OBJECTIVE: To develop a multi-instance learning (MIL) based artificial intelligence (AI)-assisted diagnosis models by using laryngoscopic images to differentiate benign and malignant vocal fold leukoplakia (VFL).
METHODS: The AI system was developed, trained and validated on 5362 images of 551 patients from three hospitals. Automated regions of interest (ROI) segmentation algorithm was utilized to construct image-level features. MIL was used to fusion image level results to patient level features, then the extracted features were modeled by seven machine learning algorithms. Finally, we evaluated the image level and patient level results. Additionally, 50 videos of VFL were prospectively gathered to assess the system\'s real-time diagnostic capabilities. A human-machine comparison database was also constructed to compare the diagnostic performance of otolaryngologists with and without AI assistance.
RESULTS: In internal and external validation sets, the maximum area under the curve (AUC) for image level segmentation models was 0.775 (95 % CI 0.740-0.811) and 0.720 (95 % CI 0.684-0.756), respectively. Utilizing a MIL-based fusion strategy, the AUC at the patient level increased to 0.869 (95 % CI 0.798-0.940) and 0.851 (95 % CI 0.756-0.945). For real-time video diagnosis, the maximum AUC at the patient level reached 0.850 (95 % CI, 0.743-0.957). With AI assistance, the AUC improved from 0.720 (95 % CI 0.682-0.755) to 0.808 (95 % CI 0.775-0.839) for senior otolaryngologists and from 0.647 (95 % CI 0.608-0.686) to 0.807 (95 % CI 0.773-0.837) for junior otolaryngologists.
CONCLUSIONS: The MIL based AI-assisted diagnosis system can significantly improve the diagnostic performance of otolaryngologists for VFL and help to make proper clinical decisions.

It has been found that progression from leukoplakia to head and neck squamous cell carcinoma (HNSCC) is a long-term process that may involve changes in the multicellular ecosystem. We acquired scRNA-seq samples information from gene expression omnibus and UCSC Xena database. The BEAM function was used to construct the pseudotime trajectory and analyze the differentially expressed genes in different branches. We used the ssGSEA method to explore the correlation between each cell subgroup and survival time, and obtained the cell subgroup related to prognosis. During the progression from leukoplakia to HNSCC, we found several prognostic cell subgroups, such as AURKB + epithelial cells, SFRP1 + fibroblasts, SLC7A8 + macrophages, FCER1A + CD1C + dendritic cells, and TRGC2 + NK/T cells. All cell subgroups had two different fates, one tending to cell proliferation, migration, and enhancement of angiogenesis capacity, and the other tending to inflammatory immune response, leukocyte chemotaxis, and T cell activation. Tumor-promoting genes such as CD163 and CD209 were highly expressed in the myeloid cells, and depletion marker genes such as TIGIT, LAG3 were highly expressed in NK/T cells. Our study may provide a reference for the molecular mechanism of HNSCC and theoretical basis for the development of new therapeutic strategies.

BACKGROUND: Laryngeal leukoplakia (LL) is a white lesion with high potential of recurrence and malignant transformation. Currently, CO2 laser has become the primary surgical treatment for LL, and the recurrence and malignant transformation rates after treatment vary widely.
OBJECTIVE: We performed a systematic review and meta-analysis dedicated to evaluating the rates of recurrence and malignant transformation of LL lesions treated with CO2 laser and exploring relevant risk factors for recurrence or malignant transformation.
METHODS: Literature searches were conducted on ProQuest, PubMed, Web of Science, Ovid Medline, Embase, and Cochrane databases. Some articles identified through hand searching were included.
RESULTS: A total of 14 articles and 1462 patients were included in this review. Pooled results showed that the overall recurrence rate was 15%, and the malignant transformation rate was 3%. Subgroup analysis showed that the dysplasia grade was not a significant risk factor for the recurrence and malignant transformation of LL (P > .05).
CONCLUSIONS: The results of this systematic review and meta-analysis suggest that the CO2 laser is a safe and effective surgical instrument for the excision of LL, which yields low rates of recurrence and malignant transformation. The risk factors relevant to recurrence or malignant transformation remain unclear and require further investigation.

