To guarantee a comprehensive and systematic retrieval of relevant literature, a structured search was conducted in strict adherence to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 Statement. Four core electronic databases—PubMed, Embase, the Web of Science Core Collection, and the Cochrane Library—were queried for literature published between January 2000 and June 2025. The search strategy integrated both Medical Subject Headings (MeSH) terms and free-text keywords pertaining to the core research themes: “nasal inflammatory diseases”, “chronic rhinosinusitis”, “allergic rhinitis”, “predictive markers”, “biomarker”, “artificial intelligence”, and “precision medicine”. Furthermore, the reference lists of all included primary studies and relevant review articles were subjected to manual screening to retrieve additional eligible publications.

Studies were selected on the basis of the following predefined criteria:

The inclusion criteria were as follows (1): Original research articles or systematic reviews published in peer-reviewed journals focused on biomarkers or AI applications in nasal inflammatory diseases (e.g., CRS, AR) (2); studies providing explicit disease diagnostic criteria (e.g., the European Position Paper on Rhinosinusitis and Nasal Polyps 2024 [EPOS 2024] for CRS) and quantitative data on biomarker performance (e.g., sensitivity, specificity, area under the curve (AUC)) or AI model efficacy (e.g., accuracy, predictive value) (3); human-subject studies, with a sample size of ≥30 for original research or ≥10 included primary studies for systematic reviews (4); studies aligned with the core research focus of predictive markers and precision medicine in nasal inflammatory diseases; and (5) publications in the English language.

The exclusion criteria were as follows (1): conference abstracts, editorials, case reports (sample<10), and narrative reviews (2); studies conducted solely on animal models or cell lines (no human data) (3); articles with unavailable full texts or unextractable key data (e.g., biomarker AUC values, AI model validation metrics) (4); identical or duplicate publications or studies irrelevant to the core research question; and (5) non-English publications.

The study selection process strictly followed the PRISMA 2020 guidelines. All the retrieved records were imported into EndNote (reference management software) for deduplication. The selection was performed in two stages: First, two authors (Jiang-Xue Liao and Xin Lin) independently screened the titles and abstracts of all the records to meet the inclusion/exclusion criteria. Second, these two authors independently retrieved and evaluated the full texts of potentially eligible studies. Any discrepancies at either stage were resolved through discussion or, if necessary, by consultation with a third author (Jing He).

To ensure consistency, a standardized data extraction form was developed. For each included study, the following information was extracted: first author, publication year, country, study design, sample size, participant characteristics (e.g., CRS subtype, AR, FRS), key methods (e.g., biomarker detection techniques, AI model algorithms), main findings (e.g., the prediction performance and predictive value of biomarkers and AI models), and conclusions. The data were extracted independently by two authors to ensure accuracy. Given the anticipated methodological and clinical heterogeneity among the included studies (e.g., diverse CRS endotypic subtypes, varied AI model architectures), a narrative synthesis approach was adopted. The findings were thematically organized (e.g., “inflammatory factors”, “inflammatory cell markers”, “metabolites”) and presented in subsequent sections, incorporating cross-study comparisons and contrasts to highlight consistencies, discrepancies, and underlying factors driving divergent results.

