Gynecologic Oncology 106 (2007) 170 – 176

www.elsevier.com/locate/ygyno

Human papillomavirus (HPV) test and PAP smear as predictors of outcome

in conservatively treated adenocarcinoma in situ (AIS) of the uterine cervix
Silvano Costa a,⁎, Giovanni Negri b , Mario Sideri c , Donatella Santini d , Giuseppe Martinelli d ,
Simona Venturoli e , Carla Pelusi a , Stina Syrjanen f , Kari Syrjanen g , Giuseppe Pelusi a
a
Department of Obstetrics and Gynecology, S. Orsola-Malpighi Hospital, via Massarenti 13, 40138 Bologna, Italy
b
Department of Pathology, Central Hospital, Bolzano, Italy
c
Department of Gynecology Oncology, European Institute of Oncology, Milano, Italy
d
Department of Pathology, S. Orsola-Malpighi Hospital, Bologna, Italy
e
Department of Clinical and Experimental Medicine, Section of Microbiology, University of Bologna, Bologna, Italy
f
Department of Oral Pathology, Institute of Dentistry, University of Turku, Finland
g
Department of Oncology and Radiotherapy, Turku University Hospital, Turku, Finland
Received 20 January 2007
Available online 4 May 2007

Abstract

Objective. The present study assessed (i) the clinical outcome of patients with conservatively treated cervical adenocarcinoma in situ (AIS), (ii)
the accuracy of diagnosing AIS by cytology, colposcopy and histology, as well as (iii) the performance of cervical cytology and HPV testing in
detection of residual or recurrent disease after conservatively treated AIS.
Methods. A series of 42 consecutive women (mean age 40.5 years; range 27–63 years) underwent conservative (cone) treatment of AIS and
were prospectively followed up for a mean of 40 months (median 42 months), using colposcopy, PAP smear, biopsy and HPV testing (with hybrid
capture II) repeated at 6-month intervals.
Results. In their referral PAP test, only 42.9% of patients had atypical glandular cells (AGC) smear. Colposcopy was unsatisfactory in 54.8%
cases and negative in 16.7%. Twenty four patients (57.1%) had AIS as a pure lesions and 18 combined with squamous cell lesion (four had
invasive SCC). Persistent or recurrent disease was observed in 17 (40.4%) cases, 19% in patients with free margins, and 65% among those with
involved margins on the first conization. In four patients, an adenocarcinoma (AdCa) stage IA1 was diagnosed during the follow-up. HPV testing
significantly predicted disease persistence/clearance with OR 12.6 (95% CI 1.18–133.89), while the predictive power of PAP smear did not reach
statistical significance at any of the follow-up visits. The combination of PAP smear and HPV testing gives SE of 90.0%, SP 50.0%, PPV 52.9%
and NPV 88.9% at first follow-up, and 100% SE and 100% NPV at the second follow-up visit.
Conclusions. These results suggest that HR-HPV test in conjunction with cytology offers clear advantages over single cytology in monitoring
the women conservatively treated for cervical AIS.
© 2007 Elsevier Inc. All rights reserved.

Keywords: Adenocarcinoma in situ; Uterine cervix; Conservative treatment; Follow-up; Prospective; PAP smear; HPV testing; Outcome; Test performance

Introduction accounted for only 5% of cervical cancer cases, while in the

Over the past 50 years, the relative proportion (compared
with squamous cell carcinoma) as well as absolute incidence of
invasive and preinvasive glandular lesions of the uterine cervix
have been changing in western countries. Reports from the
1950s and 1960s indicated that adenocarcinomas (AdCa)
⁎ Corresponding author. Fax: +39 51 6364422.
E-mail address: scosta@med.unibo.it (S. Costa).
0090-8258/$ - see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ygyno.2007.03.016
1970s AdCa represented 20–25% of all cervical carcinomas
[1,2].
This change in the ratio of AdCa relative to all cervical
carcinomas may be due to several reasons. The introduction of
the PAP test as a screening tool allowed early detection of
cervical squamous cell lesions, but despite well organized
screening programs, many studies have shown that adenocarci-
noma in situ (AIS) and AdCa are frequently missed by
conventional cytology [3,6]. Reasons for missing AdCa and
 S. Costa et al. / Gynecologic Oncology 106 (2007) 170–176
its precursors might be (a) their location and topography in the
endocervical canal with less accessibility by the spatula and
brush, or (b) failure of cytologists to recognize these lesions
[5,6]. Furthermore, AdCa and its precursor lesions have been
recently included in the growing list of human papillomavirus
(HPV) associated genital tumors [7]. Indeed, HPV DNA has
been detected in 95% of invasive AdCa lesions when
polymerase chain reaction (PCR)-based assays and serology
are used, suggesting that the increase of this disease may be a
consequence of the increasing incidence of HPV infections
[8–10].
