Breast carcinoma is one of the most common carcinomas in pregnant women. The incidence of breast carcinoma may increase in the future because of the trend toward delayed childbearing and increased screening. However, very few contemporary studies have attempted to identify the combined histopathologic and immunohistochemical features of breast carcinoma in these patients.
The authors evaluated 39 patients with breast carcinoma occurring coincident with pregnancy. This was comprised of a critical histologic review and immunohistochemical evaluation to determine the status of prognostic and predictive markers including estrogen receptor (ER), progesterone receptor (PR), HER-2/neu, Ki-67, and p53.
The mean age at presentation was 33 years (range, 24–44 years). Densities and/or masses were noted on mammograms in 14 of 16 patients with available radiographic information. The primary tumors were a mean of 4.5 cm in greatest dimension (range, 0.1–13.5 cm). Two of the 39 patients had clinical (American Joint Committee on Cancer) Stage I disease, 19 patients had Stage II disease, 16 had Stage III disease, and 2 patients had Stage IV disease at the time of presentation. Histologically, high-grade invasive ductal carcinomas were found in 32 of 38 patients. The primary tumor was not available for review in one patient. A predominantly solid pattern of growth was observed in nine patients. Lymphovascular invasion was identified in 61% of cases. Ductal carcinoma in situ was identified in 72% of tumors and was high grade in all cases. Of the 25 patients tested, ER positivity was found in 7 patients, PR positivity was found in 6 patients, HER-2/neu positivity was found in 7 patients, and p53 positivity was found in 12 patients. The proliferation rate as shown by Ki-67 staining was high in 60% of the cases. Follow-up information was available for 35 patients and the mean follow-up period was 43 months (range, 2–163 months). Distant metastasis occurred in seven patients. The mean time to disease recurrence was 20.4 months (range, 10–33 months). Of 35 patients, 4 have died, 22 were alive with no evidence of disease, and 9 were alive with disease at the last follow-up. The remaining four patients died of unknown causes.
The prevalence of breast carcinoma occurring during pregnancy is increasing as women are delaying the onset of childbearing. Currently, breast carcinoma during pregnancy is diagnosed in approximately 1 in 3000 pregnancies and breast carcinoma occurring during pregnancy accounts for approximately 2% of all new cases of breast carcinoma.1 Breast carcinoma is the most common cancer associated with pregnancy and its diagnosis during pregnancy can potentially have a devastating effect not only on the mother, but also on the unborn child. The risk of pregnancy-associated breast carcinoma is age related. Women who have their first term pregnancy after the age of 30 years old have a two to three times higher risk of developing breast carcinoma than women who have their first pregnancy before the age of 20 years old.2
Breast carcinoma is optimally treated when diagnosed early. However, breast carcinoma diagnosed during pregnancy usually presents at an advanced stage. In spite of multiple opportunities for clinical breast examinations arising from the increased frequency of physician visits, examination of the breast during pregnancy is hampered by hypertrophy, engorgement, and indistinct nodularity of the gland. Moreover, densities and nodularities in the breast of pregnant women are often overlooked or ascribed to benign proliferative changes. The hyperestrogenic state of pregnancy may also contribute to the development and rapid growth of breast carcinoma in these women. The current study attempted to determine whether distinctive pathologic factors in breast carcinoma coincident with pregnancy might contribute to the more advanced stage of cancer at presentation.
MATERIALS AND METHODS
The current study included 39 patients whose breast carcinoma was detected during pregnancy and one on post-partum day 12. These patients were treated at The M. D. Anderson Cancer Center between 1986 and 2001. Either the breast carcinoma was diagnosed at the institution or the pathology was reviewed at the institution before treatment. Most patients belonged to a cohort of women who were participating in an institutionally approved protocol in which informed consent was obtained. The patients received systemic chemotherapy consisting of 5-fluorouracil (5-FU), doxorubicin, and cyclophosphamide for a median of four cycles.3 In all cases, glass slides were reviewed and histologic parameters were assessed. Histologic grade was determined according to the modification of the Bloom Richardson scoring system.4 Using that system, Grade 3 corresponded to poorly differentiated tumors, Grade 2 corresponded to moderately well-differentiated tumors, and Grade 1 corresponded to well differentiated tumors. Each case was assessed carefully for the presence of ductal carcinoma in situ (DCIS) and lymphovascular invasion.
