Mensurações radiográficas computadorizadas da silhueta cardíaca em cães da raça Poodle portadores de insuficiência valvar mitral

Abstract

BANON, G. P. R. Computed-aid radiographic measurements of cardiac silhouette in Poodle dogs with mitral valve insufficiency. [Mensuracoes radiograficas computadorizadas da silhueta cardiaca em caes da raca Poodle portadores de insuficiencia valvar mitral]. 2012. 271 p. Tese (Doutorado em Ciencias) – Faculdade de Medicina Veterinaria e Zootecnia, Universidade de Sao Paulo, Sao Paulo, 2012. Radiographic measurement methods, Vertebral Heart Size (VHS), angle between long axis of the heart and sternum (angle) and normalized cardiac area (NCA), were evaluated in clinical trials of Poodles with mitral valve insufficiency (MVI). Seventy four sets of thoracic radiographs of Poodles were allotted to two groups: (I) normal cardiopulmonary structures (n = 18) and (II) from patients with proven cardiac disease (n = 56), confirmed by echocardiography to be caused by MVI. According to ecocardiographic classification, group II was subdivided in four subgroups: (IIa) mild (n = 4), (IIb) moderate (n = 7), (IIc) severe (n = 28) and (IId) absent (n = 17) left atrial enlargement (LAE). Measurements were performed by one experienced veterinary radiologist (A) and one trainee (B) in right lateral view. The measurements of A were used as reference. For VHS, there were significant differences between group I and groups II (P < 0.001), IIb (P < 0.05) and IIc (P < 0.001). The optimal threshold value mean was 9.85v and the median accuracy of 78.34%. There were significant differences in the measurements between A and B in groups I (P < 0.05) and II (P < 0.001). For angle, there were significant differences between group I and groups II (P < 0.05) and IIc (P < 0.05). The optimal threshold value mean was 60.57 and the median accuracy of 77.57%. There were significant differences in the measurements between A and B in groups I (P < 0.001) and II (P < 0.001). For NCA, there were significant differences between group I and groups II (P < 0.001), IIb (P < 0.05), IIc (P < 0.001) and IId (P < 0.05). The optimal threshold value mean was 18.81v and the median accuracy of 80.42%. There were significant differences in the measurements between A and B in groups I (P < 0.05) and II (P < 0.001). The VHS’ performance in clinical trials of Poodles with MVI was regular. It mainly identified generalized cardiac enlargement. The angle’s performance in clinical trials was low. It only recognized generalized cardiac enlargement. The NCA’s performance in clinical trials was good. It identified specific LAE and early changes in heart shape. To determine optimal threshold, sensitivity was prioritized, to the detriment of specificity. The reproducibility seemed to be related to the level of experience.