Objective:To develop a deep learning model based on 3D super-resolution reconstruction technology and to analyze its feasibility and effectiveness in predicting paraglottic space invasion in hypopharyngeal cancer.Methods:A retrospective study was conducted involving 382 patients with hypopharyngeal squamous cell carcinoma treated at Qilu Hospital of Shandong University between January 2014 and December 2020. The cohort included 364 males and 18 females, with a mean age of 62±7 years. Patients were divided into a training set ( n=300) and a test set ( n=82) based on enrollment time. A generative adversarial network was used to perform 3D super-resolution reconstruction on contrast-enhanced CT images, improving spatial resolution by 16 times. A 2.5D deep learning strategy was employed to construct Resnet-NR and Resnet-SR models based on conventional and super-resolution images, respectively, to predict whether the paraglottic space was invaded. Model performance was evaluated using receiver operating characteristic (ROC) curves and area under the curve (AUC). A multi-reader multi-case study was conducted to assess the impact of the artificial intelligence (AI) model on clinicians′ diagnostic capabilities. Results:The super-resolution model Resnet-SR achieved the highest accuracy in both the training set (AUC=0.87, 95% CI: 0.84-0.90) and the test set (AUC=0.88, 95% CI: 0.81-0.96), significantly outperforming traditional clinical indicators (T stage, N stage, tumor diameter, and pathological differentiation degree) (AUC range: 0.55-0.70, all P<0.05). In comparison, the conventional-resolution model Resnet-NR achieved AUCs of 0.81 (95% CI: 0.77-0.84, P=0.005) and 0.80 (95% CI: 0.71-0.89, P=0.184) in the training and test sets, respectively. Using Resnet-SR to assist clinical decision-making improved the diagnostic accuracy of junior physicians (AUC=0.793 without AI assistance vs. AUC=0.871 with AI assistance, P=0.012) and significantly reduced diagnosis time for clinicians of all experience levels (86.5 s without AI assistance vs. 82.5 s with AI assistance, t=2.01, P=0.032). Conclusion:This study successfully develops a deep learning model based on 3D super-resolution reconstruction technology, which can assist in preoperative prediction of paraglottic space invasion in hypopharyngeal cancer. The AI-assisted tool improves diagnostic accuracy for junior physicians and enhances diagnostic efficiency for clinicians across all experience levels.

Objective:To develop a deep learning model based on 3D super-resolution reconstruction technology and to analyze its feasibility and effectiveness in predicting paraglottic space invasion in hypopharyngeal cancer.Methods:A retrospective study was conducted involving 382 patients with hypopharyngeal squamous cell carcinoma treated at Qilu Hospital of Shandong University between January 2014 and December 2020. The cohort included 364 males and 18 females, with a mean age of 62±7 years. Patients were divided into a training set ( n=300) and a test set ( n=82) based on enrollment time. A generative adversarial network was used to perform 3D super-resolution reconstruction on contrast-enhanced CT images, improving spatial resolution by 16 times. A 2.5D deep learning strategy was employed to construct Resnet-NR and Resnet-SR models based on conventional and super-resolution images, respectively, to predict whether the paraglottic space was invaded. Model performance was evaluated using receiver operating characteristic (ROC) curves and area under the curve (AUC). A multi-reader multi-case study was conducted to assess the impact of the artificial intelligence (AI) model on clinicians′ diagnostic capabilities. Results:The super-resolution model Resnet-SR achieved the highest accuracy in both the training set (AUC=0.87, 95% CI: 0.84-0.90) and the test set (AUC=0.88, 95% CI: 0.81-0.96), significantly outperforming traditional clinical indicators (T stage, N stage, tumor diameter, and pathological differentiation degree) (AUC range: 0.55-0.70, all P<0.05). In comparison, the conventional-resolution model Resnet-NR achieved AUCs of 0.81 (95% CI: 0.77-0.84, P=0.005) and 0.80 (95% CI: 0.71-0.89, P=0.184) in the training and test sets, respectively. Using Resnet-SR to assist clinical decision-making improved the diagnostic accuracy of junior physicians (AUC=0.793 without AI assistance vs. AUC=0.871 with AI assistance, P=0.012) and significantly reduced diagnosis time for clinicians of all experience levels (86.5 s without AI assistance vs. 82.5 s with AI assistance, t=2.01, P=0.032). Conclusion:This study successfully develops a deep learning model based on 3D super-resolution reconstruction technology, which can assist in preoperative prediction of paraglottic space invasion in hypopharyngeal cancer. The AI-assisted tool improves diagnostic accuracy for junior physicians and enhances diagnostic efficiency for clinicians across all experience levels.

