CHIEF EDITOR’S NOTE: This article is part of a series of continuing education activities in this Journal through which a totalof 36 AMA/PRA category 1 creditsTM can be earned in 2006. Instructions for how CME credits can be earned appear on thelast page of the Table of Contents.

Managing Ovarian MassesDuring Pregnancy

Gary S. Leiserowitz, MD

Professor, Chief, Division of Gynecologic Oncology, Division of Gynecologic Oncology, Department ofObstetrics and Gynecology, University of California, Davis Medical Center, Sacramento, California

The management of adnexal masses during pregnancy can be challenging for the patient andthe clinician. The specter of a possible malignancy can sway the decision for interventionversus expectant management. The etiologies of ovarian masses are reflective of the patient’sage; and, therefore, benign entities such as functional ovarian cysts, benign cystic teratomas,and serous cystadenomas predominate. In the unusual cases when cancer is present, they aretypically germ cell and borderline ovarian tumors, and are commonly low stage and low grade.Ultrasound is the primary modality used to detect ovarian masses and to assess the risk ofmalignancy. Morphologic criteria more accurately identify benign cysts compared with malig-nant tumors. Tumor markers are used primarily to monitor disease status after treatment ratherthan establish the ovarian tumor diagnosis as a result of lack of specificity, because severalmarkers can be elevated inherent to the pregnancy itself (eg, CA-125, �-hCG). Expectantmanagement is recommended for most pregnant patients with asymptomatic, nonsuspiciouscystic ovarian masses. Surgical intervention during pregnancy is indicated for large and/orsymptomatic tumors and those that appear highly suspicious for malignancy on imaging tests.The extent of surgery depends on the intraoperative diagnosis of a benign versus a malignanttumor. Conservative surgery is appropriate for benign masses and borderline ovarian tumors.More aggressive surgery is indicated for ovarian malignancies, including surgical staging.Although rarely necessary, chemotherapy has been used during pregnancy with minimal fetaltoxicity in patients with advanced-stage ovarian cancer in which the risk of maternal mortalityoutweighs the fetal consequences.

Target Audience: Obstetricians & Gynecologists, Family PhysiciansLearning Objectives: After completion of this article, the reader should be able to recall the prevalence

of ovarian tumors during pregnancy; explain the risk of malignancy, the use of diagnostic testing, andmanagement options; and summarize potential maternal and fetal outcomes.

With the advent of nearly routine prenatal ultrasound(US), the detection of ovarian masses has become com-monplace. The presence of adnexal masses in preg-

nancy is not unusual with an estimated incidencebetween 1% and 2% (1–3). In the era before routineprenatal US, adnexal masses were discovered eitherincidentally on physical examination or as a result ofsymptoms. In either case, detection would tend to favormasses that were problematic as a result of either size orcomplications. This historic mindset tended to push forsurgical intervention to avoid the risks of rupture, tor-sion, or obstructed labor and as well as a concern forpossible ovarian malignancy. However, a contemporary

The author has disclosed that he has no financial relationshipswith or interests in any commercial companies pertaining to thiseducational activity.

Lippincott Continuing Medical Education Institute, Inc. hasidentified and resolved all faculty conflicts of interest regardingthis educational activity.

Reprint requests to: Gary S. Leiserowitz, MD, 4860 Y Street, Suite2500, Sacramento, CA 95817. E-mail: gsleiserowitz@ucdavis.edu.

CME REVIEWARTICLEVolume 61, Number 7OBSTETRICAL AND GYNECOLOGICAL SURVEY

Copyright © 2006by Lippincott Williams & Wilkins 19

463

understanding of the benign nature and uncomplicatedcourse of adnexal masses diagnosed incidentally by UShas led to a more conservative approach to managementof adnexal masses in pregnancy.

The scope of this review discusses the etiology ofadnexal masses, including the risk of malignancy,diagnostic testing, management options during preg-nancy, maternal and fetal outcomes, and manage-ment of ovarian cancer during pregnancy.

ETIOLOGY

Just like in the nonpregnant state, adnexal masseshave various gynecologic and nongynecologic causes.However, most are gynecologic and are usually ovarianor uterine in origin. Table 1 lists the most commontypes of ovarian tumors encountered during pregnancy.Functional cysts are particularly common in pregnancy,including theca-lutein cysts, which are usually related topresence of gestational trophoblastic disease (4). Therelative frequency of all other nonfunctional ovariantumors is largely a reflection of the patient’s age ratherthan something intrinsic to the pregnancy itself.

