Human error is usually evaluated using statistical descriptions during radiographic annotation. The technological advances popularized the “non-human” landmarking techniques, such as deep learning, in which the error is presented in a confidence format that is not comparable to that of the human method. The region-based landmark definition makes an arbitrary “ground truth” point impossible. The differences in patients’ anatomies, radiograph qualities, and scales make the horizontal comparison difficult. There is a demand to quantify the manual landmarking error in a probability format. Taking the measurement of pelvic tilt (PT) as an example, this study recruited 115 sagittal pelvic radiographs for the measurement of two PTs. We proposed a method to unify the scale of images that allows horizontal comparisons of landmarks and calculated the maximum possible error using a density vector. Traditional descriptive statistics were also applied. All measurements showed excellent reliabilities (intraclass correlation coefficients > 0.9). Eighty-four measurements (6.09%) were qualified as wrong landmarks that failed to label the correct locations. Directional bias (systematic error) was identified due to cognitive differences between observers. By removing wrong labels and rotated pelves, the analysis quantified the error density as a “good doctor” performance and found 6.77°-11.76° maximum PT disagreement with 95% data points. The landmarks with excellent reliability still have a chance (at least 6.09% in our case) of making wrong landmark decisions. Identifying skeletal contours is at least 24.64% more accurate than estimating landmark locations. The landmark at a clear skeletal contour is more likely to generate systematic errors. Due to landmark ambiguity, a very careful surgeon measuring PT could make a maximum 11.76° random difference in 95% of cases, serving as a “good doctor benchmark” to qualify good landmarking techniques

Plots are an elegant and effective way to represent data. At their best they encourage the reader and promote comprehension. A graphical representation can give a far more intuitive feel to the pattern of results in the study than a list of numerical data, or the result of a statistical calculation.

The temptation to exaggerate differences or relationships between variables by using broken axes, overlaid axes, or inconsistent scaling between plots should be avoided.

A plot should be self-explanatory and not complicated. It should make good use of the available space. The axes should be scaled appropriately and labelled with an appropriate dimension.

Plots are recognised statistical methods of presenting data and usually require specialised statistical software to create them. The statistical analysis and methods to generate the plots are as important as the methodology of the study itself. The software, including dates and version numbers, as well as statistical tests should be appropriately referenced.

Following some of the guidance provided in this article will enhance a manuscript.

Cite this article: Bone Joint J 2015;97-B:1593–1603.

Evaluation of patient specific spinopelvic mobility requires the detection of bony landmarks in lateral functional radiographs. Current manual landmarking methods are inefficient, and subjective. This study proposes a deep learning model to automate landmark detection and derivation of spinopelvic measurements (SPM). A deep learning model was developed using an international multicenter imaging database of 26,109 landmarked preoperative, and postoperative, lateral functional radiographs (HREC: Bellberry: 2020-08-764-A-2). Three functional positions were analysed: 1) standing, 2) contralateral step-up and 3) flexed seated. Landmarks were manually captured and independently verified by qualified engineers during pre-operative planning with additional assistance of 3D computed tomography derived landmarks. Pelvic tilt (PT), sacral slope (SS), and lumbar lordotic angle (LLA) were derived from the predicted landmark coordinates. Interobserver variability was explored in a pilot study, consisting of 9 qualified engineers, annotating three functional images, while blinded to additional 3D information. The dataset was subdivided into 70:20:10 for training, validation, and testing. The model produced a mean absolute error (MAE), for PT, SS, and LLA of 1.7°±3.1°, 3.4°±3.8°, 4.9°±4.5°, respectively. PT MAE values were dependent on functional position: standing 1.2°±1.3°, step 1.7°±4.0°, and seated 2.4°±3.3°, p< 0.001. The mean model prediction time was 0.7 seconds per image. The interobserver 95% confidence interval (CI) for engineer measured PT, SS and LLA (1.9°, 1.9°, 3.1°, respectively) was comparable to the MAE values generated by the model. The model MAE reported comparable performance to the gold standard when blinded to additional 3D information. LLA prediction produced the lowest SPM accuracy potentially due to error propagation from the SS and L1 landmarks. Reduced PT accuracy in step and seated functional positions may be attributed to an increased occlusion of the pubic-symphysis landmark. Our model shows excellent performance when compared against the current gold standard manual annotation process

