Exploring the transformative role of personalised medicine in cancer treatment, we delve into the dynamic interplay between human expertise and artificial intelligence (AI), revealing how this synergy is reshaping oncology care.
Cancer, a multifaceted disease, demands tailored treatment strategies. Personalised medicine in cancer treatment represents a paradigm shift, moving away from one-size-fits-all approaches to customised care plans based on individual genetic profiles. AI’s emergence enhances this personalisation, offering unparalleled data analysis capabilities.
AI’s Role in Enhancing Personalised Cancer Therapies
- AI-Driven Cancer Drug Discovery: Researchers have developed tools like CancerOmicsNet, an AI-driven cancer drug discovery engine. This system, adapted from algorithms initially designed for mapping social networks, can identify crucial genes in cells by analysing genetic changes in tumours. This advancement enables the creation of personalised treatment plans, including customised radiation therapy doses, based on the molecular and genetic alterations observed in tumour images across various cancer types.
- Tailoring Treatment to Cancer Subtypes: Stanford scholars have crafted an algorithm that can identify the most effective treatments for different cancer subtypes. This approach ensures that patients receive the most suitable treatment based on the specific characteristics of their cancer, improving the effectiveness and efficiency of cancer care.
- Enhancing Diagnostic Accuracy: AI-based systems are aiding pathologists in diagnosing cancer more accurately and consistently, which reduces errors. Predictive AI models can also estimate an individual’s likelihood of developing cancer by identifying risk factors. This capability is crucial in developing more personalised and effective treatment strategies for each patient.
- Precision Medicine Applications: AI techniques in precision medicine are utilized to explore novel genotypes and phenotypes data. The primary goals here include early diagnosis, screening, and creating personalised treatment regimes based on genetic features. This approach is pivotal in tailoring treatments to individual patients’ genetic profiles, leading to more effective and targeted therapies.
- Cumulative Advances in AI and Oncology: The field of AI in oncology has seen significant advancements, including the refinement of machine learning and deep learning algorithms. The expansion in the variety and depth of databases, including multiomics (the integration of various “omic” data), is crucial. These developments are bringing the promise of highly personalised oncology care closer to realisation, ensuring that treatments are more accurately aligned with each patient’s unique cancer profile.
Human Insight: The Indispensable Element
The integration of AI in personalized cancer therapy, while groundbreaking, highlights the indispensable role of human insight. Oncologists’ expertise and experience are vital in interpreting the data generated by AI, ensuring that treatment decisions are not only scientifically sound but also empathetically delivered.
- Interpreting AI-Generated Data: AI algorithms can process vast amounts of data to identify patterns and predict treatment responses. However, the human expertise of oncologists is critical in interpreting these outputs. They consider factors beyond the data, such as patient preferences, social circumstances, and ethical considerations, ensuring a holistic approach to treatment.
- Balancing Science and Empathy: AI brings precision to cancer treatment, but it is the human touch that adds empathy. Oncologists use AI data to inform their decisions but also rely on their clinical experience and understanding of patient needs. This balance ensures that treatments are not just effective, but also compassionate and patient-centric.
Case Studies: Success Stories of AI and Human Collaboration
The collaboration between AI and human expertise in cancer treatment has led to several notable successes, as illustrated by different case studies:
Skin Cancer Diagnosis
A study published in Nature Medicine explored the impact of AI in skin cancer diagnosis. It showed that AI-based support, specifically multiclass probabilities, improved the accuracy of human raters from 63.6% to 77.0%.
The study emphasised the importance of how AI outputs are presented, finding that AI-based support was most beneficial when it was aligned with the specific diagnostic task at hand. Interestingly, the least experienced clinicians gained the most from AI support. However, the study also cautioned that faulty AI could mislead clinicians across the spectrum, including experts.
AI in Telemedicine
The same study examined the role of AI in telemedicine, particularly during the COVID-19 pandemic. It found that AI-based support could help in managing workloads and expanding performance capabilities in healthcare delivery. In scenarios involving telemedicine, AI was able to recognize 95.2% of patients with skin cancer at a specificity of 59.2%, indicating its potential to triage high-risk cases and extend intervals between face-to-face visits in low-risk cases.
