Contribution of individual projections alone and in combination for radiographic detection of ankle fractures

Fusion of electronic health records and radiographic images for a multimodal deep learning prediction model of atypical femur fractures

Atypical femur fractures (AFF) represent a very rare type of fracture that can be difficult to discriminate radiologically from normal femur fractures (NFF). AFFs are associated with drugs that are administered to prevent osteoporosis-related fragility fractures, which are highly prevalent in the elderly population. Given that these fractures are rare and the radiologic changes are subtle currently only 7% of AFFs are correctly identified, which hinders adequate treatment for most patients with AFF. Deep learning models could be trained to classify automatically a fracture as AFF or NFF, thereby assisting radiologists in detecting these rare fractures. Historically, for this classification task, only imaging data have been used, using convolutional neural networks (CNN) or vision transformers applied to radiographs. However, to mimic situations in which all available data are used to arrive at a diagnosis, we adopted an approach of deep learning that is based on the integration of image data and tabular data (from electronic health records) for 159 patients with AFF and 914 patients with NFF. We hypothesized that the combinatorial data, compiled from all the radiology departments of 72 hospitals in Sweden and the Swedish National Patient Register, would improve classification accuracy, as compared to using only one modality. At the patient level, the area under the ROC curve (AUC) increased from 0.966 to 0.987 when using the integrated set of imaging data and seven pre-selected variables, as compared to only using imaging data. More importantly, the sensitivity increased from 0.796 to 0.903. We found a greater impact of data fusion when only a randomly selected subset of available images was used to make the image and tabular data more balanced for each patient. The AUC then increased from 0.949 to 0.984, and the sensitivity increased from 0.727 to 0.849. These AUC improvements are not large, mainly because of the already excellent performance of the CNN (AUC of 0.966) when only images are used. However, the improvement is clinically highly relevant considering the importance of accuracy in medical diagnostics. We expect an even greater effect when imaging data from a clinical workflow, comprising a more diverse set of diagnostic images, are used.

Evaluation of ensemble strategy on the development of multiple view ankle fracture detection algorithm

OBJECTIVE

To identify the feasibility and efficiency of deep convolutional neural networks (DCNNs) in the detection of ankle fractures and to explore ensemble strategies that applied multiple projections of radiographs.Ankle radiographs (AXRs) are the primary tool used to diagnose ankle fractures. Applying DCNN algorithms on AXRs can potentially improve the diagnostic accuracy and efficiency of detecting ankle fractures.

METHODS

A DCNN was trained using a trauma image registry, including 3102 AXRs. We separately trained the DCNN on anteroposterior (AP) and lateral (Lat) AXRs. Different ensemble methods, such as "sum-up," "severance-OR," and "severance-Both," were evaluated to incorporate the results of the model using different projections of view.

RESULTS

The AP/Lat model's individual sensitivity, specificity, positive-predictive value, accuracy, and F1 score were 79%/84%, 90%/86%, 88%/86%, 83%/85%, and 0.816/0.850, respectively. Furthermore, the area under the receiver operating characteristic curve (AUROC) of the AP/Lat model was 0.890/0.894 (95% CI: 0.826-0.954/0.831-0.953). The sum-up method generated balanced results by applying both models and obtained an AUROC of 0.917 (95% CI: 0.863-0.972) with 87% accuracy. The severance-OR method resulted in a better sensitivity of 90%, and the severance-Both method obtained a high specificity of 94%.

CONCLUSION

Ankle fracture in the AXR could be identified by the trained DCNN algorithm. The selection of ensemble methods can depend on the clinical situation which might help clinicians detect ankle fractures efficiently without interrupting the current clinical pathway.

ADVANCES IN KNOWLEDGE

This study demonstrated different ensemble strategies of AI algorithms on multiple view AXRs to optimize the performance in various clinical needs.

Weber Ankle Fracture Classification System Yields Greatest Interobserver and Intraobserver Reliability Over AO/OTA and Lauge-Hansen Classification Systems Under Time Constraints in an Asian Population

