Behind the hedges: how to convey uncertainty in imaging reports

Actionable reporting versus unwanted advice in PET-CT reports

AIM

To investigate the opinions of clinician referrers on the presence and type of management advice in reports of positron-emission tomography (PET)-computed tomography (CT) examinations.

MATERIALS AND METHODS

Eighteen clinicians were interviewed using a structured questionnaire about their opinions on further management advice in PET-CT reports.

RESULTS

Opinions varied amongst clinicians, but some themes predominated: (1) advice on further imaging tests and areas outside the referrer's area of expertise are more welcome than other types of advice; (2) a differential diagnosis or clinical significance is of greater value than specific management advice; (3) some referrers do not want any further advice.

CONCLUSION

Before advising on further management, reporters should consider clinicians' opinions regarding this. Seeking the opinion of clinicians in other hospitals may be necessary. A review of the Royal College of Radiologists Actionable Reports standard to ensure it is more in line with the opinions of clinicians may be useful.

Basic Artificial Intelligence Techniques: Natural Language Processing of Radiology Reports

Natural language processing (NLP) is a subfield of computer science and linguistics that can be applied to extract meaningful information from radiology reports. Symbolic NLP is rule based and well suited to problems that can be explicitly defined by a set of rules. Statistical NLP is better situated to problems that cannot be well defined and requires annotated or labeled examples from which machine learning algorithms can infer the rules. Both symbolic and statistical NLP have found success in a variety of radiology use cases. More recently, deep learning approaches, including transformers, have gained traction and demonstrated good performance.