Gynecologic Cytopathology

Acta Cytologica Received: March 20, 2020

Accepted: April 18, 2020
DOI: 10.1159/000508018 Published online: June 9, 2020

Comparison between Conventional Cytology and

Liquid-Based Cytology in the Tertiary Brazilian
Navy Hospital in Rio de Janeiro
Antônio Carlos Almeida de Oliveira a, b Miguel Fontes Domingues b
   

Paulo Murilo Neufeld c Marcos Fleury c José Firmino Nogueira Neto d

     

aLaboratory
Medicine and Forensic Technology, Rio de Janeiro State University (UERJ), Rio de Janeiro, Brazil;
bHospital
Naval Marcílio Dias, Pathology Service, Rio de Janeiro, Brazil; cFaculdade de Farmácia, Universidade
Federal do Rio de Janeiro, Rio de Janeiro, Brazil; dLipids Laboratory-LabLip, Faculty of Medical Sciences, Rio de
Janeiro State University (UERJ), Rio de Janeiro, Brazil

Keywords sensitive than conventional cytology (5.28%) and ThinPrep®

Cervical cancer screening · Liquid-based cytology · (3.73%). Conclusions: The results show the advantage of im-
Conventional cytology plementing LBC in routine screening for cervical lesions. In
this study, BD SurePath® achieved the overall best perfor-
mance considering the studied variables.
Abstract © 2020 S. Karger AG, Basel
Introduction: Cervical cancer screening is an important tool
in public health. Liquid-based cytology (LBC) has been per-
formed at the studied hospital for 7 years. The present study Introduction
compares the performance of 2 LBC techniques with con-
ventional cytology. Objective: Our objective is to verify the Cervical cancer is the third most common cancer type
sensitivity for the detection of neoplastic and preneoplastic amongst women in Brazil. In 2017, 6,385 women died
epithelial atypia, as well as the positive predictive value of from this disease. The crude mortality rate was 6.17 per
the 3 methodologies. Methods: We analyzed retrospectively 100 thousand. For 2020, 16,590 new cases of cervical can-
24,529 cases and evaluated the conventional cytology, Thin- cer are estimated, corresponding to 7.4% of new cancer
Prep®, and BD SurePath® performance categorizing the re- cases [1]. To prevent cervical cancer, cytology screening
sults according to the Bethesda system. We also compared is a remarkable tool, especially in non-high-income coun-
the level of unsatisfactory samples, the presence of elements tries.
from the squamocolumnar junction, and the detection of In the 1920s, Babes [2] and Papanicolaou [3] were the
pathogenic microorganisms. Results: ThinPrep® (1.43%) precursors of cancer screening using exfoliative cytologi-
showed superior sensitivity over BD SurePath® (0.91%) and cal techniques. The development of these techniques al-
conventional cytology (0.71%) in terms of the detection of lowed for early diagnosis and the expansion of cervical
high-grade lesions; however, in terms of squamous atypia as cancer screening. The pioneering spirit of these research-
a whole (ASC-US+), BD SurePath® (6.44%) proved to be more ers resulted in an evolutionary leap in public health. Pre-
177.79.118.4 - 6/10/2020 1:02:13 PM
International Academy of Cytology

[email protected] © 2020 S. Karger AG, Basel Antônio Carlos Almeida de Oliveira

www.karger.com/acy Laboratory Medicine and Forensic Technology, Rio de Janeiro State University
Rua César Zama, 185
Rio de Janeiro 20750-090 (Brazil)
Downloaded by:

