We examined multiple samples of 65 primary breast carcinomas larger than I cm in diameter for thymidine labelling index (TLI), DNA index (DNAI, a measure of cellular DNA content by flow cytometry), and estrogen (ER) and progesterone (PgR) receptors by radioligand-binding. One or more axillary metastases were also assayed in I I patients. Two to IS samples were successfully assayed for TLI from 59 tumors, 2-31 samples for DNA1 from 61 tumors, and 2-15 samples from 55 tumors for ER and PgR Criteria for heterogeneity were excess inter-sample variance in comparison with intrasample variance at the p < 0.05 level for TLI and DNAI, and variation of clinically significant magnitude in assay results for ER and PgR Sixty-one percent of tumors were heterogeneous for TLI, 26% for DNAI, 24% for ER and 40% for PgR. High TLI disposed toward heterogeneity for TLI itself (p = 0.06), for ER (p = 0.04), and for PgR (p = 0.007). Young age favored heterogeneity for TLI (p = O.I2), ER (p = 0.002), and PgR (p = 0.04). Heterogeneity for DNA1 was not related to age and TLI status but was more common in larger tumors (p = 0.08).
After consideration of relationships between TLI, age, size, ER and PgR, TLI rather than age appears to be the more important determinant of heterogeneity for receptors. High TLI could lead to heterogeneity through increased numbers of cell divisions that favor emergence of variant stemlines, or by causing local vascular and humoral disparities through rapid growth. Regional heterogeneity can explain erroneous prognostic predictions in approximately 10% to 20% of breast carcinoma patients. We recommend multiple sampling of large breast carcinomas and analysis of axillary metastases for study of tumor markers.