Sequential identification of factors predicting distant relapse in breast cancer patients treated by conservative...

Predictors of distant relapse following conservative surgery for breast cancer were studied in a review of 425 women. Five steps of breast cancer patient management were defined in which increasing amounts of information, potentially relevant to prognosis for metastasis-free survival (MFS), were available: (1) clinical, (2) biopsy, (3) tumorectomy, (4) axillary dissection, and (5) adjuvant treatment. At each step, a prognosis study based on the Cox model was carried out using all acquired information from the first step.

Among the 21 studied variables, 5 were independent stable risk factors in predicting MFS: (1) clinical node status, (2) modified Scarff-Bloom-Richardson (MSBR) histoprognostic grade, (3) progesterone receptor (PR), (4) anatomic tumor size, and (5) histologic lymph node status. These factors were progressively identified throughout the successive prognostic analyses and kept their significance at the reference step (axillary dissection step where all information is acquired). According to the prognostic score based on the significant variables, a stratification of the patients had been built at each step, identifying three risk groups (low, moderate, high). Even at biopsy step, the mere knowledge of clinical information, such as clinical node status, and biopsy information, such as MSBR grade and PR status, would enable 68% of the patient to be well classified according to the stratification of reference. Knowledge of an additional factor, such as anatomic tumor size, would bring the rate up to 88%. Some subsets of patients with stable prognosis throughout the steps were identified and their profiles were described. It is noticeable that 95% of the patients, classified low risk at the biopsy step, were patients that were stable.

The early recognition of patients, highly curable by local therapy alone, would obviate aleatoric neoadjuvant treatment.