In descending order, the most common ovarian tu-mors found during pregnancy include: functional ovar-ian cysts (follicular, corpus luteum, and theca-lutein),benign cystic teratomas, serous cystadenomas, parao-varian cysts, mucinous cystadenomas, endometriomas,and malignant tumors (1,3,5–8). Surgery (excision orbiopsy) followed by a pathologic examination is re-quired to definitively determine the etiology of theadnexal mass. Consequently, the frequency of tumorsthat either resolve spontaneously (ie, functional cysts)or are managed nonsurgically tends to be underre-ported. Similarly, the timing of detection of the adnexalmass during pregnancy influences the likely etiology ofthe mass. Cystic adnexal masses less than 5 cm that aredetected in the first 16 weeks are usually functional andalmost always resolve spontaneously (2). Ovarian tu-

mors that persist beyond 16 weeks are more likely to beneoplastic; and, therefore, are more likely to result insurgical intervention. This issue is further discussedsubsequently in the section on management of adnexalmasses.

Ovarian malignancy is estimated to occur in approx-imately 2% to 3% of the masses identified during preg-nancy (1,3,5–7,9,10), but the frequency reported in thedifferent case series is quite variable from study tostudy. Ovarian tumors of low malignant potential (LMPor borderline ovarian tumors) are usually included inthe category of ovarian malignancies, although theirbiologic aggressiveness is usually bland and indolent.The frequency of various ovarian malignancies ishighly age-dependent and reflects the younger cohort ofreproductive-age women compared with postmeno-pausal women who are more commonly diagnosed withepithelial ovarian cancer.

We recently investigated the frequency of ovarianmasses during pregnancy and analyzed the risk ofovarian malignancy during pregnancy (11). We com-bined several California databases for our analysis,including the California Vital Statistics database, theCalifornia Office of Statewide Health Planning andDevelopment (OSHPD), and the California CancerRegistry (CCR). We identified 4,846,505 obstetricpatients from the vital statistical database in the pe-riod of 1991 to 1999. Nine thousand three hundredseventy-five pregnant women were identified with ahospital discharge diagnosis that included any type ofovarian tumor (0.19% of all obstetric deliveries).These included 8267 ovarian cysts, 5910 benignovarian neoplasms, 144 ovarian LMP tumors, 174ovarian malignancies, and 19 secondary ovarian ma-lignancies (the total was greater than 9375, becausesome patients had more than one diagnosis). His-topathologic confirmation was obtained in only 117LMP and 89 ovarian invasive cancers (see Table 2),resulting in a conservatively estimated occurrence

TABLE 1Etiology of ovarian tumors during pregnancy

Benign Tumors Malignant Tumors

Functional cysts Borderline ovarian tumorsFollicular cysts Malignant epithelial tumorsCorpus luteum cysts Malignant germ cell tumorsTheca-lutein cysts Sex cord/stromal tumorsHemorrhagic cysts Granulosa cell tumorParaovarian cysts Cancer metastatic to the ovaryBenign cystic teratomas Krukenberg tumorsSerous cystadenoma Pseudomyxoma peritoneiMucinous cystadenomaEndometriomas

TABLE 2Histology of ovarian malignancies occurring in California womenduring pregnancy, 1991–1999

Histology CancerLow MalignantPotential Tumor

Serous (including papillary) 14 83Mucinous 10 34Endometrioid 5Clear cell 3Other epithelial 14Pseudomyxoma peritonei 8Germ cell 34Granulosa cell 1Total 89 117

464 Obstetrical and Gynecological Survey

rate of ovarian malignancy in ovarian masses duringpregnancy of 89 of 9375 (0.93%). This is lower thanthe 3% figure commonly quoted in the literature. TheCalifornia statistics suggest an occurrence of onemalignancy for every 54,644 deliveries.

Notably, the majority of the ovarian malignanciesare actually LMP tumors. Ninety-five percent of theLMP tumors and 83.8% of the ovarian cancers arestage I. Nearly 40% of the ovarian cancers are germcell tumors, which is typical for younger women whoare in their reproductive years. Interestingly, 8 wereclassified as pseudomyxoma peritonei, which arenow generally believed to be metastatic mucinoustumors of gastrointestinal (usually appendiceal) ori-gin. Although incompletely reported in the CCR da-tabase, low-grade (grades 1 and 2) tumors were morecommon in the cohort of pregnant women than in themore typical group of other women with ovariancancer.

The finding that most ovarian cancers are low stageand low grade is consistent with previous reports(12,13). Most germ cell tumors are dysgerminomas,which are predominantly low stage (1,12). Conse-quently, the overall prognosis of pregnant patientswith either an LMP tumor or an ovarian cancer ishighly favorable. These findings should strongly in-fluence the management of adnexal masses duringpregnancy.

DIAGNOSTIC EVALUATION

The vast majority of adnexal masses during preg-nancy will be identified by US either as an incidentalfinding or to evaluate a symptomatic pelvic mass.The safety of diagnostic US as well as its ability todelineate anatomic relationships and characterize themorphology of pelvic masses makes it ideal as theprimary initial evaluation tool. A variety of algo-rithms exist to differentiate between benign and ma-lignant ovarian tumors, or to stratify the risk ofmalignancy, using elements such as tumor size, mor-phology, and color flow Doppler. These algorithmshave a high specificity to identify benign ovarianmasses, but are less specific when used to distinguishbenign from malignant when applied to complexovarian masses.