Formalin-fixed, paraffin-embedded tissue blocks of tumor specimens were available for immunohistochemical analysis in 25 cases. Immunohistochemical staining was performed on tissue sections using the methodologies recommended by the manufacturers of the antibodies being used based on the method of Hsu et al.5
Antibodies used were directed against the estrogen receptor (ER; 1D5, Zymed Laboratories, San Francisco, CA), against the progesterone receptor (PR; 1A6, Novacastra Laboratories, Newcastle, UK), against HER-2/neu (AB8, Neomarkers, Fremont, CA), against Ki-67 (Dako, Carpinteria, CA), and against p53 (D07, Dako). Slides were counterstained with Mayer's hematoxylin, dehydrated, and mounted in mounting media. To evaluate the specificity of antibodies, known positive and negative tissue specimens were used as controls.
Immunoreactivity for ER and PR was positive if greater than 10% of the tumor cell nuclei stained. Immunoreactivity for HER-2/neu was positive if moderate to strong complete membranous staining was observed in greater than 10% of the tumor cells. Immunoreactivity for Ki-67 was categorized by the percentage of nuclear staining: low (< 10%), moderate (10–35%), and high (> 35%). Immunoreactivity for p53 was positive if greater than 5% of the tumor cell nuclei were stained.
The patients' medical records were also reviewed retrospectively to determine the patients' age and the gestational age of the fetus at the time of the initial diagnosis, pertinent history and clinical presentation, radiographic results, clinical stage, and follow-up.
The 39 patients ranged in age from 24 to 44 years (mean and median age, 33 years). The gestational age of the fetuses at the time of initial presentation ranged from 2 to 37 weeks (mean gestational age, 18 weeks; median gestational age, 20 weeks). Thirty-two patients self-discovered the tumors as palpable masses, one patient presented with a bloody nipple discharge, and one patient presented with a diffuse erythema (inflammatory carcinoma). Abnormalities were observed on mammograms for 14 of 16 patients and breast ultrasonography confirmed tumor masses in all 21 patients on whom it was performed. At presentation, 2 patients were classified as having American Joint Committee on Cancer clinical Stage I disease, 19 were classified as having Stage II disease, 16 were classified as having Stage III disease, and 2 were classified as having Stage IV disease (Table 1).6
Table 1. Clinical Stage of Disease in Pregnant Women at Presentation
AJCC: American Joint Committee on Cancer. cancer staging manual, 5th ed.
Of the 39 patients, 14 patients had an initial diagnosis of breast carcinoma after a core needle biopsy of the breast mass and 8 patients underwent a fine-needle aspiration biopsy (FNAB) as the initial diagnostic procedure. Surgeons performed a lumpectomy in 9 of 35 patients without a previous needle biopsy. In addition, 4 of 39 patients had an initial needle biopsy or aspiration of an ipsilateral axillary lymph node.
Thirty of the 39 patients had definitive surgical treatment (lumpectomy or mastectomy). Axillary lymph node clearance was performed in 29 of 39 patients. Of the remaining 10 patients, 3 underwent diagnostic needle aspirations of lymph nodes and 7 did not have lymph node sampling.
