Artificial Intelligence and Nasal Cytology: Toward a Comprehensive Cytological Endotyping of Rhinitis

To the Editor: We read with great interest the recent article by Zhang and colleagues describing the development of an artificial intelligence system for the quantitative assessment of nasal inflammatory cytology based on whole-slide imaging. The authors should be congratulated for this important contribution, which represents a significant technological advancement in rhinology and digital pathology [1]. Nasal cytology is a simple, minimally invasive, and repeatable diagnostic technique that allows direct visualization of inflammatory cells within the nasal mucosa. Over the past decades, it has progressively emerged as a valuable tool for identifying inflammatory phenotypes in patients with rhinitis and other airway diseases [2]. Despite its clinical relevance, nasal cytology is still underutilized in routine clinical practice, partly because of concerns regarding operator dependence and the lack of standardized quantitative approaches. In this context, the development of automated systems capable of identifying inflammatory cells on digital slides represents a promising step toward the wider adoption of cytological diagnostics. By applying deep learning algorithms to whole slide images, Zhang et al. demonstrate the feasibility of automated detection and quantification of eosinophils and neutrophils, providing a reproducible and scalable framework for cytology-based diagnostics. Importantly, this work reflects the growing recognition of nasal cytology as a key method for the characterization of airway inflammation. Recent efforts toward methodological standardization, including the recent expert-based Delphi consensus on nasal cytology, further highlight the increasing interest of the scientific community in this diagnostic approach and the need for reproducible analytical strategies [3]. While the present study focuses on eosinophils and neutrophils, the next natural step in the evolution of AI-assisted cytological analysis will likely involve the recognition of additional cellular components and morphological patterns that are crucial for a comprehensive interpretation of nasal cytology. In particular, mast cells represent an essential component of type 2 inflammation and play a central role in several rhinologic conditions, including mixed inflammatory patterns such as nonallergic rhinitis with eosinophilia syndrome (NARES) and nonallergic rhinitis with eosinophils and mast cells (NARESMA). Recent evidence has highlighted the importance of the eosinophil–mast cell interaction in chronic inflammatory diseases of the upper airways and in chronic rhinosinusitis with nasal polyps [4]. The identification of mast cells, together with the recognition of mixed cytotypes, would therefore significantly enhance the diagnostic and prognostic value of automated cytological analysis. Beyond inflammatory cell counts, nasal cytology also provides important morphological information. Cytological patterns such as ciliocytophthoria associated with viral infections, as well as the presence of bacterial biofilm and other chromatic and structural alterations of epithelial cells, contribute to a more complete understanding of mucosal pathology. Integrating these morphological features into artificial intelligence-based analysis could further expand the diagnostic potential of digital cytology. Ultimately, combining cytological findings with clinical parameters may allow the development of integrated models capable of stratifying disease severity and predicting clinical outcomes. In this perspective, the integration of cytological patterns with clinical comorbidities has been proposed in the clinical-cytological grading system, which has demonstrated prognostic relevance in chronic inflammatory diseases of the upper airways [4, 5]. The work by Zhang and colleagues therefore represents an important milestone. By introducing artificial intelligence into nasal cytology, it opens the door to a new era in which traditional morphological diagnostics and advanced digital technologies may converge to improve the understanding, classification, and management of airway inflammatory diseases. We congratulate the authors for their innovative work and believe that continued collaboration between cytologists, clinicians, and data scientists will further advance the field toward a comprehensive cytological endotyping of airway diseases. Sincerely The author has nothing to report. The author declares no conflicts of interest.