Management and outcome of cervical cancer

diagnosed
in pregnancy
BACKGROUND: Cervical cancer is the third most common gynecologic malignancy in
the United States. Approximately 1-3% of cervical cancers will be diagnosed in
pregnant and peripartum women; optimal management in the setting of pregnancy
is not always clear.

OBJECTIVE: We sought to describe the management of patients with cervical cancer

diagnosed in pregnancy and compare their outcomes to nonpregnant women with
similar baseline characteristics.

STUDY DESIGN: We conducted a retrospective chart review of all patients diagnosed

with cervical cancer in pregnancy and matched them 1:2 with contemporaneous
nonpregnant women of the same age diagnosed with cervical cancer of the same
stage. Patients were identified using International Classification of Diseases, Ninth
Revision codes and the Dana-Farber/Massachusetts General Hospital Cancer
Registry. Data were analyzed using Stata, Version 10.1 (College Station, TX).

RESULTS: In all, 28 women diagnosed with cervical cancer during pregnancy were
identified from 1997 through 2013. The majority were Stage IB1. In all, 25% (7/28)
of women terminated the pregnancy; these women were more likely to be
diagnosed earlier in pregnancy (10.9 vs 19.7 weeks, P= .006). For those who did not
terminate, mean gestational age at delivery was 36.1 weeks. Pregnancy
complications were uncommon. Complication rates in pregnant women undergoing
radical hysterectomy were similar to those outside of pregnancy. Time to treatment
was significantly longer for pregnant women compared to nonpregnant patients
(20.8 vs 7.9 weeks, P = .0014) but there was no survival difference between groups
(89.3% vs 95.2%, P = .08). Women who underwent gravid radical hysterectomy had
significantly higher estimated blood loss than those who had a radical hysterectomy
in the postpartum period (2033 vs 425 mL, P = .0064), but operative characteristics
were otherwise similar. None of the pregnant women who died delayed treatment
due to pregnancy.

CONCLUSION: Gestational age at diagnosis is an important determinant of

management of cervical cancer in pregnancy, underscoring the need for expeditious
workup of abnormal cervical cytology. Of women who choose to continue the
pregnancy, most delivered in the late preterm period without significant obstetric
complications. For women undergoing radical hysterectomy in the peripartum
period, complication rates are similar to nonpregnant women undergoing this
procedure. Women who died were more likely to have advanced stage disease at
the time of diagnosis. This information may be useful in counseling women facing
the diagnosis of cervical cancer in pregnancy.
Key words: cancer in pregnancy, cervical cancer, gravid hysterectomy, oncology in
pregnancy, radical hysterectomy

Introduction

Cancer diagnosis during pregnancy has risen over the last 50 years, now
complicating up to 1 in 1000 pregnancies. Some authors suggest that this may be
attributable to delay in childbearing to third or fourth decade of life. Cervical cancer
is the third most common gynecologic cancer in the United States, with >12,000
new cases of invasive cervical cancer diagnosed annually and approximately 4000
cervical cancer related deaths every year.

Cervical cancer is most frequently diagnosed at an early stage in both pregnant and
nonpregnant populations, and typically treated surgically with radical hysterectomy
plus bilateral pelvic lymphadenectomy. Approximately 1-3% of cases of cervical
cancer will be diagnosed in pregnant or peripartum women. However, the optimal
treatment for cervical cancer diagnosed in pregnancy is not always clear. Providers
must consider stage at the time of diagnosis, trimester of pregnancy, and patient
preferences regarding pregnancy outcome. We sought to describe the management
and subsequent outcomes of patients with cervical cancer diagnosed in pregnancy
compared to nonpregnant women with similar baseline characteristics.

Materials and Methods

This was a retrospective study of all patients diagnosed with cervical cancer in
pregnancy at Brigham and Womens Hospital and Massachusetts General Hospital.
Patients were identified using International Classification of Diseases, Ninth Revision
codes and the Dana Farber and Massachusetts General Hospital Cancer Registry.
These women were then matched 1:2 with contemporaneous nonpregnant patients
diagnosed within the same 5-year period. Patients were stage- and age-matched for
comparison.

