Muss, Hyman B.

PURPOSE In light of evolving evidence that some patients with node-positive estrogen receptor–positive (ER+) disease may receive less benefit from chemotherapy, this study reports 12-year outcomes of the C9741 trial overall, and by the sensitivity to endocrine therapy (SET2,3) test index, a biomarker measuring endocrine transcriptional activity, to identify patients most likely to benefit from dose-dense chemotherapy. METHODS In all, 1,973 patients were randomly assigned to dose-dense versus conventional chemotherapy. Hazard ratios (HRs) for prognosis and for predictive interaction with chemotherapy schedule were estimated from Cox models of long-term disease-free survival (DFS) and overall survival (OS). SET2,3 was tested on the 682 banked RNA samples from ER+ cancers. RESULTS Dose-dense chemotherapy improved DFS in the overall study population by 23% (HR, 0.77 [95% CI, 0.66 to 0.90]) and OS by 20% (HR, 0.80 [95% CI, 0.67 to 0.95]); the benefits of dose-dense therapy were seen for ER+ and ER-negative subsets, without significant interaction between treatment arm and ER status. Low SET2,3 status was highly prognostic, but also predicted improved outcomes from dose-dense chemotherapy (interaction P = .0998 for DFS; 0.027 for OS), independent of menopausal status. Specifically, low endocrine transcriptional activity predicted benefit from dose-dense chemotherapy, whereas tumor burden and proliferation-driven signatures for molecular subtype classification did not. CONCLUSION At 12-year follow-up, C9741 confirmed the sustained long-term benefit of adjuvant dose-dense chemotherapy for node-positive breast cancer. SET2,3 identified patients with ER+ breast cancer who benefited from dose-dense chemotherapy, and specifically, this benefit was predicted by low endocrine activity in the cancer, rather than tumor burden, molecular subtype, or menopausal status.

ImportanceBlack women in the United States have higher breast cancer (BC) mortality rates than White women. The combined role of multiple factors, including body mass index (BMI), age, and tumor subtype, remains unclear.ObjectiveTo assess the association of race and ethnicity with survival among clinical trial participants with early-stage BC (eBC) according to tumor subtype, age, and BMI.Design, Setting, and ParticipantsThis cohort study analyzed survival data, as of November 12, 2021, from participants enrolled between 1997 and 2010 in 4 randomized adjuvant chemotherapy trials: Cancer and Leukemia Group B (CALGB) 9741, 49907, and 40101 as well as North Central Cancer Treatment Group (NCCTG) N9831, legacy groups of the Alliance of Clinical Trials in Oncology. Median follow-up was 9.8 years.ExposuresNon-Hispanic Black and Hispanic participants were compared with non-Hispanic White participants within subgroups of subtype (hormone receptor positive [HR+]/ERBB2[formerlyHER2] negative [ERBB2−],ERBB2+, and HR−/ERBB2−), age (<50, 50 to <65, and ≥65 years), and BMI (<18.5, 18.5 to <25.0, 25.0 to <30.0, and ≥30.0).Main Outcomes and MeasuresRecurrence-free survival (RFS) and overall survival (OS).ResultsOf 9479 participants, 436 (4.4%) were Hispanic, 871 (8.8%) non-Hispanic Black, and 7889 (79.5%) non-Hispanic White. The median (range) age was 52 (19.0-89.7) years. Among participants with HR+/ERBB2− tumors, non-Hispanic Black individuals had worse RFS (hazard ratio [HR], 1.49; 95% CI, 1.04-2.12; 5-year RFS, 88.5% vs 93.2%) than non-Hispanic White individuals, although the global test for association of race and ethnicity with RFS was not significant within any tumor subtype. There were no OS differences by race and ethnicity in any subtype. Race and ethnicity were associated with OS in young participants (age <50 years; globalP = .008); young non-Hispanic Black participants (HR, 1.34; 95% CI, 1.04-1.71; 5-year OS, 86.6% vs 92.0%) and Hispanic participants (HR, 1.62; 95% CI, 1.16-2.29; 5-year OS, 86.2% vs 92.0%) had worse OS than young non-Hispanic White participants. Race and ethnicity were associated with RFS in participants with BMIs of 25 to less than 30, with Hispanic participants having worse RFS (HR, 1.81; 95% CI, 1.23-2.68; 5-year RFS, 83.2% vs 87.3%) than non-Hispanic White participants.Conclusions and RelevanceIn this cohort study, racial and ethnic survival disparities were identified in patients with eBC receiving standardized initial care, and potentially at-risk subgroups, for whom focused interventions may improve outcomes, were found.

