Introduction:Neoadjuvant chemotherapy is being increasingly used in operable breast cancer. There are limited data on the safety of bevacizumab (Bev) in combination with chemotherapy in the neoadjuvant setting. We sought to explore the safety and efficacy of neoadjuvant cisplatin/Bev in a protocol for triple negative breast cancer (TNBC), with post operative dose dense AC or AC/Taxol plus Bev. We present our operative experience, including postoperative complications. We also compared this to a smaller single arm neoadjuvant cisplatin alone study.Methods:51 patients with confirmed TNBC provided informed consent and were enrolled in a single arm phase II trial of neoadjuvant cisplatin 75 mg/m2 plus Bev15 mg/kg q 3 weeks x 3 cycles with the 4th cycle cisplatin alone prior to definitive surgery. This design allowed at least 6 weeks between the last dose of Bev and surgery. Research biopsies were obtained before treatment and at surgery. Postoperatively, patients received doxorubicin and cytoxan (AC) plus bevacizumab or AC/Taxol plus Bev. Median age was 50 yrs (range 30-65 yrs); tumors were clinical stage II 48 (94%) and stage III 3 (6%).28 patients with confirmed TNBC were enrolled in our phase II trial of neoadjuvant cisplatin. Single-agent cisplatin 75mg/m2 q3weeks x 4cycles was given prior to definitive surgery. Median age was 50 yrs (range 29-69yrs); tumors were clinical stage II (75%), stage III (25%). Postoperatively, patients received AC or AC/Taxol. Two-sided Fisher exact test were used for comparing the 2 trials.Results:Fifty-one patients received neoadjuvant protocol therapy with cisplatin/Bev (all received at least one cycle of bevacizumab, 7 did not complete all cycles of preop bevacizumab) and underwent definitive local therapy to the breast. Breast conserving therapy was performed in 29 (57%) and mastectomy with or without reconstruction in 22 (43%). Postoperative complications were reported in 22 (43%) of patients: 4 (8%) required wound debridement and explantation of expanders, 8 (16%) wound breakdowns, 5 (10%) developed hematomas requiring operative intervention and 5 (10%) seromas requiring multiple aspirations.Twenty-eight patients completed neoadjuvant cisplatin therapy and underwent definitive local therapy to the breast. Breast conserving therapy was performed in 13 (46%) and mastectomy with or without reconstruction in 16 (54%). Postoperative complications were reported in 11 (39%) of patients: 5 seromas requiring drainage (18%), 2 hematomas (7%), 2 abscesses (7%), and 2 (7%) wound breakdowns. No reconstructions were lost including the 3 expanders and 2 TRAM flaps.We compared all toxicities between the two trials, (p= 0.81 NS), and wound healing related complications between the two trials (p= 0.10 NS).Conclusions:Cisplatin/Bevacizumab and cisplatin alone neoadjuvant therapy resulted in a significant number of postoperative complications. Specifically use of expanders or implants may be problematic for patients treated with Bev with a trend towards more wound related events for this group. However, this was a single arm trial, a randomized controlled studies will be necessary in determine the optimal use of bevacizumab in the timing of breast cancer surgery. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 43.

#4122 Introduction: Statins are safe, reduce cardiovascular risk, and impact pathways critical to cancer progression. We and others have shown lipophilic statins cause apoptosis and growth suppression in vitro and in vivo, and though epidemiologic data are mixed, statin effect appears most evident in estrogen receptor (ER) negative or grade 3 disease. To look for a direct biologic effect of lipophilic statins, we conducted a perioperative pilot window trial in women with breast cancer (BC).
 Methods: 40 subjects with stage 0,1 BC were randomized to high dose (80mg/day) or low dose (20mg/day) fluvastatin for 3-6 weeks prior to surgery. Paired tissue (core biopsy and surgical specimen), peripheral blood and MRI were obtained. Primary endpoint was Ki-67 (proliferation) change. Secondary endpoints included cleaved caspase-3 (CC3, apoptosis), longest diameter (LD) by MRI, and C-reactive protein (CRP) change. Subgroup analyses was planned by grade (3 vs. 1,2), statin dose; and ER status. Immunohistochemistry (IHC) on paraffin tissue used standard streptavidin biotin methods. A single breast pathologist reviewed all slides; a single radiologist read all MRIs, both blinded to timepoint.
 Results: Median serum cholesterol decreased by 16% (-23% and -12% for high and low dose, respectively p=0.012), indicating drug effect and compliance. 29 patients had sufficient tumor for paired IHC, 14 and 15 were grade 3 and 1,2, and 10 and 19 were ER - and +, respectively. In grade 3 (73% of which were ER-) vs. 1,2 tumors, there was a significant decrease in Ki-67, -7.2% (interquartile range (IQR) -13.4%, 0% ) vs. -0.3% (IQR -3%, .8%), respectively, p=0.04. CC3 (apoptosis) increased, 60% vs. 13% for grade 3 vs. 1,2 tumors, respectively, p=0.015. ER- and ER+ cases had a similar reduction in Ki67 with a median drop of 2% (IQR -13.4%, 1%) and 1.2% (IQR -6.6%,0.8%), respectively, p=0.56. While CC3 was increased in ER- vs. + (55% vs. 29%), the difference was not statistically significant. There was no dose dependent effect on Ki-67or CC3.There was no evidence of Ki67 or CC3 change when all grades were analyzed together (median drop 1.2%) and no change in CRP. Of 14 subjects with paired MRIs, 4 grade 3 cases showed a significant decrease in LD, marked ductal dilatation and increased necrosis.with statin exposure.
 Conclusions: A lipophilic statin, fluvastatin, reduced cholesterol and had measurable biologic changes (reduced proliferation, size and increased apoptosis) in stage 0,1 BC after only 3-6 weeks of exposure, specifically in the grade 3 subset. Results support the study of statins for chemoprevention for women at risk for or with stage 0 grade 3 BC, where new agents are needed. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 4122.

