Applicability of ARFI elastography in the evaluation of canine prostatic alterations detected by b-mode and Doppler Ultrasonography

The aim of this study was to assess the applicability of ARFI elastography for the complementary diagnosis of canine prostatic alterations. Twenty-two healthy dogs and 45 with prostatic alterations were diagnosed by the association of general and specific physical exams. Complete blood count, B-mode and Doppler ultrasound were performed. For the elastography study, tissue homogeneity and deformity were verified using the elastogram, and the shear velocities of the left and right lobes of all prostates were obtained. The change in tissue homogeneity was observed in 62.2% of the animals, while the shear speed was significantly higher in dogs with prostatic alterations, with a cut-off point > 2.35m/s as an indication of change. All animals in both groups were non-deformable. It was concluded that ARFI elastography is capable of providing qualitative and quantitative results that assist in the diagnosis of canine prostatic alterations in a non-invasive way.


INTRODUCTION
Prostate diseases have a high casuistry in dogs, most frequently: benign prostatic hyperplasia, prostatitis, prostatic and paraprostatic cysts, abscesses and neoplasms (Souza et al., 2019). The Arq. Bras. Med. Vet. Zootec., v.72, n.6, p.2135-2140, 2020 invasive and with considerable risk of complications during the collection of material for analysis (Souza et al., 2019). Thus, non-invasive methods that provide information regarding the organ's morphology and hemodynamics, such as B-mode and Doppler ultrasound, are routinely used to assist the diagnosis and therapeutic choices, for allowing the detection of abnormalities in the prostate and the monitoring of the disease (Matton and Nyland, 2005;Cruzeiro et al., 2008;Alonge et al., 2017). However, despite being used on a large scale, these ultrasound techniques provide unspecific results (Souza et al., 2017).
Acoustic radiation force impulse elastography (ARFI) is an ultrasound technique that allows assessment of tissue elasticity in a safe and noninvasive way, providing quantitative and qualitative measures of the rigidity of various organs (Dudea et al., 2011;Gennisson et al., 2013). In veterinary medicine, the use of ARFI elastography is novel, and reference values for prostatic evaluation in young, adult and senior dogs have already been determined (Feliciano et al., 2015). In medicine, elastography has been used to assist in the diagnosis of prostatic neoplasms (D'anastasi et al., 2011;Dudea et al., 2011), with these tissues showing greater rigidity than the normal parenchyma and benign lesions (Ophir et al., 2002;Zaleska et al., 2014).
Taking into account that the definitive diagnosis of canine prostatic alterations is often not obtained and that a large share of the cases are presumptively diagnosed by the association of the history and the findings of physical and ultrasound examination, the use of other noninvasive methods that can increase the diagnostic accuracy is justified. Thus, the objective of this study was to verify the applicability of ARFI elastography in the diagnostic complementation of canine prostatic alterations verified by means of B-mode and Doppler ultrasonography.

MATERIALS AND METHODS
This study was approved by the Animal Ethic and Welfare Committee of the institution where the study took place, under protocol No. 023705/12. Sixty-seven dogs, whole, of different breeds, aged between one and 14 years and weighing between five and forty kilograms were included. Forty-five came from the institution's Veterinary Reproduction and Obstetrics Service and were already diagnosed with prostatic diseases, while 22 healthy dogs were given under free and informed consent by the tutors.
The animals' health was established based on the results of physical, specific exams (palpation of the prostate), blood count, B-mode and Doppler ultrasonography (color and pulsed wave). Animals were considered healthy if they did not have a history of prostatic changes or in parameters evaluated in physical examination or palpation, as well as any change in blood count. According to B-mode ultrasonography, healthy prostates were hypoechogenic compared to adjacent tissues, homogeneous and of regular shape (Matton and Nyland, 2005;Smith, 2008;Moxon et al., 2016) and dimensions, as already described in the literature by Atalan et al. (1999). According to Color Doppler, vascularization is normal when seen around the prostate or not identified. The spectral doppler obtained the values of peak systolic velocity, end diastolic velocity, and vascular resistivity index of the prostatic artery, taking as reference the values described by Freitas et al. (2015). Prostates that showed any change in echogenicity, echotexture, regularity, dimensions, vascularity and doppler velocimetric indices were classified abnormal.
After being classified as healthy or sick, the dogs were referred for evaluation with ARFI elastography. The animals were subjected to a previous 8-hour fast, so that there was no presence of fecal or gaseous content in the colonic portion so there would not interfere on the evaluation of the prostate. Bladder retention was necessary for at least two hours before examination. Sedation of the animals was not necessary. Wide trichotomy of the abdominal region and application of ultrasound conductive gel was done to facilitate the procedure and promote greater contact between the transducer and the animal's skin.
The exams were performed by a single experienced operator (15 years of experience), using the ACUSON S2000 ultrasound equipment (Siemens, Munich, Germany) and a 9.0mHz linear multifrequency transducer. Software for qualitative characterization and quantification ARFI (Virtual Touch Tissue Quantification, 2D-SWE technique) was used. Firstly, images of the prostate were taken in longitudinal sections to obtain a gray scale elastogram. Tissue homogeneity (homogeneous soft, homogeneous hard or heterogeneous hard) and deformity (deformable or non-deformable) were verified, based on the relative rigidity of the parenchyma, where lighter tones (white) corresponded to more elastic tissues (soft) and darker tones (black) represented more rigid tissues (hard).
Secondly, a quantitative assessment of tissue stiffness was performed by measuring the shear velocity (SWVm/s) of the prostatic parenchyma in a transversal orientation. The average was calculated for six samples from the right lobe and six from the left lobe, with depth 1.5 to 2.5cm (Feliciano et al., 2015).
The statistical analysis was performed with IBM SPSS Statistics 20 software (International Business Machines Corp, São Paulo, Brazil) and for all of the analyzes, a significance level of 95% (P≤0.05) was used. The association between elastography homogeneity and the presence of disease was verified by logistic regression and sensitivity, specificity and accuracy were calculated for the result obtained. Tissue deformity was demonstrated in percentages. Mann Whitney test was performed to check if there was an association between SWV of both lobes and the presence of prostatopathy, so that their cutoff point and predictive values (sensitivity, specificity and area under the curve) were obtained using the ROC curve.

