Aims. The purpose of this study was to develop a convolutional neural network (CNN) for fracture detection, classification, and identification of greater tuberosity displacement ≥ 1 cm, neck-shaft angle (NSA) ≤ 100°, shaft translation, and articular fracture involvement, on plain radiographs. Methods. The CNN was trained and tested on radiographs sourced from 11 hospitals in Australia and externally validated on radiographs from the Netherlands. Each radiograph was paired with corresponding CT scans to serve as the reference standard based on dual independent evaluation by trained researchers and attending orthopaedic surgeons. Presence of a fracture, classification (non- to minimally displaced; two-part, multipart, and glenohumeral dislocation), and four characteristics were determined on 2D and 3D CT scans and subsequently allocated to each series of radiographs. Fracture characteristics included greater tuberosity displacement ≥ 1 cm, NSA ≤ 100°, shaft translation (0% to < 75%, 75% to 95%, > 95%), and the extent of articular involvement (0% to < 15%, 15% to 35%, or > 35%). Results. For detection and classification, the algorithm was trained on 1,709 radiographs (n = 803), tested on 567 radiographs (n = 244), and subsequently externally validated on 535 radiographs (n = 227). For characterization, healthy shoulders and glenohumeral dislocation were excluded. The overall accuracy for fracture detection was 94% (area under the receiver operating characteristic curve (AUC) = 0.98) and for classification 78% (AUC 0.68 to 0.93). Accuracy to detect greater tuberosity fracture displacement ≥ 1 cm was 35.0% (AUC 0.57). The CNN did not recognize NSAs ≤ 100° (AUC 0.42), nor fractures with ≥ 75% shaft translation (AUC 0.51 to 0.53), or with ≥ 15% articular involvement (AUC 0.48 to 0.49). For all objectives, the model’s performance on the external dataset showed similar accuracy levels. Conclusion. CNNs proficiently rule out proximal humerus fractures on plain radiographs. Despite rigorous training methodology based on CT imaging with multi-rater consensus to serve as the reference standard, artificial intelligence-driven classification is insufficient for clinical implementation. The CNN exhibited poor diagnostic ability to detect greater tuberosity displacement ≥ 1 cm and failed to identify NSAs ≤ 100°, shaft translations, or articular fractures. Cite this article: Bone Joint J 2024;106-B(11):1348–1360

We describe the routine imaging practices of Level 1 trauma centres for patients with severe pelvic ring fractures, and the interobserver reliability of the classification systems of these fractures using plain radiographs and three-dimensional (3D) CT reconstructions. Clinical and imaging data for 187 adult patients (139 men and 48 women, mean age 43 years (15 to 101)) with a severe pelvic ring fracture managed at two Level 1 trauma centres between July 2007 and June 2010 were extracted. Three experienced orthopaedic surgeons classified the plain radiographs and 3D CT reconstruction images of 100 patients using the Tile/AO and Young–Burgess systems. Reliability was compared using kappa statistics. A total of 115 patients (62%) had plain radiographs as well as two-dimensional (2D) CT and 3D CT reconstructions, 52 patients (28%) had plain films only, 12 (6.4%) had 2D and 3D CT reconstructions images only, and eight patients (4.3%) had no available images. The plain radiograph was limited to an anteroposterior pelvic view. Patients without imaging, or only plain films, were more severely injured. A total of 72 patients (39%) were imaged with a pelvic binder in situ. Interobserver reliability for the Tile/AO (Kappa 0.10 to 0.17) and Young–Burgess (Kappa 0.09 to 0.21) was low, and insufficient for clinical and research purposes. Severe pelvic ring fractures are difficult to classify due to their complexity, the increasing use of early treatment such as with pelvic binders, and the absence of imaging altogether in important patient sub-groups, such as those who die early of their injuries. Cite this article: Bone Joint J 2013;95-B:1396–1401

