Artificial intelligence is rapidly transforming healthcare, often elevating user experience and improving operational efficiency. In some domains, however, AI is no longer just a convenience — it has become essential. Radiotherapy is one such area. As patient volumes surge, digital oncology systems require significantly higher levels of autonomy to sustain safe, timely, and high‑quality care. This includes not only treatment management, but increasingly, elements of treatment planning as well.

In this context, cybersecurity is no longer a supporting function; it is a foundational prerequisite for reliable cancer care. The encouraging news is that robust protection is achievable by applying mature, well‑established security standards — adapted thoughtfully for AI‑driven clinical environments. Frameworks such as MITRE ATLAS offer structured approaches tailored to adversarial threats against machine learning systems.

This presentation explores the unique cybersecurity challenges introduced by AI‑enabled healthcare applications and demonstrates a practical strategy for addressing them through a focused use case: an AI Agent designed to support radiotherapy management.