CRSwNP, a refractory subtype of CRS, is closely related to type 2 inflammation and has a poor prognosis, including poor disease control and recurrence. Analysis of the biomarker profiles of blood and nasal secretion samples from CRSwNP patients before and after surgery revealed that the expression profile of type 2 inflammatory cytokines may be related to CRSwNP-related postoperative recurrence and can predict the efficacy of optimal or targeted drugs and surgical treatment (17). Some studies have suggested that immunoglobulin E (IgE) expression in tissues is correlated with disease severity, clinical and pathological features, and the speed of postoperative recurrence of CRSwNP and that the peripheral blood specific immunoglobulin E (sIgE) expression level (AUC = 0.786) can be used as a reliable indicator of postoperative recurrence of CRSwNP (18) (Table 1). Numerous studies have shown that increased IL-5 levels in nasal secretions or polyp tissues are significantly positively correlated with the risk of uncontrolled disease conditions (19) and recurrence (20) of CRSwNP; furthermore, the IL-5 level is a strong predictor for the early diagnosis of eosinophilic chronic rhinosinusitis with nasal polyps (ECRSwNP) (21). Bai et al. (22) reported that the levels of IL-5 and eosinophilic cationic protein (ECP), the preoperative Lund–Mackay score, and the combination of asthma and anti-double-stranded DNA (anti-dsDNA) IgG have a good ability to predict polyp recurrence after endoscopic surgery (AUC = 0.89) (Table 1). Several studies have shown that IL-33 and soluble ST2 (sST2) levels in the nasal polyp tissues and serum of CRSwNP patients are correlated with mucosal eosinophil (EOS) infiltration and postoperative recurrence, and these levels might be objective biomarkers for the differentiation of the CRSwNP endotype and the prediction of recurrence (19, 23). Jiang et al. (24) reported that serum CD39 and IL-33 levels, whose cutoff values were 125.9 pg/ml and 242.3 pg/ml, respectively, had a strong diagnostic ability for AR patients, and the receiver operating characteristic (ROC) curves suggested that CD39 can also be used to distinguish AR cases on the basis of severity (Table 1). Thymic stromal lymphopoietin (TSLP), a master regulator of type 2 immune responses and atopic diseases, can be used as a predictive factor for atopic diseases such as AR (25). One study indicated that the level of TSLP in the nasal mucosal tissue of CRSwNP patients was associated with a greater improvement in the SNOT-22 score after functional endoscopic sinus surgery, suggesting that the TSLP level can serve as a predictor of postoperative outcomes in CRSwNP patients (26). Another recent study indicated that a baseline plasma TSLP concentration >330 fg/mL can serve as a predictive biomarker for the efficacy of targeted therapy in CRSwNP patients (27) (Figure 1) (Table 1).

Serum periostin levels can be used not only in the differentiation of CRS endotypes but also in the evaluation of the severity, comorbidities, prognosis and treatment response, especially in the prediction of recurrence after surgery (28–31) (Table 1). Laury et al. (32) reported that the increase in periostin in the tissues of allergic fungal rhinosinusitis (AFRS) patients was related to strong eosinophilic inflammation. Some studies have shown that the concentration of cystatin SN (CST1) in exosomes and nasal secretions is correlated with the severity of CRSwNP and could be used as an indicator of poor outcomes in patients with CRSwNP (33). Mueller et al. (34) reported that mucus and serum levels of periostin, mucus, CST1 and IgE might be the best combinations of inflammatory factors for tracking nasal polyps and predicting recurrence. Charcot–Leyden Crystal (CLC)/Gal-10 is considered a marker for the involvement of EOS in allergic reactions and type 2-related inflammation. Studies have shown that the CLC level in nasal secretions is positively correlated with the EOS percentage in nasal polyp tissues and peripheral blood and is negatively correlated with the olfactory threshold (35), and a crystalline CLC structure higher than 1 per high-power field in tissue or a CLC level in nasal secretions higher than 34.243 ng/mL can predict postoperative polyp recurrence of CRSwNP, which is associated with refractory CRSwNP (36, 37). ECP and eotaxin, as markers of EOS, have been suggested as predictors of postoperative recurrence in patients with CRSwNP (38), and eotaxin-3 expression may also be an important marker for ECRSwNP typing and the prediction of disease severity in ECRSwNP patients (Figure 1) (Table 1).

Gao et al. (18) reported that in patients with ECRSwNP, higher IL-6 levels (AUC = 0.707) could increase the risk of postoperative recurrence and have good value for predicting postoperative recurrence (Table 1). Findings from some studies have suggested that transforming growth factor-β1 (TGF-β1), matrix metalloproteinase 9 (MMP-9), tissue inhibitor of metalloproteinase 1 (TIMP-1), IL-1, IL-4, IL-6, IL-17 and tumor necrosis factor-α (TNF-α) are factors influencing the development of refractory CRS (39). IL-36α may be a biomarker for predicting refractory CRSwNP in Asians (40). Xie et al. (41) reported that chemerin and IL-17 are potential predictors of CRS and that chemerin silencing can reduce inflammation and bone remodeling in CRS patients. Hussien et al. (42) reported that an IL-17 concentration greater than 9.2 pg/ml and a pentraxin-3 concentration greater than 4.12 ng/ml were considered important biological indicators with high sensitivity and specificity for the diagnosis and follow-up of CRSwNP patients (Table 1). Tang et al. (43) reported that IL-22 and IL-17A may play important roles in regulating T helper type 2 (Th2) inflammatory skewing in AR patients (Figure 1).

