Costantino, J. P.

Abstract Background: Conventional WBI after lumpectomy for early-stage breast cancer decreases ipsilateral breast tumor recurrence (IBTR), yielding comparable results to mastectomy. Accelerated PBI appears effective in reducing IBTR by treating only the tumor bed area. As the majority of IBTR occur at or in the vicinity of the tumor bed, we hypothesized that PBI would be as effective as WBI in controlling IBTR. The primary aim of NSABP B-39/RTOG 0413 was to determine if PBI provides equivalent local tumor control post lumpectomy compared to WBI in pts with early-stage breast cancer. The equivalency test was based on a 50% margin of increase in the hazard ratio (HR=1.5). Secondary endpoints included: overall survival (OS), recurrence-free interval (RFI), distant disease-free interval (DDFI), and toxicity. Methods: Eligible pts had lumpectomy with histologically-free margins and 0-3 positive axillary nodes. Pts were stratified by stage, menopausal status, hormone receptor status, and intent to receive chemotherapy and then randomized to PBI or WBI. PBI was 10 fractions of 3.4-3.85 Gy, given twice daily with either brachytherapy or 3D external beam radiation. WBI was 50 Gy in 2 Gy fractions given daily with a sequential boost to the surgical cavity. Follow-up was every 6 mos for 5 yrs and then annually. All analyses were by intent-to-treat. Results: From 3-21-05 to 4-16-13, 4216 pts were randomized: 2107 PBI; 2109 WBI. 61% were postmenopausal; 81% were hormone receptor-positive; 29% intended to receive chemotherapy. Stage distribution was: DCIS, 24%; invasive pN0, 65%; invasive pN1, 10%. As of 7-31-18, median follow-up was 10.2 yrs. There were 161 IBTRs as first events: 90 PBI v 71 WBI (HR 1.22; 90%CI 0.94-1.58). Per protocol-defined margin, to declare PBI and WBI equivalent regarding IBTR risk, the 90% CI for the observed HR had to lie entirely between 0.667 and 1.5. The percent of pts IBTR-free at 10 yrs was 95.2% PBI v 95.9% WBI. A statistically significant difference in the 10-yr RFI rate favored WBI (91.9% PBI v 93.4% WBI; HR 1.32; 95%CI 1.04-1.68; p=0.02). No statistically significant differences existed between PBI and WBI in DDFI (HR 1.31; 95%CI 0.91-1.91; p=0.15), OS (HR 1.10; 95%CI 0.90-1.35; p=0.35), or DFS (HR 1.12; 95%CI 0.98-1.29; p=0.11). Grade 3 toxicity was 9.6% PBI v 7.1% WBI, and grade 4-5 toxicity was 0.5% v 0.3%, respectively. Discussion: PBI did not meet the criteria for equivalence to WBI in controlling IBTR based on the upper limit of the hazard ratio confidence interval. However, the absolute difference in 10-yr rate of IBTR was <1% (4.8% PBI v 4.1% WBI). The risk of an RFI event was statistically significantly higher for PBI compared to WBI, but the absolute difference in 10-yr RFI rate was also small (8.1% PBI v 6.6% WBI). DDFI, OS, and DFS were not statistically different for PBI v WBI. Grade 3-5 toxicities, although low, were more common for PBI than WBI. The trial population was heterogeneous, ranging from Stage 0-2 breast cancer, and outcome by risk categories are being analyzed. Support: U10CA180868, -180822, UG1CA189867. Citation Format: Vicini FA, Cecchini RS, White JR, Julian TB, Arthur DW, Rabinovitch RA, Kuske RR, Parda DS, Ganz PA, Scheier MF, Winter KA, Paik S, Kuerer HM, Vallow LA, Pierce LJ, Mamounas EP, Costantino JP, Bear HD, Germaine I, Gustafson G, Grossheim L, Petersen IA, Hudes RS, Curran, Jr. WJ, Wolmark N. Primary results of NSABP B-39/RTOG 0413 (NRG Oncology): A randomized phase III study of conventional whole breast irradiation (WBI) versus partial breast irradiation (PBI) for women with stage 0, I, or II breast cancer [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr GS4-04.

