'Precursory symptoms, awareness or progression of facial palsy' are more useful than 'forehead wrinkling ability' in differentiating central facial palsy examined in the emergency department.

BACKGROUND: Forehead wrinkling ability has been considered to be the sign of the central facial palsy (CFP). AIMS/OBJECTIVES: To identify characteristics of peripheral FP (PFP) patients in the emergency room (ER), differentiate PFP from central FP (CFP), and assess the utility of forehead wrinkling for this purpose. MATERIALS AND METHODS: ER patients with FP were clinically split into PFP (72 patients) and CFP (161 patients) groups. Factors like age, sex, medical history, time from onset to consultation, symptom awareness or progression, precursory symptoms, forehead wrinkling, and imaging history were compared. Multivariate analysis differentiated PFP from CFP, examining misdiagnosis risks based on forehead wrinkling. RESULTS: Precursory symptoms and symptom awareness or progression had the highest odds ratios. Some PFP patients could wrinkle their foreheads, typically examined within 1 day of symptoms. PFP patients had more same-day imaging than those assessed a day later. CONCLUSIONS AND SIGNIFICANCE: Forehead wrinkling, a traditional CFP sign, is also common in early-stage PFP, decreasing its diagnostic reliability. Patients with solely CFP unable to wrinkle the forehead are very rare at a single institution. Evaluating precursors symptoms, and FP awareness and progression is crucial for differentiation.

Preoperative prediction of sinonasal papilloma by artificial intelligence using nasal video endoscopy: a retrospective study.

Sinonasal inverted papilloma (IP) is at risk of recurrence and malignancy, and early diagnosis using nasal endoscopy is essential. We thus developed a diagnostic system using artificial intelligence (AI) to identify nasal sinus papilloma. Endoscopic surgery videos of 53 patients undergoing endoscopic sinus surgery were edited to train and evaluate deep neural network models and then a diagnostic system was developed. The correct diagnosis rate based on visual examination by otolaryngologists was also evaluated using the same videos and compared with that of the AI diagnostic system patients. Main outcomes evaluated included the percentage of correct diagnoses compared to AI diagnosis and the correct diagnosis rate for otolaryngologists based on years of practice experience. The diagnostic system had an area under the curve of 0.874, accuracy of 0.843, false positive rate of 0.124, and false negative rate of 0.191. The average correct diagnosis rate among otolaryngologists was 69.4%, indicating that the AI was highly accurate. Evidently, although the number of cases was small, a highly accurate diagnostic system was created. Future studies with larger samples to improve the accuracy of the system and expand the range of diseases that can be detected for more clinical applications are warranted.

Preoperative prediction by artificial intelligence for mastoid extension in pars flaccida cholesteatoma using temporal bone high-resolution computed tomography: A retrospective study.

Cholesteatoma is a progressive middle ear disease that can only be treated surgically but with a high recurrence rate. Depending on the extent of the disease, a surgical approach, such as microsurgery with a retroarticular incision or transcanal endoscopic surgery, is performed. However, the current examination cannot sufficiently predict the progression before surgery, and changes in approach may be made during the surgery. Large amounts of data are typically required to train deep neural network models; however, the prevalence of cholesteatomas is low (1-in-25, 000). Developing analysis methods that improve the accuracy with such a small number of samples is an important issue for medical artificial intelligence (AI) research. This paper presents an AI-based system to automatically detect mastoid extensions using CT. This retrospective study included 164 patients (80 with mastoid extension and 84 without mastoid extension) who underwent surgery. This study adopted a relatively lightweight neural network model called MobileNetV2 to learn and predict the CT images of 164 patients. The training was performed with eight divided groups for cross-validation and was performed 24 times with each of the eight groups to verify accuracy fluctuations caused by randomly augmented learning. An evaluation was performed by each of the 24 single-trained models, and 24 sets of ensemble predictions with 23 models for 100% original size images and 400% zoomed images. Fifteen otolaryngologists diagnosed the images and compared the results. The average accuracy of predicting 400% zoomed images using ensemble prediction model was 81.14% (sensitivity = 84.95%, specificity = 77.33%). The average accuracy of the otolaryngologists was 73.41% (sensitivity, 83.17%; specificity, 64.13%), which was not affected by their clinical experiences. Noteworthily, despite the small number of cases, we were able to create a highly accurate AI. These findings represent an important first step in the automatic diagnosis of the cholesteatoma extension.