Cell proliferation in 3,800 node-negative breast cancers: consistency over time of biological and clinical information provided by 3H-thymidine labelling index.

For breast cancer, many prognostic markers that initially appeared promising have failed to maintain their clinical predictive value. Few reports have analyzed the consistency over time of biological and clinical information provided by biomarkers. Tumour cell proliferation has acquired relevance as an indicator of prognosis and of response to treatment. Since its clinical role has been investigated for some decades, cell proliferation represents an ideal marker for an over-time validation. In 3,800 node-negative breast cancers recruited between 1972 and 1991, the consistency of information provided by the 3H-thymidine labelling index (TLI), in terms of basic relations with other clinico-pathological and biological variables and clinical predictivity, was evaluated using a combined analysis of results previously published by our group for distinct series of patients. Clinical predictivity was analyzed on a subset of 2,067 patients given local-regional therapy until relapse and followed for a median time from 6 to 10 years. Over the entire period TLI maintained a weak direct relation with tumour size and an inverse strong relation with steroid receptors. An increase in TLI was observed for tumours in post-menopausal patients up to the mid 1980s. During 3 different accrual periods (1972-1983; 1984-1987; 1988-1991), TLI was a consistent and independent predictor of relapse-free time, distant metastasis and overall survival, regardless of its consideration as a continuous variable or with a cutoff value of 3%. The reproducibility of our results over time provides support to the consistency of the methodology used and of the biological and clinical information obtained when using TLI as an indicator of breast cancer cell proliferation.