Marino et al (3) used a modification of a weightedscoring system of ovarian tumors developed by Le-rner et al (14) to determine the risk of malignancy(Table 3). Using Lerner’s scoring system against acohort of 302 benign masses, 31 malignant masses,and one LMP tumor, they found a positive predictivevalue of 29.4% and a negative predictive value of

99.6% (14). Thus, it is easier to make a confidentdiagnosis of a benign tumor than a malignant one,because benign and malignant neoplasms often sharesimilar complex morphologic features.

Some authors feel that various ovarian tumor typesare sufficiently characteristic on US that they couldaccurately predict the histology. Bromley and Ben-acerraf (15) found that they could accurately identify95% of dermoid cysts, 80% of endometriomas, and71% of simple cysts seen during pregnancy. Zanettaet al (16) used US criteria to develop the followingcategories: simple cyst, endometriosis or corpusluteum-like, dermoid-like, complex benign, borderline-like, and suspicious. They used their system to triagepregnant patients with adnexal masses into those whowarranted surgical intervention (suspicious masses)versus those who were candidates for expectant man-agement (all the other categories). In the limitednumber of patients who had adnexal surgery eitherantepartum or postpartum, the final pathology gen-erally matched well with their US diagnoses.

Use of color flow Doppler has been offered asanother tool to help differentiate between a benign ormalignant ovarian tumor. Unfortunately, there is suf-ficient overlap in blood flow patterns such that thefalse-positive rate is nearly 50% (17), which offers noadvantage over use of US morphology indexing alone.

Magnetic resonance imaging (MRI) can be safelyused during pregnancy to evaluate adnexal masses(3). The primary advantages of MRI relate to itscapacity to develop 3-dimensional planar images,delineate tissue planes, and characterize tissue com-position. This is particularly helpful in the pelviswhere US has a limited role in assessing the bony andmuscular structures. For example, Weinreb et alfound that MRI was able to identify leiomyomata,bowel loops, and an abdominal pregnancy; differen-tiate between a solid mass and hemorrhagic fluid-filled cyst; and evaluate the parametria in a patientwith cervical cancer (18). Keir et al (19) found in

TABLE 3Risk of ovarian malignancy based on sonographic criteria

Risk of Ovarian Cancer Sonographic Criteria

Low Cystic, unilocularSize �5 cm

Intermediate Cystic, multilocularComplexThin septations

High Solid massNodulesThick septationsSize �5 cm

Managing Ovarian Masses During Pregnancy Y CME Review Article 465

their limited series of pelvic masses during preg-nancy that MRI correctly identified the etiology in 17of 17 (100%) pelvic masses, whereas US was accu-rate in 12 of 17 (71%). Although MRI can providevaluable diagnostic information beyond the ability ofUS, the number of situations in which this is clini-cally important is limited. Both modalities are verydependent on the experience of the physician whointerprets the scan. MRI probably adds minimally tothe evaluation of most ovarian masses compared withUS but may be valuable when the US diagnosis isuncertain and when a radiologist experienced in in-terpreting MRI of adnexal masses and pregnancy isavailable.

Serum tumor markers are primarily used for surveil-lance of known, treated ovarian malignancies but are ofvariable benefit in the initial assessment of ovarianmasses. CA-125 is elevated in 80% of epithelial ovarianmalignancies with mucinous adenocarcinomas being anotable exception. CA-125 has limited diagnostic accu-racy in premenopausal women, because multiple benigngynecologic conditions are associated with elevated val-ues such as menses, uterine fibroids, and especiallypregnancy (20). When elevated, the CA-125 can pro-vide a baseline value before treatment of ovarian cancerbut will not help differentiate between benign and ma-lignant masses during pregnancy. Various other tumormarkers are used to monitor germ cell tumors (AFP—endodermal sinus tumor, �-hCG—choriocarcinoma,lactic dehydrogenase—dysgerminoma) (21). Althoughgerm cell tumors are among the most common ovarianmalignancies seen in pregnancy, both �-hCG and AFPhave very limited use as tumor markers during preg-nancy. Tumor markers should be obtained before anysurgical intervention when there is a suspicion ofovarian malignancy to provide a baseline should amalignancy be diagnosed. Any elevation in thetumor markers should be considered in conjunc-tion with the results of the imaging tests to avoidunnecessary intervention when possible.

MANAGEMENT OPTIONS

Conservative

The main consideration in choosing interventionversus expectant management centers on the risks tothe mother and fetus. The specter of malignancy isquickly raised whenever an ovarian mass is detected.However, a rational decision should be based on anaccurate assessment of the malignancy risk. As notedpreviously, ovarian masses in pregnancy have a can-cer risk that ranges between 0.9% and 3%. More

importantly, sonographic criteria can successfullystratify the risk. For example, those few masses thathave suspicious complex features can be separatedfrom the majority that is usually benign. Therefore,treatment considerations are best divided into thosemasses that warrant conservative (expectant) versussurgical management.

