Point-wise spatial network for identifying carcinoma at the upper digestive and respiratory tract

Sep 24, 2026

Uploaded by Cipher

Research on artificial intelligence for diagnosing oral pharyngeal, nasopharyngeal, or laryngeal carcinoma is extensive, yet comprehensive models for these regions remain limited. This study tests the hypothesis that these cancerous appearances share a common pattern that can be integrated into one framework. A point-wise spatial attention network is used to perform semantic segmentation. Results show strong performance, with average mIoU of 86.3% and average pixel accuracy of 96.3%. The findings suggest a single deep learning model can recognize tumors across these mucosal regions.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages8

Reading Time20 min

AccessFree to download

Abstract

Introduction

Background and clinical importance

Related AI research

Methods and model approach

Results

Conclusion

Zhou et al. BMC Medical Imaging (2023) 23:140 [https://doi.org/10.1186/s12880-023-01076-5](https://doi.org/10.1186/s12880-023-01076-5) BMC Medical Imaging RESEARCH Open Access Point-wise spatial network for identifying carcinoma at the upper digestive and respiratory tract Lei Zhou1†, Huaili Jiang1†, Guangyao Li1, Jiaye Ding1, Cuicui Lv1, Maoli Duan2,3, Wenfeng Wang4, Kongyang Chen4,5, Na Shen1* and Xinsheng Huang1* Abstract Problem Artificial intelligence has been widely investigated for diagnosis and treatment strategy design, with some models proposed for detecting oral pharyngeal, nasopharyngeal, or laryngeal carcinoma. However, no comprehensive model has been established for these regions. Aim Our hypothesis was that a common pattern in the cancerous appearance of these regions could be recognized and integrated into a single model, thus improving the efficacy of deep learning models. Methods We utilized a point-wise spatial attention network model to perform semantic segmentation in these regions. Results Our study demonstrated an excellent outcome, with an average mIoU of 86 . 3%, and an average pixel accuracy of 96. 3% . Conclusion The research confirmed that the mucosa of oral pharyngeal, nasopharyngeal, and laryngeal regions may share a common appearance, including the appearance of tumors, which can be recognized by a single artificial intelligence model. Therefore, a deep learning model could be constructed to effectively recognize these tumors. Keywords Artificial intelligence, Oral pharynx, Hypopharynx, Larynx, Nasopharynx †Lei Zhou and Huaili Jiang these two authors contribute equally to this manuscript. *Correspondence: Na Shen [shen.na@zs-hospital.sh.cn](shen.na@zs-hospital.sh.cn)[ ](shen.na@zs-hospital.sh.cn)Xinsheng Huang [huang.xinsheng@zs-hospital.sh.cn](huang.xinsheng@zs-hospital.sh.cn) 1 Department of Otorhinolaryngology-Head and Neck Surgery, Zhongshan Hospital Affiliated to Fudan University, Xuhui District, 180 Fenglin Road, , Shanghai 200032, P. R. China 2 Department of Clinical Science, Intervention and Technology, Karolinska Institutet, Stockholm, Sweden 3 Department of Otolaryngology Head and Neck Surgery, Karolinska University Hospital, 171 76 Stockholm, Sweden 4 Institute of Artificial Intelligence and Blockchain, Guangzhou University, Guangzhou 510006, P. R. China 5 Pazhou Lab, Guangzhou 510330, P. R. China Introduction In recent years, artificial intelligence (AI) technology has made significant strides in various fields of medicine, including the diagnosis of oral cancer [1, 2], dermatology disease [3], ocular fundus disease [4], lung cancer [5], pathological slices diagnosis [6] even the prediction of gene editing results [7]. AI is playing an increasingly important role in medicine, surpassing what was previously possible [8]. It may soon replace tedious or dangerous work with machines equipped with AI systems. Early detection of cancer has always been associated with a good prognosis. Therefore, detecting cancer in its early stages is crucial [1], regardless of whether it is done in the hospital or through self-diagnosis. © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit [http://creati

FAQs

What problem does the study address?

Existing AI models typically target specific regions such as oral pharynx, nasopharynx, or larynx, but no comprehensive single model covers these regions together. The study aims to fill this gap.

What method is used to analyze the mucosal regions?

A point-wise spatial attention network is applied to perform semantic segmentation on tissues from the upper digestive and respiratory tract regions.

How effective is the proposed model?

The model achieves an average mIoU of 86.3% and an average pixel accuracy of 96.3%, indicating strong segmentation performance.

Sep 24, 2026

Uploaded by Cipher