Abstract Background The clinical results of NSABP B-35, phase III trial comparing 1 mg/day A to 20 mg/day tam, each given for 5 years, were reported at ASCO 2015. B-35 demonstrated a statistically significant benefit in breast cancer free interval for women assigned to A, primarily in women <60 years. A secondary endpoint of B-35 was quality of life (QOL) and symptom (SX) outcomes in the two treatment groups. The primary hypotheses of the PRO study were that there would be no differences in QOL between the two treatments, and that patients receiving A would report higher rates of hot flashes compared to patients receiving tam. Other SX comparisons were secondary endpoints. Methods QOL and SX were assessed at baseline (prior to randomization), and every 6 months thereafter for 5 years of treatment and in the following 12 months. QOL was measured with the SF-12 Physical Component Summary (PCS) and Mental Component Summary (MCS). SX were measured with selected scales from the BCPT symptom-checklist, and other standardized instruments. Stratification was by age (<60 v ≥60) as in the main trial. Study hypotheses and endpoints were examined by comparing PROs in the two treatment arms using a mixed model for repeated measures analysis with adjustment for the baseline scores, time point and age category, using an intention-to-treat principle and including only patients who completed the baseline and at least one follow-up questionnaire. Patients with protocol events were censored. Only data through 60 months are reported here. The accrual goal for the sub-study was 1,150 consecutive patients. Results Between January 6, 2003 and June 15, 2006, a total of 3,104 patients were entered and randomly assigned to NSABP Protocol B-35. Accrual to the PRO study of B-35 closed on December 28, 2004, at which time 1,275 patients were entered, with 1,193 patients included in this analysis. There were no medical or demographic differences between patients assigned to A or tam in the PRO sub-study, and they reflected the characteristics of the parent trial. Adherence to data collection across the 60 months was 87%. There were no significant differences in QOL outcomes by treatment for the PCS (p=0.16) or the MCS (p=0.38). SX subscales: hot flash scale was greater in tam group and this difference varied over time (p=0.001); musculoskeletal pain was significantly greater in A group for time points 6-24 months (all p<.001); vaginal problems were greater in A group (p=0.03). Hot flash and vaginal problems were significantly worse in women <60 years. Additional SX outcomes (depression, fatigue, sexual function) will be reported at presentation. Conclusion In this large, double-blind, placebo-controlled trial comparing A to tam in patients with DCIS, there was no significant difference in QOL between the two treatments. However, there were important treatment differences in SX outcomes, which should be considered as part of treatment decision-making discussions, along with the clinical breast cancer outcome results. Support: CA-180868, 180822, 189867, 196067, 114732; AstraZeneca Pharmaceuticals LP. Citation Format: Ganz PA, Cecchini RS, Julian TB, Margolese RG, Costantino JP, Vallow LA, Albain KS, Whitworth PW, Cianfrocca ME, Brufsky A, Gross HM, Soori GS, Hopkins JO, Fehrenbacher L, Sturtz K, Wozniak TF, Seay TE, Mamounas EP, Wolmark N. Patient-reported outcome (PRO) results, NRG Oncology/NSABP B-35: A clinical trial of anastrozole (A) vs tamoxifen (tam) in postmenopausal patients with DCIS undergoing lumpectomy plus radiotherapy. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr S6-04.

