Magnetic Resonance Imaging, Computed Tomography, and Gross Anatomy of the Canine Tarsus Kirsten J. Deruddere1, BVSc, Marjorie E. Milne2, BVSc, FANZCVS (Radiology), Kane M. Wilson2, BAppSci, and Sam R. Snelling1, BVSc, FANZCVS (Small Animal Surgery) 1
Advanced Vetcare, Kensington, Australia and
Corresponding Author Kirsten J. Deruddere, BVSc, Level 1, 26 Robertson St, Kensington, VIC 3031, Australia. E‐mail: kirsten.deruddere@advancedvetcare. com.au Submitted March 2013 Accepted October 2013 DOI:10.1111/j.1532-950X.2014.12194.x
2
University of Melbourne, Werribee, Australia
Objectives: To describe the normal anatomy of the soft tissues of the canine tarsus as identified on computed tomography (CT) and magnetic resonance imaging (MRI) and to evaluate specific MRI sequences and planes for observing structures of diagnostic interest. Study Design: Prospective descriptive study. Animals: Canine cadavers (n ¼ 3). Methods: A frozen cadaver pelvic limb was used to trial multiple MRI sequences using a 1.5 T superconducting magnet and preferred sequences were selected. Radiographs of 6 canine cadaver pelvic limbs confirmed the tarsi were radiographically normal. A 16‐slice CT scanner was used to obtain 1 mm contiguous slices through the tarsi. T1‐weighted, proton density with fat suppression (PD FS) and T2‐weighted MRI sequences were obtained in the sagittal plane, T1‐weighted, and PD FS sequences in the dorsal plane and PD FS sequences in the transverse plane. The limbs were frozen for one month and sliced into 4–5 mm thick frozen sections. Anatomic sections were photographed and visually correlated to CT and MR images. Results: Most soft tissue structures were easiest to identify on the transverse MRI sections with cross reference to either the sagittal or dorsal plane. Bony structures were easily identified on all CT, MR, and gross sections. Conclusions: The anatomy of the canine tarsus can be readily identified on MR imaging.
Magnetic resonance imaging (MRI) is the most accurate modality for assessing tendons and soft tissue structures of the human ankle and foot.1 Indications for MRI of the foot and ankle in people include ligament injury, tendon pathology, impingement, plantar fasciitis, plantar plate injury, os trigonum or sinus tarsi syndrome, occult fracture, osteochondritis dissecans (OCD), osteonecrosis, or degenerative disease.2 Reports on advanced imaging of the normal canine tarsus are limited to a computed tomography (CT) study by Gielen et al.3 that examined only transverse sections, and the canine CT and MRI atlas by Assheuer and Sager4 which used a magnet with field strength of 1 T, and only identified a small number of soft tissue structures. The canine tarsus is a complex joint which is difficult to assess on radiographs because of superimposition of various bony structures and lack of soft tissue contrast. It consists of seven tarsal bones that articulate with one another as well as with the tibia, fibula, and metatarsal bones. The articulations are held together by multiple ligaments and are enveloped by one continuous fibrous joint capsule.5 For MRI to become a useful modality for investigation of soft tissue injuries in the canine tarsus, a detailed description of normal canine tarsal anatomy on MR images must be available. Our purpose was to provide a detailed description of the anatomy of the normal
912
canine tarsus as depicted on CT and MR images, and correlated with gross anatomy. We also determined MRI acquisition sequences and planes for identification of structures of diagnostic interest.
MATERIALS AND METHODS A frozen pelvic limb was obtained from a 25 kg dog euthanatized for reasons unrelated to the study. Various MRI sequences were trialed using a 1.5 T superconducting magnet (GE Signa HD, Milwaukee, WI, 1.5 T) based on sequences commonly used in MRI of the human ankle. Two separate coils were used; a receive‐only single channel small flex coil (GE Medical Systems, Milwaukee, WI) and a transmit/receive 8 channel knee coil (Invivo Corporation, Gainesville, FL). Sequences trialed for dorsal, sagittal, and transverse planes included T1‐weighted, T1‐weighted with fat suppression (T1 FS), proton density (PD), proton density with fat suppression (PD FS), T2‐weighted, T2‐weighted with fat suppression (T2 FS), short tau inversion recovery (STIR), and T2 gradient echo for each coil, with a field of view (FOV) of 10 cm 10 cm and 14 cm 14 cm. The acquisition matrix was 256 192 and was consistent between the coils. Fat suppression was achieved
Veterinary Surgery 43 (2014) 912–919 © Copyright 2014 by The American College of Veterinary Surgeons
Deruddere et al.
