Ultrasonography methods for predicting malignancy in canine mammary tumors

Abstract

The aim of this study was to evaluate and compare the efficacy of B-mode, Doppler, contrast-enhanced ultrasonography (CEUS), and Acoustic Radiation Force Impulse (ARFI) elastography in predicting malignancy in canine mammary masses. This was a prospective cohort study from 2014 to 2016, which included 153 bitches with one or more mammary masses. A total of 300 masses were evaluated by ultrasonography (B-mode, Doppler, CEUS, and ARFI) and subsequently classified as benign or malignant by histopathology. Each ultrasound parameters studied were compared between benign and malignant masses by Chi-square or Student tests and differences were considered significant when P < 0.01. For the variables that proved significant differences were estimated the cut-off point, sensitivity, specificity, accuracy, and area under curve (AUC) by receiver-operating characteristic curve (ROC) analysis in a logistic regression model using histopathological classification as reference, to assess and compare diagnostic performance of each technique. Out of 300 mammary masses evaluated 246 were classified as malignant and 54 as benign. B-mode measurements showed sensitivity 67.9%, and specificity 67.6% as malignancy predictors on canine mammary masses; Doppler indexes systolic (>21.2 m/s) and diastolic velocity (>4.8 m/s) sensitivity 79.2% and specificity 70.8%; CEUS wash-out time (<80.5 s) sensitivity 80.2% and specificity 16.7%; and ARFI elastography shear velocity (SWV > 2.57 m/s) sensitivity 94.7% and specificity 97.2% In conclusion B-mode and Doppler ultrasound evaluations may assist in malignancy prediction of canine mammary masses with moderate sensitivity and specificity, already the SWV was an great accurate predictor. Therefore, ARFI elastography exam inclusion in veterinary clinic oncology and research is highly recommended, since it allows fast, non-invasive, and complication-free malignancy prediction of canine mammary masses.

Accuracy of four ultrasonography techniques in predicting histopathological classification of canine mammary carcinomas

“…However, in the present study, only width:length ratio greater than 0.53 obtained by B‐mode ultrasonography enabled the identification of special type carcinomas with acceptable accuracy. These findings are in agreement with previous reports, which have identified this ratio to be an indicator of malignancy in mammary tumors in women and female dogs . Thus, this parameter can be used as a simple and economical tool in the diagnosis of mammary neoplasms in female dogs.…”

“…Free and informed consent was obtained from the owners prior to the animals being included in the primary study. For this secondary study, the data from physical and ultrasound examinations prior to mastectomy and histopathological classification of the canine mammary masses diagnosed as carcinomas in the primary study were retrieved and reanalyzed . A total of 246 mammary carcinomas were analyzed.…”

“…As part of the inclusion criteria for this secondary study, ultrasonography for each included dog had been carried out by a single experienced sonographer (15 years), using virtual touch system, contrast agent technology, and a linear array 4–18 MHz transducer (Acuson S2000TM, eSie TouchTM, and CadenceTM; Siemens, Munich, Germany) according to the methodology previously described . The variables analyzed by B‐mode ultrasonography were echotexture (homogenous or heterogeneous); contours/margins (defined or undefined); invasiveness (present or absent); length (cm), width (cm), and width:length ratio in longitudinal sections; and height (cm), width (cm), and width:height ratio in transverse sections (Figure ).…”

“…Histopathological diagnosis and classification for each of the carcinomas were based on interpretations by a single experienced pathologist (25 years) following the methodology previously described . Carcinomas were classified into simple, complex, or special types and grade I, II, or III according to literature …”

“…A recent study involving 153 female dogs and 300 mammary masses evaluated and compared the efficacy of the ultrasound methods currently available, B‐mode, Doppler, contrast enhanced ultrasonography, and elastography–acoustic radiation force impulse in predicting tumor malignancy. B‐mode measurements showed 67.9% sensitivity and 67.6% specificity; Doppler, 79.2% sensitivity and 70.8% specificity; contrast enhancement ultrasonography, 80.2% sensitivity and 16.7% specificity; and acoustic radiation force impulse shear wave velocities, 94.7% sensitivity and 97.2% specificity …”