When AI Meets Oncology

Many experts believe that human history can be divided into four distinct eras: the Prehistoric Era, the Agricultural/Industrial Era, the Information Age, and now … the Artificial Intelligence Era. AI is no longer the future– it is the present. From ChatGPT to exploring the depths of space colonization, AI is doing things we never thought possible. In the status quo, we are witnessing an upward trend in AI in medicine and healthcare as it offers more tools for faster and more accurate diagnoses, personalized treatments, and much more. One of the leading causes of death in the world is cancer, and to this day, there is still no known cure for the many types of cancer. We must use AI to guarantee that future generations do not need to suffer from cancer as much as the world has in the past.

AI diagnosis

Unfortunately, a universal cure for cancer has yet to be discovered; however, advancements in medical treatment, particularly when administered early, can significantly combat the disease and improve patient outcomes[1]. A diagnosis must be made as soon as possible to create the best chance at a successful treatment, and AI is key to making a diagnosis as quickly as possible. Harvard Medical School predicts a 40% improvement in health outcomes when using AI for diagnoses[2].

The American Cancer Society revealed that the new leading cause of cancer in women is breast cancer, and the only cure is to be treated earlier by having mammograms annually[3]. Radiologists play a pivotal role in early detection through interpreting imaging studies, including mammograms, ultrasounds, and MRIs, which can reveal changes in the breast before a patient or physician can speak to them. Imagine you are a woman walking into your mammogram and getting diagnosed with breast cancer … by AI. Research has demonstrated that the use of AI resulted in a significant reduction in false-positive rates by 37.3% and biopsy requests by 27.8% while maintaining sensitivity [4]. Lower false-positive rates are desirable because they reveal that the system has high accuracy and reliability. The study ultimately suggests that AI can not only assist radiologists in improving accuracy, consistency, and reliability, but it also proposes that AI is one step ahead of human experts. 

Personalized Treatment Plans

The American Cancer Society defines precision medicine as how health care providers can offer and plan specific care for their patients, based on their particular genes, proteins, and other substances in a person’s body[5]. Doctors would usually use MRIs or genomic sequencing machines to find a personalized treatment plan. Still, now they are utilizing AI to customize treatment and offer a more targeted and effective approach to healthcare, specifically cancer. 

It is crucial to understand that due to its insane complexity and diverse nature, cancer diagnosis and treatment must be taken with the utmost vigilance and attentiveness. In head and neck cancer, radiotherapy must be carefully targeted to destroy the tumor without damaging nearby healthy tissues, like the eyes, spinal cord, or salivary glands. This process, called contouring, is extremely time-consuming and complex. Google DeepMind, an artificial intelligence platform, developed an AI model that automates the segmentation of healthy tissues and tumors in CT scans for radiotherapy planning[6]. The AI successfully reduced the time needed for planning from hours to just minutes, saving valuable time for other patient care. AI has assisted radiotherapists by developing AI models to shorten the time needed, but AI has also heavily helped pathologists by detecting cancer cells with high accuracy. PathAI, the global leader in AI-powered pathology, identified specific biomarker expressions, which helped recognize patients more likely to benefit from immunotherapy by detecting PD-L1 expression levels in lung and bladder cancer[7]. This advancement not only speeds up diagnosis but also supports more personalized treatment plans, improving outcomes for patients. As AI continues to evolve, its role in oncology is becoming indispensable, enhancing both precision and efficiency in cancer care.

Drug Development

Cancer drugs are crucial because they are the building blocks of modern cancer treatment, offering a range of options to target and combat the disease, and now AI is expanding the drug field. According to medical research, SIGX1094 is the first drug candidate developed using organoid models and AI to enter clinical trials, addressing a critical unmet need in DGC treatment[8]. DGC is a highly aggressive form of stomach cancer that has very limited treatment options, but luckily, AI has expanded the constraints. AI is directly contributing to finding many cancer drugs, not just SIGX1094, but it also discovers other therapeutic agents, like medical isotopes. Medical isotopes, also known as radioisotopes or radionuclides, are radioactive substances used in medicine for both diagnosis and treatment. They are a crucial part of nuclear medicine, a field that utilizes radioactive materials to understand and treat various diseases. Research shows that AI is enhancing its role in nuclear medicine, specifically medical isotopes [9]. The Argonne National Laboratory further proves how important medical isotopes are in the cancer field, and how AI could also help doctors pair radioisotope candidates with individual tumors. AI is essential to expanding the nuclear medicine field because every year, doctors perform more than 40 million medical procedures that rely on radioisotopes[10]. AI is not only key to drug development, but also crucial to finding other materials that aid in cancer treatment. In the field of drug development and therapeutic agents, artificial intelligence is crucial to making any sort of change. 