EOSs are key effector cells that affect the prognosis of type 2 CRS, and EOS infiltration is a characteristic feature observed in most CRSwNP patients. At least 1 year after surgery, a tissue EOS count greater than 48/HPF, a blood EOS percentage greater than 2.5% or a tissue EOS percentage greater than 20.6% were independent risk factors for uncontrolled CRS (6) (Table 1). Lou et al. (44) reported that a tissue EOS count greater than 55/HPF or an EOS ratio greater than 27% could predict the recurrence of nasal polyps (Table 1). Brescia et al. (45) reported that increased serum EOS and basophil counts were associated with an increased risk of recurrence of nasal polyps, and the likelihood of developing nasal polyps in patients with a serum EOS percentage ≥ 3.7% was more than twice that of patients with a serum EOS percentage < 3.7% (Table 1). Other studies have demonstrated that the presence of EOSs in tissue or blood is significantly correlated with the development or recurrence of CRSwNP (20) and may be significantly associated with olfactory dysfunction (46). A study on the role of blood EOS and basophil counts in predicting AFRS revealed that the EOS and basophil counts in patients with recurrent AFRS were significantly greater than those in patients without recurrence of AFRS (47). However, there is currently no consensus on the tissue or serum EOS count thresholds and their effects on CRS patient outcomes. Large-scale multicenter studies should be performed to establish a widely acceptable EOS cutoff to predict definite treatment outcomes (Figure 1).

The importance of neutrophils in the pathogenesis of CRS has received increasing attention. Some studies have shown that a high level of subepithelial neutrophil infiltration may be used as a cellular biomarker of refractory CRSwNP (48). Kim et al. (40) revealed that in South Korean patients with CRSwNP, at 1 year after surgery, a subepithelial human neutrophil elastase (HNE)-positive neutrophil number greater than 45/HPF could be used as a cell marker of refractory CRS (Table 1). Cha et al. (49) reported that tissue neutrophilia and neutrophil extracellular trap formation (NETosis) can serve as prognostic biomarkers for neutrophilic CRSwNP. Subash et al. (50) conducted a prospective study and reported that the neutrophil–lymphocyte ratio (NLR) could be used as a marker for the severity and prognosis of nasal polyps and AFRS. In type 3 CRS, higher levels of neutrophil markers are correlated with disease severity and the CRSwNP phenotype (51) (Figure 1).

Polyunsaturated fatty acids (PUFAs) can be used as monitoring markers or novel treatments for AR. PUFAs include n-3 PUFAs (mainly eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA), and linolenic acid (LNA)) and n-6 PUFAs (mainly AA and LA). Docosapentaenoic acid (DPA) is a product of EPA. If pregnant women have high DPA levels in red blood cells, both the women and their offspring are more likely to be protected against the AR response caused by particulate matter ≤2.5 µm in size (PM2.5) (52). Zheng et al. (53) conducted a prospective study and reported that the levels of AA and its downstream metabolites could be used as markers of the efficacy of subcutaneous immunotherapy (SCIT) for AR and that the levels of these metabolites were correlated with improvements in the Rhinoconjunctivitis Quality of Life Questionnaire (RQLQ) score. Magnusson et al. (54) reported that AA levels and the presence of AR were related to the likelihood of achieving remission in asthma patients. Arachidonate 15-lipoxygenase (ALOX15) is a key metabolic enzyme that converts AA into biologically active metabolites, and some studies have shown that ALOX15 expression is upregulated in CRSwNP, where ALOX15 is involved in the recruitment and activation of EOSs; therefore, ALOX15 can be used as a predictive biomarker of CRS in combination with nasal polyps. Moreover, the ALOX15 mRNA level alone (AUC = 0.909) or in combination with the blood EOS count (AUC = 0.933) could be a reliable biomarker for predicting the diagnosis of ECRSwNP (55) (Table 1). The prostaglandin (PG) E2 produced in the metabolism of AA is a key component of IL-13-induced epithelial remodeling in nasal polyps, is correlated with the clinical severity of CRS, and can be used as a predictor of poor CRS prognosis (56). Findings from some studies have suggested that α-linoleic acid (ALA) may be a metabolic marker that is predictive of AR complicated with asthma or enteritis (57) (Figure 2).

