Quantitative Real-Time Three-Dimensional Echocardiography Provides New Insight into the Mechanisms of Mitral Valve Regurgitation Post-Repair of Atrioventricular Septal Defect


Mechanisms of mitral valve regurgitation after atrioventricular septal defect repair are unclear.


To gain further insight into mitral valve regurgitation, real-time three-dimensional echocardiography was performed in 53 patients after atrioventricular septal defect repair (30 partial and 23 complete) and 40 controls. Mitral valve {x, y, z} coordinates from the annulus, leaflet surface, papillary muscle, and chordal attachments were recorded. Vena contracta area of the regurgitant jet(s) and volume of leaflet prolapse and tethering were measured.


Twenty-three patients had mild (group 1) and 30 moderate (group 2) mitral valve regurgitation. Patients in both groups 1 and 2 had more circular annuli than controls. Annular area was greater in group 2 than in group 1 and controls ( P < .01). Group 2 had more frequent segmental prolapse in the superior–mural leaflet segment. The anterolateral papillary muscle was more laterally displaced in group 2 than in controls and group 1 at end-diastole ( P = .01 and P = .05) and formed a more acute angle with the mitral valve annulus than in controls or group 1 ( P = .01).


In patients with atrioventricular septal defects, significant mitral valve regurgitation is associated with leaflet prolapse, larger annular area, and lateral papillary muscle displacement.

Although current surgical strategies have led to an improvement in survival after atrioventricular septal defect (AVSD) repair, progressive mitral valve regurgitation represents an ongoing challenge. Several studies have addressed risk factors for the development of postoperative mitral valve regurgitation, but precise mechanisms remain unclear. Anomalies of the papillary muscles, the mitral valve, and an open cleft have been believed to accentuate the degree of regurgitation. Despite this, the main focus has been on follow-up of surgical results and qualitative assessment.

The term left atrioventricular valve is commonly used in the setting of an AVSD because the left-sided inflow valve in no way represents a normal mitral valve. The normal mitral valve is bileaflet and consists of a triangular anterior and a broader posterior leaflet. In “all hearts” with AVSDs, the more posterior leaflet is much narrower and smaller and is called the mural leaflet (ML). The remaining two other left-sided leaflets are referred to as the superior bridging leaflet (SBL) and inferior bridging leaflet (IBL), with a cleft separating them as they bridge the interventricular septum ( Figure 1 ). For the sake of clarity and to avoid confusion, the left-sided atrioventricular valve in patients with AVSDs is referred to here as a “mitral valve.” The papillary muscle location in an AVSD is different from the normal mitral valve, with counterclockwise rotation of the posteromedial papillary muscle when viewed in short axis from the apex, a feature linked to the narrower ML ( Figure 1 ).

Figure 1

Schematic diagram of the comparison between a normal mitral valve ( top ) and the left-sided inflow valve after repair of an AVSD. Note the smaller ML in the diagram of the AVSD, as well as the different position of the papillary muscles. AO , Aorta.

Real-time three-dimensional (RT3D) echocardiography permits a qualitative and quantitative evaluation of atrioventricular valve function in both the acquired and congenital populations. We have previously applied RT3D echocardiography to evaluate the mitral valve in AVSD but in a more qualitative fashion. The purpose of this study was to apply quantitative RT3D echocardiography to gain further insight into the mechanisms of mitral valve regurgitation in patients after AVSD repair.


Patient Characteristics

Between January 2006 and November 2007, we prospectively acquired two-dimensional (2D) transthoracic echocardiographic and transthoracic RT3D echocardiographic images from patients with AVSDs who had undergone complete repair and were >3 years of age. Patients were sequentially recruited from the cardiac outpatient clinic at the Stollery Children’s Hospital. The inclusion criteria were ability to cooperate for the study, technically adequate RT3D echocardiographic images of the “mitral valve” and its subvalvar apparatus, laminar flow across the left ventricular outflow tract, and sinus rhythm. Exclusion criteria were any form of left ventricular outflow tract obstruction, unbalanced AVSD, and right ventricular hypertension (right ventricular pressure greater than half systemic). We collected mitral valve data from normal subjects of similar age for comparison. The study was approved by the University of Alberta ethics committee.

