Recent Advances in Adjuvant Endocrine Therapy for Early Breast Cancer Clinical Updates Optimizing Endocrine Therapy for Early Breast Cancer Recent Advances in Adjuvant Endocrine Therapy for Early Breast Cancer Ruth M. O’Regan, MD Associate Professor, Hematology and Medical Oncology Emory University School of Medicine Director, Translational Breast Cancer Research Program Emory Winship Cancer Institute Chair, The Louisa and Rand Glenn Family Chair in Breast Cancer Research Director, Hematology and Medical Oncology Fellowship Program Atlanta, Georgia
Adjuvant Hormonal Therapy Options for Postmenopausal Patients AI ATAC, BIG-1-98 (upfront AI vs TAM) IES, ABCSG, BIG-1-98 (TAM to AI) TAM AI BIG-1-98, TEAM (AI to TAM) AI TAM MA.17 Extended AI TAM AI All superior to 5 years of TAM but which approach is best? ATAC Trialists Lancet Onc 2008, Mouridsen NEJM 2009, Goss JCO 2008, Coombes R Lancet 2007, Jakesz R JNCI 2007
ATAC Recruitment July 1996 – March 2000 Median follow-up 100 months Tamoxifen (n = 3,116) Surgery +/- RT +/- Chemo (20%) Anastrozole (n = 3,125) Discontinued following initial analysis as no efficacy or tolerability benefit compared with tamoxifen arm 84% HR positive 61% Node negative 5 years ATAC Trialists Lancet Onc 2009
Efficacy of Anastrozole vs Tamoxifen in HR+ Patient Population of ATAC Study 9-Year Follow-up Anastrozole Tamoxifen Recurrence 17.0% 21.8% HR = 0.76; P = .0001 Distant Recurrence 13.2% 15.6% HR = 0.84; P = .022 Contralateral Breast Cancer 2.5% 4.2% HR = 0.60; P = .004 Disease-Free Survival HR = 0.85; P = .003 Death After Recurrence HR = 0.90; P = .2 Significant long-term carryover effect for anastrozole: - Absolute difference in recurrence increased from 2.8% after 5 years to 4.8% after 9 years - Hazard ratio anastrozole vs tamoxifen for years 5-9 = 0.75; P = .01 ATAC Trialists Lancet Onc 2009
Death: All Causes HR+ Patients 30 30 25 20 15 10 5 HR 0.97 95% CI (0.86, 1.11) P-value 0.70 HR+ 25 20 Tamoxifen (T) Anastrozole (A) 15 10 5 1 2 3 4 5 6 7 8 9 Follow-up time (years) 2618 2598 2567 2549 2511 2504 2445 2432 2389 2339 2274 2227 2102 2068 1911 1888 1586 1551 659 620 At risk: A T ATAC Trialists Lancet Onc 2009
BIG 1-98 Design R A N D O M I Z E A Tamoxifen B Letrozole C Tamoxifen 2-Arm Option 3/98 to 3/00 1,835 pts B Letrozole 4-Arm Option 9/99-5/03 6,193 pts C Tamoxifen Letrozole D Letrozole Tamoxifen 2 5 YEARS Mouridsen NEJM 2009
Letrozole vs Tamoxifen Disease-free Survival Yearly DFS % 97.7 97.6 95.1 93.4 90.5 89.0 86.8 84.6 84.0 81.4 T 20 40 60 80 100 1 2 3 4 5 Percent Alive and Disease-Free Years from Randomization L N HR (95% CI) P 8010 0.81 (.70-.93) 0.003 No. at Risk 4003 4007 3892 3896 2964 2926 1261 1238 892 866 567 544 Mouridsen NEJM 2009
Cumulative Incidence Breast Cancer Relapse 20 5-year diff (L-T) = -3.4% (S.E. 1.2) cuminc P=0.0002 13.6% 15 T L 10 8.1% Proportion Failure (%) 10.2% 5 6.2% 1 2 3 4 5 Years from Randomization Mouridsen NEJM 2009
IES: Trial Design RANDOMI ZE Exemestane 5,162* Post Treatment Follow-up 10,335* Tamoxifen Tamoxifen 5,294* 2-3 years study treatment 2-3 years Start of study Diagnosis Total 5 years endocrine therapy * Total women years Coombes R Lancet 2007
Exemestane After Tamoxifen Disease Free Survival ITT ER+/Unknown E = 354 / 2352 E = 339 / 2296 treatment End of treatment End of T = 454 / 2372 T = 438 / 2306 HR=0.76 (95% CI: 0.66-0.88) Log-rank test: P=0.0001 HR=0.75 (95% CI: 0.65-0.87) Log-rank test: P=0.0001 Year % abs. diff. (95% CI) 2.5 5 3.2 3.4 (1.6 – 4.9) (0.1 – 6.8) 2.5 5 3.4 3.5 (1.8 – 5.1) (0.1 – 6.9) Coombes Lancet 2007
