ORIGINAL ARTICLE
Predictive value of computed tomography in the recurrence of chronic
rhinosinusitis with nasal polyps
Yifan Meng, MD1 , Luo Zhang, MD, PhD1,2,3
Background: Chronic rhinosinusitis with nasal polyps (CR-
SwNP) is a nasal disease with a high tendency for recur-
rence. The aim of this study was to compare the use of com-
puted tomography (CT) scan with other clinical parameters
in predicting the recurrence of CRSwNP.
Methods: A total of 272 consecutive CRSwNP patients
undergoing endoscopic functional sinus surgery were re-
cruited. The demographic characteristics and clinical pa-
rameters, including CT scores, level of exhaled nitric oxide,
and peripheral eosinophilia, were recorded. The degree of
infiltration of inflammatory cells in the sinus mucosa was
evaluated.
Results: Two hundred thirty of the 272 patients completed
the study (118 patients with recurrence and 112 patients with
no recurrence). The average follow-up time was 24 months
aer the first surgery. The 2 groups were not significantly
different with respect to age, gender distribution, comorbid
allergy, exhaled oral fractional exhaled nitric oxide levels,
nasal obstruction/runny nose/headache/facial pain scores,
Lund-Mackay score, peripheral eosinophil percentage, and
peripheral eosinophil absolute count. The onset of surgi-
1 Department of Otolaryngology Head and Neck Surgery, Beijing
TongRen Hospital, Capital Medical University, Beijing, China; 2 Beijing
Key Laboratory of Nasal Diseases, Beijing Institute of Otolaryngology,
Beijing, China; 3 Department of Allergy, Beijing TongRen Hospital,
Capital Medical University, Beijing, China
Correspondence to: Chengshuo Wang, MD, Department of Otolaryngology
Head and Neck Surgery, Beijing TongRen Hospital, Capital Medical
University, No. 1, Dongjiaominxiang, Dongcheng District, Beijing 100730,
China; e-mail: wangcs830@126.com
The first 2 authors (Y.M. and L.Z.) contributed equally to this work.
Funding sources for the study: National Key R&D Program of China
(2016YFC20160905200); Program for the Changjiang Scholars and
Innovative Research Team (IRT13082); the National Natural Science
Foundation of China (81420108009, 81630023, 81400444, and 81470678);
Beijing Municipal Administration of Hospitals’ Mission Plan (SML20150203);
Beijing Municipal Administration of Hospitals’ Youth Programme
(QML20150202); Beijing Advanced Innovation Center for Food Nutrition and
Human Health (Beijing Technology and Business University, 20181045);
Beijing Municipal Administration of Hospitals Clinical Medicine
Development of Special Funding Support (XMLX201816); and the Capital
Health Development Foundation (2016-1-2052).
Potential conflict of interest: None provided.
, Hongfei Lou, MD1 and Chengshuo Wang, MD1
cal history and asthma, visual analog scores of CRS, anos-
mia score, ratio of total ethmoid sinus scores for both sides
and maxillary sinus score for both sides (E/M ratio), Lund-
Kennedy score, tissue eosinophil percentage, and tissue
eosinophil absolute count were significantly higher in the
recurrence group. The E/M ratio showed high accuracy as
a predictor for CRSwNP recurrence. The cut-off point of
2.55 for E/M ratio indicated the highest predictive value of
CRSwNP recurrence.
Conclusion: The E/M ratio is a useful predictor for the re-
currence of CRSwNP in the Chinese population.  C 2019
ARS-AAOA, LLC.
Key Words:
chronic rhinosinusitis; computed tomography; diagnosis;
nasal polyps; recurrence
How to Cite this Article:
Meng Y, Zhang L, Lou H, Wang C. Predictive value of
computed tomography in the recurrence of chronic rhi-
nosinusitis with nasal polyps. Int Forum Allergy Rhinol.
2019;00:1-8.
C hronic rhinosinusitis (CRS) is defined as chronic
nasal/sinus mucosa inflammation lasting at least
12 weeks without complete resolution.1 CRS is a hetero-
geneous disease and is generally classified as CRS with
nasal polyps (CRSwNP) or as CRS without nasal polyps
(CRSsNP).2 CRS has a high prevalence and affects about
12% of adults in United States and 11% of adults in
Europe.3, 4 In China, the prevalence of CRS has been es-
timated to be 8%, based on face-to-face interviews.5 At the
socioeconomic level, CRS adversely impacts the quality of
life and work productivity of affected individuals, and is
often associated with lower airway disease.6
More recently, studies of CRS in China have focused
on determining different endotypes of CRS by employing
cluster analysis of inflammatory endotypes.7–9 A common
Received: 27 October 2018; Revised: 26 March 2019; Accepted:
28 April 2019
DOI: 10.1002/alr.22355
View this article online at wileyonlinelibrary.com.
International Forum of Allergy & Rhinology, Vol. 00, No. 0, xxxx 2019 1
Meng et al.
feature of these studies has been that, irrespective of
how the phenotypes or endotypes were defined, recurrence
of CRS, eosinophil-dominated type 2 inflammation, and
difficult-to-treat CRS were key factors for clustering.7–9
Moreover, Wei and colleagues9 showed that, in an 8-year
follow-up period after sinus surgery, the total recurrence
rate of CRSwNP was 21.8%.9 Collectively, the data from
these studies suggest that, over the last few years, CRS pa-
tients in several Asian countries had a T-helper 2 (Th2)-
biased “eosinophilic shift”6 from their usual Th1/Th17
signature.10
In recent years, several markers have also been used for
predicting nasal polyp recurrence. Lou and colleagues11