Reports on persistent or invasive disease following a
conservative treatment of AIS have been published by several
groups, implicating extra-fascial hysterectomy as the definitive
treatment [11–13]. In a review of 14 studies comprising 157
AIS patients with negative conization margins, 26% harbored
residual AIS and in 2% an unsuspected invasive cancer was
disclosed [12]. Adding to the complexity of management is the
fact that many patients with AIS are in childbearing age and
desire preservation of fertility [14,15]. The conservative
alternative to hysterectomy is cone biopsy with a close
surveillance after treatment to ensure an early diagnosis of
persistent/recurrent disease. Such a follow-up is traditionally
performed with repeated cytology, colposcopy, and, eventually,
using punch biopsy and endocervical curettage (ECC).
Unfortunately, however, these methods have a substantial
false negative rate for glandular lesions either in the primary
diagnosis or in the follow-up of treated patients [4–6,12,13].
Recently, an increased interest has been focused on HPV
detection in cervical lesions. It has been reported that detection
of HPV DNA in the absence of cytological abnormalities may
indicate persistent high-grade squamous lesion missed by PAP
test in the follow-up of CIN3 treated patients [16,17]. However,
to our knowledge, there are no prospective cohort studies
performed by cytology/histology, colposcopy and HPV detec-
tion from pre- to post-conservative treatment of AIS, and
assessment of the potential predictive value of the post-
treatment HPV status for persistent or recurrent cervical
glandular neoplasia.
The present study was designed to assess (i) the clinical
outcome of conservatively treated AIS as diagnosed in pre-and
post-cone biopsies, (ii) the accuracy of diagnosing AIS by
cytology, colposcopy, and histology, as well as (iii) the
performance of cervical cytology and HPV testing (with
commercially available hybrid capture II) in detection of
histologically confirmed residual or recurrent AIS (or invasive
AdCa) during the follow-up of women with conservatively
treated AIS.
Material and methods
Patients and their management
The material of the present study consists of 42 consecutive women; 41 were
referred for colposcopy due to an abnormal PAP smear (the Bethesda System,
TBS 2001), while one patient presented for an abnormal vaginal bleeding. All
were treated for cervical adenocarcinoma in situ (AIS) at our Institutions during
the period 2001–2004. All patients gave informed consent to participate in the
171
study before the study entry. Pertinent clinical data were collected and all
women were prospectively followed-up until May 2006, with the mean FU time
of 40.0 months (median 42 months).
Colposcopic examination was performed using a Zeiss OPM1F (Karl Zeiss,
Jena, Germany). All examinations were done using both acetic acid and iodine
application, and the terminology used is equivalent to the 2001 Colposcopy
Nomenclature of the IFCPC, grading the lesions as abnormal transformation
zone (AnTZ) grade 1 (G1), grade 2 (G2) or cancer, according to the pattern of
the detected colposcopic abnormalities. Visibility of the squamous–columnar
junction (SCJ) was always recorded. Directed punch biopsies and/or an
endocervical curettage (ECC) were taken from all patients at baseline.
All the patients underwent cervical conization, for both diagnostic or
therapeutic purposes. This conservative approach was performed in patients
with a glandular lesion desiring non-radical treatment or preservation of fertility,
after being carefully counseled regarding their risk of persistent disease, or in
women with a squamous intraepithelial lesion (SIL) on punch biopsy. Nineteen
had cold knife cervical cone (CKC), 15 underwent electrosurgical loop excision
(LEEP) using loop size 20 mm wide and 12, 15, or 20 mm deep (Utah Medical
Products, Midvale, Utah, USA), depending on the size of atypical transforma-
tion zone, and 8 had LASER conization performed using a Zeiss 50 W
equipment (Karl Zeiss, Jena, Germany).
After treatment, all women were prospectively followed up and scheduled to
attend the colposcopic clinic at 6-month intervals. All patients were examined by
colposcopy, brush cytology (exo- and endocervical), and ECC. A directed punch
biopsy was performed in case of colposcopic abnormalities. In addition,
repeated samples for HPV DNA detection were taken at each control visit.
Histopathology
The biopsies were fixed in 10% neutral formalin and processed into HE
sections for light microscopy. For histological diagnosis and grading of the
lesions, the commonly accepted morphological criteria were used [18,19]. The
diagnosis of AIS was confirmed by the presence of cervical glandular pseudo-
stratified epithelial cells, enlarged, hyperchromatic nuclei, and mitotic figures
without stromal invasion. Cervical glandular intraepithelial neoplasia (CGIN)
was defined as a spectrum of dysplastic changes that are less severe than those of
AIS but included at least evidence of cellular proliferation and turnover, e.g.
mitotic figures and nuclear atypia to distinguish from benign reactive glandular
atypia. The cytological and histological slides were reviewed by the pathologists
involved in the study, who previously had a consensus meeting in order to define
the morphological criteria.