Background: The need for radiation therapy (RT) in conservatively managed DCIS is a source of ongoing debate. This is an updated analysis of a phase II prospective study of wide excision alone for DCIS. The study was activated in May 1995 and closed in July 2002 following accrual of 158 patients because the number of local recurrences (LR) met the predetermined stopping rules. The objective of the analysis is to update the distribution and cumulative incidence of events (LR, contralateral breast cancer [CBC], second malignancy and death from other causes). Materials and Methods: A total of 158 patients had DCIS with predominant nuclear grade 1 or 2, a mammographic extent of ≥2.5 cm, and excision with final microscopic margins of ≥1 cm or a re-excision without residual DCIS. Tamoxifen was not permitted. The results presented are from the 8-year analysis (8-year minimum potential follow-up time). Twenty-six patients without recurrence who were followed less than 8 years were excluded from the analysis as were 7 first events (4 LR) that occurred beyond 8 years of follow-up; the analysis thus includes 132 patients and 36 first events. Cumulative incidence curves were generated to assess the rates of LR or other events. Median follow up time was 10 years. Results: Overall, 36/132 patients (27%) had a first event as of April 2010. Of these 36 events, 19 were LR, 13 were CBC, 1 was a second malignancy, and 3 were deaths from other causes. Of the 19 LR, 13 (68%) were DCIS only and 6 (32%) were invasive. Fourteen occurred in the same quadrant and 5 were elsewhere in the ipsilateral breast. The 8-year estimated cumulative incidence of LR was 14.4% (95% CI: 8.4-20.4%). For all other events, the 8-year estimated cumulative incidence was 12.9% (95% CI: 3.6-13.1%). The estimated annual percentage rates of LR, CBC, and other events were 2.1%, 1.5% and 0.4%, respectively. Discussion: The results of this prospective study demonstrate a substantial and ongoing risk of LR and CBC in patients with small, nuclear grade 1 or 2 DCIS treated with wide excision with margins of ≥1cm in the absence of RT. Most LRs occurred in the same quadrant, rather than elsewhere in the breast, suggesting that excision alone is inadequate even for this highly selected population. Further study is warranted to determine if there is a subgroup of DCIS patients with nuclear grade 1 or 2 disease who are at low enough risk of LR following wide excision that RT can be omitted safely. Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P1-15-03.

#1110 Background: The majority of women who develop a BRCA1-related breast cancer develop an ER-/PR-/HER2- breast cancer. Endocrine prophylaxis does not prevent these ER- breast cancers. Therefore, there is a need to develop novel chemoprevention agents in this population. 67% of BRCA1-associated ER- breast cancers also overexpress epidermal growth factor receptor (EGFR). We therefore examined EGFR as a potential target for chemoprevention
 Methods: We isolated primary mammary epithelial cells (HMECs) from women with a germline BRCA1 mutation who underwent prophylactic mastectomies and from age-matched controls without a mutation who underwent reduction mammoplasties. We used a three-dimensional matrigel-based colony formation assay to assess clonality and proliferative capacity of these cells as well as their response to EGFR-inhibition. Flow cytometry was used to determine the number of EGF binding sites per cell. As a corresponding mouse model we used conditional MMTV-Cre BRCA1-/-p53+/- mice. Results: HMECs from BRCA1 mutation carriers and from controls express EGFR to a similar extent as HCC1937 BRCA1-associated triple-negative breast cancer cells (5x103 binding sites/cell). In ex vivo 3D-cultures we observed that HMECs derived from BRCA1 mutation carriers showed greater clonal and proliferative capacity when compared to normal controls. However while the HMECs derived from BRCA1 mutation carriers and normal controls were equally sensitive to the growth-inhibitory effect of Erlotinib at concentrations as low as 0.2 μM, the ID50 for HCC1937 breast cancer cells was > 10 μM. Similar findings were observed in murine MECs derived from normal control mice as well as BRCA1-/-p53+/- mice which develop breast cancer at the age of 7 to 8 months. Both groups of murine MECs were equally sensitive to Erlotinib growth inhibition. Conclusion: MECs derived from breast tissue of women and mice with a germline BRCA1 mutation express EGFR and are highly sensitive to growth inhibition with the EGFR inhibitor Erlotinib. In contrast, BRCA1-associated HCC 1937 breast cancer cells are more resistant to Erlotinib inhibition despite EGFR expression. We are now studying whether Erlotinib given via oral gavage daily has the potential to delay or prevent breast cancer in this mouse model of BRCA1-related breast cancer. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 1110.