Selected prognostic variables for mammographic parenchymal patterns.

Abstract

Conjecture exists about the influence of numerous risk-factors for breast cancer on mammographic parenchymal patterns. To allow more precise documentation of the common variables considered influential in alterations of breast parenchyma, we conducted a randomized retrospective analysis. Of 10,132 women participants in the Louisville Breast Cancer Detection Demonstration Project, every tenth participant was randomly selected for evaluation using SPSS statistical programming. Each accessioned patient had discriminant analysis for the risk factors of age, parity, age at birth of first child, family history, personal history, previous history of breast biopsy, and exogenous estrogen therapy. One-thousand-and-two women were examined for the significance of the selected prognostic variable association with Wolfe mammographic parenchymal patterns (WMPP). Each prognostic factor was tested by chi-square analysis for the low-risk pattern (N1P1) versus the high-risk pattern (P2DY). A high correlation existed between the age of patient and WMPP (P = 0.0002) in the subjects evaluated (50--85 years, mean 60). Similarly, a very significant correlation was evident between WMPP and parity (P = 0.0002), age at birth of first child (P = 0.0014), family history of breast cancer (P = 0.097), and history of previous breast biopsy (P = 0.0066). Little correlation existed between the Wolfe parenchymal pattern classification and a personal history of breast cancer (P = 0.7779) or the use of exogenous estrogens (P = 0.5776).