Rimawi, Mothaffar F.

Abstract Background: We have previously shown in animal models that combination anti-HER2 therapy leads to complete tumor regression of HER2+ breast cancer (BC). We translated these findings in a neoadjuvant trial of 12 weeks (wks) of L+T (TBCRC006) that demonstrated a meaningful pathologic complete response (pCR) and near pCR (in ER+ tumors) in patients with locally advanced HER2+ BC. In the present trial, we sought to determine whether longer treatment would lead to a higher rate of pCR without the use of chemotherapy by converting near pCR to pCR especially in the ER+ subset. Methods: TBCRC023 (NCT00999804) is a randomized phase II trial combining a Simon Phase 2 design in the experimental arm with a pick-the-winner design, not powered for direct comparison. Women with HER2+ BC measuring 2 cm or larger were eligible and were randomized in a 1:2 ratio to 12 vs. 24 wks of L+T. Letrozole (along with ovarian suppression if premenopausal) was also administered in patients whose tumors were also ER+. Serial tumor biopsies were obtained at baseline, wk 1, wk 12, and at the time of surgery. All evaluable patients were assessed for pCR, defined as no residual invasive carcinoma in the breast. Patients did not undergo surgery, withdrew consent, or received additional neoadjuvant therapy were counted as non-responders. Results: Ninety-seven patients were enrolled (33 in 12-wk arm and 64 in 24-wk arm), of whom 95 were evaluable. Seventy seven percent of patients were white and 18% were black. Twenty percent were of Hispanic ethnicity. Median age was 51 and 55% were postmenopausal. Median tumor size was 5 cm and 65% were ER+. Study treatment was well tolerated with grade 1-2 diarrhea (24% in 12-wk arm, 31% in 24-wk arm) and grade 1-2 acneform rash (12% in 12-wk arm, 19% in 24-wk arm) being the most common toxicities. Grade 3 toxicities were uncommon and were mostly in the 24-wk arm (elevated liver function test: 9%, diarrhea 2%, mucositis 2%), while the 12-wk arm had one grade 3 anemia and the study’s only serious adverse event (acute kidney injury). There were no grade 4 toxicities. The experimental arm completed stage 2 accrual and pCR rate was numerically superior to control, entirely due to better results in ER+ (Table 1), but lower than the expected pCR rate of 36% needed in the planned analysis to conclude in favor of enhanced efficacy with extended therapy. pCR rates were also lower in the control arm than in the previous study. Table 1. pCR rates 12-week arm % (n)24-week arm % (n)ER-positive8.7% (2/23)33.2% (13/39)ER-negative20% (2/10)8.7% (2/23)Overall pCR12.2% (4/33)24.2% (15/62) Conclusions: Treatment with L+T (with endocrine therapy in ER+ tumors) for 24 weeks leads to doubling of the pCR rate in women with HER2+ breast cancer without using cytotoxic chemotherapy. This approach is effective and well tolerated and warrants study as part of a de-escalation strategy that may spare some patients the cost and toxicity of chemotherapy. Tissue obtained on this trial will provide a valuable resource to validate correlative findings from our prior studies and discover new biomarkers to help guide proper patient selection for treatment. Citation Format: Mothaffar F Rimawi, Polly A Niravath, Tao Wang, Brent Rexer, Andres Forero, Antonio C Wolff, Rita Nanda, Anna M Storniolo, Ian Krop, Matthew P Goetz, Julie R Nangia, Sao Jiralerspong, Anne C Pavlick, Carolina Gutierrez, Rachel Schiff, Susan G Hilsenbeck, C Kent Osborne. TBCRC023: A randomized multicenter phase II neoadjuvant trial of lapatinib plus trastuzumab, with endcorine therapy and without chemotherapy, for 12 vs. 24 weeks in patients with HER2 overexpressing breast cancer [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr S6-02.