Glucose metabolism has been shown to promote glucose uptake and mainly glycolysis by epithelial cells, which strengthens the proinflammatory function of epithelial cells in CRS (57). A P-glycoprotein (P-gp) secretion level greater than 250 pcg/μg is associated with a high incidence of CRSwNP and can be used as an indicator of disease severity (58) (Table 1). Serum iron levels and the overexpression of glucose-regulated proteins can increase the sensitivity of endothelial cells to fungi and induce fungal invasion and injury in patients with diabetic ketoacidosis (59); moreover, the hyperglycemic and acidic status of these cells can lead to the rapid progression of invasive fungal rhinosinusitis (IFRS) (60). Barbara et al. (61) studied the prognostic factors of acute invasive fungal rhinosinusitis (AIFRS) associated with coronavirus disease 2019 (COVID-19) and reported that the level of glycated hemoglobin (HbA1c) can be used as a marker of a poor prognosis (Figure 2) (Table 1).

Studies have revealed that the threshold ratio of total ethmoid sinus (E) and total maxillary (M) scores (E/M ratio) for both sides of the Lund–Mackay computed tomography (CT) score has the highest predictive value for the diagnosis of eosinophilic chronic rhinosinusitis (ECRS) and the recurrence of CRSwNP (65); when the threshold ratio of total ethmoid sinus (E) and total maxillary (M) scores (E/M) is greater than 2.55, the likelihood of CRSwNP recurrence is increased (66) (Table 1). In another study on the development of a radiomic identification model for ECRSwNP, a CT radiomic model based on 10 radiomic features was found to be promising for the identification of ECRSwNP, and this model may provide new insights to address other clinical issues, such as guiding personalized treatment and predicting CRSwNP patient outcomes (67). In a magnetic resonance imaging (MRI) study on vision and mortality outcomes associated with AIFRS, when orbital apex or cerebral artery involvement or both were present, the findings of MRI performed before the initial intervention could predict poor visual outcomes; facial soft tissue, the nasolacrimal duct, intracranial involvement or a combination of these three were associated with an increased risk of death, and hard palate involvement was associated with a poor prognosis (68). A study in AIFRS revealed that if the extranasal lesion with loss of contrast enhancement (LoCE) found on Gd-enhanced MR images could not be completely removed by surgery, the prognosis might be poor (69). Among the various other clinical and radiological factors, the LoCE was an independent prognostic factor (70). Clinical imaging is a noninvasive and efficient means to predict the development of nasal inflammatory diseases, but the accuracy of imaging-based predictions still needs to be confirmed through additional studies.

Studies have revealed that CRSwNP patients with asthma, high Lund–Kennedy scores, high visual analog scale (VAS) scores and high olfactory loss scores are more likely to recur (66). In a clinical study of indicators of IFRS outcomes in children, survival was determined by the absolute neutrophil count, recovery time, brain involvement, fungal types, patient condition in the intensive care unit (ICU), and hyperglycemia (71). A study on the survival of patients with AIFRS revealed that the shorter the duration of prediagnostic symptoms was, the worse the overall survival of patients with AIFRS (72). Owing to the variety of clinical symptoms of nasal inflammatory diseases and large individual differences, the use of a clinical score to predict the poor prognosis of nasal inflammatory diseases is controversial. In the future, a joint prediction model of clinical scores together with other predictors may be established to prevent unfavorable outcomes of nasal inflammatory diseases.

Some studies have shown that a naive Bayes model based on the combination of environmental PM exposure and the EOS count can predict the risk of CRSwSP recurrence (73). In reference to existing international diabetes management programs, Bousquet et al. (74) proposed a model in which the MASK-air app was used to manage AR patient symptoms according to an electronic daily Combined Symptom and Medication Score (CSMS) or an electronic Daily Control Score for Asthma (e-DASTHMA). The AI Chronic Rhinosinusitis Evaluation Platform 2.0 (AICEP 2.0) was established to determine the proportion of inflammatory cells in the clinical diagnosis of nasal polyp cells and to explore the clinical significance of different nasal polyp phenotypes (75). In addition, a multitask deep learning network based on a deep learning radiology nomogram can be used for sinus segmentation and noninvasive prediction of CRS recurrence (65).