Most ovarian cysts discovered during pregnancywill resolve spontaneously and/or require no inter-vention. In their series of over 18,000 obstetric USscans, Bernhard (2) found that 76% of 432 masseswere simple cysts less than 5 cm in diameter. Theydid not intervene in this group, and there were nocomplications. The remaining 24% (102) of themasses were either complex and/or greater than 5cm. Nearly 69% of these masses also resolved spon-taneously. Hogston (22) and colleagues evaluated26,110 prenatal US examinations and detected 137asymptomatic ovarian cysts. One hundred twenty-threepatients were treated conservatively, but 3 requiredsubsequent intervention during pregnancy as a result ofpain. Ninety-six percent of the remaining 120 womenhad repeat scans, and 89% of these had complete res-olution, including 82% of the cysts �6 cm.

Zanetta (16) used their sonographic criteria to iden-tify a group of women with nonsuspicious ovarianmasses who were candidates for expectant manage-ment during pregnancy (see previously for details).Only cysts �3 cm were included in their study. In6636 US scans, they found 82 cysts in 79 women. Nomasses suspicious for malignancy were identified.Three cysts required surgery for torsion. The remain-ing 79 were followed expectantly during the preg-nancy. They noted either complete disappearance or areduction of �50% in 69% of the simple cysts, 77% ofthe endometriosis-like cysts, 57% of the complexbenign cysts, and none of either the dermoid-like orborderline-like cysts (54.5% total). Thirty-one massespersisted after pregnancy and 19 patients had surgerywith a variety of histologies, including 3 borderlinetumors (all of which were identified by their UScriteria), but no ovarian cancer. There were also nopregnancy losses. Thus, observation of the major-ity of ovarian masses appears to be a viable man-agement option.

Surgical

Ovarian masses that warrant intervention usuallyhave at least one of the following indications: 1) astrong suspicion of malignancy and/or large size(�8–10 cm), 2) symptomatic complaints, or 3) anincreased risk of torsion/rupture/obstruction of labor.

466 Obstetrical and Gynecological Survey

The likelihood of malignancy with a complex ovar-ian mass is relatively low (29.4% in Lerner’s study[14]), but increases if there are other associated find-ings such as ascites or omental thickening. Histori-cally, many authors have emphasized the risks oftorsion, cyst rupture, and obstruction of labor, butthese usually occurred in large, symptomatic masses.Struyk et al (8) reported their experience with ovar-ian tumors in pregnancy and noted torsion in 12%,rupture in 9%, and obstruction of labor in 17%. InWhitecar’s series (7), 16 of 130 women underwenturgent laparotomy for acute abdominal pain, 14 be-fore delivery. Eleven of the 16 either had docu-mented or presumptive torsion (11 of 130 [8.5%]).Bromley and Benacerraf in their series of 131 ovar-ian masses in 125 pregnant women reported a farlower rate of antepartum problems (15). Only onepatient had an ovarian torsion and one patient wasexplored for an ovarian cyst that was not found atlaparotomy. Bernhard reported a 1% risk of torsionin their series of 102 ovarian masses (2) and thatone occurred in a patient with a palpable mass.Overall, it appears that later studies report lowerrisks of torsion and rupture than earlier studies.This probably reflects a higher proportion ofasymptomatic, US-detected ovarian masses thatare less prone to complicate the pregnancy.

The choice of laparotomy versus laparoscopy isdependent on the risks of malignancy, the urgencyof the procedure, and the skills of the surgeon. Assurgeons gain training and experience, there hasbeen acceleration in the frequency of laparoscopicoperations during pregnancy. Laparoscopic sur-gery has been commonly reported for treatment ofappendicitis and cholecystitis during pregnancywith generally excellent results and minimal riskof fetal loss and preterm delivery (23,24). Lapa-roscopy for adnexal masses has become increas-ingly standard management for benign ovarianmasses in nonpregnant women (25), and it hasbeen used in selected cases of ovarian cancer (26).Consequently, it has also been adopted in manage-ment of some pregnant patients with adnexal tu-mors (24,27–31). The presumptive benefits oflaparoscopy in pregnancy include a minimally in-vasive approach with decreased recuperative timeand risk of fetal loss/preterm delivery comparedwith laparotomy (23). The most recent case seriesdemonstrate that experienced laparoscopic sur-geons are able to manage a spectrum of adnexalpathologies, including ovarian cysts, adnexal tor-sion, heterotopic pregnancy, and bleeding ovariancysts (29–31).

