Search

1090Background: CTCs play a critical role in BCa metastasis and worse prognosis even after systemic treatments such as chemotherapy and biological therapy. The mechanisms that contribute to CTCs mi...

TPS1110 Background: HER2+ breast cancers have the propensity to metastasize to the CNS. Most systemic therapy does not cross the blood-brain barrier (BBB) effectively, limiting treatment options. Palbociclib is a small molecule inhibitor of cyclin dependent kinases (CDK) 4 and 6, which has been studied in hormone positive breast cancers and found to have significant anti-tumor activity. Preclinical models demonstrate HER2+ tumors require cyclin D1 and CDK4 for progression and maintenance, and palbociclib has been found to have single agent activity in transgenic HER2+ models. Preclinical studies have also found that palbociclib crosses the BBB and exerts anti-cancer effects; indeed, clinical trials investigating its activity in primary brain tumors are ongoing. The summary of these data provide rationale for investigation of palbociclib in patients with HER2+ breast cancer and brain metastasis. Methods: A single arm study of palbociclib in patients with metastatic HER2+ breast cancer with brain metastasis was designed. The primary objective is to determine overall radiographic response rate in the CNS using modified RANO-BM criteria. Secondary objectives include PFS, OS, systemic response rates, safety, and tolerability. Exploratory objectives include assessment of mutational profiles in tissue and ctDNA, and quality of life endpoints using FACT-Cog and FACT-Br. Eligible patients must have metastatic HER2+ breast cancer with measurable brain metastasis (at least 1 lesion > = 5mm). Patients with rapidly progressive symptoms or with leptomeningeal disease are excluded. Concurrent therapy with trastuzumab is allowed. Eligible patients are treated with palbociclib 125mg PO (21 days on, 7 days off) until progressive disease or unacceptable toxicity. Available archival tissue will be obtained, ctDNA/FACT-Cog/Br will be assessed at baseline, 2, and 4 months. A total of 20 patients will be enrolled. To detect a radiographic response rates increase from 23% to 40%, a Bayesian posterior probability will be calculated after the response status in 20 patients has been assessed. The goal is that there is at least an 80% probability that the response rate exceeds 40%. Clinical trial information: NCT02774681.

3052 Background: PD-1/PD-L1 inhibitors produce modest responses in MBC; adding CTLA-4 inhibitors can augment anti-tumor activity in other cancers. Immunopharmacogenomics characterize immune-cancer cell interactions and may predict response. Methods: A single arm study was designed to determine the efficacy of durvalumab (PD-L1 inhibitor) and tremelimumab (CTLA-4 inhibitor), and immunopharmacogenomics in pts with metastatic ER+ or TNBC. The primary clinical endpoint was to assess ORR using a Simon 2-stage design (28 pts needed for type 1 error rate of 4%, 80% power). 18 pts were enrolled in the first stage; ≥4 responses were needed to proceed with the second stage. Pts were treated with durvalumab 1500mg IV and tremelimumab 75mg IV monthly for 4 doses, then durvalumab 750mg every 2 weeks to complete 1 year of therapy (option to renew for an additional year); biopsies at baseline and 2 months were collected. T-cell receptor (TCR) sequencing using mRNA isolates was conducted at baseline and 2 months, whole exome sequencing and immune-gene expression profiling were conducted at baseline. Results: From 01-09/2016, 18 evaluable pts were accrued (11 ER+; 7 TNBC). Responses are shown in the table below; the ORR did not meet criteria to proceed to the second stage. Notably, one pt with TNBC with PD had pseudoprogression, thus 5/7 (71%) pts with TNBC had clinical benefit. Median PFS was 3.8 months (95%CI 2.2-11.2) for the entire cohort (TNBC not reached), and median OS has not been reached. No grade 4 AEs were observed; grade 3 immune-related AEs included hepatitis (n = 3), nephritis (n = 1), and myocarditis (n = 1). Proportions of abundant TCR-β clonotypes (≥ 0.5% frequency) were increased in 2-month samples compared to baseline in TNBC compared to ER+, p = 0.004, and associated with responses. Remaining correlative analysis is ongoing and will be presented at ASCO. Conclusions: Response rates to PDL-1 and CTLA-4 inhibition were low in unselected MBC, however, high rates of clinical benefit were observed in TNBC. Immunopharmacogenomic may help identify phenotypes most likely to respond. Future studies in TNBC are warranted to confirm findings. Clinical trial information: NCT02536794. [Table: see text]

e12050 Background: There is emerging evidence that a 21 gene expression assay recurrence score (RS) is prognostic independent of age. In practice, pathological markers may influence decision to recommend adjuvant chemotherapy. Methods: The primary objective of this study is to investigate the relationship between the RS and pathological markers between younger (29-49) and older (50-79) women with early stage ER+ breast cancer. Pathological markers investigated included the progesterone receptor (PR), grade, Ki-67, and P53. Patients who underwent 21 gene assay testing between 2002 through 2012 were sequantially identified. Data was extracted via the institutional tumor registry or chart reviewing. For each pathological feature, mean RS was compared between younger and older patients by t-tests. Trends in chemotherapy recommendation were assessed between younger and older patients within each RS risk category (≤10, 11-25, ≥26). Results: Between 2002 and 2012, 344 eligible patients were identified. 133 were ≤49 years of age, and 211 ≥50. There was no difference in distribution of RS across age (R2=3x10-4). Between younger and older patients, there was no difference in mean RS for any pathological marker (table 1). Within each age group, mean RS was always higher in tumors that were PR negative, grade 2/3, Ki67 >10%, and P53 ≥10% (p