Following the histological confirmation, all women underwent cervical
conization. Gross and histological processing of the cone specimens was based
on previously described standard criteria [20]. Additional sections were obtained
in case when the margin status was ill defined. All cones were serially sectioned
and analyzed for the clearance of the three margins (exocervical, endocervical,
and deep stromal) that were categorized as free or involved by any grade of
glandular neoplasia. The cone was also measured in three dimensions (width,
height, depth), used to calculate the cone volume (cm3) (W × H × D). All the
histological analyses were processed in the Laboratory of Pathology at our
Institutions and diagnosed by a pathologist blinded to the cytological diagnoses.
HPV DNA testing
The specimens for HPV testing with commercially available hybrid capture
II (HCII) assay were collected at the time of colposcopy and tested with probe B
(high-risk HPVs: types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68).
The test was classified positive at the relative light unit–cut-off (RLU–CO) ratio
(RLU of specimen–mean RLU of two positive controls) of 1 pg/ml. The storage
of the specimens and all reagents as well as conduction of the tests was done
following the manufacturer's instructions (Digene Diagnostics Inc., USA). In
addition, HPV DNA detection by PCR assay was performed in 22 cytological
specimens, as previously described [17].
Data analysis
Statistical analyses were performed using the SPSS® and STATA software
packages (SPSS for Windows, Version 14.0.1. and STATA/SE 9.2). Frequency
172 S. Costa et al. / Gynecologic Oncology 106 (2007) 170–176

tables for categorical variables were analyzed using the Chi-square test, with while 42.9% had an atypical glandular cells (AGC) or AIS
likelihood ratio (LR) or Fisher's exact test to assess the significance of the report, and 1 was negative. Colposcopy was abnormal in 24/42
correlation. Odds ratios (OR) and their 95% CI were calculated where
appropriate. Differences in the means of continuous variables were analyzed
(57.2%) of the women, but normal in 7/42 (16.7%) cases.
using non-parametric tests (Mann–Whitney) or ANOVA (analysis of variance). Importantly, squamo-columnar junction (SCJ) was not visible in
Performance indicators for cytology and HCII test were calculated using the 23/42 (54.8%) cases making the colposcopic examination
contingency tables for sensitivity (SE), specificity (SP), positive (PPV) and unsatisfactory.
negative predictive value (NPV), with 95% CI based on the F-distribution. In all
In directed punch biopsy, pure AIS was detected in 11 cases
tests, the values p < 0.05 were regarded statistically significant.
(26.2%), while different grades of CIN were diagnosed in 8
women. AIS was associated with CIN3 in 8 (19.0%) cases,
Results while 2 presented CGIN. Noteworthy is that cervical biopsy was
negative in 5 (11.9%) women.
The mean age of the women is 40.5 years (SD 8.4 years; Cone histology showed AIS in all 42 patients: 24 (57.1%) as
range 27–63 years). The baseline characteristics of the patients a pure lesion and in 18 combined with squamous cell pathology.
are shown in Table 1. In their referral PAP smear 52.4% of the Thus, four patients (9.5%) had an associated squamous cell
patients showed abnormal squamous abnormalities, 33.3% of cancer (SCC) (two at stage 1A1, one 1A2, and one 1B1 small
which were high-grade squamous intraepithelial lesions (HSIL), volume 1.1 cm in diameter) and 14 (33.3%) had coexistent
CIN3 lesion confirmed. Concordance between colposcopic
Table 1
biopsy and cone histology was only fair to moderate (weighted
Baseline key characteristics of the patients and their lesions kappa or ICC = 0.351; 95% CI −0.934–0.782) (p = 0.214). The
mean height of the cones was 1.8 cm (range 1–2.9) and volume
Characteristics Number (n = 42) Percent
was 3.8 cm3 (range 0.64–12.4). Cone margins were free in 21
Referral PAP smear
(51.2%) cases, while in 15 patients (36.6%) the endocervical
Negative 1 2.4
ASCUS 2 4.8 margin was involved. Altogether, 46.6% of LEEP treated
ASC-H 1 2.4 women, 50% of those treated by LASER, and 61.1% of the
LSIL 5 11.9 CKC patients were disease-free after their initial treatment
HSIL 14 33.3 (Table 1).
AGC 13 31.0
In HCII assay, 1 (2.3%) woman was HPV negative, all others
AIS 5 11.9
Cancer 1 2.4 had HPV DNA detected. In 19 women, HPV type was
Colposcopy determined also by PCR following a HCII positive test, and
Negative 7 16.7 the single most frequent type was HPV18 (8/19; 42.1%)
AnTZ1 4 9.5 followed by HPV16 (4/19; 21.1%). In addition, a double
AnTZ2 20 47.6
infection by HPV16 and HPV18 was found in 15.8%. HPV type
NA 11 26.2
SCJ visible distribution was significantly (p = 0.008) related to cervical
Yes 8 19.0 biopsy but not to PAP smear results.
No 23 54.8 The women were followed up for a mean of 40.0 months (SD
NA 11 26.2 26.5 months; range 3–84 months), with up to 6 clinical visits.
