• canine.
• right atrial size
• right pulmonary artery
• three-dimensional echocardiography
• tricuspid regurgitation

The echocardiographic evaluation of right heart and the pulmonary artery (PA) size plays a role in the noninvasive detection and characterization of pulmonary hypertension (PH) both in humans and in dogs. Nowadays, several studies have described reference intervals (RIs) for right ventricular size and function in healthy dogs and their applicability in PH. Differently, only one study described the RIs of the right atrium (RA) and the main PA (MPA) size in a small number of healthy dogs, despite their routinely echocardiographic evaluation in dogs with PH. Furthermore, no studies aimed to describe the echocardiographic RIs of PA branches size in dogs.
Based on this, the present research project was divided into three clinical studies. The aim of the first study was to describe the body weight (BW)-prediction intervals (PIs) of the MPA and the right PA (RPA) size in a sample composed by at least 120 dogs. The aim of the second study was to evaluated the diagnostic accuracy of MPA and RPA size for the detection of PH in dogs. Finally, the aim of the third study was to describe BW-PIs of the RA diameter (RAD), RA area (RAA) and two and three-dimensional (2D and 3D) RA volume (RAV) in a sample composed by at least 120 dogs.
Healthy dogs (control group) and dogs with naturally occurring PH were prospectively included in the studies. Dogs with PH were classified according to the precapillary or postcapillary PH origin and on the basis of the tricuspid regurgitation pressure gradient as mild (36-50 mmHg), moderate (51-75 mmHg), and severe (>75 mmHg). In the first two studies, the MPA to aorta ratio (MPA/Ao), the RPA maximum (RPAmax) and minimum (RPAmin) to the aortic annulus ratio (RPAmax/Aod and RPAmin/Aod, respectively) and right pulmonary artery distensibility (RPAD) index were measured. In the third study, the RAD, the RAA and 2D and 3D RAV were measured.
In the first study, among a total of 269 healthy dogs, MPA, RPAmax, RPAmin showed a positive linear relationship with BW after logarithmic transformation (P<0.0001). According to allometric scales, the PI for the MPA normalized for BW (MPA_N) was between 5.50 and 8.07, the PI for the RPAmax normalized for BW (RPAmax_N) was between 3.23 and 5.62, whilst the PI for the RPAmin normalized for BW (RPAmin_N) was between 1.62 and 3.30. The median MPA/AO was 0.92 (RI, 0.78-1.01), the median RPAmax/Aod was 0.70 (RI, 0.53-0.98) and the median RPAmin/Aod was 0.40 (RI, 0.29-0.61).
In the second study, a total of 404 dogs, of which 136 composed the control group and 268 had PH, were included. The RPAmin_N, the RPAmin/Aod and the RPADi showed similar diagnostic accuracy for the detection of PH (AUC=0.971, AUC=0.953 and AUC=0.975, respectively), but higher than MPA/Ao (AUC=0.926; P<0.05), MPA_N (AUC=0.880; P<0.001), RPAmax_N (AUC=0.814; P<0.001) and RPAmax/Aod (AUC=0.803; P<0.001). No differences were found between precapillary and postcapillary PH dogs considering all the PA variables, except for RPAmax_N and RPAmax/Aod which were higher among precapillary PH dogs.
In the third study, among a total of 195 healthy dogs, RAD, RAA , 2D RAV and 3D RAV showed a positive linear relationship with BW after logarithmic transformation (P<0.0001). According to the allometric scales, the PI of RADn (from right parasternal long axis view; RPLA) was between 5.12 and 9.04, the PI of RADn (from left apical four chamber view; LA4C) was between 5.29 and 8.79, the PI of RAAn was between 0.26 and 0.58, the PI of RAVn 2D (Simpson method of discs; SMOD) and RAVn 2D (area-length method; A-L) was between 0.09 and 0.32, whilst the PI of RAV 3D was between 0.18 and 0.56. Reference intervals for RAD(RPLA)/Ao, RAD(LA4C)/Ao, RAAi, RAV 2D(SMOD)/BW, RAV 2D(A-L)/BW and RAV 3D/BW were also described.
In conclusion, we described the reference values of the MPA and RPA dimensions in a large sample of healthy dogs. In dogs with PH, although MPA/Ao showed an excellent sensitivity and specificity for the detection of PH, the RPAmin proved to have a higher diagnostic accuracy and thus could represent a new useful parameter for the non-invasive detection of PH in dogs. Finally, the 2D and 3D echocardiographic reference values of RA size were described among a large sample of healthy dogs and those are made available for clinical use.