Impact of Rapid On-Site Cytological Evaluation (ROSE) on the Diagnostic Yield of Transbronchial Needle Aspiration During Mediastinal Lymph Node Sampling: Systematic Review and Meta-Analysis

Abstract

Background:
Whether the use of rapid onsite cytological evaluation (ROSE) increases the diagnostic yield of transbronchial needle aspiration (TBNA) remains unclear. Herein, we perform a systematic review of studies describing the utility of ROSE in subjects undergoing TBNA.

Methods:
Systematic review of PubMed, EmBase and Scopus databases for RCTs investigating the diagnostic yield of conventional TBNA (c-TBNA) or endobronchial ultrasound (EBUS)-TBNA, with or without ROSE, in subjects with mediastinal lymphadenopathy.

Results:
Five (618 subjects; two EBUS-TBNA, two c-TBNA, one both) studies were identified. Overall, the studies were of good quality. The pooled risk difference (95% CI) of the diagnostic yield of EBUS-TBNA and c-TBNA was 0.04 (-0.01, 0.09) and 0.12 (-0.08, 0.33), respectively suggesting no added benefit with ROSE. The use of ROSE during EBUS-TBNA (but not c-TBNA) resulted in a significantly lesser number of needle passes (mean difference, [95% CI], -1.1 [-2.2, -0.005; p<0.001]). There was no difference in the procedure time during EBUS-TBNA. The complication rate was significantly lower (OR [95% CI], 0.26 [0.10, 0.71], p=0.009]), when ROSE was used during c-TBNA, due to fewer additional procedures required to make a diagnosis. There was evidence of heterogeneity in the studies involving c-TBNA but not EBUS-TBNA. There was no publication bias.

Conclusion:
The use of ROSE neither improved the diagnostic yield nor reduced the procedure time during TBNA. However, the use of ROSE was associated with lesser number of needle passes during EBUS-TBNA, and overall fewer requirement of additional bronchoscopy procedures during TBNA to make a final diagnosis.

Full text

Impact of Rapid On-Site Cytological
Evaluation (ROSE) on the Diagnostic Yield
of Transbronchial Needle Aspiration During
Mediastinal Lymph Node Sampling
Systematic Review and Meta-Analysis
Inderpaul Singh Sehgal, MD, DM; Sahajal Dhooria, MD, DM; Ashutosh Nath Aggarwal, MD, DM;
and Ritesh Agarwal, MD, DM
BACKGROUND: Whether the use of rapid on-site cytologic evaluation (ROSE) increases the
diagnostic yield of transbronchial needle aspiration (TBNA) remains unclear. This article is a
systematic review of studies describing the utility of ROSE in subjects undergoing TBNA.
METHODS: The study included a systematic review of the PubMed, Embase, and Scopus
databases for randomized controlled trials investigating the diagnostic yield of conventional
transbronchial needle aspiration (c-TBNA) or endobronchial ultrasound (EBUS)-TBNA,
with or without ROSE, in subjects with mediastinal lymphadenopathy.
RESULTS: Five studies (618 subjects; two EBUS-TBNA, two c-TBNA, and one both) were
identified. Overall, the studies were of good quality. The pooled risk difference (95% CI) of
the diagnostic yield of EBUS-TBNA and c-TBNA was 0.04 (–0.01 to 0.09) and 0.12 (–0.08 to
0.33), respectively, suggesting no added benefit with ROSE. The use of ROSE during EBUS-
TBNA (but not c-TBNA) resulted in significantly fewer needle passes (mean difference
[95% CI], –1.1 [–2.2 to –0.005]; P<.001). There was no difference in the procedure time
during EBUS-TBNA. The complication rate was significantly lower (OR [95% CI], 0.26 [0.10
to 0.71]; P¼.009) when ROSE was used during c-TBNA due to fewer additional procedures
required to make a diagnosis. There was evidence of heterogeneity in the studies involving
c-TBNA but not EBUS-TBNA. There was no publication bias.
CONCLUSIONS: The use of ROSE neither improved the diagnostic yield nor reduced the
procedure time during TBNA. However, the use of ROSE was associated with fewer number
of needle passes during EBUS-TBNA and overall lower requirement for additional bron-
choscopy procedures during TBNA to make a final diagnosis.
TRIAL REGISTRY: PROSPERO; No.: CRD42017058937; URL: www.crd.york.ac.uk/prospero/.
CHEST 2017; -(-):---
KEY WORDS: cytology; EBUS; endoscopic ultrasound; endosonography; EUS; lung cancer;
rapid on-site evaluation; sarcoidosis
ABBREVIATIONS: c-TBNA = conventional transbronchial needle
aspiration; EBUS = endobronchial ultrasound; ROSE = rapid on-site
cytologic evaluation; TBNA = transbronchial needle aspiration
AFFILIATIONS: From the Department of Pulmonary Medicine, Post-
graduate Institute of Medical Education and Research, Chandigarh,
India.
CORRESPONDENCE TO: Ritesh Agarwal, MD, DM, Department of
Pulmonary Medicine, Postgraduate Institute of Medical Education and
Research, Sector-12, Chandigarh-160012, India; e-mail: agarwal.
ritesh@outlook.in
Copyright Ó2017 American College of Chest Physicians. Published by
Elsevier Inc. All rights reserved.
DOI: https://doi.org/10.1016/j.chest.2017.11.004
[Original Research ]
chestjournal.org 1