Results of punch biopsy
During the follow-up, 12 women had two cones and 1 had 3
Negative 5 11.9
CIN1 1 2.4 cones performed. In these 13 patients residual disease was
CIN2 2 4.8 found in five; four AIS and one AdCA stage IA2.
CIN3 5 11.9 Hysterectomy was performed in 18 (42.8%) cases based on
CGIN 2 4.8 patients' specific request, older age, multiparity, persistently
AIS 11 26.2
positive cone margins, or diagnosis of invasive carcinoma. Nine
AIS and CIN3 8 19.0
NA 8 19.0 of these procedures were performed after the first follow-up
Cone histology visit, seven at the second follow-up visit, and one each at the
CIN3 14 33.3 third and fourth follow-up control. In these 18 cases, cone
SCC, stage IA1 2 4.8 margins were involved in 13 (72.2%) while in the remaining 5
SCC, stage IA2 1 2.4
were negative. Residual disease was found in 12 patients
SCC, stage IB1 1 2.4
AIS 42 100.0 (66.6%), all with involved cone margins (one case each of CIN3
Involvement of cone margins and CGIN, two cases of AIS, one case of AIS with CIN3
Margins free 21 51.2 associated, three cases of AIS combined with stage IA1 SCC,
Endocervical margin involved 15 36.6 and four AdCa stage IA1). Taken together, persistent or
Exo- and endocervical margin involved 2 4.9
recurrent disease was observed in 17 cases (40.4%): 19% in
Endo-stromal margin involvement 1 2.4
Deep margins involved 1 2.4 patients with free margins and 65% with involved margins at
All margins involved 1 2.4 initial conization.
AnTZ1, atypical transformation zone, grade 1; AnTZ2, atypical transformation At the end of the follow-up, all 42 women were shown to be
zone, grade 2; NA, not available; SCJ, squamo-columnar junction; SCC, disease-free. Similarly, HPV was shown to be cleared in 40
squamous cell carcinoma. women, while 2 remained HPV-negative throughout the follow-
 S. Costa et al. / Gynecologic Oncology 106 (2007) 170–176
up period. PAP smear was not of any predictive value for the
number of cones or hysterectomy decision. Colposcopy result
was significantly related to hysterectomy decision (p = 0.029
Fisher's exact test), which was never done in cases with Grade 1
ATZ, but mostly in cases with negative colposcopy (57.1%).
Similarly, if colposcopy was unsatisfactory, there was a good
change (52.2%) for 2 or 3 cones (p = 0.007). Of the HPV-
negative women, all had only one cone, while 2 or more
positive HPV tests were invariably associated with 2 or more
cones (p = 0.0001). On the other hand, results of the biopsy or
cone histology were not related to number of cones or
hysterectomy decision. The same was true with the margin
involvement in the first cone, which was not related to number
of cones or decision to perform hysterectomy.
During the follow-up, the women were monitored using
colposcopy, PAP smear, biopsy and HPV testing. The
performance of PAP smear and HPV testing in detection of
disease clearance/persistence was monitored at the first four
follow-up visits, when still enough events were recorded for
meaningful analysis. The results are shown in Table 2. HPV test
significantly predicts disease persistence/clearance at the 1st FU
visit only, with OR 12.6 (95% CI 1.18–133.89). Despite ORs
ranging from 3.3 to 5.6, the predictive power of PAP smear does
not reach statistical significance at any of the FU visits. In
subsequent FU visits, also HPV test looses its significance, due
to the rapidly declining cases of persistent disease. At the 1st
and 2nd FU visit, HPV test is more sensitive and PAP is more
specific in detecting residual AIS. Positive predictive value
(PPV) of PAP test is somewhat higher, while HCII shows a
better negative predictive value (NPV). Due to the disappear-
ance of residual AIS, the calculations of these performance
indices are not possible at the 4th, 5th, and 6th FU visit.
When both PAP smear and HPV test are used together, such
a combined test detects persistent lesions at the 1st FU visit with
OR = 9.0 (95% 0.91–88.57) and gives sensitivity (SE) of 90.0%,
specificity (SP) 50.0%, PPV 52.9%, and NPV 88.9%. At the
2nd FU visit, this combination gives SE 100%, SP 52.6%, PPV
40.0%, and NPV 100% (OR not computable). At the 3rd FU
visit, when most of the disease has disappeared, SE is 0%, SP
91.7%, PPV 0%, and NPV 91.7%. After that, all AIS lesions
Table 2
Performance of PAP smear and HPV detection in the post-treatment follow-up of endocervical AIS
Test/FU visit Sensitivity a Specificity
First FU visit
PAP smear 60.0 (26.2–87.8) 68.7 (41.3–88.9)
HPV test 90.0 (55.5–99.7) 58.3 (27.7–84.8)
Second FU visit
PAP smear 66.7 (22.2–95.6) 73.7 (48.8–90.8)
HPV test 83.3 (35.8–99.5) 58.8 (32.9–81.5)
Third FU visit
PAP smear 0.0 (0.0–84.2) 95.4 (77.2–99.8)
HPV test 0.0 (0.0–84.3) 91.7 (73.0–98.9)
Fourth FU visit NC NC
Fifth FU visit NC NC
Sixth FU visit NC NC
NC, non-computable, all cases regressed.
a
Biopsy and colposcopy used as the gold standard.
