Hillman, David W.

5053 Background: Genomic aberrations associated with resistance/response to AA/P are not known. In a prospective study we assessed whole-exome/RNA-seq based aberrations in CRPC metastatic biopsies for identifying molecular markers associated with primary resistance and response duration. Methods: Sequencing of metastatic biopsies was performed for analyzing molecular aberrations that predict primary resistance (defined as progression at 12-weeks of therapy (non-responders) using PSA, RECIST, bone scan criteria per PCWG2). Gene network analysis was performed in genes mutated more frequently in non-responders and in genes differentially expressed between non-responders and responders using a “risk ratio” (RR) of ≥2. Cox regression models with multiple gene network pathways were used for determining association with time to treatment change (TTTC). Results: Of 92 enrolled pts 82 had complete whole-exome, RNA-seq & 12-week outcome data available for analysis. At 12-weeks 33/82 had progressed. Using a RR of ≥2, 113 genes were more frequently mutated in non-responders & 292 in responders. In non-responders, gene network analysis revealed frequent mutations in Wnt/β-catenin pathway genes; frequent deletion of negative regulators of Wnt pathway ( DKK4, SFRP2, LRP6). Gene expression analyses revealed significantly reduced expression levels of Wnt/β-catenin pathway inhibitors and increased expression levels of cell cycle proliferation (CCP) genes in non-responders. Median study follow up was 32 months during which time 58/82 pts progressed and switched treatments. Median TTTC was 10.1 months (IQR:4.4-24.1). In multivariate analysis CCP scores of ≥50 predicted shorter TTTC (HR = 2.11, 95% CI: 1.17-3.80; p = 0.01). Conclusions: In metastases Wnt/β-catenin pathway activation is associated with primary AA/P resistance and increased CCP with acquired drug resistance. These findings offer molecular based predictive biomarkers in CRPC stage treatment. Clinical trial information: NCT#01953640.

533 Background: Current guidelines define TNBC as complete absence of estrogen (ER) and progesterone receptor (PR), without HER2 amplification. However, the prognostic impact of clinical and histopathological factors, sTILs, and AdjCT in TNBC meeting these strict criteria is unknown. Methods: From a cohort of 9985 women who underwent upfront surgery for M0 breast cancer (BC) at Mayo Clinic Rochester from 1985-2012, 1159 pts with ER negative or low (≤10%) BC were identified for central ER/PR/HER2 staining and HER2 FISH (IHC2+ only) to select those with TNBC by modern definitions. Cox proportional hazards models were used to assess the impact of clinicopathological variables on invasive disease-free (IDFS) and overall survival (OS). Results: Tumors from 605 pts (median age 56.3 yrs) met criteria for TNBC (ER < 1%, PR < 1% and HER2 0, 1 or 2+ and FISH negative). 51% were T1, 65% N0, 88% grade 3, and 75% had Ki67 > 15%. Histologically, 39% were anaplastic, 26% invasive ductal, 16% medullary, 8% metaplastic, 6% apocrine and 5% others. Median sTILs was 20% (0-90%). 55% pts received AdjCT [21% anthracycline (A), 19% A + taxane, and 15% other]. Median follow-up for IDFS and OS were 7.4 and 10.6 yrs, respectively. Multivariate analyses demonstrated that higher N stage (p < 0.01), lower sTILs (p = 0.01) and no AdjCT (p < 0.01) were independently associated with worse IDFS and OS. Histology (medullary subtype) was associated with better IDFS in univariate (HR 0.56, 95% CI, 0.35-0.89) but not in multivariate analyses, once sTILs were accounted for. Among systemically untreated pts (n = 182), higher N (p < 0.01) and lower sTILs (p = 0.04) were associated with worse IDFS. For systemically untreated T1N0 pts (n = 111), the 5-yr IDFS was 70% (95% CI, 61-81) [T1a: 83% (95% CI, 63-100), T1b: 68% (95% CI, 52-88) and T1c: 67% (95% CI, 55-83)], compared to 78% (95% CI, 68-84) for T1N0 pts treated with AdjCT. Conclusions: In TNBC pts, N stage, sTILs and receipt of AdjCT were independently prognostic for IDFS and OS. sTILs remained prognostic for IDFS in systemically untreated TNBC. In N0 TNBC, the risk of recurrence or death was substantial in the absence of chemotherapy, even for those with T1 tumors.