antonio-carlos.almeida @ marinha.mil.br
vention and early diagnosis are essential to reduce can- creased from 3% (conventional cytology) to 5.7% (LBC).
cer-related morbidity and mortality. In 2009, Beerman et al. [10] analyzed 86,469 gynecologi-
Cervical cancer screening has been primarily carried cal cytology records from a database, in which 51,154
out by conventional cytology. Several institutions are used conventional cytology and 35,315 used BD Sure-
continually evaluating the results obtained after the im- Path®. They observed a reduction in unsatisfactory sam-
plementation of liquid-based cytology (LBC). Tao et al. ples with the use of LBC (0.30% for LBC vs. 0.89% for
[4] compared 1,248,785 results from patients from an ac- conventional cytology). They also noted a higher detec-
ademic hospital in China between 2009 and 2014. The tion of cellular atypia when using LBC (2.97% for LBC vs.
samples were processed using LBC, either BD SurePath® 1.64% for conventional cytology).
or ThinPrep®, and conventional cytology. They observed Simonsen et al. [11] studied 2 devices (Cervex-Brush®
a significant increase in the detection of all categories Combi vs. cervical brush + Ayre’s spatula) for cervical
considered abnormal after the establishment of LBC. harvesting in LBC, aiming to evaluate their effectiveness
Also, one of the topics evaluated was the adequacy of the in obtaining endocervical cells, as well as the accuracy for
sample. Conventional cytology showed a higher occur- detection of NIC 2+. They observed that Cervex-Brush®
rence of unsatisfactory cases than LBC. In Turkey, Budak Combi was superior to the cervical brush + Ayre’s spatu-
et al. [5] published a retrospective study to compare the la for endocervical sampling (82.7 vs. 74.6%) and more
results of cytological exams from 47,954 patients between sensitive (48.6 vs. 33.9%) for detecting NIC 2+ lesions.
2008 and 2014, performed by conventional cytology and Kituncharoen et al. [12] compared the rates of unsat-
LBC. There was no significant difference regarding the isfactory samples and the detection of epithelial abnor-
occurrence of abnormalities. malities amongst 23,030 cervical samples collected by
Jeong et al. [6] studied 38,956 gynecologic cytopathol- conventional cytology and BD SurePath®. They conclud-
ogy reports using conventional cytology and LBC (from ed that there was no difference in the occurrence of un-
3 different manufacturers). It only addressed the inci- satisfactory samples (0.1 vs. 0.1%). The detection of high-
dence of unsatisfactory samples. The study concluded grade and glandular squamous abnormalities was also the
that LBC resulted in 1.26% of unsatisfactory samples, same. However, they observed that the detection of low-
whereas conventional cytology resulted in 3.31%. In 2016, grade abnormalities in squamous cells was significantly
Rozemeijer et al. [7] retrospectively studied more than 6 greater with the LBC (7.7 vs. 11.5%). In 2019, Gupta et al.
million cytological reports on conventional cytology and [13] compared the performance of a low-cost LBC tech-
2 LBC (BD SurePath® and ThinPrep®). BD SurePath® nique (EziPREP®) with conventional cytology. It was a
showed a 12% increase in the detection of epithelial atyp- cross-sectional split-sample study conducted on 515
ia (moderate dysplasia or more) compared with conven- women. The rate of unsatisfactory samples favored con-
tional cytology. ThinPrep® did not show any advantage ventional cytology (1 vs. 1.3%), but the differences in the
compared with conventional cytology in terms of sensi- detection of epithelial abnormalities and pathogenic mi-
tivity. The incidence of unsatisfactory samples or the de- croorganisms were not statistically significant, nor was
tection of pathogenic microorganisms was not addressed. the cytohistological correlation (96% for both tech-
In Brazil, Etlinger-Colonelli et al. [8] analyzed 41,264 niques).
gynecological samples comparing conventional cytology Pun et al. [14] conducted a comparative study between
and BD SurePath®. The increase in the detection of ab- conventional cytology (1,180 samples) and LBC (1,160
normal results rose from 7.80 to 11.57% when using LBC. samples). The percentage of unsatisfactory samples was
The incidence of low-grade lesions (LSIL) was the main lower in the LBC group (1.2 vs. 3.9%), and there was an
reason for the increase in sensitivity. As for the incidence increase in the detection of epithelial atypia (ASC-US+)
of unsatisfactory samples, it was decreased from 3.50 to when LBC was implemented (3.77 vs. 2.71%). Kaza et al.
0.25%. Longatto-Filho et al. [9] analyzed 218,594 conven- [15] compared conventional cytology and BD SurePath®
tional cytology and BD SurePath® samples. When using in terms of their sensitivities for detecting pathogenic mi-
conventional cytology, 3% of the samples were unsatis- croorganisms. Their findings suggested that both meth-
factory for analysis, whereas when using BD SurePath®, ods are equally valuable for this purpose.
the rate was only 0.3%. As for the detection of high-grade The main purpose of this study is to compare the sen-
lesions (HSIL), they observed the same percentage for sitivity for the detection of neoplastic and preneoplastic
both methods; however, the LSIL percentage was higher epithelial atypia, as well as the positive predictive value of
for LBC (2.2 vs. 0.7%). The index of positive results in- 3 methodologies, conventional cytology and LBC (BD
177.79.118.4 - 6/10/2020 1:02:13 PM
International Academy of Cytology

