Read the following passage and mark the letter A, B, C or D on your answer sheet to indicate the option that best fits each of the numbered blanks from 11 to 15.
One of the most promising recent applications of artificial intelligence in medicine is rare-disease diagnosis. This field has long challenged clinicians because symptoms are often diverse, inconsistent, multisystem, and easily mistaken for more common disorders. In addition, the relevant knowledge base changes quickly as new genetic conditions continue to be identified. (11) ______. In 2026, Nature reported DeepRare, an agentic system that combines clinical descriptions, genetic evidence, specialised tools, and literature retrieval to generate ranked differential diagnoses with traceable reasoning. Rather than producing a single obscure answer, the system links each hypothesis to verifiable medical evidence, (12) ______.
This feature matters in practice because clinicians are unlikely to trust recommendations they cannot inspect, challenge, or justify. (13) ______. Used carefully within specialist workflows, such tools may help hospitals shorten diagnostic delay and support patients who currently face long, uncertain journeys. Even so, recent reviews of AI agents in healthcare stress that progress will depend on more than technical accuracy alone. (14) ______, these systems must be evaluated for safety, governance, and real-world reliability before they are widely deployed. In the long term, the aim is not to replace clinical judgment, (15) ______.
(Adapted from https://www.nature.com)