suggested that a tissue eosinophil proportion of >27% of
total cells or a tissue eosinophil absolute count of >55
eosinophils per high-power field may be reliable as mark-
ers for nasal polyp recurrence within 2 years after sinus
surgery, whereas Younis and colleagues12 suggested that
the presence of eosinophilic mucin was a strong predic-
tor of CRS recurrence.12 Similarly, other studies suggested
CRSwNP recurrence to be associated with occupational
dust exposure,13 serum periostin,14 and polypoid change
of anterior free border of middle turbinate.15 However,
use of these markers is limited because they are either in-
vasive or impossible to obtain before surgery. In recent
years, computed tomography (CT), the most commonly
used imaging modality for evaluation of CRS, has received
increased attention. Sakuma and colleagues16 and Zuo and
colleagues17 demonstrated that CT characteristics could be
used for diagnosing eosinophilic CRS (ECRS). Similarly,
Hong and colleagues18 suggested that the sinus CT scan
could be used in predicting glucocorticoid sensitivity in CR-
SwNP patients,18 whereas our own earlier studies suggested
that the CT scan was a more useful predictor for diagnosis
of eosinophilic CRSwNP compared with the other clinical
parameters.19 However, the value of CT scans for predict-
ing the recurrence of CRSwNP is unknown. The aim of
the present study was therefore to investigate the predic-
tive value of the CT scan for recurrence of CRSwNP in
Chinese subjects. In particular, the threshold ratio of total
ethmoid sinus (E) and total maxillary (M) scores for both
sides (E/M ratio), determined in a CT scan according to the
Lund-Mackay scoring system, was assessed for predicting
the recurrence of disease. Recurrence was defined as the
presence of nasal polyps with/without CRSwNP symptoms
after adequate surgery and management with normative
intranasal corticosteroid and up to 2 short courses of an-
tibiotics or systemic corticosteroids.
Patients and methods
Patients with a diagnosis of CRSwNP, confirmed accord-
ing to the criteria recommended by the European Posi-
tion Paper on Rhinosinusitis and Nasal Polyps,1 were en-
rolled prospectively into this study from June 2014 to May
2015 in the Department of Otolaryngology Head and Neck
Surgery, Beijng TongRen Hospital. None of the patients
2 International Forum of Allergy & Rhinology, Vol. 00, No. 0, xxxx 2019
had been treated with antibiotics or corticosteroids within
the 4-week period before surgery, and patients with cys-
tic fibrosis, fungal sinusitis, allergic fungal rhinosinusitis,
or primary ciliary dyskinesia were excluded. All patients
underwent functional endoscopic sinus surgery (FESS) and
the sinuses opened based on severity of the sinusitis. For
standardization, all procedures and postoperative follow-
up endoscopy examinations were performed by the same
surgeon. All patients enrolled provided Informed consent
documentation to ensure compliance.
At enrollment after FESS, the demographic characteris-
tics of all patients were recorded, and all patients adhered
the same postoperative follow-up schedule, which included
weekly visits during the first month, monthly visits during
the second to third months, and visits every 2-3 months
until repair to the nasal epithelium was complete. During
follow-up, all patients were asked to use intranasal budes-
onide, oral steroid if the patient showed a high degree of
eosinophil infiltration before surgery,11 and nasal saline ir-
rigation as ongoing treatment. Nasal steroid irrigation was
not used in any patient. To ensure that patients complied
with the intranasal steroid and rinses, reminders were sent
to all patients in the morning and evening as a daily short
message service via mobile phone, as we have previously
shown this to be effective.20 All patients were reexamined
from June 2017 to April 2018 for recurrence of CRSwNP,
based on the presence of nasal polyps with/without CR-
SwNP symptoms by nasal endoscopy after adequate surgery
and management with normative intranasal corticosteroid
and up to 2 short courses of antibiotics or systemic corti-
costeroids.
The study was approved by the medical ethics commit-
tee of Beijing TongRen Hospital and all patients provided
written informed consent before entry into the study and
collection of data.
Assessment of preoperative clinical characteristics
Before surgery, all patients were evaluated for CT
scores, the level of exhaled nitric oxide, and peripheral
eosinophilia. Each patient underwent CT scanning (Philips
Health Care, Best, The Netherlands), and the CT of the
affected sinuses was graded according to the Lund-Mackay
staging system.21 The maxillary sinus score (M score), an-
terior ethmoid (AE) sinus score, and posterior ethmoid (PE)
sinus score were determined and used to additionally calcu-
late total ethmoid (E) sinus score (total of AE and PE scores
for both sides), total maxillary (M) score for both sides, and
E/M ratio (ratio of E and M scores). The CT scans were
scored by 2 independent radiologists. Endoscopic scoring
was completed by an independent observer according to
the Lund-Kennedy scoring system.22
History of allergic disease was determined based on the
characteristic symptoms of disease, medications taken for
the disease, and the results of serum immunoglobulin E
(IgE) testing. Serum IgE levels for common aeroallergens,
including Dermatophagoides farinae, Dermatophagoides