2 Acta Cytologica de Oliveira/Domingues/Neufeld/Fleury/

DOI: 10.1159/000508018 Nogueira Neto
Downloaded by:
SurePath® and ThinPrep®), in a tertiary hospital in Rio For cases in which there was a violation between observed and
de Janeiro, Brazil. We also assessed the samples’ satisfac- expected values ≥5, we applied Fisher’s exact test. Table 1 presents
statistical analysis.
tory rate, the presence of the elements of the squamoco-
lumnar junction (SCJ), and the detection of pathogenic
microorganisms.
Results

Materials and Methods Between January 2007 and December 2018, during an
11-year study period, 24,529 cytology reports were re-
Samples viewed and categorized, according to TBS. Table 2 pres-
This is a retrospective study where the cytology samples and ents all the reporting results for the 3 techniques we em-
reports were obtained from the Pathology Service in the tertiary ployed. We also assessed the rate of unsatisfactory sam-
care center of the Brazilian Navy Hospital, Hospital Naval Marcílio
Dias (HNMD), Rio de Janeiro. This study obtained Institutional ples and the sampling of the SCJ.
Ethics Committee clearance before the commencement of the re- Interestingly, when we examined the specimen ade-
search. The reference numbers are CONEP/Plataforma Brasil, quacy, the percentage of unsatisfactory samples was
CAAE 06468819.2.000.5256, and authorization 3.208.007, Sep- smaller for conventional cytology (0.20%) when com-
tember 9, 2019. pared with BD SurePath® (1.08%). Nonetheless, there
The reports/samples were divided into 3 groups. The first
group consisted of 10,742 samples prepared by conventional cytol- was no statistical difference when comparing the 2 LBC,
ogy from January 2007 to December 2016. The second group stud- ThinPrep® (0.56%) and BD SurePath® (1.08%). The rate
ied was represented by 1,258 samples from January 2012 to July of unsatisfactory samples we observed for conventional
2013. They were processed using LBC ThinPrep® 2000 Hologic cytology is significantly smaller, insofar as reflecting the
(Marlborough, MA, USA), and Ayre’s spatula, endocervical sam- continuing training of the gynecology staff of HNMD for
pling brush, and the PreservCyt® Solution were employed for har-
vesting the samples. In the third group, we analyzed 12,529 reports a very long time. There is a residency medical program
of gynecological cytology using LBC BD SurePath® from August and a fellowship program in HNMD. We acknowledge
2013 to December 2018. For harvesting the samples, BD Sure- that this rate might not be reproducible in all cases. Con-
Path® Vial and Brush/Spatula Collection Kit were used. tinuous education and quality management programs for
The 24,529 cytology results were plotted in a table and catego- sample takers may not be a reality for some institutions
rized according to the Bethesda system (TBS) [16]. The following
variables were considered: negative for malignancy, ASC-US, and hospitals. Therefore, we believe this is the key point
LSIL, ASC-H, HSIL, squamous cell carcinoma, AGC, adenocarci- why our unsatisfactory rate for conventional cytology was
noma, and other malignant neoplasms. The presence of the follow- considerably small. Occasionally, the sample taker might
ing organisms was also evaluated: Trichomonas vaginalis, fungal not, for various reasons, accurately access the SCJ or even
organisms morphologically consistent with Candida spp., the shift the cervix itself. It is a critical step in terms of the institu-
in flora suggestive of bacterial vaginosis, bacteria morphologically
consistent with Actinomyces spp., and cellular changes consistent tions’ quality management program as a whole. The sam-
with herpes simplex virus. We verified the sampling of the SCJ (the ple taker skills should not be taken for granted and might
presence of endocervical cells and metaplastic cells) and the rate of be assessed periodically.
unsatisfactory samples. In our daily routine, we observed that the main causes
Whenever necessary, the patients were referred to colposcopy of unsatisfactory samples for conventional cytology are
and biopsied, according to TBS and the Brazilian guidelines for
cervical cancer screening [17]. Therefore, we correlated the cytol- obscuring blood/white blood cells, interfering substanc-
ogy results of the patients that were biopsied with their histopath- es, and cellularity excess. The percentage of unsatisfac-
ological result counterpart. The aim was to compare the positive tory samples we observed for LBC is comparable to the
predictive value for each of the cytology techniques (conventional literature [6, 13, 14]. They were higher when comparing
cytology, ThinPrep®, and BD SurePath®). The data were obtained to LBC, possibly to the adaptation of the gynecology staff
through direct search in registry books, Excel® sheets, medical re-
cords, cytology results plotted in the laboratory software Com- to the new technique when it was implemented. As for the
plab®, and printed medical files. LBC, we noticed that scanty squamous cellularity was the
main reason for rejection.
Statistics The next evaluated point was the presence of organ-
We used Pearson’s χ2 test to check the existence of a significant isms in the samples. We perceived no statistical difference
difference between the techniques studied (conventional cytology,
LBC ThinPrep®, and BD SurePath®) as to the proportion of oc- between conventional cytology (7.52%) and ThinPrep®
currence of the test results. In cases where the difference was sig- (5.80%), notwithstanding, BD SurePath® showed consid-
nificant, we performed a post hoc test with Bonferroni correction. erable superiority (16.94%).
177.79.118.4 - 6/10/2020 1:02:13 PM
International Academy of Cytology