Abstract Background: Although the spectrum of genomic alterations in primary, treatment-naïve breast tumors has been described, the genomic landscape of HER2+ MBC remains underexplored. Furthermore, tumor genomic alterations that arise after progression on anti-HER2 therapy are largely unknown. Methods: We prospectively collected metastatic tumor biopsies from patients (pts) enrolled on TBCRC003 (NCT00470704), a phase II study evaluating the combination of lapatinib (L) and T in pts with HER2+ MBC who had varying degrees of prior T exposure. We performed WES on baseline metastatic biopsies and normal DNA from 57 pts. In 36 pts, we also performed WES on pre-treatment primary tumors. Tumors were analyzed for point mutations, insertions/deletions, and copy number alterations. Results: Total accrual was 116 pts. 87 pts were registered in one of two efficacy cohorts: Cohort 1 included pts w no prior T for MBC. Pts with prior adjuvant T were included if the interval from last T to 1st recurrence > 12 months. Cohort 2 included pts with 1-2 prior lines of T for MBC or recurrence within 12 months of adjuvant T. An additional 29 pts were enrolled in a biomarker cohort (Cohort 3). Per-protocol efficacy analyses for 85 pts deemed evaluable are shown below: Objective Response RateClinical Benefit RateMedian Time to ProgressionCohort 150% (90% CI 33.8-66.2%)57.5% (95% CI 40.9-73.0)7.4 monthsCohort 222.2% (90% CI 11.2-37.1%)42.2% (95% CI 27.7-57.8)5.3 months As we previously reported (Wagle et al, ASCO 2014), across 57 metastatic tumors, significant recurrently mutated genes were TP53 (n=30; 53%) and PIK3CA (n=19; 33%). The frequency of mutant TP53 and PIK3CA was not significantly different from 119 primary, treatment-naïve HER2+ tumors sequenced in the TCGA study (50%, p=0.8 and 27%, p=0.5, respectively). Recurrent copy number alterations were also similar to TCGA data. Comparing the 38 pts who received any prior T with the 19 pts who did not, there was no significant difference in the incidence of mutant TP53 (53% vs 53%, p=1.0) and PIK3CA (37% vs 26%, p=0.6). We identified mutations in the HER2 kinase domain in 4/38 pts who received prior T (11%), as compared to 0/19 T-naïve pts. HER2 kinase domain mutations have been identified in ∼2% of HER2-negative cancers but <1% of primary HER2+ cancers. 3 of the mutations were L755S, which has been shown to be resistant to L and sensitive to irreversible HER2 inhibitors. The 4th mutation was D742N, a novel kinase domain mutation. None of the 4 pts with HER2 kinase domain mutations had an objective response, though 1 pt had stable disease for 29 weeks. An analysis comparing paired archival primary tumors and baseline metastatic biopsies from 36 pts to identify genomic alterations acquired or enriched in the metastatic tumors will be presented. Conclusions: We present an analysis of the genomic landscape of HER2+ MBC, including comparisons between matched primary tumors and metastatic biopsies. Somatic HER2 kinase mutations in pts with HER2+ MBC treated with prior T suggests that these mutations may be involved in resistance to T, and may predict poor response to additional anti-HER2 therapy with combined L and T. Novel therapeutic approaches may be required for these pts. Citation Format: Nikhil Wagle, Nancy U Lin, Andrea L Richardson, Ignaty Leshchiner, Ingrid A Mayer, Andres Forero-Torres, Timothy J Hobday, Elizabeth C Dees, Rita Nanda, Mothaffar F Rimawi, Hao Guo, William T Barry, Ron Bose, Wei Shen, Antonio C Wolff, Stacey B Gabriel, Levi A Garraway, Eric P Winer, Ian E Krop. Whole exome sequencing (WES) of HER2+ metastatic breast cancer (MBC) from patients with or without prior trastuzumab (T): A correlative analysis of TBCRC003 [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr PD3-5.

Abstract We have recently reported that in patients with HER2-positive breast cancer, neoadjuvant targeted therapy with lapatinib and trastuzumab to more completely block the HER receptor layer, combined with endocrine therapy (in ER-positive tumors) and without chemotherapy led to a substantial 27% pathologic complete response (pCR) rate in the breast. Activation of downstream signaling pathways may lead to resistance to therapies targeting the HER pathway receptors. Aberrant activation of the PI3K pathway via decreased levels of PTEN and/or the presence of activating PIK3CA mutations has been implicated in resistance to targeted anti-HER2 therapy, but results of clinical trials are all confounded by the co-administration of chemotherapy and are inconsistent. We sought to clarify the role of these variables in predicting pCR, a surrogate for long-term outcome, in patients treated with potent targeted therapy alone in a prospective Phase II neoadjuvant trial in patients with HER2-positive breast cancer. Patients with large tumors (median 6 cm) were given 12 weeks of lapatinib plus trastuzumab followed by surgery (Rimawi et al. JCO, 2013). Serial tissue biopsies were obtained from study participants. For this study, we focused on baseline pre-treatment characteristics. PTEN protein levels were