Emergency Radiology最新文献

Purpose: Missed fractures are the primary cause of interpretation errors in emergency radiology, and artificial intelligence has recently shown great promise in radiograph interpretation. This study compared the diagnostic performance of two AI algorithms, BoneView and RBfracture, in detecting traumatic abnormalities (fractures and dislocations) in MSK radiographs.

Methods: AI algorithms analyzed 998 radiographs (585 normal, 413 abnormal), against the consensus of two MSK specialists. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and interobserver agreement (Cohen's Kappa) were calculated. 95% confidence intervals (CI) assessed robustness, and McNemar's tests compared sensitivity and specificity between the AI algorithms.

Results: BoneView demonstrated a sensitivity of 0.893 (95% CI: 0.860-0.920), specificity of 0.885 (95% CI: 0.857-0.909), PPV of 0.846, NPV of 0.922, and accuracy of 0.889. RBfracture demonstrated a sensitivity of 0.872 (95% CI: 0.836-0.901), specificity of 0.892 (95% CI: 0.865-0.915), PPV of 0.851, NPV of 0.908, and accuracy of 0.884. No statistically significant differences were found in sensitivity (p = 0.151) or specificity (p = 0.708). Kappa was 0.81 (95% CI: 0.77-0.84), indicating almost perfect agreement between the two AI algorithms. Performance was similar in adults and children. Both AI algorithms struggled more with subtle abnormalities, which constituted 66% and 70% of false negatives but only 20% and 18% of true positives for the two AI algorithms, respectively (p < 0.001).

Conclusions: BoneView and RBfracture exhibited high diagnostic performance and almost perfect agreement, with consistent results across adults and children, highlighting the potential of AI in emergency radiograph interpretation.

Purpose: The purpose of this retrospective study is to evaluate imaging features of sinonasal NK/T-cell lymphoma, potentially aiding early diagnosis and improving patient outcomes through timely biopsy evaluation and treatment.

Materials & methods: This study was designed as a retrospective review. Imaging characteristics in a total of 28 patients with histology confirmed sinonasal NK/T-cell lymphoma were evaluated by two radiologists. Demographic data and clinical features were extracted from the electronic medical record.

Results: Among 28 patients, 84.6% presented with nasal cavity involvement and 88.5% with an isoattenuating lesion on CT and 63.6% with T2 hypointensity on MRI. Notably, 57.7% showed osseous destruction and 27.3% had devascularized tissue, highlighting significant radiologic features that may overlap with non-neoplastic etiologies.

Conclusion: Sinonasal NK/T-cell lymphoma should be in the differential diagnosis in patients presenting with chronic sinonasal symptoms and imaging findings of tissue necrosis, osseous destruction, and adjacent fat infiltration. Imaging features overlap with entities such as aggressive fungal sinusitis and granulomatosis with polyangiitis.

Among radiologists, the scalp is a frequently overlooked region, and the literature offers little guidance regarding the imaging spectrum of scalp abnormalities in acute trauma. The severity of scalp abnormalities on imaging can be a harbinger of significant intracranial trauma. In this pictorial essay, we discuss a myriad of traumatic scalp lesions that are infrequently encountered on imaging.

Multi-energy detector spectral CT is a specialized form of dual-energy CT, an imaging technique that takes full advantage of the entire X-ray spectrum applied during CT scans, enabling superior differentiation of materials that may be indistinguishable on traditional CT. Contrary to source-based dual-energy CT, the benefits of detector-based spectral CT are always available with every scan, without needing to preemptively manually select dual source technique at the time of scanning. In cases of questionable findings on conventional CT images that require enhanced material differentiation, "spectral" data collected alongside CT images scanned utilizing this technique can be instantly analyzed to improve diagnostic confidence, without needing to re-scan the patient. While applicable to various patient populations and clinical scenarios, spectral CT is particularly useful for neuroradiology in the often fast-paced emergency department. The authors' facility uses the Philips Healthcare IQon Spectral CT in its own emergency department. This paper strives to briefly introduce some ways that spectral CT can be utilized to improve patient care in this setting.

Spontaneous hemoperitoneum (SH) is defined as the presence of blood within the peritoneal cavity from a non-traumatic origin. The absence of clinical suspicion and its low frequency render it a challenging diagnosis in the emergency setting. As it is a potentially life-threatening condition, radiologists play a fundamental role in the identification and location of the source of bleed, both essential to be able to deliver the best care possible for the patient. This review article aims to facilitate a practical approach to SH: We will begin by reviewing key imaging modalities of significant value for the detection and characterization of SH. Relevant anatomy and important CT signs to know will be mentioned. We will summarize the potential causes of SH in order of frequency, with a brief mention of differential diagnoses to consider. Finally, we will illustrate this review with a collection of cases from our hospital to showcase relevant diagnoses to consider when facing a patient with SH, focusing on key imaging elements that should not be overlooked.

