Citations

5008 Background: Androgen receptor (AR) signaling is an important growth mechanism in mCRPC, providing the rationale for treatment with AR axis inhibitors such as ENZ and AAP. Targeting AR with anti-androgens such as ENZ can result in compensatory autocrine and paracrine androgenic stimulation. Therefore, using ENZ with the androgen biosynthesis inhibitor AAP to dampen these resistance mechanisms could improve clinical outcomes relative to ENZ alone. Methods: Men with progressive mCRPC by Prostate Cancer Working Group 2 criteria were eligible. Prior treatment with taxanes for mCRPC and any prior treatment with ENZ or AAP was exclusionary. Patients (pts) were randomized 1:1 to ENZ or ENZ/AAP at standard FDA-approved doses. Randomization was stratified by prior chemotherapy and Halabi prognostic three risk groups. Castrating therapy was maintained. The primary endpoint was overall survival (OS) defined as the date of randomization from date of death or last follow-up. The log-rank test had 90% power to detect a hazard ratio for OS of 0.77 with a one-sided type I error rate of 0.025. Secondary endpoints included radiographic progression free survival (rPFS) and on-treatment PSA declines. Exploratory endpoints included imaging changes, and changes in serum biomarkers such as androgens, angiokines, and circulating microRNA and RNA. The primary analysis was based on the stratified log-rank test adjusting on the stratification factors. Results: Between January 2014 and August 2016, 1311 men were randomized: 657 to ENZ and 654 to ENZ/AAP. Groups were well balanced between arms, including stratification variables. 15.6% of pts were high risk, 35.3% intermediate, and 48.1% low. Median OS was 33.6 mo (95% CI 30.5-36.4) and 32.7 mo (29.9-35.4) respectively, two-sided p = 0.53. Fifty percent PSA decline rate was 80% vs. 76.5%. Grade 3-5 adverse events (AE) (all attributions) were 55.6% and 68.8% respectively. Treatment discontinuation due to AEs occurred in 5% and 12%, pt withdrawal in 5% and 13%, and progression or death in 57% and 48% of pts respectively. Conclusions: Addition of abiraterone acetate to enzalutamide did not prolong survival in men with mCRPC. The combination resulted in more AEs than enzalutamide alone. Support: U10CA180821, U10CA180882, U24CA196171; https://acknowledgments.alliancefound.org . Clinical trial information: NCT01949337.

4502 Background: Patients with no evidence of disease (NED) after metastasectomy for metastatic renal cell carcinoma (mRCC) are at high risk of recurrence, but no systemic therapy has been shown to benefit this population. Pazopanib is an inhibitor of VEGFR and other kinases that improves progression-free survival in patients with measurable RCC metastatic disease. We performed a randomized, double-blind, placebo-controlled multicenter study to test the hypothesis that pazopanib would improve disease-free survival in patients with mRCC rendered NED after metastasectomy Methods: Patients with NED following metastasectomy were randomized 1:1 to receive pazopanib starting at 800 mg daily vs. placebo for 52 weeks. Patients were stratified by 1 vs. > 1 site of resected disease, and by disease-free interval ≤ vs. > 1 year. Clinical assessment for toxicity and patient-reported outcomes were performed every 4 weeks, and restaging scans every 12 weeks. The study was designed to observe a 42% improvement in disease-free survival (DFS) from 25% to 45% at 3 years. Results: From August 2012 to July 2017, 129 patients were enrolled. The study was unblinded after 83 DFS events had been observed (92% information). The median follow-up from randomization was 30 months (range 0.4 – 66.5 months). The study did not meet the primary endpoint: hazard ratio (95% CI) for DFS was 0.85 (0.55, 1.31) p= 0.47 in favor of pazopanib. At the time of unblinding, 22/129 (17%) of subjects had died. The HR for overall survival (OS) was 2.65 (1.02, 6.9) in favor of placebo ( p= 0.05). Patient-reported outcomes and laboratory correlates will be reported separately. Conclusions: 52 weeks of pazopanib did not improve DFS compared to blinded placebo in patients with mRCC who were NED after metastasectomy. There was a trend toward worse overall survival with pazopanib. Clinical trial information: NCT01575548.

