1. Introduction MarkerPolymorphisms in CYP2C19 (*2 and *17) assessed by Taqman allelic discrimination ObjectivesTo evaluate the predictive capacity of CYP2C19 variants with regard to tamoxifen response in postmenopausal breast cancer. HypothesisPolymorphisms in estrogen metabolizing enzymes, like CYP2C19, may affect breast cancer molecular subtype as well as response to anti-estrogens. We hypothesize that patients carrying variant CYP2C19 alleles may derive differential benefit from tamoxifen compared to patients who carry the wild type alleles. Methods (1)Patients CharacteristicsFrom 1982 until 1994 a randomized clinical trial was conducted in the Netherlands, studying the benefit of adjuvant tamoxifen (IKA-trial) in postmenopausal breast cancer patients. Inclusion criteriaIn the original study, 1662 breast cancer patients were included who were were post-menopausal, less than 76 years of age and had a T 1– 4, N 0–3, M 0 breast tumor. Primary tumor blocks were recollected and after revision for the presence of invasive tumor cells and positive ER status (≥10%), tumor material of 563 patients was used for this translational study. Exclusion criteriaMastitis or palpable supra- or infraclavicular lymph nodes TreatmentPatients were randomized in a 2:1 distribution between 1 year tamoxifen (30 mg per day) versus no adjuvant therapy. After 1 year a second randomization was performed to receive another 2 years of tamoxifen or to stop further treatment. From 1989, based on two interim analyses showing a significant improvement in recurrence free-free survival in lymph node positive patients, these node positive patients were all allocated to the tamoxifen treatment arm (ie skipped the first randomization). Methods (2)Specimen characteristics Material usedFormalin-fixed paraffin-embedded (FFPE) breast tumor tissue of the primary tumor. DNA was isolated from FFPE material. No control samples were used Preservation/storageFormalin fixation and paraffin embedding. Storage at room temperature. Tumor DNA was stored at 4ºC. Table S3. Specifications of REMARK recommendations 1/3

2. Methods (3)Assay methods AssayGenotyping for CYP2C19*2 (681G>A, rs4244285), CYP2C19*17 (806C>T, rs12248560) and CYP2D6*4 (1846G>A, rs3892097) was performed on genomic DNA. ProtocolTaqman allelic discrimination assays were used (Applied Biosystems, Nieuwerkerk ad IJssel, The Netherlands) on an ABI Prism 7500 Sequence Detection system (Applied Biosystems). The assays IDs are C_25986767_70 (CYP2C19*2), C_469857_10 (CYP2C19*17) and C_27102431_D0 (CYP2D6*4). Thermal profile for cyp2D6*4 genotyping consisted of 95 ○ C for 10 minutes, followed by 60 cycles of 15 seconds at 92 ○ C and 30 seconds at 60 ○ C. Thermal profile for cyp2DC19*2 and CYP2C19*17 genotyping consisted of 95 ○ C for 10 minutes, followed by 50 cycles of 15 seconds at 92 ○ C and 90 seconds at 60 ○ C. Control experimentsEach (96 well) plate that was run for allelic discrimination, included 1-2 water samples as negative control ReproducibilityFor each genotyping assays a series of 50 tumor samples were run in duplo, showing > 95% concordance. QuantificationGenotypes were scored by allele-specific fluorescence using 7500 fast system SDS software (Applied Biosystems). BlindingScoring of the genotypes was done without knowledge regarding both the recurrence-free-interval survival as well as the treatment arm at the time of scoring. Methods (4)Study design I Case selectionA randomized controlled trial. The translational study presented her was performed retrospectively. The median duration of follow-up for patients without a recurrence event was 7.8 years. Patient records were re-evaluated for recurrence until 2000. Clinical endpointsThe improvement of recurrence free interval (RFI) with tamoxifen versus no nil was assessed according to the presence or absence of CYP2C19*2 and CYP2C19*17 variants. RFI included local, regional, distant recurrences and breast cancer-specific death, but not contra- lateral breast cancer, as the primary event. Variables examined or considered Multivariate Cox models were included age (≥ 65 versus < 65), grade (grade 3 versus grade 1-2), tumor size (T3-4 versus T1-T2), HER2 status (positive versus negative), estrogen receptor expression (10-99% versus 100 %) and progesterone status (positive versus negative) as covariates. Rational for sample sizeThe sample size of the translational study is based on the amount of available tumor blocks containing invasive, ER positive tumor cells, that could be recollected. 2/3

3. Methods (5)Statistical analysis Statistical methods and variable selection procedure Genotypes were tested for Hardy Weinberg equilibrium using a Chi-square test. The distribution of clinico-pathological characteristics by the CYP2C19*2 and CYP2C19*17 variants was evaluated using Chi-square tests. Survival curves were constructed using the Kaplan Meier method and compared using log-rank tests. To determine whether the benefit from tamoxifen was different in CYP2C19 *2 carriers ( one or two *2 alleles versus no *2 allele) or CYP2C19*17 carriers (( one or two *17 alleles versus no *17 allele), (covariate unadjusted) Cox proportional hazard regression models were constructed with an interaction between treatment and the genotype. To account for the change in randomization of node positive patients, all survival analyses are stratified by nodal status. All p-values are based on a two- sided test. All calculations were made with Statistical Package for the Social Sciences (SPSS) 15.0 Inc., IL, USA. Missing dataCases with a missing value for one of the variables were excluded from the multivariate analysis, with the exception of missing HER2 and PR data for which a separate level was created Marker handling in analysisThe following genotypes were compared to detect a difference in tamoxifen effect : CYP2C19 (one or two *2 alleles versus no *2 allele and one or two CYP2C19*17 alleles versus no CYP2C19*17 allele). Results (1)Data Flow of patientsSee Figure S1 for description of patients excluded for this translational study. See Table S1 for characteristics of total study patients versus the 739 patients with sufficient tumor material included in TMA. CharacteristicsSee Table 1. 3/3 Results (2)Analysis and presentation Relation to standard prognostic variables See Table 1 and S4-6. Univariate analysisAll presented HRs are multivariate tested Multivariate analysis:See table 2 and S7-10. Estimated effects with CIs for marker and all other variables in the model. Discussion Interpretation, limitations and implication See discussion section