Contemporary multidisciplinary treatment of pregnancy-associated breast cancer

It is estimated that cancer complicates as many as 1 in 1000 pregnancies, with breast cancer as the most common associated malignancy (Smith et al. 2001). Contemporary management of pregnancy-associated breast cancer (PABC) encourages continuation of pregnancy during treatment. Limitations in published data describing the safety of cancer therapy in pregnancy create challenges and dilemmas in clinical decision-making for providers caring for these patients. Prospective evaluation of the safety of antineoplastic therapy during pregnancy is methodologically difficult; therefore, most published data consist of case and cohort series (Mir et al. 2008; Cardonick et al. 2012a; Loibl et al. 2012). Given the apparent rising incidence of PABC, perhaps due to trends in delaying pregnancy (Andersson et al. 2009; Litton & Theriault 2010), a greater understanding of the pathologic features, treatment options, and maternal and fetal outcomes in PABC is critical.

PABC has been defined as breast cancer diagnosed during pregnancy and up to one year postpartum. Although much PABC research has considered this group as a whole, it is worth considering two distinct subsets – those diagnosed during pregnancy, and those diagnosed in the postpartum period (Borges & Schedin 2011). Physicians caring for a pregnant patient with active breast cancer must select treatment options which maximize benefits to the patient while minimizing risks to the developing fetus. Given the advances in contemporary management of breast cancer, including use of taxanes, “dose dense” schedules, and sentinel lymph node biopsy, it is imperative to understand whether or not these new treatments can be safely administered during pregnancy.

With these considerations, a cohort of breast cancer patients treated during pregnancy was identified at an academic comprehensive cancer center. This study examined the distribution of tumor subtypes, treatment decisions, and maternal and fetal outcomes within the cohort, and evaluated the contemporary management of PABC supported by a multidisciplinary team.

A retrospective cohort of patients diagnosed with breast cancer during pregnancy was created through a search of the Dana Farber/Harvard Cancer Center (DF/HCC) clinical database. Eligibility criteria included: a history of pathologically confirmed breast cancer diagnosed during pregnancy, at least two follow-up visits for treatment documented within the DF/HCC system (including the Dana-Farber Cancer Institute, Massachusetts General Hospital, Brigham and Women’s Hospital, and Beth Israel Deaconess Medical Center) to identify continuous care, and age > 18 years. Medical oncologists, breast surgeons, and maternal-fetal medicine specialists within the DF/HCC institutions were also queried to identify eligible patients treated in their practices. Comprehensive review of the DF/HCC database was used to gather patient information as well as data on short-term maternal and fetal outcomes. Statistical analysis was exploratory and descriptive in nature. Institutional Review Board approval was obtained prior to initiation of this study.

In this cohort of patients with PABC, contemporary therapies for breast cancer including “dose-dense” scheduling, taxane chemotherapy, growth factor support, immediate breast reconstruction, and sentinel lymph node biopsy, were utilized without significant rates of adverse maternal or fetal events. There was a low rate of maternal complications, and the overall rate of fetal abnormalities, 4.4%, was consistent with expected population rates (Parker et al. 2010). This series is one of the most modern cohorts of PABC patients in the current literature, with the majority of patients receiving their diagnosis during or after the year 2005.

Consistent with prior reports, and likely due to the young age of the patients, this PABC cohort was enriched for ER-negative tumors as well as BRCA mutations (Middleton et al. 2003; Elledge et al. 1993). However, patients’ clinicopathologic details did not differ greatly from those noted in other studies of non-pregnant young women diagnosed with breast cancer (Collins et al. 2012).

Almost all patients in the cohort underwent imaging to stage their disease while pregnant. Notably, nearly half of patients underwent mammography and over 10% underwent MRI scanning, with no apparent adverse effect on the fetus. The safety and feasibility of mammography during pregnancy have been reported (Kopans 1998; Yang et al. 2006), and it is unclear why rates of mammography were not higher in this population. The findings from this study support the safety of mammography during pregnancy; and suggest that MRI may have a role in select cases (Taylor et al. 2011).

Reports suggest that mastectomy or lumpectomy can be safely pursued at any point during a pregnancy with minimal risk to the fetus (Rosenkranz & Lucci 2005; Amant et al. 2010). However, first trimester surgery is often deferred, out of concern for higher risks of fetal complications (Moran et al. 2007). In this cohort, nearly one in five patients underwent surgery in the first trimester without apparent complications, suggesting the safety of surgical intervention during this time. Additionally, both sentinel lymph node biopsy and immediate reconstruction were utilized without apparent complication. Overall, very few surgical complications were observed, consistent with previous observations (Dominici et al. 2010). These data support the safety of contemporary surgical treatment paradigms in PABC, as long as practitioners are aware of the unique physiologic changes associated with pregnancy (Hill & Pickinpaugh 2008).

Although chemotherapy administration during the first trimester has been associated with risk of fetal defects, the use of certain chemotherapeutics in the second and third trimesters is generally considered safe without increase in the rates of fetal abnormalities (Loibl et al. 2012; Amant et al. 2010; Cardonick & Iacobucci 2004; Ring et al. 2005). The chemotherapy data presented in this study support the safety of administering traditional as well as more contemporary treatment regimens in this population. Dose-dense scheduling has an important role in modern adjuvant breast cancer treatment, however is not typically offered to women during pregnancy due to concerns regarding potential harm to the fetal bone marrow. A prior study of 10 pregnant patients treated with dose-dense AC found no difference in maternal or neonatal outcomes when compared with every three week dosing (Cardonick et al. 2012b). In this study, in the patients who received AC, including the significant minority who received a “dose-dense” schedule, there were no apparent short-term adverse effects on offspring.