No previous studies have evaluated the intra- and interobserver reliability between the Weber, Lauge-Hansen and AO Foundation/Orthopaedic Trauma Association (AO/OTA) classification systems under time constraints. This study compares the interobserver and intraobserver reliability of the aforementioned classification systems under simulated time constraints. Anteroposterior and lateral radiographs of ankle malleolar fractures from 80 consecutive patients from 2015 to 2016 were classified by 2 independent observers according to Weber, Lauge-Hansen and AO/OTA. Classifications were conducted over 4 successive weeks under timed (25-seconds) and untimed conditions, with 1-week gaps between each classification. Cohen's kappa and percentage agreement were calculated. Cohen's kappa for interobserver agreement ranged 0.67 to 0.67 and 0.59 to 0.73 for untimed and timed classifications for Weber; 0.38 to 0.47 and 0.44 to 0.50 for Lauge-Hansen; 0.28 to 0.49 and 0.13 to 0.37 for AO/OTA. Intraobserver agreement ranged from 0.83 to 0.85 and 0.78 to 0.79 for untimed and timed classifications for Weber; 0.46 to 0.65 and 0.59 to 0.73 for Lauge-Hansen; 0.42 to 0.63 and 0.40 to 0.51 for AO/OTA. Based on the Landis and Koch's benchmark scale, there was substantial agreement in the inter- and intraobserver variables for Weber; moderate agreement in inter- and intraobserver variables for Lauge-Hansen; fair and moderate agreement in inter- and intraobserver variables respectively for AO/OTA. Interobserver and intraobserver reliability was the most substantial for Weber, followed by Lauge-Hansen and AO/OTA. Time constraint did not have a statistically significant effect on the reliability of classifications. We recommend concurrent usage of the Weber and Lauge-Hansen system, since they demonstrate the greatest reliability and reproducibility, and confer better understanding of the fracture type, respectively.

Are 2 Radiographic Views As Good as 3 Views to Diagnose Ankle Fractures in Children and Adolescents?

OBJECTIVES

Ankle radiographs are among the most commonly obtained trauma images in the pediatric population, with the standard 3 views (AP/mortise [M]/lateral [L]) routinely ordered in the emergency department. The purpose of this study was to compare the diagnostic accuracy, sensitivity, and specificity of sets of 2 views (AP/L or M/L) with the standard 3 views.

METHODS

One hundred twenty sets of ankle radiographs of skeletally immature patients obtained in the emergency department of a level 1 pediatric trauma center were used. These included sets with and without fractures. Sets of 3 and 2 views were reviewed by pairs of pediatric-trained orthopedic surgeons, radiologists, and emergency physicians. Each completed 3 randomized viewing sessions where all possible combinations for each set of radiographs were reviewed. Diagnoses for the 3 sets of views were compared for accuracy, sensitivity, and specificity.

RESULTS

Overall accuracy, sensitivity, and specificity for all reviewers were as follows: AP/M/L 74%, 94%, and 90%, AP/L 71%, 90%, and 94%, as well as M/L 69%, 90%, and 91%. P values for accuracy, sensitivity, and specificity of AP/L and M/L compared with 3 views were 0.34, 0.04, and 0.52, as well as 0.04, 0.004, and 1.00, respectively.

CONCLUSIONS

In skeletally immature patients, statistically significant differences in accuracy were obtained when comparing the standard 3 AP/M/L views with more limited M/L views, suggesting that this set of 2 views is not as accurate. Differences in sensitivity of limited views were also statistically significant. Conversely, differences in accuracy between the standard 3 views and AP/Lateral views were not statistically significant. While more limited AP/L views may be comparable in accuracy and specificity and lead to dramatically decreased radiation and costs, this can be at the expense of less diagnostic sensitivity and increased risk of misdiagnosing or missing certain fractures.

LEVEL OF EVIDENCE

Level III.

Sensitivity and specificity of ultrasound in the diagnosis of traumatic ankle injury

Background:

This study was performed to determine the sensitivity and specificity of ultrasound in the diagnosis of traumatic ankle injury in comparison with magnetic resonance imaging (MRI).

Materials and Methods:

This cross-sectional study was performed on 31 patients with soft-tissue injury or fracture, referring to the MRI imaging center of Alzahra and Kashani Hospitals in Isfahan from October 2018 to March 2019. After an MRI, an ultrasound of the affected ankle was performed for all patients. Sonography and MRI were performed by two radiologists who were blinded to the results of each other's reports. The sensitivity, specificity, positive predictive value, and negative predictive value of sonography were determined.

Results:

In this study, 31 patients with ankle trauma were studied. The mean age of the patients was 30.73 ± 10.15 years; 32.3% were male and 67.7% were female. The sensitivity of ultrasound relative to MRI to detect damage to the anterior talofibular ligament (ATFL), posterior talofibular ligament (PTFL), and calcaneofibular ligament (CFL) was 66.67%, 50%, and 100%, respectively, and the corresponding specificity was 92.86%, 93.10%, and 93.10%, respectively. According to Kappa test, the agreement between ultrasound and MRI methods for detecting injury to ATFL (κ = 0.51), PTFL (κ = 0.35), and CFL (κ= 0.63) was statistically significant (P < 0.05).