Patients who potentially benefit the most by laparo-scopic surgery of the adnexal mass should fit the fol-lowing criteria: first or second trimester of pregnancyand an ovarian mass that is not suspicious for malig-nancy. Sound clinical judgment is critical for patientselection and is clearly tempered by the surgeon’s skilland experience. Caution is strongly advised when con-sidering laparoscopic management of possible ovariancancer. Port site recurrences are noted in 2.3% of pa-tients treated laparoscopically for their malignancy in arecent review (32), although these were most commonin patients with primary peritoneal cancer and recurrentcancer. Ovarian masses, especially suspicious ones,must be removed intact when possible. Although itremains controversial (25), spillage or rupture of amalignant ovarian cyst was associated with decreasedsurvival in a recent study (33). The risk of adverse fetaloutcomes is not eliminated with a minimally invasiveapproach. Soriano reported that the rates of spontaneousabortion were 12.8% (5 of 39) in the first trimester and8% (2 of 25) in the second trimester, although allthese miscarriages occurred in women with ovar-ian torsion (30).

Patients with very large ovarian masses fall into 2groups: those with large but simple unilocular cystsand those with complex cysts. In both groups, con-sideration can be given to expectant managementwith surgical intervention reserved for symptoms re-sulting from possible torsion or rupture or if the massrisks obstructing vaginal delivery. Alternatively,multiple case series report that aspiration of simpleunilocular cysts can avoid the need for major surgeryand provide symptomatic relief (16,34,35). However,aspiration of a complex ovarian cyst runs the poten-tial risk of malignant fluid spillage. Surgical inter-vention for large complex ovarian masses should beby laparotomy because these masses will not fit intoendoscopic bags.

Whether by laparoscopy or laparotomy, consider-ation can be given to ovarian cystectomy if the UScriteria for a benign mass are met. Otherwise, oopho-rectomy is appropriate. If there is a risk of disruptinga corpus luteum cyst up to 12 weeks gestation, thenprogesterone support is indicated.

Surgical management of ovarian cancer is dis-cussed separately subsequently.

MATERNAL AND FETAL OUTCOMES

The adverse consequences to mother and fetus areprimarily a result of complications from the ovarianmass and/or the interventions for the mass. If anovarian malignancy is present, then there are also

Managing Ovarian Masses During Pregnancy Y CME Review Article 467

risks of the cancer and the consequences of its treat-ment as well. In reviewing the literature, it is oftendifficult to determine if the adverse effects were theresult of the adnexal mass, the treatment of the mass,or unrelated (eg, spontaneous abortion of fetus withmultiple anomalies in a patient with an ovarianmass). Nevertheless, surgical intervention for benignadnexal masses in pregnancy is associated with ahigher risk of preterm deliveries and low neonatalbirth weights compared with those patients who didnot have surgery (11).

Pain is the most common symptom in pregnantpatients with adnexal masses (26% in the Struykstudy [8]). This ranges from mild (which can bemanaged expectantly) to severe (requiring emergentlaparotomy). The etiology of the pain is usually tor-sion, although ovarian rupture also occurs. Whitecarreported in his series that nearly half of the patientswith acute abdominal pain required emergency lap-arotomy for ovarian masses and uterine leiomyomas(7). The rate of torsion is quite variable in manyseries, from �1% to 22% (5,15). Rupture appears tobe less common, ranging from 0% to 9% (7,8,15).Obstruction of labor is also reported to occur in 2%to 17% of patients (8,10). Other less frequent prob-lems include bleeding and infection. Struyk noted therelationship between tumor size and the risk of com-plications as 35% for tumors between 5 and 6 cm indiameter and up to 85% for larger tumors (8). How-ever, no other authors reported such a high maternalcomplication rate. Observational US studies by Bern-hard and Zanetta report far lower complication ratesresulting from problems of torsion and obstruction oflabor (2,16).

Adverse fetal outcomes are most commonly theresult of an abdominal catastrophe from ovarian tor-sion or rupture associated with abdominal surgery. Inmany cases, the relationship of poor fetal outcomesto the adnexal mass is not apparent. Elective surgicalintervention is preferably timed for the second tri-mester in which the risk of subsequent fetal loss isminimized (3). Whitecar found that adverse preg-nancy outcomes, including preterm deliveries andfetal loss, were significantly less frequent if laparot-omy occurred before 23 weeks gestational age (oddsratio � 0.15, P � .005) (7). The effectiveness oftocolytics for suppression of preterm delivery is un-clear. In Whitecar’s series, tocolytics were adminis-tered in 13 patients who had surgeries in the secondand third trimesters. Six of 13 had preterm de-liveries, although only 2 occurred within 2 weeksof laparotomy.