Most of the women were treated during their second and third trimester with 5-FU, doxorubicin, and cyclophosphamide for a median of four cycles with minimal complications. Follow-up ranged from 2 to 163 months (mean follow-up, 43.7 months; median follow-up, 32 months). Two patients had distant metastasis at the time of initial presentation and distant metastasis subsequently developed in an additional seven patients. The sites of subsequent metastasis included the liver (four patients), brain (three patients), bones (three patients), ipsilateral axillary lymph nodes (one patient), and nonaxillary lymph nodes (three patients). A second carcinoma developed in the opposite breast in one patient 1 year after the initial diagnosis. At their last follow-up, 22 of 39 patients (56%) were alive with no evidence of disease, 9 (23%) were still alive with disease, and 4 (10%) had died of disease. The other 4 patients (10%) died of unknown causes (Table 2).
Table 2. Clinical Follow-Up of Pregnant Women with Breast Carcinoma
Clinical status at last follow-up
Alive with no evidence of disease
Alive with disease
Died of disease
Died of unknown causes
The primary tumors ranged from 0.1 to 13.5 cm in diameter (mean, 4.5 cm; median, 4 cm). Thirty-seven patients presented with a solitary tumor mass, 1 patient had multifocal disease with 3 separate foci of invasive ductal carcinoma (IDC), and 1 patient had extensive multicentric tumors that involved all 4 quadrants of the breast.
Histologically, all of the tumors were IDC. The invasive carcinoma in 16% of the patients was moderately to well differentiated and the other 84% of patients had poorly differenitiated invasive carcinomas (Fig. 1). In the majority of cases, the architectural patterns of the invasive carcinoma was comprised of solid cell nests, sheets, and cords of tumor cells infiltrating within a desmoplastic stroma. Glandular formation was relatively infrequent, and when present, it was focal. Histologic evaluation of the tumor in nine cases revealed a syncytial-sheeting pattern of growth with prominent areas of necrosis (Fig. 2). One patient had a tumor composed almost entirely of high-grade DCIS with multiple foci of microinvasive carcinoma (each focus measuring ≤ 1 mm).
One patient had focal spindle cell differentiation of the tumor cells and one patient with a poorly differentiated ductal carcinoma had areas of squamous differentiation. Secondary Paget involvement of the nipple was present in one patient.
DCIS was present in 72% of the tumor specimens. Of the 26 patients who had an excisional biopsy, the DCIS was focal in 9 (35%) and was prominent in 17 (65%). In all patients, the DCIS was a high nuclear grade and the architectural patterns were solid and/or cribriform with luminal necrosis. Lymphovascular invasion was present in 61% of cases. Four of these cases had extensive tumor emboli within lymphovascular spaces and three cases had tumor emboli in the dermal lymphatics.
Axillary lymphadenectomies were performed in 29 of the 39 patients. Of these 29 patients, 23 (79 %) had lymph node metastasis. The number of lymph nodes evaluated ranged from 7 to 34 (mean, 15 lymph nodes; median, 14 lymph nodes) and the number of lymph nodes containing metastatic carcinoma ranged from 1 to 19. The size of the lymph node metastasis ranged from microscopic foci to 2.5 cm. In six patients, extracapsular extension of the tumor was confirmed. FNABs were performed on the lymph nodes of three patients, two of whom had metastatic carcinoma. The FNAB results were negative in one patient with palpable lymph nodes. However, a concurrent FNAB of the liver showed metastatic carcinoma.
Of the 25 patients for whom immunohistochemical results were available, ER immunoreactivity was noted in 7 (28%) and PR immunoreactivity was observed in 6 (24%). Both ER and PR positivity were identified in only 4 of the 25 patients (16%). Of the remaining 14 patients, 3 (21%) were positive for ER according to the referring hospital pathology reports. Similarly, 3 of the 14 patients (21%) were positive for PR. Staining for the Her-2/neu oncoprotein in 7 of the 25 patients (28%) revealed 2+ and 3+ positive membranous staining. The remaining 18 patients were negative for HER-2/neu. Positive p53 immunostaining of the tumor nuclei was observed in 12 of the 25 patients (48%). In 15 of 25 cases (60%), the proliferation rate as determined by Ki-67 staining was high (> 35% of the tumor nuclei stained positive), the proliferation rate was moderate (10–35% of tumor nuclei stained positive) in 4 cases (16%), and the proliferation rate was low (< 10% of the tumor nuclei stained positive) in the remaining 6 cases (24%) (Table 3).