Comparison between Conventional Acta Cytologica 3

Cytology and Liquid-Based Cytology DOI: 10.1159/000508018
Downloaded by:
Table 1. Statistical analysis – TBS results

Interpretation/result Multiple comparison Methodology Post hoc

NILM <0.0001 Conventional × Thinprep® 0.2700

Conventional × BD Surepath® <0.0001
Thinprep® × BD Surepath® <0.0001
ASC-US <0.0001 Conventional × Thinprep® 0.0042
Conventional × BD Surepath® 0.0042
Thinprep® × BD Surepath® <0.0001
LSIL 0.0985 Conventional × Thinprep® Not calculated
Conventional × BD Surepath® Not calculated
Thinprep® × BD Surepath® Not calculated
ASC-H 0.0002a Conventional × Thinprep® 0.0002
Conventional × BD Surepath® 0.1126
Thinprep® × BD Surepath® 0.0035
HSIL 0.0385a Conventional × Thinprep® 0.7772a
Conventional × BD Surepath® 0.0113a
Thinprep® × BD Surepath® 0.5201a
Carcinoma 0.0047a Conventional × Thinprep® 1.0000a
Conventional × BD Surepath® 0.0016a
Thinprep® × BD Surepath® 0.3810a
AGC (either NOS or favor neoplastic) 0.4678a Conventional × Thinprep® 0.5810a
Conventional × BD Surepath® 0.5033a
Thinprep® × BD Surepath® 0.3186a
Adenocarcinoma 0.0102a Conventional × Thinprep® 0.2466a
Conventional × BD Surepath® 0.0044a
Thinprep® × BD Surepath® 1.0000a
Other malignant neoplasms 0.1478a Conventional × Thinprep® Not calculated
Conventional × BD Surepath® Not calculated
Thinprep® × BD Surepath® Not calculated
Unsatisfactory <0.0001 Conventional × Thinprep® 0.0271a
Conventional × BD Surepath® <0.0001
Thinprep® × BD Surepath® 0.1083
Organisms <0.0001 Conventional × Thinprep® 0.0940
Conventional × BD Surepath® <0.0001
Thinprep® × BD Surepath® <0.0001
SCJ sampling <0.0001 Conventional × Thinprep® <0.0001
Conventional × BD Surepath® <0.0001
Thinprep® × BD Surepath® <0.0001

Bold: statistically significant difference. ASC-US, epithelial atypia; LSIL, low-grade lesions; HSIL, high-grade lesions. a Fisher’s exact
test.

Cervical intraepithelial neoplasia most commonly oc- Path® (67.08%), followed by ThinPrep® (50.40%) and
curs at the SCJ, a transitional zone. Sometimes, it is a chal- conventional cytology (39.53%). These results endorse
lenge to obtain an adequate sample. When we analyzed the better sensitivity for detecting atypia when using BD
the sampling of the SCJ, meaning the presence of well- SurePath®.
preserved endocervical and squamous metaplastic cells, From the total of 24,529 cytology reports, we recog-
the technique that showed the best result was BD Sure- nized 1,419 results considered atypical/abnormal, mean-
177.79.118.4 - 6/10/2020 1:02:13 PM
International Academy of Cytology

4 Acta Cytologica de Oliveira/Domingues/Neufeld/Fleury/

DOI: 10.1159/000508018 Nogueira Neto
Downloaded by:
 Table 2. Number of Pap tests from 2007 to 2018 according to results (TBS) using conventional cytology, BD
SurePath®, and ThinPrep®

Conventional cytology BD SurePath® ThinPrep®

(10,742), n (%) (12,529), n (%) (1,258), n (%)