We evaluated the impact of stereo-visualisation of three-dimensional volume-rendering CT datasets on the inter- and intraobserver reliability assessed by kappa values on the AO/OTA and Neer classifications in the assessment of proximal humeral fractures. Four independent observers classified 40 fractures according to the AO/OTA and Neer classifications using plain radiographs, two-dimensional CT scans and with stereo-visualised three-dimensional volume-rendering reconstructions. Both classification systems showed moderate interobserver reliability with plain radiographs and two-dimensional CT scans. Three-dimensional volume-rendered CT scans improved the interobserver reliability of both systems to good. Intraobserver reliability was moderate for both classifications when assessed by plain radiographs. Stereo visualisation of three-dimensional volume rendering improved intraobserver reliability to good for the AO/OTA method and to excellent for the Neer classification. These data support our opinion that stereo visualisation of three-dimensional volume-rendering datasets is of value when analysing and classifying complex fractures of the proximal humerus

Introduction. We aim to assess whether radiographic characteristics of the greater tuberosity fragment can predict rotator cuff tears inpatients with anterior shoulder dislocations combined with an isolated fracture of the greater tuberosity. Methods. A retrospective single-centre case series of 61 consecutive patients that presented with anterior shoulder dislocations combined with an isolated fracture of the greater tuberosity between January 2018 and July 2022. Inclusion criteria: patients with atraumatic anterior shoulder dislocation associated with an isolated fracture of the greater tuberosity with a minimum follow-up of 3-months. Exclusion criteria: patients with other fractures of the proximal humerus or glenoid. Rotator cuff tears were diagnosed using magnetic resonance or ultrasound imaging. Greater tuberosity fragment size and displacement was calculated on plain radiographs using validated methods. Results. The case series was composed of 22 men and 39 women with a mean age of 65 years (29 - 91 years). The mean follow-up was 15months and median follow up 8.5 months (3 – 60 months). A rotator cuff tear was diagnosed in 14 patients (16%) and involved the supraspinatus (13), infraspinatus (4) and subscapularis (2). Full-thickness tears occurred in 6 patients and partial-thickness tears in 8patients. The mean time from initial injury to rotator cuff tear diagnosis was 5 months (2 – 22 months). The mean greater tuberosity fragment length was 23.4 mm in rotator cuff tear patients versus 32.6 mm in those without a tear (p = 0.006, CI: -15 - -2). The mean greater tuberosity. fragment width was 11.1 mm in rotator cuff tear patients versus 17.8 mm in those without a tear (p = 0.0004, CI: -10 - -2). There was no significant difference in the super inferior and anteroposterior fragment displacement between the two groups. Conclusion. In patients with shoulder dislocations combined with an isolated fracture of the greater tuberosity, rotator cuff tears are associated with a smaller sized greater tuberosity fragment

An MR scan was performed on all patients who presented to our hospital with a clinical diagnosis of a fracture of the proximal femur, but who had no abnormality on plain radiographs. This was a prospective study of 102 consecutive patients over a ten-year period. There were 98 patients who fulfilled our inclusion criteria, of whom 75 were scanned within 48 hours of admission, with an overall mean time between admission and scanning of 2.4 days (0 to 10). A total of 81 patients (83%) had abnormalities detected on MRI; 23 (23%) required operative management. The use of MRI led to the early diagnosis and treatment of occult hip pathology. We recommend that incomplete intertrochanteric fractures are managed non-operatively with protected weight-bearing. The study illustrates the high incidence of fractures which are not apparent on plain radiographs, and shows that MRI is useful when diagnosing other pathology such as malignancy, which may not be apparent on plain films

Interobserver reliability of the AO system of classification of fractures of the distal radius was assessed using plain radiographs and CT. Five observers classified 30 Colles’-type fractures using only plain radiographs; two months later they were reclassified using CT in addition. Interobserver reliability was poor in both series when detailed classification was used. By reducing the categories to five, interobserver reliability was slightly improved, but was still poor. When only two AO types were used, the reliability was moderate using plain radiographs and good to excellent with the addition of CT. The use of CT as well as plain radiographs brings interobserver reliability to a good level in assessment of the presence or absence of articular involvement, but is otherwise of minor value in improving the interobserver reliability of the AO system of classification of fractures of the distal radius

