BackgroundA rare subtype of breast cancer, atypical medullary carcinoma of the breast (AMCB), shows a highly adverse prognosis compared to medullary carcinoma of the breast (MBC). The current study aimed to establish a correlated nomogram for the identification of the prognostic factors of AMCB and MBC.MethodsKaplan–Meier and Cox regression analyses were applied to data acquired from the Surveillance, Epidemiology and End Results (SEER) database for 2004 to 2013 to analyse tumour characteristics and overall survival. Propensity score matching (PSM) analysis was performed to determine the overall survival (OS) among those with AMCB and MBC. A predictive nomogram was created, and the concordance index (C-index) was used to predict accuracy and discriminative ability.ResultsA total of 2,001 patients from the SEER database were diagnosed with MBC between 2004 and 2013, including 147 patients diagnosed with AMCB. The number of diagnoses gradually increased in both groups. Cox analysis of multivariate and Kaplan–Meier analysis showed that older age (HR = 3.005, 95% CI 1.906–4.739) and later stage were significantly associated with poor prognosis, while cancer-directed surgery was an independent protective factor (HR = 0.252, 95% CI 0.086–0.740). In the HR-negative stratification analysis, older age (HR = 2.476, 95% CI 1.398–4.385), later stage and histological type (HR=0.381, 95% CI 0.198-0.734) were found to be independent prognostic factors for low standard survival. The log-rank analysis demonstrated significantly worse prognostic factors for patients with AMCB than for those with MBC (P = 0.004). A nomogram (C-index for survival = 0.75; 95% CI 0.69–0.81) was established from four independent prognostic factors after complete identification.ConclusionsMBC is rare, and cancer-directed surgery, older age, and later stage are independently linked with prognosis. In the HR negative population, AMCB patients show a worse survival gain than those with MBC.