Disparities in the Way Radiologists, Referring Physicians, and Patients See the Language Used to Report Incidental Findings

Abstract

Background: Incidental findings on cross-sectional imaging are common, often causing concern due to the possibility, however remote, that they could indicate malignancy. While radiologists aim to document these findings thoroughly, the language used may lead to varying interpretations by referring physicians and patients, potentially resulting in unnecessary follow-up tests, increased costs, and patient anxiety. The language radiologists use to describe low-risk incidental findings, especially those with minimal clinical significance, is underexplored, particularly from the perspectives of radiologists, referring physicians, and patients. Methods: This study involved three anonymous surveys administered to radiologists, referring physicians, and patients at a large academic medical center. The surveys assessed how each group interprets 10 expressions used to describe a 5-mm incidental liver lesion, based on its perceived likelihood of malignancy. Radiologists, referring physicians, and patients rated their perceived likelihood of malignancy for these expressions and provided feedback on whether follow-up imaging was warranted. Data were analyzed using median responses for malignancy likelihood, follow-up imaging decisions, and descriptive statistics. Results: A total of 13 radiologists (93%), 59 referring physicians (6% response rate), and 51 patients participated. Referring physicians and patients consistently rated the likelihood of malignancy higher than radiologists, particularly for phrases like "tumor not excluded," which were associated with the highest perceived likelihood of malignancy. Regarding follow-up imaging, 75-83% of referring physicians indicated they would order follow-up for phrases with uncertainty about malignancy, while only 2-46% chose to do so for phrases indicating benign findings. Radiologists favored "cyst" to describe the lesion, while referring physicians favored "benign cyst." Radiologists unanimously agreed that no follow-up was necessary for the benign lesion. Conclusion: Both referring physicians and patients expressed greater concern than radiologists regarding incidental findings, especially when the language used suggested uncertainty about malignancy. Ambiguous terms such as "tumor not excluded" increased anxiety and the likelihood of follow-up tests. This highlights the need for radiologists to refine their terminology to improve communication clarity and reduce unnecessary patient anxiety and healthcare costs