RESULTS
The elastogram revealed that 100% of healthy prostates were characterized with homogeneous hard parenchyma, while only 35.5% (n=16) of sick dogs had such a classification. Only one sick prostate presented a soft homogeneous parenchyma (Table 1) and the other animals were classified as heterogeneous hard (62.2%). There was a significant association between elastographic heterogeneity and presence of disease (P<0.001), with a sensitivity of 64.44%, specificity of 100% and accuracy of 76.11%. All prostates in both groups were classified as nondeformable.
Quantitative elastography showed that the average shear velocity (SWV) obtained in the healthy group was 1.93±0.42m/s (left prostatic lobe) and 1.88±0.32m/s (right prostatic lobe), while the sick group (Table 1.) presented 2.24±0.88m/s (left lobe) and 2.35±0.78m/s (right lobe), with a significant difference between groups (P <0.001) (Figure 1). Thus, it is indicative of prostatic disease when the SWV of the prostate is greater than or equal to 2.35m/s, with a sensitivity of 84.4%, specificity of 68.2% and accuracy of 70.1%.

DISCUSSION
This is the first description of the use of elastography (ARFI) in the evaluation of dogs with a presumptive diagnosis of prostatic pathologies. The changes observed in B-mode and Doppler are nonspecific and may be present in most diseases that affect the prostate (Lévy et al., 2014;Souza et al., 2017) making it difficult to discriminate between different prostatic alterations due to the lack of pathognomonic characteristics (Zelli et al., 2013). Therefore, further investigation is needed through other diagnostic methods, such as the ARFI elastography, which demonstrated applicability through qualitative and quantitative assessments in this study.
In this study, the elastogram of all of the healthy animals showed homogeneity compatible with that described by Feliciano et al. (2015) who demonstrated the normal elastographic characteristics of dogs' prostates of different ages. On the other hand, there was a change in this characteristic only in sick animals. This change in homogeneity and increased tissue stiffness verified by the elastogram in sick prostates may be related to tissue repair resulting from a prostatic lesion (whether of physiological or pathological origin) which involves several mechanisms, and tissue regeneration (recovery of functional activity may occur) or healing, forming a fibrotic matrix. In cases of chronic or neoplastic lesions, the healing process interferes negatively, causing functional tissue loss (Balbino et al., 2005).
The tissue repair characteristics also justify the increase in SWV observed in the diseased prostates of this study. Although SWV values of altered prostate in dogs are not described, this change has already been observed in men with different types of prostatic alterations for example, in cases of benign prostatic hyperplasia, where it was found that the increase in SWV is directly related with the severity of the disease (Alan et al., 2016). Sigrist et al. (2017) cite that the greater the stiffness of the tissue, the greater the shear speed.
The presence of cystic structures and prostatic atrophy present in advanced cases of prostatic hyperplasia (as well as cases of prostatitis and neoplasia) are easily identified in B-mode (Shimomura et al., 2009) and according to Feliciano et al. (2015), these can alter the stiffness of the prostatic parenchyma tissue and subsequently promote changes in the shear velocity values. It is important to highlight that, like other organs, the prostate also has some limitations capable of affecting elastographic results, due to the structures present in the tissue, such as cystic components, or even the experience of the operator (Hoskins, 2011;Jeon et al., 2015;Nowicki and Dobruch-Sobczak, 2016).
In addition, even if predictive values were obtained to identify prostatic alterations through ARFI elastography in this study, it should be emphasized that this method does not replace the performance of cyto and histopathological exams, since only these are able to provide the definitive diagnosis and to discriminate the types of prostatic alterations (Domostawska et al., 2018). Therefore, elastography should be used together, and can assist in the early diagnosis of different prostatic changes that affect dogs. Based on the results obtained, ARFI elastography proved to be an appropriate and accurate method for detecting prostatic alterations, with high predictive performance variables. The shear speed was significantly higher in the tissues of the group of animals that presented clinical symptoms of prostatic disease when compared to the group of healthy animals.

CONCLUSIONS
The findings of this study may provide a baseline for clinical investigation, as ARFI elastography brings additional information to B-mode Doppler, assisting in the presumptive diagnosis of canine prostatic alterations.

ACKNOWLEDGMENTS
The present study was supported by FAPESP (São Paulo State Foundation Research) in the form thematic project (2017/14957-6 and 2019/09197-8) and CNPq (National Council for Scientific and Technological Development) in the form of productivity scholarship (process number 309199/2017-4). The authors also thank Jair Matos and Siemens Healthineers for technical assistance.