Abstract Background: Previous studies have demonstrated poorer survival of Black women with breast cancer. We assessed whether race/ethnicity was associated with disease-free (DFS) and overall survival (OS) among women with breast cancer enrolled in clinical trials for early-stage breast cancer according to tumor subtype, age, and body mass index (BMI). Methods: 10,011 women enrolled in one of four adjuvant chemotherapy trials: CALGB 9741, CALGB 49907, CALGB 40101, or NCCTG N9831. 9918 participants had available DFS and/or OS data and were included in the analysis. Cox models were used to estimate multivariable-adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between race/ethnicity and DFS and OS. We compared Non-Hispanic (NH) Black (n=871), Hispanic (n=436), and other race participants (n=283) to NH Whites (n=7889). We assessed associations within strata of age group (<50, 50-<65, or ≥65), tumor subtype (hormone receptor (HR)+/HER2-, HR-/HER2+, or HR-/HER2-), and BMI (<25, 25-<30, or ≥30). Results: In multivariable-adjusted models, NH Black patients under 50 years of age had worse DFS compared to NH White patients (HR: 1.34, 95% CI: 1.10-1.62) and worse OS (HR: 1.64, 95% CI: 1.30-2.07). The differences in DFS and OS persisted in patients ages 50 to <65, though there were no significant differences in DFS or OS between NH Black and NH White patients ages ≥65. Among Hispanic and NH White participants, younger age at diagnosis was associated with greater DFS compared with older age overall while this was not true for NH Black patients. Among patients with HR+/HER2- tumors, NH Black patients when compared to NH White patients had worse DFS (HR 1.33, 95% CI: 1.04-1.70) but there was not a significant difference in OS (HR 1.35, 95% CI: 1.00-1.83). DFS and OS for other tumor subtypes did not significantly differ by race. Among patients with BMI <25, NH Black patients had significantly worse DFS (HR: 1.70, 95% CI: 1.25-2.30) and OS (HR:1.76, 95% CI:1.20-2.58) compared to NH White patients. There was no difference in survival between different race/ethnicity groups among individuals with BMI ≥25. Conclusions: Our results identified subgroups that may contribute to the observed disparities in survival between NH Black and NH White women with early-stage breast cancer. The greatest disparities are among individuals <50 years of age, those with HR+/HER2-, and those with BMI <25. These differences exist even within clinical trial populations with similar initial therapy, suggesting that disparities may be influenced by inequities in survivorship care and long-term treatment, such as endocrine therapy adherence and persistence, and/or differences in tumor or host biology. Support: U10CA180821 and U10CA180882; ClinicalTrials.gov Identifiers: NCT00003088, NCT00005970, NCT00024102, NCT00041119; https://acknowledgments.alliancefound.org Citation Format: Marla Lipsyc-Sharf, Karla V. Ballman, Jordan D. Campbell, Hyman B. Muss, Edith A. Perez, Lawrence N. Shulman, Lisa A. Carey, Ann H. Partridge, Erica T. Warner. Role of age, BMI, and tumor subtype in racial/ethnic disparities in breast cancer survival: A pooled analysis of four Alliance adjuvant clinical trials [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 494.

