Application of deep learning algorithms in classification and localization of implant cutout for the postoperative hip.

OBJECTIVE: This study aims to explore the feasibility of employing convolutional neural networks for detecting and localizing implant cutouts on anteroposterior pelvic radiographs. MATERIALS AND METHODS: The research involves the development of two Deep Learning models. Initially, a model was created for image-level classification of implant cutouts using 40191 pelvic radiographs obtained from a single institution. The radiographs were partitioned into training, validation, and hold-out test datasets in a 6/2/2 ratio. Performance metrics including the area under the receiver operator characteristics curve (AUROC), sensitivity, and specificity were calculated using the test dataset. Additionally, a second object detection model was trained to localize implant cutouts within the same dataset. Bounding box visualizations were generated on images predicted as cutout-positive by the classification model in the test dataset, serving as an adjunct for assessing algorithm validity. RESULTS: The classification model had an accuracy of 99.7%, sensitivity of 84.6%, specificity of 99.8%, AUROC of 0.998 (95% CI: 0.996, 0.999) and AUPRC of 0.774 (95% CI: 0.646, 0.880). From the pelvic radiographs predicted as cutout-positive, the object detection model could achieve 95.5% localization accuracy on true positive images, but falsely generated 14 results from the 15 false-positive predictions. CONCLUSION: The classification model showed fair accuracy for detection of implant cutouts, while the object detection model effectively localized cutout. This serves as proof of concept of using a deep learning-based approach for classification and localization of implant cutouts from pelvic radiographs.

Application of a deep learning algorithm in the detection of hip fractures.

This paper describes the development of a deep learning model for prediction of hip fractures on pelvic radiographs (X-rays). Developed using over 40,000 pelvic radiographs from a single institution, the model demonstrated high sensitivity and specificity when applied to a test set of emergency department radiographs. This study approximates the real-world application of a deep learning fracture detection model by including radiographs with sub-optimal image quality, other non-hip fractures, and metallic implants, which were excluded from prior published work. The study also explores the effect of ethnicity on model performance, as well as the accuracy of visualization algorithm for fracture localization.

Acute Hemiballismus as the Initial Manifestation of Ischemic Stroke: A Case Report.

INTRODUCTION: Cerebrovascular disease often presents with "negative" symptoms such as weakness with reduced movement of body parts or sensory loss. Rarely do "positive" symptoms such as abnormal movements manifest in acute stroke, with hemichorea being a very rare manifestation. CASE REPORT: This is a case report of a 62-year-old chronic smoker with no known past medical history who presented with choreatic movements of his arm and leg. Magnetic resonance imaging of the brain showed changes consistent with an infarct in the right centrum semiovale. He was treated with dual antiplatelets and was noted to have subsequent improvement in symptoms. CONCLUSION: Recognition and awareness of stroke presenting as movement disorders in the emergency department can help prevent delays in diagnosis and treatment.

Factors associated with inappropriate attendances at the emergency department of a tertiary hospital in Singapore.

INTRODUCTION: Inappropriate attendances (IAs) at emergency departments (ED) are contributed by patients with mild or moderate medical conditions that can be effectively managed by primary care physicians. IAs strain limited ED resources and have an adverse impact on efficiency. This study aimed to identify factors associated with IA at the ED of a tertiary hospital in Singapore. METHODS: We conducted a retrospective cohort study of all eligible visits to the aforementioned ED between 1 January 2015 and 31 December 2015. The appropriateness of each attendance was estimated using criteria based on investigations or procedures that were performed on the attendee and the discharge type of that attendance. IAs were then compared against appropriate attendances in these areas: attendee demographics; referral source; time of ED visit; proximity to ED and 24-hour general practitioner clinics; and history of ED visits in 2014. Multivariate analysis was performed on significant variables associated with IAs. RESULTS: Among 120,606 attendances, 11,631 (9.6%) were IAs. Multivariate analysis showed that gender, ethnicity, referral source, time of ED visit, nationality and history of frequent visits to the ED were factors associated with IAs. Moreover, the odds of IA were found to be higher among attendees who were younger, were self-referred, or had at least one IA in 2014. CONCLUSION: This study identified subgroups in the population who were more likely to contribute to IAs at the ED. These findings offer relevant insights into future research directions and strategies that might potentially reduce avoidable IAs.

Utilization of a Glidescope videolaryngoscope for orotracheal intubations in different emergency airway management settings.

