The project was approved by the Ethics Committee on the Use of Animals (CEUA) of the Faculty of Veterinary Medicine and Animal Science (FMVZ) at the University of São Paulo (USP), process 4143100220, and by the Federal University of Espírito Santo, process 31/2019, as well as by the Biodiversity Authorization and Information System (Sisbio), processes 31762-5, and 78227-3.

In this study, the pelvic and thoracic limbs of four roadkilled animals were used. The limbs were frozen shortly after the animals died and stored for up to a year. Before performing the procedures, the limbs were thawed at 7–10°C for 48 h.

CT and MRI scans were performed on the thoracic and pelvic limbs. The limbs were positioned in the lateral and ventral decubitus positions on the examination table in the most symmetrical manner, and scans were performed. The images were saved in the DICOM format for later analysis.

For CT, a Philips Brilliance device with 64 channels was used, calibrated at 120 kV and 332 mA with a 512 × 512 matrix. Two window settings were applied: for bones, window width (WW) of 4,095 and window level (WL) of 600; and for soft tissues, WW of 350 and WL of 70. The slice thickness was set to 1 mm.

For MRI, a SIEMENS 7 Tesla machine was used to acquire images with T2W sequences (TR: 500, TE: 86, matrix 640 × 320, WW 402, and WL 169). A coil was positioned over the limbs of the animal with a slice thicknesses of 0.6 mm.

The thoracic and pelvic limbs of three animals were sectioned and subjected to the silicone plastination technique through forced impregnation at a low temperature (−25°C), following the method described by von Hagens, Tiedemann, and Kriz (7).

To analyze shrinkage in the plastinated specimens, weight (g) was measured on a calibrated scale, volume (mL) in graduated vats, and thickness (cm) using an analog caliper at the four edges of the sections (mean values of the four extremities were used). Measurements were performed at two time points: t1 (before dehydration) and t2 (after curing). The arithmetic mean of each variable was calculated. The difference in means between t1 and t2 was calculated with a 95% confidence interval.

Prior to applying the paired Student’s t-test, data normality was assessed using the Shapiro–Wilk W test. A paired Student’s t-test was applied to the samples, reporting one-tailed p-values to test for shrinkage, where Ha: t1 > t2, indicating that the difference between t1 and t2 was greater than zero for volume, weight, and thickness. Statistical significance was set at p < 0.05. Additionally, percentage reduction for each sample was calculated as (t2 − t1) ÷ t1 × 100, and the mean reduction and standard deviation were reported.

The proximal surface of each segment was photographed using a Canon EOS Rebel T3i digital camera with an EF-S 18–135 mm lens. The images were correlated and identified with CT and MRI scans using the RadiAnt DICOM Viewer (12) and 3D Slicer (13) software. These programs have been employed for multiplanar reconstruction, 3D rendering, filtering, and segmentation of structures. Finally, virtual dissection and identification of the anatomical structures in each correlated image were conducted according to Veterinary Anatomical Nomina (14). All images were presented in a proximal view, with the left side of the image corresponding to the viewer’s left side for easier identification.

Ninety-three sections corresponding to the plastinated anatomical regions of the thoracic and pelvic limbs were examined. Table 1 shows an average reduction of 3.20 g and 2.59 g, respectively, with weight reductions of 29.80 and 25.12% in the right and left limbs, respectively. The average reduction for the right and left pelvic limbs was 3.40 g (23.53%) and 3.04 g (23.89%), respectively. The p-value was less than 0.001 for all four limbs (Table 1).

The data distribution in Figure 1A shows the maximum and minimum values for the anatomical regions, with the minimum being greater than zero and the maximum exceeding 30 g identified in the left pelvic limb. Outliers appear in all charts but are only present at the upper limit. The median shows variations but remains close to the first quartile in all analyzed regions (Figure 1A).

The weight reduction in the thoracic and pelvic section showed an average retention of 2.91 g (27.53%) and 3.22 g (23.71%), respectively. The combined average reduction for all thoracic and pelvic limb sections was 3.08 g, representing a total weight reduction of 25.41%. The p-value was less than 0.001 for the analysis of weight retraction of all thoracic and pelvic limb sections (Table 1). Figure 1B shows the maximum and minimum values for the anatomical regions, with a minimum greater than zero and a maximum exceeding 30 g identified in the pelvic limb. Outliers appear in all charts but are only present at the upper limit. The median shows variations but remains close to the first quartile in all analyzed regions.

The retraction volume in milliliters for the right and left antimeres showed a mean retention of 1.41 mL (14.72%) and 1.13 mL (12.80%), respectively. The average volume reduction for the right and left limbs was 2.37 mL (19.08%) and 1.60 mL (13.85%), respectively. The p-value was less than 0.001 for all four limbs (Table 2). Figure 1C shows the maximum and minimum values for the anatomical regions, with a minimum greater than zero and a maximum exceeding 30 mL identified in the left pelvic limb. Outliers appeared in the charts for the right thoracic and pelvic limbs, as well as the grand total but only at the upper limit. The median shows variations but remains close to the first quartile in all analyzed regions.