173
have disappeared, precluding these calculations at the 4th, 5th,
and 6th FU visit.
Discussion
In screened populations, early detection and treatment of
squamous precancer lesions have decreased the incidence of
SCC, whereas many studies have shown that glandular lesions
are frequently missed by conventional cytology [21–23]. One
of the main reasons for the low sensitivity of PAP smear for
glandular lesions is their location in the endocervical canal
which is less well accessible for adequate cytological sampling.
Furthermore, because of the relatively rarity of the disease,
cytologists rarely gather sufficient experience with glandular
neoplasms [5,6,24–27]. Difficulty in diagnosing AIS even in
biopsy is further increased by the fact that this lesion does not
always show any evident abnormalities on colposcopy, and
minimally invasive biopsy procedure or scraping of endocervi-
cal lesions may be inadequate, leading to false negative
diagnoses. As pointed out by several authors [28–30] all
these contribute to the fact that only 40% to 60% of AIS cases
are correctly diagnosed before conization.
It is well established that glandular neoplasms frequently
coexist with invasive squamous cell carcinoma or its precursors.
Prevalence of HPV in AdCa lesions is equally high as in SCC,
HPV playing an etiological role in both diseases [7–10].
Moreover, squamous cell lesions develop at the SCJ and AIS is
often observed adjacent to the upper part of the transformation
zone [25,26,28–30]. These general observations were con-
firmed in our series, where 33% of AIS cases were associated
with CIN3 and 9.5% cases with invasive SCC.
In the present series of 42 AIS cases, an abnormal PAP smear
almost invariably initiated the workup, even if referral cytology
was consistent with AIS in less than 50% of cases [27,31,32].
Importantly, colposcopy was abnormal in only 57% of the cases
and entirely negative in 16%. In 55% of the patients,
colposcopic examination was unsatisfactory due to the non-
visible SCJ. It has been suggested that colposcopic subtlety
[14,15,27] is probably due to the multifocality, topography and
possible location of AIS into the deep clefts of the endocervical
PPV NPV OR (95% CI)
54.6 (23.4–83.2) 73.3 (44.9–92.2) 3.3 (0.63–17.16)
64.3 (35.1–87.2) 87.5 (47.3–99.6) 12.6 (1.18–133.89)
44.4 (13.7–78.8) 87.5 (61.6–98.4) 5.6 (0.77–40.59)
41.7 (15.1–72.3) 90.9 (58.7–99.7) 7.1 (0.67–75.21)
0.0 (0.0–97.5) 91.3 (71.9–98.9) 1.1 (0.96–1.24)
0.0 (0.0–84.2) 91.7 (73.0–98.9) 1.1 (0.97–1.23)
NC NC NC
NC NC NC
NC NC NC
174 S. Costa et al. / Gynecologic Oncology 106 (2007) 170–176
glands with consequently less accessibility than squamous
lesions that are more prominent and localized in the
transformation zone [31–33]. The inadequacy of colposcopy
in recognizing endocervical lesions implies that punch biopsy
cannot be considered an adequate endpoint in the pre-surgical
diagnosis of AIS. This is well demonstrated in this series, where
19% of directed cervical biopsies showed SIL only, and,
importantly, 12% were totally negative. Given the inability of
routine diagnostic tests to detect glandular lesions in up to half
of the cases, conization of the cervix and histopathological
confirmation are essential for a definitive diagnosis of AIS.
In this prospective series, the cone modalities were at
discretion of the referring clinical centers with the newer
techniques (LASER or LEEP) being used more frequently
(54.7%) than CKC [28,32]. As previously reported [25–34],
cone margins of the first cone were more frequently positive
with LEEP than CKC cases (53.4% of LEEP versus 38.9% of
CKC cases). This agrees with the reported data comparing the
margin status with the type of cone. Nevertheless, comparing
LEEP and CKC might not be fully justified because LEEP is
more often used as a diagnostic method, whereas CKC is an
intended therapeutic procedure following the confirmation of
AIS diagnosis [29].
Several previous studies have addressed the cone margin
status as a predictor of residual disease [11,21,34–36]. As
pointed out before [11,33–36], incomplete cone is a frequent
cause of treatment failure in AIS. In fact, residual AIS or
invasive AdCa has been demonstrated in 20%–40% of patients
with uninvolved cone margins and in 50%–80% of those with
involved margins. In the present series, 19% (4/21) of the
patients with negative cone margins and 65% (13/20) of those
with involved margins had histology-confirmed residual
disease, implicating a significant risk of persistence/recurrence
of intraepithelial or invasive glandular lesion in cases where
cone margins are compromised [36–38]. When margins are not
free and future fertility is desired, a rigorous follow-up must be
arranged and a second cone should be performed in case of
lesion persistence [36,40].