pteronyssinus, mugwort, Penicillium notatum, Candida
albicans, Alternaria, Cladosporium, and Aspergillus, were
determined using a fluoroenzyme immunosorbent assay
(UniCAP, Uppsala, Sweden), with serum IgE ࣙ0.35 kU/L
value considered positive.
Exhaled oral fractional exhaled nitric oxide (FeNO) was
measured using a nitric oxide analyzer (NIOX; Aereocrine,
Solna, Sweden) at a flow rate of 50 mL/s through the oral
cavity, with the gas continuously routed into the analyzer
via a side port. The procedure was repeated 3 times and a
mean concentration of FeNO was calculated.
A complete blood cell count with the differential cell
count vs patient’s regular condition without infection was
performed within 1 week before surgery. The percentage
of eosinophils and absolute blood eosinophil counts were
calculated.
Histologic evaluation of nasal mucosa
Samples of nasal mucosa were obtained from each pa-
tient during FESS and then processed for histologic eval-
uation using standardized procedures. Paraffin-embedded
samples were sectioned at 4-μm thickness and stained with
hematoxylin-eosin (H&E). All samples were observed by
light microscopy at ×400 magnification for distribution
and numbers of eosinophils, neutrophils, plasma cells, and
lymphocytes. The samples were assessed by 2 indepen-
dent observers who were blinded to the study protocol
and recorded the cell counts as the mean of counts for 10
nonoverlapping fields in the lamina propria. The numbers
of cells were calculated as both the absolute number per
high-power field and the percentage of total inflammatory
cells present.
Statistical analysis
A 2-sample t test and χ 2 test were used for 2-group com-
parisons of age, gender ratio, history of surgery, onset of
asthma, and history of allergy. A 1-way analysis of variance
was followed by a Mann-Whitney test for 2-group com-
parisons of E/M ratio, Lund-Mackay score, Lund-Kennedy
score, visual analog scale (VAS) scores of CRS, FeNO, tis-
sue eosinophil percentage, tissue eosinophil absolute count,
blood eosinophil percentage, and blood eosinophil abso-
lute count. All statistical analyses were performed using
SPSS for Windows version 22.0 (SPSS, Inc, Chicago, IL)
and GraphPad Prism version 7.0 (GraphPad Software, Inc,
La Jolla, CA). p < 0.05 was considered significant.
To determine parameters associated with CRSwNP re-
currence, logistic regression analysis was conducted for sev-
eral predefined demographic parameters (age, gender ratio,
history of surgery, the onset of asthma, history of allergy)
and clinical parameters (E/M ratio, Lund-Mackay score,
Lund-Kennedy score, VAS score, FeNO, tissue eosinophil
percentage, tissue eosinophil absolute count, blood
eosinophil percentage, blood eosinophil absolute count),
which may be useful predictors for recurrence of CRSwNP.
Furthermore, receiver operating characteristic (ROC)
Predicting recurrence of rhinosinusitis
curves were compared for 2 combinations of predictor: one
of which included all clinical parameters measured, except
E/M ratio (model 1), and the other included all 9 clinical
parameters measured, including E/M ratio (model 2).
Odds ratios (ORs) and 95% confidence intervals were
calculated for each parameter. The ROC curve was plotted
by calculating the sensitivity and specificity of the predic-
tor to find the best cut-off point. The best cut-off points
with optimal sensitivity and specificity were identified as
those that yielded a maximum Youden index. The diagnos-
tic ability of each predictor was calculated based on the
area under the curve (AUC), with an AUC value close to
1 indicating high predictability. An AUC value of >0.9 is
considered to represent high accuracy, and AUC values of
0.7-0.9 and 0.5-0.7 represent moderate and low accuracy,
respectively.
Results
Overall, a total of 42 of 272 patients enrolled in the study
were lost during follow-up. The average follow-up time was
24 months after the first surgery and, of the 230 remaining
patients, 118 exhibited recurrence (51.3%).
The demographic and clinical characteristics of CRSwNP
patients in the recurrence group (n = 118) and non-
recurrence group (n = 112) are shown in Table 1. The
2 groups were not significantly different with respect to age,
gender distribution, comorbidity of allergy, level of FeNO,
nasal obstruction score, runny nose score, headache/facial
pain score, Lund-Mackay score, peripheral eosinophil per-
centage, and peripheral eosinophil absolute count. How-
ever, the onset of surgical history and asthma, total VAS
score and anosmia score, E/M ratio, Lund-Kennedy score,
tissue eosinophil percentage, and tissue eosinophil ab-
solute count were significantly higher in the recurrence
group.
ROC curve analysis of factors associated with CR-
SwNP recurrence and the AUC of each parameter are pre-
sented in Table 2. The E/M ratio and tissue eosinophil
percentage showed high accuracy as predictors for CR-
SwNP recurrence (AUC = 0.947 and 0.916, respectively),
whereas tissue eosinophil absolute count, VAS score, and
Lund-Kennedy score showed moderate accuracy as pre-
dictors for CRSwNP recurrence (AUC = 0.801, 0.801,
and 0.762, respectively). In contrast, the Lund-Mackay
score, peripheral eosinophil absolute count, FeNO, and
peripheral eosinophil percentage exhibited low accuracy
(AUC = 0.542, 0.540, 0.535, and 0.578, respectively)
(Fig. 1A and B). The AUC of combination model 1 was
0.871, indicating moderate accuracy, whereas the AUC of
model 2 was 0.954, indicating high accuracy for CRSwNP
recurrence (Fig. 1C and Table 3). The subgroup analysis for
only patients undergoing primary surgery indicated that the
E/M ratio was still a strong predictor for recurrence in this
group (AUC = 0.935).
The best cut-off points for E/M ratio and tissue eosinophil
percentage, as determined using the Youden index to
International Forum of Allergy & Rhinology, Vol. 00, No. 0, xxxx 2019 3
Meng et al.

