Can CT texture analysis parameters be used as imaging biomarkers for prediction of malignancy in canine splenic tumors?

Splenic hemangiosarcoma has morphological similarities to benign nodular hyperplasia. Computed tomography (CT) texture analysis can analyze the texture of images that the naive human eye cannot detect. Recently, there have been attempts to incorporate CT texture analysis with artificial intelligence in human medicine. This retrospective, analytical design study aimed to assess the feasibility of CT texture analysis in splenic masses and investigate predictive biomarkers of splenic hemangiosarcoma in dogs. Parameters for dogs with hemangiosarcoma and nodular hyperplasia were compared, and an independent parameter that could differentiate between them was selected. Discriminant analysis was performed to assess the ability to discriminate the two splenic masses and compare the relative importance of the parameters. A total of 23 dogs were sampled, including 16 splenic nodular hyperplasia and seven hemangiosarcoma. In each dog, total 38 radiomic parameters were extracted from first-, second-, and higher-order matrices. Thirteen parameters had significant differences between hemangiosarcoma and nodular hyperplasia. Skewness in the first-order matrix and GLRLM_LGRE and GLZLM_ZLNU in the second, higher-order matrix were determined as independent parameters. A discriminant equation consisting of skewness, GLZLM_LGZE, and GLZLM_ZLNU was derived, and the cross-validation verification result showed an accuracy of 95.7%. Skewness was the most influential parameter for the discrimination of the two masses. The study results supported using CT texture analysis to help differentiate hemangiosarcoma from nodular hyperplasia in dogs. This new diagnostic approach can be used for developing future machine learning-based texture analysis tools.

3.0-Tesla MRI of normal canine adrenal glands.

Adrenal disease is a common problem in dogs and MRI is increasingly being used as an adjunctive diagnostic test. To date, the MRI features of normal canine adrenal glands have only been reported using 1.5-Tesla (T) MRI. The aims of this prospective, methods-comparison, exploratory study were to evaluate the effects of pulse sequence on the appearance of normal canine adrenal glands using 3 Tesla MRI. Six research beagle dogs were sampled and the following pulse sequences were acquired for each: (1) T2-weighted images using two-dimensional (2D) turbo spin-echo (TSE), single-shot spin-echo (SSTSE), and three-dimensional (3D) TSE, (2) T1-weighted images using 2D TSE, 3D TSE, and 3D turbo field echo sequences, (3) post-contrast T1-weighted images, and (4) chemical shift imaging. The signal-to-noise ratio and contrast-to-noise ratio were measured for each dog and each pulse sequence. The signal intensity, clarity of the contour, distinction of the corticomedullary junction, degrees of motion, partial volume, and chemical shift artifact, and homogeneity of the contrast enhancement were evaluated qualitatively. In all sequences, except for chemical shift imaging, the adrenal glands were visualized in both planes with successful control of motion artifacts by manual ventilation. The adrenal contour was considered to be most clearly visualized with 2D TSE. Adrenal images were acquired within the shortest time using SSTSE although the contour was less clearly visualized than with TSE. Findings from this study in normal dogs can serve as background for further 3.0-T MRI studies of dogs with adrenal disease.

Comparison of sulfur hexafluoride microbubbles and perfluorobutane for performance of contrast-enhanced ultrasonography of the pancreas in dogs.