MANAGEMENT OF OVARIAN CANCERDURING PREGNANCY

Both ovarian malignancies and LMP ovarian tu-mors should be surgically managed as in the non-pregnant patient. If there is a preoperative suspicionof malignancy, then a surgeon who is both knowl-edgeable and capable should be available to completethe standard surgical staging that includes peritonealwashings, peritoneal biopsies, omentectomy, pluspelvic and paraaortic lymphadenectomy (1). Thereare situations in which complete surgical staging isnot feasible (eg, large gravid uterus that obscures thesurgical field, no qualified surgeon available), andthen the subsequent treatment must be based onincomplete information. Fortunately, the majority ofboth ovarian malignancies and borderline tumors arelow grade and low stage, which may allow for de-finitive surgical staging to be completed either at thetime of cesarean section or postpartum.

Conservative, fertility-sparing surgery is appropriatewhen the malignancy is apparently unilateral. Survivaldoes not appear to be compromised by sparing thenormal contralateral ovary (36,37). Borderline ovariantumors are candidates for either unilateral salpin-gooophorectomy or even ovarian cystectomy. The re-currence rates for borderline tumors are higher withovarian cystectomy, but almost all recurrences are sal-vageable with further surgery (38,39). In the unusualcase in which gross metastatic disease is present, thenaggressive surgical debulking of the extraovarian dis-ease is indicated. A decision regarding sparing of theintrauterine pregnancy is based on gestational age. Inthe first trimester, sacrifice of the pregnancy may be thebest choice, because exposure to subsequent chemo-therapy may be teratogenic. In the second and thirdtrimesters, preservation of the pregnancy is gener-ally recommended because limited clinical experi-ence has failed to demonstrate an adverse fetaleffect with chemotherapy given during the preg-nancy (see subsequently).

In all cases, use of expert frozen section pathologyis required for critical decision-making. If there isany doubt regarding the frozen section pathologicdiagnosis, then it is best to defer definitive surgicaltreatment until the final pathological report, espe-cially if the tumor appears limited to a single ovary.Fortunately, the accuracy of frozen section diagnosisof ovarian tumors is over 90%, with one study re-porting an overdiagnosis (false-positive) rate of 2.2%and an underdiagnosis (false-negative) rate of 5.4%(40). The frozen section accuracy rates are best for

468 Obstetrical and Gynecological Survey

benign tumors followed by malignant and then bor-derline tumors.

The maternal and fetal outcomes associated withovarian cancer are a special situation. The adverse con-sequences of the symptomatic malignant ovarian neo-plasm and surgical intervention are applicable to theprevious discussion. There are also the unique risks ofthe malignancy and its treatment to the mother andfetus. In our own series derived from the CCR (11), thefollowing maternal outcomes were statistically morefrequent in patients with ovarian cancer and LMP tu-mors compared with noncancer cases: cesarean section,blood transfusions, hysterectomy, postpartum stay �5days, and total hospital charges �$10,000. Interest-ingly, neonatal outcomes were not adversely affectedby the presence of the ovarian neoplasm, including lowbirth weight, prematurity, neonatal death, readmissionto the hospital, prolonged hospital stay, and excessivehospital charges. The risk of maternal death wasstrongly associated with the timing of diagnosis. Themortality rate of ovarian cancer was zero if the diagno-sis was made 9 to 12 months before delivery, 5.6% ifmade zero to 9 months before delivery, 6.3% if made atdelivery, and 18.5% if made zero to 12 months afterdelivery.

The need for adjuvant chemotherapy after a diag-nosis of ovarian cancer is based on the tumor stage,histologic type, and grade. The potential benefits ofcancer control versus the risks to the mother andfetus must be carefully balanced. In general, chemo-therapy should not be delayed to allow for deliverywhen the risk of cancer progression threatens mater-nal survival. The findings of extraovarian metastasesand/or highly aggressive germ cell tumors usuallyrequire urgent administration of chemotherapy.However, patients with low-staged malignancieswith intermediate risk factors for recurrence (such asadverse histology or high-grade tumors) might bemanaged conservatively and then given chemother-apy after delivery.

The risks of chemotherapy during pregnancy arewell described (41,42). Although almost all chemo-therapeutic agents are toxic to animals, the risks tothe developing fetus appear to be less ominous ex-cept for certain classes such as the antimetabolites(42). The U.S. Food and Drug Administration assignsrisk categories to drug use during pregnancy, and mostchemotherapeutic agents are placed in the C, D, or Xcategories (12). Chemotherapy should be avoided dur-ing the first trimester, because the teratogenic risks offetal malformation are the greatest (43). In thesecond and third trimesters, impairment of fetalgrowth and functional development are a greater

concern than malformation. There does not appearto be a greater risk of adverse fetal effects frommultiagent versus single agent chemotherapy (43).

