针对原木CT图像的一种中值滤波

在对传统的中值滤波和原木CT图像的特点的研究基础上,发现单纯地用中值滤波算法或均值滤波难以同时尽可能地消除混合噪声。针对这一问题,提出了一种改进的自适应中值滤波算法。对于不同的图像区域,算子也相应地有所不同,其中算子中的权值选取依赖于区域的灰度中值,且当某点的灰度越接近灰度中值,其权值就相应地越大。实践证明,新算法的处理结果优于传统的滤波算法和均值滤波,具有准确性和实时性。     

ASSESSMENT OF RESPIRATION‐INDUCED DISPLACEMENT OF CANINE ABDOMINAL ORGANS IN DORSAL AND VENTRAL RECUMBENCY USING MULTISLICE COMPUTED TOMOGRAPHY

Respiratory‐induced organ displacement during image acquisition can produce motion artifacts and variation in spatial localization of an organ in diagnostic computed tomography (CT) examinations. The purpose of this prospective study was to quantify respiratory‐induced abdominal organ displacement in dorsal and ventral recumbency using five normal dogs. All dogs underwent CT examinations using 64 multidetector row CT (64‐MDCT). A “3‐dimensional (3D) apneic CT exam” of the abdomen was acquired followed by a “4‐dimensional (4D) ventilated CT exam.” The liver, pancreas, both kidneys, both medial iliac lymph nodes, and urinary bladder were delineated on the 3D‐apneic examination and the organ outlines were compared to the maximum alteration in organ position in the 4D‐ventilated examination. Displacement was measured in dorsal‐to‐ventral (DV), right‐to‐left (RL), and cranial‐to‐caudal (CC) directions. Respiratory‐induced displacement of canine abdominal organs was not predictable and showed large variability in the three directions evaluated. For most canine abdominal organs, dorsal recumbency provided overall the least amount of displacement among all directions evaluated except for liver and urinary bladder. For liver, a large variability was found for all directions and a statistically significant difference was found only in the RL direction with ventral recumbency exhibiting less displacement (P = 0.0099). For the urinary bladder, ventral recumbency also provided less displacement but this was statistically significant only in the RL direction (P < 0.0001). Findings from this study indicated that dorsal recumbency may be preferred for minimizing respiratory motion artifacts in whole abdomen studies, but ventral recumbency may be preferred for liver and urinary bladder studies when respiration cannot be controlled.     

COMPUTED TOMOGRAPHIC FINDINGS IN 57 CATS WITH PRIMARY PULMONARY NEOPLASIA

Primary pulmonary neoplasia is relatively uncommon in cats and generally has a poor prognosis. In this multicenter, retrospective study of 57 cats with pulmonary neoplasia, the most frequent presenting signs were anorexia/inappetence (39%) and cough (37%). The pulmonary tumors were considered to be incidental findings in 9% cats. In computed tomographic (CT) images, primary pulmonary tumors appeared as a pulmonary mass in 55 (96%) cats and as a disseminated pulmonary lesion without a defined mass in two (4%) cats. Most pulmonary tumors were in the caudal lobes, with 28 (49%) in the right caudal lobe and 17 (30%) in the left caudal lobe. CT features associated with pulmonary tumors included mass in contact with visceral pleura (96%), irregular margins (83%), well‐defined borders (79%), bronchial compression (74%), gas‐containing cavities (63%), foci of mineral attenuation (56%), and bronchial invasion (19%). The mean (range) maximal dimension of the pulmonary masses was 3.5 cm (1.1–11.5 cm). Additional foci of pulmonary disease compatible with metastasis were observed in 53% cats. Pleural fluid was evident in 30% cats and pulmonary thrombosis in 12% cats. The histologic diagnoses were 47 (82%) adenocarcinomas, six (11%) tumors of bronchial origin, three (5%) adenosquamous cell carcinomas, and one (2%) squamous cell carcinoma. In this series, adenocarcinoma was the predominant tumor type, but shared many features with less common tumor types. No associations were identified between tumor type and CT features. Prevalence of suspected intrapulmonary metastasis was higher than in previous radiographic studies of cats with lung tumors.     

Effects of two different anesthetic protocols on 64‐MDCT coronary angiography in dogs