Transbronchial needle aspiration (TBNA) is a routine
procedure for sampling mediastinal lymph nodes.
Conventionally, TBNA is performed using a flexible
bronchoscope and, recently, employing the
endobronchial ultrasound (EBUS). The yield of
conventional transbronchial needle aspiration
(c-TBNA) and EBUS-TBNA depends on several factors,
including the etiology (benign or malignant), lymph
node size (<1or>1 cm), type of sedation used
(conscious sedation or general anesthesia), number of
passes per lymph node station ($3or<3), and the
lymph node station being sampled (station 4R, 7, or
others).
1-3
Another factor that can potentially increase
the diagnostic yield of TBNA is rapid on-site cytologic
evaluation (ROSE).
ROSE provides immediate feedback regarding the
adequacy of the specimens obtained and can thus
increase the diagnostic yield.
4-7
In case of an
inadequate sample, ROSE may guide the operator to
modify the technique of TBNA by changing the
lymph node, the puncture site, the depth and angle of
puncture, and the use of suction.
4,8
Intuitively, the use
of ROSE during TBNA has the potential to reduce the
number of needle passes and thus the procedure time.
Furthermore, it can reduce the need for additional
procedures. Despite a sound logic, ROSE is not widely
used, and its utility remains unclear.
9,10
Several
observational studies have shown that ROSE increases
the yield of TBNA.
11,12
However, most of these studies
have been small and retrospective, in which ROSE was
performed in nonconsecutive subjects based on the
physician’s discretion, the underlying diagnosis, and
the size of lymph nodes (choosing smaller lymph
nodes). All these factors can introduce a selection bias,
wherein the real benefit of ROSE remains unclear. We
conducted a systematic review and meta-analysis of
randomized controlled trials investigating the
diagnostic yield of c-TBNA or EBUS-TBNA with or
without ROSE in the evaluation of patients with
mediastinal lymphadenopathy.
Materials and Methods
This review was conducted in accordance with guidelines of Preferred
Reporting Items for Systematic Reviews and Meta-Analyses
statement.
13
An ethics committee approval was not required because
this study was a systematic review of published data.
PICO (Patients, Intervention[s], Control, Outcomes)
Question
The PICO question was whether ROSE (vs no ROSE) improved
diagnostic yield in patients with mediastinal lymphadenopathy
undergoing EBUS-TBNA or c-TBNA.
Search Strategy
The PubMed, Embase, and Scopus databases (from inception until
March 31, 2017) were searched by using the following free text terms:
(ebus OR eus OR endosono* OR “endobronchial ultrasound”OR
“endoscopic ultrasound”OR “ebus-tbna”OR “eus-fna”) AND
(“transbronchial needle aspiration”OR “tbna”OR “needle aspiration”)
AND (“rapid onsite evaluation”OR “rose”OR “rapid onsite cytological
evaluation”). The reference list of all the included articles and previous
review articles were searched. In addition, we searched our personal files.
Inclusion Criteria
Studies meeting the following criteria were included: (1) randomized
controlled trials in which the subjects underwent mediastinal lymph
node sampling using either c-TBNA or EBUS-TBNA, with or
without ROSE; and (2) studies providing outcome of the procedures
with or without ROSE, thereby allowing calculation of diagnostic
yield from the study observations.
The following type of studies were excluded: (1) observational studies;
(2) case reports, abstracts, comments, editorials, and reviews; (3)
studies not providing the diagnostic yield of procedures performed
by using ROSE separately; (4) studies describing the use of ROSE in
sampling peripheral lung lesions; and (5) studies describing the use
of ROSE in transthoracic sampling of thoracic lesions.
Initial Review of Studies
The electronic searches were assimilated in a reference manager
package (Endnote [version X8; Clarivate Analytics]), and all
duplicate citations were discarded. Two authors (I. S. S. and R. A.)
screened the citations by review of the title and abstract to identify
the relevant studies. Any disagreement was resolved by consensus
between the authors. This database was then scrutinized again to
include only primary articles. The full text of each of these studies
was obtained and reviewed in detail.
Study Selection and Data Abstraction
Two authors (I. S. S. and R. A.) independently extracted the data into a
standard data extraction form. The following information was retrieved:
(1) publication details (authors, year of publication, and country where
the study was conducted); (2) number of patients, inclusion criteria, and
demographic profile of patients; (3) the type of sedation or anesthesia
used; (4) lymph node stations sampled by c-TBNA or EBUS-TBNA; (5)
diameter of conventional and EBUS needle, number of passes made
through conventional and EBUS-TBNA, with or without ROSE; (6) the
adequacy (preponderance of lymphocytes) and diagnostic yield
(detection of malignant cells or granuloma or abnormality in the lymph
nodes resulting in a specific diagnosis) of c-TBNA and EBUS-TBNA;
(7) reagent used for rapid staining of the cytology specimens; (8)
duration of procedure; and (9) complications associated with the
procedures. Any differences in the data extraction process were resolved
by discussion.
Assessment of Study Quality
The quality of each included study was independently evaluated by two
authors (I. S. S. and R. A.) using the Cochrane risk of bias tool.
14
This
tool assesses the risk of bias and applicability judgment based on
randomization sequence generation, allocation concealment and
blinding of participants and personnel, attrition of participants,
selective reporting of results, and other sources of bias. Each item is
rated as low, high, or unclear risk of bias.
2Original Research [-#-CHEST -2017 ]