Mirels’ score predicts the likelihood of sustaining pathological fractures using pain, lesion site, size and morphology. The aim is to investigate its reproducibility, reliability and accuracy in upper limb bony metastases and validate its use in pathological fracture prediction. A retrospective cohort study of patients with upper limb metastases, referred to an Orthopaedic Trauma Centre (2013–18). Mirels’ was calculated in 32 patients; plain radiographs at presentation scored by 6 raters. Radiological aspects were scored twice by each rater, 2-weeks apart. Inter- and intra-observer reliability were calculated (Fleiss’ kappa test). Bland-Altman plots compared variances of individual score components &total Mirels’ score. Mirels’ score of ≥9 did not accurately predict lesions that would fracture (11% 5/46 vs 65.2% Mirels’ score ≤8, p<0.0001). Sensitivity was 14.3% &specificity was 72.7%. When Mirels’ cut-off was lowered to ≥7, patients were more likely to fracture (48% 22/46 versus 28% 13/46, p=0.045). Sensitivity rose to 62.9%, specificity fell to 54.6%. Kappa values for interobserver variability were 0.358 (fair, 0.288–0.429) for lesion size, 0.107 (poor, 0.02–0.193) for radiological appearance and 0.274 (fair, 0.229–0.318) for total Mirels’ score. Values for intraobserver variability were 0.716 (good, 95% CI 0.432–0.999) for lesion size, 0.427 (moderate, 95% CI 0.195–0.768) for radiological appearance and 0.580 (moderate, 0.395–0.765) for total Mirels’ score. We showed moderate to substantial agreement between &within raters using Mirels’ score on upper limb radiographs. Mirels’ has poor sensitivity &specificity predicting upper limb fractures - we recommend the cut-off score for prophylactic surgery should be lower than for lower limb lesions

Fractures of the distal humeral articular surface which do not involve the medial and lateral columns are often more extensive than is apparent from plain radiographs. This retrospective study describes the epidemiology of this injury using modern classification systems and compares pre-operative radiography with operative findings. The study group included 79 patients with a mean age of 47 years (13 to 91). The annual incidence was 1.5 per 100 000 population, and was highest in women over the age of 60. The majority of the fractures (59; 75%) were sustained in falls from standing height. Young males tended to sustain more high-energy injuries with more complex fracture patterns. In 24% of cases (19) there was a concomitant radial head fracture. Classification from plain radiographs often underestimates the true extent of the injury and computed tomography may be of benefit in pre-operative planning, especially in those over 60 years of age

Aim. The aim of this study was to determine whether the absence of periosteal reaction on plain radiographs was predictive of exchange nail failure in lower limb diaphyseal fracture non-unions. Methods. A consecutive cohort of 20 femora and 35 tibiae undergoing exchange nailing for diaphyseal aseptic (n=39) and septic (n=16) fracture non-union at a single centre from 2003 to 2010. Multiple causes of non-union were found in 29 patients (53%) with infection present in 16 cases (29.1%). Of this cohort 49 fracture non-unions had complete radiographic records (19 femora and 30 tibiae) allowing evaluation of the periosteal callus. The primary outcome was the number of number of revision procedures required to achieve union. Failure was defined was as the requirement of >two revision procedures to achieve union. Results. The absence of periosteal callus from the fracture site on all four cortices conferred a relative risk ratio (RRR) 5.00 (p=0.006) of exchange nail failure in septic fracture non-unions. Receiver operator characteristic curve analysis of number of cortices with periosteal reaction for predicting exchange nail (>two exchange nail procedures) failure in both septic and aseptic cases found an area under the curve of 0.79 (95% confidence interval 0.675–0.904, p<0.0001). The complete absence of callus within 5 mm of the fracture site had a 75% chance of requiring 3 or more exchange nails to obtain union. Conversely the presence periosteal callus on at least one cortex within 5mm of the fracture site had a 100% chance the fracture would unite following 1 or 2 exchange nail procedures. Conclusion. Careful radiographic analysis of lower limb diaphyseal non-union should be undertaken before exchange nailing. The absence of periosteal reaction on plain radiographs in septic fracture non-union is predictive of exchange nail failure