Interpretation/result
NILM 10,154 (94.52) 11,588 (92.49) 1,204 (95.71)
ASC-US 295 (2.75) 436 (3.48) 15 (1.19)
LSIL 143 (1.33) 197 (1.57) 12 (0.95)
ASC-H 29 (0.27) 49 (0.39) 13 (1.03)
HSIL 30 (0.28) 61 (0.49) 4 (0.32)
Carcinoma 17 (0.16) 4 (0.03) 1 (0.08)
AGC (either NOS or favor neoplastic) 34 (0.32) 47 (0.38) 2 (0.16)
Adenocarcinoma 17 (0.16) 5 (0.04) 0 (0.00)
Other malignant neoplasms 1 (0.01) 7 (0.06) 0 (0.00)
Unsatisfactory 22 (0.20) 135 (1.08) 7 (0.56)
Organisms 808 (7.52) 2,122 (16.94) 73 (5.80)
SCJ sampling 4,246 (39.53) 8,405 (67.08) 634 (50.40)

ASC-US, epithelial atypia; LSIL, low-grade lesions; HSIL, high-grade lesions.

Table 3. Grouped atypical/abnormal results

ASC-US LSIL, ASC-H HSIL AGC (either NOS or Other malignant Total, n (%)
n (%) carcinoma, n (%) favor neoplastic) neoplasms, n (%)
adenocarcinoma, n (%)

Conventional cytology 438 (4.08) 76 (0.71) 51 (0.47) 1 (0.01) 566 (5.28)

BD SurePath® 633 (5.05) 114 (0.91) 52 (0.42) 7 (0.06) 806 (6.44)
ThinPrep® 27 (2.15) 18 (1.43) 2 (0.16) 0 (0.00) 47 (3.73)
Total 1,098 (11.27) 208 (3.05) 105 (1.06) 8 (0.07) 1,419 (15.45)

ASC-US, epithelial atypia; LSIL, low-grade lesions; HSIL, high-grade lesions.

ing ASC-US or higher (Table 3). They were categorized conventional cytology (0.71%). We did not observe any
into 4 different groups: ASC-US and LSIL; ASC-H, HSIL, statistical difference between ThinPrep® versus BD Sure-
and carcinoma; AGC (either NOS or favor neoplastic) Path®, both techniques proved to be superior to conven-
and adenocarcinoma, and other malignant neoplasms. tional cytology.
According to the protocols adopted in the studied hos- We assume the higher percentage of the ThinPrep®
pital, the patients are followed after 6 months whenever atypical results (1.43%) resemble false-positive results
they have results of ASC-US or LSIL. We observed the when compared with conventional cytology and BD
following rates of detection for this group: BD SurePath® SurePath®. This is critical when choosing the ideal tech-
(5.05%), followed by conventional cytology (4.08%) and nique, but it should be correlated to the positive predic-
ThinPrep® (2.15%). tive value (PPV).
According to TBS [16] and the Brazilian guidelines for In our study, 321 patients were reported with a result
cervical cancer screening [17, 18], whenever the patients of ASC-H or higher. They were referred to colposcopy
have a result of ASC-H or higher, they are referred to col- and biopsy, but only 200 patients returned to the hospital
poscopy and biopsy. Hence, in the next group of results, and were effectively biopsied. Accordingly, we compared
we investigated ASC-H, HSIL, and squamous cell carci- their cytology reports with their histopathological result
noma, and we verified the following results for each tech- counterpart (Table 4). Then, we assessed the PPV for each
nique: ThinPrep® (1.43%), BD SurePath® (0.91%), and of the techniques. We observed the following PPV: Thin-
177.79.118.4 - 6/10/2020 1:02:13 PM
International Academy of Cytology

Comparison between Conventional Acta Cytologica 5

Cytology and Liquid-Based Cytology DOI: 10.1159/000508018
Downloaded by:
 Table 4. PPV and correlation between cytology and histology

Patients biopsied, Correlated histology × Discrepant correlation PPV, %

n (%) cytology, n (%) histology × cytology, n (%)

Conventional cytology 93 (0.87) 74 (0.69) 19 (0.18) 79.6

BD SurePath® 86 (0.69) 68 (0.54) 17 (0.14) 79.1
ThinPrep® 21 (1.67) 13 (1.03) 8 (0.64) 61.9

PPV, positive predictive value.