OBJECTIVE: To analyse the comparative clinical outcomes and clinicopathological significance of vocal fold leukoplakia lesions treated by appearance classification and traditional methods.
METHODS: A total of 1442 vocal fold leukoplakia patients were enrolled. Group A patients were treated according to appearance classification and Group B patients were treated according to traditional methods.
RESULTS: In Group A, 24.4, 14.9 and 60.6 per cent of patients had grade I, II and III dysplasia, respectively. Grade I dysplasia (63.4 per cent) was more than twice as frequent in Group B patients than in Group A patients, while grade II dysplasia (20.4 per cent) and grade III dysplasia (16.2 per cent) were significantly less frequent in Group B patients than in Group A patients (p = 0.000). There was a significant correlation between vocal fold leukoplakia appearance and the degree of dysplasia (p = 0.000). The recurrence and malignant transformation rates (17.6 and 31 per cent, respectively) in Group B were significantly greater than those in Group A (10.8 and 25.9 per cent, respectively) (p = 0.000).
CONCLUSIONS: Vocal fold leukoplakia appearance classification is useful for guiding treatment decision-making and could help to improve therapeutic accuracy.

OBJECTIVE: This study evaluated vocal fold leukoplakia using i-scan combined with laryngovideostroboscopy for risk assessment prediction.
METHODS: A total of 141 patients with 218 lesions were enrolled in this study. Morphological characteristics of leukoplakia, assessment of the vascular pattern using i-scan, and vocal fold vibratory function were analyzed.
RESULTS: The number of patients with no, mild, moderate, severe dysplasia, and invasive carcinoma were 68, 40, 17, 46 and 47, respectively. The sensitivity of morphological characteristic, vascular pattern, vibratory function and predictive model were 77.4%, 72%, 69.9%, and 82.8%, respectively. Receiver operating characteristic curve analysis of morphological characteristic, vascular pattern, vibratory function and predictive model were 0.771, 0.824, 0.769, and 0.923, respectively. The results of logistic regression analysis showed that rough morphological types, perpendicular vascular pattern, severe decrease and absence of mucosal waves increased the risk of malignancy (OR = 5.531, 4.973, and 16.992, respectively; P < 0.001).
CONCLUSIONS: I-scan combined with laryngovideostroboscopy can improve the differential diagnosis of low-risk and high-risk vocal fold leukoplakia.

OBJECTIVE: Accurate vocal cord leukoplakia classification is instructive for clinical diagnosis and surgical treatment. This article introduces a reliable very deep Siamese network for accurate vocal cord leukoplakia classification.
METHODS: A study of a classification network based on a retrospective database.
METHODS: Academic university and hospital.
METHODS: The white light image datasets of vocal cord leukoplakia used in this article were classified into 6 classes: normal tissues, inflammatory keratosis, mild dysplasia, moderate dysplasia, severe dysplasia, and squamous cell carcinoma. The classification performance was assessed by comparing it with 6 classical deep learning models, including AlexNet, VGG Net, Google Inception, ResNet, DenseNet, and Vision Transformer.
RESULTS: Experiments show the superior classification performance of our proposed network compared to state-of-the-art methods. The overall accuracy is 0.9756. The values of sensitivity and specificity are very high as well. The confusion matrix provides information for the 6-class classification task and demonstrates the superiority of our proposed network.
CONCLUSIONS: Our very deep Siamese network can provide accurate classification results of vocal cord leukoplakia, which facilitates early detection, clinical diagnosis, and surgical treatment. The excellent performance obtained in white light images can reduce the cost for patients, especially those living in developing countries.

OBJECTIVE: To propose a scoring system based on laryngoscopic characteristics for the differential diagnosis of benign and malignant vocal fold leukoplakia.
METHODS: Laryngoscopic images from 200 vocal fold leukoplakia cases were retrospectively analysed. The laryngoscopic signs of benign and malignant vocal fold leukoplakia were compared, and statistically significant features were assigned and accumulated to establish the leukoplakia finding score.
RESULTS: A total of five indicators associated with malignant vocal fold leukoplakia were included to construct the leukoplakia finding score, with a possible range of 0-10 points. A score of 6 points or more was indicative of a diagnosis of malignant vocal fold leukoplakia. The sensitivity, specificity and accuracy values of the leukoplakia finding score were 93.8 per cent, 83.6 per cent and 86.0 per cent, respectively. The consistency in the leukoplakia finding score obtained by different laryngologists was strong (kappa = 0.809).
CONCLUSIONS: This scoring system based on laryngoscopic characteristics has high diagnostic value for distinguishing benign and malignant vocal fold leukoplakia.

