Ultrasonographic quantification of the endometrium echogenicity and heterogeneity in canine physiological and pathological conditions using computer-assisted analysis

The aims of this study were: To ultrasonographically describe and compare endometrial echogenicity and heterogeneity using digital analysis in normal and bitches suffering from pyometra, cystic endometrial hyperplasia (CEH) and endometritis; and to evaluate the effect of clinical, bacteriological and histopathological uterine parameters on endometrial echogenicity and heterogeneity. Forty-one post pubertal intact bitches were included. According to clinical, ultrasonographic, anatomopathological and histopathological uterine evaluation, the animals were classified as: Pyometra (PYO; n=6); CEH (n=8); Endometritis (END; n=13); Normal group (NG; n=14). Endometrial images were analyzed with ImageJ software to obtain echogenicity and heterogeneity, represented as the mean gray value (MGV) and standard deviation of gray (SDG), respectively. The effect of the group, clinical, bacteriological, ultrasonographic and histological variables on MGV and SDG were analyzed by a generalized linear model. PYO exhibited higher MGV (P<0.01) and SDG (P<0.01) than the other groups. No differences were found among CEH, END and NG for both parameters (P>0.1). Body weight decreased MGV (P<0.01), while increasing degrees of inflammatory reaction (P<0.01), edema (P<0.01), hemorrhages (P<0.01) and vascular congestion (P<0.01) were associated with higher MGV. Inflammatory reaction (P<0.01) and ulceration (P<0.01) increased SDG. Ultrasonographic images evaluated using computer assisted image analysis were useful to differentiate pyometra from other uterine conditions in dogs. However, this technique could not differentiate among CEH, END and NG. Uterine echogenicity and echotexture, which clearly represent the different histopathological patterns, contribute to the diagnosis of the definite diagnosis of some canine uterine diseases.