using chemical pre‐saturation. A veterinary radiologist (M.M.) reviewed the sequences and designed the MRI protocol used, balancing optimal signal‐to‐noise ratio, spatial resolution, acquisition time, and ability to identify anatomic structures. The pelvic limbs of 3 hound‐type dogs, euthanatized for reasons unrelated to the study or any form of lameness, were examined and mediolateral and dorsoplantar radiographs of all tarsi were performed. Only tarsi without gross or radiographic evidence of abnormalities were selected. The pelvic limbs were removed from the cadavers by disarticulation at the level of the stifle joint. The limbs were evaluated using CT and MRI within 10 hours of euthanasia. CT CT was performed on each tarsus with the pelvic limb placed in a position mimicking a dog in sternal recumbency, with the tarsus in 140° flexion. Lateral and dorsal topogram views were obtained using a 16‐slice CT scanner (Somatom Emotion, Siemens, Erlangen, Germany). One millimeter thick contiguous slices were obtained from the distal third of the tibia to the distal third of the metatarsal bones. The CT settings used were: 108 mAs, 130 kVp. An acquisition matrix of 512 512 was used. The FOV differed slightly between individual limbs but ranged from 30 cm 25 cm to 34 cm 28.3 cm. The raw data were reconstructed with sharpening and smoothing algorithms and viewed in a bone window (window width ¼ 2,000 HU; window level ¼ 500 HU) and soft tissue window (window width ¼ 400 HU; window level ¼ 50 HU). Data were reformatted into 2 mm thick slices in transverse (relative to the tarsometatarsal joint), sagittal (parallel to the sagittal plane of the calcaneus), and dorsal (perpendicular to sagittal and transverse planes) planes. Images were viewed simultaneously in multiple planes in soft tissue and bone windows using proprietary viewing software (Osirix version 3.9.4, 32‐bit, Pixmeo, Geneva, Switzerland). MRI MRI was performed on each tarsus using a 1.5 T superconducting magnet (GE Signa HD, 1.5 T) and a receive‐only single channel
MRI, CT, and Gross Anatomy of the Canine Tarsus
small flex coil (GE Medical Systems). Limbs were placed in 140° flexion mimicking a dog in sternal recumbency. Two‐millimeter contiguous images were acquired from the proximal aspect of the calcaneus to the proximal third of the metatarsal bones. Alignment of the transverse, sagittal, and dorsal scan planes were the same as those described for CT acquisition. The parameters for each sequence were as follows: in the transverse plane ¼ PD FS (repetition time [TR]: 3,200 milliseconds, echo time [TE]: 30 milliseconds), in the dorsal plane ¼ T1‐weighted (TR: 500 milliseconds, TE: 13 milliseconds), PD FS (TR: 3,200 milliseconds, TE: 30 milliseconds), in the sagittal plane ¼ T1‐weighted (TR: 500 milliseconds, TE: 13 milliseconds), T2‐ weighted (TR: 5,900 milliseconds, TE: 100 milliseconds), and PD FS (TR: 3,200 milliseconds, TE: 30 milliseconds) with a FOV of 10 cm 10 cm. The acquisition matrix used was 256 192. Acquisition time per limb was
Advanced Vetcare, Kensington, Australia and
Corresponding Author Kirsten J. Deruddere, BVSc, Level 1, 26 Robertson St, Kensington, VIC 3031, Australia. E‐mail: kirsten.deruddere@advancedvetcare. com.au Submitted March 2013 Accepted October 2013 DOI:10.1111/j.1532-950X.2014.12194.x
2
University of Melbourne, Werribee, Australia
Objectives: To describe the normal anatomy of the soft tissues of the canine tarsus as identified on computed tomography (CT) and magnetic resonance imaging (MRI) and to evaluate specific MRI sequences and planes for observing structures of diagnostic interest. Study Design: Prospective descriptive study. Animals: Canine cadavers (n ¼ 3). Methods: A frozen cadaver pelvic limb was used to trial multiple MRI sequences using a 1.5 T superconducting magnet and preferred sequences were selected. Radiographs of 6 canine cadaver pelvic limbs confirmed the tarsi were radiographically normal. A 16‐slice CT scanner was used to obtain 1 mm contiguous slices through the tarsi. T1‐weighted, proton density with fat suppression (PD FS) and T2‐weighted MRI sequences were obtained in the sagittal plane, T1‐weighted, and PD FS sequences in the dorsal plane and PD FS sequences in the transverse plane. The limbs were frozen for one month and sliced into 4–5 mm thick frozen sections. Anatomic sections were photographed and visually correlated to CT and MR images. Results: Most soft tissue structures were easiest to identify on the transverse MRI sections with cross reference to either the sagittal or dorsal plane. Bony structures were easily identified on all CT, MR, and gross sections. Conclusions: The anatomy of the canine tarsus can be readily identified on MR imaging.
