Background Breast density and mammographic tumor features of breast cancer may carry prognostic information. The potential benefit of using the combined information obtained from breast density, mammographic tumor features, and pathological tumor characteristics has not been extensively studied. Purpose To investigate how mammographic tumor features relate to breast density and pathological tumor characteristics. Material and Methods This retrospective study was carried out within the Malmö Diet and Cancer Study: a population-based cohort study recruiting 17,035 women during 1991–1996. A total of 826 incident breast cancers were identified during follow-up. Mammography images were collected and analyzed according to breast density and tumor features at diagnosis. Pathological data were retrieved from medical reports. Mammographic tumor features in relation to invasiveness, tumor size, and axillary lymph node involvement were analyzed using logistic regression yielding odds ratios (OR) with 95% confidence intervals (CI) and adjusted for age at diagnosis, mode of detection, and breast density. Results Tumors presenting as an ill-defined mass or calcifications were more common in dense breasts than tumors presenting as a distinct mass or with spiculated appearance. Invasive cancer was more common in tumors with spiculated appearance than tumors presenting as a distinct mass (adjusted OR, 5.68 [1.81–17.84]). Among invasive tumors, an ill-defined mass was more often large (>20 mm) compared with a distinct mass, (adjusted OR, 3.16 [1.80–5.55]). Conclusion Tumors presenting as an ill-defined mass or calcifications were more common in dense breasts. Spiculated appearance was related to invasiveness, and ill-defined mass to larger tumor size, regardless of mode of detection and breast density. The potential role of mammographic tumor features in clinical decision-making warrants further investigation.