Exemestane After Tamoxifen Overall Survival ITT ER+/Unknown E=222 / 2352 treatment End of treatment End of E=210 / 2296 T=261 / 2372 T=251 / 2306 HR=0.85 (95% CI: 0.71-1.02) Log-rank test: P=0.08 HR=0.83 (95% CI: 0.69-1.00) Log-rank test: P=0.05 year % abs. diff. (95% CI) 2.5 5 0.8 1.2 (-0.4 – 1.9) (-1.5 – 3.9) 2.5 5 0.7 1.6 (-0.4 – 1.9) (-1.2 – 4.3) Coombes Lancet 2007
BIG 1-98 Design A B C D R A N D O M I Z E Tamoxifen 2-Arm Option 3/98 to 3/00 1,835 pts B Letrozole C Tamoxifen Letrozole 4-Arm Option 9/99-5/03 6,193 pts D Letrozole Tamoxifen 2 5 YEARS Mouridsen NEJM 2009
BIG 1-98 Sequential Therapy Two Pairwise Comparisons Letrozole 3 blinded arms Sequential vs letrozole monotherapy Evaluated from randomization Median Follow-up 71 mos. 99% confidence intervals to account for multiple comparisons N = 3,094 Tamoxifen Letrozole 2 5 Letrozole N = 3,086 Letrozole Tamoxifen 2 5 YEARS Mouridsen NEJM 2009 13
BIG 1-98 Sequential Treatment Disease-Free Survival Mouridsen NEJM 2009 14
Breast Cancer Events TamLet vs Let Overall By Nodal Status* *42% of the population is node positive; 58% node negative Mouridsen NEJM 2009 15
Breast Cancer Events TamLet vs Let Overall By Nodal Status* *42% of the population is node positive; 58% node negative Mouridsen NEJM 2009
TEAM Trial Upfront AI vs Switching Postmenopausal ER+ and/or PgR+ invasive breast cancer Current analysis (2.75 year follow-up) R A N D O M I Z E Tamoxifen 20 mg/day (2-3 years) Exemestane 25 mg/day (2-3 years) Primary Surgery Exemestane 25 mg/day (5 years) Primary endpoint: DFS for tamoxifen vs exemestane (2.75 years) Following IES: protocol modified - DFS for exemestane vs tamoxifen→exemestane (5 years) Jones et al. SABCS 2008. Abstract 15. 17 17
TEAM Trial Efficacy of Tamoxifen vs Exemestane HR (95% CI) P Value Disease-Free Survival* 0.89 (0.77-1.03) .12 Disease-Free Survival (on Study Drug) 0.83 (0.71-0.97) .02 Recurrence-Free Survival 0.85 (0.72-1.00) .05 Time to Distant Metastases 0.81 (0.67-0.98) < .03 * Events: Exemestane 7% vs tamoxifen 8% Jones et al. SABCS 2008. Abstract 15.
Extended Adjuvant Therapy MA.17: Trial Design Randomization (Disease-free) Letrozole 2.5 mg daily n = 2,575 Tamoxifen Placebo n = 2,582 5 years early adjuvant 5 years extended adjuvant Node(–) 2,581 Node(+) 2,370 Follow-up 30 months Goss JCO 2008
Extended Adjuvant Therapy Letrozole After 5 Years of TAM DFS* Distant* DFS OS Node– pts Node+ HR=0.61* (0.45-0.84) HR=0.45* (0.27-0.75) HR=0.63 (0.31-1.27) HR=0.53* (0.36-0.78) HR=1.52 (0.76-3.06) (0.38-0.98) *Statistically significant benefit of LET. A similar reduction in local recurrences, new primaries, and distant recurrences occurred in node-positive and node-negative patients Goss et al JNCI 2005
Ongoing First Generation AI Adjuvant Trials BIG 1-98 (BIG FEMTA) NSABP B33 IES Trial ITA Tamoxifen Anastrozole Placebo Letrozole Exemestane MA.17 ATAC Trialists Lancet Onc 2008, Mouridsen NEJM 2009, Goss JCO 2008, Coombes R Lancet 2007, Jakesz R JNCI 2007
Upfront vs Sequencing Approach? BIG-1-98 suggests that an upfront AI may be superior to a TAM-AI sequencing approach But patients seem to do just as well with an AI followed by TAM (unclear what clinical relevance this has) Major issue is the fact that ER-positive cancers are not all the same and therefore one size may not fit all
Key Questions Can we use biomarkers/molecular profiling to decide how to treat patients with ER+ cancers? Do all or just a subset of ER+ cancers require more than 5 years of endocrine treatment? Can we use biomarkers/molecular profiling/pharmacogenomics to individualize endocrine therapy (AI vs TAM)?
