RESUMEN
BACKGROUND: Using prior mammograms from patients with delays in their breast cancer diagnoses, we sought to describe in-vivo growth kinetics of untreated breast cancer to determine if the time they became clinically apparent can be predicted. METHODS: Patient and tumor characteristics were collected from those who presented with "missed," untreated breast cancer to a breast center in a single institution. Only patients whose biopsied masses revealed estrogen receptor-positive, Her2-negative (ER+/Her2-) invasive cancers were included. Two attending radiologists reviewed images from prior mammograms. Rates of change in volume were calculated in mm3/day, and a logarithmic equation was used to calculate tumor volume doubling time (TVDT). A Spearman's Rho correlation was performed for the continuous variables, and the Mann-Whitney U and Kruskal-Wallis tests were used to compare categorical data. A p value < 0.05 was considered statistically significant. Logistic regression was performed to determine if patient or tumor characteristics were correlated to tumor growth velocity. RESULTS: Of the 36 ER+/Her2- invasive breast cancers included in the analysis, 13 (36%) were at least cT2 (of TNM), 7 (19%) were grade 3, and 7 (19%) were node positive at diagnosis. Grade (p = 0.043) and pathologic invasive tumor size (p = 0.001) were positively correlated to tumor growth velocity. Median TVDT was 385 days (23-1897). Age, nodal positivity, Oncotype Dx® Recurrence Score, time of diagnostic delay, and spheroid-ellipsoid discrepancy (SED) were not related to tumor growth velocity in this sample. CONCLUSION: In this cohort of patients with untreated ER+/Her2- invasive breast cancers, grade and pathologic tumor size were found to be positively correlated to growth velocity. The growth rates in a homogeneous group of tumors varied widely and could not be predicted. One possible explanation for this finding is that other difficult-to-measure biologic factors such as tumor microenvironment may play a greater role in tumor progression than traditional clinicopathologic characteristics.