TABLE 1. Demographics and clinical characteristics

AUC = area under the curve; CRSwNP = chronic rhinosinusitis with nasal polyps; FeNO = oral fractional exhaled nitric oxide; PEAC = peripheral eosinophil absolute count; PEP = peripheral eosinophil percentage;
of study patients*

Youden index
Recurrent Non-recurrent

0.868
0.744
0.612
0.507
0.392
0.096
0.093
0.092
0.051
(n = 118) (n = 112) p value

Age, mean ± SD (y) 44.7 ± 3.4 43.7 ± 7.1 ns

TABLE 2. ROC curve analysis of factors associated with CRSwNP recurrence and sensitivity and specificity at different thresholds
Gender (M/F) 66/52 62/50 ns
Surgical history (%) 26 (22.0) 14 (12.5) 0.022
Allergy (%) 32 (27.1) 37 (33.0) ns

Specificity
Asthma (%) 36 (30.5) 20 (17.9) 0.043

0.911
0.964
0.866
0.812
0.875
0.562
0.161
0.812
0.170
FeNO, mean ± SD 17.8 ± 4.4 17.3 ± 6.2 ns
Total VAS, mean 24.7 (9-35) 15.7 (8-35) 0.035
(range)
Nasal obstruction, 8.6 (2-10) 4.6 (1-10) ns
mean (range)

Sensitivity
Runny nose, mean 4.6 (0-7) 5.6 (0-9) ns

0.958
0.780
0.746
0.695
0.517
0.534
0.932
0.280
0.881

ROC = receiver operating characteristic; TEAC = tissue eosinophil absolute count; TEP = tissue eosinophil percentage; VAS = visual analog scale.
(range)
Anosmia, mean 7.9 (2-10) 3.4 (0-6) 0.026
(range)
Headache/facial pain, 3.6 (0-8) 2.1 (0-5) ns
mean (range)
Lund-Mackay score, 18.3 (4-24) 17.8 (4-24) ns Cut-off

2.55
0.26

0.12
83.5
17.5
6.5
19.5

22.5
3.3
mean (range)
E/M ratio, mean ± SD 3.6 ± 1.3 2.0 ± 1.2 0.040
Lund-Kennedy score, 6.4 (2-12) 3.8 (3-10) 0.037
mean (range)a
TEP, mean ± SD (%) 25.0 ± 7.8 15.9 ± 3.8 0.041
Uppera

0.979
0.953
0.860
0.858
0.823
0.617
0.615
0.610
0.615
TEAC, mean ± SD 138.9 ± 13.4 60.8 ± 9.1 0.010
(× 109 )
PEP, mean ± SD (%) 7.4 ± 3.3 6.6 ± 4.0 ns
PEAC, mean ± SD 0.5 ± 1.3 0.4 ± 2.2 ns
(× 109 )
*
Lowera

Data obtained using t test for equality of means (2-tailed).

