Incidence and Survival of Pregnancy Associated Breast Cancer: A Contribution from the IMI ConcePTION Project Piloted in Finland and Norway

Background: Pregnancy-associated breast cancer (PABC) diagnosed during pregnancy or within one year postpartum has poorer prognosis than breast cancer diagnosed at another time (non-PABC). To what extent this is attributable to adverse prognostic factors is uncertain. We developed methods to assess medication exposure in hospital setting in secondary use of health data, and to handle missing data in multinational studies to better understand worse survival of PABC patients.

Objectives: To study incidence and survival of PABC and non-PABC in women aged 15 to 54 years in Finland (FI) and Norway (NO).

Methods: Data from cancer, patient, prescription, birth and cause of death registries were transformed into the ConcePTION common data model. Breast cancers identified from cancer registries (FI 1953-2019, NO 2004-2020) were linked to data from birth registry. Valid pregnancies had a gestational age between 22 and 44 weeks and ended in a live or stillbirth (FI 1996-2018, n=1 295 995; NO 2004-2020, n=1 043 141). From multiple deliveries, we chose the one closest to the cancer diagnosis. We calculated the age standardized incidence rates per 100 000 person-years for PABC and non-PABC, and incidence of PABC per 10 000 deliveries. We applied multiple imputation to improve measurement of clinical prognostic factors with highly varying proportion of missing data [range from 0.7% to 89.5%]. We analyzed overall survival using Cox proportional hazard regression. The crude and adjusted risk estimates were combined using random effect meta-analysis.

Results: Study material included 26 651 breast cancer cases with 5632 deaths from Finland and 16 471 breast cancer cases with 2064 deaths from Norway. The incidence of PABC was 12.2 per 105 person-years in Finland and 11.4 per 105 in Norway. The incidence of non-PABC during the study period was 1058.4 per 105 person-years in Finland and 720.2 per 105 in Norway. PABC occurred from 2.2 to 2.5 per 10 000 deliveries in these countries. Roughly one third of PABC cases were diagnosed during pregnancy and about two thirds postpartum in both countries. PABC patients had higher risk of death compared to non-PABC patients (crude HR 2.08; 95% CI 1.65-2.59). Following adjustment for prognostic factors, the hazard ratios for PABC patients were attenuated but remained elevated (fully adjusted HR 1.30; 95% CI 1.07-1.57). Breast cancers diagnosed postpartum had somewhat poorer prognosis than those diagnosed during pregnancy.

Conclusions: Missing data reduces study power and may lead to survival estimates that are substantially different from those that would have been obtained had no data been missing. Patients with PABC are at an increased risk of death which is largely, but not fully, explained by more adverse tumor characteristics.