OBJECTIVE: To describe the initial experience of a group of emergency department (ED) physicians, utilizing a Glidescope videolaryngoscope (GVL) for orotracheal intubations in the ED. METHODOLOGY: A 6-month, single center, prospective observational study from 19 Feb 2007 to 18 Aug 2007 was conducted on all orotracheal intubations, which involved utilization of the original GVL in different emergency airway management scenarios. RESULTS: Overall success of GVL intubation was 15 out of 21 (71.4%) cases. The GVL was able to provide at least Cormack-Lehane grade I or II laryngoscopy views in all cases. All the operators highlighted difficulty in angulating and maneuvering the endotracheal tube for insertion through the glottis as the primary difficulty encountered. CONCLUSION: We found the GVL to be an effective device in our ED's emergency airway control repertoire. Its role in the anticipated difficult airway in the ED will need further studies.

Spinal abscess: a case series.

OBJECTIVE: Spinal abscess is a rare but potentially devastating condition. We present a case series, looking into its presentation, risk factors, management and outcome. METHODS: Five patients over a 10 year period were identified, with a discharge diagnosis code of 'spinal abscess, 324.1' from the hospital computer database. RESULTS: Four out of five patients presented atypically. Prognosis corresponded to early diagnosis and surgical decompression. Risk factors included intravenous drug abuse, a compromised immune system and infection in another organ system. Magnetic resonance imaging was an important diagnostic tool for all. CONCLUSION: The early diagnosis and immediate surgical treatment of spinal abscesses remain cornerstones in improving the outcomes of the disease. From our series, risk factor assessment appear to be more useful than the classical triad of fever, spine pain and neurological deficits to screen ED patients with spine pain for spinal abscess.

Monteggia 'fracture'-dislocation with bowing of the ulna: a pitfall for the unwary emergency physician.

We recently encountered a child who fell on the left outstretched hand. She complained of left elbow pain subsequently and radiographs showed a supposedly isolated radial head dislocation. This was spontaneously reduced when the left forearm was extended for another radiograph. She was discharged from the emergency department with her left elbow in a long backslab, but returned within half an hour with recurrence of the radial head dislocation without repeat trauma. Only when the radiograph was reviewed was a subtle plastic bowing deformity of the ulna seen. This proved the axiom, that in forearm parallel long bone injuries, always consider the possibility of Monteggia or Galeazzi lesions when there is apparently isolated radial or ulna head dislocations. The clinician must always look very carefully for a fracture of the accompanying long bone, as even bowing 'fractures' might occur.

Disaster preparedness: Experience from a smoke inhalation mass casualty incident.

OBJECTIVE: We describe a modified triage system used in managing a smoke inhalation mass casualty incident that we recently encountered at our community hospital. MATERIALS AND METHODS: The patients were triaged as priority 1, 2 or 3 on the basis of their symptoms, signs and circumstances at scene. In addition, the use of fibre-optic examinations of the upper airway, chest radiography and carboxyhaemoglobin levels with arterial blood gas analyses were used to aid in disposal plans. RESULTS: Of the 22 patients evacuated, 15 were triaged as priority 2 and the remaining seven as priority 3. None of the patients was identified as priority 1. All the priority 2 patients underwent further investigations. Those with mild upper airway oedema (four patients) or raised carboxyhaemoglobin levels (two patients) were admitted. Only one patient had both. Another patient who was a known asthmatic developed bronchospasm and was admitted as well. All six were admitted to the general ward with subsequent good recovery and were discharged within 3 days. The remaining nine priority 2 and seven priority 3 patients were discharged from the emergency department. CONCLUSIONS: These modified triage criteria, with selective use of fibre-optic examinations, chest radiography and arterial blood gas analyses with carboxyhaemoglobin levels, are useful in smoke inhalation mass casualty incidents without dermal burns. Systemic injury and poisoning by toxic fumes often coexist with airway burns and should not be overlooked. Lastly, disaster planning and frequent drills at both local and national levels will optimize the response to future mass casualty incidents.

Sigmoid volvulus: diagnostic twists and turns.