PURPOSE This study determined whether an electronic version of the geriatric assessment is feasible in a multi-institutional, diverse setting. METHODs Ten sites within the Alliance for Clinical Trials in Oncology participated. Patients who had active cancer or a history of cancer and were 65 years of age or older were eligible. The geriatric assessment was completed with an electronic data capture system that had been loaded onto iPads. Feasibility was defined a priori as completion in at least 70% of patients either with or without help. To enhance racial diversity, the original sample size was later changed and augmented by 50% with the intention of increasing enrollment of older minority patients. RESULTS A total of one hundred fifty-four patients were registered with a median age of 72 years (range, 65-91 years). Forty-three (28%) identified themselves as African American or Black. One hundred forty-one patients (92%) completed the electronic geriatric assessment. Feasibility was observed across all subgroups, regardless of race, education, performance status, comorbidities, and cognition; 124 patients (81%) completed the geriatric assessment without help. Reasons for not completing the geriatric assessment are as follows: clinic visit did not occur (n = 6), no iPad connection to the Internet (n = 3), patient declined (n = 2), prolonged hospitalization (n = 1), and patient died (n = 1). Reasons for needing help, as reported by study personnel, were as follows: the patient preferred that research personnel ask the questions (n = 9), vision problem (n = 3), lack of comfort with the iPad (n = 2), questions were not clear (n = 1), less proficient in English (n = 1), and challenge in pressing the green button to go to the next question (n = 1). CONCLUSION The electronic geriatric assessment is feasible in a multi-institutional setting that includes a notable proportion of African American or Black patients.

Abstract Cancer is a disease of aging and, as the world's population ages, the number of older persons with cancer is increasing and will make up a growing share of the oncology population in virtually every country. Despite this, older patients remain vastly underrepresented in research that sets the standards for cancer treatments. Consequently, most of what we know about cancer therapeutics is based on clinical trials conducted in younger, healthier patients, and effective strategies to improve clinical trial participation of older adults with cancer remain sparse. For this systematic review, the authors evaluated published studies regarding barriers to participation and interventions to improve participation of older adults in cancer trials. The quality of the available evidence was low and, despite a literature describing multifaceted barriers, only one intervention study aimed to increase enrollment of older adults in trials. The findings starkly amplify the paucity of evidence‐based, effective strategies to improve participation of this underrepresented population in cancer trials. Within these limitations, the authors provide their opinion on how the current cancer research infrastructure must be modified to accommodate the needs of older patients. Several underused solutions are offered to expand clinical trials to include older adults with cancer. However, as currently constructed, these recommendations alone will not solve the evidence gap in geriatric oncology, and efforts are needed to meet older and frail adults where they are by expanding clinical trials designed specifically for this population and leveraging real‐world data.

9515 Background: CC 49907 showed superiority of standard therapy (cyclophosphamide/doxorubicin [AC] or cyclophosphamide/methotrexate/5-fluorouracil [CMF]) over capecitabine[C]. Dose adjustments made for renal insufficiency (RI) for methotrexate and C; ideal body weight used. Purpose was to analyze the relationship between RF at baseline and 5 endpoints: toxicity, dose modification, therapy completion, relapse-free survival [RFS] and overall survival [OS]. Methods: Pre-treatment RF was calculated (Cockcroft-Gault). Endpoints assessed by regimen. RF was tested as a dichotomous and continuous variable of stages 1,2 vs. 3,4 kidney disease (National Kidney Foundation). Logistic regression modeled the relationship between renal stage and the first three endpoints of toxicity, dose modification and therapy completion. Toxicity divided by hematologic or not. The relationship of RFS and OS with RF was assessed with the logrank test and as a continuous variable with Wald chi square. Results: 619 patients; incidence of stage 3/4 RI(<60 ml/min) was: CMF=72%; AC=64%; C=75%. With AC the incidence of toxicity differed by renal function. 31% of patients with poorer function >grade 3 non-hematologic toxicity vs. 14% with better function(p=0.011). There was a suggestion of effect of RF on OS and RFS for C-treated patients. RF was not associated with dose modification, premature therapy termination, RFS or OS for the CMF-treated patients. Conclusions: 1) AC: declining RF was associated with increased non-hematologic toxicity. 2)Patients with RI who received dose modifications were not at increased risk for complications in comparison to those who did not have renal insufficiency and received full dose. 3)Declining RF did not affect therapy completion. 4)C: suggestion that worse RF was related to poorer RFS or OS. 5)Exclusion of patients from clinical trials with RI based on concern of excessive hematologic toxicity may not be justified with appropriate modification. 6)Results should be considered in the design of clinical trials for older patients.

