Abstract P1-14-18: Overall survival results of a multicenter randomized phase II study in locally advanced breast cancer patients treated with or without celecoxib for HER2 negative tumor (Remagus 02 trial)

Background: Cox 2 is frequently over expressed in breast cancers. Celecoxib is a COX-2 inhibitor with anti angiogenic and pro-apoptotic activities. There are few data of anti-COX2 treatment in breast cancers. and no data on the impact of neoadjuvant anti COX 2 agent on survival.

Patients and methods/: From May 2004 to October 2007, 340 stage II-III breast cancer patients were included in a phase II randomized trial and received 4 cycles (c) of epirubicin (75 mg/m2)–cyclophosphamide (750 mg/m2) q 3 w followed by 4 (c) of docetaxel (100 mg/m2) q 3 w. Pts with HER2 negative tumors (220 pts) were randomized to receive or not neoadjuvant celecoxib (200 mg bid) combined with docetaxel. All pts with hormone receptors positive tumor received hormonal treatment according to menopausal status (Pierga et al BCRT 2010). We report here overall survival (OS) and disease free survival (DFS) data and prognostic factors analyses at 5 years.

Results/: At a median follow up of 49 months, the median DFS and OS are not reached for the whole population and none of them is significantly different between pts who received celecoxib or who did not (p = respectively 0.62 and 0.36). Celecoxib had no impact either on clinical and pathological complete response rate (pCR). DFS is significantly higher in patients who achieved pCR as compared to those who did not (p = 0.017; RR = 0.21 [0.051–0.88], whereas OS is borderline significant [p = 0.07; RR = 0.19 (0.026–1.4)]. Patients with triple negative (TN) tumors (78 pts) achieved worst DFS (p = 0.02) and OS (p <0.001) than non triple negative ones despite their higher pCR rate, 29.5 % [95 % CI 19;7-40;9 %] vs 11.4 % in all the other subgroups. At multivariate analysis including 7 factors [tumor size, clinical lymph node, grade (1 vs. 2and 3), pCR, ER, PR, TN], factors which influenced OS and DFS were pCR (p = 0.034 & 0.013 and PgR expression (p = 0.046 & 0.01).

Conclusion/: Celecoxib had no influence on pCR, DFS or OS. Despite higher pCR rate triple negative breast cancer patients' subgroup remains with the poorest outcome.

Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P1-14-18.

Sex moderates circadian chemotherapy effects on survival of patients with metastatic colorectal cancer: a meta-analysis

BACKGROUND

Molecular circadian clocks can modify cancer chemotherapy effects, with a possible moderation according to sex differences. We investigated whether sex determine the optimal delivery schedule of chemotherapy for metastatic colorectal cancer.

PATIENTS AND METHODS

A meta-analysis was performed using individual data from three international Phase III trials comparing 5-fluorouracil, leucovorin and oxaliplatin administered in chronomodulated (chronoFLO) or conventional (CONV) infusions. The data from 345 females and 497 males were updated at 9 years. The main end point was survival.

RESULTS

Overall survival was improved in males on chronoFLO when compared with CONV (P = 0.009), with respective median values of 20.8 (95% CL, 18.7 to 22.9) and 17.5 months (16.1 to 18.8). Conversely, median survival was 16.6 months (13.9 to 19.3) on chronoFLO and 18.4 months (16.6 to 20.2) on CONV in females (P = 0.012). The sex versus schedule interaction was a strong predictive factor of optimal treatment schedule, with a hazard ratio of 1.59 (1.30 to 1.75) for overall survival (P = 0.002) in multivariate analysis.

CONCLUSIONS

Males lived significantly longer on chronomodulated chemotherapy rather than on conventional chemotherapy. The current chronoFLO schedule deserves prospective assessment as a safe and more effective first-line treatment option than conventional delivery for male patients.

P2-09-11: Role of 18FDG-PET/CT in the Staging of Large Primary Operable Breast Cancer

Background: Prospective evaluation of the role of 18FDG-PET/CT in patients with large primary operable breast cancer.