OBJECTIVE: The purpose of this study was to describe the clinical profile of patients with sigmoid volvulus and highlight the diagnostic difficulties and pitfalls. METHODS: This was a retrospective descriptive study of patients with sigmoid volvulus seen at a community hospital over a 5-year period from July 1999 to July 2004. RESULTS: Twenty-eight cases of sigmoid volvulus were identified, of which 16 were male and 12 female. The median age was 74 years. Twenty-seven were Chinese and one was Malay. Seven out of 28 patients were from nursing institutes while the rest came from their own homes. Eleven were bedridden and chronic constipation was a problem for 17 patients. Co-existing medical illnesses, including neuropsychiatric conditions, were common. Abdominal distension and pain were the most frequent presentations (26 and 21 patients, respectively). Twelve patients had constipation and 11 had diarrhoea. Only nine had the complete typical triad of abdominal distension, pain and constipation. The classical coffee-bean sign was identified in nine out of 26 initial plain abdominal radiographs done at the emergency department and a further 11 were diagnosed after review by the surgical team in the ward. Another four were identified only after computed tomography scan of the abdomen and pelvis. The remaining two patients were diagnosed intra-operatively after they underwent emergency laparotomy. CONCLUSIONS: As our aged population continues to grow, sigmoid volvulus may be more commonly encountered. Emergency physicians should maintain a high index of suspicion and avoid the pitfalls so as to prevent delay in diagnosis and treatment.

Clinical predictors of bleeding esophageal varices in the ED.

OBJECTIVES: Some authors have found that thrombocytopenia (<118,000/mm3), splenomegaly, and ascites are useful predictors of large esophageal varices in cirrhotic patients. We decide to see whether these factors could also be used to predict bleeding esophageal varices in patients known to have chronic liver disease in the ED. METHODS: A case record review was done of all patients admitted to the ED of Changi General Hospital with upper gastrointestinal bleeding from esophageal varices from October 1999 to April 2004. The criteria of thrombocytopenia, splenomegaly, and ascites were applied retrospectively to these patients to see how accurately they performed in predicting bleeding esophageal varices. RESULTS: Only 55% of patients had thrombocytopenia, whereas 45% had splenomegaly, and 27.5% had ascites. Combining thrombocytopenia with the presence of either ascites or splenomegaly did not improve the yield (only 40%), and only 6 patients had all 3 criteria. Twelve patients with bleeding varices did not have any of the criteria. CONCLUSIONS: Thrombocytopenia, splenomegaly, or ascites is an unreliable predictor of bleeding esophageal varices. Urgent or emergent endoscopy is still advocated to accurately diagnose bleeding esophageal varices.

Radiologic screening for midfacial fractures: a single 30-degree occipitomental view is enough.

BACKGROUND: The standard radiologic screening views for maxillofacial trauma in the accident and emergency department have been the Waters (30-degree occipitomental [OM]) view, the Caldwell (posteroanterior) view, and the lateral view. We wanted to see if a single 30-degree occipitomental view would be sufficient to screen for maxillofacial trauma (excluding nasal fractures). METHODS: Over a 1-year period, each of 730 patients had three radiologic views obtained for maxillofacial trauma at our emergency department. A consultant emergency physician read the single OM view for injury, and then compared his radiologic findings with those made by other doctors on the basis of the three views. The need for further computed tomographic scanning was also studied. RESULTS: There were 730 patients recruited, of which 104 patients had abnormal findings on three-view screening. All the radiologic abnormalities were seen on the OM view; however, for the same 104 patients, only 48 had abnormal findings on the Caldwell views. Radiologic detail for the latter 48 patients was less obvious on the Caldwell views compared with the OM views. Other than two mandible fractures, no facial abnormality was demonstrable on any of the lateral views, even with significant bony injury on the OM views. CONCLUSION: A single 30-degree occipitomental view would be sufficient to screen for maxillofacial trauma. Further views are unnecessary and add little clinical value. Computed tomographic scanning with three-dimensional reconstruction of the facial skeleton should be the next preferred investigation for any patient found to have fractures on the Waters views.

Emergency department organisation for disasters : A review of emergency department disaster plans in public hospitals of Singapore

Disaster management plans of emergency departments (EDs) in four major public hospitals were reviewed. A comparison was made between these plans, and they were analyzed to gain an understanding of the differing objectives and doctrines behind the practices. These were summarized into five major management concepts, which are considered to be critical to the success of a disaster plan: 1)staff mobilization systems (cascading vs batch mobilization); 2)staff deployment systems; 3)team organization (surgeons vs residents); 4)area management (the role of the area manager); 5)casualty volume management (accommodation vs expansion vs extension concepts). the concepts derived should serve as a useful guide to the development of an ED disaster plan and potentially influence how new ED facilities could be planned(AU)