Fracture classification of femoral trochanteric fracture is usually based on plain X-ray. However, complications such as delayed union, non-union, and cut out are seen in stable fracture on X-ray. In this study, fracture was classified by 3D-CT and relationship to X-ray classification was investigated. 48 femoral trochanteric fractures (15 males, 33 female, average age: 82.6) treated with PFNA-II were investigated. Fracture was classified to 2part, 3part(5 subgroups), and 4part with combination of 4 fragments in CT; Head (H), Greater trochanter (G), Lesser trochanter (L), and Shaft (S). 5 subgroups of 3 part fracture were (1) H+G (S: small fragment) + L-S, (2) H + G (B:big fragment) + L-S, (3) H + G-L + S, (4) H + G (W:whole) + S, and (5) H + L + G-S. Numbers of each group were as follows; 2 part: 11, 3 part (1) : 7, 3 part (2) : 12, 3 part (3) : 10, 3 part (4) : 2, 3 part (5) : 3, 4 part : 3. 3 part (3), (4), (5) and 4 part are considered as unstable, however, 6 cases in these groups were classified in A1–1 or A1–2 stable fracture in AO classification. 10 fractures in Evans and 5 fractures in Jensen classification classified as stable were unstable in CT evaluation. It is sometimes very difficult to classify the femoral trochanteric fracture by plain X-ray. Classification with 3D-CT is very useful to distinguish which fracture is stable or unstable

We treated 50 patients with fractures of the waist of the scaphoid in a below-elbow plaster cast for up to 13 weeks. Displacement of the fragments was assessed independently by two observers using MRI and radiographs performed within two weeks of injury. The MRI assessments showed that only the measurement of sagittal translation of the fragments and an overall assessment of displacement had satisfactory inter- and intra-observer reproducibility and revealed that nine of the 50 fractures were displaced. Only three of the 49 fractures with adequate follow-up failed to unite, and all were displaced with more than 1 mm of translation in the sagittal plane. If the MRI assessment of displacement of the fracture was used as the measurement of choice, assessment of displacement on the initial scaphoid series of radiographs showed a sensitivity of between 33% and 47% and a positive predictive value of between 27% and 86%. Neither observer was able correctly to identify more than 33% to 47% of the displaced fractures from the plain radiographs. Although the overall assessment of displacement and gapping and translation in the coronal plane on the plain radiographs influenced the rate of union, none of these parameters identified all three fractures which failed to unite. We conclude that the assessment of displacement of scaphoid fractures on MRI can probably be used to assess the likelihood of union although the small number of nonunions limits the power of the study. In contrast, the assessment of displacement on routine radiography is inaccurate and of less value in predicting union

We have studied prospectively the effect of indomethacin on the development of heterotopic ossification (HO) after the internal fixation of acetabular fractures. After operation 107 patients randomly received either a six-week course of indomethacin or no treatment against HO. Plain radiographs of 101 patients at a mean of 7.9 months after surgery showed HO in 47.4% of the 57 patients who received indomethacin and in 56.8% of the 44 who did not. This difference was not statistically significant. Heterotopic ossification of Brooker class II or more was seen in four patients (7%) with prophylaxis and in one without (p = 0.51). Measurements of the volume of HO on 3-D CT reconstructions showed a median value of 1.5 cm. 3. in patients with indomethacin and 4.0 cm. 3. in those without (p = 0.28). When only the 57 patients in whom the operation was carried out through either a Kocher-Langenbeck or an extended iliofemoral approach were included the indomethacin group showed a median volume of 1.7 cm. 3. compared with 3.6 cm. 3. On plain radiographs Brooker class II or above was seen in 9.4% of the patients receiving indomethacin and in 4.8% of those who did not. Indomethacin was therefore not effective in preventing ectopic bone formation after surgery for acetabular fractures. There was a significant association of male gender with volume of HO using a non-parametric analysis of variance

