Validation of the Institut Curie simplified prognostic score for breast cancer meningeal carcinomatosis.

Breast cancer is the leading nonhematologic cause of meningeal carcinomatosis (MC), a unusual metastatic site. Its treatment generally relies on intrathecally and/or systematically delivered chemotherapy, combined with radiotherapy or surgery when needed. However, the overall survival (OS) of patients with proven MC is short, generally no longer than 6 months after the diagnosis of MC [1]. A retrospective study on 91 breast cancer MC diagnosed and treated homogeneously by a high-dose intrathecal methotrexate regimen [2] at the Institut Curie (Paris, France) from 2000 to 2007 has been recently reported [3]. This study retrieved four adverse independent prognostic factors at MC diagnosis: performance status (PS) of three to four, elevated Cyfra 21-1 marker in the cerebrospinal fluid (CSF), negative hormone receptors (HR) status and more than three previous chemotherapy lines for metastatic disease. Two prognostic scores were derived from these results, based on the number of adverse prognostic factors: the complete score included the four prognostic factors, whereas the simplified score did not include CSF Cyfra 21-1 as its dosage is not made routinely in this setting. Both prognostic scores allowed to classify patients into three groups with poor, intermediate and good prognosis (median OS of 2, 4 and 12 months, respectively). These scores required further validation. A retrospective validation of the simplified prognostic score has been conducted in patients diagnosed with breast cancer MC at the Institut Jules Bordet (Brussels, Belgium), after approval of the local ethic committee. Briefly, the electronic file of each of the 926 patients who underwent CSF cytological analysis from 01/2000 to 07/2010 were screened manually for MC diagnosis (tumor cells in CSF): 88 patients had both breast letters to the editor Annals of Oncology