Data Analysis
The statistical software package RevMan (Review Manager, version 5.3:
The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) was
used to perform the statistical analyses.
Determination of the Pooled Effect
For each study, the risk difference (RD), with 95% CIs, was calculated
for the diagnostic yield of EBUS-TBNA and c-TBNA with or without
ROSE. We also calculated the mean difference along with 95% CI, for
the number of needle passes attempted to obtain specimens and the
procedure time for c-TBNA and EBUS-TBNA procedures; the OR
(95% CI) of any complication (due to TBNA or additional
procedures) encountered in the two groups was also calculated. All
the individual study estimates were pooled by using a random
effects model. Forest plots were generated to display the
RD estimates.
Assessment of Heterogeneity and Subgroup Analysis
Heterogeneity for the individual outcomes was assessed by using the I
2
test, with a value $50% indicating significant heterogeneity.
15
Heterogeneity was also assessed by using the Cochran Q statistic,
and a Pvalue <0.1 was considered to be significant. To determine
the causes of heterogeneity, sensitivity analysis stratified according to
predefined covariates was planned. This method included the type of
sedation used (conscious sedation or general anesthesia), indication
for TBNA (benign or malignant disorder), gauge of needle used, use
of stylet, use of suction, number of revolutions in the lymph node,
and the lymph node stations sampled.
Estimation of Publication Bias
The presence of publication bias was evaluated by using a funnel plot
(log RD on the horizontal axis against SE of RD on the vertical axis).
16
Publication bias was also investigated by using the Egger test
17
and the
Begg and Mazumdar test.
16
Results
The initial database search yielded a total of
782 citations, of which five studies (618 subjects;
three EBUS-TBNA, and three c-TBNA) were
included in the present analysis (Fig 1).
4-7,18
One study compared the diagnostic yield (with
or without ROSE) during both EBUS-TBNA
and c-TBNA.
18
Databases searched: PubMed, EmBase and Scopus
Search Terms: (“EBUS” OR “endobronchial ultrasound” OR
“endosonography” OR “bronchial ultrasound” AND
“transbronchial needle aspiration” OR “tbna” OR “needle
aspiration” AND “rose” OR “rapid onsite evaluation” OR “rapid
onsite cytological evaluation”)
Citations found after initial search (N = 782)
Studies assessed for eligibility (n = 132)
Studies included in the systematic review (n = 5)a
Studies describing ROSE during
EBUS (n = 3)
Studies describing ROSE in
conventional TBNA (n = 3)
Studies excluded after initial review (n = 650)
Studies excluded
No comparison between ROSE and no-ROSE
(n = 50)
Abstract/letter only (n = 46)
Duplicate (n = 4)
Review article (n = 3)
EBUS/ROSE not used (n = 15)
Observational studies (n = 9)
Figure 1 –Study selection process for the systematic review.
a
One study included both EBUS-TBNA and conventional TBNA. EBUS ¼endobronchial
ultrasound; ROSE ¼rapid on-site cytologic evaluation; TBNA ¼transbronchial needle aspiration.
chestjournal.org 3

Studies Describing ROSE in EBUS-TBNA
Three studies (345 subjects) described the utility of ROSE
during EBUS-TBNA (Tables 1,2).
4,5,18
Two studies
4,5
included subjects with undiagnosed mediastinal
lymphadenopathy or suspected lung cancer; one study
included subjects with suspected sarcoidosis.
18
EBUS was
performed under conscious sedation in all the studies.
ROSE was performed with either Diff-Quik (Baxter
Diagnostics, Inc; and Kokusaishiyaku)
4,5
or toluidine
blue.
18
None of the trials provided information regarding
the discrepancy between the ROSE and the final diagnosis.
The RD of the diagnostic yield during EBUS-TBNA
varied from –0.02 to 0.12 (Fig 2), with the pooled RD
(95% CI) being 0.04 (–0.01 to 0.09), suggesting
equivalence of the two strategies. The use of ROSE
resulted in a significantly fewer number of passes (two
studies
4,5
) during EBUS-TBNA vs without ROSE (pooled
mean difference [95% CI], –1.1 [–2.2 to –0.005; P<
.001]). The procedure time was similar with or without
ROSE (two studies
5,18
): pooled mean difference (95% CI),
0.06 (–2.5 to 2.7) min; P¼.95. One study found a
significant decrease in additional procedures required to
make the final diagnosis in the ROSE arm.
5
The rate of
sample adequacy (one study) was similar with or without
ROSE.
4
Studies Describing ROSE in Conventional TBNA
Three studies (n ¼273) have described the utility of
ROSE during c-TBNA (Tables 1,2).
6,7,18
ROSE was
performed by using Diff-Quik
6,7
or toluidine blue.
18
Two studies were performed in the evaluation of
undiagnosed mediastinal abnormalities, and one was
conducted in suspected sarcoidosis. c-TBNA was
performed under conscious sedation in all studies. One
study reported a discordance rate of 10.8% between the
ROSE and the final pathology results.
7
The RD of the diagnostic yield during c-TBNA varied
from 0.03 to 0.41 (Fig 2) with a pooled RD (95% CI) of
0.12 (–0.08 to 0.33), suggesting no difference between the
two strategies. Two studies provided the procedure time,
which was similar between the two groups (pooling of
results not possible).
6,18
Two studies provided the
number of passes made for lymph node sampling, and it
was similar between the two groups.
6,7
The rates of
adequate samples were similar in both arms on a per-
patient
6
and a per-lymph node basis.
7
Quality of Studies Included
The studies were generally of good quality with a low
risk of bias (Fig 2).
Complications
There were minor complications (minor bleeding,
hypoxemia, transient laryngospasm, and others) during
EBUS-TBNA and were not different with or without
ROSE. In one study, three pneumothoraces were
observed following lung biopsy and were equally
distributed between the two arms.
18
On pooling the
complications, there was no significant difference
between the odds for complication with or without
ROSE (OR [95% CI], 1.67 [0.38 to 7.29]; P¼0.49)
during EBUS (Fig 3). The odds for complication was
significantly lower when ROSE was used during c-TBNA
(OR [95% CI], 0.26 [0.10, 0.71]; P¼.009). The
complications were attributed to additional procedures
required to make a diagnosis without ROSE, rather than
due to c-TBNA itself.
Heterogeneity
There was clinical heterogeneity as the studies had
included subjects with different underlying conditions,
used different needle types, and included varying
reagents for performing ROSE (Tables 1,2). There was no
statistical heterogeneity observed in the trials on EBUS-
TBNA. However, there was significant heterogeneity in
the randomized trials (I
2
, 64) describing c-TBNA.
Subgroup Analysis
A subgroup analysis was performed in subjects with lung
cancer. The pooled RD (95% CI) of the diagnostic yield
of EBUS-TBNA for diagnosis and staging of lung cancer
was 0.06 (–0.05 to 0.16).
4,5
The pooled RD (95% CI) of
the diagnostic yield of c-TBNA for lung cancer was 0.03
(–0.08 to 0.14).
6,7
For other prespecified covariates, a
sensitivity analysis could not be performed because the
diagnostic yield was not available separately for these
variables.
Publication Bias
There was no evidence of publication bias on visual
analysis of the funnel plots (Fig 4) for both EBUS-TBNA
and c-TBNA. There was no evidence of publication bias
on either the Egger test or the Begg and Mazumdar test
for studies describing EBUS-TBNA (P¼.40 and .48,
respectively, for the Egger and Begg and Mazumdar
tests) and c-TBNA (P¼.62 and .99 for the Egger and
Begg and Mazumdar tests).
Discussion
The results of this meta-analysis suggest that the use of
ROSE does not enhance the diagnostic yield of c-TBNA
and EBUS-TBNA during mediastinal lymph node
4Original Research [-#-CHEST -2017 ]

TABLE 1 ]Description of Studies on TBNA With or Without ROSE
Author/Year
Place
of Study
No. of
Patients
Underlying
Disease
Type of
Sedation
Used
Lymph
Node
Stations
Sampled
Needle
Size
Staining
Agent
Person
Performing
ROSE
Consecutive
Patients
Randomization
Computer
Generated Age (y)
Sex
Distribution
Discordant
Results,
No. (%)
EBUS-TBNA
Madan
et al,
2017
18
India 80 (20 in
each
group)
Sarcoidosis Conscious
sedation
4R or 7 or
both
21 G Toluidine
blue
stain
Pathologist Yes Yes 37.7 9.7 M:F
(1.5:1)
.
Trisolini
et al,
2015
4
Italy 197 Known or
suspected
advanced
lung cancer
.2, 3, 4, 7,
10, 11,
lung
lesions
.Diff-Quik Pathologist Yes Yes Mean,
68.4
M:F
(1.7:1)
.
Oki et al,
2013
5
Japan 108 Suspected
lung cancer
Conscious
sedation
2, 3, 4, 7,
10, 11,
lung
lesions
.Diff-Quik Technician No Yes Range,
34-84
M:F
(3.3:1)
.
c-TBNA
Madan
et al,
2017
18
India 80 (20 in
each
group)
Sarcoidosis Conscious
sedation
4R or 7 or
both
21 G Toluidine
blue
Pathologist Yes Yes 37.7 9.7 M:F
(1.5:1)
.
Yarmus
et al,
2011
6
United
States
68 Undiagnosed
hilar or
mediastinal
lymph-
adenopathy
Conscious
sedation
2, 4, 7,
10, 11,
mass
19 G or
22 G
Diff-Quik .Yes Yes Median
(range),
68 (32-
88)
M:F
(1.4:1)
.
Trisolini
et al,
2011
7
Italy 168 Undiagnosed
hilar or
mediastinal
lymph-
adenopathy
Conscious
sedation
4, 7, 10,
11,
mass
19 G or
22 G
Diff-Quik Pathologist Yes Yes Mean,
63.8
M:F
(2.4:1)
9 (10.8)
c-TBNA ¼conventional transbronchial needle aspiration; EBUS ¼endobronchial ultrasound; F ¼female; M ¼male; ROSE ¼rapid on-site cytologic evaluation; TBNA ¼transbronchial needle aspiration.
chestjournal.org 5

TABLE 2 ]Comparison of Outcomes With or Without ROSE During TBNA
Author/Year
With ROSE Without ROSE
No. of
Patients
Diagnostic
Yield (n/N)
Procedure
Time
(minutes)
No. of
Passes
Complications
(n/N)
Additional
Procedures
Needed
Sample
Adequacy
Rate (n/N
[%])
No. of
patients
Diagnostic
Yield (n/N)
Procedure
Time
(minutes)
No. of
Passes
Complications
(n/N)
Additional
Procedures
Needed
(n/N)
Sample
Adequacy
Rate (n/N
[%])
EBUS-TBNA
Madan
et al,
2017
18
20 12/18 25 4.8 NA 2/18 (Ptx) .17/20
(85%)
20 13/19 25 4.9 3 1/19 (Ptx) ..
Trisolini
et al,
2015
4
98 96/98 NA 4 1.5 3/65 .92/98
(94.3%)
99 94/99 NA 4 1.5 2/61 .96/99
(97.1%)
Oki et al,
2013
5
55 47/55 22.3 
15.9
2.2 0.9 0 6/55
a
FP: 2 FN:
2
b
53 39/53 22.1 7.7 3.1 0.4 0 30/53 .
c-TBNA
Madan
et al,
2017
18
20 13/18 25 4.8 NA 0 .17/20
(85%)
20 6/19 254.9 3 1/19 (Ptx) ..
Yarmus
et al,
2011
6
34 19/34 29.5 Mean,
4.5
NR .32/34
(94.1%)
34 18/34 27.6 4.1 NR .30/34
(88%)
Trisolini
et al,
2011
7
83 65/83 .1.3 0.8 1/83
(major
bleeding);
5/83
.80/102
b
(78.4%)
85 64/85 .1.7 0.8 2/85
(major
bleeding);
17/85
.109/
126
b
(86.5)
Values presented as mean standard deviation unless otherwise stated. FN ¼false-negative; FP ¼false-positive; NA ¼not available; NR ¼not reported; Ptx ¼pneumothorax. See Table 1 legend for expansion of other
abbreviations.
a
Within same setting.
b
Analysis is per lesion.
6Original Research [-#-CHEST -2017 ]

-1 -0.5
No ROSE ROSE
0 0.5 1
Madan et al 2017
Trisolini et al 2015
Oki et al 2013
12
96
47
18
98
55
13
94
39
19
99
53
2.5%
87.3%
10.1%
–0.02 [–0.32 to 0.28]
0.03 [–0.02 to 0.08]
0.12 [–0.03 to 0.27]
Subtotal (95% CI)
1.1.1 EBUS-TBNA
Study Events Total
ROSE NO ROSE
Events Total Weight M-H, Random, 95% CI
Risk Difference
M-H, Random, 95% CI A B C D E F G
Risk Difference Risk of Bias
171 171 100.0% 0.04 [–0.01 to 0.09]
Trisolini et al 2011
Madan et al 2017
Yarmus et al 2011
65
13
19
83
18
34
64
6
18
85
19
34
43.8%
25.2%
30.9%
0.03 [–0.10 to 0.16]
0.41 [0.11 to 0.70]
0.03 [–0.21 to 0.27]
Subtotal (95% CI)
1.1.2 c-TBNA
135 138 100.0% 0.12 [–0.08 to 0.33]
++
++ ++
+++++
+
++++
+
+
+
++
+++
+
++++
Heterogeneity: Tau2 = 0.02; Chi2 = 5.50, df = 2 (P = .06); I2 = 64%
Test for overall effect: z = 1.20 (P = .23)
Test for subgroup differences: Chi2 = 0.66, df = 1 (P = .42); I2 = 0%
(A) Random sequence generation (selection bias)
(B) Allocation concealment (selection bias)
(C) Blinding of participants and personnel (performance bias)
(D) Blinding of outcome assessment (detection bias)
(E) Incomplete outcome data (attrition bias)
(F) Selective reporting (reporting bias)
(G) Other bias
Risk of bias legend
Total events 97 88
Heterogeneity: Tau2 = 0.00; Chi2 = 1.70, df = 2 (P = .43); I2 = 0%
Test for overall effect: z = 1.54 (P = .12)
Total events 155 146
Figure 2 –Forest plot of the risk difference comparing the diagnostic yield of EBUS-TBNA and c-TBNA with or without ROSE. The risk difference of
individual studies is represented by a square through which runs a horizontal line (95% CI). The diamond with horizontal lines represents the pooled
risk difference with 95% CI. Also depicted is the risk of bias of the individual studies. c-TBNA ¼conventional TBNA; M-H ¼Maentel-Hanszel test. See
Figure 1 legend for expansion of other abbreviations.
0.01
No ROSEROSE
0.1 1 10 100
1.3.1 EBUS-TBNA
Study Events Total
ROSE NO ROSE
Events Total Weight M-H, Random, 95% CI
OR
M-H, Random, 95% CI
OR
Madan et al 2017
Trisolini et al 2015
Oki et al 2013
0
3
2
55
65
18
0
2
1
53
61
19
65.1%
34.9%
Not estimable
1.43 [0.23 to 8.85]
2.25 [0.19 to 27.22]
Subtotal (95% CI) 138 133 100.0% 1.67 [0.38 to 7.29]
Heterogeneity: Tau2 = 0.00; Chi2 = 0.08, df = 1 (P = .77); I2 = 0%
Test for overall effect: z = 0.69 (P = .49)
Test for sub
g
roup differences: Chi2 = 4.16, df = 1 (P = .04), I2 = 76.0%
Total events 5 3
1.3.2 c-TBNA
Madan et al 2017
Trisolini et al 2011
Subtotal (95% CI)
5
0
83
18
17
1
85
19
90.6%
9.4%
0.26 [0.09 to 0.73]
0.33 [0.01 to 8.73]
101 104 100.0% 0.26 [0.10 to 0.71]
Heterogeneity: Tau2 = 0.00; Chi2 = 0.02, df = 1 (P = .88); I2 = 0%
Test for overall effect: z = 2.62 (P = .009)
Total events 5 18
Figure 3 –Forest plot of the OR comparing the complication of EBUS-TBNA and c-TBNA with or without ROSE. The OR of individual studies is
represented by a square through which runs a horizontal line (95% CI). The diamond with horizontal lines represents the pooled risk difference with
95% CI. See Figure 1 and 2legends for expansion of abbreviations.
chestjournal.org 7

sampling. There was no difference in the sample
adequacy rates with or without ROSE. Although the use
of ROSE was associated with significantly fewer number
of needle passes per patient during EBUS, its use did not
decrease the overall procedure time. The requirement
for additional procedures was lower with the use of
ROSE in both EBUS and conventional TBNA, which
translated into a significant reduction in the
complication rate when ROSE was used during c-TBNA
but not during EBUS-TBNA.
EBUS has revolutionized the diagnosis of mediastinal
lymphadenopathy, as it allows sampling of mediastinal
lymph nodes under real-time guidance.
19
Despite this
advantage, there is still a variation in the diagnostic
yield of EBUS-TBNA across different centers.
1,3
In
addition to the operator experience, the variation in yield
has been attributed to several factors, including the
etiology, lymph node size, number of lymph nodes
sampled, and others. One way to ensure uniformity in the
diagnostic yield would be to utilize ROSE, which provides
immediate assurance of an adequate specimen. This
approach was suggested in a few observational studies in
which the use of ROSE resulted in a higher diagnostic
yield.
8,20
The use of ROSE, however, was not shown to increase
the diagnostic yield of EBUS-TBNA in the present
analysis. The likely reason could be abolition of
confounding factors (a limitation of observational
studies) that could affect the diagnostic yield by
using a randomized study design. Another reason was
that the overall yield of EBUS-TBNA in most of the
studies included in this analysis was >85%, as the
studies involved high-volume centers. Moreover, at
least three needle passes were made in the group
without ROSE across studies. Previously, it has been
shown that the yield of EBUS-TBNA is maximal at
three needle passes.
21
Thus, the real benefitofROSE
would be reduction in the number of needle passes,
the procedure time, or the requirement of additional
procedures for making a diagnosis. The present
analysis found that the use of ROSE resulted in a
reduction in the number of needle passes during
EBUS-TBNA and, in one study, the number of
additional procedures required to make a diagnosis.
However, there was no effect on the duration of the
EBUS-TBNA procedure. This could be due to the time
required to process and review the slides, which might
have negated the time-saving benefits of ROSE.
5
Moreover, in staging a patient with lung cancer who
has a normal PET-CT scan of the mediastinum, ROSE
will increase the duration of the procedure due to slide
preparation and reading for several lymph node
stations.
Conversely, ROSE should affect the yield of c-TBNA
where the yield is extremely variable.
2
However,
this finding was also not seen in the present analysis.
The likely reason is again that all studies included in this
analysis were conducted at high-volume centers with
expert operators performing c-TBNA. This is supported
by almost similar rates of adequate samples
(preponderance of lymphocytes) with (78%-94%) and
without (86%-88%) ROSE.
6,7
These findings might have
abrogated the beneficial effect of ROSE in increasing the
procedural yield. However, in one study, the utilization
of ROSE during c-TBNA led to a decline in the
0
0.1
0.05
0.15
0.2
–1 –0.5
SE(RD) EBUS-TBNA
0 0.5
RD
1
0
0.1
0.05
0.15
0.2
–1 –0.5
SE(RD) c-TBNA
0 0.5
RD
1
Figure 4 –Funnel plot comparing the log risk difference vs the SE of RD. The circles represent the individual studies included in the meta-analysis.
The line in the center indicates the summary risk difference. RD ¼risk difference. See Figure 1 and 2legends for expansion of other abbreviations.
8Original Research [-#-CHEST -2017 ]