Table 5. Statistical analysis for grouped atypia

Grouped atypia Technique Frequency, p value Post hoc Bonferroni

% (p value)

ASC-US, LSIL Conventional cytology 4.08 <0.0001 Conventional × ThinPrep® 0.0031

BD SurePath® 5.05 Conventional × BD SurePath® 0.0012
ThinPrep® 2.15 ThinPrep® × BD SurePath® <0.0001
ASC-H, HSIL, carcinoma Conventional cytology 0.71 0.0007 Conventional × ThinPrep® 0.004
BD SurePath® 0.91 Conventional × BD SurePath® 0.012
ThinPrep® 1.43 ThinPrep® × BD SurePath® 0.393
AGC (either NOS or favor Conventional cytology 0.48 0.2560 Conventional × ThinPrep® Not calculated
neoplastic), adenocarcinoma BD SurePath® 0.42 Conventional × BD SurePath® Not calculated
ThinPrep® 0.16 ThinPrep® × BD SurePath® Not calculated
Epithelial total atypia Conventional cytology 5.28 <0.0001 Conventional × ThinPrep® 0.0908
BD SurePath® 6.44 Conventional × BD SurePath® 0.0001
ThinPrep® 3.73 ThinPrep® × BD SurePath® 0.0005

Bold: statistically significant difference. ASC-US, epithelial atypia; LSIL, low-grade lesions; HSIL, high-grade lesions.

Prep® (61.9%), conventional cytology (79.6%), and BD fixation, and preparation of slides. The fixation process
SurePath® (79.1%). consists of collecting cervical cells with a brush and dis-
As we summarized our data considering the presence persing the material in a preservative liquid right on site,
of atypical/abnormal cells, whether squamous or glandu- for further processing and assembly of the slides in the
lar, we verified that BD SurePath® detected more of these laboratory, using an automated and controlled method-
cells (6.44%) in comparison with conventional cytology ology. It was a milestone considering advances in the fol-
(5.28) and ThinPrep® (3.73%) (Table 5). These results as- lowing aspects:
sociated with the PPV indicate that, in our study, the most 1. Higher potential for detecting epithelial lesions
sensitive technique for cervical cancer screening is BD 2. Diminishing interferences (such as red blood cells and
SurePath®. We did not observe any statistical difference leukocytes) that obscure the slides in conventional cy-
between the results observed when using ThinPrep® and tology
conventional cytology. 3. Better distribution of epithelial cells in the preservative
liquid medium in the slide, thereby reducing the over-
lap of cells
Discussion/Conclusion 4. Better sampling of the cells onto the slide as the entire
collection of cells is transferred to the flask with the
In 2012, the Pathology Service of Hospital Naval Mar- preservative medium
cílio Dias implemented LBC, an extraordinary technique 5. Cells are better preserved, therefore better dyed and
promoting a breakthrough in cervical harvesting, sample more likely to be morphologically examined
177.79.118.4 - 6/10/2020 1:02:13 PM
International Academy of Cytology