In fact the main objective of such a follow-up is an early
detection of residual/recurrent disease because of the significant
risk of progression to invasive carcinoma, if an effective
treatment is not administered [11]. Current follow-up protocols
are mainly based on serial cytology and, eventually, ECC,
colposcopy adding little to the detection rate of glandular
neoplasia. However, there is evidence that glandular lesions are
frequently missed by conventional cytology. In management of
squamous cell lesions, detection of high-risk HPV (HR-HPV)
has been recently proposed in the follow-up of patients treated
for CIN2–3 because of its high SE and NPV value in detecting
residual/recurrent disease, i.e., HPV detection is as a good index
of disease clearance [16]. Because HR-HPV is intimately
involved in etiology of these glandular lesions as well, we
studied the performance of HPV testing in conjunction with
cervical cytology in detecting histologically confirmed residual
or recurrent disease during the follow-up of women conserva-
tively treated for AIS. Our data show that HPV test significantly
predicts disease clearance/persistence, in contrast to PAP smear,
whose predictive power does not reach statistical significance at
any of the FU visits. Furthermore, HPV test shows a better NPV
while PAP smear is more specific in detecting residual AIS. The
combination of PAP smear and HPV test has SE of 90.0% and
NPV of 88.9% for persistent lesions at the 1st FU visit. At the
2nd FU visit, HR-HPV test in conjunction with cytology gives
SE 100% and NPV 100%.
AIS of the cervix is a premalignant condition that is being
diagnosed with increasing frequency. Historically, an extra-
fascial or even radical hysterectomy has been considered as the
standard treatment for all glandular in situ lesions. This
management was recommended by several authors in the past
(1) because of the possibility of a persistent disease in excisional
biopsy with negative borders (due to multi-focal disease “skip
lesions”) and (2) because of the concern of occult invasive
AdCa in the deep clefts of the endocervical glands, with normal-
appearing overlying epithelium [15].
In recent years, a trend towards less radical, fertility-sparing
surgery has been proposed as a treatment option, considering
that in the majority of cases clinicians have to face with young
women who desire conservative therapy. The solid argument for
this new trend is provided by our better understanding of the
natural history of cervical glandular lesions, including the data
on the duration of these pre-invasive stages [38,39], and their
association with HR-HPV infections [24–26]. Nevertheless, the
management of AIS continues to be a controversial issue
because conflicting data are being constantly reported. While
some authors underline the high incidence of residual in situ or
invasive disease following various forms of cone biopsy, there
are others who emphasize the risk of “unnecessary” hyster-
ectomies, if further radical treatment is indicated solely on the
basis of the margin status [28,38–40]. Our experience from the
present series clearly substantiates the latter view because 33%
of our patients who underwent radical surgery had a subsequent
negative pathology report.
Evidently, the safety of conservative management of AIS
adopted in specifically defined circumstances depends on the
accuracy of screening methods utilized in predicting the
residual disease [41,42]. Our study showed that the combination
of PAP smear and HPV test gives SE of 90% in detecting
persistent lesions at the 1st FU visit, and 100% SE at the 2nd FU
visit. Moreover, the NPV of 100% seems to be very useful in
preventing unnecessary hysterectomies. To conclude, the results
of our study suggest that HR-HPV test in conjunction with
cytology offers clear advantages over single cytology in
monitoring the women conservatively treated for glandular
intraepithelial neoplasia. This is because of their high sensitivity
and NPV that help ensuring an early detection of any patients at
increased risk for disease recurrence and progression, thus
preventing unnecessary radical treatment [43]. A multi-center
trial based on a larger cohort would be warranted to confirm
these preliminary data.
On the basis of this limited number of cases, however, it is
impossible to formulate a strategy that could satisfy the strict
requirements of commonly agreed guidelines of management
of AIS patients. However, in simple terms and substantiated by
the above discussed data, we suggest that (i) after treatment,
 S. Costa et al. / Gynecologic Oncology 106 (2007) 170–176
these women should be followed up at 6-month intervals, for a
minimum of 2 years, (ii) both PAP test and HPV testing should
be done; (iii) in case of both tests being negative, no other
diagnostic test is needed, and after 2 years, the interval of
controls could be extended to 12 months; (iv) if either PAP or
HPV test is positive, however, colposcopy and ECC are
mandatory; (v) additional management decisions must be then
based on the results of these two. Rather than speaking about
guidelines, we would call this a potential strategy that could be
useful in monitoring the women treated due to AIS, for early
detection of the patients at risk to develop a progressive
disease.
Acknowledgments
This study was supported by the research grant from
Fondazione Cassa di Risparmio, Bologna, No. 2005/0457,
2006.