502 Background: Pre-specified 5-year analyses of the phase III Adjuvant Lapatinib and/or Trastuzumab Treatment Optimisation (ALTTO) Trial defined in Amendments 11&12. Methods: From June 2007 to July 2011, 8381 patients (pts) were randomised from 946 sites in 44 countries to receive either L+T, T→L, L, or T. In 2011, due to futility the L arm was closed and is not included in this analysis. The primary end point is disease-free survival (DFS). Secondary objectives include treatment comparisons with respect to overall survival (OS), time to recurrence (TTR), time to distant recurrence (TDR), cardiac and overall safety and tolerability. Primary analysis results of the study were published in JCO 2015 34:1034-1042. This updated analysis occurs at a 6.9 yrs median follow up (MFU). Results: All patients have reached 5-years of follow-up. 705 DFS events for L+T vs T have been observed. HR for DFS was 0.86 (95% CI, 0.74-1.00; 6-yr DFS%=85% vs 82%) for L+T vs T and 0.93 (95% CI, 0.81-1.08; 6-yr DFS%=84% vs 82%) for T→L vs. T. The 6-year OS was 93%, 92%, and 91% for L+T, T→L, and T, respectively. HR for OS was 0.86 (95% CI, 0.70-1.06) for L+T vs. T and 0.88 (95% CI, 0.71-1.08) for T→L vs. T. DFS differences for L+T vs. T were slightly higher for the hormone-receptor(ER)-negative [HR 0.80 (95% CI, 0.64-1.00; 6-yr DFS%=84% vs. 80%)] and the sequential chemotherapy [HR 0.83 (95% CI, 0.69-1.00; 6-yr DFS%=83% vs.79%)] subgroups. There were no differences in sites of first DFS events according to treatment arm for CNS, loco-regional, or distant recurrences. There were more AEs related to study treatment (L+T 93% vs T 64%). The incidence of primary cardiac end points was low: 1% for L+T, 0.5% for T→L and 0.9% for T. Conclusions: The HRs for this updated analysis are similar to those from the Primary Analysis and the event rate remains lower than anticipated (705 vs 850 planned). Cardiac toxicity remains low. This analysis suggests that HER2+/ER- tumors may have a different biology than HER2+/ER+ and may benefit more from dual HER2 blockade. Long-term follow up continues. Clinical trial information: NCT00490139.

262 Background: Inherited and somatic aberrations in DNA repair genes in castrate resistant prostate cancer (CRPC) are associated with poor prognosis, but respond well to poly ADP ribose polymerase (PARP) inhibitors. We evaluated the prevalence and prognostic impact of harboring germline DNA repair variants in hormone sensitive prostate cancer (HSPC). Methods: Germline DNA from buffy coat was sequenced on HiSeq4000 with a median coverage of 200X for DNA repair variants in 20 genes in HSPC and CRPC patients (pts) enrolled in a hospital registry. Pts were divided into two groups; Group A: pts enrolled at the time of CRPC stage; Group B: treatment naïve HSPC stage pts. The primary endpoints were to determine any impact of harboring DNA repair variants on time to progression from HSPC to CRPC and, from CRPC to death. Group A pts were retrospectively analyzed for time to progression from HSPC to CRPC while Group B patients were followed prospectively for outcomes. Statistical analysis included Cox proportional hazard models and Wilcoxon Rank sum test with significance at p≤0.05. Results: In Group A, 51/562 CRPC pts (9.07%) had variants in the 20 genes (most frequently in BRCA2; n = 15). 44/51 pts with variants and 399/511 without variants had died. Median time of progression from HSPC to CRPC with/without variants was 22.1 vs. 25.1 months (mths); p-value = 0.679. Median time from CRPC to death with/without variants was 32.2 Vs. 27.7 mths (p = 0.6). In HSPC Group B, 14/100 pts were identified with germline variants in ATM (n = 5), CHEK2 (n = 3), BRCA1 (n = 2), BRCA2 (n = 2), RAD50 (n = 1), and MSH2 (n = 1). 31/100 have died and median time to progression from HSPC to CRPC with/without variants was 15.6 vs.11.8 mths, p-value = 0.76. Conclusions: Pts with germline DNA repair variants detected in HSPC stage were not associated with poor prognosis. Presence of additional somatic DNA repair gene aberrations in cell-free DNA, not investigated in this cohort may add to the prevalence of DNA repair gene variations in HSPC and together impact prognosis adversely so as to provide a rationale for PARP inhibitor therapy in select HSPC stage pts.