Abstract Background Trastuzumab (T) combined with chemotherapy improves survival of women with HER2-positive breast cancer. Dual therapy with T and other HER-2 targeted agents (lapatinib/pertuzumab) has demonstrated improved efficacy compared to T alone in advanced breast cancer, and is in clinical trials in early stage disease. We hypothesize that subgroups of patients with HER2-positive tumors and limited tumor burden (tumor size and no or few positive lymph nodes) have a favorable outcome and would not be candidates for dual therapy. This study will analyze the disease-free survival (DFS)and overall survival (OS) of patients with early stage breast cancer treated with chemotherapy and T in the seminal randomized studies (RCTs). Methods RCTs were identified by a Medline search from 2004-2013. Trial groups from five phase III RCTs agreed to provide data. TrialNumberEligibilityRegimenNSABP B312130T1-3; pN1, pN2a, or pN3a; M0ACx4 then Tx4 VSACx4 then THx4 then Hx40wkNCCTG N98313505T1-3; pN1-2; M0ACx4 then Tx12wk VS ACTx4 then Hx52wk VS ACx4 then Tx12wk then Hx40wkHERA5102T1-3; pN1-2; M0Any (neo)adjuvant regimen, then H q3wk x1yr VS H q3wk x 2yr VS observationFinHER232T1-4; N0-3; M0Vx3 or T*x3 then FECx3 VS Vx3 or T*x3 with Hx9wk then FECx3PACS043010(528 HER2+)T1-3; pN1-2; M0FE(100)Cx6 VSFE(100)Cx6 then Hx1yr VS T*E(75) x6 VS T*E(75) x6 then Hx1yrA, adriamycin; C, cyclophosphamide; T, paclitaxel; T*, docetaxel; H, trastuzumab; P, carboplatin; F, 5-fluorouracil; E, epirubicin; Treatment regimen also included 5 years of hormonal therapy for ER positive patients. The ER+/HER2+ and the ER-/HER2+ cohorts will be analyzed separately. Patients with tumors 3 cm or smaller (T1a, T1b, T1c and >2 cm- 3cm) and 0, 1, 2 and 3 positive lymph nodes will be included. Initially a log rank test will be used to determine if the survival results are different by study. If there is no difference across studies, Kaplan-Meier survival curves with confidence intervals will be estimated. If there is a difference between studies, individual Kaplan-Meier curves will be estimated and the mean of the curves will be estimated along with confidence intervals using the adjusted standard deviation. Estimates at 3, 5 and 8 years will be presented as well as estimates of the event rates within each subpopulation defined by tumor size, number of positive lymph nodes, and ER status. Survival endpoints of DFS, and OS as defined in the original studies will be analyzed. Results and Conclusion Data from the five trials will be analyzed and an abstract provided prior to the meeting. Subgroups of patients with a very favorable outcome when treated with adjuvant chemotherapy and T will be identified. Our survival estimates might serve as benchmarks for future trials evaluating therapy reduction. Our results will complement those of the Dana Farber single arm multicenter trial in patients with “low risk” HER2-positive disease (APT trial; NCT00542451), being submitted to SABCS 2013. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr S6-03.

Abstract Bisphosphonates reduce the incidence of skeletal-related events (fractures, pain, hypercalcemia) in patients (pts) with bone metastases from breast cancer. By inhibiting osteoclast function and subsequent bone turnover they may inhibit the growth of bone (and other) metastases. Their role in preventing or delaying the development of bone (or other) metastases in pts with early breast cancer is uncertain. Three previous trials of oral clodronate have given mixed results. The largest trial (placebo-controlled) and a smaller open-label trial suggested that oral clodronate benefitted pts with improved bone metastases-free survival (BMFS) and overall survival (OS), but a third open-label study showed no benefit with an apparent detrimental effect on survival. Studies of IV zoledronate in open-label trials in early breast cancer have also shown mixed results: the AZURE trial in pts with node positive breast cancer showed no benefit in disease-free survival (DFS) or OS with a possible effect in pts aged over 60. The ABCSG-12 trial in GnRH (plus tamoxifen or anastrozole) treated women with early breast cancer showed a small DFS benefit for those receiving IV zoledronate but no BMFS or OS benefit. Methods: B-34 is a prospective, randomized, double-blind, phase III clinical trial in pts with stage 1, 2 or 3 breast cancer assessing oral clodronate 1600mg daily for 3 years compared to placebo given alone or in addition to adjuvant chemo- or hormone therapy. Stratification is by age (<50, +50), number of positive nodes (0, 1–3, 4+) and ER/PR status. The primary end-point is DFS. Secondary endpoints