Why reporting matters.

Dear colleagues,

Artificial intelligence (AI) has expanded the possibilities of diagnostic imaging, from detecting and classifying findings to supporting clinical interpretation. Yet reporting remains inconsistent. The origin and composition of datasets, model development, validation procedures and performance are often insufficiently described. These omissions limit reproducibility, cross-study comparisons and critical appraisal. Without this information, how can the real-world usefulness of an AI model be adequately assessed?

Technical novelty and high performance under controlled conditions are no longer sufficient contributions in themselves. Neither establishes clinical usefulness. Reporting guidelines such as CLAIM, TRIPOD+AI and STARD-AI provide important recommendations; nevertheless, a gap persists between their availability and the information routinely reported in imaging AI research.

Each letter of CLAIRE reflects a reporting need:

Initiated in 2025, CLAIRE seeks to bring technical, methodological and clinical reporting requirements into a unified, operational structure. Its purpose is to support transparent reporting and structured appraisal of the information presented in a study—not to replace assessments of methodological quality or risk of bias. Incomplete reporting limits what can be judged; it does not, by itself, establish that the methods were inadequate.

Following the publication of the CLAIRE Framework, an international consensus process is being developed to examine and refine its components from complementary clinical, scientific and computational perspectives. Our aim is not to replace existing guidelines, but to help authors communicate their methods and results more clearly, so that readers can better understand what the evidence does—and does not—support.