249 Background: Serum CGA has been identified as a candidate prognostic biomarker for mCRPC. In a two cohort study, we compared the prognostic value of serum CGA with a validated CTC assay. Methods: In the discovery cohort (DC), blood samples were collected from 256 men with mCRPC. In an independent validation cohort (VC), 92 men with mCRPC were enrolled in a biospecimen collection study. In both cohorts, men receiving proton pump inhibitors and those with non-castrate levels of testosterone ( > 50ng/dl) were excluded. Serum CGA was measured in a homogeneous automated immunofluorescent assay using time-resolved amplified cryptate emission. In the VC, CTC enumeration was performed using the FDA cleared CELLSEARCH CTC test prior to treatment with abiraterone acetate/ prednisone. Cox proportional hazard regression and Kaplan-Meier analysis were performed for associations with elevated CGA (above reference range), unfavorable (≥ 5) CTCs, and overall survival (OS). Results: In the DC, 200 men were eligible for analysis. The median age was 72 years (yrs), 81/200 pts had a Gleason score (GS) ≥ 8, 34/200 had an elevated CGA. At a median follow up of 2.2 yrs, 156/200 were deceased. In the subset of men with GS ≥ 8, elevated CGA was associated with shorter OS [hazard ratio (HR) 2.19, p = 0.017]. In the VC, 71 men were eligible for analysis. The median age was 71 yrs, 36/71 tumors were GS ≥ 8, 31/71 pts had an elevated CGA, 26/66 had unfavorable CTCs (≥ 5). At a median follow up of 1.8 yrs, 31/71 were deceased. Elevated CGA (HR 1.91, p = 0.043) and unfavorable CTC counts (HR 2.97, p = 0.0012) were adversely associated with OS. In the high GS group, both CTCs and CGA had the same area under the curve (AUC) of 0.72. Pts with elevated CTC and CGA had the poorest OS (HR 3.76, p = 0.008). Conclusions: Elevated serum CGA was negatively associated with OS in men with mCRPC. Serum CGA represents a prognostic biomarker that may complement CTC enumeration. Clinical trial information: NCT#01953640.

Abstract Background: While TNBC remains the most aggressive type of breast cancer (BC), substantial heterogeneity in biology and outcomes exists among TNBC subtypes. Historically, risk stratification of TNBC has been based on anatomic factors such as tumor size, nodal involvement and presence of distant metastases. However, these features alone fail to accurately predict outcomes. Tumor immune infiltration (sTILs) and distribution of immune cell subsets in the perip heral blood (NLR) have emerged as variables reported to be associated with outcomes in TNBC. We sought to evaluate whether NLR and sTILs provided independent prognostic information in TNBC. Methods: From a cohort of 9,982 women who underwent BC surgery at Mayo Clinic, Rochester, MN between Jan 1985 and Dec 2012, we identified 605 centrally-confirmed TNBC tumors. Patients (pts) with prior BC, bilateral BC, non-invasive disease, stage IV, neoadjuvant therapy, endocrine therapy, or adenoid cystic histology were excluded. For eligible tumors, clinical and pathologic variables were evaluated, including peripheral blood NLR and central assessment of sTILs per the 2014 International TILs Working Group recommendations. We calculated the Pearson correlation coefficient (PCC) between NLR and sTILs and constructed Cox Proportional Hazards Models to evaluate their association with invasive-disease free (IDFS) and overall survival (OS). NLR and sTILs were both analyzed as continuous variables. Results: Most pts had T1-2 (95%) and N0-1 disease (86%). Median OS follow-up was 10.6yrs. Median IDFS was 12yrs (95%CI 10.2-16.7) and median OS was 18.8yrs (95%CI 15.6-20.8). NLR and sTILs were available in 408 and 599 pts, respectively. The median NLR and sTIL content were 2.29 (0.14-10.50) and 20% (0-90%), respectively. NLR and sTILs were poorly correlated (PCC 0.0237). On univariate analysis (UVA), a higher NLR was associated with worse IDFS (HR 1.13; 95%CI 1.02-1.26, p=0.02) and OS (HR 1.17; 95%CI 1.04-1.31, p=0.01). Each 1% increment in sTILs was associated with improved IDFS (HR 0.99; 95%CI 0.98-0.99, p&lt;0.001) and OS (HR 0.99, 95%CI 0.98-1.00, p&lt;0.001). Among pts with high sTILs (≥20%), a higher NLR remained significantly associated with worse IDFS (HR 1.21; 95%CI 1.05-1.38, p=0.007) and OS (HR 1.25; 95%CI 