Table 2 shows the retraction volume of the thoracic and pelvic limbs in milliliters, corresponding to 1.28 mL (13.80%) and 1.97 mL (16.42%), respectively. The average reduction for all limbs was 1.66 mL (15.24%). The p-value was less than 0.001 for all analyses (Table 2).

Figure 1D shows the maximum and minimum values for the anatomical regions, with a minimum greater than zero and a maximum exceeding 30 mL identified in the pelvic limb. Outliers appeared in the charts but only at the maximum limit, except for the thoracic limb post-cure analysis, where no outliers were present. The median shows variations but remains close to the first quartile in all analyzed regions.

Table 3 shows the reduction in thickness of the thoracic limb for both right and left antimeres, with a mean reduction of 0.06 cm (6.40%) and 0.05 cm (5.14%), respectively. The mean reduction for the right and left pelvic limbs was 0.06 cm (5.34%) and 0.07 cm (6.93%), respectively. The p-values were 0.002 for the right thoracic limb, 0.000 for the left thoracic limb, and less than 0.001 for the right and left pelvic limbs (Table 3).

The data distribution in Figure 1E shows the maximum and minimum values for the anatomical regions, with a minimum >0.5 cm for both pelvic limbs and the total for all limbs and a minimum greater than 1.0 cm for both thoracic limbs. The maximum exceeded 1.0 cm in all the analyses. Outliers appeared in all chart analyses. The median shows variations but is generally close to the third quartile in the analyses (Figure 1E).

The reduction in thickness of the thoracic and pelvic limb sections showed an average reduction of 0.05 cm (5.80%) and 0.07 cm (6.15%), respectively. The average reduction for all limbs was 0.06 cm (5.99%). The p-value was less than 0.001 for all analyses (Table 3). Figure 1F shows the maximum and minimum values for the anatomical regions, with minimum greater than 1.0 cm for the thoracic limb and 0.5 cm for the pelvic limb. For all limbs, the maximum exceeded 1.0 cm in all analyses except during the post-healing period for the thoracic limb. Outliers appeared in all the Figures. The median shows variations across all the analyzed regions.

The anatomical structures can be identified by correlating plastination techniques with the images obtained from the proximal surface of the plastinated sections and the corresponding CT and MRI scans (Figures 2, 3). Three-dimensional reconstruction from volumetric CT scans enabled the identification of anatomical structures and provided a three-dimensional view of the thoracic and pelvic limbs as well as the animal’s skeleton (Supplementary Videos 1, 2), with the articular regions highlighted in blue (Supplementary Videos 3, 4).

The brachialis and triceps brachii muscles were observed in the proximal region of the arm (Figure 2A), including their long, medial, lateral, and accessory heads. Because this is the most proximal region of the limb, the superficial and deep pectoral muscles, cutaneous muscle of the trunk, and large dorsal muscle are also visible in the plastinated section as well as in the CT and MRI scans. The proximal third of the humerus was visible near the humeral neck region. Owing to the varying tissue densities on the CT and MRI images and in the plastinated metamer, muscle surfaces were clearly delineated, facilitating the distinction of each structure in this region (Figure 2A).

The accessory, long, lateral, and medial heads of the triceps brachii extend into the distal third of the arm. In this region, the biceps brachii muscle exhibits a distinctive morphology, making it easily identifiable in the plastinated specimen as well as in CT and MRI scans. The brachialis muscle, which wraps around the humerus, also appears larger than that described in the previous section, and is easily distinguished using imaging techniques. Regarding vascularization, major vessels, such as the cephalic vein, brachial artery, and vein, were clearly identifiable in all imaging modalities (Figure 2B).

In the forearm, the biceps brachii muscle was identified in the plastinated specimen as well as in the CT and MRI scans, positioned lateral to the brachialis muscle, followed by the extensor carpi radialis muscle. The humeral condyle is distinctly visible along the lateral and medial epicondyles at the apex. The olecranon lies posterior to the humerus, with muscles such as the anconeus muscle laterally and the flexor carpi ulnaris muscle medially. The vascular structures were also well visualized, with the cephalic vein appearing superficial and cranial, whereas the median vein was observed in the medio caudal region (Figure 2C).

In the proximal third of the forearm, the extensor muscles of the fingers are positioned dorsally and dorsolaterally, with the extensor carpi radialis being particularly prominent. Flexor muscles, including the deep and superficial flexors of the fingers, are also evident in this region and are easily differentiated on MRI and plastinated sections. The radius and ulna support the leverage of the forearm and are identifiable on CT, MRI, and plastinated sections (Figures 2D–E).