Objective:To obtain the prevalence laryngopharyngeal reflux disease (LPRD), anxiety and depression in otorhinolaryngology outpatients and to explore the role of mental and psychological factors (anxiety and depression) in their pathogenesis.Methods:A questionnaire survey of reflux symptom index(RSI) scale and hospital anxiety and depression (HAD) scale were used to report 1 111 cases of outpatients in Department of Otorhinolaryngology Head and Neck Surgery, Peking University People′s Hospital, from July 2017 to June 2018 (486 males, 625 females, age of 18-96 years old, median age of 38[30,53] years old) and to obtain the prevalence of LPRD, anxiety and depression. RSI-positive patients were selected in the case group, and RSI-negative patients were selected in the control group. The differences in HAD scores between the two groups were compared, and the risk factors of laryngopharyngeal reflux were analyzed. Statistical analysis was performed using SPSS 20.0 software.Results:There were 151 cases in the case group and 960 cases in the control group. The prevalence of LPRD was 13.59% (151/1 111).There was no significant difference in the prevalence of LPRD between different genders ( P>0.05). The prevalence rate was the highest in the 18-40 age group, and the difference in the prevalence of all age groups (18~ 40 years old; 41-65 years old; >65 years old) was statistically significant ( P<0.05). The prevalence of LPRD among smokers and non-drinkers was higher than that of non-smokers and non-drinkers and the prevalence of the two groups was statistically significant ( P<0.05). The most common symptoms of the RSI scale were pharyngeal foreign body sensation (92.72%,140/151), persistent clearing throat (88.74%,124/151), excessive sputum or nasal reflux (82.12%, 124/151). There were significant statistical differences between the two groups ( P<0.05). Ninty-one patients with anxiety, the prevalence was 8.19%(91/1 111); 76 patients with depression, the prevalence was 6.84%(76/1 111).Among the LPRD patients, the hospital anxiety scale scored 29.14% (44/151), and the hospital depression scale scored 17.22% (26/151). The scores of anxiety symptoms and depressive symptoms in the LPRD group were higher than those in the non-LPRD group. The above scores were statistically significant ( P<0.05). Logistic regression analysis showed that smoking, anxiety and symptoms of gastroesophageal reflux disease were independent risk factors for laryngopharyngeal reflux. Conclusions:The prevalences of LPRD, anxiety and depression in the otorhinolaryngology clinic are 13.59%, 8.19% and 6.84%, respectively. Among patients with laryngopharyngeal reflux, the prevalence of anxiety is 29.14%, and the prevalence of depression is 17.22%. Age, smoking, drinking, alcohol consumption, education level, course of disease, symptoms of gastroesophageal reflux disease, pharyngeal foreign body sensation, etc. are related to LPRD. Mental factors (anxiety and depression) may play a role in LPRD. Smoking, anxiety symptoms and symptoms of gastroesophageal reflux disease are closely related to the incidence of LPRD.

Objective: To establish a New Zealand rabbit animal model of laryngopharyngeal reflux disease (LPRD) using esophageal balloon together with metal internal stent dilation and to investigate the changes of mucosa. Methods: 20 New Zealand rabbits were randomly divided into experimental group and control group, with 10 in each group. Balloon dilatation and metal internal stent dilation were carried out in experimental group to reproduce the animal model of LPRD.The middle of balloon was placed at the lower esophageal sphincter (LES) while the stent was placed at the upper esophageal sphincter (UES). The guide wire was placed in the control group, but the balloon was not expanded and the stent was not placed. The general condition, pH value of hypopharynx, laryngeal histopathology and changes of pepsin content of New Zealand rabbits were observed regularly. The difference between experimental group and control group was compared. Results: The 24-hour Dx-pH monitoring results showed that the number of reflux episodes(20.0[9.5, 35.0], 13.0[6.5, 22.0]), and the percent time below pH 5.5 (1.36%[0.60%, 4.57%], 1.36%[0.43%, 2.77%]) in the experimental group at the 2nd and 4th week were significantly different from those in the control group (0[0,3.0], 1.0[0.5, 3.8]; 0[0, 0.01%], 0[0, 0], respectively, all P<0.01), suggesting that the experimental group New Zealand rabbits developed LPRD. Compared with the control group under microscope, lymphocytes infiltration and submucosal gland hyperplasia increased in the mucosa of the throat of the experimental group. The results of pepsin immunohistochemical staining between the two groups were statistically significant (P=0.014). Conclusion The use of balloon dilatation of the LES combined with metal stent dilatation of the UES can successfully establish a laryngopharyngeal reflux model, and lesions in the throat tissue can be observed.

Objective: To investigate the pathogenic factors of vocal leukoplakia and its clinical and pathological features. Methods: Eighty-one patients with vocal cord leukoplakia who underwent surgery between February 2010 and December 2016 and 160 volunteers without pharyngeal symptoms designed as controls were included in this case control study. The clinicopathological characteristics of 81 patients were summarized and analyzed synthetically. Results: There was statistical significance in reflux symptom index(RSI), reflux finding score(RFS), smoking index, and drinking index between case group and control group(Z=-5.35, -4.82, -4.76, -2.44, P<0.05). The voice-using duration per day in case group was significantly longer than that of control group.There was no statistical significance in hospital anxiety and depression scale for anxiety(HADA) scores、hospital anxiety and depression scale for depression(HADD) scores between case group and control group(P>0.05). In 42 patients who received 24-hour dual probe pH monitoring the prevalence of pathologic LPR was 42.8%. In 81 patients, 39(48%)patients were pathologically diagnosed as squamous cell hyperplasia, 18(22%)patients as mild dysplasia, 12(15%)sides as moderate dysplasia , 10(12%)patients as severe dysplasia and 2(2%)patients as carcinoma in-situ. The average age of high-risk pathological vocal leukoplakia was significantly higher than that of low-risk leukoplakia(t=-2.73, P<0.01). The propotion of speckled leukoplakia in high-risk leukoplakia was significantly higher than that of low-risk leukoplakia(χ²=23.81, P<0.01). There was no statistical significance between high-risk leukoplakia and low-risk leukoplakia in the prevalence of pathologic LPR(P>0.05). The bilateral lesions, speckled leukoplakia were more likely to relapse(χ²=4.27, 12.17, P<0.05). The more serious the pathology, the more likely it was to relapse (Z=-2.168, P=0.03). There was no statistical significance between recurrence group and non-recurrence group in the prevalence of pathologic LPR(P>0.05). Conclusions LPR, smoke constitute the risk factors of vocal cord leukoplakia. Drinking, voice abuse are related to vocal cord leukoplakia. Senile, speckled leukoplakia are more likely to be malignancy. A speckled leukoplakia, bilateral leukoplakia, severe pathological degree are important factors to predict recurrence.