Although the aforementioned AI applications for predicting poor outcomes of nasal inflammatory diseases are promising, their inherent limitations and risks in clinical translation warrant serious attention, with drawbacks being highly intertwined with the core research challenges in this field. At the data foundation level, several core AI models (65, 73) are trained on single-center homogeneous cohorts or are restricted to specific detection equipment, resulting in limited generalizability of the models. In terms of functional limitations, the phenotypic diagnosis platform (75) lacks longitudinal validation capabilities, as it does not meet the research needs for assessing dynamic changes in biomarkers; the symptom management tool (74) relies on patient-reported data with quality biases, which may trigger cascading risks if directly used as input for prognostic models.

Although medical and surgical treatments for nasal inflammatory diseases are currently available, the treatment effects remain unsatisfactory; therefore, an increasing number of studies have been conducted on targeted biological therapy for nasal inflammatory diseases. Targeting type 2 inflammatory cytokines, such as IL-4, IL-5, IL-13, and IgE, is currently considered a promising treatment approach. Through indirect comparisons of placebo-controlled trials, mepolizumab (anti-IL-5), reslizumab (anti-IL-5), benralizumab (anti-IL-5Rα), omalizumab (anti-IgE), and dupilumab (anti-IL-4Rα, inhibition of IL-4 and IL-13 signal transduction) have been shown to reduce nasal airway obstruction in patients with CRSwNP and to reduce the likelihood of requiring emergency medical treatment and/or surgical removal of polyps (76–79) (Table 2).

Numerous studies have revealed that mepolizumab and reslizumab, which bind to IL-5 with high affinity and specificity, can reduce the EOS count, nasal and peripheral IL-5 levels, soluble IL-5Rα levels, and ECP levels in CRSwNP patients (80–82). The anti-IL-5 monoclonal antibody mepolizumab has been shown to reduce the need for surgical intervention in patients with severe, recurrent bilateral nasal polyps; improve nasal polyp size and nasal congestion severity scores; and decrease systemic levels of type 2 inflammatory markers, including IgE, periostin, MMP-9, myeloperoxidase, PGD2, PGF2α, LTB4, and thrombin (77, 80, 83–85). A phase III randomized study of benralizumab (anti-IL-5Rα antibody) revealed that, compared with placebo, benralizumab could reduce the nasal polyp score, nasal obstruction score and olfactory score in patients with CRSwNP; furthermore, benralizumab was beneficial for the control of asthma symptoms (77, 82, 86, 87). Accordingly, serum IL-5 levels serve as a key predictive biomarker for anti-IL-5/IL-5Rα therapy with mepolizumab, reslizumab, and benralizumab. Numerous studies have shown that omalizumab can alleviate clinical symptoms; reduce the SNOT-22 score and nasal polyp score; and reduce serum periostin, ECP and soluble IL-5Rα levels in CRSwNP patients (77, 78, 80, 88, 89). In AR patients, omalizumab reduced the daily nasal and eye symptom scores, the medication symptom score, and the proportion of days of emergency drug use and improve the safety of immunotherapy for AR (90, 91). The application of this therapy is specifically guided by elevated serum IgE levels, establishing IgE as its primary predictive biomarker (82, 92). Multiple studies have shown that a dupilumab +/-mometasone nasal spray can improve outcomes, alleviate symptoms and reduce nasal polyp scores in patients with severe CRSwNP and refractory ECRS (77–79, 93–95); furthermore, this combined treatment can decrease the levels of type 2 biomarkers (such as eotaxin-2, ECP, IL-5, IL-13, and IgE) and significantly alleviate AR-related symptoms in patients with comorbid asthma (82, 96–100). Therefore, the efficacy of dupilumab is correlated with a broader type 2 inflammatory signature, and biomarkers such as IL-4, IL-1, eosinophils, and IgE can inform its predictive value. Tezepelumab can increase the efficacy of SCIT in patients with AR (101). Shen et al. (102) reported that CM310, a monoclonal antibody targeting IL-4Rα, was safe and effective at reducing the size of nasal polyps, improving symptom scores, and improving the QoL of patients with severe ECRSwNP. The IL-4Rα chain (also known as CD124) is the major component of the IL-4 receptor (103). Injections of an anti-CD124 monoclonal antibody (αCD124) have long been used to treat CRSwNP (101). Wu et al. (104) reported that the use of protamine nanostructures to encapsulate αCD124 led to increased efficacy of the localized intranasal administration of αCD124, which revealed that the local administration of drugs via nanomaterials may become a future research direction.