Trauma surgeries in the pelvic area are often difficult and prolonged processes that require comprehensive preoperative planning based on a CT scan. Preoperative planning is essential for the appreciation and spatial visualisation of the bone fragments, for planning the reduction strategy, and for determining the optimal type, size, and location of the fixation hardware. We have developed a novel haptic-based patient specific preoperative planning system for pelvic bone fractures surgery planning. The system provides a virtual environment in which 3D bone fragments and fixation hardware models are interactively manipulated with full spatial depth and tactile perception. It supports the choice of the surgical approach and the planning of the two mains steps of bone fracture surgery: reduction and fixation. The purpose of the tool is to provide an intuitive haptic spatial interface for the manipulation of bone fracture 3D models extracted from CT images, to support the selection of bone fragments, the annotation of the fracture surface, the selection and placement of fixation screws, and the creation and placement of fixation plates with an anatomically fit shape. The system incorporates ligament models that constrain the bone fragments motions and provides a realistic interactive fracture reduction support feeling to the surgeon. It allows the surgeon to view the fracture from various directions, thereby allowing fast and accurate fracture reduction planning. Two haptic devices, one for each hand, provide tactile feedback when objects touch without interpenetrating. To facilitate the reduction, the system provides an interactive, haptic fracture surface annotation tool and a fracture reduction algorithm that automatically minimises the pairwise distance between the fracture surfaces. For fracture fixation, the system provides a screw creation and placement capability as well as custom anatomical-fit fixation plate creation and placement. The screw placement is facilitated by the transparent viewing mode that allows the surgeon to navigate the screws inside the bone fragments while constraining them to remain within the bone fragments with haptic forces. Our experimental results with five surgeons show that the method allows highly accurate reduction planning to within 1 mm or less. To evaluate the alignment in terms of quantity, we created a model of an artificial fracture in a healthy pelvis bone. The created model is placed in its anatomic location thus allowing us to measure the error in relation to its initial position. We calculate the anatomic alignment error by measuring the Hausdorff distance in mm between the fragment positioned in the desired location and the fragment placed by the surgeon. The new haptic-based system also supports patient-specific training of pelvic fracture surgeries

The ability of mesenchymal stem cells (MSCs) to differentiate in vitro into chondrocytes, osteocytes and myocytes holds great promise for tissue engineering. Skeletal defects are emerging as key targets for treatment using MSCs due to the high responsiveness of bone to interventions in animal models. Interest in MSCs has further expanded in recognition of their ability to release growth factors and to adjust immune responses. Despite their increasing application in clinical trials, the origin and role of MSCs in the development, repair and regeneration of organs have remained unclear. Until recently, MSCs could only be isolated in a process that requires culture in a laboratory; these cells were being used for tissue engineering without understanding their native location and function. MSCs isolated in this indirect way have been used in clinical trials and remain the reference standard cellular substrate for musculoskeletal engineering. The therapeutic use of autologous MSCs is currently limited by the need for ex vivo expansion and by heterogeneity within MSC preparations. The recent discovery that the walls of blood vessels harbour native precursors of MSCs has led to their prospective identification and isolation. MSCs may therefore now be purified from dispensable tissues such as lipo-aspirate and returned for clinical use in sufficient quantity, negating the requirement for ex vivo expansion and a second surgical procedure. In this annotation we provide an update on the recent developments in the understanding of the identity of MSCs within tissues and outline how this may affect their use in orthopaedic surgery in the future. Cite this article: Bone Joint J 2014;96-B:291–8

Aims

This study explored the shared genetic traits and molecular interactions between postmenopausal osteoporosis (POMP) and sarcopenia, both of which substantially degrade elderly health and quality of life. We hypothesized that these motor system diseases overlap in pathophysiology and regulatory mechanisms.