Are Heterogenous in Outcome ER+ Cancers Are Heterogenous in Outcome ER+ ER+ Copyright ©2003 by the National Academy of Sciences Sorlie et al PNAS 2003
Efficacy of Tamoxifen Varies in ER+ Cancers 2 4 6 8 14 16 Years 0.0 0.2 0.4 0.6 0.8 1.0 DRFS Placebo Tamoxifen 12 10 2 4 6 8 14 16 Years 0.0 0.2 0.4 0.6 0.8 1.0 DRFS Placebo Tamoxifen 12 10 Low Risk (RS<18) Int Risk (RS 18-30) N 171 142 N 85 69 High Risk (RS≥31) N 99 79 2 4 6 8 14 16 Years 0.0 0.2 0.4 0.6 0.8 1.0 DRFS Placebo Tamoxifen 12 10 TAM-resistant Interaction P = 0.06 Paik et al SABCS 2004
Correlation Between Recurrence Score and Intrinsic Subtype Luminal A (n = 123) Luminal B (n = 55) Low 62 1 Intermediate 25 4 High 36 50 Fan et al NEJM 2006
Oncotype DX 21 Gene Recurrence Score (RS) Assay 16 Cancer and 5 Reference Genes From 3 Studies PROLIFERATION Ki-67 STK15 Survivin Cyclin B1 MYBL2 ESTROGEN ER PR Bcl2 SCUBE2 GSTM1 BAG1 INVASION Stromolysin 3 Cathepsin L2 CD68 Category RS (0 – 100) Low risk RS < 18 Int risk RS ≥ 18 and < 31 High risk RS ≥ 31 Reference Beta-actin GAPDH RPLPO GUS TFRC HER2 GRB7 Paik et al NEJM 2004
Poor Outcome for Patients with HER2+ MBC Treated with Endocrine Agents Alone Clinical Benefit PFS Anastrozole 28% 2.4 months Letrozole 29% 3.0 months *PFS in first-line aromatase inhibitor trials ranges from 9 to 11 months Kaufmann ESMO 2006, Johnston SABCS 2008
PR is prognostic or predictive for patients treated with tamoxifen Overall Survival According to Tumor Receptor Status in Women Treated with Tamoxifen PR is prognostic or predictive for patients treated with tamoxifen Cui, X. et al. J Clin Oncol; 23:7721-7735 2005
Disease-free Survival by Ki-67 LI in BIG-1-98 (Letrozole and Tamoxifen) 1,252 (47%) expressed Ki-67 LI > 11% (high) Viale G et al. SABCS 2007 Abs 64.