Solid pseudopapillary neoplasm (SPN) of the pancreas, are rare low-grade malignant pancreatic tumors, most commonly occurring in young women during their second or third decade of life. They can present with vague abdominal symptoms like pain, discomfort, and nausea, or it can be incidentally discovered on imaging studies. There has been limited literature on ruptured SPN, which can be spontaneous or traumatic. In this paper, we present three cases of ruptured SPN, one traumatic and two spontaneous, to add to the scarcity of knowledge regarding this condition and its radiologic features, along with a detailed review of current literature.

Distal radius fractures are one of the most prevalent fracture types. Dorsal tilt of the articular surface of the distal radius as assessed on the lateral radiograph is often used in determining the correct treatment method. This report seeks to highlight the importance of correct forearm positioning when obtaining distal radius radiographs, and the effects of pronation and supination on dorsal tilt measurements of the radius. It presents a case of a 35-year-old male with a malaligned left-sided distal radius fracture sustained during a sports-related fall.

Purpose: MRI is the preferred imaging modality for patients with acute dizziness when a central etiology is possible. Abbreviated protocols may improve access in urgent settings. This study assesses the diagnostic yield and utility of an abbreviated MRI protocol for patients presenting with dizziness to the emergency department (ED).

Method: This retrospective study included 613 adult patients presenting to the ED with dizziness from August 1, 2019 to August 31, 2023. The protocol included 3 mm coronal and axial DWI, axial FLAIR, and SWI sequences, with a duration of approximately 11 min. MRI findings were categorized as negative or positive for intracranial pathology; etiology and location were recorded. Charts were reviewed for concurrent CTA during the ED visit, and findings were assessed for correlation with MRI results.

Results: Of the 613 patients, clinically significant intracranial pathology was identified in 52 cases (8%), including 42 (7%) acute infarcts. Of these infarcts, 19 (45%) were infratentorial, 16 (38%) supratentorial, and 7 (17%) involved both regions. The cerebellum was the most common infratentorial site (38%), followed by the brainstem (24%). Infarcts ranged from 1-84 mm, with 48% measuring less than 1 cm. TOAST classification revealed strokes as cardioembolic (36%), large vessel (26%), cryptogenic (19%), and lacunar (19%). Statistical analysis showed no significant relationship between vertigo and infarct characteristics (P > 0.05).

Conclusion: Abbreviated protocol MRI demonstrated a 8% diagnostic yield for detecting intracranial pathology and more often positive than concurrent CT/CTA in identifying acute findings. Supratentorial pathology can present with symptoms of dizziness as well. The abbreviated protocol offers a rapid, efficient diagnostic tool for urgent care settings and MRI identifies more acute findings than concurrent CT/CTA.

The gallbladder can be affected by various acute conditions beyond uncomplicated cholecystitis, with the prevalence of gallstones having risen over the past three decades. Management of biliary diseases is complex and necessitates careful case-by-case consideration. Radiologic imaging plays a crucial role in the evaluation, diagnosis, and management planning of biliary pathologies. This review explores acute gallbladder conditions such as simple cholecystitis, gangrenous cholecystitis, emphysematous cholecystitis, hemorrhagic cholecystitis, perforated cholecystitis, gallbladder fistulas, gallstone ileus, Bouveret syndrome, Mirizzi syndrome, gallbladder herniation, volvulus, trauma, pseudoaneurysm, portal venous thrombosis, and gallbladder carcinoma. It highlights the utility of various imaging modalities including ultrasound, CT, dual-energy CT, MRI, and MRCP in diagnosing these conditions. Advancements in imaging techniques have enhanced the ability to detect and characterize gallbladder diseases, facilitating timely surgical interventions and improving patient outcomes. This review emphasizes the importance of close collaboration between radiologists and clinicians to optimize diagnosis and management strategies, underscoring the indispensable role of radiologic imaging in modern medicine.

Purpose: This study aimed to develop an automated early warning system using a large language model (LLM) to identify acute to subacute brain infarction from free-text computed tomography (CT) or magnetic resonance imaging (MRI) radiology reports.

Methods: In this retrospective study, 5,573, 1,883, and 834 patients were included in the training (mean age, 67.5 ± 17.2 years; 2,831 males), validation (mean age, 61.5 ± 18.3 years; 994 males), and test (mean age, 66.5 ± 16.1 years; 488 males) datasets. An LLM (Japanese Bidirectional Encoder Representations from Transformers model) was fine-tuned to classify the CT and MRI reports into three groups (group 0, newly identified acute to subacute infarction; group 1, known acute to subacute infarction or old infarction; group 2, without infarction). The training and validation processes were repeated 15 times, and the best-performing model on the validation dataset was selected to further evaluate its performance on the test dataset.

Results: The best fine-tuned model exhibited sensitivities of 0.891, 0.905, and 0.959 for groups 0, 1, and 2, respectively, in the test dataset. The macrosensitivity (the average of sensitivity for all groups) and accuracy were 0.918 and 0.923, respectively. The model's performance in extracting newly identified acute brain infarcts was high, with an area under the receiver operating characteristic curve of 0.979 (95% confidence interval, 0.956-1.000). The average prediction time was 0.115 ± 0.037 s per patient.

Conclusion: A fine-tuned LLM could extract newly identified acute to subacute brain infarcts based on CT or MRI findings with high performance.