Material and Methods: During 56 months, consecutive patients with large (>2cm) breast cancer and clinical stage IIA/IIB/IIIA (based on clinical examination, mammography, breast MRI and ultrasonography) underwent 18FDG-PET/CT. The nuclear physician was blind to the results of any other procedure (bone scan, chest X-ray, liver ultrasound, or thoraco-abdominal CT scan). Results: Out of the 131 examined patients, 36 had clinical stage IIA (34 T2 N0, 2 T1 N1), 48 stage IIB (20 T3 N0, 28 T2 N1), and 47 stage IIIA (29 T3 N1, 9 T2 N2, 9 T3 N2). 18FDG-PET/CT modified staging for 5.6% of stage IIA patients, for 14.6% of stage IIB patients, and for 27.6% of stage IIIA patients. However, within stage IIIA, the yield was specifically high among the 18 patients with N2 disease (56% stage modification). When considering stage IIB and primary operable IIIA (T3 N1) together, the yield of 18FDG-PET/CT was 13% (10/77); extra-axillary regional lymph nodes were detected in 5 and distant metastases in 7 patients. In this series, 18FDG-PET/CT outperformed bone scan with only 1 misclassification versus 8 for bone scan (p=0.036).

Discussion: 18FDG-PET/CT provided useful information in 13% of patients with T3 N0 / T2 N1 / T3 N1 disease. The yield was more modest in patients with T2 N0 disease. The very high yield in the case of lymph nodes classified N2 demonstrates that stage IIIA comprises two quite distinct groups of patients.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P2-09-11.

Chemotherapy-induced neutropenia association with survival in metastatic colorectal cancer (MCC): Schedule dependency

454

Background: Circadian clocks control cellular proliferation and drug metabolism over 24 h. However, circadian chronomodulated chemotherapy with 5-fluorouracil, leucovorin and oxaliplatin (chronoFLO4) offered no survival benefit as compared to the non-time stipulated FOLFOX2 in an international randomized trial involving patients (pts) with previously untreated MCC (EORTC05963). We hypothesized that treatment near maximum tolerated dose could disrupt circadian clocks thus impairing the efficacy of chronoFLO4 but not of FOLFOX2.

Methods: Pts with available data (N=556) were categorized into three subgroups according to the worst grade of neutropenia experienced during treatment. Distinct multivariate models were constructed for each treatment schedule.

Results: Neutropenia (all grades) occurred in 39% of the pts on chronoFLO4 as compared to 67% of those on FOLFOX2 (p< 0.0001), with G3-4 being encountered in 7% and 25%, respectively (p< 0.0001). In both schedules, neutropenia was more frequent and more severe in women than in men (p<0.04).The occurrence (but not the severity) of neutropenia was significantly associated with improved objective response rate, progression-free and overall survival in pts on FOLFOX2 (p< 0.0001), confirming previous results by others. In pts on chronoFLO4, the occurrence of neutropenia was not associated with any efficacy advantage (p=0.36), and worst survival was observed in pts developing severe neutropenia.

Conclusions: Neutropenia was more frequent and severe in women than in men, and on FOLFOX2 than on chronoFLO4. Neutropenia was positively correlated with survival in pts on FOLFOX2 supporting intra-patient dose escalation to achieve toxicity for conventional chemotherapy. No survival prolongation was found in pts with neutropenia on chronoFLO4. Thus, reaching maximum tolerated dose was unnecessary to achieve optimal effectiveness of circadian-timed therapy, and should even be avoided.

[Table: see text]

No significant financial relationships to disclose.

Abstract P1-17-03: Comparison of Long Term Outcome of Locally Advanced and Inflammatory Breast Cancers Treated with Dose Dense Neoadjuvant Chemotherapy

Background: Inflammatory breast cancers (IBC) have a worse prognosis than locally advanced breast cancer (LABC) (KW Hance, JNCI 2005; 97:966-75). We report a series of LABC and IBC treated with dose dense anthracycline and cyclophosphamide and report the correlation between phenotypic features, pathological complete response (pCR) and disease free survival (DFS).