Conclusion:

Ultrasound is an appropriate modality for the diagnosis of injuries to CFL and ATFL and has shown acceptable results for PTFL. It could be used as an alternative in cases where access to MRI is not available.

ACR Appropriateness Criteria® Acute Trauma to the Ankle

Acute injuries to the ankle are frequently encountered in the setting of the emergency room, sport, and general practice. This ACR Appropriateness Criteria defines best practices for imaging evaluation for several variants of patients presenting with acute ankle trauma. The variants include scenarios when Ottawa Rules can be evaluated, when there are exclusionary criteria, when Ottawa Rules cannot be evaluated, as well as specific injuries. Clinical scenarios are followed by the imaging choices and their appropriateness with an accompanying narrative explanation to help physicians to order the most appropriate imaging test. The American College of Radiology Appropriateness Criteria are evidence-based guidelines for specific clinical conditions that are reviewed annually by a multidisciplinary expert panel. The guideline development and revision include an extensive analysis of current medical literature from peer reviewed journals and the application of well-established methodologies (RAND/UCLA Appropriateness Method and Grading of Recommendations Assessment, Development, and Evaluation or GRADE) to rate the appropriateness of imaging and treatment procedures for specific clinical scenarios. In those instances where evidence is lacking or equivocal, expert opinion may supplement the available evidence to recommend imaging or treatment.

Ankle torque-related fractures and its echo-fast diagnosis protocol

Ankle torque is the most frequent injury mechanism encountered in traumatology. It usually develops as an ankle sprain, and often produces an ankle fracture. Signs and symptoms are very unspecific on the differential diagnosis, and conventional radiographs must be obtained to confirm it. This study compares ultrasound and standard X-ray images values found in ankle sprain-related fracture diagnoses. A 3-month prospective study was carried out on 52 patients with ankle torque. First, an ultrasound diagnosis was made by the first researcher at their arrival to the emergency department, usually consisting of a longitudinal section of the fibula, tibia, and V metatarsal. Then, a blinded independent investigator carried out the usual diagnosis protocol using traditional radiographs. When the required number was obtained, a third independent investigator evaluated the results. A Chi-squared test was used to contrast the outcomes between sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) comparing a non-inferiority hypothesis of our protocol against standard X-ray images screening. Echography showed to be at least as good as standard radiographs on the screening of ankle torque-related fractures (p = 0.02). In our series of patients, nearly 8% of fractures are misdiagnosed with plain radiographies, which concurs with the lower limit found in the bibliography. All the false negatives on radiographs were true positives in the first ultrasound assessment. The average time for the ultrasound protocol was 42 s. An ultrasound is at least as effective as conventional X-ray images when used in diagnosing ankle sprain-related fractures, regardless of age or sex and could in fact be better. The use of echography could reduce the number of radiographies performed when diagnosing these kinds of fractures, thereby reducing the amount of radiation exposure and expediting the diagnostic process as well as the "in situ diagnosis".Level of clinical evidence 2 [prospective cohort (N > 30 patients) study].

Preoperative radiography versus computed tomography for surgical planning for ankle fractures

PURPOSE

To review preoperative radiography and computed tomography (CT) of the ankle in 69 patients who underwent surgery for ankle fractures to determine the value of CT in diagnosis and surgical planning.

METHODS

Preoperative radiography and CT of the ankle of 46 women and 23 men aged 17 to 90 (mean, 48.8) years were reviewed. CT was deemed necessary when radiographs showed the following features: (1) comminuted fracture of the medial malleolus involving the tibial plafond, (2) comminuted fracture of the posterior malleolus, (3) presence of loose bodies, and/or (4) suspected Chaput or Volkman fracture fragment. Two orthopaedic surgeons independently reviewed the radiographs to look for any of the above features for which CT was indicated. In patients whose radiographs did not show any of the above features, each surgeon formulated a surgical plan based on radiographs alone and decided if any modification was needed after reviewing the CT scan.

RESULTS

Based on radiographs of the 69 patients, 19 (28%) patients had features of posterior malleolar comminution (n=7), medial malleolar comminution (n=7), suspected Chaput fracture fragment (n=1), suspected Volkman fracture fragment (n=1), and combination of 2 lesions (n=3), and were deemed to require CT. In 10 (20%) of the remaining 50 patients, the surgical plan was modified after review of the CT scan. The intra- and inter-observer agreement was good to excellent.

CONCLUSION

Radiography alone is not adequate for surgical planning for ankle fractures. More accurate imaging tools such as CT are needed to enable a more accurate diagnosis and surgical planning.