Two-Dimensional Echocardiography

Two-dimensional transthoracic echocardiography was performed using an iE33 system (Philips Medical Systems, Andover, MA). Left ventricular global function was assessed by shortening fraction and corrected velocity of circumferential fiber shortening. Left ventricular volume was measured using the biplane Simpson method. Mitral valve inflow mean Doppler velocity was recorded, as was the right ventricular pressure from tricuspid valve regurgitation or interventricular septal shape. The degree of mitral valve regurgitation was assessed by vena contracta (VC) width measured at the orifice in the parasternal long-axis view and characterized on a 0 to 3+ grading scale, defined as follows: 0 = no mitral valve regurgitation, 1 = mild (VC width < 0.3 cm), 2 = moderate (VC width, 0.3–0.69 cm), and 3 = severe (VC width > 0.7 cm), per the published guidelines of the American Society of Echocardiography. In a similar fashion to our previous study, we combined grades 0 and 1 to represent mild (group 1) and grades 2 and 3 (group 2) to represent moderate mitral valve regurgitation.

Acquisition of RT3D Data Sets

Images were obtained using a RT3D ultrasound system (iE33) with a transthoracic matrix X3-1 or X7-2 probe. Full-volume images of the mitral valve and papillary muscles at end-expiration were obtained from a position as close to the apex as possible, including a color Doppler data set from the same position to permit precise anatomic location of any regurgitant jets. During acquisition, the high–frame rate mode was selected to optimize temporal resolution, and high density was selected to optimize image spatial resolution. The apical position was chosen for image acquisition, because this provided an image of the mitral valve in the axial plane, hence again optimizing leaflet resolution.

Mitral Valve Three-Dimensional Analysis

Mitral valve assessment software (TomTec Imaging Systems, Munich, Germany) was used to analyze the RT3D data sets. Data points were recorded at end-diastole and midsystole. Exploration of the RT3D data set was performed using multiplanar 2D views. This enabled markers to be placed to delineate the annular hinge points ( Figures 2 A and 2 B). We then obtained 15 vertical radial and equally spaced 2D planes (24°) passing through the center of the mitral valve ( Figure 2 C). Our reference within each RT3D data set was marked between the aortic and mitral valves. On these 2D planes at midsystole, we placed nine makers on the mitral valve leaflets to delineate its shape in three-dimensional (3D) space for subsequent reconstruction ( Figures 2 A and 2 C). Using the multiplanar navigation tool, we located the tip of the anterolateral and posteromedial papillary muscle ( Figures 2 A, 2 B, and 2 D) and marked the subvalvar apparatus position with five points ( Figures 3 A, 3 B, and 3 C): (1) the base of the papillary muscle, (2) the tip of the papillary muscle, (3) the chordal attachments on the leaflet that were closest to the tip of the papillary muscle, (4) the annular point closest to the papillary muscle tip, and (5) the annular point furthest from the papillary muscle tip. All 3D marked locations were converted into spatial coordinates {x, y, z} and exported to MATLAB (The MathWorks Inc, Natick, MA) to extract measures of annular, leaflet, and papillary muscle geometry as detailed below.

Figure 2

(A,B) Annular coordinates as indicated by the black dots in four-chamber and two-chamber views. The white dot represents the tip of the anterior papillary muscle. The red dots are the coordinates of the leaflet seen in one individual plane. (D) Three-dimensional reconstruction of the same case. (C) En face view of the reconstructed annulus as indicated by the green outer line , with the inner line indicating the line of leaflet coaptation. The green dots represent some of the 15 planes at 24° to one another. The red dots represent the coordinates from the leaflets in one of these 15 planes. APM , Anterolateral papillary muscle; LA , left atrium; LV , left ventricle; PPM , posteromedial papillary muscle.

Figure 3

(A) Posteromedial papillary muscle (PPM) ( red ) and anterolateral papillary muscle (APM) ( green ) measurements. The {x, y, z} coordinates of the chords, papillary muscles, and their attachment to the leaflets were identified within the RT3D data sets ( Figure 2 ). The subvalvar apparatus variables derived are illustrated in the diagram. (B) Reconstructed coordinates obtained in (a) in a 3D model seen in different views: above, from the side, and in a 3D model. (C) Two volume-rendered images showing examples of papillary muscle position. ( Left ) A patient in whom the APM angle with the annulus was normal (72°) without lateral displacement. ( Right ) A patient in whom the APM had a more acute angle with the mitral valve annulus (64°) and was displaced laterally. BC , Base to center of left ventricular distance; CL , chordal length; LV , left ventricular length in diastole; PMH , papillary muscle height; PML , papillary muscle length; TC , tip of papillary muscle to center of left ventricular distance.

