Table 4.

Studies using XR for diagnostics including measurement (excluding case reports or series with fewer than 5 patients)

First author	Year	Cohort	Lesion(s) / Procedure	Pt no.	Control	XR type	Modality	Study description	Results
Van den Bosch et al.[49]	2005	NS	MV prolapse, stenosis, HCM with SAM of MV, LAVV	6	N	VR	3DE	Retrospective/proof of concept 10 observers reviewed 3DE images in VR and identify normal and pathological MVs	All correctly identified normal and pathological mitral valves.
Bol Raap et al.[50]	2007	Paed.	TV post-VSD closure	5	N	VR	3DE	Retrospective Observers reviewed 12 3DE datasets in VR for assessment of post-op TV function compared to 2DE	In 3 patients - VR analysis identified restriction of septal leaflet not appreciated on 2DE
Xue et al.[56]	2010	Paed.	ASD, VSD, TOF, DORV	40	Y	VR	3DE	Retrospective 3 observers reviewed datasets as 2DE, and as 3DE in VR. Asked to determine if abnormal intracardiac anatomy present. Compared to intra-op findings.	Diagnostic accuracy of VR > 92% Significantly higher accuracy in VR for TOF and DORV. ROC curve for VR closer to the optimal performance point than 2DE (AUC 0.96 in VR vs. 0.92 2DE)
Chan et al.[59]	2013	Paed.	PA + MAPCAs	9	N	MR	CT	Retrospective Participant identified MAPCA anatomy in standard 2D platform and in MR. Comparison made to gold standard (angiography).	Sensitivity, specificity, accuracy of standard CT review: 81%, 93%, 91% respectively; MR review 90%, 91% and 91%. Mean interpretation time shorter with MR: 13 ± 4 min vs. 22 ± 7 min (p = 0.0004)
Beitnes et al.[51]	2015	Adult	Degenerative MV disease	40	N	VR	3DE	Retrospective 2 observers assessed mitral valve segments in VR and compared to findings at intra-op inspection (35) or 3D TOE (5) as gold-standard.	Diagnosis sensitivity/specificity was 87/99% and accuracy/precision was 96/95% in VR Inter-observer agreement very good (Cohen’s Kappa 0.95)
Bruckheimer et al.[57]	2016	Adult /Paed.	ASD, percutaneous PVI, Glenn shunt	8	N	Holo-gram	3DE, RA	Proof of concept Assessment of assessing anatomy using 3D hologram in catheter lab. 4 observers asked to identify anatomical landmarks.	All anatomical landmarks identified by all participants. No adverse events reported.
Ballocca et al.[52]	2019	Adult	Degenerative MV disease	40	Y	MR	3DE	Comparison of measurements on normal and pathological mitral valves using standard 3DE platform and MR platform	Poor agreement between systems for abnormal MVs in annular area (ICC 0.58) & circumference (0.5). Good agreement in normal valves for annular area (0.95), circumference (0.91) and diameter (0.88–0.97). Poor agreement for scallop length in both normal and abnormal valves. Inter- & intra-observer agreement good (> 0.8) for all measurements except scallop length.
Wheeler et al.[53]	2019	Paed./ Phantom	Phantom, AV valves	5	Y	VR	3DE	Comparison of measurements in VR in phantom and patient data to 2 standard 3DE platforms	VR system more accurate in phantom and clinical measurements (lower mean difference) Precision higher in standard software (lower standard deviation)
Narang et al.[54]	2020	NS	Degenerative MV disease and MVR	30	N	VR	3DE, CT	Comparison of measurements in VR and conventional software Diagnostic quality of the VR models assessed	Measurement variability reduced in VR (20.1% standard vs. 12.2% in VR for 3DE; 15.3–10.1% for CT) Reduced measurement time (mean 61s in standard, 42s in VR for 3DE; 37s standard, 23 s in VR) for CT

NS: not stated, Paed.: paediatric.

2DE: 2 dimensional echocardiography, 3DE: three-dimensional echocardiography, ASD: atrial septal defect, AUC: area under the curve, AVSD: atrioventricular septal defect, CHD: congenital heart disease, DORV: double outlet right ventricle, LAVV: left atrioventricular valve, MAPCA: major aortopulmonary collateral artery, MV: mitral valve, MVR: mitral valve replacement; RA: rotational angiography, ROC: receiver operator characteristics, PA: pulmonary atresia, PVI: pulmonary valve implantation, TOF: tetralogy of Fallot, TV: tricuspid valve, VSD: ventricular septal defect.