are the incidence of skeletal metastases, OS, relapse-free survival, incidence of non-skeletal metastases, and incidence of skeletal morbid events. Results: 3323 patients were accrued — 1662 in Group A and 1661 in Group B. As of March 31, 2011, 54 pts were declared ineligible: 23 in Group A and 31 in Group B. The average time on study is 101.3 months. Patient characteristics were evenly distributed throughout both groups. Approximately 75% of pts in each group had negative axillary lymph nodes and some 78% in each group were ER and/or PR positive. About 64% of pts were age 50 or over. Oral clodronate was generally tolerable, toxicities observed being mainly due to the concomitant systemic chemotherapy. One pt in Group A had a 1 mm area of osteonecrosis on the palatal taurus. Compliance, as expected with trials of oral medications, has been a problem. 1910 pts completed 3 years of therapy. Event rates were slower than anticipated. This final analysis assesses a total of 598 events in the two arms. Conclusions: This trial is the largest placebo-controlled study of an oral bisphosphonate in patients with early breast cancer and will provide further information on the role of bisphosphonates in breast cancer management. Supported for NSABP by NCI U10CA12027, −37377, 69651, 69974; and Bayer Schering Pharma Oy; for ECOG by U10CA021115; for NCCTG by U10CA25224; for SWOG by U10CA38926. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr S2-3.

Abstract Background: NSABP P1 and P2 are the largest prevention trials of tamoxifen and raloxifene in women at high risk of developing breast cancer. Both agents have been shown to decrease, but not eliminate breast cancer risk in these women. The goal of our study was to identify single nucleotide polymorphisms (SNPs) associated with development of breast cancer in high-risk women treated with tamoxifen or raloxifene. Methods: We performed a nested matched case-control genome-wide association study (GWAS) in subjects entered on the P2, which included only postmenopausal women, and subjects at least 50 years old entered on the P1 phase III prevention trials who were treated with tamoxifen or raloxifene. Cases were defined as subjects who developed invasive breast cancer or ductal carcinoma in situ and each was matched to two controls based on trial (P1, P2), age at trial entry, Gail risk score, history of lobular carcinoma in situ, and time on study. The study was restricted to the 94% of subjects self-identified as Caucasian. Genotyping was performed with the Illumina Human610-Quad chip. Eigenstrat analyses were performed to control for population stratification. SNPs were imputed within 200 Kb on either side of regions containing SNPs with p ≤ 3E-05. CYP2D6 genotyping was performed on all subjects. Results: The GWAS included 592 cases and 1171 controls. 547,356 SNPs were used in the analyses after removing 14,019 for call rates less than 98%, 30,560 for minor allele frequencies <1%, and 301 with Hardy-Weinberg equilibrium p-value <1E-06. Eigenvectors did not alter the results when used as covariates. Eleven genotyped SNPs with p-values ranging from 2.12E-06 to 9.98E-06 (conditional logistic regression with eigenvector adjustment) were identified; and 22 SNPs with p < 4E-05 were identified by imputation around these SNPs on chromosomes (Chr) 4, 8, 9, 13, 16 and 22, which were then genotyped with the Invader platform. Following imputation and genotyping, initial functional genomic studies have focused on 8 SNPs on Chr 16 (p-values 1.81E-06 to 9.55E-06, with the lowest p-value for an imputed and fine mapped SNP), all of which were in ZNF423, a gene known to encode a protein that associates with SMAD family members. In lymphoblastoid cell lines genotyped for the 2 top SNPs in ZNF423 (rs8060157 and rs11076499) and stably transfected with estrogen receptor-α, the variant and wild type SNPs displayed striking differences in estradiol-induced expression of ZNF423, SMAD3, BRCA1, and BRCA2. Conclusions: This GWAS revealed SNPs associated with the development of breast cancer in high-risk women treated with tamoxifen or raloxifene. Genome-wide significance was not achieved, but SNPs in ZNF423 identified on Chr 16 were associated with differential expression of ZNF423, SMAD3, BRCA1, and BRCA2 in response to estradiol in vitro, providing direction for future research. (Supported in part by NIH grants U01GM61388, U01GM63173, P50CA116201, U10CA77202, U10CA37377, U10CA69974, U24CA114732, and the Biobank Japan Project funded by the Ministry of Education, Culture, Sports, Science and Technology, Japan) Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr PD05-02.