6 Acta Cytologica de Oliveira/Domingues/Neufeld/Fleury/

DOI: 10.1159/000508018 Nogueira Neto
Downloaded by:
6. Shorter slide scrutiny time with the others. This assumption might corroborate to
7. Use of the residual sample for molecular tests for HPV the better sensitivity observed.
detection Ultimately, BD SurePath® detected more than twice
Implementing HPV testing and LBC is still a chal- the presence of organisms compared with conventional
lenge to non-high-income countries. Besides the clear re- cytology, but we did not observe any advantage in this
lationship between HPV and cervical cancer, imple- scope when analyzing the results from ThinPrep®. This
menting both methodologies in these countries is part of situation surprised us, considering that some studies do
a complex strategy of public health in the context of not show an advantage when comparing LBC with con-
screening programs. The World Health Organization ventional cytology. It is an important observation, espe-
(WHO) considers that better screening tests may result cially in non-high-income countries, where cervical can-
in more effective programs [19]. In public health, it cer screening is still not as effective as it should be. In
should be affordable considering the financial invest- these countries, women usually seek gynecologists when
ment, the training of personnel involved, and easy access they have complaints related to infectious processes. In
for patients to the screening and treatment [20]. The Brazil, the price of LBC is still a barrier when we think of
global outcome of this investment may overcome the implementing this technology in our Public Health Sys-
burden of the incidence and mortality of cervical cancer, tem; on the other hand, it is more important to evaluate
as soon as these technologies become more and more af- the cost benefits of how this technology can improve cer-
fordable. vical cancer screening in our society.
When the hospital decided to implement LBC at the
Pathology Service, we chose 2 different technologies. Ini-
tially, we used ThinPrep® and then we switched to BD Acknowledgement
SurePath®. This study compared the performance be- The authors wish to thank Giuliana Vasconcelos de Souza Fon-
tween these 2 LBC and conventional cytology, searching seca for statistical advice and Dr. Cesar de Sousa Bastos Júnior and
for answers that could corroborate to better results pro- Dr. Ana Lúcia Guimarães Arêas from the Pathology Service of
posed by LBC. HNMD.
Both ThinPrep® and BD SurePath® showed unsatis-
factory rates compared with other studies. However,
these results surprisingly did not excel conventional cy- Statement of Ethics
tology since it already had a low occurrence of unsatisfac-
The study obtained Institutional Ethics Committee clearance
tory samples in the studied institution (0.2%). before the commencement of the research, assuring the ethical and
According to the literature, LBC detected more low- legal aspects of the study. The reference numbers are CONEP/Plat-
grade lesions (ASC-US + LSIL) compared with conven- aforma Brasil, CAAE 06468819.2.000.5256, and authorization
tional cytology; notwithstanding, in our study, we ob- 3.208.007, September 9, 2019.
served this result only with BD SurePath®. In contrast,
ThinPrep® distinguished more grouped high-grade le-
sions (ASC-H + HSIL + carcinoma) compared with BD Disclosure Statement
SurePath® and conventional cytology. Despite this re- The authors have no conflicts of interest to declare.
gard, when we analyzed the PPV, ThinPrep® presented
more false-positive cases since its PPV was lower than BD
SurePath® and conventional cytology. We did not ob- Funding Sources
serve any difference when detecting glandular atypia.
Considering that the noblest purpose of cervical screen- The funding for this research was provided by the Pathology
ing is the detection of neoplastic and preneoplastic le- Service of the Brazilian Navy Hospital, Hospital Naval Marcílio
sions, we conclude the superiority of LBC over conven- Dias (HNMD), Rio de Janeiro.
tional cytology as a screening method. Embodied in this
conclusion is the advantage of BD SurePath® regarding
the sensitivity for detecting epithelial atypia as a whole Author Contributions
(ASC-US+ and glandular atypia) and ThinPrep® for the Antônio Carlos Almeida de Oliveira: screening/analyzing all
detection of high-grade lesions (ASC-H+). BD SurePath® the samples studied and substantial contribution to the concep-
was the technique that better sampled the SCJ compared tion of the work and acquisition, analysis and interpretation of
177.79.118.4 - 6/10/2020 1:02:13 PM
International Academy of Cytology

Comparison between Conventional Acta Cytologica 7

Cytology and Liquid-Based Cytology DOI: 10.1159/000508018
Downloaded by:
data for the work, and writing of the draft and final version of of project definition and article evaluation. Marcos Fleury: guid-
the article. Miguel Fontes Domingues: screening/analyzing all ance in the initial phase of project definition and article evalua-
the samples studied and substantial contribution to the concep- tion. José Firmino Nogueira Neto: conception of the study from
tion of the work and acquisition, analysis and interpretation of the draft, planning, monitoring the execution, discussion of sta-
data for the work, and writing of the draft and final version of tistical analysis, and contribution to the final version of the ar-
the article. Paulo Murilo Neufeld: guidance in the initial phase ticle.