additional procedures required to make the diagnosis,
with an overall reduction in the complication rate.
5
What is the current role of ROSE? Whether ROSE
should be used in all TBNA procedures is debatable;
however, it should be used in high-risk patients such
as those with multiple comorbid illnesses because it
can reduce the need for additional bronchoscopic
procedures. Future trials should investigate the role
of ROSE in patients with sarcoidosis and those
requiring EBUS in the staging of lung cancer. In the
latter, once a N3 node is positive for malignant
cells, further sampling of other lymph nodes is not
required, thereby considerably reducing the procedure
time. In sarcoidosis, the yield of endosonography is
approximately 75%,
22,23
and ROSE might help in
reducing the need for additional bronchoscopic
procedures, especially bronchoscopic lung biopsy.
24
Our meta-analysis has a few limitations. Only a few
studies could be included in the present meta-analysis,
with most studies conducted at high-volume centers.
Thus, the results of this analysis cannot be generalized.
Whether the results of ROSE are concordant with the
final cytopathologic diagnosis could not be ascertained
from the present analysis. One study did suggest a high
concordance of ROSE with final cytology results.
7
The
studies included in the present analysis also did not
provide information on the extra costs that would be
incurred with the addition of ROSE during EBUS, which
remains an important consideration for starting this
facility. It is likely that by reducing the need for
additional procedures (including a repeat procedure)
and avoiding complications, the use of ROSE during
TBNA may be cost-effective. Finally, the results of this
meta-analysis are only pertinent to diagnosis and not to
molecular testing in lung cancer.
Conclusions
The addition of ROSE for mediastinal lymph node
sampling did not enhance the diagnostic yield or reduce
the procedure time of EBUS-TBNA and c-TBNA. There
was, however, a suggestion that utilization of ROSE may
decrease both the number of passes required per patient
and the need for additional bronchoscopic procedures in
making a final diagnosis.
Acknowledgments
Author contributions: R. A. serves as
guarantor of the paper and takes
responsibility for the integrity of the work as
a whole, from inception to published article.
I. S. S. and R. A. conceived the idea and was
responsible for the systematic review and
meta-analysis. I. S. S., S. D., R. A., and A. N.
A. drafted and revised the manuscript.
Financial/nonfinancial disclosures: None
declared.
References
1. Agarwal R, Srinivasan A, Aggarwal AN,
Gupta D. Efficacy and safety of convex
probe EBUS-TBNA in sarcoidosis: a
systematic review and meta-analysis.
Respir Med. 2012;106(6):883-892.
2. Agarwal R, Aggarwal AN, Gupta D.
Efficacy and safety of conventional
transbronchial needle aspiration in
sarcoidosis: a systematic review and
meta-analysis. Respir Care. 2013;58(4):
683-693.
3. Chandra S, Nehra M, Agarwal D,
Mohan A. Diagnostic accuracy of
endobronchial ultrasound-guided
transbronchial needle biopsy in
mediastinal lymphadenopathy: a
systematic review and meta-analysis.
Respir Care. 2012;57(3):384-391.
4. Trisolini R, Cancellieri A, Tinelli C,
et al. Randomized trial of endobronchial
ultrasound-guided transbronchial needle
aspiration with and without rapid on-
site evaluation for lung cancer
genotyping. Chest. 2015;148(6):
1430-1437.
5. Oki M, Saka H, Kitagawa C, et al.
Rapid on-site cytologic evaluation
during endobronchial ultrasound-
guided transbronchial needle aspiration
for diagnosing lung cancer: a
randomized study. Respiration.
2013;85(6):486-492.
6. Yarmus L, Van der Kloot T,
Lechtzin N, Napier M, Dressel D,
Feller-Kopman D. A randomized
prospective trial of the utility of rapid
on-site evaluation of transbronchial
needle aspirate specimens.
J Bronchology Interv Pulmonol.
2011;18(2):121-127.
7. Trisolini R, Cancellieri A, Tinelli C, et al.
Rapid on-site evaluation of transbronchial
aspirates in the diagnosis of hilar and
mediastinal adenopathy: a randomized
trial. Chest. 2011;139(2):395-401.
8. Guo H, Liu S, Guo J, et al. Rapid on-site
evaluation during endobronchial
ultrasound-guided transbronchial needle
aspiration for the diagnosis of hilar and
mediastinal lymphadenopathy in patients
with lung cancer. Cancer Lett.
2016;371(2):182-186.
9. Boyan W. On-site cytopathologic analysis
of bronchoscopic needle aspiration: con:
on-site analysis is not indicated.
J Bronchology Interv Pulmonol. 2003;10:
152-154.
10. Chin R. On-site cytopathologic analysis of
bronchoscopic needle aspiration: pro: on-
site analysis is indicated. J Bronchology
Interv Pulmonol. 2003;10:150-151.
11. Bruno P, Ricci A, Esposito MC, et al.
Efficacy and cost effectiveness of rapid
on site examination (ROSE) in
management of patients with
mediastinal lymphadenopathies. Eur
Rev Med Pharmacol Sci. 2013;17(11):
1517-1522.
12. Cardoso AV, Neves I, Magalhaes A,
Sucena M, Barroca H, Fernandes G. The
value of rapid on-site evaluation during
EBUS-TBNA. Rev Port Pneumol (2006).
2015;21(5):253-258.
13. Liberati A, Altman DG, Tetzlaff J, et al.
The PRISMA statement for reporting
systematic reviews and meta-analyses of
studies that evaluate health care
interventions: explanation and
elaboration. Ann Intern Med. 2009;151(4):
W65-W94.
14. Alderson P, Green S, Higgins JPT, eds.
Assessment of study quality. Cochrane
Reviewers’Handbook 4.2.2 [updated
March 2004]; Section 6. In: The Cochrane
Library, Issue 1, 2004. Chichester, UK:
John Wiley & Sons, Ltd. Accessed April
12, 2017.
15. Deeks JJ, Higgins JPT, Altman DG. A.
Analysing and presenting results. In:
Alderson P, Green S, Higgins JPT, eds.
Cochrane Reviewers’Handbook 422
[updated March 2004]. Chichester, UK:
John Wiley & Sons, Ltd; 2004:68-139.
chestjournal.org 9