0.914
0.879
0.741
0.743
0.701
0.468
0.466
0.461
0.466
a
Postoperatively.
E/M ratio = ratio of computed tomography score for ethmoid sinus and maxillary
sinus; FeNO = oral fractional exhaled nitric oxide; ns = not statistically signifi-
Data expressed as boundary of the 95% confidence interval.

cant; PEAC = peripheral eosinophil absolute count; PEP = peripheral eosinophil

percentage; SD = standard deviation; TEAC = tissue eosinophil absolute count;
TEP = tissue eosinophil percentage; VAS = visual analog scale.
0.947
0.916
0.801
0.801
0.762
0.542
0.540
0.535
0.578
AUC

discriminate CRSwNP recurrence, demonstrated that a

value of 2.55 for E/M ratio resulted in a Youden index re-
sult of 86.8%, with a sensitivity of 95.8% and a specificity
of 91.1%. This cut-off point of 2.55 for E/M ratio indicated
the CT scan showed ethmoid sinus–dominant opacification
(Fig. 2A). Similarly, setting the optimal cut-off point for
Lund-Kennedy score
Lund-Mackay score

tissue eosinophil percentage at 26%, using histologic as-

sessment (Fig. 3), yielded a Youden index of 74.4%, with
78.0% sensitivity and 96.4% specificity.
Predictors

E/M ratio

Figure 4 shows representative endoscopic images of a

FeNO
PEAC
TEAC

patient with CRS recurrence (Fig. 4A) and a patient without

PEP
VAS
TEP

recurrence (Fig. 4B), after 2 years postsurgery.

a

4 International Forum of Allergy & Rhinology, Vol. 00, No. 0, xxxx 2019
 Predicting recurrence of rhinosinusitis

FIGURE 1. ROC curves of individual clinical predictors. (A) E/M ratio, tissue eosinophil percentage, tissue eosinophil absolute count, blood eosinophil
percentage, and blood eosinophil absolute count. (B) E/M ratio, Lund-Kennedy score, Lund-Mackay score, VAS scores, and FeNO. The E/M ratio had highest
accuracy for CRSwNP recurrence (AUC = 0.947). (C) ROC curves of model 1 (E/M ratio excluded; blue line) and model 2 (E/M ratio included; red line). The
AUCs of these 2 models were 0.871 and 0.954, respectively. AUC = area under the receiver operating characteristic curve; E/M ratio = ratio of the computed
tomography scores for ethmoid sinus and maxillary sinus; FeNO = fractional exhaled nitric oxide; ROC = receiver operating characteristic; VAS = visual analog
scale.

Discussion if a patient is at increased risk of recurrence and a choice has

to be made between surgical treatment or just controlling
CRS is a heterogeneous nasal mucosal disease with a high
the symptoms with currently available medical treatment.
prevalence worldwide.3–5 Thus, the prognosis of CRS and
To date, several methods have been described for predict-
prediction of recurrence are important factors, particularly
ing of the recurrence of CRS postsurgery,11–15 but these
approaches are often invasive.
TABLE 3. ROC curve analysis of models with CRS In the present study, the average follow-up time in 230
recurrence of 272 patients who completed the study was 24 months
after the first surgery and, of these, 51.3% (118 patients)
AUC Lowera Uppera
had CRSwNP recurrence. This recurrence rate is in accor-
dance with our previous finding of 55.3% recurrence after
Model 1 0.871 0.826 0.915 a minimum of 24 months for all patients,11 and comparable
Model 2 0.954 0.929 0.978 to the recurrence 60-70% demonstrated by DeConde and
a
colleagues,23 at 18 months after FESS.
Data expressed as boundary of the 95% confidence interval.
AUC = area under the curve; CRS = chronic rhinosinusitus; ROC = receiver CT has been used for diagnosis of CRS in the recent
operating characteristic. years, especially in the diagnosis of ECRS, which was