3144 Background: Updated ASCO/CAP Guidelines for HER2 testing in breast cancer have been most impactful on the resolution of certain challenging groups of FISH results. We review the change in assignment of HER2 status in a large series of breast cancers referred to a large national reference laboratory for FISH testing since the introduction of the 2018 updated guidelines. Methods: Patient samples submitted to the Mayo Clinic Cytogenetics Laboratory (N = 2208) were analyzed by FISH. Samples with Group 2, Group 3, or Group 4 FISH results were reflexed to immunohistochemistry (IHC) in our central laboratory; FISH slides for those cases with equivocal 2+ IHC results were re-scored in the regions of invasive cancer showing more intense membranous staining. A subset of 202 samples with Group 4 FISH results were also reflexed to the previously employed reflex FISH assay (HER2/D17S122), and these were also re-analyzed according to the new reflex IHC/FISH process. Results: 382 of 2208 breast cancer samples tested (17.3%) had FISH results categorized as Group 2 (N = 17, 0.8%), Group 3 (N = 34, 1.5%), or Group 4 (N = 331, 15%) and required reflex IHC testing, and of those, 75% were 2+ equivocal and required targeted re-analysis of the FISH slide according to the 2018 updated guidelines. Re-analysis of the FISH slide resulted in switching between Groups 1-5 in 19.4% of cases, but HER2 status was changed by FISH re-scoring in only 7.7% of cases re-scored (1.0% of all samples), generally due to only minor shifts in HER2 copy number and HER2/control ratios between the initial and IHC-guided reflex FISH scores. In the subset of 202 cases tested by both reflex methods, the previously employed HER2/D17S122 reflex probe set was positive in 123 cases (60.9%), whereas reflex IHC/FISH was positive in only 10 cases (7.9%). Including positive reflex IHC (0.4%) and positive reflex FISH results (2.1%), the overall assignment of positive HER2 status on our series of 2208 cases was 11.5%. Conclusions: Overall rates of HER2 positive FISH results have declined under the most recent ASCO/CAP guideline update as a consequence of new recommendations for reflex testing for Groups 2-4. This change is largely due to reassignment of Group 2 and Group 4 results as negative in the absence of positive IHC.

1562 Background: In Caucasian-predominant populations, overweight or obese NSCLC pts (BMI≥25-kg/m2) have better prognosis while underweight (BMI≤18.5-kg/m2) pts have a worse prognosis. A large pooled sample allowed us to evaluate the role of ethnicity in this BMI-NSCLC OS relationship. Methods: Using individual data, survival analysis was performed on 15 ILCCO studies, assessing the interactions between ethnicity and BMI on overall survival (OS). Adjusted Hazard Ratios (aHR) from Cox models and adjusted penalized smoothing spline plots were generated. Results: Among 13416 (77%) Caucasian, 2975 (17%) Asian, and 935 (5%) Black NSCLC pts analyzed, we confirmed that for all pts, being underweight at NSCLC diagnosis was associated with worse OS (aHR=1.68, CI 1.5-1.8, p<0.001), while overweight/obese pts had improved survival (aHR 0.89, CI 0.8-0.9, p<0.001), when compared to pts with normal BMI. In general, Black pts had poorer OS than Caucasian pts (aHR 1.26 CI 1.1-1.4, p<0.001). However, the BMI-OS relationship differed according to ethnicity (BMI-ethnicity interaction, p=0.009): Caucasian underweight pts had poorer OS (aHR 1.67, CI 1.5-1.8, P<0.001) while overweight/obese pts had improved OS (aHR 0.89, CI 0.8-0.9, P<0.001). In Asian pts, these two associations were aHR 1.15, CI 0.8-1.5, P=0.33, and aHR 0.91, CI 0.7-1.1, P=0.32, respectively. In Black pts, these two associations were aHR 1.06, CI 0.7-1.5, p=0.73 and 0.75, CI 0.6-0.8 p<0.001, respectively. For overweight/obese patients, as BMI rises from 25-kg/m2 through 45-kg/m2, the prognosis worsens for Caucasian pts, remains stable for Asian pts, but improves for Black pts. Conclusions: InCaucasian pts, being underweight had a greater negative impact on OS than for Asian or Black pts, while being obese had a greater beneficial impact in Blacks than in other ethnic groups. These ethnic differences likely reflect genetically-informed muscle/adipose tissue distributions, where Black pts may have less sarcopenic obesity than other ethnicities. In future prognostic studies, BMI relationships must account for ethnic differences.

