暂无摘要。

The \"puffed cheek\" technique is routinely performed during CT neck studies in patients with suspected oral cavity cancers. The insufflation of air within the oral vestibule helps in the detection of small buccal mucosal lesions, with better delineation of lesion origin, depth, and extent of spread. The pitfalls associated with this technique are often underrecognized and poorly understood. They can mimic actual lesions, forfeiting the technique\'s primary purpose. This review provides an overview of the puffed cheek technique and its associated pitfalls. These pitfalls include pneumoparotid, soft palate elevation that resembles a nasopharyngeal mass, various tongue displacements or distortions that obscure tongue lesions or mimic them, sublingual gland herniation, an apparent exacerbation of the airway edema, vocal cord adduction that hinders glottic evaluation, and false indications of osteochondronecrosis in laryngeal cartilage. Most stem from a common underlying mechanism of unintentional Valsalva maneuver engaged in by the patient while trying to perform a puffed cheek, creating a closed air column under positive pressure with resultant surrounding soft-tissue displacement. These pitfalls can thus be avoided by instructing the patient to maintain continuous nasal breathing while puffing out their cheek during image acquisition, preventing the formation of the closed air column. Keywords: CT, Head/Neck © RSNA, 2024.

Purpose To explore the potential benefits of deep learning-based artifact reduction in sparse-view cranial CT scans and its impact on automated hemorrhage detection. Materials and Methods In this retrospective study, a U-Net was trained for artifact reduction on simulated sparse-view cranial CT scans in 3000 patients, obtained from a public dataset and reconstructed with varying sparse-view levels. Additionally, EfficientNet-B2 was trained on full-view CT data from 17 545 patients for automated hemorrhage detection. Detection performance was evaluated using the area under the receiver operating characteristic curve (AUC), with differences assessed using the DeLong test, along with confusion matrices. A total variation (TV) postprocessing approach, commonly applied to sparse-view CT, served as the basis for comparison. A Bonferroni-corrected significance level of .001/6 = .00017 was used to accommodate for multiple hypotheses testing. Results Images with U-Net postprocessing were better than unprocessed and TV-processed images with respect to image quality and automated hemorrhage detection. With U-Net postprocessing, the number of views could be reduced from 4096 (AUC: 0.97 [95% CI: 0.97, 0.98]) to 512 (0.97 [95% CI: 0.97, 0.98], P < .00017) and to 256 views (0.97 [95% CI: 0.96, 0.97], P < .00017) with a minimal decrease in hemorrhage detection performance. This was accompanied by mean structural similarity index measure increases of 0.0210 (95% CI: 0.0210, 0.0211) and 0.0560 (95% CI: 0.0559, 0.0560) relative to unprocessed images. Conclusion U-Net-based artifact reduction substantially enhanced automated hemorrhage detection in sparse-view cranial CT scans. Keywords: CT, Head/Neck, Hemorrhage, Diagnosis, Supervised Learning Supplemental material is available for this article. © RSNA, 2024.

BACKGROUND: Thyroid storm is a rare but potentially lethal manifestation of thyrotoxicosis. Guidelines recommend nonoperative management of thyroid storm, but thyroidectomy can be performed if patients fail medical therapy or need immediate resolution of the storm. Outcomes of thyroidectomy for management of thyroid storm remain ill-defined.
METHODS: Using the National Inpatient Sample from 2016 to 2020, a retrospective analysis was conducted of patients admitted with thyroid storm. Outcomes of interest included operative complications and mortality. Multivariable logistic regression was performed to assess factors associated with receiving thyroidectomy and mortality.
RESULTS: An estimated 16,175 admissions had a diagnosis of thyroid storm. The incidence of thyroid storm increased from .91 per 100,000 people in 2016 to 1.03 per 100,000 people in 2020, with a concomitant increase in mortality from 2.9% to 5.3% (P < .001). Operative intervention was pursued in 635 (3.9%) cases with a perioperative complication rate of 30%. On multivariable regression, development of acute decompensated heart failure (adjusted odds ratio [AOR] 1.66, 95% Confidence Interval [CI] 1.03-2.68, P = .037) and acute renal failure (AOR 2.10, 95% CI 1.17-3.75, P = .013) increased odds of receiving surgery. The same multivariable model did not show a significant association between thyroidectomy and mortality.
CONCLUSIONS: The incidence of thyroid storm and associated mortality increased during the study period. Thyroidectomy is rarely performed during the same admission, with an overall perioperative complication rate of 30% and no effect on mortality. Patients with acute decompensated heart failure and renal failure were more likely to receive an operative intervention.