Measurements of 3D Echocardiographic Data

Annular Area

Annular area was calculated as the sum of 30 triangles composed of two adjacent annular points and a mathematically calculated gravity center of the annulus. The annular area was indexed to body surface area (BSA).

Anterior-Posterior (AP)/Commissure Ratio

The AP diameter was automatically measured as the diameter which passed the point closest to the aortic valve. The commissure-commissure (CC) diameter was measured as the diameter that was perpendicular to the AP diameter, passing the center of the annulus. The AP/CC ratio was calculated as AP diameter/CC diameter. The percentage change of the AP/CC ratio at midsystole was calculated as (AP/CC ratio at end-diastole − AP/CC ratio at midsystole)/(AP/CC ratio at end-diastole) × 100.

Mitral Valve Annular Height

The anatomic location of the high (toward the left atrium) and low (toward the left ventricle) points of the plane of least squares fit was plotted as a function of annular position from the reconstruction of the annulus. In all, two positive peaks were observed at the anterior and posterior aspects of the annulus, with two negative peaks at the lateral and septal aspects. The height of all four points from the least squares fit plane was measured. The degree of nonplanarity was measured by the distance between high and low points calculated as follows: mean height = (height at anterior + height at posterior − height at septal − height at lateral)/2. These measurements were adjusted for left ventricular length obtained from the 2D four-chamber image.

Bending Angle of the Mitral Valve Annulus

A best fit (nonnegative least squares) plane was fit to the annular points in each of the two sections, from which the bending angle was determined as the angle between the normal lines for each plane ( Figure 4 ).

Figure 4

Image demonstrating the saddle shape of the left atrioventricular annulus and the calculation of the bending angle by a best fit ( nonnegative least squares ) plane that was fit to the annular points in each of the two sections. LA , Left atrium; LV , left ventricle.

Measurements of Prolapse and Tethering Volume

For quantitative analysis, the mitral valve annular plane was determined on the basis of a saddlelike annular shape in each subject ( Figure 5 ). A volume was calculated at mid systole, which represented the volume enclosed between the annular plane and the mitral valve leaflets. The measured total tethered and prolapse volume was indexed to BSA. To evaluate regional prolapse and tethering, the circumference of the mitral valve was divided into four quadrants using the aorta as the reference point ( Figure 6 ).

Figure 5

( Top left ) A patient from group 1 with mild regurgitation and ( top middle ) a patient from group 2 with prolapse in the region of the superior leaflet and ML. ( Top right ) Tethering in the inferior leaflet and ML region from a patient in group 2. The tethering in this example is most likely related to the fact that the IBL is often attached by short chordae as it bridges the interventricular septum. The red area represents the area above the least fit annular plane (closest to the left atrium ) and the blue below , closest to the apex of the left ventricle. The black arrow represents the septal region of the left atrioventricular valve. ANT , Anterior; POST , posterior.

Figure 6

RT3D cropped images (from QLAB) of an AVSD ( right ) and a mitral valve ( left ) seen from a left atrial view to demonstrate the shape of the annulus and the division of the leaflets used to assess regional prolapse and tethering. The top two images were taken during late diastole and thus do not demonstrate all of the leaflets but best demonstrate the commissure regions shape of the annulus and the residual cleft (indicated by the black arrow in the top right panel). The bottom two images were taken in systole and show the leaflets. The black dots indicate where the papillary muscles are located in the left ventricle. The circumference of the mitral valve was divided into four quadrants, and regional prolapse or tethering was evaluated in each of the four quadrants. AL , Anterior leaflet; AO , aorta; PL , posterior leaflet; RAVV, right atrioventricular valve; TV , tricuspid valve.

Measurements of Papillary Muscle Geometry

The length of the papillary muscles and their supporting chordae were measured. As well, the lateral distance of the base and the tip of the papillary muscle from the central axial line of the left ventricle were measured at end-diastole and midsystole ( Figure 3 A). The vertical distance of the anterolateral and posteromedial papillary muscle was measured from the annulus to their base. All chord and papillary muscle measurements were divided by left ventricular length at end-diastole to correct for left ventricular size. The angle between the annular plane and the line connecting the tip of the papillary muscles and their supporting chordae to the leaflet was measured ( Figures 3 A, 3B, and 3C).