International Forum of Allergy & Rhinology, Vol. 00, No. 0, xxxx 2019 5
Meng et al.
FIGURE 2. Computed tomography imaging of ethmoid sinus–dominant opacification (A) and a pan-sinusitis pattern on CT scans (Lund-Mackay scores = 24)
(B). Ethmoid sinus–dominant opacification was observed mostly in recurrence CRS patients. CRS = chronic rhinosinusitis; CT = computed tomography.
FIGURE 3. Histologic assessment of sinus mucosa in a patient with CR-
SwNP recurrence. The sample from this patient demonstrated infiltration
by a large number of eosinophils and an edematous basement membrane
(hematoxylin-eosin stain; magnification ×200). CRSwNP = chronic rhinosi-
nusitis with nasal polyps.
highly associated with CRS recurrence.16, 17, 24 Ishitoya and
colleagues24 indicated that CT scans of ECRS patients typ-
ically show “ethmoid sinus-dominant opacification,” with
opacification of the olfactory cleft also being a feature. Sim-
ilarly, Sakuma and colleagues16 demonstrated that the cut-
off values of Lund-Mackay scores on CT scan (olfactory
cleft score ࣙ1 and posterior ethmoid score ࣙ1) may indi-
cate a highly accurate diagnostic ability for ECRS, whereas
Zuo and colleagues17 suggested that ethmoid osteitis in-
dex on CT scan may be a suitable indicator in the dif-
ferential diagnosis of ECRS.17 Indeed, our own study has
shown that the ratio of CT scores for the ethmoid sinus and
maxillary sinus (E/M ratio) based on Lund-Mackay scores
6 International Forum of Allergy & Rhinology, Vol. 00, No. 0, xxxx 2019
had the highest predictive value for ECRS compared with
other clinical parameters.19 Although these studies clearly
demonstrate the benefit and accuracy of CT scan in diag-
nosis of ECRS, and several studies have demonstrated a
close association between ECRS and CRS recurrence,25–29
none have directly assessed the association between specific
CT scan characteristics and CRSwNP recurrence. Thus, in
the present study we investigated the correlations between
findings of CT scans and CRSwNP recurrence and found
that a cut-off value of 2.55 for E/M ratio on CT scan had
the highest predictive value for CRSwNP recurrence, with
a sensitivity of 95.8% and a specificity of 91.1%, suggest-
ing an E/M ratio >2.55 was indicative of the likelihood of
CRSwNP recurring, even after adequate surgery and post-
surgical management with normative intranasal corticos-
teroid and up to 2 short courses of antibiotics or systemic
corticosteroids. This finding also suggests that CRSwNP
patients who show ethmoid sinus–dominant opacification
beyond a certain level on CT scan have a high possibil-
ity for CRS recurrence (Fig. 3). Pathologically, this can be
explained on the basis that mucosal edema and polypo-
sis are often present around the middle turbinate in ECRS
patients, and this type of disease seen predominantly the
anterior ethmoid sinus, posterior ethmoid sinuses, and ol-
factory cleft usually presents as opacification of the areas
around ethmoid sinuses in CT scans.16, 24 Moreover, it has
been suggested that the maxillary sinus may not be involved
in ECRS, especially in the early stage of the disease.16, 24 In
the present study there were 40 (17.4%) patients who had
a history of surgery. The E/M ratio of the AUC of the pa-
tients in the predictive model was similar to that observed
for the entire eligible study cohort (AUC = 0.904 vs 0.947,
data not shown), suggesting that E/M ratio is applicable in
all CRSwNP patients irrespective of whether they have had
previous sinus surgery.