Information was collected regarding sociodemographic characteristics as well as

medical and pregnancy history. Pregnancy outcomes included termination of
pregnancy, mode of delivery for women who did not undergo termination,
gestational age (GA) at delivery, birthweight, and 5-minute Apgar scores. Oncologic
outcomes of interest included time from diagnosis to treatment, method of
treatment (surgical vs medical), operative time, estimated blood loss (EBL), surgical
complications, and 5- year survival. Surgical complications were defined as
intensive care unit admission, unexpected injury to adjacent organs, or reoperation.
In a planned subgroup analysis among women diagnosed with cervical cancer in
pregnancy, we compared outcomes of women who elected termination of
pregnancy with those who continued their pregnancy. Finally, we evaluated surgical
outcomes for women who underwent gravid radical hysterectomy and compared
them to women who had surgical management in the postpartum period. Study
data were collected and managed using Research Electronic Data Capture (REDCap)
tools hosted at Partners Health System. REDCap is a secure, World Wide
Webebased application designed to support data capture for research studies.

Data were analyzed using software (Stata, Version 10.1; StataCorp LP, College
Station, TX). The X2 test was used for categorical data, while a 2-sided t-test was
used for continuous variables. A P value of <.05 was considered significant. This
study was approved by the Partners Health System Institutional Review Board.

Results

A total of 28 women diagnosed with cervical cancer during pregnancy were

identified from 1997 through 2013. These patients were matched with 52
nonpregnant control patients. The majority of patients diagnosed with cancer in
pregnancy were stage IB1 at the time of diagnosis (Figure). Demographic
characteristics were similar between the pregnant and nonpregnant groups, as were
medical comorbidities (Table 1).

On average, patients were diagnosed in the 17th week of gestation (mean GA at

diagnosis 17.4 weeks). Of the pregnant patients, 25% (7/28) terminated their
pregnancy after learning of their diagnosis. These women were significantly more
likely to be diagnosed in the first trimester (mean GA at diagnosis 10.9 vs 19.7
weeks, P < .005) (Table 2). While most pregnant women were early stage (79% of
women who continued, 63% of women who terminated), all those who terminated
were stage IB1 or greater at the time of the diagnosis (5/7 women stage IB1, 2/7
stage IIB). There were no differences in demographics, medical comorbidities,
parity, or treating hospital between the 2 groups. Their oncologic outcomes were
also similar (Table 2).

For the women who continued their pregnancies, there were no significant obstetric
complications, including cervical insufficiency, preterm labor, preterm premature
rupture of membranes, or fetal growth restriction. The mean GA at delivery was in
the late preterm period at 36.1 weeks (SD 5.1 weeks) with a mean birthweight of
2820 g, which is appropriate for this GA.12 Apgar scores for all neonates were 7 at
5 minutes. Three patients with known invasive cervical cancer had planned vaginal
deliveries without reported obstetric complications (Table 3); all of these patients
had microinvasive disease. An additional 7 patients had vaginal deliveries in the
setting of cervical cytology and biopsy findings suggestive of cancer during
pregnancy for which the definitive diagnostic procedure (conization or
hysterectomy) was deferred until the immediate postpartum period either because
the patient declined the procedure or because the managing physician opted to
delay.

Oncologic outcomes were similar between pregnant patients and nonpregnant

control subjects (Table 3). The mean time from diagnosis to initiation of treatment in
the pregnant cohort was 20.8 weeks (range 2-62.2 weeks). This was statistically
significantly higher than the nonpregnant patients (20.8 vs 7.9 weeks, P = .0014),
although no significant difference in survival was seen (89.3%vs 95.2%, P = .08)
and follow-up time was similar (3.4 vs 3.7 years, P = .73). Importantly, none of the
women who died delayed treatment due to pregnancy status; all of these women
underwent pregnancy termination or had a spontaneous loss.

There were 6 women who underwent gravid radical hysterectomy (radical

hysterectomy performed at the time of cesarean delivery) and 8 women who had
surgical treatment at 6-8 weeks postpartum (Table 4). Women undergoing cesarean
radical hysterectomy were similar demographically to those who had surgery in the
postpartum period, although there was a trend that white women were more likely
to have a gravid radical hysterectomy (5/6 [83.3%] vs 3/8 [37.5%], P = .09). As
expected, these women also had statistically significantly higher EBL (2033 vs 425
mL, P = .0064), although were not more likely to receive a blood transfusion (50%
vs 25%, P = .33) (Table 4). All womenwere still alive at the time of this analysis.