References
 1 Instituto Nacional de Câncer José Alencar   9 Longatto-Filho A, Levi JE, Martins TR, Cohen 15 Kaza S, Renuka IV, Kantham L, Srinath S,
Gomes da Silva. Estimate/2020: Cancer Inci- D, Cury L, Villa LL, et al. Critical analyses of Babu R. Comparative study of conventional
dence in Brazil. Rio de Janeiro: INCA; 2019. the introduction of liquid-based cytology in a papanicolaou smears and liquid-based direct-
  2 Babes A. Diagnostic du cancer du col utérin public health service of the state of São Paulo, to-vial thin-layer preparation in the detection
par les frottis. Presse Méd. 1928;36:451–4. Brazil. Acta Cytol. 2015;59(3):273–7. of microorganisms in cervical smears. IP
 3 Papanicolaou GN. New cancer diagnosis. 10 Beerman H, van Dorst EB, Kuenen-Bou- Arch Cytol Histopathol Res. 2018;3(2):69–75.
Proceedings of the 3rd Race Betterment Con- meester V, Hogendoorn PC. Superior perfor- 16 Nayar R, Wilbur D, editors. The Bethesda
ference. Battle Creek: RBF; 1928. p. 528–34. mance of liquid-based versus conventional System for Reporting Cervical Cytology. 3rd
  4 Tao X, Austin RM, Zhang H, Zhang L, Xiao J, cytology in a population-based cervical can- ed. Cham: Springer; 2014.
Wang L, et al. Pap test reporting rates for con- cer screening program. Gynecol Oncol. 2009; 17 Diretrizes brasileiras para o rastreamento do
ventional smear and liquid-based cervical cy- 112(3):572–6. câncer do colo do útero/Instituto Nacional de
tology from the largest academic women’s 11 Simonsen M, Tavares Guerreiro Fregnani JH, Câncer José Alencar Gomes da Silva. Coorde-
hospital in China: analysis of 1,248,785 pap Possati Resende JC, Antoniazzi M, Longatto- nação de Prevenção e Vigilância. Divisão de
test reports. Acta Cytol. 2015;59(6):445–51. Filho A, Scapulatempo-Neto C. Comparison Detecção Precoce e Apoio à Organização de
 5 Budak MŞ, Senturk MB, Kaya C, Akgol S, of the Cervex-Brush® combi and the cyto- Rede. – 2. ed. rev. atual. Rio de Janeiro: INCA;
Bademkiran MH, Tahaoğlu AE, et al. A com- brush + ayres spatula combination for cervi- 2016.
parative study of conventional and liquid- cal sampling in liquid-based cytology. PLoS 18 Instituto Nacional de Câncer José Alencar
based cervical cytology. Ginekol Pol. 2016; One. 2016;11(10):e0164077. Gomes da Silva. Coordenação-Geral de Pre-
87(3):190–3. 12 Kituncharoen S, Tantbirojn P, Niruthisard S. venção e Vigilância. Divisão de Detecção Pre-
  6 Jeong H, Hong SR, Chae SW, Jin SY, Yoon Comparison of unsatisfactory rates and de- coce e Apoio à Organização de Rede. Nomen-
HK, Lee J, et al. Comparison of unsatisfactory tection of abnormal cervical cytology between clatura brasileira para laudos citopatológicos
samples from conventional smear versus liq- conventional papanicolaou smear and liquid- cervicais/Instituto Nacional deCâncer José
uid-based cytology in uterine cervical cancer based cytology (Sure Path®). Asian Pac J Can- Alencar Gomes da Silva, Coordenação-Geral
screening test. J Pathol Transl Med. 2017; cer Prev. 2015;16(18):8491–4. de Prevenção e Vigilância,Divisão de De-
51(3):314–9. 13 Gupta R, Yadav R, Sharda A, Kumar D, Sand- tecção Precoce e Apoio à Organização de
  7 Rozemeijer K, Penning C, Siebers AG, Naber eep D, Mehrotra R, et al. Comparative evalu- Rede. – 3. ed. Rio de Janeiro: INCA; 2012.
SK, Matthijsse SM, van Ballegooijen M, et al. ation of conventional cytology and a low-cost 19 Cervical cancer screening in developing
Comparing SurePath, ThinPrep, and conven- liquid-based cytology technique, EziPREP™, countries: a report of a WHO consultation.
tional cytology as primary test method: Sure- for cervicovaginal smear reporting: a split WHO Library Cataloguing-in-Publication
Path is associated with increased CIN II+ de- sample study. Cytojournal. 2019;16:22. Data. World Health Organization; 2002 [In-
tection rates. Cancer Causes Control. 2016; 14 Pun C, Shrestha S, Bhatta R, Pandey G, Up- ternet]. Available from: https://apps.who.int/
27(1):15–25. rety S, Bastakoti S, et al. Comparative analysis iris/handle/10665/42544.
  8 Etlinger-Colonelli D, Oliveira MP, Basso MC, between liquid based cytology and conven- 20 Comprehensive cervical cancer control: a
Iglezias SA, Yamamoto LSU, Sakai YI, et al. tional Pap smears at B P Koirala Memorial guide to essential practice – 2nd ed. WHO Li-
Comparison of performance of liquid-based Cancer Hospital in Bharatpur, Chitwan, Ne- brary Cataloguing-in-Publication Data.
versus conventional cytology in Brazilian pal. Nep J Cancer. 2018;2(1):68–70. World Health Organization; 2014 [Internet].
public health system of Vale do Ribeira from Available from: https://www.who.int/repro-
2009 to 2012. Secretaria Estadual de Saúde de ductivehealth/publications/cancers/cervical-
São Paulo, 2014 [Internet]. [cited 2017 Sep 4]. cancer-guide/en/.
Available from: http://ses.sp.bvs.br/lildbi/
docsonline/get.php?id=5323.

177.79.118.4 - 6/10/2020 1:02:13 PM

International Academy of Cytology

8 Acta Cytologica de Oliveira/Domingues/Neufeld/Fleury/

DOI: 10.1159/000508018 Nogueira Neto
Downloaded by: