Impact of clinical information on CT diagnosis by radiologist and subsequent clinical management by physician in acute abdominal pain

• Published in: European radiology 2021-08, Vol.31 (8), p.5454-5463
• Main Authors: Hattori, Shinya, Yokota, Hajime, Takada, Toshihiko, Horikoshi, Takuro, Takishima, Hazuki, Mikami, Wakako, Mukai, Hiroki, Funakoshi, Hiraku, Naganuma, Toru, Yano, Tetsuhiro, Terada, Kazuhiko, Inokuchi, Ryota, Tsuchiya, Satoshi, Akutsu, Akira, Saiga, Atsushi, Kubota, Yoshihiro, Wada, Takeshi, Takada, Akiyo, Fujimoto, Hajime, Uno, Takashi
• Format: Article
• Language: English
• Subjects: Abdomen; Clinical decision making; Computed tomography; Confidence intervals; Diagnostic Radiology; Diagnostic systems; Emergency Radiology; Imaging; Internal Medicine; Interventional Radiology; Laboratory tests; Management; Medical diagnosis; Medicine; Medicine & Public Health; Neuroradiology; Pain; Patients; Physicians; Radiology; Ultrasound
• ISSN: 0938-7994; 1432-1084
• DOI: 10.1007/s00330-021-07700-8

Objective The impact of clinical information on radiological diagnoses and subsequent clinical management has not been sufficiently investigated. This study aimed to compare diagnostic performance between radiological reports made with and without clinical information and to evaluate differences in the clinical management decisions based on each of these reports. Methods We retrospectively reviewed 410 patients who presented with acute abdominal pain and underwent unenhanced ( n = 248) or enhanced CT ( n = 162). Clinical information including age, sex, current and past history, physical findings, and laboratory tests were collected. Six radiologists independently interpreted CTs that were randomly assigned with or without clinical information, made radiological diagnoses, and scored the diagnostic confidence level. Four general and emergency physicians simulated clinical management (i.e., followed up in the outpatient clinic, hospitalized for conservative therapy, or referred to other departments for invasive therapy) based on reports made with or without the clinical information. Reference standards for the radiological diagnoses and clinical management were defined by an independent expert panel. Results The radiological diagnoses made with clinical information were more accurate than those made without clinical information (93.7% vs. 87.8%, p = 0.008). Median interpretation time for radiological reporting with clinical information was significantly shorter than that without clinical information (median 122.0 vs. 139.0 s, p < 0.001). Clinical simulation better matched the reference standard for clinical management when radiological diagnoses were made with reference to clinical information (97.3% vs. 87.8%, p < 0.001). Conclusion Access to adequate clinical information enables accurate radiological diagnoses and appropriate subsequent clinical management of patients with acute abdominal pain. Key Points • Radiological interpretation improved diagnostic accuracy and confidence level when clinical information was provided. • Providing clinical information did not extend the interpretation time required by radiologists. • Radiological interpretation with clinical information led to correct clinical management by physicians.