Feasibility of novel robot-assisted, remotely-performed echocardiographic examination

Abstract

Abstract Background Access to echocardiography services outside metropolitan areas is persistently limited. We sought to determine the feasibility and efficiency of performing remotely-controlled, semi-automated robot-assisted echocardiographic examinations (robot-assisted echo) by cardiac sonographers as a potential avenue to provide standard echocardiographic specialist services to remote and rural areas as well as permit expedited diagnoses enabling earlier management of disease. Methods Two accredited cardiac sonographers (Cardiac Investigations Unit) undertook 6 hours virtual training, and 10 hours hands-on robot-assisted echo simulation training; using a novel, semi-automated robotic arm system to remotely manipulate an ultrasound probe to acquire echocardiographic images. Sonographers then performed remotely-performed robot-assisted echo examinations on staff volunteers, and consenting inpatients in whom echocardiography was clinically indicated. Cameras, probe pressure sensors and an ultrasound platform remote console were utilised to support remote operation. Traditional echocardiography (manipulation of probe by hand) was also performed as a reference. Feasibility to obtain a comprehensive robot-assisted echo was compared to traditional echocardiography. Sonographer learning curve and efficiency (image acquisition time and offline measurement time) was also investigated. Results Total cohort was n=78 (age 51±15yrs; 57% male) with 63% patients. All participants had robot-assisted echo and traditional echocardiograms performed within 72hrs. Average feasibility (%) of robot-assisted echo as (a) 11 averaged assessments and (b) 38 averaged measures was 92±9 (a) and 86±11 (b), compared to (a) 99±1 and (b) 97±7 by traditional examination (all respectively). Sonographer learning curve was identified via 1:1 block subcohort analysis, with averaged assessment feasibility (n=33:31) being 86±19 (first block) to 98±3 (last block) and averaged measurement feasibility (n=32:31) being 77±11 (first block) to 92±7 (last block), all respectively. In the staff participant cohort (n=31), average image acquisition efficiency in robot-assisted echo was poorer than that of traditional examination (47+9 minutes compared to 19+2 minutes; p<0.0001), respectively. Average efficiency of offline measurement time was 13+7 minutes by robot-assisted echo; comparable (p=0.33) to traditional examination at 12+6 minutes. There was no learning curve effect on image acquisition efficiency; 42+10 vs 51+7 minutes (1:1; n=16:15; p=0.007). Conclusion Remotely-performed, robot-assisted echo using a commercially-available ultrasound platform performed by briefly-trained, cardiac sonographers is feasible, with reasonable efficiency even though less efficient in image acquisition compared to traditional echocardiograms. Further investigation of performance as a remote service is warranted in larger, clinically-indicated populations.