Ratio of Each Leaflet to Mitral Valve Area

From an en face view, the program demonstrated the annular line, the cleft line, and the coaptation line between the ML and SBL and IBL. We manually measured the area of each leaflet. The leaflet/mitral valve area ratio was calculated as leaflet area/mitral valve leaflet area. The program did not permit the inclusion of leaflet undulations in the measurements and thus represents the area of a flat surface.

Mitral Valve Residual Cleft Measurement

We used the multiplanar reformatting mode in QLAB (Philips Medical Systems) during ventricular diastasis to navigate the RT3D data sets. Two reference landmarks were identified: the green plane was parallel to the cleft suture line ( Figure 7 C), with the red plane running along the interventricular septum on the right ventricular endocardial surface ( Figures 7 A and 7B). A third plane at right angles to the green plane was used to locate and mark the beginning of the residual cleft, defined as the point at which the edges of the SBL-IBL begin to separate ( Figure 7 C, yellow line). A second yellow line was placed at the most distal margins of the residual cleft ( Figure 7 C). The distance between leaflet separation and the most distal margin of the cleft was measured, and the ratio with the sutured cleft length was calculated ( Figures 7 C and 7 D).

Figure 7

These images demonstrate how the residual cleft measurements were made. (A) The red plane running parallel along the right side of the interventricular septum (IVS). The surface of the right side of the IVS is seen in (B) . (C) En face view of the left atrioventricular valve seen in the blue plane and shows the green plane running through the IVS at the level of the sutured and unsutured component of the cleft. The two yellow lines represent how the measurements were made of the sutured (D1) and unsutured (D2) cleft. (D) RT3D cropped image in the same orientation as (C) . Note the small residual cleft between the SBL and IBL. AO , Aorta; D1 , distance between IVS and apex of cleft; D2 , distance from apex of cleft to free edge; LA , left atrium; LAVV, left atrioventricular valve; LV , left ventricle; MV , mitral valve; RAVV, right atrioventricular valve; RV , right ventricle; TV , tricuspid valve.

Measurements of VC Area by 3D Echocardiography

Systematic cropping of the RT3D color data set was performed to measure VC area using CardioView software (TomTec Imaging Systems). When more than one VC was present, areas were summed.

Statistical Analysis

Continuous variables are presented as median (range) or as mean ± SD as appropriate. Comparisons of continuous variables between groups 1 and 2 and controls were determined using one-way analysis of variance with Scheffé’s F test as the post hoc comparison. Proportions were compared using χ 2 or Fisher’s exact tests as appropriate. Stepwise linear regression was performed to select the variables most related to mitral valve 3D VC area. Statistical significance was set at P < .05. Interobserver variability testing was performed for novel 3D variables on 10 randomly selected patients by two blinded observers (N.S.K. and T.C.). N.S.K. is a junior staff echocardiographer with prior training in RT3D echocardiography, while T.C. is a second-year research fellow in general and RT3D echocardiography. Because the measurements are complex, both observers were trained in the analysis using the TomTec system and then practiced on different data sets that were not used in the analysis. We chose to determine the variability of the reconstructed data from MATLAB, which were derived from the various x, y, and z coordinates obtained by the two observers, rather than the raw data points. Bland-Altman methods were used to assess the limits of agreement, and intraclass correlation coefficients were calculated in reliability testing. Analyses were performed using Stata version 9.2 (StataCorp LP, College Station, TX).


We studied 78 subjects with repaired AVSDs, of whom 53 met the eligibility criteria. Twenty-five were excluded because of inadequate RT3D data sets. Subject characteristics are summarized in Table 1 . All patients were in New York Heart Association functional class I. Using VC size, 23 patients had trivial or mild regurgitation (group 1), and 30 had moderate or greater mitral valve regurgitation (group 2). There was no significant difference of age at repair or age at echocardiography ( Table 1 ). Both VC width and VC area were larger in group 2 compared with group 1 ( P = .0001 for both; Table 2 ). There was a good linear correlation between the square of VC width by 2D echocardiography and 3D VC area ( R = 0.79, P = .01). There was no difference in left ventricular size, geometry, or function between controls or patients ( Table 2 ). In this population, the severity of regurgitation was not related to patient age.

