In global conversations about the future of radiology, artificial intelligence (AI) is often positioned as the solution to the field’s most pressing challenges. From improving diagnostic accuracy to reducing physician workload, the possibilities seem extensive. Yet, as a medical student training in Nigeria, I have come to realize that the conversation around sustainability in radiology must begin much earlier, specifically at the point of access. In my training, I have learned the gold-standard investigations for many conditions: advanced imaging modalities that provide definitive answers. However, in practice, I have come to realize that the first-line investigation is rarely the gold standard; it is simply the most accessible and affordable option available to the patient.
I remember a ward round during my surgery rotation where a computed tomography (CT) scan was being considered for a patient. The conversation shifted instantly not to clinical indications or urgency, but to financial feasibility. In the end, the team proceeded with more basic imaging and clinical judgment. That was a defining moment that reshaped my understanding of what it means to practice medicine in my setting, but for my consultants and residents, it was just another day. In many parts of the world, considerations before requesting imaging extend far beyond clinical need. Not every facility has access to advanced equipment, often requiring patients to travel to distant centers for scans they may not even be able to afford once they arrive. These barriers introduce delays that significantly affect outcomes. As a result, clinicians frequently rely on a combination of basic imaging and rigorous clinical reasoning to arrive at a working diagnosis.
This reality has shifted how I think about sustainability in health care. While often framed in terms of environmental impact or technological advancement, in low-resource settings, sustainability also means building systems that are accessible, efficient, and reliable over time. This is not a campaign against the global enthusiasm for AI in radiology. I believe it is valid and justified. However, its integration must be thoughtful and context-specific. If AI is developed primarily for systems that already function efficiently, it risks optimizing care for some while remaining inaccessible to others. For me, this has sparked a growing interest in how AI can be adapted to meet the realities of low-resource environments. Rather than focusing solely on high-performance models trained on datasets from well-resourced systems, there is an opportunity to design solutions that prioritize accessibility. This could include AI tools that can derive high-quality insights from lower-cost, “noisy” imaging equipment, or mobile platforms that eliminate the need for expensive patient referrals to distant centers.
Equally important is the role of education. As medical students entering a rapidly evolving field, our understanding of sustainability must evolve with it. This includes not only learning to use emerging technologies but also developing the awareness to question how and where they are applied. Sustainable progress requires a balance between innovation and equity, ensuring that advancements do not leave certain populations behind. As I consider my future in radiology, I am increasingly drawn to the intersection of technology, access, and global health. The question is no longer just how to make radiology more advanced, but how to make it more inclusive, because, ultimately, the success of any innovation is measured not by its sophistication, but by its impact.
Amarachi Amanda Dukor is a physician in Nigeria.
