EdgeNeXt-SEDP for cervical adenocarcinoma HPV-associated and non-HPV-associated diagnosis and decision support

IF 5.1 2区医学 Q1 MEDICINE, RESEARCH & EXPERIMENTAL

Life sciences Pub Date : 2025-08-27 DOI:10.1016/j.lfs.2025.123931

Qi Chen , Hao Wang , Hao Zhang , Zhenkun Zhu , Xi Wei

{"title":"EdgeNeXt-SEDP for cervical adenocarcinoma HPV-associated and non-HPV-associated diagnosis and decision support","authors":"Qi Chen , Hao Wang , Hao Zhang , Zhenkun Zhu , Xi Wei","doi":"10.1016/j.lfs.2025.123931","DOIUrl":null,"url":null,"abstract":"<div><h3>Aims</h3><div>Adenocarcinoma of the uterine cervix exhibits substantial biological and histological heterogeneity, with subtype-specific differences in prognosis and therapeutic response. Conventional classification—based on histopathology, immunohistochemistry, and molecular testing—remains subjective, labor-intensive, and challenging to standardize. This study introduces EdgeNeXt-SEDP, a lightweight deep-learning framework for automated differentiation of HPV-associated (HPVA) and non-HPV-associated (NHPVA) subtypes from histopathological whole-slide images (WSIs).</div></div><div><h3>Materials and methods</h3><div>EdgeNeXt-SEDP integrates three synergistic components: a Squeeze-and-Excitation (SE) module to recalibrate channel-wise feature importance, dual-pooling feature fusion to enrich spatial representation, and progressive stochastic depth decay to enhance generalization. The model was trained and evaluated on 49 WSIs from 21 patients using standardized preprocessing, augmentation, and evaluation protocols. Performance metrics included accuracy, precision, specificity, and macro-averaged F1 score, benchmarked against DilateFormer, RepVIT, and EdgeNeXt architectures.</div></div><div><h3>Key findings</h3><div>EdgeNeXt-SEDP achieved 97.63% accuracy, 97.61% precision, 96.98% specificity, and a 97.58% macro-averaged F1 score, while maintaining computational efficiency with 1.9M parameters and 0.2G FLOPs. Ablation analyses confirmed that each module significantly contributed to performance, with the SE module yielding the largest gains. The proposed model consistently surpassed baseline methods without incurring additional computational cost.</div></div><div><h3>Significance</h3><div>By delivering high diagnostic accuracy in an efficient architecture, EdgeNeXt-SEDP offers a scalable and reliable solution for reducing interobserver variability and facilitating timely, individualized management of cervical adenocarcinoma. Its compact design supports integration into diverse clinical and resource-limited settings, advancing the application of AI in digital pathology.</div></div>","PeriodicalId":18122,"journal":{"name":"Life sciences","volume":"380 ","pages":"Article 123931"},"PeriodicalIF":5.1000,"publicationDate":"2025-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Life sciences","FirstCategoryId":"3","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0024320525005661","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"MEDICINE, RESEARCH & EXPERIMENTAL","Score":null,"Total":0}

引用次数: 0

Abstract

Aims

Adenocarcinoma of the uterine cervix exhibits substantial biological and histological heterogeneity, with subtype-specific differences in prognosis and therapeutic response. Conventional classification—based on histopathology, immunohistochemistry, and molecular testing—remains subjective, labor-intensive, and challenging to standardize. This study introduces EdgeNeXt-SEDP, a lightweight deep-learning framework for automated differentiation of HPV-associated (HPVA) and non-HPV-associated (NHPVA) subtypes from histopathological whole-slide images (WSIs).

Materials and methods

EdgeNeXt-SEDP integrates three synergistic components: a Squeeze-and-Excitation (SE) module to recalibrate channel-wise feature importance, dual-pooling feature fusion to enrich spatial representation, and progressive stochastic depth decay to enhance generalization. The model was trained and evaluated on 49 WSIs from 21 patients using standardized preprocessing, augmentation, and evaluation protocols. Performance metrics included accuracy, precision, specificity, and macro-averaged F1 score, benchmarked against DilateFormer, RepVIT, and EdgeNeXt architectures.

Key findings

EdgeNeXt-SEDP achieved 97.63% accuracy, 97.61% precision, 96.98% specificity, and a 97.58% macro-averaged F1 score, while maintaining computational efficiency with 1.9M parameters and 0.2G FLOPs. Ablation analyses confirmed that each module significantly contributed to performance, with the SE module yielding the largest gains. The proposed model consistently surpassed baseline methods without incurring additional computational cost.

Significance

By delivering high diagnostic accuracy in an efficient architecture, EdgeNeXt-SEDP offers a scalable and reliable solution for reducing interobserver variability and facilitating timely, individualized management of cervical adenocarcinoma. Its compact design supports integration into diverse clinical and resource-limited settings, advancing the application of AI in digital pathology.

Abstract Image

查看原文本刊更多论文

EdgeNeXt-SEDP用于宫颈癌hpv相关和非hpv相关的诊断和决策支持

目的子宫颈腺癌表现出明显的生物学和组织学异质性，在预后和治疗反应方面存在亚型特异性差异。基于组织病理学、免疫组织化学和分子检测的传统分类仍然是主观的、劳动密集型的，并且很难标准化。本研究引入了EdgeNeXt-SEDP，这是一个轻量级的深度学习框架，用于从组织病理学全片图像（wsi）中自动区分hpv相关（HPVA）和非hpv相关（NHPVA）亚型。sedgenext - sedp集成了三个协同组件：用于重新校准通道特征重要性的挤压和激励（SE）模块，用于丰富空间表示的双池特征融合，以及用于增强泛化的渐进随机深度衰减。采用标准化的预处理、增强和评估方案，对来自21例患者的49个wsi进行了模型训练和评估。性能指标包括准确性、精密度、特异性和宏观平均F1分数，以DilateFormer、RepVIT和EdgeNeXt架构为基准。关键发现sedgenext - sedp的准确率为97.63%，精密度为97.61%，特异性为96.98%，宏观平均F1评分为97.58%，同时保持了1.9M参数和0.2G FLOPs的计算效率。烧蚀分析证实，每个模块都对性能有显著贡献，其中SE模块的收益最大。所提出的模型在不产生额外计算成本的情况下始终优于基线方法。EdgeNeXt-SEDP通过高效的架构提供高诊断准确性，为减少观察者之间的差异和促进及时、个性化的宫颈腺癌管理提供了可扩展和可靠的解决方案。其紧凑的设计支持集成到不同的临床和资源有限的设置，推进人工智能在数字病理学中的应用。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Life sciences 医学-药学

CiteScore

12.20

自引率

1.60%

发文量

841

审稿时长

6 months

期刊介绍： Life Sciences is an international journal publishing articles that emphasize the molecular, cellular, and functional basis of therapy. The journal emphasizes the understanding of mechanism that is relevant to all aspects of human disease and translation to patients. All articles are rigorously reviewed. The Journal favors publication of full-length papers where modern scientific technologies are used to explain molecular, cellular and physiological mechanisms. Articles that merely report observations are rarely accepted. Recommendations from the Declaration of Helsinki or NIH guidelines for care and use of laboratory animals must be adhered to. Articles should be written at a level accessible to readers who are non-specialists in the topic of the article themselves, but who are interested in the research. The Journal welcomes reviews on topics of wide interest to investigators in the life sciences. We particularly encourage submission of brief, focused reviews containing high-quality artwork and require the use of mechanistic summary diagrams.