Possible Surrogate Markers for Hormone Resistance LUM A LUM B ER PR HER2 RS Low Ki-67 High Ki-67 Hormone-sensitive Hormone-resistant
Luminal B Cancers Most Likely to Relapse in First 5 Years Copyright ©2003 by the National Academy of Sciences Sorlie et al PNAS 2003
Extended Adjuvant Therapy is Beneficial in Luminal A But Not B Cancers (using PR as a surrogate)? ER+PR- LUMINAL B ER+PR+ LUMINAL A Disease free survival Distant disease free survival Goss et al. J Clin Oncol; 2007
ATAC Trial: Disease-free Survival HR+ patients 30 29.9% 30 25 20 15 10 5 HR 0.85 95% CI (0.76, 0.94) P-value 0.003 HR+ 25 25.8% 20 Tamoxifen (T) Anastrozole (A) 16.4% 15 10 13.9% 5 Absolute difference 2.5% 4.1% 1 2 3 4 5 6 7 8 9 Follow-up time (years) 2618 2598 2541 2516 2453 2400 2361 2306 2278 2196 2159 2075 1995 1896 1801 1711 1492 1396 608 547 At risk: A T HR, hazard ratio; CI, confidence interval ATAC Trialists Lancet Onc 2008
Aromatase Inhibitor vs Tamoxifen? TransATAC is a subset of patients from the ATAC trial in whom tumor blocks are available for molecular profiling 21-gene recurrence score performed (n = 1,308) None of these patients received chemotherapy Will 21-gene recurrence score be prognostic for post-menopausal patients treated with anastrozole? Can the recurrence score define a group of patients who would do just as well with tamoxifen? Dowsett, SABCS 2008, Abstract 53
21-gene Recurrence Score To Predict Risk of Distant Recurrence In Patients Treated With Anastrozole or Tamoxifen Results Node- (N=872) Node+ (N=306) % pts 9-year DR rate Low RS <18 59% 4% 52% 17% Int RS 18-30 26% 12% 31% 28% High RS ≥ 30 15% 25% 49% High vs. Low RS: HR 5.2 Int vs. Low RS: HR 2.5 High vs. Low RS: HR 2.7 Int vs. Low RS: HR 1.8 P<.001 for RS in predicting time to distant recurrence (DR) in N+ and N- patients Dowsett, SABCS 2008, Abstract 53 36 36
Percent with Distant Recurrence at 9 Years Node Negative # of events All Patients, Low RS (n=513) 20 All Patients, Int. RS (n=229) 24 All Patients, High RS (n=130) 28 All Patients (n = 872) 72 Tamoxifen, Low RS (n=245) 8 Tamoxifen, Int. RS (n=117) 12 Tamoxifen, High RS (n=70) 21 All Tamoxifen (n = 432) 41 Anastrozole, Low RS (n=268) 12 Anastrozole, Int. RS (n=112) 12 Anastrozole, High RS (n=60) 7 All Anastrozole (n = 440) 31 Dowsett, SABCS 2008, Abstract 53 10 20 30 40 50 60 70 Percent with Distant Recurrence at 9 Years 37
Percent with Distant Recurrence at 9 Years Node Positive # of events All Patients, Low RS (n=160) 25 All Patients, Int. RS (n=94) 25 All Patients, High RS (n=52) 24 All Patients (n = 306) 74 Tamoxifen, Low RS (n=79) 11 Tamoxifen, Int. RS (n=47) 13 Tamoxifen, High RS (n=26) 11 All Tamoxifen (n = 152) 35 Anastrozole, Low RS (n=81) 14 Anastrozole, Int. RS (n=47) 12 Anastrozole, High RS (n=26) 13 All Anastrozole (n = 154) 39 Dowsett, SABCS 2008, Abstract 53 10 20 30 40 50 60 70 Percent with Distant Recurrence at 9 Years 38
Summary Postmenopausal Patients Available data suggests that upfront AI approach is optimal for unselected ER+ breast cancers But luminal A and B are clearly different and may need different therapeutic approaches Luminal A (low RS): May do fine with tamoxifen or sequencing approach Require extended adjuvant therapy? Luminal B (high RS): Five years of treatment probably sufficient AIs may be better in some patients (or they could be hormone-refractory)
Premenopausal Patients Tamoxifen for 5 years is standard of care SOFT trial is evaluating ovarian suppression plus tamoxifen vs exemestane Should we evaluate CYP2D6 in premenopausal patients and what should we do if they are poor metabolizers?
