Remote imaging via FaceTime for potential long-distance diagnosis of suspected cardiac structural and shunt anomalies on contrast echocardiography (The RIFLE Study)

Abstract

Certain cardiac anomalies can be missed by standard transthoracic echocardiography and are usually confirmed through contrast echocardiography, a specialized procedure that warrants supervision by a cardiologist with advanced echo training. However, the maldistribution or lack of echo specialists in certain countries, especially in archipelagic settings, deprives many patients of such procedures and probably contributes to delayed diagnosis and treatment. We wanted to find out whether onsite interpretations of contrast echocardiograms of patients with suspected cardiac anomalies are comparable to offsite readings made via the real-time videoconferencing technology of FaceTime. Contrast echocardiograms done on adults or grown-up patients with suspected cardiac structural or shunt anomalies during a one-year period at the Non-invasive Laboratory of the Philippine Heart Center were included in the study. Using the internet-enabled technology of FaceTime, prerecorded echo videos from DVD files were simultaneously viewed and interpreted by two echocardiographers, with the onsite reader using a TV monitor and offsite reader using an Ipad device. Interpretations were compared with respect to 22 pre-specified echocardiographic paramaters, which included chamber size, presence of defect on color Doppler and contrast study, and final diagnosis. Agreement in echo readings was reported as overall agreement rate (% of congruent readings divided by total number of data points for comparison) and Kappa coefficient. The impact of contrast echocardiography on the final diagnosis at both sites was likewise assessed (accuracy rates for pre- and post-contrast diagnoses, with the official test report as reference standard). Video and audio quality were evaluated using a subjective scoring scale, while internet connection was assessed in terms of lag time to transmission (milliseconds), as well as download and upload speeds (Megabytes per second). Sixty-eight (68) videoconferencing sessions were conducted on patients 37 ± 14 years (62% females). Overall reading agreement rate was 89% (1333 congruent readings out of 1496 data points), with most disagreements arising from chamber size evaluation (left ventricle, left atrium and coronary sinus dimensions). Good concordance (kappa > 0.60) was demonstrated between onsite and offsite readings in 14 of 22 parameters. Accuracy rates were comparable for precontrast onsite (51%) and offsite (54%) readings, with a substantial improvement in post-contrast readings (88% onsite and 79% offsite). Only 11 of 68 sessions (16%) had discordant final diagnoses, most of which consisted of overdiagnosis by the offsite reader. Video (3.6) and audio (3.8) quality scores were high (maximum of 4) for both sites, with acceptable lag time (PING < 250 ms), download (2.1 Mbps) and upload (0.3 Mbps) speeds. Onsite interpretations of contrast echocardiograms are comparable to offsite readings made via real-time videoconferencing, with contrast echocardiography enhancing accuracy of diagnosis at both ends. Moreover, FaceTime was feasible, generating acceptable internet speeds as well as obtaining high audio and video quality scores at both ends of the transaction. These findings suggest a role for tele-echocardiography in long-distance, point-of-care diagnosis of patients with structural or shunt anomalies, especially in areas of the country where echocardiographic expertise is very much lacking.