References
[1] Smith HO, Tiffany MF, Qualls CR, Key CR. The rising incidence of
adenocarcinoma relative to squamous cell carcinoma of the uterine cervix
in the united States. A 24-year population-based study. Gynecol Oncol
2000;78:97–105.
[2] Horowitz IR, Jacobson LP, Zucker PK, Currie J, Rosenshein NB.
Epidemiology of adenocarcinoma of the cervix. Gynecol Oncol 1988;31:
25–31.
[3] Vizcaino AP, Moreno V, Bosch FX, Munoz N, Barros-Dios XM, Parkin
DM. International trends in the incidence of cervical cancer: I.
Adenocarcinoma and adenosquamous cell carcinomas. Int J Cancer
1998;75:536–45.
[4] Raab SS. Can glandular lesions be diagnosed in pap smear cytology?
Diagn Cytopathol 2000;23:127–33.
[5] Krane JF, Granter SR, Trask CE, Hogan CL, Lee KR. Papanicolaou smear
sensitivity for the detection of adenocarcinoma of the cervix: a study of 49
cases. Cancer 2001;93:8–15.
[6] Ruba S, Schoolland M, Allpress S, Sterrett G. Adenocarcinoma in situ of
he uterine cervix. Screening and diagnostic errors in Papanicolau smears.
Cancer Cytopath 2004;102:280–7.
[7] Walboomers JM, Jacobs MV, Manos MM, Bosch FX, Kummer JA, Shah
KV, et al. Human papillomavirus is a necessary cause of invasive cervical
cancer worldwide. J Pathol 1999;189:12–9.
[8] Zielinski GD, Snijders PJ, Rozendaal L, Daalmeijer NF, Risse EK,
Voorhorst FJ, et al. The presence of high-risk HPV combined with specific
p53 and p16INK4a expression patterns points to high-risk HPVas the main
causative agent for adenocarcinoma in situ and adenocarcinoma of the
cervix. J Pathol 2003;201:535–43.
[9] Derchain SFM, Rabelo-antos SH, Sarian LO, Zeferino LC, de Olivera
Zambeli ER, do Amaral Westin MC, et al. Human papillomavirus DNA
detection and histological findings in women referred for atypical
glandular cells or adenocarcinoma in situ in their Pap smears. Gynecol
Oncol 2004;95:618–23.
[10] Andersson S, Rylander E, Larsson B, Strand A, Silfversvard C, Wilander
E. The role of human papillomavirus in cervical adenocarcinomna
carcinogenesis. Eur J Cancer 2001;37:246–50.
[11] Cohn DE, Morrison CD, Zanagnolo VL, Goist M, Copeland LJ. Invasive
cervical adenocarcinoma immediately following a cone biopsy for
adenocarcinoma in situ with negative margins. Gynecol Oncol 2005;98:
158–60.
[12] Shipman SD, Bristow RE. Adenocarcinoma in situ and early invasive
adenocarcinoma of the uterine cervix. Curr Opin Oncol 2001;13:394–8.
[13] Krivak TC, Rose GJ, McBroom JW, Carlson JW, Winter WE, Kost ER.
Cervical adenocarcinoma in situ: a systematic review of therapeutic
175
options and predictors of persistent or recurrent disease. Obstet Gynecol
Surv 2001;56:567–75.
[14] ACOG Practice Bulletin. Clinical management guidelines for obstetrician-
gynaecologist. Management of abnormal cervical cytology and histology.
Obstet Gynecol 2005;106:645–63.
[15] Bryson P, Stulberg R, Shepherd L, McLelland K, Jeffrey J. Is
electrosurgical loop excision with negative margins sufficient treatment
for cervical ACIS? Gynecol Oncol 2004;93:465–8.
[16] Costa S, De Simone P, Venturoli S, Cricca M, Zerbini M, Musiani M, et al.
Factors predicting human papillomavirus clearance in cervical intraepithe-
lial neoplasia lesions treated by conization. Gynecol Oncol 2003;90:
358–65.
[17] Cricca M, Venturoli S, Morselli-Labate AM, Costa S, Santini D, Ambretti
M, et al. HPV DNA patterns and disease implications in the follow-up of
patients treated for HPV16 high-grade carcinoma in situ. J Med Virol
2006;78:494–500.
[18] Clement BP, Young RH. Atlas of Gynecologic Surgical Pathology.
Philadelphia: W.B. Saunders; 2000. p. 109–13.
[19] Zaino RJ. Symposium part I: adenocarcinoma in situ, glandular dysplasia,
and early invasive adenocarcinoma of the uterine cervix. Int J Gynecol
Pathol 2002;21:314–26.
[20] Costa S, De Nuzzo M, Infante FE, Bonavita B, Marinelli M, Rambelli V,
et al. Disease persistence in patients with cervical intraepithelial neoplasia
undergoing electrosurgical conization. Gynecol Oncol 2002;85:119–24.
[21] Sherman ME, Wang SS, Carreon J, Devesa SS. Mortality trends for
cervical squamous and adenocarcinoma in the United States. Cancer
2005;103:1258–64.