Most prior studies evaluating chemotherapy in PABC have examined AC or FAC, but few studies have examined the safety of taxanes, which play an integral role in adjuvant chemotherapy regimens. In this cohort, 8.3% of patients who received AC also received 12 cycles of weekly paclitaxel during pregnancy without evident undesirable effects. To date, the use of taxanes in PABC has been reported primarily as case reports (Potluri et al. 2006; Nieto et al. 2006). In a recent review of this literature, it was noted that of 42 infants born to 40 patients exposed to taxanes during pregnancy, only one had a malformation possibly related to taxane use (Mir et al. 2010). Further expansion of this data is desired to confirm the safety of taxane-containing chemotherapy regimens after the first trimester.

In non-pregnant women, growth factors may be required during adjuvant chemotherapy, particularly with “dose-dense” schedules. There are few reports in the literature of patients receiving growth factors during pregnancy (Dale et al. 2003). Consistent with the prior publications, 6 patients in this study received growth factor support without apparent adverse effect on the fetus.

Although neither tamoxifen nor trastuzumab has been rigorously studied in pregnant patients, these important systemic agents are traditionally avoided in pregnancy out of concern for teratogenic effects. Case reports have described multiple fetal complications after prenatal exposures to these agents (Gottschalk et al. 2011; Braems et al. 2011; Witzel et al. 2008). Trastuzumab has been associated with oligohydramnios, and a boxed warning for teratogenicity has been issued, contraindicating the use of trastuzumab in pregnancy. As a short-term delay in the initiation of endocrine therapy likely does not adversely affect survival outcomes, particularly in early stage breast cancer, it is feasible to consider delayed initiation of tamoxifen until completion of pregnancy. However, given the growth kinetics of HER2 positive breast cancer, a delay in HER2-directed therapy may be more clinically significant and introduce potential risk of compromise in the care of these patients. Fortunately, current adjuvant regimens for HER2 positive breast cancer can begin with anthracycline-based chemotherapy, facilitating postponement of trastuzumab therapy until after delivery. Analysis of unintentional pregnancies during adjuvant trastuzumab clinical trials is in process (Azim et al. 2012), and an ongoing prospective study (MotHER pregnancy registry) is collecting data on pregnancies complicated by trastuzumab exposure.

In general, increased rates of maternal complications of pregnancy were not observed in this cohort, possibly reflecting early inclusion of a maternal-fetal medicine specialist in the multidisciplinary team. A notable finding was the observation that over half of deliveries occurred at < 37 weeks, and were considered preterm. In contrast, the current expected rate of preterm delivery in the United States is 12.2% (Martin et al. 2011). With the exception of one series suggesting rates of preterm delivery comparable to general population norms (Cardonick et al. 2010), most other published studies have also reported an increased risk of preterm delivery in patients treated with chemotherapy for PABC (Smith et al. 2001; Loibl et al. 2012; Van Calsteren et al. 2010). In this series, the majority of preterm deliveries were planned to facilitate maternal therapy, e.g. to allow continuation of chemotherapy. Only 16.2% experienced unplanned preterm delivery, a figure more consistent with national averages for non-PABC patients.

The question of planned premature delivery deserves examination. Although preterm delivery in this cohort did not appear to result in increased harm to the fetus, recent obstetrical data has suggested even late preterm delivery (34–37 weeks) or near term delivery (37–39 weeks) can be associated with adverse outcomes in the fetus (Hibbard et al. 2010; Shapiro-Mendoza et al. 2008). In addition, a recent evaluation of children exposed to chemotherapy in utero has noted lower cognitive development for those born preterm than for those born full term (Amant et al. 2012). Given the apparent safety of chemotherapy during pregnancy, especially in the third trimester, the risks of late preterm delivery must be carefully considered when planning the timing of delivery. In these situations, a careful discussion is encouraged between the patient, the maternal-fetal medicine specialist, and the medical oncologist so that delivery timing minimizes risks to both mother and fetus.

Strengths of this study include utilization of a large academic comprehensive cancer center, and evaluation of a sizeable number of pregnant breast cancer patients treated with more contemporary treatment regimens than have been previously published. In addition, a systematic approach was used to assemble the cohort rather than solely relying on physician recall or patient self-enrollment, eliminating a source of bias. Finally, the majority of patients were treated for their breast cancer and delivered their infants at the same institution, allowing for data capture for peripartum as well as maternal outcomes.

A limitation of this study is the lack of follow-up data in exposed infants; data is limited to in utero and perinatal outcomes. Obtaining retrospective follow-up data is complicated by loss to follow-up and limitations of self-reported data. However, prospective long-term follow-up studies have generally demonstrated favorable outcomes in exposed offspring (Aviles & Neri 2001; Hahn et al. 2006). Another limitation of this study is sample size; while this cohort is large compared to other published cohorts of PABC, the sample size is small in absolute terms, making it difficult to generate definitive conclusions. In addition, the study was limited to women treated in a large academic tertiary care center, which introduces the possibility of selection bias. Data on socioeconomic status was not available, but it is possible that a higher degree of social and financial support may have impacted not only treatment choices but also patient outcomes.

This retrospective study of PABC demonstrates the feasibility of optimal treatment through multidisciplinary care team management. Continued prospective data collection and case series publication may improve comfort with using contemporary treatment paradigms to care for women in the difficult situation of having been diagnosed with breast cancer while pregnant. When making treatment decisions for these patients, attention to multidisciplinary team management is necessary to appropriately balance perinatal care of mother and infant. Further data collection and investigation will improve understanding of and treatment paradigms for breast cancer diagnosed during pregnancy.