Materials and methods: Between 1990 and 2003, 196 patients (pts) with LABC (105 pts) and IBC (91 pts) treated at St Louis hospital received

6 cycles (c) of dose dense Cyclophophamid (1.2g/m2 d1)-Epirubicin (E) (75mg/m2d1) q2w (SIM regimen). 133 patients had frozen biopsy allowing p53 determination. Surgery was done after chemotherapy. pCR was defined as no residual invasive tumor in breast and lymph nodes.

Results: Median follow-up was 110 months. IBC were more often ER negative than LABC(52 % vs 34%), more often triple negative (33 % vs 22 %) and more often TP53 mutations .(60% vs 39 %).

pCR in breast and lymph nodes was identical in LABC (17 pts,16%) and in IBC (14 pts, 16 %).

Table I Pathological response and factors associated with pCR

Negative ER and presence of TP53 mutations were strong predictive factors of chemotherapy efficacy in LABC and less in IBCS. histologically grade 3 was not predictive in IBC.

Median disease free survival (DFS) was higher in LABC [103 months (m)] than in IBC (54 m), p=.031. Median Overall survival (OS) was not reached in LABC and was of 134 m in IBC (p = .015). A significant effect of ER status on DFS was found in LABC and IBC. When adjusting for ER, IBC still had a significantly shorter DFS (HR 1.64, 95%CI 1.12 to 2.39, P=.011). Similar results were found for OS. For TP 53, a significant interaction was found with LABC/IBC (p=0.017) with a risk for mutated patients in LABC (p=0.004) and a slight and nonsignificant increased risk in IBC patients (p=0.62).

Estimated disease free survival at 15 years

Conclusion: High risk patients treated with dose dense cyclophosphamide and anthracyclin have very long survival rate. Negativity of ER and p53 mutations are strong predictive factors of good outcome in LABC but less in IBC. This approach in ER-LABC should be prospectively studied..

Abstract P2-09-26: Circadian Clock Genes in Primary Breast Cancer: Strong Predictors of Pathologic Response on Neoadjuvant Chemotherapy

Background: Circadian clocks involve 15 specific genes, which control cell cycle, apoptosis, DNA repair and metabolism. Circadian disruption is associated with both an increased risk of breast cancer, and poor survival in patients (pts) with metastatic cancer (IARC#98; Innominato et al. Cancer Res 2009).

Objective: to provide first evidence of clinical relevance of clock genes expression for the therapeutic sensitivity of primary breast cancer. Method: Affymetrix Hgu133plus2 microarrays data were derived from 189 primary breast cancers. All the pts had non metastatic disease and were registered in multicenter Phase II trial of neoadjuvant chemotherapy with epirubicin-cyclophosphamide (4 courses) then docetaxel ± herceptin (4 courses). In this study, negative hormonal receptor status and limited tumor size predicted for pathological complete response rate (pCR) (Pierga JY, et al Breast Cancer Res Treat. 2010).The expression data of 15 core clock genes, within the whole tumor transcriptome, were related both to tumor mitotic index (MI) and estrogen receptors (ESR1), and to chemotherapy-induced pCR. Statistical validation involved T-test statistics, with corrections for multiple testing.

Results:

In the tumor biopsy obtained before treatment onset, both Per2 and Cry2 expressions correlated positively with ESR1 (r≥0.43; P&lt;0.001) and inversely with (MI (r≤−0.28; P&lt;0.007). In addition, a positive correlation was found for Dec 1 with ESR1 (r=0.65, P&lt;0.001). The mean expressions of Per2, Cry2 and Dec 1 before chemotherapy were lower in the 32 pts who subsequently achieved pathologic complete response (pCR) as compared to the 157 who did not (P&lt;0.001). No other clock gene expression was significantly related to pCR. The odds ratio confirmed the decreased chance of achieving pCR with increased expression of clock genes Per2, Cry2 or Dec1.