Table 3. Immunohistochemical Analysis of Breast Carcinoma Occurring during Pregnancy
No. positive (%)
ER: estrogen receptor; PR: progesterone receptor.
ER and PR
The incidence of breast carcinoma occurring in association with pregnancy ranges from 0.2% to 4.0%.1, 7, 8 Of particular concern, breast carcinoma occurring in the physiologic setting of pregnancy tends to be diagnosed at an advanced stage. Several factors may contribute to this. First, it is difficult to evaluate tumors in the gravid female breast because of engorgement secondary to hormonal changes. In addition, obstetricians frequently direct their attention to the developing fetus rather than perform a comprehensive physical examination. These factors often contribute to a delay in diagnosis and advanced disease presentation. The prognosis and response to endocrine treatment are often predicated on the ER and PR content of the breast tumor. In the current study evaluating the tumor histology and the prognostic and predictive markers ER, PR, HER-2/neu, p53, and Ki-67 in breast carcinomas occurring during pregnancy, we found that the steroid receptor content, HER-2/neu status, proliferation rate, and the presence of the p53 mutation did not differ significantly from those of age-matched, nonpregnant women.
Similar to our findings, most studies evaluating breast carcinoma occurring in pregnant women reported that the predominant histology is IDC.9–11 Invasive lobular carcinoma (ILC) and other special histologic types are infrequent. For example, Parente et al.9 reported that only one of the eight patients in their series had ILC. This was also observed by Tobon and Horowitz10 who reported 1 of 14 patients with ILC and King et al.11 who reported 2 of 63 pregnant patients with ILC. ILC is also an uncommon histologic finding in the tumors of young women.12 In further agreement with the observations of others studying breast carcinoma in pregnancy,11, 13 the majority of tumors in our series were high-grade tumors. Similar to the results of Ishida et al.,8 lymphovascular invasion was a frequent observation in our series.
Several researchers have reported an increased tumor nuclear grade in all women younger than 40 with breast carcinoma.12, 14–16 In a review of 1869 breast carcinoma patients, Fisher et al.12 found a statistically significant difference in the number of nuclear Grade 3 carcinomas between unselected women younger than 40 and older women (P < 0.0001). In the same study, ILC was observed more frequently in women 70 years old or older. In an evaluation of breast carcinoma in women age 35 years or younger, Rosen et al.16 found a high incidence of poorly differentiated tumors (53%) in their young cohort of patients. Kollias et al.14 reported similar findings in an evaluation of 2897 women with breast carcinoma. In their study, high nuclear grade and lymphovascular invasion were observed more frequently in women younger than 35 years of age than in older women.
Seventy-nine percent of the patients in the current study presented with metastatic carcinoma to the ipsilateral axillary lymph nodes. This high percentage of lymph node involvement is similar to that observed in pregnant patients by Holleb and Farrow17 (72% lymph node involvement) and King et al.11 (62% lymph node involvement). In fact, the incidence of lymph node involvement in pregnant and lactating women has been variably reported to be between 56% and 89%.8, 9, 11, 18–20 Such lymph node involvement can signal clinical aggressiveness and can be an indicator of poor prognosis. However, this finding is not specific to pregnant women with breast carcinoma. In general, a higher percentage of young patients present with lymph node metastasis.13, 21–23
Only 28% of the women in our series had ER-positive tumors and 24% had PR-positive tumors. These percentages are slightly lower that those found in young women in general with breast carcinoma. Maru et al.24 found that 45% of the young women studied had ER-positive tumors and 36% had PR-positive tumors. ER positivity is strongly associated with age at diagnosis and is more prevalent in postmenopausal women.
