Nasal Swab Cytokine Predicts Smell Recovery After Sinus Surgery

For patients with severe chronic rhinosinusitis with nasal polyps, losing the sense of smell is often the most devastating part of the disease. Surgery can clear blocked sinuses and stop the relentless discharge, yet whether smell returns afterward has long been a clinical gamble. Now, a research team at Sun Yat-sen University in Guangzhou, China, reports a way to tip those odds in patients’ favor before the first incision is made: a simple, non-invasive measurement of a single cytokine in nasal secretions, folded into a practical prognostic tool that estimates each patient’s probability of olfactory recovery six months after surgery.

The study, published in Frontiers in Allergy, tackled a stubborn translational problem. The inflammatory state of the deep nasal mucosa is known to shape whether the delicate olfactory neuroepithelium can regenerate, but reaching that state has traditionally required invasive tissue biopsies, which are impractical for routine pre-operative screening or repeated outpatient monitoring. Nasal secretions, which directly bathe the olfactory cleft, offer an ideal biological window, yet the precise mapping between deep-tissue inflammation and superficial secretory chemistry has remained largely unknown.

To bridge that gap, the team led by Hangtian Zhang, Linlu Wang, Tong Lu, and Yilin Hou, under the senior authorship of Jian Li, Weiping Wen, and Hua Zhong, designed a tissue-to-secretion dual-cohort strategy. In a discovery cohort of 67 patients, a Random Forest machine learning algorithm screened deep-tissue proteomes quantified with a Luminex multiplex platform, ranking candidates by their contribution to predictive accuracy. An independent Random Forest model then evaluated non-invasive nasal secretions, profiled with high-sensitivity Olink proximity extension assay technology, in a validation cohort of 62 patients. The two cohorts were well matched in age, sex, CT-based disease severity, blood inflammatory counts, and comorbid asthma and allergic rhinitis, ensuring that the cross-compartment molecular mapping was not confounded by clinical disparities.

The machine learning screen revealed a compelling pathophysiological landscape. Rather than a homogenous Type 2 inflammatory burden, the canonical signature long blamed for polyp growth and olfactory damage, the top 20 tissue candidates included potent Type 1 and Type 3 mediators such as tumor necrosis factor-alpha and interleukin-17, alongside the macrophage chemoattractant MCP-1 and classical Type 2 cytokines like IL-13 and IL-5. This mixed-endotype picture, in which neutrophilic inflammation coexists with Type 2 signals, is increasingly recognized as a driver of severe polypoid remodeling, particularly in Asian cohorts.

The critical test was whether these deep-tissue signals could survive translation to the secretory compartment. Eight top-tier tissue candidates, defined by a Mean Decrease Gini index above 1.0, were cross-referenced with the Olink Target 96 Inflammation panel. Five, MCP-1, TNF-alpha, IL-13, IL-5, and IL-17A, were successfully matched and quantified in nasal secretions. The results were striking: the canonical Type 2 cytokines IL-5 and IL-13, and even the dominant tissue chemoattractant MCP-1, lost all predictive significance in secretions. Only the non-Type 2 axes retained clinical relevance, with secretory IL-17A demonstrating the strongest continuous dose-response relationship with absolute olfactory improvement, a Spearman correlation of 0.334 with a P value of 0.0003, roughly two orders of magnitude stronger than the next best candidate.

That discrepancy between tissue and secretion is itself a scientific finding. In bulk tissue homogenates, IL-17’s standalone linear predictive capacity was notably weak, likely because its independent prognostic signature is masked by the overwhelming Type 2 background and complex stromal interactions within the polypoid bed. The secretory compartment, by contrast, appears to act as a biological filter. The team postulates that high concentrations of IL-17A in nasal secretions serve as a direct, functional readout of epithelial barrier collapse and tight junction disruption, isolating terminal pathogenic signals of neuroepithelial leakage from the convoluted deep-tissue inflammatory background.

Multivariate logistic regression confirmed that secretory IL-17A was an independent protective factor for olfactory recovery, with an odds ratio of 1.408 per unit increase, after adjusting for baseline disease severity. Higher pre-operative visual analog scale scores, by contrast, independently decreased the likelihood of achieving a clinically meaningful recovery threshold, defined as an absolute improvement of at least 3 points or a final 6-month score of 3 or less. The counterintuitive protective role of IL-17A, a cytokine conventionally recognized for disrupting epithelial tight junctions and driving neutrophilic infiltration, suggests a paradigm shift. Within an extreme, recalcitrant CRSwNP cohort, the presence of an IL-17A-associated mixed inflammatory endotype may reflect a higher degree of neuroepithelial plasticity and post-operative reversibility than a homogenous, end-stage Type 2 milieu, which is thought to lead to terminal, irreversible neural degradation.

Translating these insights into point-of-care practice, the team constructed a clinical-molecular nomogram integrating baseline VAS and secretory IL-17A. The joint model achieved an optimism-corrected area under the receiver operating characteristic curve of 0.793 after 1,000 bootstrap resamples, with adequate calibration and, as demonstrated by decision curve analysis, substantially superior net clinical benefit across a broad spectrum of threshold probabilities compared to treat-all or treat-none default strategies or a model relying solely on baseline symptom severity. Classical severity markers, including peripheral blood eosinophil percentage, Lund-Mackay CT scores, and olfactory cleft scores, did not exhibit independent predictive significance in the current cohort, underscoring the parsimony of the final model.

The authors acknowledge limitations, including the absence of an independent external validation cohort and the need for objective psychophysical testing such as Sniffin’ Sticks in future multicenter studies. Yet the framework offers rhinologists a highly actionable tool for personalized pre-operative counseling, empowering clinicians to preemptively identify refractory patients who might benefit from intensified perioperative medical adjuncts, such as targeted biologics or systemic corticosteroids, rather than relying solely on surgical intervention. By filtering out the profound Type 2 background noise predominant in severe CRSwNP, the study identifies secretory IL-17A as a specific, non-invasive indicator of mucosal plasticity and olfactory reversibility, marking a significant step toward precision medicine in one of otolaryngology’s most recalcitrant symptoms.

Subject of Research: Non-invasive prediction of olfactory recovery after sinus surgery using secretory IL-17A in chronic rhinosinusitis with nasal polyps

Article Title: Non-invasive prediction of olfactory recovery in severe CRSwNP: a cross-compartment translational nomogram based on secretory IL-17A

Article References: Zhang, H., Wang, L., Lu, T., Hou, Y., Wei, Y., Huang, W., Liu, S., Huang, Y., Xu, Z., Li, J., Wen, W., & Zhong, H. (2026). Non-invasive prediction of olfactory recovery in severe CRSwNP: a cross-compartment translational nomogram based on secretory IL-17A. Frontiers in Allergy, 7, Article 1978136. https://doi.org/10.3389/falgy.2026.1978136

Image Credits: AI Generated

DOI: 10.3389/falgy.2026.1978136

Keywords: olfactory recovery, CRSwNP, IL-17A, nasal secretions, biomarker, machine learning, nomogram, Type 2 inflammation, sinus surgery, liquid biopsy, Non-invasive, prediction