Electric scooter (ES)-related injuries are increasing but poorly described. Clinicians need more information to be prepared for these patients. We supposed two prevalent patterns of patients: mildly injured (predominant upper-limb injuries) and severely injured (predominant head trauma). This study aims to understand the frequency of ES-related injuries and patients\' characteristics despite the heterogeneity of data currently available. A systematic review with a proportion meta-analysis was conducted on studies with a multidisciplinary description of ES-related injuries in adult patients (PROSPERO-ID: CRD42022341241). Articles from inception to April 2023 were identified in MEDLINE, Embase, and Cochrane\'s databases. The risk of bias was evaluated using ROBINS-I. Twenty-five observational studies with 5387 patients were included in the meta-analysis, depending on reported data. Upper-limb (31.8%) and head (19.5%) injuries are the most frequent (25/25 studies included). When injured while riding, 19.5% of patients are intoxicated with drugs/alcohol, and only 3.9% use a helmet, increasing the possibility of severe injuries. About 80% of patients are victims of spontaneous falls. Half of the patients self-present to the ED, and 69.4% of cases are discharged directly from the ED. Studies\' limitations include an overall moderate risk of bias and high heterogeneity. Electric scooter-related accidents are commonly associated with upper-limb injuries but often involve the head. Spontaneous falls are the most common mechanism of injury, probably related to frequent substance abuse and helmet misuse. This hot topic is not adequately investigated due to a lack of data. A prospective registry could fill this gap.

Purpose To investigate the role of quantitative US (QUS) radiomics data obtained after the 1st week of radiation therapy (RT) in predicting treatment response in individuals with head and neck squamous cell carcinoma (HNSCC). Materials and Methods This prospective study included 55 participants (21 with complete response [median age, 65 years {IQR: 47-80 years}, 20 male, one female; and 34 with incomplete response [median age, 59 years {IQR: 39-79 years}, 33 male, one female) with bulky node-positive HNSCC treated with curative-intent RT from January 2015 to October 2019. All participants received 70 Gy of radiation in 33-35 fractions over 6-7 weeks. US radiofrequency data from metastatic lymph nodes were acquired prior to and after 1 week of RT. QUS analysis resulted in five spectral maps from which mean values were extracted. We applied a gray-level co-occurrence matrix technique for textural analysis, leading to 20 QUS texture and 80 texture-derivative parameters. The response 3 months after RT was used as the end point. Model building and evaluation utilized nested leave-one-out cross-validation. Results Five delta (Δ) parameters had statistically significant differences (P < .05). The support vector machines classifier achieved a sensitivity of 71% (15 of 21), a specificity of 76% (26 of 34), a balanced accuracy of 74%, and an area under the receiver operating characteristic curve of 0.77 on the test set. For all the classifiers, the performance improved after the 1st week of treatment. Conclusion A QUS Δ-radiomics model using data obtained after the 1st week of RT from individuals with HNSCC predicted response 3 months after treatment completion with reasonable accuracy. Keywords: Computer-Aided Diagnosis (CAD), Ultrasound, Radiation Therapy/Oncology, Head/Neck, Radiomics, Quantitative US, Radiotherapy, Head and Neck Squamous Cell Carcinoma, Machine Learning Clinicaltrials.gov registration no. NCT03908684 Supplemental material is available for this article. © RSNA, 2024.

Purpose To compare the effectiveness of weak supervision (ie, with examination-level labels only) and strong supervision (ie, with image-level labels) in training deep learning models for detection of intracranial hemorrhage (ICH) on head CT scans. Materials and Methods In this retrospective study, an attention-based convolutional neural network was trained with either local (ie, image level) or global (ie, examination level) binary labels on the Radiological Society of North America (RSNA) 2019 Brain CT Hemorrhage Challenge dataset of 21 736 examinations (8876 [40.8%] ICH) and 752 422 images (107 784 [14.3%] ICH). The CQ500 (436 examinations; 212 [48.6%] ICH) and CT-ICH (75 examinations; 36 [48.0%] ICH) datasets were employed for external testing. Performance in detecting ICH was compared between weak (examination-level labels) and strong (image-level labels) learners as a function of the number of labels available during training. Results On examination-level binary classification, strong and weak learners did not have different area under the receiver operating characteristic curve values on the internal validation split (0.96 vs 0.96; P = .64) and the CQ500 dataset (0.90 vs 0.92; P = .15). Weak learners outperformed strong ones on the CT-ICH dataset (0.95 vs 0.92; P = .03). Weak learners had better section-level ICH detection performance when more than 10 000 labels were available for training (average f1 = 0.73 vs 0.65; P < .001). Weakly supervised models trained on the entire RSNA dataset required 35 times fewer labels than equivalent strong learners. Conclusion Strongly supervised models did not achieve better performance than weakly supervised ones, which could reduce radiologist labor requirements for prospective dataset curation. Keywords: CT, Head/Neck, Brain/Brain Stem, Hemorrhage Supplemental material is available for this article. © RSNA, 2023 See also commentary by Wahid and Fuentes in this issue.