We have examined the accuracy of reduction and the functional outcomes in elderly patients with surgically treated acetabular fractures, based on assessment of plain radiographs and CT scans. There were 45 patients with such a fracture with a mean age of 67 years (59 to 82) at the time of surgery. All patients completed SF-36 questionnaires to determine the functional outcome at a mean follow-up of 72.4 months (24 to 188). All had radiographs and a CT scan within one week of surgery. The reduction was categorised as ‘anatomical’, ‘imperfect’, or ‘poor’. Radiographs classified 26 patients (58%) as anatomical,13 (29%) as imperfect and six (13%) as poor. The maximum displacement on CT showed none as anatomical, 23 (51%) as imperfect and 22 (49%) as poor, but this was not always at the weight-bearing dome. SF-36 scores showed functional outcomes comparable with those of the general elderly population, with no correlation with the radiological reduction. Perfect anatomical reduction is not necessary to attain a good functional outcome in acetabular fractures in the elderly

Incomplete intertrochanteric fractures do not extend across to the medial femoral cortex and are stable, without rotational deformity or shortening of the lower limb. The aim of our study was to establish whether they can be successfully managed conservatively. A total of 68 patients over a five-year period presented with a suspected fracture of the femoral neck and underwent an MRI scan for further assessment. From these, we retrospectively reviewed eight patients with normal plain radiographs but with an incomplete, intertrochanteric fracture on MRI scan. Five were managed conservatively and three operatively. The mean length of hospital stay was 16 days for the conservatively-treated group and 15 days for those who underwent surgery; this was not statistically significant (p > 0.5) and all patients were mobilised on discharge. Although five patients were readmitted at a mean of 3.2 years after discharge, none had progressed to a complete fracture. We believe that patients with incomplete intertrochanteric fractures should be considered for conservative treatment

Objectives. Small animal models of fracture repair primarily investigate indirect fracture healing via external callus formation. We present the first described rat model of direct fracture healing. Methods. A rat tibial osteotomy was created and fixed with compression plating similar to that used in patients. The procedure was evaluated in 15 cadaver rats and then in vivo in ten Sprague-Dawley rats. Controls had osteotomies stabilised with a uniaxial external fixator that used the same surgical approach and relied on the same number and diameter of screw holes in bone. Results. Fracture healing occurred without evidence of external callus on plain radiographs. At six weeks after fracture fixation, the mean stress at failure in a four-point bending test was 24.65 N/mm. 2. (. sd. 6.15). Histology revealed ‘cutting-cones’ traversing the fracture site. In controls where a uniaxial external fixator was used, bone healing occurred via external callus formation. Conclusions. A simple, reproducible model of direct fracture healing in rat tibia that mimics clinical practice has been developed for use in future studies of direct fracture healing

Specific radiological features identified by Brandser and Marsh were selected for the analysis of acetabular fractures according to the classification of Letournel and Judet. The method employs a binary approach that requires the observer to allocate each radiological feature to one of two groups. The inter- and intra-observer variances were assessed. The presence of articular displacement, marginal impaction, incongruity, intra-articular fragments and osteochondral injuries to the femoral head were analysed by a similar method. These factors were termed ‘modifiers’ and are generally considered when planning operative intervention and, critically, they may influence prognosis. Six observers independently assessed 30 sets of plain radiographs and CT scans on two separate occasions, 12 weeks apart. They were asked to determine the presence or absence of specific radiological features. This simple binary approach to classification yields an inter- and intra-observer agreement which ranges from moderate to near-perfect (κ = 0.49 to 0.88 and κ = 0.57 to 0.88, respectively). A similar approach to the modifiers yields only slight to fair inter-observer agreement (κ = 0.20 to 0.34) and slight to moderate intra-observer agreement (κ = 0 to 0.55)