[22] Stockton D, Cooper P, Lonsdale RN. Changing incidence of invasive
adenocarcinoma of the uterine cervix in East Anglia. J Med Screen 1997;4:
40–3.
[23] Negri G, Egarter-Vigl E, Kasal A, Romano F, Haitel A, Mian C. p16INK4a
is a useful marker for the diagnosis of adenocarcinoma of the cervix uteri
and ist precursors. Am J Surg Pathol 2003;27:187–93.
[24] Soutter WP, Haidopoulos D, Gornall RJ, McIndoe GA, Fox J, Mason WP,
et al. Is conservative treatment for adenocarcinoma in situ of the cervix
safe? BJOG 2001;108:1184–9.
[25] Widrich T, Kennedy AW, Myers TM, Hart WR, Wirth S. Adenocarcinoma
in situ of the uterine cervix: management and outcome. Gynecol Oncol
1996;61:304–8.
[26] Muntz GM. Can cervical adenocarcinoma in situ be safely managed by
conizations alone. Gynecol Oncol 1996;61:301–3.
[27] Ostor AG, Duncan A, Quinn M, Rome R. Adenocarcinoma in situ of the
uterine cervix: an experience with 100 cases. Gynecol Oncol 2000;79(2):
207–10.
[28] Akiba Y, Kubushiro K, Fukuchi T, Fujii K, Tsukazaki M, Mukai S. Is laser
conization adequate for therapeutic excision f adenocarcinoma in situ of
the uterine cervix? J Obstet Gynaecol Res 2005;31:252–6.
[29] Kennedy AW, Biscotti CV. Further study of the management of cervical
adenocarcinoma in situ. Gynecol Oncol 2002;86:361–4.
[30] Krivak TC, Rose GS, McBroom JW, Carlson JW, Winter WE, Kost ER.
Cervical adenocarcinoma in situ: a systematic review of therapeutic
options and predictors of persistent or recurrent disease. Obstet Gynecol
Survey 2001;56:179–82.
[31] Azodi M, Chambers SK, Rutherford TJ, Kohorn EI, Scwartz PE,
Chambers JT. Adenocarcinoma in situ of the cervix: management options.
Gynecol Oncol 1999;73:348–53.
[32] Andersen ES, Nielsen K. Adenocarcinoma in situ of the cervix: a
prospective study of conization as definitive treatment. Gynecol Oncol
2002;86:365–9.
[33] Levenback WJK, Malpica A, Morris M, Burke T, Mitchell MF.
Adenocarcinoma in situ of the cervix: significance of cone biopsy
margins. Obstet Gynecol 1996;88:82–6.
[34] Goldstein NS, Mani A. The status and distance of cone biopsy margins as a
predictor of excision adequacy for endocervical adenocarcinoma in situ.
Am J Clin Pathol 1998;109:727–32.
[35] Bertrand M, Lickrish GM, Colgan TJ. The anatomic distribution of
cervical adenocarcinoma in situ: implications for treatment. Am J Obstet
Gyecol 1987;157:21–5.
176 S. Costa et al. / Gynecologic Oncology 106 (2007) 170–176
[36] McHale MT, Le TD, Burger RA, Gu M, Rutgers JL, Monk BJ. Fertility
sparing treatment for in situ and early invasive adenocarcinoma of the
cervix. Obstet Gynecol 2001;98:726–31.
[37] Im DD, Duska LR, Rosenheim NB. Adequacy of conization margins in
adenocarcinoma in situ of the cervix as a predictor of residual disease.
Gynecol Oncol 1995;59:179–82.
[38] Plaxe SC, Saltzstein L. Estimation of the duration of the preliminary phase
of cervical adenocarcinoma suggests that there is ample opportunity for
screening. Gynecol Oncol 1999;75:55–61.
[39] Andersen ES, Arfmann E. Adenocarcinoma in situ of the uterine cervix: a
clinico-pathologic study of 36 cases. Gynecol Oncol 1989;35:1–7.
[40] Shin CH, Schorge JO, Kenneth RL, Sheets EE. Conservative management
of adenocarcinoma in situ of the cervix. Gynecol Oncol 2000;79:6–10.
[41] Saqi A, Gupta PK, Erroll M, Babiac A, Blackmun D, Mansukhani M, et al.
High-risk human papillomavirus DNA testing: a marker for atypical
glandular cells. Diagn Cytopathol 2006;34:235–9.
[42] Syrjänen KJ. Improved detection of adenocarcinoma in situ (AIS) by
screening. A key to reducing the incidence of cervical adenocarcinoma?
Editorial. Acta Cytol 2004;48:591–4.
[43] Wang SS, Sherman ME, Silverberg SG, Carreon JD, Lacey Jr JV, Zaino R,
et al. Pathological characteristics of cervical adenocarcinoma in a multi-
center US-based study. Gynecol Oncol 2006;103:541–6.