The interscapular region is complex to investigate, with layers of overlaid muscles whose spatial relationships are difficult to evaluate from conventional imaging and anatomic references. Moreover, this region is characterized by high mobility that in some instances can over-or underestimate the relationships among muscles and between muscles and scapulae. A relevant aspect not considered in traditional literature together with a rapid identification of muscles themselves and of their spatial relationship with surrounding structures.

To our knowledge, this is the first atlas that provides an exhaustive description of the interscapular region of the cat and considers its dynamic behaviour. Data were obtained by comparing cross-sectional imaging of gross, CT and MRI anatomy, which permitted to better identify the musculoskeletal and myofascial components of the region while taking full advantage of the potential of the three methods.

From a technical point of view, this multiple approach represents a novelty in its considering both techniques simultaneously while previous studies in rabbits [20–23] dogs [6–8] and cats [9, 10] compared either cross-sectional anatomy and CT or cross-sectional anatomy and MRI. Further, unlike other studies [6–8, 10, 20–23], in our work the sections were 3 mm-thick on CT, 3 mm-thick on MRI, and approximately 6 mm-thick on cross-sectional anatomy, allowing improved comparison between gross anatomy slices and advanced imaging techniques. In this contest, the choice to maintain vertebral bodies as the reference pivot proved to be a useful tool in detecting anddescribing the anatomical structures.

All images obtained in this study were of excellent quality and no abnormalities due to sample preservation were observed.

As expected, soft tissue details were superior on MRI than on CT, and the sequence identified as High Resolution Gradient Echo showed better rendering of the morphology of the musculoskeletal structures. Low field MRI required a long acquisition time when 3 mm thickness was applied (total average acquisition time: 50 min); an aspect that might represent a problem only in clinical activities requiring a prolonged anaesthesia time, but not in our investigation carried on dead animals. Conversely, CT acquisition required a short resolution time; depending on the tomography used, it ranged from 35 to 15 s for single or multi-detector, respectively. In subject C we decreased thickness of the CT-slices down to 1.25 mm and no significant improvement in image quality was observed in a transverse plane. Opposite to CT, in the same subject we increased thickness of MRI slices up to 4 mm. Data demonstrated that the quality of images kept unchanged, but time taken was significantly shorter (total average acquisition time: 35 min), which makes clinical application more feasible.

The corpses of the cats were positioned in sternal recumbency with their forelimbs extended cranially or caudally along their bodies. As a consequence, the scapula and its related muscles moved against the chest wall in two different ways even thanks to the characteristic loose connective tissue presents between the limb and the trunk, where the axilla is located.

We observed that changes in the position of the scapula can modify the shape of the muscles, as clearly appreciable by CT and MRI and confirmed by gross anatomy images. Although Travetti et al. [12] provided exhaustive images of the interscapular region, it is important to underline that those images derived from cats affected by Feline Injection-Site Sarcoma (FISS), with manifest tumours in between their muscles. A situation where not only the movements of the limbs, but also the volume of the neoplastic masses may change the shape of the muscles implied, thus complicating the interpretation of the images. One of the key strengths of our work is that its images were derived from cats with no lesion, and therefore their consultation should avoid any misinterpretation.

In veterinary medicine_, sternal recumbency is the standard positioning commonly recognized for CT studies of the thoracic region [11]; however, in the cat, tomographic references available in literature are all obtained with animals positioned in dorsal recumbency [9, 10].

In animals, sternal recumbency should be preferred to dorsal recumbency to reduce possible lung congestion occurring in prolonged dorsal recumbency, with the forelimbs extended cranially to avoid including them in the visual field. In our study, sternal positioning associated to forelimbs either flexed along the body or pulled cranially was aimed at better visualizing the musculoskeletal structures by enhancing their mobility. Moreover, since double positioning made the myofascial system and tissue adipose infiltrations visible, the images acquired can support a more accurate estimation of the relationship between both infiltrating and neoplastic lesions and surrounding tissues.

In particular, in cats affected by FISS, a deep knowledge of the sectional anatomy of the region is mandatory in order to reduce post-surgical complications. It has recently been demonstrated that surgery time is the best predictor of wound healing complications and that they are influenced by excision pattern and mass size [24].

Currently, establishing the best relationship possible between a mass and its surrounding tissues while considering the mobility of the region is critical as extensive surgery, including ostectomy if necessary is still the treatment of choice [12–14, 17, 24].

Cost implication and availability of CT and MRI scanners are the most important limitations in the integrating cross-sectional anatomy. Moreover these techniques require a training and assessment in the interpretation of post-mortem imaging. In fact, by understanding the benefit and limit of each imaging technique we can employ CT and MRI to their maximal advantages.