Pregnancy and breast cancer : risk patterns, tumour characteristics and prognosis

Abstract

Pregnancy-associated breast cancer (PABC) is commonly defined as a breast cancer (BC) diagnosed during pregnancy or within one or two years after delivery. The risk of BC is lower than expected during pregnancy with a small rebound within 2 years after delivery, compared to non-pregnant women of the same age. Although a rare event, there is evidence that women with PABC have a poorer prognosis compared to women with BC diagnosed not near a pregnancy (non-PABC). It is important to improve the understanding of underlying mechanisms and identify factors that may influence outcome, including the role of detection and tumour biology. The possible influence of pregnancy exposures on detection and tumour biology may be time-varying, with different effects during the pregnancy period compared to time windows further away from delivery. Since a transient risk has been reported 5-10 years post-delivery, it is also of interest to compare pregnancy-associated tumours to tumours diagnosed further away, say up to ten years, from delivery with respect to tumour biology and prognosis.

Having a relative with breast cancer is an established risk factor for breast cancer, in particular at a young age. Women with a family history of breast cancer may be more likely to have pre-malignant breast cells at a young age, and could be more susceptible to pregnancy-related exposures, such as elevated levels of endogenous hormones. Study III assessed the risk pattern for breast cancer during pregnancy and up to ten years after delivery in relation to family history of breast cancer using a cohort of 3.5 million women aged 15-44 years identified in the Swedish Multi-Generation Register and the Swedish Cancer Register. In total, 15,548 women had a recorded BC diagnosis of which 1,208 were PABC. To reduce the comparison population and improve computational efficiency, a case-cohort design was utilised. The results showed that a family history of BC did not affect the risk pattern around pregnancy and up to 10 years post-delivery, indicating that pregnancy-related exposures do not interact with the familial pre-disposition for BC in these women.

Results from several studies have suggested that BC tumours diagnosed around delivery have adverse tumour characteristics, such as advanced TNM stage and triple-negative subtype, but it remains unclear if adverse tumour characteristics can fully explain the worse prognosis in women with PABC. Studies II and IV assessed TNM stage in cohorts of patients diagnosed with BC during pregnancy and up to ten years post delivery using information from the Swedish Cancer Register for years 2002-2009 (study II) and the Swedish Breast Cancer Quality Register for years 1992-2009 (study IV). Study IV also assessed histological grade and tumour biology, such as hormone receptor status. The results showed that tumours detected during pregnancy and within one year of delivery were more often advanced, of high grade and hormone receptor negative (triple-negative and non-luminal Her2 positive subtypes) compared to nulliparous women. These associations were most pronounced in tumours diagnosed during the first six months post-delivery. In combination with the previously observed lower risk during pregnancy, these findings could reflect diagnostic delays, resulting in more advanced tumours at diagnosis, or a suppression of hormone responsive tumours around delivery.

With the aim to investigate prognosis in women with PABC, two partly over-lapping cohorts of BC patients were identified in the Swedish Cancer Register for years 1963-2002 (study I) and in the Breast Cancer Quality Register for years 1992-2009 (study IV), including 1,110 and 778 women with PABC, respectively. The association between PABC and mortality was analysed using Cox regression and flexible parametric models. The results from study I and IV corroborated previous findings that women with PABC have a poorer prognosis compared to women with non-PABC. The mortality was highest in women diagnosed 0-6 months after delivery, and remained elevated up to eight years after diagnosis, in comparison to women diagnosed not near a delivery. However, following adjustment for tumour characteristics, the difference in mortality between women with PABC and women with non-PABC was attenuated and non-significant (study IV). These findings suggest that although PABC is associated with more advanced and aggressive tumours, the outcome is similar to that in women with non-PABC when taking tumour characteristics into account. For comparison, we also assessed tumour characteristics and prognosis in women diagnosed with BC 2-10 years after delivery (study IV). A transient risk of BC around 5-10 years after delivery has been reported previously. Compared to nulliparous women, women diagnosed 2-5 years post-delivery did not differ in tumour size, grade and ER status, but more often presented with spread to lymph nodes, negative PR status and non-luminal Her2 positive subtype. Women with BC diagnosed 2-5 years post-delivery also exhibited a poorer prognosis compared to nulliparous women, which remained significant after adjustment for T/N stage and hormone receptor status. These findings could indicate different mechanisms for risk, detection and prognosis in women diagnosed 2-5 years after delivery compared to women with PABC.

In conclusion, the findings indicate that pregnant and lactating women exhibit different patterns of BC risk, tumour biology and prognosis compared to other pre-menopausal women with BC. The poorer prognosis observed in women with PABC can largely be explained by adverse tumour characteristics at diagnosis.