Accurate vocal cord leukoplakia classification is critical for the individualized treatment and early detection of laryngeal cancer. Numerous deep learning techniques have been proposed, but it is unclear how to select one to apply in the laryngeal tasks. This article introduces and reliably evaluates existing deep learning models for vocal cord leukoplakia classification.
We created white light and narrow band imaging (NBI) image datasets of vocal cord leukoplakia which were classified into six classes: normal tissues (NT), inflammatory keratosis (IK), mild dysplasia (MiD), moderate dysplasia (MoD), severe dysplasia (SD), and squamous cell carcinoma (SCC). Vocal cord leukoplakia classification was performed using six classical deep learning models, AlexNet, VGG, Google Inception, ResNet, DenseNet, and Vision Transformer.
GoogLeNet (i.e., Google Inception V1), DenseNet-121, and ResNet-152 perform excellent classification. The highest overall accuracy of white light image classification is 0.9583, while the highest overall accuracy of NBI image classification is 0.9478. These three neural networks all provide very high sensitivity, specificity, and precision values.
GoogLeNet, ResNet, and DenseNet can provide accurate pathological classification of vocal cord leukoplakia. It facilitates early diagnosis, providing judgment on conservative treatment or surgical treatment of different degrees, and reducing the burden on endoscopists.

To investigate the role of H+ /K+ ATPase in the proliferation of pepsin-induced vocal cord leukoplakia (VCL) cells.
Translation research.
Affiliated Hospital of University.
Immunohistochemistry was used to detect pepsin, H+ /K+ ATPase (ATP4A and ATP4B subunits) in VCL cells with varying degrees of dysplasia. After primary cultures of VCL cells had been established, the effects of acidified pepsin on the proliferation, autophagy, and H+ /K+ -ATPase distribution of VCL cells were investigated.
The levels of pepsin, ATP4A, and ATP4B were significantly higher in VCL tissue with moderate-to-severe dysplasia than in normal tissue (p < .05); these levels gradually increased according to dysplasia severity. The expression levels of ATP4A and ATP4B were significantly correlated with the amount of pepsin in VCL cells (p < .01). Acidified pepsin enhanced the levels of proliferation and autophagy in human VCL epithelial cells. The cloning- and autophagy-promoting effects of acidified pepsin on VCL cells were partially reversed by pantoprazole; these effects were completely blocked by the autophagy inhibitor chloroquine. Finally, acidified pepsin promoted the colocalization of H+ /K+ -ATPase and lysosomes in VCL cells; it also mediated lysosome acidification.
Pepsin and H+ /K+ -ATPase may contribute to the progression of VCL. Specifically, acidified pepsin may regulate lysosome acidification by promoting lysosomal localization of H+ /K+ -ATPase.

OBJECTIVE: To evaluate risk factors for poor prognosis in vocal fold leukoplakia.
METHODS: Clinical data were collected for 344 patients with vocal fold leukoplakia who received surgical treatment in our otolaryngology department from October 2010 to June 2019. Univariate and multivariate logistic regression analyses of the relevant factors were conducted.
RESULTS: Among the 344 patients, 98 exhibited recurrence and 30 underwent a malignant change. Multivariate logistic regression analysis showed that size of the lesion (p = 0.03, odds ratio = 2.14), form of the lesion under white light (p < 0.001), surgical method (p < 0.001, odds ratio = 0.28) and pathological type (p < 0.001) were independent factors that affected the recurrence of vocal fold leukoplakia. In both univariate and multivariate analyses, the sole independent risk factor for malignant transformation of vocal fold leukoplakia was pathological type (p < 0.001).
CONCLUSIONS: The outlook for vocal fold leukoplakia depends on several clinical factors, especially pathological type. The more severe the pathological type, the more likely it is to recur or become cancerous.