Use of chemotherapy to treat ovarian cancer duringpregnancy has limited experience. Ebert noted only 11cases reported from 1983 to 1995 of chemotherapygiven to patients with ovarian cancer (42). Chemother-apy use has been reported in the treatment of both germcell and epithelial ovarian malignancies. Cytotoxicagents reportedly used for ovarian cancer during preg-nancy include cyclophosphamide, doxorubicin, vincris-tine, bleomycin, etoposide, cisplatin, carboplatin, andpaclitaxel (41,44,45). In epithelial ovarian cancer, therehas been more experience with cisplatin than carbopla-tin and only 2 case reports of paclitaxel use. Perinatalcomplications included preterm delivery, gesta-tional diabetes, and neonatal hyaline membranedisease/respiratory distress syndrome. Long-terminfant outcomes (such as neurologic and physicaldevelopment) appear to be good for those fetusesexposed to chemotherapy in utero (41). Therefore,when there are strong indications for chemother-apy use during pregnancy (in the second and thirdtrimesters), it should not be withheld as a result ofconcerns of fetal effects.

Overall, the management of pregnant patients witha malignant ovarian neoplasm is similar to what isrecommended in the nonpregnant state. The primarydifference lies in considering adjustments in the sur-gical and/or chemotherapy treatment to allow forfetal viability if the patient desires this. A limiteddelay in the timing of definitive surgical resection orchemotherapy until after delivery is unlikely to resultin a worse prognosis unless the patient has obviousmetastatic disease. In the setting of aggressive can-cer, however, consideration can be given to pretermdelivery and/or use of chemotherapy during preg-nancy if the maternal life is threatened.

REFERENCES

1. Goff BA, Paley PJ, Koh W-J, et al. Cancer in the pregnantpatient. In: Hoskins WJ, Perez CA, Young RC, eds. Principlesand Practice of Gynecologic Oncology, 3rd ed. Philadelphia:Lippincott Williams & Wilkins, 2000:501–528.

2. Bernhard LM, Klebba PK, Gray DL, et al. Predictors of persis-tence of adnexal masses in pregnancy. Obstet Gynecol 1999;93:585–589.

3. Marino T, Craigo SD. Managing adnexal masses in preg-nancy. Contemp Obstet Gynecol 2000;45:130–143.

4. Montz FJ, Schlaerth JB, Morrow CP. The natural history oftheca lutein cysts. Obstet Gynecol 1988;72:247–251.

5. Hermans RH, Fischer DC, van der Putten HW, et al. Adnexalmasses in pregnancy. Onkologie 2003;26:167–172.

6. Usui R, Minakami H, Kosuge S, et al. A retrospective survey ofclinical, pathologic, and prognostic features of adnexal

Managing Ovarian Masses During Pregnancy Y CME Review Article 469

masses operated on during pregnancy. J Obstet GynaecolRes 2000;26:89–93.

7. Whitecar MP, Turner S, Higby MK. Adnexal masses in preg-nancy: a review of 130 cases undergoing surgical manage-ment. Am J Obstet Gynecol 1999;181:19–24.

8. Struyk AP, Treffers PE. Ovarian tumors in pregnancy. ActaObstet Gynecol Scand 1984;63:421–424.

9. Agarwal N, Parul, Kriplani A, et al. Management and outcomeof pregnancies complicated with adnexal masses. Arch Gy-necol Obstet 2003;267:148–152.

10. Ueda M, Ueki M. Ovarian tumors associated with pregnancy.Int J Gynaecol Obstet 1996;55:59–65.

11. Leiserowitz GS, Xing G, Cress R, et al. Adnexal masses inpregnancy: how often are they malignant? Gynecol Oncol2005.

12. Boulay R, Podczaski E. Ovarian cancer complicating preg-nancy. Obstet Gynecol Clin North Am 1998;25:385–399.

13. Dgani R, Shoham Z, Atar E, et al. Ovarian carcinoma duringpregnancy: a study of 23 cases in Israel between the years1960 and 1984. Gynecol Oncol 1989;33:326–331.

14. Lerner JP, Timor-Tritsch IE, Federman A, et al. Transvaginalultrasonographic characterization of ovarian masses with animproved, weighted scoring system. Am J Obstet Gynecol1994;170:81–85.

15. Bromley B, Benacerraf B. Adnexal masses during pregnancy:accuracy of sonographic diagnosis and outcome. J Ultra-sound Med 1997;16:447–452; quiz 453–454.

16. Zanetta G, Mariani E, Lissoni A, et al. A prospective study ofthe role of ultrasound in the management of adnexal massesin pregnancy. BJOG 2003;110:578–583.

17. Wheeler TC, Fleischer AC. Complex adnexal mass in preg-nancy: predictive value of color Doppler sonography. J Ultra-sound Med 1997;16:425–428.

18. Weinreb JC, Brown CE, Lowe TW, et al. Pelvic masses inpregnant patients: MR and US imaging. Radiology 1986;159:717–724.

19. Kier R, McCarthy SM, Scoutt LM, et al. Pelvic masses inpregnancy: MR imaging. Radiology 1990;176:709–713.

20. Look KY. Epidemiology, etiology, and screening of ovariancancer. In: Rubin SC, Sutton GP, eds. Ovarian Cancer, 2nded. Philadelphia: Lippincott Williams & Wilkins, 2001:167–180.