1007 Background: Determining optimal adjuvant chemotherapy for early stage breast cancer depends on efficacy and toxicity. We sought to determine if T is equivalent to AC but with reduced toxicity. Methods: Pts with operable breast cancer with 0-3 positive nodes were enrolled on a 2x2 factorial design study which addressed (1) superiority of 6 vs. 4 cycles of therapy (previously reported, Shulman, JCO 2012) and (2) equivalence of single-agent T to standard AC, defined as upper bound of 95% confidence interval (CI) of hazard ratio (HR) of T vs. AC < 1.30 for the primary endpoint of relapse-free survival (RFS). A planned target of 567 RFS events required 4,646 pts with 4 yrs FU. At activation in 2002, T (80mg/m2) was q1wk for 12 or 18 wks and AC (60/600 mg/m2) was q3wk for 4 or 6 cycles. In 2003 (570 pts enrolled) schedules were revised to 4 or 6 cycles q2wk for both T (175 mg/m2) and AC. The 6-cycle arms were dropped in 2008 (3,171 pts enrolled) due to slow accrual. Relative effectiveness of T to AC is shown by hazard ratio (HR). Logrank p-values are measures of discordance but are not relevant for the equivalence question and are not adjusted for multiple comparisons. Results: After enrolling 3,871 pts, the study closed in 2010 due to slowing accrual. With a median follow-up of 6.1 yrs there are 437 RFS events. The HR of 1.26 (95% CI: 1.05-1.53; p = 0.02) does not allow a conclusion of equivalence of T with AC. With 266 deaths the HR for overall survival (OS) is 1.27 (95% CI=1.00-1.62; p = 0.05), favoring AC. The estimated absolute advantage of AC at 5 yrs is 3% (91 vs. 88%) for RFS and 1% (95 vs. 94%) for OS. All 9 treatment-related deaths were in pts receiving AC and are included in the survival analysis. The incidence of Grade 3+ toxicity for AC vs T was 33% vs. 4% for hematologic toxicity and 36% vs 22% for non-hematologic toxicity. Conclusions: This trial did not show equivalence of T to AC, a conclusion that is very unlikely to change with additional follow-up. T was less toxic than AC. Clinical trial information: CDR0000069444/NCT00041119.

6015 Background: Comorbidity increases with age. We evaluated how comorbidity was associated with toxicity and clinical outcomes among older women with BC who received adjuvant chemotherapy. Methods: Cancer and Leukemia Group B (CALGB 49907) enrolled 633 women ≥65 years with early stage BC and randomized them to standard adjuvant chemotherapy (AC or CMF) or capecitabine (N Eng J Med 360:2055, 2009). 329 women on the Quality of Life companion study completed a pre-treatment health survey (Older American Resources and Services Questionnaire, physical health subscale). The survey evaluates 14 conditions and the degree to which each interferes with daily activities (rated from 1 “not at all” to 3 “a great deal”). Comorbidity was analyzed as follows: 1) total number 2) comorbidity burden score (summed conditions multiplied by interference rating); and 3) individual diseases. Primary outcomes were: 1) grade 3-5 toxicity (incident and cumulative); 2) time to relapse (TTR), and 3) overall survival (OS). Contingency table methods were used to evaluate the association between comorbidity and toxicity. Cox proportional hazards models were used to evaluate TTR and survival outcomes. Results: Among 329 women [median age 71 years (range 65-89), 86% white, 98% ECOG performance score 0-1, 75% stage 1-2] the median number of comorbidities was 2 (0-10), median comorbidity burden score was 3 (0-25), and most common conditions were arthritis (58%) and hypertension (54%). Few patients had diabetes (17%), heart disease (16%), and pulmonary disease (9%). No comorbidity measure was associated with toxicity or TTR. With a median follow-up of 5.4 years, increasing comorbidity was associated with shorter OS. For each additional comorbid condition the hazard of death increased by 18% (HR 1.18 [95% CI; 1.06-1.33]) after adjusting for age, tumor size, treatment, node status and receptor status. Comorbidity burden score was similarly associated with OS (HR 1.08 [95% CI; 1.03-1.14]), while no specific condition was independently associated. Conclusions: Among older women with good functional status, comorbidity was not associated with toxicity or BC relapse. However, comorbidity burden was associated with shorter OS.