In view of the influence of the microbiota on nasal inflammatory diseases, changing the characteristics of the nasal microbial environment may prevent negative clinical outcomes of nasal inflammatory diseases. According to the 2020 European document on rhinosinusitis and nasal polyps, antibiotics (usually macrolides) are often used to treat CRS to change the characteristics of the nasal microbial environment (105). Some studies have shown that probiotics can relieve the symptoms of CRS and prevent poor prognoses (106), and the topical use of probiotics prevented infection by Corynebacterium tuberculostearicum in a CRS mouse model (107). Since probiotics are less harmful and effective at alleviating AR symptoms and improving QoL (107), in the International Consensus Statement on Allergy and Rhinology: Allergic Rhinitis-2023, it is recommended that probiotics be used as adjuvant therapy for symptomatic seasonal or perennial AR (108). For fungal infections in patients with IFRS, the current antifungal treatments are oral amphotericin B lipid preparations and voriconazole (109). However, the use of antibiotics is controversial because of the wide range of microbial populations involved in nasal inflammatory diseases. Owing to the individualized efficacy of probiotic treatment, large-scale studies are needed to confirm the efficacy of changing the nasal microbial environment as a treatment.

Other treatments for nasal inflammatory diseases include hormones, antihistamines, and immunotherapy, and hormones and antihistamines are often used to control symptoms. Hormones are beneficial for the control of olfactory dysfunction in CRS patients (110). Some scholars believe that steroid hormones can control the allergic response of AR patients by inhibiting Th2 cytokines (IL-4, IL-5, IL-13, etc.) (111). H1 antihistamines (such as chlorpheniramine, loratadine, and fexofenadine) and intranasal antihistamines (such as azelastine and olopatadine) can reduce the expression of proinflammatory cytokines and cell adhesion molecules; therefore, they are often used to treat AR (112). Treatment with loratadine combined with glucocorticoids has been confirmed to reduce the incidence of adverse reactions and nasal symptom scores in children with AR (113). Allergen immunotherapy (AIT) is a treatment for AR and has a long history of clinical application. The AITs used for AR mainly include SCIT, sublingual immunotherapy (SLIT), intralymphatic immunotherapy (ILIT), and local immunotherapy. SCIT can reduce the symptom score, drug score, disease-specific QoL score and serum sIgG4 level of AR patients (114). SLIT is important for preventing asthma development or reducing the risk of asthma exacerbation in AR patients (115). The use of ILIT helps relieve rhinitis symptoms, reduce drug dosage, and reduce the VAS score and RQLQ score (116).

As a metabolite of EPA, 15-hydroxyeicosapentaenoic acid (15-HEPE) has also been shown to be an EOS-dependent antiallergic metabolite that can be applied in the prevention and treatment of AR (117). ALA was shown to significantly relieve nasal symptoms, reduce serum OVA-sIgE levels, and correct the Th1/Th2 imbalance in AR mice (118). Short-chain fatty acid (SCFA) supplementation or SCFA-producing bacteria administration helps relieve AR symptoms (119). Alfalfa (Medicago sativa) was used in AR treatment because of its anti-inflammatory effects and ability to inhibit key enzymes involved in AA metabolism.

TCMs such as crocin can reduce the levels of Th2 cytokines such as IL-4 and inhibit the NF-κB signaling pathway, thereby inhibiting inflammation in ECRS (120). Xiao-qing-long-tang (a TCM) was shown to relieve nasal symptoms and reduce the levels of type 2 cytokines and OVA-sIgE in AR mice, thereby reducing type 2 inflammation in AR (121).

The role of dietary factors in managing chronic diseases like type 2 diabetes has been widely explored. Just as specific interventions are used for glycemic control in diabetic patients, predictive biomarkers in inflammatory diseases can help stratify patient risk and optimize therapeutic approaches. Biomarkers, much like dietary interventions, serve as a tool for personalized treatment, improving patient outcomes through targeted approaches (122).