21. Hurteau JA, Williams SJ. Ovarian germ cell tumors. In: RubinSC, Sutton GP, eds. Ovarian Cancer, 2nd ed. Philadelphia:Lippincott, Williams & Wilkins, 2001:371–382.

22. Hogston P, Lilford RJ. Ultrasound study of ovarian cysts inpregnancy: prevalence and significance. BJOG 1986;93:625–628.

23. Al-Fozan H, Tulandi T. Safety and risks of laparoscopy inpregnancy. Curr Opin Obstet Gynecol 2002;14:375–379.

24. Fatum M, Rojansky N. Laparoscopic surgery during preg-nancy. Obstet Gynecol Surv 2001;56:50–59.

25. Canis M, Rabischong B, Houlle C, et al. Laparoscopic man-agement of adnexal masses: a gold standard? Curr OpinObstet Gynecol 2002;14:423–428.

26. Tozzi R, Kohler C, Ferrara A, et al. Laparoscopic treatment ofearly ovarian cancer: surgical and survival outcomes. GynecolOncol 2004;93:199–203.

27. Neiswender LL, Toub DB. Laparoscopic excision of pelvicmasses during pregnancy. J Am Assoc Gynecol Laparosc1997;4:269–272.

28. Nezhat F, Nezhat C, Silfen SL, et al. Laparoscopic ovariancystectomy during pregnancy. J Laparoendosc Surg 1991;1:161–164.

29. Stepp KJ, Tulikangas PK, Goldberg JM, et al. Laparoscopy foradnexal masses in the second trimester of pregnancy. J AmAssoc Gynecol Laparosc 2003;10:55–59.

30. Soriano D, Yefet Y, Seidman DS, et al. Laparoscopy versuslaparotomy in the management of adnexal masses duringpregnancy. Fertil Steril 1999;71:955–960.

31. Mathevet P, Nessah K, Dargent D, et al. Laparoscopic man-agement of adnexal masses in pregnancy: a case series. EurJ Obstet Gynecol Reprod Biol 2003;108:217–222.

32. Nagarsheth NP, Rahaman J, Cohen CJ, et al. The incidence ofport-site metastases in gynecologic cancers. Jsls 2004;8:133–139.

33. Vergote I, De Brabanter J, Fyles A, et al. Prognostic impor-tance of degree of differentiation and cyst rupture in stage Iinvasive epithelial ovarian carcinoma. Lancet 2001;357:176–182.

34. Khaw KT, Walker WJ. Ultrasound guided fine needle aspira-tion of ovarian cysts: diagnosis and treatment in pregnant andnon-pregnant women. Clin Radiol 1990;41:105–108.

35. Platek DN, Henderson CE, Goldberg GL. The management ofa persistent adnexal mass in pregnancy. Am J Obstet Gynecol1995;173:1236–1240.

36. Schilder JM, Thompson AM, DePriest PD, et al. Outcome ofreproductive age women with stage IA or IC invasive epithelialovarian cancer treated with fertility-sparing therapy. GynecolOncol 2002;87:1–7.

37. Morice P, Camatte S, Wicart-Poque F, et al. Results of con-servative management of epithelial malignant and borderlineovarian tumours. Hum Reprod Update 2003;9:185–192.

38. Zanetta G, Rota S, Chiari S, et al. Behavior of borderlinetumors with particular interest to persistence, recurrence, andprogression to invasive carcinoma: a prospective study. J ClinOncol 2001;19:2658–2664.

39. Donnez J, Munschke A, Berliere M, et al. Safety of conserva-tive management and fertility outcome in women with border-line tumors of the ovary. Fertil Steril 2003;79:1216–1221.

40. Gol M, Baloglu A, Yigit S, et al. Accuracy of frozen sectiondiagnosis in ovarian tumors: is there a change in the course oftime? Int J Gynecol Cancer 2003;13:593–597.

41. Cardonick E, Iacobucci A. Use of chemotherapy during hu-man pregnancy. Lancet Oncol 2004;5:283–291.

42. Ebert U, Loffler H, Kirch W. Cytotoxic therapy and pregnancy.Pharmacol Ther 1997;74:207–220.

43. Doll DC, Ringenberg QS, Yarbro JW. Antineoplastic agentsand pregnancy. Semin Oncol 1989;16:337–346.

44. Mendez LE, Mueller A, Salom E, et al. Paclitaxel and carbo-platin chemotherapy administered during pregnancy for ad-vanced epithelial ovarian cancer. Obstet Gynecol 2003;102:1200–1202.

45. Sood AK, Shahin MS, Sorosky JI. Paclitaxel and platinumchemotherapy for ovarian carcinoma during pregnancy. Gy-necol Oncol 2001;83:599–600.

470 Obstetrical and Gynecological Survey