Heart rate is a major factor influencing diagnostic image quality in computed tomographic coronary artery angiography (MDCT‐CA), with an ideal heart rate of 60–65 beats/min in humans. The purpose of this prospective study was to compare effects of two different clinically applicable anesthetic protocols on cardiovascular parameters and 64‐MDCT‐CA quality in 10 healthy dogs. Scan protocols and bolus volumes were standardized. Image evaluations were performed in random order by a board‐certified veterinary radiologist who was unaware of anesthetic protocols used. Heart rate during image acquisition did not differ between protocols (P = 1), with 80.6 ± 7.5 bpm for protocol A and 79.2 ± 14.2 bpm for protocol B. Mean blood pressure was significantly higher (P > 0.05) using protocol B (protocol A 62.9 ± 9.1 vs. protocol B 72.4 ± 15.9 mmHg). The R‐R intervals allowing for best depiction of individual coronary artery segments were found in the end diastolic period and varied between the 70% and 95% interval. Diagnostic quality was rated excellent, good, and moderate in the majority of the segments evaluated, with higher scores given for more proximal segments and lower for more distal segments, respectively. Blur was the most commonly observed artifact and mainly affected the distal segments. No significant differences were identified between the two protocols for optimal reconstruction interval, diagnostic quality and measured length individual segments, or proximal diameter of the coronary arteries (P = 1). Findings indicated that, when used with a standardized bolus volume, both of these anesthetic protocols yielded diagnostic quality coronary 64‐MDCT‐CA exams in healthy dogs.     

MULTIDETECTOR‐ROW COMPUTED TOMOGRAPHY PATTERNS OF BRONCHOESPHAGEAL ARTERY HYPERTROPHY AND SYSTEMIC‐TO‐PULMONARY FISTULA IN DOGS

Anomalies involving arterial branches in the lungs are one of the causes of hemoptysis in humans and dogs. Congenital and acquired patterns of bronchoesophageal artery hypertrophy have been reported in humans based on CT characteristics. The purpose of this retrospective study was to describe clinical, echocardiographic, and multidetector computed tomography features of bronchoesophageal artery hypertrophy and systemic‐to‐pulmonary arterial communications in a sample of 14 dogs. Two main vascular patterns were identified in dogs that resembled congenital and acquired conditions reported in humans. Pattern 1 appeared as an aberrant origin of the right bronchoesophageal artery, normal origin of the left one, and enlargement of both the bronchial and esophageal branches that formed a dense network terminating in a pulmonary artery through an orifice. Pattern 2 appeared as a normal origin of both right and left bronchoesophageal arteries, with an enlarged and tortuous course along the bronchi to the periphery of the lung, where they communicated with subsegmental pulmonary arteries. Dogs having Pattern 1 also had paraesophageal and esophageal varices, with the latter being confirmed by videoendoscopy examination. Authors conclude that dogs with Pattern 1 should be differentiated from dogs with other congenital vascular systemic‐to‐pulmonary connections. Dogs having Pattern 2 should be evaluated for underlying pleural or pulmonary diseases. Bronchoesophageal artery hypertrophy can be accompanied by esophageal venous engorgement and should be included in the differential diagnosis for esophageal and paraesophageal varices in dogs.     

INTRAARTERIAL INJECTION OF IODINATED CONTRAST MEDIUM FOR CONTRAST ENHANCED COMPUTED TOMOGRAPHY OF THE EQUINE HEAD

Minimizing the volume of contrast administered for contrast‐enhanced computed tomography (CT) of the equine head is desirable for reducing costs and risks of adverse reactions, however evidence‐based studies on the effects of varying volumes on image quality are currently lacking. The objective of the current study was to determine whether low‐volume intraarterial administration of contrast medium would result in an equivalent image quality and tissue attenuation vs. high‐volume intravenous bolus administration. A prospective cross‐over experimental design was used in a sample of six horses. After anesthetic induction, the right carotid artery was exposed surgically and catheterized. Four CT scans of the cranium were performed for each horse: baseline, immediately following intraarterial contrast injection, five‐min postinjection (return to baseline) and a final scan after intravenous contrast administration. Soft tissue attenuation in predetermined regions of interest (ROI); and length, width, and height measurements of the pituitary gland were recorded at each time point. Horses were euthanized and measurements of the pituitary gland were repeated postmortem. No adverse reactions to contrast administration were observed. Intraarterial and intravenous administration of contrast medium resulted in significantly greater soft tissue enhancement of some brain ROI's and the pituitary gland vs. baseline values. Pituitary gland measurements made on postcontrast CT images did not differ from those obtained during postmortem examination. Findings indicated that low‐dose intraarterial administration of contrast material in the equine head resulted in comparable soft tissue enhancement vs. high volume intravenous administration.     

High-resolution fluorodeoxyglucose positron emission tomography and magnetic resonance imaging findings of a pituitary microtumor in a dog

A 16-year-old, castrated, male English cocker spaniel dog was presented due to generalized alopecia. Routine clinical pathology, endocrine and abdominal ultrasonography results were consistent with a diagnosis of pituitary-dependent hyperadrenocorticism. The adenohypophyseal lesion was clearly visualized on both 3 T and 7 T magnetic resonance imaging (MRI) of the pituitary gland. Although biochemical and MRI findings were consistent with a functional pituitary microtumor, a pituitary lesion was not detected using ¹⁸F-fluorodeoxyglucose positron emission tomography (FDG-PET). This report firstly describes the application of high-resolution FDG-PET to a spontaneous pituitary microtumor in a dog.