1079 Background: Brain metastases (BM) are a common and fatal complication of breast cancer but survival varies widely based on various prognostic factors (PF). Hence, patient counseling and therapeutic decisions should be individualized. We previously published a prognostic index (Breast GPA) based on cohort A (1985-2007, n = 642), updated it with tumor subtype in cohort B (1993-2010, n = 400) and are now updating it with a larger contemporary cohort (C). Methods: A multi-institutional (19) multi-national (3) retrospective database of 2473 breast cancer patients with BM diagnosed from 1/1/2006-12/31/2017 was created and compared to our prior cohorts. Demographic, clinical, molecular factors, tumor subtype and treatment were correlated with survival. Kaplan-Meier survival estimates were calculated and compared with log-rank tests. Results: The median survival (MS) for cohorts A, B and C improved over time [12, 14 and 16 mo, respectively ( < 0.01)] despite the subtype distribution becoming less favorable: Luminal B (ER/PR/HER2+) decreased from 26% to 21%; HER2 (HER2+/ER/PR-) decreased from 31% to 17%, Luminal A (ER/PR+/HER2-) increased from 20% to 31%; Basal (ER/PR/HER2-) was unchanged at 24%.MS by subtype improved from 21 to 27 mo in Luminal B, 18 to 25 mo in HER2, 10 to 14 mo in Luminal A and 6 to 9 mo in Basal tumors. The number of BM was 1 in 35%, ≤4 in 67% and > 10 in 18%. PF significant for survival were tumor subtype, age, KPS, number of BM and extracranial metastases (ECM) (all < 0.01). Surprisingly, Hispanic women (7%) showed improved survival (p < 0.01). BRCA1 was mutated in 57/533 (11%) and those patients showed a trend (0.16) toward improved survival. Treatment patterns have changed: the use of whole brain radiation therapy decreased from 71% to 67% to 47% in cohorts A, B and C, respectively. Conclusions: Despite the shift to less favorable tumor subtypes, MS has improvedbut varies widely by diagnosis-specific PF. Compared to prior cohorts, number of BM and ECM were identified as new PF. Ethnic, genetic and treatment differences between the eras are apparent. The updated Breast GPA, based on these data, and the correlation between BRCA1 and tumor subtype will be presented.

1030 Background: Previous studies have shown that 44-71% of trastuzumab (T)-treated pts develop HER2-specific immunity (Clin Cancer Res 2007, 13:5133; Cancer Res 2016, 76:3702; Breast Cancer Res 2018, 20:52). M is an Fc-engineered mAb that shares similar HER2 binding and antiproliferative activity as T. The Fc region of M has been engineered for increased affinity to the activating FcγRIIIA (CD16A) and lower binding to the inhibitory FcγRIIB (CD32B), attributes that may enhance the mAb’s immune function, such as antigen presentation. Methods: HER2+ cancer pts who progressed on prior therapy received M (0.1-6 mg/kg QW for 3 of every 4 weeks [N = 34]; or 10-18 mg/kg Q3W [N = 32]) in phase 1 trial NCT01148849. PBMC and plasma were collected pre-dose and 50 days post-dose for 46 pts and > 4 years for 3 pts on long-term treatment. Response to HER2 or control antigens (Ag) was assessed by IFNγ ELISpot and antibody (Ab) ELISA. In 14 pts, T-cell antigen receptor (TCR) repertoire was assessed by immunosequencing. Results: Following M treatment, mean frequencies of IFNγ-producing T cells specific for intra- or extracellular fragments of HER2 increased by 2.5 to 6-fold (p < 0.0027, paired t test). Most (95%) of subjects responded to ≥2 of 6 (median = 5) HER2 Ag. Mean HER2-specific Ab concentration increased by 19-54% (p < 0.0001), with all subjects responding to ≥2 (median = 5) of the 6 Ag. A small 1.6-fold increase in IFNγ response to control CMV/EBV/Flu (but not tetanus or cyclin D1) peptides was observed; no increase in Ab response to control Ag was noted. Subsets of HER2-specific T-cell and Ab responses persisted during long-term treatment. Median TCR clonality increased by 54% (p = 0.003), with an average of 125 unique clones expanding, while overall TCR diversity remained unchanged (p = 0.19). Conclusions: Treatment of HER2+ cancer with M was associated with enhanced HER2-specific T-cell and Ab responses together with increased TCR clonality, indicative of a more focused T-cell repertoire. The high frequency of HER2-specific immunity in M-treated patients ( > 95%) is consistent with its enhanced Fc region contributing to linkage of innate and adaptive immune responses.