16. Begg CB, Mazumdar M. Operating
characteristics of a rank correlation test
for publication bias. Biometrics.
1994;50(4):1088-1101.
17. Egger M, Davey Smith G, Schneider M,
Minder C. Bias in meta-analysis detected
by a simple, graphical test. BMJ.
1997;315(7109):629-634.
18. Madan K, Dhungana A, Mohan A, et al.
Conventional transbronchial needle
aspiration versus endobronchial
ultrasound-guided transbronchial needle
aspiration, with or without rapid on-site
evaluation, for the diagnosis of
sarcoidosis: a randomized controlled trial.
J Bronchology Interv Pulmonol.
2017;24(1):48-58.
19. Dhooria S, Sehgal IS, Aggarwal AN,
Agarwal R. Convex-probe endobronchial
ultrasound: a decade of progress. Indian J
Chest Dis Allied Sci. 2016;58(1):21-35.
20. Sørhaug S, Hjelde H, Hatlen P, et al.
Learning EBUS-TBNA—a 6-year
experience at a single institution
[published online ahead of print March 8,
2016]. Clin Respir J. https://doi.org/1
0.1111/crj.12475.
21. Lee HS, Lee GK, Lee HS, et al. Real-time
endobronchial ultrasound-guided
transbronchial needle aspiration in
mediastinal staging of non-small cell lung
cancer: how many aspirations per target
lymph node station? Chest. 2008;134(2):
368-374.
22. von Bartheld MB, Dekkers OM,
Szlubowski A, et al. Endosonography
vs conventional bronchoscopy for
the diagnosis of sarcoidosis:
the GRANULOMA randomized
clinical trial. JAMA. 2013;309(23):
2457-2464.
23. Gupta D, Dadhwal DS, Agarwal R,
Gupta N, Bal A, Aggarwal AN.
Endobronchial ultrasound-guided
transbronchial needle aspiration
vs conventional transbronchial needle
aspiration in the diagnosis of sarcoidosis.
Chest. 2014;146(3):547-556.
24. Kumar S, Chandra S. A “ROSE”in every
“EBUS”keeps transbronchial lung biopsy
away. Chest. 2014;146(3):e97.
10 Original Research [-#-CHEST -2017 ]