Table 1

Subject characteristics

Variable Group 1: mild MVR
( n = 23)
Group 2: moderate MVR
( n = 30)
( n = 40)
Men 11 16 28
AVSD type
Partial 11 19
Complete 12 11
Down syndrome 6 9
Age at echocardiography (y) 13.7 (3.3–29.8) 12.7 (4.0–32.6) 12.8 (3.4–32.0)
Age at AVSD repair (years) 0.8 (0.1–5.8) 1.3 (0.1–17.4)

MVR , Mitral valve regurgitation.

Data are expressed as numbers or as median (range).

Table 2

Left ventricular function and vena contracta measurements

Variable Group 1: mild MVR Group 2: moderate MVR Controls P
FS% 40.1 ± 7.1 42.6 ± 11.8 44.7 ± 5.8 .12
VCFc (circ s-1) 1.53 ± 0.43 1.79 ± 0.53 1.75 ± 0.40 .11
EDV/BSA (mL/m 2 ) 61.3 ± 12.5 65.7 ± 21.6 62.7 ± 11.0 .54
ESV/BSA (mL/m 2 ) 28.7 ± 6.7 29.6 ± 9.2 27.2 ± 5.9 .39
End-diastolic sphericity index 0.721 ± 0.477 0.676 ± 0.09 .31
End-systolic sphericity index 0.568 ± 0.09 0.583 ± 0.107 .70
VC width on 2D echocardiography (mm) 0.24 ± 0.02 0.56 ± 0.03 <.0001
VC area on 3D echocardiography (mm 2 ) 0.19 ± 0.02 0.69 ± 0.08 <.0001

BSA , Body surface area; EDV , end-diastolic volume; ESV , end-systolic volume; FS , fractional shortening; MVR , mitral valve regurgitation; VCFc , velocity circumferential shortening corrected for heart rate.

Data are expressed as mean ± SD.

The median mitral Doppler inflow gradient was 2.2 mm Hg (range, 0.7–8.4 mm Hg), with higher gradients among patients with greater mitral valve regurgitation. Tricuspid valve regurgitation velocity was available in 50 patients, with an estimated median right ventricular systolic pressure of 21.5 mm Hg (range, 13–36 mm Hg); the other three patients had normal interventricular septal shapes.

Three-Dimensional Echocardiographic Measurements

The median frame rate of RT3D data sets was 22 Hz (range, 17–55 Hz).

Annular Geometry

At end-diastole, mitral valve area adjusted by BSA was larger in groups 1 and 2 compared with controls ( Table 3 ). However, at midsystole, mitral valve area was significantly larger in group 2 than in group 1 and controls ( P < .01). On the basis of the AP/CC ratio, annuli in both groups 1 and 2 were more circular than in controls ( Table 3 ). At end-diastole, the septal minimal height was less than in controls for both groups 1 and 2. The remainder of the annular geometry in groups 1 and 2 was similar to controls ( Table 3 ).

Table 3

Three-dimensional echocardiography: annular geometry

Variable AVSD Controls P
Group 1: mild MVR Group 2: moderate MVR
MV area/BSA at ED (cm 2 /m 2 ) 568 ± 147 688 ± 196 510 ± 123 <.01 ; <.05
MV area/BSA at MS (cm 2 /m 2 ) 515 ± 143 634 ± 179 460 ± 90.5 <.01
AP/CC ratio at ED (%) 100.3 ± 13.3 104.5 ± 15.8 90.8 ± 12.2 <.01 ; <.05
AP/CC ratio at MS (%) 104.4 ± 14.0 103.1 ± 10.0 99.2 ± 8.0 .11
Anterior maximal height/LV length at ED 0.39 ± 0.19 0.36 ± 0.15 0.33 ± 0.10 .31
Lateral minimal height/LV length at ED −0.29 ± 0.12 −0.31 ± 0.17 −0.26 ± 0.11 .29
Posterior maximal height/LV length at ED 0.23 ± 0.38 0.35 ± 0.20 0.20 ± 0.10 <.05
Septal minimal height/LV length at ED −0.37 ± 0.15 −0.40 ± 0.16 −0.26 ± 0.09 <.01 §
Bending angle at ED (°) 133.8 ± 17.3 139.9 ± 18.4 137.2 ± 16.3 .45
Bending angle at MS (°) 139.3 ± 16.4 146.8 ± 16.8 140.9 ± 11.9 .13

BSA , Body surface area; ED , end-diastole; LV , left ventricular; MS , midsystole; MV , mitral valve; MVR , mitral valve regurgitation.

