2D PISA Limitations

Previous clinical and in vitro investigations have demonstrated that the 2D PISA method may significantly underestimate regurgitant flow volume. A study of patients with functional MR demonstrated that the 2D PISA method underestimated EROA by as much as 24%. For MR quantification, the 2D PISA method has 2 potential sources of significant error. The first is the geometric assumption of a hemispheric isovelocity surface, which is often not correct. The magnitude of this error varies according to both the severity and mechanism of MR. In organic MR where the regurgitant orifice can be assumed to be roughly circular, the 2D PISA method may have merit. On the other hand, in patients with functional MR the geometric assumptions of the 2D PISA method are often not valid. Recently, computational fluid dynamics modeling has clearly demonstrated the importance of this relationship between regurgitant orifice shape, converging isovelocity geometry, and the accuracy of derived flow measures. A volumetric depiction of an isovelocity shape with a direct measurement of its surface area (at a known velocity) should be the solution to this variable error. The second common error (albeit of a smaller magnitude) is the determination of the radius of the “hemispheric” isovelocity. Measuring the radius from the isovelocity shell to the regurgitant orifice can be technically challenging and difficult to reproduce. This limitation is important because this radial measure (r) is squared to derive the isovelocity surface area and it is likely that as the size of the flow convergence zone increases, so does the magnitude of error in determination of flow convergence radius.

3D—Adding Volume and Value

In general 3D PISA methods are based on computer algorithms that perform automated volume element (voxel) segmentation of the isovelocity surface area as opposed to 2D methods which depend on manual picture element (pixel) recognition. Recently developed automated 3D methods add another layer of software programming to differentiate isovelocity surface voxels from background voxels so that little human input is required for the measurement of the 3D PISA. Since these techniques require no geometric assumption they could be applied in patients with organic or functional MR with similar accuracy expected. By directly measuring the isovelocity surface area, irrespective of its shape, the 3D PISA method should provide a solution to the principal limitations of the 2D method. A practical advantage of the 3D PISA method is that the lower boundary of the flow convergence zone (the mitral leaflet) is relatively easy to identify by manual trace or by an automated software algorithm. This is in stark contrast to the 2D PISA method where the user must define the center of the regurgitant orifice to permit accurate measurement of the PISA radius. Since the leaflet surface is easier to identify than the center of the regurgitant orifice, this important distinction may lead to improved accuracy of the 3D color Doppler technique. Previous studies have reported accurate applications of 3D PISA for quantifying MR severity. The limitation of these techniques was largely just the time consuming practice of repeated manual identification of the PISA zone from the 3D color Doppler data.

In this issue of the Journal, de Agustin and colleagues report the use of single beat 3D color Doppler to directly measure the PISA in 33 patients with chronic MR. To minimize the potential effect of low temporal resolution of 3D color Doppler imaging they describe a methodology using single-beat non-stitched real-time color Doppler volumes acquired for each of 5 consecutive beats. The largest single convergence zone from 1 of the 5 cycles was used for 3D PISA analysis. After manual identification of the flow convergence region and flow direction, an automated software program performed the steps of 3D PISA contour identification and area calculation. The accuracy of this novel PISA determination was assessed by comparing the PISA-derived regurgitant volume (Rvol) and EROA to reference standard measures of regurgitant flow by Doppler quantification of stroke volume, and by direct planimetry of the EROA using 3D color Doppler measurement of the vena contracta area (VCA). The study population included 25 patients with degenerative MR and 6 patients with atrial fibrillation. No patients were excluded because of poor image quality; however 5 patients were excluded because of more than one flow convergence region.

Using the classic 2D hemispheric-PISA method, the mean derived EROA was significantly smaller than that obtained by the reference methods (0.16cm ² smaller than 3D VCA; 0.15cm ² smaller than Doppler stroke volume). In contrast, 3D PISA-derived EROA was only slightly smaller than the reference methods (0.03cm ² smaller than 3D VCA; 0.02cm ² smaller than Doppler stroke volume). In 14 patients with severe MR (defined by direct 3D color Doppler measurement of an EROA ≥0.4cm ² ), this difference between 2D and 3D PISA assessment appeared important. Six patients were classified with less than severe MR by 2D PISA method, whereas only 1 patient was classified as less than severe MR by 3D PISA method. In addition, the authors demonstrated very good intra- and inter-observer agreement of 3D PISA measurements.

3D—Adding Volume and Value

In general 3D PISA methods are based on computer algorithms that perform automated volume element (voxel) segmentation of the isovelocity surface area as opposed to 2D methods which depend on manual picture element (pixel) recognition. Recently developed automated 3D methods add another layer of software programming to differentiate isovelocity surface voxels from background voxels so that little human input is required for the measurement of the 3D PISA. Since these techniques require no geometric assumption they could be applied in patients with organic or functional MR with similar accuracy expected. By directly measuring the isovelocity surface area, irrespective of its shape, the 3D PISA method should provide a solution to the principal limitations of the 2D method. A practical advantage of the 3D PISA method is that the lower boundary of the flow convergence zone (the mitral leaflet) is relatively easy to identify by manual trace or by an automated software algorithm. This is in stark contrast to the 2D PISA method where the user must define the center of the regurgitant orifice to permit accurate measurement of the PISA radius. Since the leaflet surface is easier to identify than the center of the regurgitant orifice, this important distinction may lead to improved accuracy of the 3D color Doppler technique. Previous studies have reported accurate applications of 3D PISA for quantifying MR severity. The limitation of these techniques was largely just the time consuming practice of repeated manual identification of the PISA zone from the 3D color Doppler data.

In this issue of the Journal, de Agustin and colleagues report the use of single beat 3D color Doppler to directly measure the PISA in 33 patients with chronic MR. To minimize the potential effect of low temporal resolution of 3D color Doppler imaging they describe a methodology using single-beat non-stitched real-time color Doppler volumes acquired for each of 5 consecutive beats. The largest single convergence zone from 1 of the 5 cycles was used for 3D PISA analysis. After manual identification of the flow convergence region and flow direction, an automated software program performed the steps of 3D PISA contour identification and area calculation. The accuracy of this novel PISA determination was assessed by comparing the PISA-derived regurgitant volume (Rvol) and EROA to reference standard measures of regurgitant flow by Doppler quantification of stroke volume, and by direct planimetry of the EROA using 3D color Doppler measurement of the vena contracta area (VCA). The study population included 25 patients with degenerative MR and 6 patients with atrial fibrillation. No patients were excluded because of poor image quality; however 5 patients were excluded because of more than one flow convergence region.

Using the classic 2D hemispheric-PISA method, the mean derived EROA was significantly smaller than that obtained by the reference methods (0.16cm ² smaller than 3D VCA; 0.15cm ² smaller than Doppler stroke volume). In contrast, 3D PISA-derived EROA was only slightly smaller than the reference methods (0.03cm ² smaller than 3D VCA; 0.02cm ² smaller than Doppler stroke volume). In 14 patients with severe MR (defined by direct 3D color Doppler measurement of an EROA ≥0.4cm ² ), this difference between 2D and 3D PISA assessment appeared important. Six patients were classified with less than severe MR by 2D PISA method, whereas only 1 patient was classified as less than severe MR by 3D PISA method. In addition, the authors demonstrated very good intra- and inter-observer agreement of 3D PISA measurements.