RESUMO
PURPOSE: To analyse the prognostic impact of isolated local recurrence (ILR) on long-term outcome for early-breast cancer patients treated with breast-conserving therapy. MATERIAL AND METHODS: The data of the EORTC 22881-10882 'boost-no boost' and 22922-10925 'IM-MS' trials were used to analyse the prevalence and outcome following ILR. A multistate model described the impact of intermediate events on long-term outcomes, taking into account various prognostic factors. This model was used to predict long-term outcomes after ILR. RESULTS: Of the 8367 patients included, 726 experienced an ILR, 11.6% of them within the first 2 years and 30.0% after 10 years. Ten-year cumulative breast cancer mortality rates after ILR were 58.2% in patients with an ILR within 2 years, 31.0% for ILR between 2 and 4 years, 17.6% in patients with an ILR between 4 and 10 years, and 29.7% for ILR after year 10 (p < 0.001). The multistate model showed that when tumour-free, younger breast cancer patients had a higher probability of developing ILR compared to older patients. Shorter time to ILR was associated with a higher chance to develop distant metastases (DM), and a shorter time to development of DM were associated with an increased hazard of breast cancer-related death. The multistate model enabled prediction of long-term outcome based on individual patient covariates, length of follow-up without recurrence and timing of ILR since randomisation. CONCLUSIONS: Outcome of early-breast cancer changed not only according to baseline risk factors but also according to the presence of intermediate events, time to these events, and subsequent follow-up without any further events.