FIGURE 4. (A) Endoscopic image of a 56-year-old female patient, after 2 years postsurgery. This patient had recurrence of CRSwNP, with the nasal cavity
showing an edematous mucosa with recurrence of polyps. (B) Endoscopic image of a 37-year-old female patient, after 2 years postsurgery. This patient had a
normal nasal mucosa showing no recurrence of polyps. CRSwNP = chronic rhinosinusitis with nasal polyps.
Tissue eosinophil percentage was also found to have good
predictive value for CRS recurrence, with the optimal cut-
off point being 26%, with 78.0% sensitivity and 96.4%
specificity. This result is also in accordance with our ear-
lier study, which indicated that tissue eosinophils >27%
of total cells may be a reliable prognostic indicator for
recurrence.11 However, unlike CT scans, the use of tissue
eosinophils for predicting recurrence of CRS in the present
study is limited by 2 major factors. First, obtaining nasal tis-
sue requires an invasive procedure that may not be tolerated
by some patients due to bleeding, and the complete safety of
the procedure cannot be guaranteed in the elderly or in chil-
dren. Furthermore, obtaining the tissue is not convenient in
the outpatient clinic. Second, irrespective of whether per-
centage of tissue eosinophils is employed for the diagnosis
of ECRS or prediction for CRS recurrence, the criterion
for tissue eosinophilia percentage has been shown to range
from 5% to 50%,11, 25–29 which makes it difficult to define
a reliable measure. From this standpoint, the E/M ratio ap-
pears to be superior to tissue eosinophil percentage as a
prognostic indicator for CRS recurrence. However, some
patients in the present study showed a pan-sinusitis pattern
on CT scan (Lund-Mackay scores = 24; Fig. 2B), suggesting
that, under this circumstance, the E/M ratio would be non-
applicable and less meaningful when compared with tissue
eosinophil percentage. Nevertheless, our study has indi-
cated that blood eosinophil count had low predictive value
for CRS recurrence, and therefore, unless pan-sinusitis is
noted, it would be more useful to check just the E/M ra-
tio on the CT scan for prognosis. Although this finding is
in accordance with our earlier study,11 it is not in accor-
dance with studies by Hu and colleagues30 and Sakuma and
colleagues,16 which showed AUC values of >0.8 for blood
eosinophil counts. This difference between our study and
theirs may be due to differences in the study cohorts. In
particular, both of those earlier studies focused on differ-
entiation of ECRS from non-ECRS and demonstrated that
increased blood eosinophils were indeed a good differen-
tiator between ECRS and non-ECRS patients. In contrast,
Predicting recurrence of rhinosinusitis
we have focused on recurrence of disease rather than differ-
entiation between eosinophilic/noneosinophilic status, and
therefore data for the entire cohort comprising both ECRS
and non-ECRS patients were analyzed as a single data set.
Our study has shown that Lund-Kennedy, VAS, and
anosmia scores were significantly higher in the CRSwNP
recurrence group compared with the non-recurrence group.
These findings are consistent with those of previous
studies.8, 9, 11 As mucosal edema and polyposis were usually
limited to the mucosa around the olfactory cleft, the loss of
smell was more severe for the CRSwNP recurrence patients,
with the other symptoms not significantly affected between
groups. We also found that the CRSwNP recurrence group
had a higher prevalence of asthma than the non-recurrence
group. This finding is also in accordance with earlier stud-
ies, which suggested that a higher prevalence of asthma was
closely associated with polyp recurrence.9, 11, 31, 32 Further-
more, consistent with previous study findings showing that
atopic status was not associated with recurrence of nasal
polyps,9, 11 we were also unable to show any significant
difference between groups for prevalence of allergy.
Conclusion
There is little doubt that, despite the development of dif-
ferent treatment strategies, recurrence still presents a major
challenge in the management of CRS. In this study we found
that patients with CRSwNP recurrence exhibited several
clinical characteristics, including increased frequency of
surgical history, high prevalence of asthma, high VAS score,
high anosmia score, high Lund-Kennedy score, high tissue
eosinophil count, and ethmoid sinus–dominant opacifica-
tion on CT scan. ROC analysis and linear regression sug-
gested that E/M ratio of the CT scan at a cut-off point of
2.55 has the highest predictive value for CRSwNP recur-
rence in the Chinese population, and that an E/M ratio of
>2.55 indicates a greater likelihood for recurrence of CR-
SwNP after surgery and/or appropriate therapy. This non-
invasive predictor for prognosis of CRSwNP will likely add
International Forum of Allergy & Rhinology, Vol. 00, No. 0, xxxx 2019 7
Meng et al.
to our knowledge base and lead to identification of the opti-
mal medical treatment for management of CRSwNP recur-
rence. All CRSwNP patients, apart from children, should
References
1. Fokkens WJ, Lund VJ, Mullol J, et al. European Posi-
tion Paper on Rhinosinusitis and Nasal Polyps 2012.
Rhinol Suppl. 2012;3:1-298.
2. Bachert C, Akdis CA. Phenotypes and emerging endo-
types of chronic rhinosinusitis. J Allergy Clin Immunol
Pract. 2016;4:621-628.
