5091 Background: Endometrial cancer, especially when diagnosed in stage I, is associated with an excellent prognosis. Although, special histological types or poor differentiation characterize a "high risk" group, these factors do not sufficiently explain recurrence. The present multicenter study was undertaken to investigate whether the expression of the neural cell adhesion molecule L1-CAM, even in small tumor areas, is able to identify early endometrial cancers with such a particular aggressive behavior.
METHODS: A total of 1031 endometrial cancers classified as "endometrioid" in the pathological examination were retrospectively assessed for L1-CAM expression by means of immunohistochemistry using the monoclonal mouse antibody L1 clone 14.10 [Covance, Princeton, NJ] on paraffin embedded specimens. Cancers with ≥ 5% of positive cells or circumscribed L1-CAM expressing islets were considered as L1-CAM positive. Blinded evaluation of the specimens was done by two independent pathologists.
RESULTS: 22.1% of the specimens were rated L1-CAM positive. Tumors classified as "high risk" and "low risk" in the intraoperative risk assessment were positive in 27.2% and 19.3%, respectively (p=0.0001). L1-CAM positivity was associated with histopathological grading (p=0.0001). Of special note is that L1-CAM was a reliable predictor for local (HR=2.134; p=0.05) and especially distal recurrence (HR=5.431; p=0.0001), but not for lymph node relapse. In addition to histopathological grading and age at diagnosis, L1-CAM expression was shown to be the most prominent independent predictor for disease-free (HR=5.411; p=0.0001) and overall survival (HR=4.154; p=0.0001). Subgroup analyses showed that even in "low risk" patients L1-CAM was a strong predictor for disease-free and overall survival (p=0.0001 and p=0.0001, respectively).
CONCLUSIONS: L1-CAM expression in stage I, type 1 endometrial cancers reliably identifies a subset of cancers with high risk for recurrence and with poor survival. These cancers are candidates for trials evaluating the value of adjuvant treatment. We propose L1-CAM in early type 1 endometrial cancer as a novel tool for therapeutic decision-making.
CONCLUSION: As shown by the significant influence of metastatic sites, some P die from their advanced systemic disease situation before they would experience cerebral progression, in part explaining the influence of systemic treatment. In other individuals however, intensified local treatment and systemic treatment appear to influence both cTTP and OS significantly, implicating a direct influence of systemic therapy on BM. This might result from an impaired blood brain barrier around metastatic sites, making sufficient tissue concentrations of cytotoxic agents possible. No significant financial relationships to disclose.
TREATMENT: Rx 375 mg/m(2) IV at day 1; concomitant therapy: Rasburicase, steroids only for anaphylaxis, intrathecal (IT) triple drug at days 1, 3 for CNS pts only. Begin of chemotherapy at day 5. Response evaluation: product of 2 perpendicular diameters of 1 - 3 index lesions/% blasts in BM/PB within 24 h prior to Rx and at day 5: responder (RP): at least 1 lesion with at least objective effect (decrease of ≥25%) and no progress (increase of ≥25 %) at other sites. Study plan: Simon 2-stage phase II with α and β = 5%. Response rate (RR) for poor activity was set to 45%, for good activity 65%. 33 pts entered the first stage, final evaluation after 79 pts.