There were 3 pregnant women who died of cervical cancer. One patient had a Pap
smear at 13 weeks gestation demonstrating squamous cell cancer. She was seen
for colposcopy 3 weeks after the initial Pap smear and had a biopsy that confirmed
invasive squamous cell carcinoma. She miscarried shortly after that appointment
and followed up with gynecologic oncology 1 month later (7 weeks after the initial
Pap smear). She was diagnosed with stage IVA cervical cancer at that time based on
physical exam and magnetic resonance imaging (MRI) notable for vaginal extension,
bony metastases, and distant lung metastases. She had external beam pelvic
radiation with sensitizing cisplatin and zoledronic acid for bony metastases followed
by systemic chemotherapy. She demonstrated continued progression of pulmonary
metastases despite multiple lines of chemotherapy and surgical wedge resection.
The patient ultimately developed brain metastases and died 3 years after her
diagnosis.

The second patient was diagnosedwith stage IIB cervical cancer at 19 weeks GA
after cervical biopsy demonstrated invasive cancer and MRI demonstrated complete
replacement of the cervix by tumor with bilateral parametrial invasion. She
underwent fetal intracardiac potassium chloride injection followed by laparotomy
with hysterotomy for pregnancytermination prior to initiation of treatment. She was
treated with sensitizing cisplatin and concomitant external beam radiation along
with brachytherapy. The patient developed recurrent disease after 3 years,
diagnosed based on metastatic supraclavicular nodes and symptoms consistent
with a Pancoast tumor. The patient underwent multiple lines of treatment and died 5
years after her initial diagnosis.

The final pregnant patient who died due to cervical cancer was diagnosed with
stage IIB disease in the first trimester after colposcopic biopsies demonstrated
poorly differentiated adenosquamous carcinoma, with physical exam notable for left
parametrial and upper vaginal involvement. She underwent medical termination of
pregnancy followed by external beam pelvic radiation and concomitant
radiosensitizing cisplatin. Due to marked improvement noted on MRI, she also
underwent tandem and ovoid brachytherapy. She represented 1 year later with
worsening pain and was found to have recurrent disease with paraaortic metastasis.
The patient received palliative radiation and died 17 months after her initial
diagnosis.

Comment

Management of medical complications in pregnancy, and particularly malignancies,

is challenging. Understanding the potential tradeoffs involved in altering the usual
care recommended outside of pregnancy is an essential part of helping patients and
providers make an individualized choice. We found that the oncologic outcomes and
survival of pregnant women diagnosed with cervical cancer were similar to
nonpregnant women of a similar age with a similar stage at diagnosis. For women
who opted to continue pregnancy, obstetric complications were rare; the most
common was iatrogenic preterm birth to facilitate treatment.

GA at diagnosis was an important determinant of management of cervical cancer in

pregnancy in this study. While 75% of women were diagnosed <24 weeks, we
demonstrated a significantly higher rate of pregnancy termination with earlier
diagnosis of invasive cervical cancer, underscoring the need for expeditious workup
of abnormal cervical cytology in pregnancy, which can include colposcopy, targeted
cervical biopsy, as well as cervical conization depending on the GA and clinical
scenario. Although the natural history of cervical intraepithelial neoplasia in
pregnancy demonstrates high rates of spontaneous regression, early invasive
disease is unlikely to resolve.

There were 7 patients who did not undergo a definitive diagnostic procedure until
the immediate postpartum period despite cytology or biopsies during pregnancy
concerning for invasive cancer these 7 had reportedly uncomplicated vaginal births.
While we would not recommend a delay of diagnostic procedure when concern for
cancer has been raised, these patients were included in the analysis because based
on cytology and/or biopsy findings, cervical cancer was present during pregnancy.
The opportunity to review their obstetric and oncologic outcomes is both interesting
and potentially informative with regard to the impact of delay in treatment and the
option of vaginal delivery. The literature is mixed regarding the outcome of cervical
cancer after vaginal delivery.
Most pregnant women in our cohort who had a confirmed cancer diagnosis prior to
delivery underwent planned cesarean; planned vaginal delivery in the setting of
microinvasive disease or vaginal delivery in the absence of a confirmed diagnosis
was not associated with disease progression, obstetric complications, or decreased
survival in this small sample. However, in light of literature suggestive of increased
risk of local recurrence and potential for increased risk of distant metastasis, it is
reasonable to recommend elective cesarean delivery.