1.09-1.44, p=0.001). In contrast, among pts with low sTILs (&lt;20%), NLR was not associated with IDFS (HR 1.07; 95%CI 0.89-1.28, p=0.49) or OS (HR 1.07; 95%CI 0.88-1.30, p=0.49). The interaction test between NLR and sTILs did not reach statistical significance. A multivariate analysis (MVA; including age, menopausal status, histologic subtype, grade, tumor size, nodal stage, Ki-67, NLR, sTILs, adjuvant chemotherapy, type of surgery and adjuvant radiation) showed that sTILs remained independently associated with IDFS (HR 0.99, 95%CI 0.97-1.0, p=0.019) and OS (HR 0.99, 95% CI 0.97-1.0, p=0.044), whereas NLR did not. Conclusions: A lower NLR and a higher sTIL content were each associated with improved IDFS and OS among pts with nonmetastatic TNBC on UVA. However, when evaluated on a MVA, only sTILs remained independently associated with IDFS and OS. Our data suggest that the effect of sTILs on outcomes may not be modified by the NLR. Citation Format: Leon-Ferre RA, Polley M-Y, Liu H, Gilbert J, Cafourek V, Hillman D, Negron V, Boughey JC, Liu MC, Ingle JN, Kalari K, Couch FJ, Visscher DW, Goetz MP. Prognostic value of the neutrophil-to-lymphocyte ratio (NLR) and its relation to stromal tumor infiltrating lymphocytes (sTILs) in triple negative breast cancer (TNBC) [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P3-05-06.

Abstract Purpose: Triple negative breast cancer (TNBC) is an aggressive breast cancer subtype with substantial risks of disease recurrence. While cytotoxic chemotherapy is commonly administered and reduces recurrence, disease outcomes vary considerably and few prognostic tools are available for risk stratification for TNBC patients. We constructed and validated clinical calculators for invasive-disease free survival (IDFS) and overall survival (OS) for TNBC and compared their performance against AJCC-based models which include only tumor size and nodal status. Methods: From a surgical cohort of 9,982 patients who underwent breast cancer surgery at Mayo Clinic between January 1985 and December 2012, 605 centrally reviewed TNBC patients were identified and used to construct Cox models for IDFS and OS. Patients treated with neoadjuvant chemotherapy were excluded. Variables considered included age, menopausal status, tumor size, nodal status, Nottingham grade, type of breast surgery (mastectomy vs. lumpectomy), adjuvant radiation therapy, adjuvant chemotherapy, Ki67, stromal tumor infiltrating lymphocytes (sTILs), and neutrophil-to-lymphocyte ratio (NLR). Missing values were imputed using single imputation with all variables (including outcomes) included in the imputation model. Backward step-down procedure was used for model selections. The final models were internally validated for calibration and discrimination using bootstrapping methods and compared with AJCC-based models. Results: For both IDFS and OS, higher sTIL's, less extensive nodal involvement, use of adjuvant chemotherapy, and lower NLR were significant predictors of improved clinical outcomes. Premenopausal status and younger age were additionally predictive of improved IDFS and OS, respectively. Models for IDFS and OS have good calibration and are associated with bias-corrected C-indices of 0.68 and 0.71, respectively, as compared with C-indices of 0.59 and 0.62 for AJCC-based models. Conclusions: Our data indicate that a clinical calculator that includes sTIL's, NLR, menopausal status, age, nodal involvement as well as chemotherapy use can provide significantly greater prediction of clinical risk than tumor size and nodal status alone. These tools may be used to identify TNBC patients at elevated risk of disease relapse and to aid physician's communication with patients regarding their long-term disease outlook and planning treatment strategies. External validation is required to further evaluate broader applicability of this tool, which was developed utilizing a single-institutional experience. Citation Format: Polley M-YC, Leon-Ferre RA, Liu H, Gilbert J, Cafourek V, Hillman DW, Negron V, Boughey JC, Liu MC, Ingle JN, Kalari K, Couch F, Visscher DW, Goetz MP. Mayo clinic TNBC outcome calculator: A clinical calculator to predict disease relapse and survival in women with triple-negative breast cancer [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P1-06-07.