Precision Oncology at Charité – Universitätsmedizin Berlin
Researchers at Charité in Berlin studied whether AI could assist in selecting personalised treatments for cancer patients. They found that while AI models could identify personalised treatment options in principle, they were not as capable as human experts.
The study involved creating molecular tumour profiles of fictitious patients and comparing AI-generated treatment options with those recommended by experts. Despite the promise of AI, the study highlighted the superiority of human expertise in complex decision-making processes.
Oncology Drug Development
A case study by Cognizant demonstrated how AI and data science were utilised by a major pharmaceutical company to process vast amounts of data from clinical trials and research, specifically for acute myeloid leukaemia (AML) treatment. This collaboration led to an automated solution that made identifying optimal drug doses much faster, demonstrating the powerful synergy of AI and human expertise in drug development.
Radiomics for Cancer Subtype Identification
A Stanford-led research team used AI to identify subtypes of cancer and match them to precision medicines. The study highlighted the role of radiomics, a field that combines human knowledge with AI’s computational power.
The team’s AI algorithms identified four tumour subtypes based on their radiographical features, which helped match these subtypes to favourable treatments. For example, a lung cancer subtype was found to respond better to immunotherapy, leading to improved survival rates. This research exemplifies the successful integration of AI with human expertise in the field of cancer treatment.
These case studies illustrate the critical role of human expertise in guiding and interpreting AI’s capabilities in cancer treatment. While AI offers significant advancements in diagnostic accuracy and treatment planning, human insight remains crucial for effective and ethical decision-making in patient care.
Ethical and Practical Considerations
While the benefits are significant, ethical and practical considerations must be addressed. Issues such as data privacy, algorithmic biases, and the digital divide pose challenges that need careful navigation to ensure equitable and ethical implementation of AI in cancer care.
The future of personalised medicine in cancer treatment looks promising, with ongoing research and development. The aim is to further refine AI algorithms and enhance the collaboration between technology and clinicians, ultimately leading to more innovative and effective cancer therapies.
References
- LSU Research Team Uses AI to Quickly Discover Personalized Cancer Cures. (n.d.). https://www.lsu.edu/mediacenter/news/2022/09/wfl-cancer.php
- Using AI To Personalize Cancer Care. (2021, August 9). Stanford HAI. https://hai.stanford.edu/news/using-ai-personalize-cancer-care-0
- Sebastian, A. M., & Peter, D. (2022, November 28). Artificial Intelligence in Cancer Research: Trends, Challenges and Future Directions. Life. https://doi.org/10.3390/life12121991
- Rezayi, S., Kalhori, S. R. N., & Saeedi, S. (2022, April 7). Effectiveness of Artificial Intelligence for Personalized Medicine in Neoplasms: A Systematic Review. BioMed Research International. https://doi.org/10.1155/2022/7842566
- Shreve, J. T., Khanani, S. A., & Haddad, T. C. (2022, July). Artificial Intelligence in Oncology: Current Capabilities, Future Opportunities, and Ethical Considerations. American Society of Clinical Oncology Educational Book, 42, 842–851. https://doi.org/10.1200/edbk_350652
- Tschandl, P., Rinner, C., Apalla, Z., Argenziano, G., Codella, N., Halpern, A. C., Janda, M., Lallas, A., Longo, C., Malvehy, J., Paoli, J., Puig, S., Rosendahl, C., Soyer, H. P., Zalaudek, I., & Kittler, H. (2020, June 22). Human–computer collaboration for skin cancer recognition. Nature Medicine. https://doi.org/10.1038/s41591-020-0942-0
- Berlin, C. U. (n.d.). Press reports. Charité. https://www.charite.de/en/service/press_reports/artikel/detail/personalized_cancer_medicine_humans_make_better_treatment_decisions_than_ai/
- Data science fast-tracks cancer drug. (n.d.). www.cognizant.com. https://www.cognizant.com/us/en/case-studies/data-science-ai-cancer-drug-development