Randomized Trial in Premenopausal Patients With HR+ Breast Cancer 1,803 premenopausal breast cancer patients Endocrine-responsive (ER and/or PR positive) No chemotherapy except neoadjuvant Treatment duration: 3 years Tamoxifen 20 mg daily Zoledronic acid 4 mg q6mo Anastrozole 1 mg daily Anastrozole mg daily Surgery (+ RT) Goserelin 3.6 mg q28d RANDOMI ZAT I ON Gnant et al NEJM 2009
Zoledronic Acid (ZOMETA®) Preclinical Profile Primary Endpoint: Disease-free Survival Zoledronic Acid vs No Zoledronic Acid 100 90 80 70 60 50 40 30 20 10 12 24 36 48 72 84 Time since randomization, months Disease-free survival, % No. of Hazard ratio (95% CI) events vs No ZOL P value ZOL 54 0.643 (0.46 to 0.91) .011 No ZOL 83 Gnant et al NEJM 2009 42 42
Primary Endpoint: Disease-free Survival Anastrozole vs Tamoxifen Zoledronic Acid (ZOMETA®) Preclinical Profile Primary Endpoint: Disease-free Survival Anastrozole vs Tamoxifen Follow-up 60 months 100 90 80 70 60 50 Disease-free survival, % 40 30 No. of Hazard ratio (95% CI) events vs TAM P value ANA 72 1.096 (0.78 to 1.53) .593 TAM 65 20 10 12 24 36 48 60 72 84 Time since randomization, months SOFT study continues to accrue Gnant et al NEJM 2009 43 43
Metabolism of Tamoxifen Jin et al JNCI 97: 30-39, 2005
CYP2D6 Genetic Variations Caucasians and Western Europeans: 1-2% have multiple copies (Ultra-rapids = UM) 70% have two wild type alleles (Extensive = EM) 20% have one variant (CYP2D6*4) allele (Intermediate = IM) ~8% have two variant alleles (Poor = PM) CYP2D6 is responsible for the conversion of tamoxifen to its most abundant active metabolite: endoxifen
Kaplan Meier Plots For Patient By CYP2D6*4 Genotype Disease Free Overall Survival Hot flashes 36/177 0/13 Goetz J Clin Oncol 23: 9312-9318, 2005
CYP2D6 as Predictive Marker as Prognostic Marker in the ABCSG-8 Trial ABCSG-8 : TAM vs TAM followed by anastrozole CYP2D6 and risk of breast event relative to extensive metabolizers (EM) N CYP2D6: PM CYP2D6: IM Tamoxifen (5 years) 67 3.83, P= .017 0.87, P = .689 Tam to Ana (5 years) 55 1.02, P = .985 0.81, P = .538 Tamoxifen (years 3-5) 2.81, P = .081 0.75, P = .431 Anastrozole (years 3-5) 31 0.71, P = .782 0.57, P = .269 Goetz, SABCS 2008, Abstract 57 47 47
Risk of Breast Cancer Recurrence in Women Initiating Tamoxifen With CYP2D6 Inhibitors Eligibility: Continuous eligibility 6 mos. prior to tamoxifen initiation Tamoxifen naïve (6 mos. negative history) Tamoxifen duration ≥ 24 mos. Medication possession ratio of ≥ 0.7 No CYP2D6 inhibitor therapy (n = 945) Weak CYP2D6 inhibitor therapy use or without overlap with tamoxifen (n = 355) Moderate-severe CYP2D6 inhibitor use with tamoxifen (n = 359) (n = 1,659) Retrospective cohort analysis of medical and pharmacy claims from the Medco Health Solutions integrated database Primary endpoint: hospitalization for breast cancer (event-free survival) Median duration of overlap between CYP2D6 inhibitors and tamoxifen: 287 days Aubert et al. J Clin Oncol 2009; 27 (suppl): 18s (abstract CRA508).
Risk of Breast Cancer Recurrence in Women Initiating Tamoxifen With CYP2D6 Inhibitors HR* P value No CYP2D6 inhibitors 945 7.5% reference Moderate/severe CYP2D6 inhibitors 407 14% 1.92 (1.36-2.73) .0002 SSRIs Weak 137 9% 1.07 (0.79-1.45) .677 Moderate/potent 213 16% 2.20 (1.46-3.31) * HR relative to no CYP2D6 inhibitor group Concomitant use of tamoxifen with moderate-severe CYP2D6 inhibitors significantly increases the risk of breast cancer recurrence Moderate-potent SSRIs double the risk of recurrence, while weak SSRIs were not associated with increased risk Aubert et al. J Clin Oncol 2009; 27 (suppl): 18s (abstract CRA508).
Summary Premenopausal Patients Available data do not support a role for ovarian ablation with an AI However, trials have not taken into account the luminal subtypes CYP2D6 data is compelling and should be looked at in patients on the SOFT trial Until then the role of CYP2D6 testing remains unclear since we do not have data on how to treat poor metabolizers
Conclusions HR+ cancers are clearly heterogeneous with divergent outcomes Need to identify robust predictive factors for both luminal A and B cancers Essential to establish molecular differences between luminal subtypes so that novel therapeutic approaches can be developed