Abstract Purpose Dual HER2 targeting increases pathologic complete response (pCR) rate to neoadjuvant therapy and improves outcomes in both early and metastatic HER2-positive disease. CALGB 40601 is a randomized phase III trial examining the impact of dual HER2 blockade consisting of trastuzumab (H) and lapatinib (L) added to paclitaxel (T) on pCR, considering tumor and microenvironment molecular features. We previously found that pCR was numerically but not significantly increased with dual therapy, and that tumor molecular subtype and evidence of immune activation significantly and independently affected pCR (Carey et al, JCO 2016). In this secondary analysis, we sought to evaluate the effects of treatment arm and gene expression-defined subgroups on invasive disease free survival (IDFS). Patients and Methods Patients (Pts) with stage II to III HER2-positive breast cancer underwent tumor biopsy followed by random assignment with equal probabilityto paclitaxel plus trastuzumab alone (TH) or with the addition of lapatinib (THL) for 16 weeks before surgery. A paclitaxel plus lapatinib (TL) arm was closed early based on reports of futility from other trials. A secondary endpoint was IDFS, defined as the time from surgery until local or distant recurrence, new primary, or death from any cause, whichever was first. Gene expression signatures were identified by RNA sequencing. Results Between 12/2008 and 2/2012, 305 pts were enrolled. 261 pts had IDFS and gene expression information available (THL, n = 103; TH, n =101; TL, n = 57); there were no significant differences in clinical characteristics between this subset and the entire population. The median IDFS follow-up was 4.6 years with 40 IDFS events having occurred (THL, n=7; TH, n=19; TL, n=14). IDFS was significantly longer in the THL arm compared to standard TH (HR=0.34; 95% CI: 0.14-0.82; p=0.02). IDFS was also significantly longer among pCR than non-pCR pts (HR=0.40; 95% CI: 0.19-0.81; p=0.01), and did not differ by hormone receptor (HR) status, clinical stage, tumor size, race, menopausal status or age. Among gene expression signatures, only immune activation measured by an IgG signature was associated with longer IDFS (HR=0.71; 95% CI: 0.51-0.98; p=0.04); this signature was previously also associated with pCR. Multivariate analysis showed dual therapy (HR=0.35; p=0.02), pCR (HR=0.36; p=0.01), IgG (HR=0.69; p=0.05), and molecular subtype (LumA vs HER2E, HR=0.24, p=0.005) were associated with longer IDFS. A subgroup analysis by hormone receptor status revealed that among pts with HR+ disease, pts with luminal A experienced longer IDFS (HR=0.23; p=0.02) compared to those with luminal B or HER2-enriched molecular subtypes. Conclusion Dual HER2-targeting with lapatinib added to 16 weeks of TH produced significantly longer IDFS than TH alone, despite modest effects on pCR. Similar to pts with HER2-negative disease, pts with luminal A had better IDFS than those with other molecular subtypes. Immune activation as measured by RNA-based signature independently predicted both pCR and IDFS. Support: U10CA180882, U10CA180821, U24CA196171, P50-CA58823, Susan G Komen, BCRF Citation Format: Krop IE, Hillman D, Polley M-Y, Tanioka M, Parker J, Huebner L, Henry NL, Tolaney SM, Dang C, Harris L, Berry DA, Perou CM, Partridge A, Winer EP, Carey LA. Invasive disease-free survival and gene expression signatures in CALGB (Alliance) 40601, a randomized phase III neoadjuvant trial of dual HER2-targeting with lapatinib added to chemotherapy plus trastuzumab [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr GS3-02.