53 Background: The American College of Surgeons Commission on Cancer require accredited institutions to give patients a survivorship care plan (SCPs) within six months of completing curative intent therapy. However, only a minority are receiving SCP’s. Some institutions have survivorship clinics to deliver comprehensive care, including SCP’s. Insufficient referrals to such clinics are a common barrier due to survivorship not being integrated into oncology workflows. To address this, we developed and implemented a best practice advisory (BPA) alert within our EMR to identify eligible patients and facilitate referrals to the survivorship clinic. Methods: Our pilot included breast cancer patients within medical oncology. The BPA electronic alert criteria included: stage 0-III, new patient encounter within 12 months, no SCP completed or previous referral. Upon triggering, the BPA asks “Does your patient require a SCP?” followed by 3 options: 1) SCP needed- an automatic order is generated, if signed, the BPA will not fire again, if unsigned, it will fire at the next encounter (no sooner than 30 days) and will continue until an order is placed or the response “ SCP not needed” is selected. 2) SCP not needed- the BPA will never trigger again for that patient. 3) Don’t know/still on treatment-the BPA will re-fire in 30 days. Data on frequency of BPA firing and number of referrals was compared 90 days prior and post implementation. Chi-square analysis was used. Results: Between 4/1/2015-3/31/2016, 902 patients were seen with stage 0-III breast cancer at Northwestern. Ninety days prior to implementation of the BPA, 30 patients (3.3%) were referred by 8 oncology providers. In the 90 days following implementation, the BPA fired 845 times (48.5% option 1, 24.8% option 2, 26.6% option 3) and 198 patients (22%) were referred. The difference was statistically significant ( χ2 = 141.7, p < 0.0001). Conclusions: The implementation of BPAs within an EMR is an effective way to increase referrals to a survivorship clinic, thus increasing the number of patients given SCPs. Challenges identified were having enough staff availible to deal with a rapid increase in referrals, need for refinement of BPA criteria to more precisely identify eligible patients, and provider burden.

e23030Background: Liquid biopsy provides a real-time assessment of MBC. We evaluated enumeration of CTCs and analysis of ctDNA to assess prognosis and detect molecular abnormalities in MBC. Methods...

11528Background: Cell-free DNA (cfDNA) is a non-invasive method of assessing mutations in the blood. This study compared the concordance between cfDNA and tissue biopsies across DNA genomic alterations in patients with advanced breast cancer (ABC). Methods: Forty-five patients with ABC who had both NGS testing with cfDNA (Guardant) and tissue DNA (Foundation One) for clinical management were included. 45-67 genes were common to both assays depending on when the assays were performed. Concordance was defined as the presence or absence of the identical genomic sequencing alteration(s) in a single gene. Results: Median time between assays was 3.3 months [range, 0-67 months]. Including all alterations and variants of unknown significance, the average number per patient for tissue and cfDNA was 4.50 (SD 2.96) and 1.98 (SD 2.13), respectively. Concordance between the two assays was 91.2%. Among only genes with reported genomic alterations in either assay (N=232), concordance was 10.3%. 15.0% of alterations foun...

1028 Background: MBC is a challenging clinical condition treated with palliative intent due to tumor heterogeneity. We reported in 2019 ASCO that the correlation of HER2 and ESR1 mutations of ctDNA with CTCs results in worse prognosis in MBC. Here we reported that ctDNA mutations is a key point which is different between Stage III and Stage IV, and it would be helpful to evaluate the MBC metastasis and outcome. Methods: This study included 33 Stage III and 204 Stage IV MBC patients who received systemic treatments at NMH (2016-2019). Plasma ctDNA before treatment was isolated from patients and then was analyzed by Guardant 360 Health NGS-based assay for a 73 genes panel for genomic alterations including single nucleotide variants, insertions/deletions, gene fusions/rearrangements and copy number variations. Causal Inference with Ensembel Learning was used for statistical analyses. Results: Among stage III patients, 40% are luminal, 44% are HER2+ and 16% are TNBC, while in stage IV 50% patients are luminal, 20% are HER2+ and 30% are TNBC. The major differences in ctDNA between two stages lie in several genes including PIK3CA, ERBB2 and KRAS. On the top of the list is PIK3CA, which is detected in 2 out of 33 stage III patients (1 luminal, 1 HER2+) (6.06%) in baseline, each of them carries 1 mutation on PIK3CA (E542K, E545G). In 43 out of 204 stage IV patients (21.57%) who carry this gene, 15 show 1 amplification, 34 have 1 mutation (mainly H1047R, E542K, E545K and H1047L), 11 have 2 mutations (E542K/E726K, D454N/D1029N, E545K/D1017H, E545K/L287L, H1047R/E453K, H1047R/N426S and P539R/H1047R), and 1 has 3 mutations (E542Q/D454N/D1029N and E545K/E726K/R93Q). On treatment effects, PIK3CA is found to be very detrimental on prognosis and on its effects on the treatment outcome. Patients without any mutation in PIK3CA live 2.65 times longer than those with more than one mutations on PIK3CA ( p-value = 4.47e-06, CI = [1.731, 3.926]). PIK3CA, ESR1, TP53 and ARID1A are found to significantly affect liver metastasis when RET, FBXW7, ERBB2, CCND2, BRAF and MET are found to be associated with lung metastasis for both stages. EGFR, KIT and ARID1A are associated with CNS metastasis. Conclusions: We elucidated that ctDNA mutations on PIK3CA and other genes dramatically increased in Stage IV patients compared to Stage III patients which provides a new insight on the Stage III and Stage IV MBC determination. New set of genes especially PIK3CA are identified to correlate with metastasis and affect the outcome which may also be reliably used to monitor the response to therapy.