Abstract Abstract #76 Background: QOL and MH outcomes were integrated into the NSABP B-30 trial as secondary outcomes to the efficacy analyses which are being presented separately. Explicit secondary aims of the NSABP B-30 study were 1) to compare toxicities among the regimens, 2) to compare QOL, and 3) to examine differences in amenorrhea and its relationship to symptoms, QOL, and efficacy. Here we examine the secondary aims of the study as a companion to the efficacy results that are presented separately.
 Materials and Methods: 5351 pts with cT1-3, N0-1, M0 were enrolled from 3/1/99 to 3/31/2004. 2170 were enrolled on the QOL study, and 2449 were enrolled on the MH study and were randomized to one of three treatment groups: Group 1 [doxorubicin (A) 60 mg/m2 and C 600 mg/m2 q 3 weeks (wks) x 4 followed by docetaxel (T) 100 mg/m2 q 3 wks x 4; Group 2 [A 50 mg/m2 and T 75 mg/m2 q 3 wks x 4]; Group 3 [A 50 mg/m2 T 75 mg/m2 and cyclophosphamide (C) 500 mg/m2 q 3 wks x 4]. All patients with ER-positive tumors received hormonal therapy after completing chemotherapy. Preliminary results from Group 1 have been reported previously (Swain, et al. Breast Cancer Res Treat, 2008).
 Results: The protocol specifies that 800 deaths are required for the definitive analysis of treatment, QOL, and MH outcomes, which are expected to occur by fall 2008. For this final report, results from a comparison of the three arms will be analyzed and presented. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 76.

Abstract Abstract #75 Background: The primary aims of this study were to determine 1) if a concurrent regimen of therapy containing cyclophosphamide compared to a sequential regimen containing cyclophosphamide would prolong disease-free survival (DFS) and overall survival (OS) and 2) if a regimen without cyclophosphamide was at least as efficacious as the two regimens containing cyclophosphamide. The secondary aims were to compare toxicities among the three regimens and to compare differences in amenorrhea in premenopausal women in each arm and its relationship to outcome.
 Materials and Methods: 5351 pts with cT1-3, N0-1, M0 were enrolled from 3/1/99 to 3/31/2004. 1783 pts were randomized to Group 1 [doxorubicin (A) 60 mg/m2 and C 600 mg/m2 q 3 weeks (wks) x 4 followed by docetaxel (T) 100 mg/m2 q 3 wks x 4], 1784 pts to Group 2 [A 50 mg/m2 and T 75 mg/m2 q 3 wks x 4], and 1784 to Group 3 [A 50 mg/m2 T 75 mg/m2 and cyclophosphamide (C) 500 mg/m2 q 3 wks x 4]. All patients with ER-positive tumors received hormonal therapy after completing chemotherapy. The protocol specifies that 800 deaths are required for the definitive analysis. This is expected to occur by Fall 2008.
 Results: Mean time on study is currently 79 months. At entry, 46% of pts were < 49 yrs; 54% were > 50 yrs. 82% were white, 10% black, 4% Hispanic. 66% had 1-3 positive nodes, 26% had 4-9, and 8% had 10+. 75% of pts were ER pos. 49% had lumpectomy, 51% mastectomy, and 24% local/regional radiotherapy. Mean tumor size was 2.5 cm. The presentation will include the first results of DFS, OS, and relevant toxicities from the trial. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 75.