Elledge et al.25 evaluated the histology of breast carcinoma occurring in pregnant patients. Immunohistochemical analysis showed that 6 of 12 (50%) patients were ER positive, 10 of 12 (83%) were PR positive, and 7 of 12 (58%) were Her-2/neu positive. In their study, hormonal receptors were detected by immunohistochemistry significantly more often than a ligand-binding assay on the same tumor specimens. The researchers postulated that circulating steroid receptors may interfere with the ligand-binding assay and immunohistochemistry may thus be more reliable in this setting.
Previous studies have evaluated ER and PR in pregnant women using ligand-binding assays. Tobon and Horowitz10 found that the assay for ER was negative or of low density in all seven cases studied whereas the assay for PR was of low density in two cases and positive in five cases. The clinicopathologic and immunohistochemical features of breast carcinoma occurring during pregnancy were studied in a case–control multiinstitutional study from Japan.8 Using the dextran-coated charcoal method of examination, the researchers found that 30% of patients with breast carcinoma diagnosed during pregnancy were ER positive and 29% were PR positive. These rates were less than one-half of the rates observed in the age-matched, nonpregnant control group (61% were ER positive and 60% were PR positive) and were statistically significant.
In a retrospective, multiinstitutional study from France, Bonnier et al.26 evaluated ER and PR status in the tumors of pregnant women using both a radioligand-binding assay (performed by using dextran-coated charcoal) and an enzyme immunoassay (EIA). Using the radioligand-binding assay, they found that ER and PR levels were significantly lower in pregnant women than in age-matched nonpregnant women. However, when the EIA was performed, the difference in hormonal receptor status between pregnant and nonpregnant, aged-matched women was not statistically significant. In their study, the EIA demonstrated that 46.7% of pregnant women were ER positive and 34.2% were PR positive. These results indicate that in this group of patients, immunohistochemical studies may yield the most reliable findings in terms of ER and PR status.
Shousha13 reported interesting results when evaluating HER-2/neu (c-erb-B2) overexpression by immunohistochemistry. In that study, all six tumors that occurred during pregnancy or lactation were negative, whereas all four tumors that occurred shortly after delivery or cessation of lactation were positive. The author hypothesized that HER-2/neu overexpression is suppressed in breast tumors occurring during pregnancy and lactation, but not immediately after. In the current study, 28% of the pregnant patients had tumors that expressed the HER-2/neu oncoprotein. This result, although less than the 58% observed by Elledge et al.25, is similar to that found in the general population.27
The p53 tumor suppressor protein is a negative regulator of cell growth and plays an important role in the control of apoptosis. The 48% immunoreactivity for p53 that we observed in the current series of pregnant patients is similar to the 52% found in young women by Maru et al.24 The high proliferation rate observed in 60% of the tumors in the current series of pregnant women is also similar to the proliferation rate described in breast tumors occurring in young patients.24
When current immunohistochemical methodologies are used to determine ER and PR hormonal status, the results are similar in pregnant and nonpregnant, age-matched women. It is possible that the ligand-binding assay historically used to determine ER and PR positivity is less accurate in the hormonal milieu of pregnancy because of the competitive blocking of the receptor sites resulting from hyperestrogenemia and hyperprogesteronemia.25
The findings of the current study regarding the histologic features of breast carcinoma during pregnancy are similar to the findings observed by other investigators studying breast carcinoma in young women.12, 15, 16, 18 The immunohistochemical profile that we observed in women with breast carcinoma occurring during pregnancy (low expression of ER and PR and high expression of Ki-67 and p53) and the findings of other investigators using similar methodologies25 show that the prognostic and predictive markers do not differ significantly between pregnant and age-matched, nonpregnant women. It may be the age of the breast carcinoma patient, not the pregnancy itself that affects the biologic features of the tumor. The outcome for the pregnant breast carcinoma patient is likely affected by the stage at presentation, which may be influenced by a delay in diagnosis and treatment as well as the more aggressive biologic characteristics observed among young women with breast carcinoma.