ACR Appropriateness Criteria acute trauma to the ankle

Acute ankle injuries are frequently diagnosed and treated in emergency departments and outpatient clinics. Recent evidence-based clinical treatment guidelines and systematic review of economic analyses support the use of 3-view (anteroposterior, lateral, and mortise) radiographic evaluation of patients meeting the criteria of the Ottawa ankle rules. Cross-sectional imaging has a limited secondary role primarily as a tool for preoperative planning and as a problem-solving technique in patients with persistent symptoms and suspected of having occult fractures. The ACR Appropriateness Criteria are evidence-based guidelines for specific clinical conditions that are reviewed every 3 years by a multidisciplinary expert panel. The guideline development and review include an extensive analysis of current medical literature from peer-reviewed journals and the application of a well-established consensus methodology (modified Delphi) to rate the appropriateness of imaging and treatment procedures by the panel. In those instances in which evidence is lacking or not definitive, expert opinion may be used to recommend imaging or treatment.

Late diagnosis of a McFarland fracture: imaging and treatment

McFarland fractures represent a type of oblique medial malleolar fracture in children that can be challenging to diagnose and treat. A 14-year-old junior league soccer player with a Salter Harris type IV McFarland fracture presented late, as the initial routine two views radiological assessment failed to reveal a clear fracture line. The addition of a mortise ankle view led to the correct diagnosis and subsequent MRI findings guided nonsurgical treatment with an excellent outcome. The debate between obtaining two or three views in closed pediatric ankle injuries according to the so-called Ottawa rules and the usefulness of magnetic resonance imaging (MRI) in the decision making for the choice of treatment of McFarland fractures are discussed in this case report.

Stability-based classification for ankle fracture management and the syndesmosis injury in ankle fractures due to a supination external rotation mechanism of injury

The aim of this thesis was to confirm the utility of stability-based ankle fracture classification in choosing between non-operative and operative treatment of ankle fractures, to determine how many ankle fractures are amenable to non-operative treatment, to assess the roles of the exploration and anatomical repair of the AITFL in the outcome of patients with SER ankle fractures, to establish the sensitivities, specificities and interobserver reliabilities of the hook and intraoperative stress tests for diagnosing syndesmosis instability in SER ankle fractures, and to determine whether transfixation of unstable syndesmosis is necessary in SER ankle fractures. The utility of stability based fracture classification to choose between non-operative and operative treatment was assessed in a retrospective study (1) of 253 ankle fractures in skeletally mature patients, 160 of whom were included in the study to obtain an epidemiological profile in a population of 130,000. Outcome was assessed after a minimum follow-up of two years. The role of AITFL repairs was assessed in a retrospective study (2) of 288 patients with Lauge-Hansen SE4 ankle fractures; the AITFL was explored and repaired in one group (n=165), and a similar operative method was used but the AITFL was not explored in another group (n=123). Outcome was measured with a minimum follow-up of two years. Interobserver reliability of clinical syndesomosis tests (study 3) and the role of syndesmosis transfixation (study 4) were assessed in a prospective study of 140 patients with Lauge-Hansen SE4 ankle fractures. The stability of the distal tibiofibular joint was evaluated by the hook and ER stress tests. Clinical tests were carried out by the main surgeon and assistant, separately, after which a 7.5-Nm standardized ER stress test for both ankles was performed; if it was positive, the patient was randomized to either syndesmosis transfixation (13 patients) or no fixation (11 patients) treatment groups. The sensitivity and specificity of both clinical tests were calculated using the standard 7.5-Nm external rotation stress test as reference. Outcome was assessed after a minimum of one year of follow-up. Olerud-Molander (OM) scoring system, RAND 36-Item Health Survey, and VAS to measure pain and function were used as outcome measures in all studies. In study 1, 85 (53%) fractures were treated operatively using the stability based fracture classification. Non-operatively treated patients reported less pain and better OM (good or excellent 89% vs. 71%) and VAS functional scores compared to operatively treated patients although they experienced more displacement of the distal fibula (0 mm 30% vs. 69%; 0-2 mm 65% vs. 25%) after treatment. No non-operatively treated patients required operative fracture fixation during follow-up. In study 2, AITFL exploration and suture lead to equal functional outcome (OM mean, 77 vs. 73) to no exploration or fixation. In study 3, the hook test had a sensitivity of 0.25 and a specificity of 0.98. The external rotation stress test had a sensitivity of 0.58 and a specificity of 0.9. Both tests had excellent interobserver reliability; the agreement was 99% for the hook test and 98% for the stress test. There was no statistically significant difference in functional scores (OM mean, 79.6 vs. 83.6) or pain between syndesmosis transfixation and no fixation groups (Study 4). Our results suggest that a simple stability-based fracture classification is useful in choosing between non-operative and operative treatment of ankle fractures; approximately half of the ankle fractures can be treated non-operatively with success. Our observations also suggest that relevant syndesmosis injuries are rare in ankle fractures due to an SER mechanism of injury. According to our research, syndesmotic repair or fixation in SER ankle fracture has no influence on functional outcome or pain after minimum one year compared with no fixation.