OBJECTIVE: To compare qualitative features and quantitative parameters of 2 contrast agents (sulfur hexafluoride microbubbles [SHM; SonoVue] and perfluorobutane [PFB; Sonazoid]) for performance of contrast-enhanced ultrasonography (CEUS) of the pancreas in dogs. ANIMALS: 8 healthy Beagles. PROCEDURES: In a crossover study design, CEUS of the pancreas was performed twice in each dog, once with SHM and once with PFB, in random order with at least 3 days between examinations. The recorded cine images were qualitatively assessed for homogeneity of pancreatic enhancement and conspicuity of the pancreatic signal relative to the background. For the quantitative assessment, circular regions of interest were placed over the pancreatic body, and a time-intensity curve was obtained. For each region of interest, CEUS parameters including peak intensity (PI), time to peak pancreatic enhancement, area under the curve (AUC), and wash-in rate were obtained. RESULTS: The homogeneity of the pancreatic parenchyma was not significantly different between contrast agents. The signal conspicuity relative to background noise was significantly higher with PFB than with SHM. Mean values of PI, wash-in rate, and AUC were significantly higher with PFB than with SHM. Time to peak enhancement was not significantly different between contrast agents. CONCLUSIONS AND CLINICAL RELEVANCE: Pancreatic CEUS with SHM and PFB produced similar homogeneity scores, but only PFB provided excellent signal conspicuity. Perfluorobutane produced higher values of PI, wash-in rate, and AUC. Findings indicated that PFB can provide homogeneous and strong enhancement of the pancreas during CEUS in healthy dogs and that pancreatic CEUS parameter values differ with the contrast agent used.

Efficacy of a phospholipid-stabilized sulfur hexafluoride microsphere contrast agent and water for hydrosonography of the upper portion of the gastrointestinal tract in dogs.

OBJECTIVE: To investigate the efficacy of a phospholipid-stabilized sulfur hexafluoride microsphere (SHM) contrast agent and water for hydrosonography of the upper portion of the gastrointestinal tract of dogs. ANIMALS: 12 healthy adult Beagles. PROCEDURES: In a crossover study, each dog was anesthetized and underwent noncontrast ultrasonography then hydrosonography following administration of tap water (30 mL/kg) without (water method) or with SHM (0.1 mL; SHM method) via an orogastric tube. There were at least 3 days between hydrosonographic procedures. Wall thickness, wall layer definition, conspicuity of the mucosal-luminal interface, and image quality were evaluated separately in the near and far fields for the gastric cardia, body, and pylorus and descending duodenum and compared among the 3 scanning methods. RESULTS: Mean wall thickness measurements did not differ significantly between the water and SHM methods at any location except the far-field gastric cardia where the mean wall thickness for the SHM method was less than that for the water method. In general, the SHM method improved wall layer definition and conspicuity of the mucosal-luminal interface of structures in the near field, compared with noncontrast method. The water and SHM methods both resulted in superior image quality relative to the noncontrast method for the near-field gastric cardia, far-field gastric cardia, and far-field duodenum. CONCLUSIONS AND CLINICAL RELEVANCE: Results indicated that, for dogs, gastrointestinal hydrosonography by use of the SHM method improved wall layer definition and mucosal conspicuity, particularly in near-field images of the upper portion of the gastrointestinal tract.

Radiographic assessment of splenic size and correlation with splenic measurements estimated by use of computed tomography in healthy cats.

OBJECTIVE: To investigate radiographic variables for correlation with splenic size as estimated with CT in cats. ANIMALS: 38 healthy adult cats. PROCEDURES: The width and height of the splenic head and total length, segmental length, and width of the spleen were measured on radiographic and CT images obtained from 10 cats in prospective, exploratory experiments. Distance between the splenic head and left kidney, anatomic locations of the head and tail of the spleen, and CT-derived splenic volume were also assessed. Correlation and agreement between radiographic and CT measurements and interobserver agreement for measurements with each method were determined. A retrospective evaluation of radiographs obtained without sedation or anesthesia for 28 cats was performed to establish preliminary guidelines for the measurement deemed the most reliable estimator of splenic size. RESULTS: Radiographic measurements of total and segmental splenic length were significantly correlated with the respective CT measurements and with splenic volume. Agreement between radiographic and CT measurements of segmental length was good; interobserver agreement was excellent for all variables. In retrospective evaluations, median segmental length of the spleen was 57.87 mm (range, 34.72 to 105.44 mm) on radiographs; the caudal border of the splenic head on lateral views was located from the cranial part of L1 to the caudal part of L2, and the caudal border of the splenic tail on ventrodorsal views was located from the caudal part of L2 to the caudal part of L5. CONCLUSIONS AND CLINICAL RELEVANCE: Results indicated that segmental length of the spleen on radiographs is a reliable estimator of splenic size in healthy cats.