Examination of TIMP-1 Levels and Relapse-Free Survival for Patients in NCIC CTG MA.14 Who Received Adjuvant Tamoxifen +/- Octreotide LAR.

Abstract

Abstract Background: Tissue inhibitor of metalloproteinase-1 (TIMP-1) has been shown to have diverse multifunctional roles in tumorigenesis. AL/KL/SA postulated from first line metastatic endocrine therapy that elevated pre-treatment serum TIMP-1 required additional therapy. NCIC CTG MA.14 adjuvant endocrine trial permitted examination of whether TIMP-1 serum levels post-chemotherapy was associated with RFS. Methods: NCIC CTG MA.14 is a trial where 667 postmenopausal patients were randomized to receive adjuvant tamoxifen +/- octreotide LAR with a stratification factor of no, concurrent or sequential chemotherapy. For the purposes of this work, this was simplified to whether a patient had or had not received adjuvant chemotherapy prior to the serum draw. Serum TIMP-1 was assessed on >90% of the trial patients. We examined the effect of baseline TIMP-1 levels on relapse-free survival (RFS) of 1) all types, 2) bone only, 3) all types of bone, and 4) non-bone, by timing of chemotherapy before the serum draw utilizing continuous TIMP (ng/ml) and categorical TIMP (<454, >454 ng/ml, based on 95% non-parametric cut-point for healthy post-menopausal females). Data were available on patient and tumour characteristics of age (years), pathologic tumour size (cm), pathologic lymph node status (# nodes), IGF-1, C-peptide, IGFBP-3. Step-wise forward Cox multivariate models were used where a factor was added if p<=0.05, and interaction terms were added for TIMP-1 level and timing of chemotherapy. Results: High (categorical) TIMP-1 was significantly associated with longer RFS (p=0.04) in the non-bone RFS multivariate model in the subgroup of patients who had prior chemotherapy, but not in those who did not. The interaction between TIMP-1 and administration of chemotherapy before serum draw was significant (p=0.02). High (continuous) TIMP-1 was significantly associated with longer bone only RFS (p=0.04) in subgroup who had no chemotherapy before the serum draw, but not in those who may or may not have had it after. There was no significant interaction effect. Categorical and continuous TIMP-1 were not associated with any other type of RFS. Conclusions: We found a predictive benefit with the assessment of serum TIMP-1 in hormone receptor positive patients, that high TIMP-1 levels after adjuvant chemotherapy were predictive of a reduction in non-bone RFS. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 3022.