We have evaluated retrospectively the relationship of bony injuries seen on 106 consecutive MR scans in elderly patients of a mean age of 81.4 years (67 to 101) who were unable to bear weight after a low-energy injury. There were no visible fractures on plain radiographs of the hip but eight patients (7.5%) had fractures of the pubic ramus. In 43 patients (40.5%) MRI revealed a fracture of the femoral neck and in 26 (24.5%) there was a fracture of a pubic ramus. In 17 patients (16%) MRI showed an occult sacral fracture and all of these had a fracture of the pubic ramus. No patient with a fracture of the femoral neck had an associated fracture of the pelvic ring or vice versa. Occult fracture of the hip and of the pelvic ring appear to be mutually exclusive and if an acute fracture of the pubic ramus is diagnosed radiologically further investigations are not needed to rule out an occult fracture of the hip

Our aim was to evaluate the efficacy of a two-level reconstruction technique using subchondral miniscrews for the stabilisation of comminuted posterior-wall marginal acetabular fragments before applying lag screws and a buttress plate to the main overlying posterior fragment. Between 1995 and 2003, 29 consecutive patients with acute comminuted displaced posterior-wall fractures of the acetabulum were treated operatively using this technique. The quality of reduction measured from three standard plain radiographs was graded as anatomical in all 29 hips. The clinical outcome at a mean follow-up of 35 months (24 to 90) was considered to be excellent in five patients (17%), very good in 16 (55%), good in six (21%) and poor in two (7%). The use of the two-level reconstruction technique appears to provide stable fixation and is associated with favourable results in terms of the incidence of post-traumatic osteoarthritis and the clinical outcome. However, poor results may occur in patients over the age of 55 years

Methods:. Total radiation exposure accumulated during circular frame treatment of distal tibial fractures was quantified in 47 patients treated by a single surgeon from March 2011 until Nov 2014. The radiation exposures for all relevant radiology procedures for the distal tibial injury were included to estimate the radiation risk to the patient. Results:. The median time of treatment in the frame was 169 days (range 105 – 368 days). Patients underwent a median of 13 sets of plain radiographs; at least one intra operative exposure and 16 patients underwent CT scanning. The median total effective dose per patient from time of injury to discharge was 0.025 mSv (interquartile range 0.013 – 0.162 and minimum to maximum 0.01–0.53). CT scanning is the only variable shown to be an independent predictor of cumulative radiation dose on multivariate analysis, with a 13 fold increase in overall exposure. Conclusion:. Radiation exposure during treatment of distal tibial fractures with a circular frame in this group was well within reasonably safe limits. CT was the only significant predictor of overall exposure serves as a reminder to individually assess the risk and utility of radiological investigations on an individual basis. This is consistent with the UK legal requirements (Ionising Radiation (Medical Exposure) Regulations 2000. 1. )

A series of 76 distal tibial pilon fractures treated with surgical fixation were retrospectively reviewed from case notes, plain radiographs and CT imaging. Patient demographics, injury and fracture patterns, methods and timing of fixation and clinical and radiological outcomes were assessed over a mean follow up period of 8.6 months (range 2–30). Definitive fixation was most commonly performed through an open technique (71 cases) with plate fixation. CT imaging was used to plan the most direct approach to access the fracture fragments. Single or double incision techniques were used to access the tibia, with fixation of the fibular performed when necessary. Superficial infection occurred in 5 cases (6.9%) and deep infection in 2 (2.8%). Aseptic wound breakdown occurred in 5 cases (6.9%). The rate of wound breakdown after three-incision technique was 37.5%. There were 10 cases of non-union (13.9%) and 8 of mal-union (10.5%). Post-traumatic arthritis was present on the most recent x ray in 17 cases (23.4%). Further surgery was required in 20 cases (27.8%), most commonly for metalwork related problems and also for treatment of non-union, post-traumatic arthritis and infection. This review gives comprehensive injury specific and surgical outcome data from one of the largest reported series of these complex and problematic injuries