Dog population
This study took place at Evidensia Animal Hospital Södra Djursjukhuset (“South Animal Hospital”), Stockholm, during the period August 2019 to August 2020. Dogs included in this study were selected from a larger group of dogs undergoing a CT examination for front limb orthopaedic disease. The specific group was selected because they receive sedation as a standard, and they are examined in a sternal position, which was two important criteria for this study. None of the dogs received any intravenous contrast medium as a part of their orthopaedic limb CT-exam.
Dogs with lung disease, signs of systemic infection or inflammation according to blood test results and clinical examination, brachycephalic breeds, and dogs with recently known trauma or with suspicion of neoplasia were excluded. All dogs underwent physical examination prior to the CT-scanning including cardiac and pulmonary auscultation, heart rate, respiratory rate, assessment of general condition, and examination of mucus membranes. Basic parameters such as age, breed and gender were obtained from the patients’ journal. Blood samples were collected at the day of the study and C-reactive protein (CRP) and complete blood count was analyzed. Animals with hematologic signs of systemic infection or inflammation were excluded.
Study design
All animals received an aseptically peripheral vein catheter placed into a cephalic vein before receiving any sedatives. Sedation was given intravenously using the following protocol: butorphanol (Butomidor vet® 10 mg/mL) 0.2–0.4 mg/kg in combination with dexmedetomidine (Dexdomitor® 5 mg/mL) 2.5–5 mcg/kg followed by sterile saline to flush the syringe. If additional sedation was needed, this was also administered intravenously. Sedatives were given directly prior to the first CT scan. The dog was positioned in sternal recumbency with front legs pulled forward, and head in between its front legs.
Three lung scans (Scan 1—3) were performed on each patient. Time zero was stated as the time of sedation. According to Fig. 1, Scan 1 was performed as soon as possible after the patient was sufficiently sedated and correctly positioned, Scan 2 as shown approximately 5 min later. The ordinary scan, meaning the designated orthopaedic scan of the patient, was then performed prior to the third delayed scan, Scan 3, which could vary in time depending on how long time the ordinary scan took. A protocol was used to note the time for all three scans (Table 2).
CT examination and interpretation
All dogs were scanned in the same scanner (Philips Ingenuity Core 128 powered by iPatient, scanner V4.1.7.10503, Philips Healthcare Netherland B.V, Veenpluis 4–6, 5684 PC Best, Netherlands) 128 slice. A standardized lung protocol was used with technical parameters as follows: slice thickness 1.0 mm, pitch 1.074, matrix 512, image field of view − 600 to 1600, lung reconstruction algorithm Y-Detail (YD), kV 100, mAs 182, rotation time 0,75. Window width and level could be adjusted between study animals. Scans were obtained in a cranial to caudal direction with the dog in a sternal position. The time required for doing a lung scan using the dedicated machine varies between 3 to 10 s depending on the dog’s size.
Images were viewed in a PACS working station using the same software program for measurements of all dogs (HOROS™ dicom viewer). All images were analyzed by a single observer (aspirant in the Swedish program for Specialist in Diagnostic Imaging) under supervision of a board-certified specialist (Dipl.ECVDI).
Each lung lobe: left cranial lobe pars cranialis (LCPCr) and pars caudalis (LCPCa), left caudal (LC) lobe, right cranial lobe (RCr), right middle lobe (RM), right caudal lobe (RCa) and the accessory lobe (LA) were individually examined at all three scans. An axial slice where the most ventral tip of the lobe was included, was chosen and images from the three scans were compared and propagated so that measurements were done at the same topographic level on all time intervals. If the dog was repositioned between the lung scans, anatomical landmarks such as vertebrae, bronchi and vessels were used to examine the lobe at the same level. Care was taken that the most ventral tip of the lung lobe was included in the slice since lung atelectasis was expected to develop at this level first [15].
Three measurements of lung attenuation measured in HU were collected for each lobe (Fig. 2). First, a mean attenuation for each lung lobe was determined by manually drawing out the outlines of the lung lobe using a software tool included in HOROS™, excluding the main bronchi and aligned vessels. Second, a region of interest (ROI) of a minimum of 1 cm in diameter was manually drawn in the most dorsal part of the lung lobe. And third, a circle of ROI of a minimum of 1 cm in diameter, was drawn at the most ventral part of the lobe. Care was taken to avoid larger blood vessels and bronchi and the same anatomical areas were included in all of the three scans. If areas of visually increased attenuation could be seen these were also included using a free hand circle of ROI.
Lung attenuation was measured in HU and following values were used for interpreting the degree of lung attenuation: nonaerated (− 100 to + 100 HU) indicative of atelectasis, poorly aerated (− 500 to − 101 HU), normally aerated (− 900 to − 501 HU) or hyperaerated (− 1000 to − 901 HU) [12].
Statistical analyses
The statistical analyses were carried out in Matlab R 2020a.
To determine if there was a statistically significant increase in lung attenuation over time, a paired test “Signed rank Wilcoxon test” was used. For each individual, a mean of the ventral attenuation values of Scan 1 and scan 2 respectively Scan 1 and Scan 3 were compared.