Magnetic resonance imaging (MRI) is the most accurate modality for assessing tendons and soft tissue structures of the human ankle and foot.1 Indications for MRI of the foot and ankle in people include ligament injury, tendon pathology, impingement, plantar fasciitis, plantar plate injury, os trigonum or sinus tarsi syndrome, occult fracture, osteochondritis dissecans (OCD), osteonecrosis, or degenerative disease.2 Reports on advanced imaging of the normal canine tarsus are limited to a computed tomography (CT) study by Gielen et al.3 that examined only transverse sections, and the canine CT and MRI atlas by Assheuer and Sager4 which used a magnet with field strength of 1 T, and only identified a small number of soft tissue structures. The canine tarsus is a complex joint which is difficult to assess on radiographs because of superimposition of various bony structures and lack of soft tissue contrast. It consists of seven tarsal bones that articulate with one another as well as with the tibia, fibula, and metatarsal bones. The articulations are held together by multiple ligaments and are enveloped by one continuous fibrous joint capsule.5 For MRI to become a useful modality for investigation of soft tissue injuries in the canine tarsus, a detailed description of normal canine tarsal anatomy on MR images must be available. Our purpose was to provide a detailed description of the anatomy of the normal
912
canine tarsus as depicted on CT and MR images, and correlated with gross anatomy. We also determined MRI acquisition sequences and planes for identification of structures of diagnostic interest.
MATERIALS AND METHODS A frozen pelvic limb was obtained from a 25 kg dog euthanatized for reasons unrelated to the study. Various MRI sequences were trialed using a 1.5 T superconducting magnet (GE Signa HD, Milwaukee, WI, 1.5 T) based on sequences commonly used in MRI of the human ankle. Two separate coils were used; a receive‐only single channel small flex coil (GE Medical Systems, Milwaukee, WI) and a transmit/receive 8 channel knee coil (Invivo Corporation, Gainesville, FL). Sequences trialed for dorsal, sagittal, and transverse planes included T1‐weighted, T1‐weighted with fat suppression (T1 FS), proton density (PD), proton density with fat suppression (PD FS), T2‐weighted, T2‐weighted with fat suppression (T2 FS), short tau inversion recovery (STIR), and T2 gradient echo for each coil, with a field of view (FOV) of 10 cm 10 cm and 14 cm 14 cm. The acquisition matrix was 256 192 and was consistent between the coils. Fat suppression was achieved
Veterinary Surgery 43 (2014) 912–919 © Copyright 2014 by The American College of Veterinary Surgeons
Deruddere et al.
using chemical pre‐saturation. A veterinary radiologist (M.M.) reviewed the sequences and designed the MRI protocol used, balancing optimal signal‐to‐noise ratio, spatial resolution, acquisition time, and ability to identify anatomic structures. The pelvic limbs of 3 hound‐type dogs, euthanatized for reasons unrelated to the study or any form of lameness, were examined and mediolateral and dorsoplantar radiographs of all tarsi were performed. Only tarsi without gross or radiographic evidence of abnormalities were selected. The pelvic limbs were removed from the cadavers by disarticulation at the level of the stifle joint. The limbs were evaluated using CT and MRI within 10 hours of euthanasia. CT CT was performed on each tarsus with the pelvic limb placed in a position mimicking a dog in sternal recumbency, with the tarsus in 140° flexion. Lateral and dorsal topogram views were obtained using a 16‐slice CT scanner (Somatom Emotion, Siemens, Erlangen, Germany). One millimeter thick contiguous slices were obtained from the distal third of the tibia to the distal third of the metatarsal bones. The CT settings used were: 108 mAs, 130 kVp. An acquisition matrix of 512 512 was used. The FOV differed slightly between individual limbs but ranged from 30 cm 25 cm to 34 cm 28.3 cm. The raw data were reconstructed with sharpening and smoothing algorithms and viewed in a bone window (window width ¼ 2,000 HU; window level ¼ 500 HU) and soft tissue window (window width ¼ 400 HU; window level ¼ 50 HU). Data were reformatted into 2 mm thick slices in transverse (relative to the tarsometatarsal joint), sagittal (parallel to the sagittal plane of the calcaneus), and dorsal (perpendicular to sagittal and transverse planes) planes. Images were viewed simultaneously in multiple planes in soft tissue and bone windows using proprietary viewing software (Osirix version 3.9.4, 32‐bit, Pixmeo, Geneva, Switzerland). MRI MRI was performed on each tarsus using a 1.5 T superconducting magnet (GE Signa HD, 1.5 T) and a receive‐only single channel
MRI, CT, and Gross Anatomy of the Canine Tarsus
small flex coil (GE Medical Systems). Limbs were placed in 140° flexion mimicking a dog in sternal recumbency. Two‐millimeter contiguous images were acquired from the proximal aspect of the calcaneus to the proximal third of the metatarsal bones. Alignment of the transverse, sagittal, and dorsal scan planes were the same as those described for CT acquisition. The parameters for each sequence were as follows: in the transverse plane ¼ PD FS (repetition time [TR]: 3,200 milliseconds, echo time [TE]: 30 milliseconds), in the dorsal plane ¼ T1‐weighted (TR: 500 milliseconds, TE: 13 milliseconds), PD FS (TR: 3,200 milliseconds, TE: 30 milliseconds), in the sagittal plane ¼ T1‐weighted (TR: 500 milliseconds, TE: 13 milliseconds), T2‐ weighted (TR: 5,900 milliseconds, TE: 100 milliseconds), and PD FS (TR: 3,200 milliseconds, TE: 30 milliseconds) with a FOV of 10 cm 10 cm. The acquisition matrix used was 256 192. Acquisition time per limb was