3. Hirsch AG, Stewart WF, Sundaresan AS, et al. Nasal
and sinus symptoms and chronic rhinosinusitis in a
population-based sample. Allergy. 2017;72:274-281.
4. Hastan D, Fokkens WJ, Bachert C, et al. Chronic rhi-
nosinusitis in Europe—an underestimated disease: a
GA2LEN study. Allergy. 2011;66:1216-1223.
5. Shi JB, Fu QL, Zhang H, et al. Epidemiology of
chronic rhinosinusitis: results from a cross-sectional
survey in seven Chinese cities. Allergy. 2015;70:533-
539.
6. Zhang Y, Gevaert E, Lou H, et al. Chronic rhinosinusi-
tis in Asia. Allergy Clin Immunol. 2017;140:1230-
1239.
7. Lou H, Meng Y, Piao Y, et al. Cellular phenotyping
of chronic rhinosinusitis with nasal polyps. Rhinol-
ogy. 2016;54:150-159.
8. Liao B, Liu JX, Li ZY, et al. Multidimensional en-
dotypes of chronic rhinosinusitis and their associa-
tion with treatment outcomes. Allergy. 2018;73:1459-
1469.
9. Wei B, Liu F, Zhang J, et al. Multivariate analysis of in-
flammatory endotypes in recurrent nasal polyposis in
a Chinese population. Rhinology. 2018;56:216-226.
10. Zhang N, Van Zele T, Perez-Novo C, et al. Different
types of T effector cells orchestrate mucosal inflamma-
tion in chronic sinus disease. J Allergy Clin Immunol.
2008;122:961-968.
11. Lou H, Meng Y, Piao Y, Wang C, Zhang L, Bachert C.
Predictive significance of tissue eosinophilia for nasal
polyp recurrence in the Chinese population. Am J Rhi-
nol Allergy. 2015;29,1-7.
12. Younis RT, Ahmed J. Predicting revision sinus surgery
in allergic fungal and eosinophilic mucin chronic rhi-
nosinusitis. Laryngoscope. 2017;127:59-63.
8 International Forum of Allergy & Rhinology, Vol. 00, No. 0, xxxx 2019
have had a CT scan before surgery for determination of
E/M ratio and should avoid any additional investigations,
unless pan-sinusitis has been demonstrated.
13. Veloso-Teles R, Cerejeira R. Endoscopic sinus surgery
for chronic rhinosinusitis with nasal polyps: clinical
outcome and predictive factors of recurrence. Am J
Rhinol Allergy. 2017;31:56-62.
14. Ninomiya T, Noguchi E, Haruna T, et al. Periostin
as a novel biomarker for postoperative recurrence
of chronic rhinosinitis with nasal polyps. Sci Rep.
2018;8:11450.
15. Amali A, Bidar Z, Rahavi-Ezabadi S, Mikaniki
N, Sadrehosseini SM. Polypoid change of middle
turbinate is associated to an increased risk of polyp re-
currence after surgery in patients with chronic rhinosi-
nusitis with nasal polyps. Eur Arch Otorhinolaryngol.
2018;275:2021-2025.
16. Sakuma Y, Ishitoya J, Komatsu M, et al. New clini-
cal diagnostic criteria for eosinophilic chronic rhinos-
inusitis. Auris Nasus Larynx. 2011;38:583-588.
17. Zuo K, Guo J, Chen F, et al. Clinical characteristics
and surrogate markers of eosinophilic chronic rhinosi-
nusitis in southern China. Eur Arch Otorhinolaryngol.
2014;271:2461-2468.
18. Hong H, Wang D, Tan KS, et al. Sinus computed to-
mography predicts clinical response to corticosteroids
in chronic rhinosinusitis with nasal polyps. Clin Transl
Allergy. 2018;24:1-8.
19. Meng Y, Lou H, Wang C, Zhang L. Predictive sig-
nificance of computed tomography in eosinophilic
chronic rhinosinusitis with nasal polyps. Int Forum
Allergy Rhinol. 2016;6:812-819.
20. Wang K, Wang C, Xi L, et al. A randomized con-
trolled trial to assess adherence to allergic rhinitis
treatment following a daily short message service
(SMS) via the mobile phone. Int Arch Allergy Im-
munol. 2014;163:51-58.
21. Lund VJ, Mackay IS. Staging in rhinosinusitis. Rhi-
nology. 1993;31:183-184.
22. Lund VJ, Kennedy DW. Staging for rhinosinusitis.
Otolaryngol Head Neck Surg. 1997;117(suppl):S35-
S40.
23. DeConde AS, Mace JC, Levy JM, Rudmik L, Alt JA,
Smith TL. Prevalence of polyp recurrence after en-
doscopic sinus surgery for chronic rhinosinusitis with
nasal polyposis. Laryngoscope. 2017;127:550-555.
24. Ishitoya J, Sakuma Y, Tsukuda M. Eosinophilic
chronic rhinosinusitis in Japan. Allergol Int. 2010;59:
239-245.
25. Kim JW, Hong SL, Kim YK, Lee CH, Min YG, Rhee
CS. Histological and immunological features of non-
eosinophilic nasal polyps. Otolaryngol Head Neck
Surg. 2007;137:925-930.
26. Jankowski R, Bouchoua F, Coffinet L, Vignaud JM.
Clinical factors influencing the eosinophil infiltration
of nasal polyps. Rhinology. 2002;40:173-178.
27. Jeong WJ, Lee CH, Cho SH, Rhee CS. Eosinophilic
allergic polyp: a clinically oriented concept of nasal
polyp. Otolaryngol Head Neck Surg. 2011;144:241-
246.
28. Bonfils P, Badoual C, Bonfils NA, Gallas D, Malin-
vaud D. Eosinophil infiltration of nasal polyps in pa-
tients with nasal polyposis: role in clinical evolution
after medical and surgical treatment. J Laryngol Otol.
2009;123:509-516.
29. Tecimer SH, Kasapoglu F, Demir UL, Ozmen OA,
Coskun H, Basut O. Correlation between clini-
cal findings and eosinophil/neutrophil ratio in pa-
tients with nasal polyps. Eur Arch Otorhinolaryngol.
2015;272:915-921.
30. Hu Y, Cao PP, Liang GT, Cui YH, Liu Z. Diagnostic
significance of blood eosinophil count in eosinophilic
chronic rhinosinusitis with nasal polyps in Chinese
adults. Laryngoscope. 2012;122:498-503.
31. Nakayama T, Yoshikawa M, Asaka D, et al. Mu-
cosal eosinophilia and recurrence of nasal polyps:
new classification of chronic rhinosinusitis. Rhinol-
ogy. 2011;49:392-396.
32. Pearlman AN, Chandra RK, Chang D, et al. Relation-
ships between severity of chronic rhinosinusitis and
nasal polyposis, asthma, and atopy. Am J Rhinol Al-
lergy. 2009;23:145-148.