How AI is Transforming Cancer Understanding but Challenging Doctor-Patient Trust
Artificial intelligence (AI), particularly through large language models (LLMs), is revolutionizing how patients understand cancer. While AI has made cancer information more accessible and comprehensible, it also presents challenges, especially in preserving the crucial doctor-patient relationship and medical judgment. This article explores the impact of AI on cancer education, clinical trial awareness, and the emerging complexities between AI-generated insights and expert medical advice.
The Democratization of Cancer Knowledge
Traditionally, oncologists used complex terminology to explain cancer diagnoses and treatment options, often leaving patients confused and overwhelmed. Terms like “PDL1 status” or “neoadjuvant therapy” were difficult to grasp without specialized medical knowledge, causing many patients to search the internet for explanations, sometimes landing on inaccurate or alarming information.
With advancements in AI, patients can now use LLMs like ChatGPT or Claude to decipher their pathology reports and treatment plans. This empowerment means patients arrive at consultations with foundational knowledge, enabling more meaningful and informed discussions with their doctors.
AI Enhancing Clinical Trial Awareness
One of the most exciting benefits of AI in oncology is improved patient understanding of clinical trials. Patients can comprehend complex trial protocols, endpoints, and possible side effects. This enhanced medical literacy allows patients to meaningfully consent to participation in trials, ultimately expanding access to cutting-edge treatments.
For instance, patients with advanced cancers can directly inquire about relevant clinical trials based on AI-generated information. This knowledge fosters shared decision-making between doctors and patients, centered on evidence-based science.
The Limits of AI: Judgment and Context Matter
Despite its benefits, AI has limitations, especially when it comes to context-based medical judgment. AI can offer generalized, population-based information but cannot tailor advice to the unique biology of an individual’s tumor, their values, or personal risk tolerances.
For example, an AI model may explain that radiotherapy carries side effects and that some early-stage cancers might not require it. However, it cannot communicate the precise recurrence risks based on a patient’s tumor characteristics or how treatment choices impact long-term survival.
Accuracy Challenges in Complex Cases
Research indicates that LLM accuracy drops significantly in complex medical scenarios that require nuanced reasoning. This shortfall underscores the importance of experienced oncologists whose expertise is built on years of clinical observation and judgment honed through navigating uncertainty.
The Widening Trust Gap Between AI and Physicians
A significant concern is the erosion of trust between patients and doctors due to AI’s influence. Patients may trust AI-generated explanations more than nuanced clinical advice, interpreting medical recommendations as biased or profit-driven rather than grounded in expert judgment and patient welfare.
For example, patients might focus on AI’s presentation of chemotherapy side effects without understanding the balanced benefit-risk assessment made by their oncologist. This misalignment can lead to skepticism about potentially life-extending treatments.
Mental health studies reveal similar patterns, with some patients experiencing worsening outcomes or delaying professional help due to AI-generated self-assurance. In oncology, delayed evaluations or treatments validated by AI interpretations can have catastrophic consequences.
Real Patient Stories Underline the Risks
- A patient reassured by AI that a lung nodule was benign delayed follow-up and later developed advanced disease.
- A breast cancer patient misunderstood AI advice on hormone therapy duration, leading to premature discontinuation and relapse.
These examples illustrate the potential harm of unchecked reliance on AI without clinical oversight.
Moving Forward: Balancing AI and Medical Expertise
AI is unlikely to replace oncologists because while knowledge can be replicated, the clinical judgment born of experience in managing uncertainty cannot be duplicated by algorithms. The challenge lies in integrating AI as a supportive tool without undermining the doctor-patient relationship.
Patients must understand that doctors who provide nuanced advice are leveraging extensive clinical experience and personalized assessment, not acting defensively or with ulterior motives.
Healthcare systems should track outcomes related to AI use in patient decision-making to ensure safety and efficacy. Additionally, developers must recognize that AI models are designed to be agreeable, not necessarily neutral or critically evaluative, which influences the information patients receive.
Conclusion
AI has transformed cancer education by breaking down barriers and making knowledge more accessible, especially in regions with limited specialist access. However, informed medical judgment remains irreplaceable. Ensuring patients receive contextual, personalized advice while navigating AI-generated information is critical to achieving the best outcomes in cancer care.
Author: Narayana Subramaniam, Lead Consultant, Head and Neck Surgery and Oncology, Aster Hospitals, and Adjunct Faculty, Indian Institute of Science, Bengaluru.