measured by IHC and scored using the H-score. PIK3CA mutations were identified on extracted DNA using multiplex PCR with targeted next generation sequencing (the Ion Torrent 50-gene cancer mutation panel). Of 64 evaluable patients, tissue was available on 59 for PTEN IHC, and sufficient DNA was available on 33 for the mutation panel. PTEN median H-score was 100 (range 0-300). PTEN status when dichotomized by the median was correlated with pCR (32% in high PTEN vs. 9% in low PTEN, p = 0.04). Activating PIK3CA mutations were identified in 12 out of 33 tumors (36%; 3 mutations in the helical and 9 in the catalytic domain) and were independent of ER status. None of the patients whose tumors harbored a PIK3CA mutation achieved pCR (p = 0.06). There was no association between PTEN status and PIK3CA mutation suggesting they are independent variables (p = 0.44). When PIK3CA mutations were considered together with PTEN status, there were 31 cases with data on both. The overall pCR rate in this cohort was 16% (lower than pCR rate observed in the overall trial). However, 0/17 cases (0%) with a mutation and/or PTEN low expression (<100 H score) had a pCR compared to 5/14 cases (36%) with PI3KCA wild type and high PTEN levels (p = 0.01). We conclude that PI3K pathway activation downstream of HER2 as a result of either low PTEN or activating PIK3CA mutation results in resistance to the combination of lapatinib and trastuzumab. This is the first report on patient tissue samples from a neoadjuvant trial using the combination of lapatinib and trastuzumab without chemotherapy. If validated in a larger cohort, our findings suggest that patients with HER2 positive tumors and who also harbor aberrant downstream PI3K pathway activation may benefit from the addition of PI3K/Akt/mTOR inhibitors to potent HER2 blockade. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr PD1-2.

Abstract Background: Imaging biomarkers of cellular division offer promise as non-invasive measures of tumor response. 3’-deoxy-3’[18F]-fluorothymidine ([18F]-FLT) positron emission tomography (PET) imaging generally correlates with pathology-based measurements of cancer proliferation, especially the Ki67 score. Though clinical studies have associated changes in [18F]-FLT uptake with therapeutic response, clinical studies validating its ability to assess cell proliferation are comparatively lacking. The goal of this study was to determine quantitative relationships between [18F]-FLT compared with molecular and cellular metrics of proliferation during treatment for locally advanced breast cancer (LABC). Methods: Baseline [18F]−FLT-PET scans were obtained prior to the initiation of chemotherapy for LABC from patients enrolled at several academic oncology study sites. MRI scans, and transmission CT scans were obtained. Core needle biopsies were obtained to determine Ki-67 indices using immuno-histochemistry and to assess an mRNA signature based measurement of proliferation. Prospectively specified quantitative relationships between PET, Ki67 immunohistochemistry and the mRNA signature were evaluated using image-matched tumor specimens. Correlations between volumetric MRI changes and pathologic responses were evaluated in a post-hoc exploratory analysis. Results: Motivated by the hypothesis that effective chemotherapies should decrease tumor cell proliferation, FLT-PET was compared with biomarkers of proliferation including Ki67 and the mRNA signature during neoadjuvant treatment for LABC. [18F]-FLT correlated both with the Ki67 labeling index (SUVmean r = 0.53) and with the proliferation signature (SUVmean r = 0.7), validating the principle of thymidine analogue imaging. However, variability in the [18F]-FLT PET and tumor cell proliferation measures likely contributed to correlations less than pre-specified target values considered appropriate for clinical use (r > 0.78). Moreover, none of the proliferation biomarkers predicted pathologic complete responses at the end of neoadjuvant therapy ∼16 weeks after the 3 week response scan. In contrast, an evaluation of change in tumor volume measured by MRI after 3 weeks of therapy confirmed its superior ability to predict pCR and tumor re-staging. Conclusion: With large numbers of cancer drugs entering therapeutic pipelines, early efficacy measures remain critical for drug development. The 3-4 month neoadjuvant treatment paradigm for LABC offers unique opportunities for drug evaluation. Functional imaging using [18F]-FLT has been advanced as an assessor of cellular proliferation, potentially offering a non-invasive approach to response evaluation. While [18F]-FLT generally correlated with proliferation, its lack of association with patient responses likely limit its clinical utility. On the other hand, the predictive value of MRI offers unique opportunities for future trial designs and confirms previous reports (1). 1. N. M. Hylton et al., Locally advanced breast cancer: MR imaging for prediction of response to neoadjuvant chemotherapy. Radiology 263, 663 (2012). Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P4-01-11.