Data are expressed as mean ± SD.

Group 2 versus controls.

Group 1 versus controls.

Group 2 versus group 1 and controls.

§ Groups 1 and 2 versus controls.

Mitral Valve Prolapse

Overall prolapse volume indexed to BSA was greater in group 2 compared with controls ( P < .01), with less difference between group 1 and controls ( P < .05) ( Table 4 , Figure 4 ). When assessed on a regional basis, there was greater prolapse in segment 4 in group 2 compared with group 1, which represented the superior-ML components of the mitral valve ( P < .01, group 2 vs group 1; Figure 6 ).

Table 4

Three-dimensional echocardiography: prolapse and tethering

Variable AVSD Controls P
Group 1: mild MVR Group 2: moderate MVR
Total prolapse volume/BSA (mm 3 /m 2 ) 0.05 ± 0.08 0.23 ± 0.41 0.00 ± 0.02 <.01 ; <.05
Segment 1 prolapse volume/BSA (mm 3 /m 2 ) 0.01 ± 0.01 0.06 ± 0.12 0.00 ± 0.00 <.01 ; <.05
Segment 2 prolapse volume/BSA (mm 3 /m 2 ) 0.00 ± 0.01 0.03 ± 0.09 0.00 ± 0.00 .05
Segment 3 prolapse volume/BSA (mm 3 /m 2 ) 0.01 ± 0.03 0.03 ± 0.09 0.00 ± 0.01 .06
Segment 4 prolapse volume/BSA (mm 3 /m 2 ) 0.03 ± 0.05 0.10 ± 0.13 0.00 ± 0.01 <.01
Total tethered volume/BSA (mm 3 /m 2 ) 0.62 ± 0.45 0.50 ± 0.26 0.77 ± 0.40 <.05
Segment 1 tethered volume/BSA (mm 3 /m 2 ) 0.11 ± 0.14 0.10 ± 0.09 0.18 ± 0.12 <.05
Segment 2 tethered volume/BSA (mm 3 /m 2 ) 0.22 ± 0.19 0.18 ± 0.10 0.21 ± 0.13 .04
Segment 3 tethered volume/BSA (mm 3 /m 2 ) 0.20 ± 0.14 0.17 ± 0.12 0.22 ± 0.12 .21
Segment 4 tethered volume/BSA (mm 3 /m 2 ) 0.08 ± 0.08 0.05 ± 0.06 0.13 ± 0.08 <.01 §

BSA , Body surface area; MVR , mitral valve regurgitation.

Data are expressed as mean ± SD.

Group 2 versus controls.

Group 1 versus controls.

Group 2 versus group 1 and controls.

§ Groups 1 and 2 versus controls.

Mitral Valve Tethering

Total tethering of the mitral valve was slightly less in group 2 than in controls ( P < .05; Table 4 ), with segments 1 and 4 being less than segments 2 and 3.

Anterolateral and Posteromedial Chordal and Papillary Muscle Geometry

In controls, the anterolateral angle was 71° at end-diastole and midsystole, with the posteromedial angle being 72° ( Table 5 ). In those with AVSDs, the anterolateral papillary muscle angle was similar to controls in group 1 at end-diastole (71°; Figure 3 C). On the other hand, group 2 had significantly more acute angles of their anterolateral papillary muscles with the annuli at end-diastole and midsystole than group 1 and controls (64° and 67°, respectively; Table 5 , Figure 3 C). The posteromedial papillary muscle in groups 1 and 2 had a more acute angle with the annulus compared with controls irrespective of the degree of mitral valve regurgitation ( Table 5 ).

Jun 2, 2018 | Posted by in CARDIOLOGY | Comments Off on Quantitative Real-Time Three-Dimensional Echocardiography Provides New Insight into the Mechanisms of Mitral Valve Regurgitation Post-Repair of Atrioventricular Septal Defect

Full access? Get Clinical Tree

Get Clinical Tree app for offline access