Interpretation of posttreatment imaging findings in patients with head and neck cancer can pose a substantial challenge. Malignancies in this region are often managed through surgery, radiation therapy, chemotherapy, and newer approaches like immunotherapy. After treatment, patients may experience various expected changes, including mucositis, soft-tissue inflammation, laryngeal edema, and salivary gland inflammation. Imaging techniques such as CT, MRI, and PET scans help differentiate these changes from tumor recurrence. Complications such as osteoradionecrosis, chondroradionecrosis, and radiation-induced vasculopathy can arise because of radiation effects. Radiation-induced malignancies may occur in the delayed setting. This review article emphasizes the importance of posttreatment surveillance imaging to ensure proper care of patients with head and neck cancer and highlights the complexities in distinguishing between expected treatment effects and potential complications. Keywords: CT, MR Imaging, Radiation Therapy, Ear/Nose/Throat, Head/Neck, Nervous-Peripheral, Bone Marrow, Calvarium, Carotid Arteries, Jaw, Face, Larynx © RSNA, 2024.

BACKGROUND: Blunt cerebrovascular injury (BCVI) with concurrent traumatic brain injury (TBI) presents increased risk of both ischemic stroke and bleeding. This study investigated the safety and survival benefit of BCVI treatment (antithrombotic and/or anticoagulant therapy) in this population. We hypothesized that treatment would be associated with fewer and later strokes in patients with BCVI and TBI without increasing bleeding complications.
METHODS: Patients with head AIS >0 were selected from a database of BCVI patients previously obtained for an observational trial. A Kaplan-Meier analysis compared stroke survival in patients who received BCVI treatment to those who did not. Logistic regression was used to evaluate for confounding variables.
RESULTS: Of 488 patients, 347 (71.1%) received BCVI treatment and 141 (28.9%) did not. BCVI treatment was given at a median of 31 h post-admission. BCVI treatment was associated with lower stroke rate (4.9% vs 24.1%, P < .001 and longer stroke-free survival (P < .001), but also less severe systemic injury. Logistic regression identified motor GCS and BCVI treatment as the only predictors of stroke. No patients experienced worsening TBI because of treatment.
CONCLUSIONS: Patients with BCVI and TBI who did not receive BCVI treatment had an increased rate of stroke early in their hospital stay, though this effect may be confounded by worse motor deficits and systemic injuries. BCVI treatment within 2-3 days of admission may be safe for patients with mean head AIS of 2.6. Future prospective trials are needed to confirm these findings and determine optimal timing of BCVI treatment in TBI patients with BCVI.

Purpose To evaluate and report the performance of the winning algorithms of the Radiological Society of North America Cervical Spine Fracture AI Challenge. Materials and Methods The competition was open to the public on Kaggle from July 28 to October 27, 2022. A sample of 3112 CT scans with and without cervical spine fractures (CSFx) were assembled from multiple sites (12 institutions across six continents) and prepared for the competition. The test set had 1093 scans (private test set: n = 789; mean age, 53.40 years ± 22.86 [SD]; 509 males; public test set: n = 304; mean age, 52.51 years ± 20.73; 189 males) and 847 fractures. The eight top-performing artificial intelligence (AI) algorithms were retrospectively evaluated, and the area under the receiver operating characteristic curve (AUC) value, F1 score, sensitivity, and specificity were calculated. Results A total of 1108 contestants composing 883 teams worldwide participated in the competition. The top eight AI models showed high performance, with a mean AUC value of 0.96 (95% CI: 0.95, 0.96), mean F1 score of 90% (95% CI: 90%, 91%), mean sensitivity of 88% (95% Cl: 86%, 90%), and mean specificity of 94% (95% CI: 93%, 96%). The highest values reported for previous models were an AUC of 0.85, F1 score of 81%, sensitivity of 76%, and specificity of 97%. Conclusion The competition successfully facilitated the development of AI models that could detect and localize CSFx on CT scans with high performance outcomes, which appear to exceed known values of previously reported models. Further study is needed to evaluate the generalizability of these models in a clinical environment. Keywords: Cervical Spine, Fracture Detection, Machine Learning, Artificial Intelligence Algorithms, CT, Head/Neck Supplemental material is available for this article. © RSNA, 2024.