In the proximal third of the hand, the carpal bones articulated closely with the radial and ulnar bones and distally with the metacarpals. The accessory carpal bones were arranged caudally, while the radial carpal bones were arranged cranially. The superficial and deep flexor muscles of the fingers are located lateral to the accessory carpal bone. Muscle structures were well identified in the plastinated specimen and on MRI. However, on CT, the bone and vascular aspects were more evident (Figure 2E).

The metacarpal bones were arranged proximally and distally, with the second to fourth metacarpals aligned mediolaterally. The first metacarpal was located medial and caudal to the other bones. The sesamoid bone was positioned adjacent and caudal to the first metacarpal and was identifiable on CT images (Figure 2F).

In distal sections of the hand, the proximal phalanges of the second, third, fourth, and fifth digits were visible, with the metacarpal pad located caudally in the palmar region. The visualization and identification of these structures were enhanced by the correlation of the three imaging techniques, each offering complementary perspectives.

The pelvic limb begins with the cranial sartorius muscle, followed by the quadriceps femoris, which includes the vastus lateralis, medialis, intermedius, and rectus femoris muscles. The femur is the sole bony structure supporting this section of the limb, whereas the femoral artery and vein provide caudomedial vascularization. The adductor muscle is located caudal and lateral to the biceps femoris, whereas the semimembranosus, semitendinosus, and gracilis muscles occupy the caudal region (Figure 3A).

In plastinated sections, on CT and MRI, the patella articulates with the femur, and its contours are well-defined and clearly identified. The lateral and medial patellofemoral ligaments support the patellar syntopy. The biceps femoris muscle is located in the lateral region of the section, followed by the caudal crural abductor muscles. In the medial region, the sartorius muscle was visualized cranially. Following the caudal direction of the humerus, we identified the medial head of the gastrocnemius muscle on its caudal surface. The distal femoral caudal arteries and veins are located in the caudal region between the biceps femoris and the semimembranosus muscles. The gracilis extends through the medial region of the section to its caudal apex, where the semitendinosus is located (Figure 3B).

The tibia and fibula constitute the main bone components in the proximal region of the leg as identified on the plastinated image, CT, and MRI of the cranial section. The patellar ligament is visible on the cranial aspect of the tibia. The fibula is positioned caudolateral to the tibia. The cranial tibialis and extensor tendons of the extensor digitorum longus muscle are located lateral to the tibia and contralateral to the sartorius muscle, respectively. The popliteal and flexor digitorum superficialis muscles are located caudal to the tibia. Caudal to the fibula lies the flexor digitorum longus muscle, which is associated with the first digit, followed by the lateral head of the gastrocnemius and biceps femoris muscles, extending laterally across the section. The medial head of the gastrocnemius muscle was also present, followed by the semitendinosus muscle caudomedially, whereas the gracilis muscle extended throughout the medial region of the section (Figure 3C).

In the distal region of the leg, plastinated sections, CT, and MRI revealed that the cranial tibialis muscle was positioned cranially, followed by the extensor digitorum longus and peroneus longus muscles. The tibia was located medially, whereas the fibula was in an intermediate position. Caudal to the fibula lies the flexor digitorum longus, which is associated with the first digit, followed by the flexor digitorum superficialis and the gastrocnemius muscles. The calcaneal tendon is the most caudal structure. The lateral saphenous vein was identified caudolaterally on CT and MRI (Figure 3D).

In the proximal region of the foot, vascularization is observed in the cranial section, represented by the lateral and medial saphenous veins, which belong to the cranial branches. The deepest vein follows the dorsal artery of the foot caudally, with clearer visualization on CT and MRI because of attenuation of the structures. The cranial tibialis and extensor digitorum longus muscles were positioned in the cranial region of the section, followed by the talus and calcaneus. The medial region, from cranial to caudal, includes the medial collateral ligament, medial flexor digitorum, and flexor digitorum lateralis. The lateral region, visualized cranially to caudally, contained the lateral digital extensor muscle, lateral collateral ligament, and the caudally located collateral ligament within the deep digital flexor muscle section (Figure 3E).

In the cranial metatarsal region, as observed in the plastinated CT and MRI sections, the second, third, fourth, and fifth metatarsals were followed by the caudal interosseous muscles. The common dorsal digital vein is located cranial to the metatarsal region. The deep digital flexor muscle is the most cranial section, followed by the superficial digital flexor muscle (Figure 3F). In the distal region of the foot, the medial phalanges of the fourth and fifth digits, proximal phalanx of the third digit, and distal phalanx of the second digit were identified in the plastinated section, CT, and MRI images. The deep digital flexors are caudal to the phalanges. The ungual processes of the distal phalanx were identified in the second and fifth digits. MRI allows identification of the metatarsal pad (Figure 3G).