Preventing surgical complications: A survey on surgeons' perception of intra-articular malleolar screw misplacement in a cadaveric study

BACKGROUND

Intra-articular hardware penetration can occur during osteosynthesis of ankle fractures, jeopardizing patients' outcomes. The intraoperative recognition of misplaced screws may be difficult due to the challenge of adequate interpretation of specific radiographic views. The present study was designed to investigate the diagnostic accuracy of standardized radiographic ankle views to determine the accuracy of diagnosis for intra-articular hardware placement of medial malleolar screws in a cadaveric model.

METHODS

Nine preserved human cadaveric lower extremity specimens were used. Under direct visualization, two 4.0 mm cancellous screws were inserted into the medial malleolus. Each specimen was analyzed radiographically using antero-posterior (AP) and mortise views. The X-rays were randomly uploaded on a CD-ROM and included in a survey submitted to ten selected orthopaedic surgeons. The "Standards for Reporting of Diagnostic Accuracy" (STARD) questionnaire was used to determine the surgeons' perception of accuracy of screw placement in the medial malleolus. The selection of items was based on evidence whenever possible, therefore the "inconclusive" category was added. Inter and intraobserver variations were analyzed by kappa statistics to measure the amount of agreement.

RESULTS

There was a poor level of agreement (kappa 0.4) both in the AP and in the mortise view among all the examiners. Associating the two x-rays, the agreement remained poor (kappa 0.4). In the cases in which there was a diagnosis of articular penetration, there was a poor agreement related to which of the screws was intra-articular. The number of "inconclusive" responses was low and constant, without a statistically significant difference between the subspecialists

CONCLUSION

The routine intraoperative radiographic imaging of the ankle is difficult to interpret and unreliable for detection of intra-articular hardware penetration. We therefore recommend to reposition medial malleolar screws intraoperatively if there is any doubt regarding inadequate screw placement.

Validation of the Ottawa Ankle Rules in Iran: a prospective survey

Background

Acute ankle injuries are one of the most common reasons for presenting to emergency departments, but only a small percentage of patients – approximately 15% – have clinically significant fractures. However, these patients are almost always referred for radiography. The Ottawa Ankle Rules (OARs) have been designed to reduce the number of unnecessary radiographs ordered for these patients. The objective of this study was to validate the OARs in the Iranian population.

Methods

This prospective survey was done among 200 patients with acute ankle injury from January 2004 to April 2004 in the Akhtar Orthopedics Hospital Emergency Department. Main outcome measures of this survey were: sensitivity, specificity, positive predictive value, negative predictive value, and likelihood ratios (positive and negative) of the OARs.

Results

Sensitivity of the OARs for detecting 37 ankle fractures (23 in the malleolar zone and 14 in the midfoot zone) was 100% for each of the two zones, and 100% for both zones. Specificity of the OARs for detecting fractures was 40.50% for both zones, 40.50% for the malleolar zone, and 56.00% for the midfoot zone. Implementation of the OARs had the potential for reducing radiographs by 33%.

Conclusion

OARs are very accurate and highly sensitive tools for detecting ankle fractures. Implementation of these rules would lead to significant reduction in the number of radiographs, costs, radiation exposure and waiting times in emergency departments.

Ankle fractures resulting from rotational injuries

Ankle fractures are among the most common skeletal injuries; selection of an optimal management method depends on ankle stability. Stable fractures (eg, isolated lateral malleolar) generally are managed nonsurgically; unstable fractures (eg, bimalleolar, bimalleolar equivalent) usually are managed with open reduction and internal fixation. Stress radiographs may aid in the management of incomplete deltoid injury in which there is medial swelling and tenderness without radiographic talar shift. A posterior malleolar fracture should be reduced and stabilized if it comprises >30% of the articular surface and remains displaced after fibular stabilization. Ankle fractures with syndesmotic injury have additional tibiofibular instability that can be controlled by screw fixation. However, the choice between metal and bioabsorbable screws, screw size, number of cortices fixed, and indications for screw removal remain controversial. Conditions such as diabetes or advanced age are no longer contraindications to usual management recommendations.