564 Background: Our group previously examined if baseline patient/tumor characteristics, or prior treatment affected cause of death in MA.17, a placebo controlled extended adjuvant trial of the aromatase inhibitor (AI) letrozole. We now examine factor effects on all cause mortality in MA.27. Methods: MA.27 was an adjuvant phase III superiority trial of 5 yrs of exemestane vs anastrozole, in ER+ postmenopausal breast cancer accrued between 2003 and 2008; event free survival was similar. We examined by intention-to-treat, the multivariate time-to-breast cancer-specific (BrCa), cardiovascular (Cardio), and other causes (OT) of death with log-normal survival analysis adjusted by treatment and stratification factors (lymph node status, adjuvant chemotherapy, celecoxib, aspirin, and trastuzumab). We tested whether factors were associated with 1) all cause mortality, and if so, 2) cause-specific mortality. We also fit step-wise forward cause-specific adjusted models. Results: 7,576 women (median age 64.1 years; 5417 (71.5%) <70 yrs and 2159 (28.5%) >70 yrs) were enrolled and followed for a median of 4.1 yrs. The 432 deaths comprised: 187 (43.3%) BrCa, 66 (15.3%) Cardiovascular, and 179 (41.4%) OT MA.27 therapy was not associated with mortality (p=0.84). Five baseline factors were differentially associated with cause of death. Older age was associated with greater BrCa (p=0.03), Cardio (p<0.001), and OT (p<0.001) mortality. Pre-existing cardiovascular history led to worse Cardio mortality (p<0.001). Worse ECOG performance status led to worse OT death (p<0.001). T1 tumors were associated with less BrCa mortality (p<0.001). PgR+ tumors were also associated with less BrCa mortality (p<0.001). There were fewer BrCa deaths with Node –ve disease (p<0.001), ER+ tumors (p=0.001) and without adjuvant chemotherapy (p=0.005); there was worse Cardio mortality (p=0.01) with receipt of trastuzumab; worse OT (p=0.03) for non-whites, and without adjuvant radiotherapy (p=0.003). Conclusions: 56.7% of deaths in MA.27 patients were non-breast cancer related. We showed baseline patient and tumor characteristics, and prior treatment differentially affected cause of death. Clinical trial information: NCT00066573.

Abstract Background: Contemporary approaches to non-Hodgkin lymphoma (NHL) and chronic lymphocytic leukemia (CLL) incorporate chemo-immunotherapy, biologic combinations, and immune modulating agents; toxicities in elderly pts (³65 years) receiving these therapies are not well studied. Further, clinical and biologic factors predicting these toxicities in pts receiving biologic therapy remain undefined. Methods: We reviewed data on NHL and/or CLL pts treated prospectively on 14 studies by the Alliance for Clinical Trials in Oncology from 2004-2014 (Table 1). Toxicity was assessed per the NCI CTCAE at the time of trial enrollment, and the probabilities of experiencing grade 3 and grade 4 hematologic (hem) and non-hematologic (non-hem) toxicities were modeled as a function of age (³65 years vs. < 65), time on study, treatment (biologics only vs. biologic + chemotherapy), gender, race, LDH, performance status (PS), stage, and an age-by-treatment interaction using logistic regression. Results: A total of 1199 pts (409 age ³ 65; 790 age < 65; 736=CLL and 463=NHL) were included. Among these patients, 493 received only biologic therapy (166 age ³ 65; 327 age < 65; 104 CLL; 389 NHL), and 706 received biologic + chemotherapy (243 age ³ 65; 463 age < 65; 632 CLL; 74 NHL). Among CLL pts (259 pts ³ 65), the effect of age on the probability of experiencing a grade 3 