In terms of the timing of surgical management for women who chose to continue
pregnancy, a delay in management until the postpartum period did not significantly
impact survival, with 5-year survival rates similar to those available in large
population studies. While the pregnant patients had a significantly longer delay
from diagnosis to treatment, this did not seem to impact the overall oncologic
outcome. In this setting, it may be reasonable to consider delaying delivery until at
least 37 weeks gestation, if not 39 weeks, due to increasing pediatric literature
about the risks of late preterm and early term birth. Many current studies of cervical
cancer management in the context of pregnancy describe iatrogenic preterm birth
to expedite maternal oncologic treatment.

While Xia and colleagues noted decreased survival in patients who delayed
treatment of cervical cancer in a Chinese cohort, this patient group had a high
proportion of tumors >4 cm and aggressive histopathologic subtypes. This finding
has not been described by other groups. In terms of immediate postoperative
complications, as expected, the EBL was significantly higher with peripartum
hysterectomy, although there was no significant difference in transfusion. Based on
prior literature as well as our observations, it may be reasonable to consider delay
of definitive surgical management until the postpartum period, to minimize both
neonatal and maternal immediate complications without affecting longterm
outcome.

As expected, women who died were more likely to be of advanced stage at the time
of diagnosis, in both pregnant and nonpregnant cohorts. While there was a trend
toward decreased survival in the pregnant patients, none of these patients delayed
cancer treatment in the setting of pregnancyeall 3 patients underwent termination
or had a spontaneous loss. While no patients in our study received neoadjuvant
chemotherapy in the setting of advanced stage disease to allow fetal maturation,
this would be a potential therapeutic option and could be explored with the patient
and her oncologist depending on the clinical scenario.

Our study has limitations. This study is limited by its retrospective nature, which
may be affected by confounding and reporting bias. Additionally, because cervical
cancer in pregnancy is a relatively rare event, we are limited by small numbers. Due
to the nature and scope of this study, we were not able to collect and compare
specific neonatal outcomes, although these were universally favorable. In spite of
these limitations, the ability to compare age- and stagematched pregnant and
nonpregnant cervical cancer patients provides some insight into oncologic
outcomes based on pregnancy status. The matched design in a contemporary
cohort also corrects for potential changes in practice patterns over the time period
studied. We also compared timing of radical hysterectomy in the peripartum period,
which has been a point of discussion in the literature. In this study, we did not
demonstrate a significant impact on survival with a delay in surgery to the
postpartum period, although the treatment delay in our pregnancy cohort was
significantly longer than the nonpregnant patients.

Cervical cancer diagnosed in the context of pregnancy remains a rare event,

although interaction with the medical system in pregnancy provides a unique
opportunity to screen for and diagnose this gynecologic malignancy. We would
recommend following age-based guidelines put forth by the American Society for
Colposcopy and Cervical Pathology (ASCCP) for cervical cancer screening in
pregnancy, provided obstetricians are able to confirm normal cytology with prior
records. We recognize that this may be challenging in populations that pursue
antenatal care in a different system from their primary preventative care. If prior
normal cytologic results are unable to be confirmed, screening Pap (with human
papillomavirus cotesting based on ASCCP guidelines) may be considered at the
initial prenatal visit.

Based on this series, tertiary and quaternary referral centers are likely to manage
approximately 1-2 cases of cervical cancer in pregnancy annually, many of which
will be referred after abnormal cervical cytology or biopsy results. Management
strategies among multidisciplinary teams remain variable, appropriately taking into
account the patients individual clinical characteristics and personal preferences
regarding pregnancy outcome and potential treatment delay. Results of this study
may be useful in counseling women facing the diagnosis of cervical cancer in
pregnancy. We continue to recommend a multidisciplinary approach with maternal
fetal medicine and gynecologic oncology to guide the counseling and treatment of
these patients.