Sonographic callus may enable assessment of fracture healing. The aim of this study was to establish a reliable method for three-dimensional reconstruction of sonographic callus.

Patients that underwent non-operative management of displaced midshaft clavicle fractures and intramedullary nailing of tibia fractures were prospectively recruited and followed to union. Ultrasound scanning was performed at periodical time points following injury. Infra-red tracking technology was used to map each image to a three-dimensional lattice. Criteria was fist established for two-dimensional bridging callus detection in a pilot study. Using echo intensity of the ultrasound image, semi-automated mapping was used to create an anatomic three-dimensional representation of fracture healing. Agreement on the presence of sonographic bridging callus was assessed using the kappa coefficient and intra-class-correlation (ICC) between observers.

112 clavicle fractures and 10 tibia fractures completed follow-up at six months. Sonographic bridging callus was detected in 62.5% (n=70/112) of the clavicles at six weeks post-injury. If present, union occurred in 98.6% of the fractures (n=69/70). If absent, nonunion developed in 40.5% of cases (n=17/42)(73.4%-sensitive and 100%-specific to predict union). Out of 10 tibia fractures, 7 had bridging callus of at least one cortex at 6 weeks and when present all united. Of the three patients lacking sonographic bridging callus, one went onto a nonunion (77.8%-sensitive and 100%-specific to predict union). The ICC for sonographic callus between four reviewers was 0.82 (95% CI 0.68–0.91)

Three-dimensional ultrasound reconstruction of bridging callus has the potential to identify impaired fracture healing at an early stage in fracture management.

X-ray is the standard method for monitoring fracture healing however it is not ideal; signs of healing are not normally visible on X-ray until around 6–8 weeks post fracture. Ultrasonography allows the detection of both the initial haematoma, usually formed immediately after fracture, and the small calcium deposits laid down between broken bone ends in the first stages of fracture healing. It has been reported that these early indicators of the healing process are visible as early as 1–2 weeks after fracture. We use Freehand 3D Ultrasound to monitor the early stages of fracture healing as both the bone surface and surrounding soft tissues can be imaged simultaneously.

The Freehand 3D Ultrasound system consists of a standard Ultrasound machine, a PC running STRAD-WIN (Medical Imaging Group, Cambridge University) 3D software, and an optical tracking devise (NDI Polaris) to record the position and orientation of the Ultrasound probe during scanning. Images are transferred from the Ultrasound machine to the PC using RF capture through out a scan. Calibrating the system matches up the correct image with the correct probe position to produce a 3D dataset.

We segment features of interest on the sequence of 2D images to construct a 3D model. These models are rotatable and provide views of the scanned anatomy that are not otherwise achievable using conventional Ultrasound or X-ray. The 3D data set can also be resliced through any plane to provide further views.

To conduct a 3D Ultrasound scan takes the same amount of time as a conventional 2D scan. The production of the 3D model takes between 15–60 minutes depending on the level of detail required. Distances are measurable to within ±0.4mm meaning fracture gaps of sub-millimeter width can be resolved. The system has already been evaluated on healthy volunteers and a clinical study currently underway.

An uncomplicated, quantitative method of determining density from X-rays would be of extreme value to clinicians. In this study we perform a thorough assessment of applying a step wedge to grey level calibration method to X-rays obtained using Computed Radiography (CR).

An Aluminium step wedge of ten, 5mm-thick steps was X-rayed with a Fuji CR system together with a knee phantom (3M) at various energy and Fuji processing settings. Automatic detection of the steps by means of the Hough transform was used to assess optimum CR settings. Background variation due to the anode Heel effect was evaluated by acquiring an “empty field” X-ray at different energy settings and with copper filtering. The effects of beam hardening were considered with a custom-made phantom which was also used to assess correcting for soft tissue and bone thickness.

X-rays taken at higher energy settings and with wider windowing imaged the widest number of steps (nine) and gave the best accuracy in modelling the step thickness to grey level relationship. Fitting a straight line to the log of the net grey levels gives an excellent model of the data (R2 = 0.99). X-rays of copper sheeting show that automatic histogram analysis is performed by the Fuji CR system, which can have unpredictable effects on aluminium thickness to grey level relationship. Background variation in the anode-cathode direction due to the Heel effect was corrected with a 1D exponential model (R2 = 0.99), allowing position-independent measurements to be obtained. Correcting for bone thickness, soft tissue and beam hardening further improves measurement quality.

Use of step wedge calibration to provide quantitative information on plain X-rays without altering their clinical quality is possible using digital radiography. However, a thorough assessment of the entire X-ray process is necessary to achieve accurate and comparable information.

Imaging of the musculoskeletal system is vital for delivering optimum treatment particularly in the assessment of fracture healing. X-ray and CT are adequate imaging methods for bone but, soft tissue needs other modalities such as MRI and Ultrasound. We propose the use of Freehand 3D Ultrasound to study the early stages of fracture healing by imaging the bone surfaces around the fracture site and monitoring changes in the surrounding soft tissue.

Freehand 3D ultrasound is acquired by attaching a position sensor to the probe of a conventional 2D diagnostic ultrasound machine. As the probe is moved, its position and orientation are recorded along with the 2D ultrasound images. This enables slices through the body to be viewed that would be inaccessible using a normal ultrasound system. Bone surfaces around a fracture site are scanned and the data reconstructed using the Stradx and Stradwin software developed by Cambridge University, to give a 3D visualization of the area.

To assess the feasibility of this proposed method the lower limbs of healthy volunteers were scanned using a 5–10MHz ultrasound probe. The scanning resolution of the system was evaluated using a phantom to ensure millimetre detail could be detected as would be required for imaging early fracture healing. It was found that detail down to 0.8mm could easily be resolved for measurement.

The 3D system could accurately profile the different soft tissue interfaces. The visible surfaces of the tibia were reconstructed to give 3D models. Additional layers of soft tissue interfaces could easily be added to these models to provide more detail.

This imaging modality can provided detailed 3D models of bone the bone surface and surrounding soft tissue. As ultrasound is non-ionizing, rescanning can be conducted more frequently than with CT or x-ray thus offering a more accurate assessment of a patient’s response to healing.