Purpose We previously reported the eradication of human epidermal growth factor receptor 2 (HER2)– amplified human xenografts in mice by inhibition of the HER2 pathway with lapatinib and trastuzumab to block all homo- and heterodimer signaling as well as by blockade of estrogen receptor (ER) when expressed. In this clinical trial, we sought to translate these findings to patients using targeted therapy without chemotherapy. Patients and Methods Women with stages II to III HER2-positive breast cancers were eligible. They received trastuzumab once per week (4 mg/kg loading, then 2 mg/kg) and lapatinib 1000 mg once per day for 12 weeks. Women with ER-positive tumors also received letrozole (plus a luteinizing hormone–releasing hormone [LHRH] agonist if premenopausal). Pathologic response was assessed by ER status. Biopsies were obtained at baseline, weeks 2 and 8, and time of surgery. Results Sixty-six patients were enrolled, and 64 were eligible and evaluable for response. Median tumor size was 6 cm (range, 1.5 to 30 cm). Adverse events were mainly grades 1 to 2 (GI, 63%; skin, 46%). Grade 3 metabolic, GI, and liver (18%; 12 patients) and grade 4 liver toxicities (one patient) were also observed. Overall, in-breast pathologic complete response (pCR; yp T 0 - is ) was 27% (ER positive, 21%; ER negative, 36%). The rate of low-volume residual disease ( yp T 1a-b ) was 22% (ER positive, 33%; ER negative, 4%). Conclusion In patients with locally advanced HER2-positive breast cancer, our approach of targeted therapy only resulted in a high pCR rate without chemotherapy. Our data support the hypothesis that selected patients with HER2-positive tumors may not need chemotherapy, and more-complete blockade of HER receptors and ER is an effective strategy worthy of further study.

Abstract Background We evaluated the safety and efficacy of L+T in pts with 0–2 prior lines of chemotherapy (CT) for HER2+ MBC. In the context of this phase II trial, we evaluated metabolic response by FDG-PET and explored the relationship between metabolic response and clinical outcomes. Methods: Pts with measurable, HER2+ MBC were eligible. Cohort 1: No prior T, L, or CT +T for MBC, and >1 yr from adjuvant T, if received. Cohort 2: 1–2 prior lines of CT for MBC, including T, or relapse within 1 yr of adjuvant T. Pts received L 1,000 mg QD + T (2 mg/kg weekly or 6 mg/kg Q3W). Staging studies were done with CT or MRI at baseline (BL) and every 2 cycles (1 cycle=4 weeks [wks]). Objective response was assessed by local investigator according to RECIST 1.0. FDG-PET/CT was performed at BL, Wk 1, and Wk 8 per NCI guidelines. Central quality assurance, review, and analysis were performed on FDG-PET studies. Up to 5 target lesions were identified on BL FDG-PET images based on hypermetabolic uptake. Percent change in the summed maximum standardized uptake value (SUVmax) of target lesions was calculated at Wk 1 or Wk 8, compared to BL. Metabolic response was assessed according to EORTC criteria for % change in SUVmax (progressive disease [PD]: ≥25% increase; partial response [PR]: ≥25% decrease; stable disease [SD]: <25% change). Metabolic response at Wk 1 was compared to Wk 8 as well as to clinical outcome, including objective response, clinical benefit, and progression-free survival (PFS). Results: 87 pts were registered to the study. Of these, one pt did not begin protocol therapy and one pt did not have MBC on further testing, and are not included. 81/85 pts had FDG-PET data at Wk 1; 75/85 had data at Wk 8. Metabolic PR at Wk 1 was observed in 28/39 (72%) pts in Cohort 1 and 20/42 (48%) pts in Cohort 2. Metabolic PR at Wk 8 was observed in 27/34 (79%) pts in Cohort 1 and 18/41 (44%) pts in Cohort 2. Wk 1 and Wk 8 metabolic responses were similar. In cohort 1, 18/28 (64%) pts who achieved Wk 1 metabolic PR had clinical benefit by RECIST. Of pts with Wk 1 metabolic SD, 2/9 (22%) had clinical benefit. In cohort 2, 9/20 (45%) pts who achieved Wk 1 metabolic PR had clinical benefit; 5/22 (23%) who achieved Week 1 metabolic SD had clinical benefit. Exploratory analysis of progression-free survival (PFS) showed that pts in Cohort 1 who achieved Wk 1 metabolic PR experienced a median PFS of 9.3 months ([mos]; 95% CI 5.6−22.3); for pts with metabolic SD, median PFS was 1.9 mos (95% CI 0.8−5.5). For pts in Cohort 2, Wk 1 metabolic PR was associated with median PFS of 5.6 mos (95% CI 3.7−7.8), whereas for pts with metabolic SD, median PFS was 3.7 mos (95% CI 1.8−5.5). Conclusions: L+T is associated with a high rate of early and sustained metabolic response by FDG-PET. Exploratory analyses suggest that metabolic PR may be associated with clinical benefit and longer PFS. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P2-09-07.