R E Tarone1. 1. Biostatistics Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892.
Abstract
BACKGROUND: An unexpected excess of patients with breast cancer with four or more positive lymph nodes was observed in mammographically screened women who were younger than the age of 50 years at enrollment into the Canadian National Breast Screening Study (NBSS). It has been suggested that this excess is consistent with prior screening trials evaluating mammography. METHODS: A quantitative evaluation of the distribution of patients with breast cancer with four or more positive lymph nodes in the NBSS was undertaken, and the percentages of patients with breast cancer who were at an advanced state at diagnosis in the NBSS and in previous randomized screening trials were compared. The validity of mortality analyses after eliminating advanced cases detected by physical examination at the initial screening visit is examined. RESULTS: The excess of patients with cancer with four or more positive lymph nodes in the 40-49-year mammography age group of the NBSS was statistically significant, even when expressed as a percentage of all invasive cancers diagnosed. Such an excess is inconsistent with published data on extent of disease at diagnosis from previous studies. Analysis of NBSS mortality data after eliminating advanced cases detected by physical examination at the initial visit should result in minimal, if any, bias. CONCLUSIONS: Mortality analyses eliminating advanced cases detected by physical examination at the initial screening visit may be less susceptible to bias caused by possible nonrandom allocation of study participants and should be considered in future evaluations of the NBSS cohort after longer follow-up periods and in meta-analyses that may include the NBSS in assessments of the efficacy of mammography.
BACKGROUND: An unexpected excess of patients with breast cancer with four or more positive lymph nodes was observed in mammographically screened women who were younger than the age of 50 years at enrollment into the Canadian National Breast Screening Study (NBSS). It has been suggested that this excess is consistent with prior screening trials evaluating mammography. METHODS: A quantitative evaluation of the distribution of patients with breast cancer with four or more positive lymph nodes in the NBSS was undertaken, and the percentages of patients with breast cancer who were at an advanced state at diagnosis in the NBSS and in previous randomized screening trials were compared. The validity of mortality analyses after eliminating advanced cases detected by physical examination at the initial screening visit is examined. RESULTS: The excess of patients with cancer with four or more positive lymph nodes in the 40-49-year mammography age group of the NBSS was statistically significant, even when expressed as a percentage of all invasive cancers diagnosed. Such an excess is inconsistent with published data on extent of disease at diagnosis from previous studies. Analysis of NBSS mortality data after eliminating advanced cases detected by physical examination at the initial visit should result in minimal, if any, bias. CONCLUSIONS: Mortality analyses eliminating advanced cases detected by physical examination at the initial screening visit may be less susceptible to bias caused by possible nonrandom allocation of study participants and should be considered in future evaluations of the NBSS cohort after longer follow-up periods and in meta-analyses that may include the NBSS in assessments of the efficacy of mammography.
Authors: S Heywang-Koebrunner; K Bock; W Heindel; G Hecht; L Regitz-Jedermann; A Hacker; V Kaeaeb-Sanyal Journal: Geburtshilfe Frauenheilkd Date: 2013-10 Impact factor: 2.915