status
The most common HER2 scoring approach (immunohistochemistry or IHC) suffers from poor interobserver concordance.
Artificial intelligence (AI) may optimize this scoring regarding standardization, accuracy, and efficiency.
Researchers recently reported using an AI-based HER2 IHC quantifier software to support pathologists in standardized breast cancer assessment. They gathered validation specimens from four institutions and five scanners.
As published in the Journal of Clinical Oncology, here are the promising results:
AI software showed quite high agreement when discriminating HER2-neg from HER2-low/positive cases and general HER2 scoring. AI use also appeared to cut HER-2 evaluation time by nearly half.
A separate study showed significant improvement in inter- and intraobserver agreement when the computer-aided reading mode was used to evaluate preselected image fields.
These results demonstrate AI solutions’ potential to increase the consistency and efficiency of HER2 scoring and ultimately improve patient outcomes.
Personalized Treatment with AI
Cancer is not a one-size-fits-all disease. Each patient’s cancer is unique, driven by specific genetic mutations and molecular characteristics.
AI can tailor treatment plans to individual patients, optimizing therapeutic outcomes.
One of the most promising applications of AI in cancer treatment is in the field of precision medicine. By analyzing a patient’s genetic profile, AI algorithms can identify targeted therapies that are most likely effective.
This approach minimizes broad-spectrum treatments like chemotherapy, which can have debilitating side effects, and instead focuses on treatments that specifically target the cancer’s vulnerabilities.
Furthermore, AI can continuously analyze a patient’s response to treatment, adjusting the therapy regimen in real time based on the evolving nature of the cancer.
This dynamic approach enhances the chances of treatment success while minimizing the adverse effects on the patient’s overall health.
Personalized Treatment with AI: Potential
Earlier this year, I saw this interesting headline: “New AI tool could help doctors better personalize breast cancer treatment.”
Here’s the backstory. University of Waterloo (Ontario, Canada) engineers developed an artificial intelligence (AI) tool to help cancer specialists determine whether patients with breast cancer should receive chemotherapy before surgery.
Dr. Alexander Wong, a professor of systems design engineering at the university, offered his take:
“An AI system that can help predict if a patient is likely to respond well to a given treatment gives doctors the tool needed to prescribe the best-personalized treatment for a patient to improve recovery and survival.”
The Canadian researchers trained the AI algorithm using images of breast cancer made with a new MRI method they developed called synthetic correlated diffusion imaging (CDI).
Armed with knowledge obtained from CDI images of historical breast cancer cases (and information on their outcomes), the AI tool has learned to accurately predict if patients would benefit from pre-surgical chemotherapy based on their CDI images.
The team needs to validate their findings by comprehensively testing the algorithm in a larger group of patients.
AI’s Role in Drug Discovery
Finally, developing new cancer drugs is a lengthy and expensive process.
However, AI is poised to accelerate drug discovery and development by predicting which compounds are most likely effective against specific cancer types.
Machine learning models can analyze vast datasets of molecular information, including genetic mutations and protein interactions, to identify potential drug candidates.
AI can also predict how cancer cells may develop resistance to certain treatments, helping researchers devise strategies to overcome this resistance.
This knowledge is invaluable in the ongoing battle against cancer, where the disease often evolves and adapts in response to treatment.
Enhancing Patient Care and Experience
Beyond diagnosis and treatment, AI is transforming the overall patient experience and care delivery.
AI-powered chatbots and virtual assistants can give patients 24/7 access to healthcare information, answer questions, and schedule appointments.
This process enhances convenience and reduces the burden on healthcare staff, allowing them to focus on more critical tasks.
Additionally, AI can improve the quality of life for cancer patients through symptom management and supportive care.
AI algorithms can monitor patient-reported symptoms and side effects, providing timely interventions and recommendations for symptom relief. This personalized approach helps patients cope with cancer treatment’s physical and emotional challenges.
Do Chatbots Give Reliable Cancer Information?
Two recent studies conclude that:
Artificial intelligence (AI) chatbots can give accurate information to common questions about cancer but not so much when providing evidence-based cancer treatment recommendations.
One study, which looked at common cancer-related Google searches, found that AI chatbots generally provide accurate information to consumers, but the information’s usefulness may be limited by its complexity.
The researchers discovered four chatbots generated “high-quality” responses about five cancers. The chatbots did not appear to spread misinformation.
Three of the four chatbots cited reputable sources.
However, the team also found that the information usefulness was “limited” because chatbots wrote responses at a college reading level. Moreover, AI chatbots generated concise answers with no visual aids.
A separate research study assessed cancer treatment recommendations. The AI AI chatbots missed the mark on giving treatment recommendations for breast, prostate, and lung cancers in line with national guidelines.
Ethical Considerations and Challenges
While the potential of AI in cancer detection and treatment is immense, it also raises several ethical and practical challenges.
One of the primary concerns is data privacy and security. AI systems require access to vast patient data, including medical records and genetic information. Ensuring the confidentiality and protection of this sensitive data is paramount.
Another challenge is the need for regulatory oversight and standardization. AI algorithms used in healthcare must meet rigorous quality and safety standards.
There is also the issue of equity in access to AI-powered cancer care. Not all healthcare facilities and regions have equal access to AI technologies, which can exacerbate healthcare disparities.
We must ensure that AI benefits reach underserved populations and that access is not limited to affluent communities.
My Final Thoughts — Cracking the Code: How AI is Unveiling Cancer’s Secrets
Artificial Intelligence is ushering in a new era in the fight against cancer. It transforms how we detect, diagnose, and treat this complex disease.
AI’s ability to process vast amounts of data, identify patterns, and personalize treatment plans is revolutionizing cancer care, offering new hope to patients and healthcare providers.
As AI advances, we must navigate ethical and regulatory challenges, ensuring that AI-powered cancer care is accessible, safe, and equitable for all.
With ongoing research and collaboration, AI promises to unlock more of cancer’s secrets, bringing us closer to a future where cancer is not only detected early but also effectively treated and, ideally, prevented altogether.
In this battle against cancer, AI can be a powerful ally, helping us crack the code and unveil the secrets that have long eluded us.