heme toxicity differed by treatment type (age-by-treatment interaction p = 0.047). Specifically, the adjusted odds ratio (OR) (age ³ 65 vs. < 65) for pts receiving only biologic therapy was 3.075 (95% CI: 1,15-8.25), and that for pts receiving biologic + chemotherapy was 1.044 (95% CI: 0.69 - 1.57). Similar results were seen in CLL pts for grade 4 heme toxicities (age-by-treatment interaction p = 0.033; biologic OR 6.937, 95% CI: 1.76-27.35; biologic + chemo OR 1.484, 95% CI: 1.04-2.13). No such interactions were found in CLL pts for grade 3 and 4 non-hem toxicities; however, older pts had significantly higher odds of experiencing a grade 3 non-hem toxicity than younger pts (adjusted OR 1.40; p = 0.047; 95% CI: 1.004-1.96). No age group differences were found in CLL pts for grade 4 non-hem toxicity. Similar analyses were performed for NHL pts (150 pts ³ 65), but no significant age group differences were found. Among CLL pts, women had significantly higher odds of experiencing a grade 3 heme toxicity than men (OR 2.01; p = 0.0007); no difference in grade 3 non-hem or any grade 4 toxicity was noted. Non-Caucasian CLL pts had higher odds of experiencing a grade 4 non-hem toxicity than Caucasians (OR 2.892; p = 0.0029), but no other toxicity differences were found. Worse PS was associated with increased toxicities (OR 1.871; p = 0.0009: grade 3 heme; OR 1.647; p = 0.0025: grade 3 non-hem; OR 1.410; p = 0.0448: grade 4 heme, and OR 1.931; p = 0.0252: grade 4 non-hem). CLL pts with advanced stage disease had higher odds of experiencing a heme toxicity (grade 3: OR 1.95; p = 0.0007, grade 4: OR 1.451; p = 0.0329), but no stage associations were found for non-hem toxicities. Among NHL pts, men had significantly higher odds of experiencing a grade 4 hematologic toxicity than women (OR 4.351; p = 0.0169), but no other toxicity differences were found. An increase in LDH was associated with significantly higher odds of experiencing grade 3 non-hem and grade 4 heme toxicities (grade 3 non-heme: OR 1.633; p = 0.021, grade 4 heme: OR 2.039, p = 0.0182), but no such effect was found for grade 3 heme or grade 4 non-hem. Worse PS was associated with higher odds of experiencing grade 3 toxicities (heme: OR 2.025; p = 0.0344, non-hem: OR 2.458; p = 0.0013), but no such differences were found for grade 4 toxicities. No significant stage or race effects were found in NHL patients. Conclusion: In pts ³65 years who have CLL or NHL, we identified several clinical and disease-related factors as potential predictors of developing grade 3 and/or 4 heme and non-hem toxicities (Table 2). A prognostic model is being constructed to predict toxicities in these under-studied pts to guide management and monitoring. Further, the impact of these toxicities on outcomes is being analyzed and will be presented at the meeting. Disclosures Hurria: Celgene: Other: Research; Optum Health Care SOlutions: Consultancy, Other: Conference panel, research; Boehringer Ingelheim Pharmaceuticals: Consultancy; Sanofi: Consultancy; Carevive: Consultancy; Novartis: Other: Research; GTx, Inc: Consultancy. Bartlett:Gilead: Consultancy. Cheson:Gilead: Research Funding; Acerta: Membership on an entity's Board of Directors or advisory committees, Research Funding; Pharmacyclics: Membership on an entity's Board of Directors or advisory committees, Research Funding. Nabhan:Infinity: Consultancy; Abbvie: Consultancy; Cardinal Health: Consultancy; Seattle Genetics: Research Funding; Genentech: Consultancy, Research Funding; Astellas: Research Funding; Celgene Corporation: Consultancy, Research Funding.