The Limits of Artificial Intelligence 

If AI is so efficient and accurate, why don’t we just implement artificial intelligence in the medical/healthcare field permanently? AI might be a better doctor than humans, but it still isn’t a human being. The Indian Journal of Medical Sciences claims that some notable challenges of AI in Pathology and Radiology include a lack of standardization, generalization, and regulatory/ethical considerations [11]. A lack of standardization in AI expresses the absence of uniform guidelines, protocols, and best practices for deploying and developing artificial intelligence. AI’s ability depends on the data set it was trained on, so generalization refers to its limits in the medical field, especially cancer, since it is so extensive. Another reason is that AI has no consideration for ethical and regulatory concerns, which encompass a broad range of issues related to patient rights, the responsible use of technology, and data privacy. An example of AI violating data privacy could be when it unintentionally memorizes and stores sensitive information, like a medical record, which can lead to potential data breaches. These flaws make AI a little less reliable, but the worst part of AI has not even been discussed: bias. The Virginia Law Review claims that people of color receive inferior pregnancy-related healthcare, and healthcare generally, because medical AI technologies will be developed, trained, and deployed in a country with striking and unforgivable racial disparities in health[12]. Artificial Intelligence certainly should not be implemented in the healthcare system if it creates inequality. Even though AI is strong, accurate, and essential to helping decrease cancer deaths in the world, it has its limits in the medical field. 

Interview with Science Teacher

Q: Would you put artificial intelligence into the medical field specifically for cancer?

A: “Yes, I would put artificial intelligence into the medical field for cancer. I do believe that it should be implemented with caution. AI is the future of medicine, and it does diagnose people accurately and quickly, which probably increases the number of lives saved. So, yes.”

Q: Would you trust a relative to be treated with AI?

A: “Yes, I would trust a relative to be treated, but only if there was a doctor there.”

Works Cited

1. The American Cancer Society. “Can Cancer Be Cured? | Does Treatment Cure Cancer?” American Cancer Society, 6 May 2021, https://www.cancer.org/cancer/understanding-cancer/can-cancer-be-cured.html. Accessed 28 May 2025. 

2. Harvard Medical School. “AI in Health Care: From Strategies to Implementation.” https://execonline.hms.harvard.edu/artificial-intelligence-in-health-care-from-strategies-to-implementation?utm_source=Google&utm_network=g&utm_medium=c&utm_term=healthcare%20in%20ai&utm_location=9190264&utm_campaign_id=21212474098&utm_adset_id=16084532590. Accessed 28 May 2025. 

3. The American Cancer Society. “Cancer Facts for Women | Most Common Cancers in Women.” American Cancer Society, 5 May 2025, https://www.cancer.org/cancer/risk-prevention/understanding-cancer-risk/cancer-facts/cancer-facts-for-women.html. Accessed 28 May 2025.

4. National Library of Medicine. Artificial Intelligence in Breast Cancer Diagnosis and Personalized Medicine, https://pmc.ncbi.nlm.nih.gov/articles/PMC10625863/#sec4. Accessed 28 May 2025. 

5. The American Cancer Society. “Precision or Personalized Medicine | Precision Medicine for Cancer.” American Cancer Society, 15 June 2023, https://www.cancer.org/cancer/managing-cancer/treatment-types/precision-medicine.html. Accessed 28 May 2025.

6. “Applying machine learning to radiotherapy planning for head & neck cancer.” 30 August 2016, https://deepmind.google/discover/blog/applying-machine-learning-to-radiotherapy-planning-for-head-neck-cancer/. Accessed 28 May 2025.

7. PathAI. “PathAI Introduces TumorDetect, an AI Solution to Automate Tumor Assessment and Case Prioritization For Anatomic Pathology Laboratories.” 14 December 2023, https://www.pathai.com/resources/pathai-introduces-tumordetect-an-ai-solution-to-automate-tumor-assessment-and-case-prioritization-for-anatomic-pathology-laboratories/. Accessed 28 May 2025. 

8. MedPath. “FDA Grants Fast Track Status to SIGX1094 for Diffuse Gastric Cancer.” 28 January 2025, https://trial.medpath.com/news/4eb040ebdf78d69e/world-s-first-drug-candidate-developed-by-organoid-and-ai-enters-clinical-trials?. Accessed 28 May 2028. 

9. Saboury, Babak, et al. “Artificial Intelligence in Nuclear Medicine: Opportunities, Challenges, and Responsibilities Toward a Trustworthy Ecosystem.” 2023, https://jnm.snmjournals.org/content/64/2/188. Accessed 28 May 2025.

10. Argonne National Laboratory. “Accelerating cancer treatments with the power of isotopes.” 3 May 2021, https://www.anl.gov/article/accelerating-cancer-treatments-with-the-power-of-isotopes. Accessed 28 May 2025. 

11. Shah, Rajendra M. “Overcoming diagnostic challenges of artificial intelligence in pathology and radiology: Innovative solutions and strategies.” Indian Journal of Medical Sciences, 2023, https://ijmsweb.com/overcoming-diagnostic-challenges-of-artificial-intelligence-in-pathology-and-radiology-innovative-solutions-and-strategies/#:~:text=Lack%20of%20standardization,images%20and%20identify%20potential%20abnormalities. Accessed 28 May 2025.

12. Bridges, Khiara M. “Race in the Machine: Racial Disparities in Health and Medical AI.” Virginia Law Review, vol. 110, no. 2, https://virginialawreview.org/articles/race-in-the-machine-racial-disparities-in-health-and-medical-ai/. Accessed 28 May 2025.