529 Background: The effect of metformin on sex hormones (SHs) levels that may impact breast cancer (BC) outcome, is unclear. We evaluated the effect of metformin on SHs in a subgroup of women enrolled in the placebo-controlled MA.32 trial, a phase III randomized study of nondiabetic subjects with T1-3, N0-3 BC randomized to receive metformin 850 mg po bid or placebo for 5 years. Methods: Our substudy was conducted on the group of post-menopausal women with estrogen receptor (ER) and progesterone receptor (PR) negative BC and not receiving endocrine treatment enrolled in the trial. Post-menopausal was defined as prior bilateral oophorectomy or > 12 months since last menses without prior hysterectomy. Fasting blood and adherence to study drug at baseline and 6 months was required. Sex hormone binding globulin (SHBG), free testosterone and estradiol serum levels were evaluated using Roche Modular competitive ECLIA (electrochemiluminescense immunoassay). We compared change from baseline to 6 months in estradiol, SHBG and free testosterone between study arms using Wilcoxon sum rank tests and regression models to adjust for baseline BMI and weight change. Results: 304 women were eligible, 135 metformin vs 169 placebo; tumor stage and prior (neo)adjuvant chemotherapy (98% in both arms) and HER2 targeted treatment were well balanced between arms. At baseline mean age was 58.2±6.9 vs 57.6±7.9 years, mean BMI 26.9±4.9 vs 28.6±6.2 Kg/m2, median estradiol 29.82 vs 31.01 pmol/L, SHBG 80.6 vs 73.7 nmol/L and free testosterone 0.02 vs 0.03 nmol/L in metformin vs placebo patients, respectively. In univariable analysis, median estradiol decreased significantly between baseline and 6 months on the metformin vs placebo arm (-4.81 vs 0 pmol/L, p = < 0.0001); this difference remained significant after controlling for baseline BMI (p = 0.0001) and weight change (mean weight change -1.8±3.5 vs 0.4±3.5 Kg, p = < 0.0001 for metformin vs placebo respectively) between baseline and 6 months (p = 0.0007). In contrast, there was no significant change in SHBG or free testosterone (median change SHBG -5.5 vs -5.9 nmol/L, p = 0.33; median change free testosterone 0 vs 0 nmol/L, p = 0.33 for metformin vs placebo respectively). Conclusions: Metformin lowered estradiol levels, independent of weight change in non-diabetic post-menopausal women with ER and PR negative BC enrolled onto MA.32 trial. This observation suggests a new metformin action that has potential relevance to BC management.

Survival outcomes for older adults with acute lymphoblastic leukemia (ALL) are poor and optimal management is challenging due to higher-risk leukemia genetics, comorbidities, and lower tolerance to intensive therapy. A critical understanding of these factors guides the selection of frontline therapies and subsequent treatment strategies. In addition, there have been recent developments in minimal/measurable residual disease (MRD) testing and blinatumomab use in the context of MRD-positive disease after therapy. These NCCN Guidelines Insights discuss recent updates to the NCCN Guidelines for ALL regarding upfront therapy in older adults and MRD monitoring/testing in response to ALL treatment.

Abstract Introduction Flight surgeons play a vital role in U.S. Air Force aviation operations by ensuring that pilots are medically prepared to meet the demands of military aviation. However, there is natural tension between pilots and flight surgeons. A pilot may be reluctant to share medical information with a flight surgeon who could negatively impact the pilot’s career or flight status. In this preliminary study, we sought to identify pilot-perceived strengths and weaknesses in the relationship between U.S. Air Force aviators and their flight surgeons. Materials and Methods An online survey regarding pilot–flight surgeon confidence and perceived values was distributed electronically to a convenience sample of U.S. Air Force aviators. Participants included U.S. Air Force active duty and Air Reserve Component (Air Force Reserve and Air National Guard) military aviators in addition to U.S. Air Force Academy aviation cadets. Results One hundred and seventy-three aviators participated in the survey. Respondents reported variable comfort in approaching flight surgeons with medical concerns and suggested that they believed other pilots might be withholding medical information from flight surgeons or seeking care from civilian physicians for career protection. Conclusions We sought to examine the pilot–flight surgeon relationship and its impact on daily flying operations. While limited, results suggest that there may be gaps in trust between pilots and their flight surgeons. These findings could present an opportunity to improve the pilot–flight surgeon relationship by identifying factors that contribute to closer pilot–flight surgeon relationships.