Diagnostic performance of smartphone reading of the coronary CT angiography in patients with acute chest pain at ED

Abstract

Purpose

The aims of this study were to simulate mobile consultation for the coronary computed tomography angiography (CCTA) at the emergency department (ED) and to measure the diagnostic performance of the mobile reading.

Materials and methods

A total of 107 patients with acute chest pain who underwent CCTA and coronary angiography (CAG) were included. The CCTA images were reviewed by a cardiac radiologist using a smartphone. The degree of stenosis at each coronary segment was scored with 4-point scale (score 1, < 50%; score 2, 51%-70%; score 3, 71%-90%; score 4, > 90%). The degree of stenosis at each coronary segments were also scored with preliminary CCTA report by on-call residents, final CCTA reports by in-house attending cardiac radiologists, and CAG. Interobserver agreement was measured using κ statistics. The areas under the receiver operating characteristic curves (AUCs) for diagnosing segments with obstructive stenosis were compared between each reader and CAG.

Results

The smartphone reader's reading was more similar to the CAG results and in-house radiologists' reports than reading of on-call residents. The diagnostic performance of smartphone reading for detection of obstructive stenosis was significantly greater than that of on-call residents (AUC, 0.89 vs 0.75; P < .001) and did not significantly differ from that of the in-house radiologists (AUC, 0.89 vs 0.90; P = .05).

Conclusion

Smartphone reading by the cardiac radiologist was superior to the on-call residents' reading. Further study with real-time mobile consultation needs to be investigated to evaluate whether improvement in diagnostic competency can make a difference in the outcome of patients.

Introduction

Chest pain accounts for a substantial proportion of ED visits. Approximately one third to two thirds of patients presenting with chest pain are admitted, and only 15% to 25% are diagnosed with acute coronary syndromes [1]. Therefore, efficient and accurate determination of acute coronary syndromes is important in the ED.

Coronary computed tomography (CT) angiography (CCTA) is a robust and fast tool for noninvasive evaluation of coronary artery disease (CAD) [2]. Several randomized clinical trials revealed the CCTA-based strategy for patients at low to intermediate risk of CAD is useful in the ED [1], [3]. In the ED, timely and accurate CCTA interpretation is needed to reduce crowding, which is associated with adverse cardiovascular outcome in patients with chest pain [4]. For the interpretation of CCTA, the reader's level of experience is a strong determinant of proficiency [5], [6]. However, the after-hour CT interpretations in the ED usually rely on less experienced on-call residents for night time coverage or training purpose [7], [8]. Therefore, real-time consultation of CCTA to the specialists will be beneficial in triaging the patients with chest pain in ED.

Advancement in smartphone technology with combination of specialized medical applications has proven its value in hospital environments, especially in the ED setting [9], [10]. Pocket-sized teleradiology terminals using a smartphone have been tested for its technical feasibility in consultation purposes in previous studies [11]. Many studies have suggested the potential role of off-site smartphone reading as a radiology consultant in critical situations such as acute stroke, intracranial hemorrhage, and inconclusive diagnosis of appendicitis [12], [13], [14]. To our knowledge, there is no previous study demonstrating feasibility and accuracy of off-site smartphone reading for the CCTA in the ED.

Thus, the aims of our study were to simulate mobile consultation for the CCTA at ED and to measure the diagnostic performance of the mobile consultation.