Clock gene expression and pCR

Conclusion: Low expressions of three core genes in the negative loop of the molecular circadian clock strongly predicted for the induction of a pathologic complete response of primary breast cancer with neoadjuvant chemotherapy. While circadian clock disruption seems to constitute a poor prognostic factor in cancer patients, our data suggest that the downregulation of core circadian genes in tumors significantly enhances susceptibility to chemotherapy.

Abstract P6-09-06: Oral Contraceptive Use and Overall Breast Disease Risk: Cross Sectional Retrospective Study

Background: Oral contraception (OC) is one of the most widely used means for birth control in the world. In several studies, OC has been associated with a slightly increased risk of breast cancer in current user, and with a decreased risk of benign disease. Little is known about atypical high risk lesions. Our goal was to investigate whether OC use was a risk factor or a protective factor for benign, high risk or malignant breast lesions. Material and methods : From 2001 to 2007, all non-palpable breast lesions referred to biopsy or cytology in Saint Louis hospital were prospectively registered. Demographic and clinical data including oral contraceptive pill use extent were reported. We defined benign lesions (fibroadenoma, blunt duct adenosis, fibrocystic changes, epithelial hyperplasia, others), high risk lesions (atypical ductal hyperplasia, atypical lobular hyperplasia, lobular carcinoma in situ and malignant lesions (ductal carcinoma in situ, invasive ductal or lobular carcinoma). The aim was to analyse the correlation between the duration of OC use and the occurrence of the breast disease classified by histology groups. Patients with previous history of benign breast disease or malignancy were excluded.

Results: The analysis was performed on 1329 breast lesions. The breakdown of the lesions were as follows: 819 benign lesions (fibroadenoma n=155, blunt duct adenosis n=169, fibrocystic disease n=194, epithelial hyperplasia n=132, others n=170), 104 high risk lesions (atypical ductal hyperplasia n=54, atypical lobular hyperplasia n=29, lobular carcinoma in situ n=21), and 406 malignant lesions (ductal carcinoma in situ n=158, invasive ductal or lobular carcinoma n=248). The duration of oral contraception use was not significantly associated with the occurrence of benign, high risk or malignant breast disease. When focusing on benign lesion subtypes, no association was observed either. Older age was significantly correlated to the occurrence of atypia or carcinoma. Significant difference existed in the median age of apparition 55 y.o and 57 y.o respectively, versus 53 y.o for benign lesion (p&lt; 0,0001), as well as menopausal status at diagnosis (57.7% and 68.3%, respectively, versus 53% for benign lesion) (p&lt; 0,0001).

Conclusion: In this cross sectional retrospective study, the duration of OC use was not associated with differential occurrence of benign, high risk and malignant breast lesion. Although a lack of statistical power may be invoked to explain the results, we are called to believe that the magnitude of an effect of OC use is small, if ever it exists.

Results [Table 1]

Abstract P6-05-06: Clinical Presentation of Molecular Apocrine Subgroup of Breast Cancer: A Rather Aggressive Group of Tumor

Background: Estrogen receptor negative (ER-) breast cancer represents 30% of breast cancers. This heterogeneous group comprises at least the basal and HER2+ subgroups. Recent data, as well as our own, has observed that the HER2+ subtype is highly heterogeneous. Several teams have identified a new “apocrine” molecular subgroup of cancer, characterized by androgen receptor (AR) expresion in an ER-context. Here, we have retrospectively identified, based on a transcriptionnal signature, these apocrine molecular tumours and described their clinical presentation and evolution.

Material and Methods:

We retrospectively identified 60 patients treated in St Louis Hospital (Paris) from 1995 to 2008 and presenting the signature of the molecular apocrine subgroup (ERA-, AR+, FOXA1+) by Q-RT-PCR.

Results:

Mean age at diagnosis was 53,5 y.o. Tumours size were T2 or more in 78% cases. Histological types were ductal invasive with intraductal component (n=22), histological apocrine (n=3), and paget disease (n=4). Tumor grade was 3 in 68%, and 2 in 21%, with lymphovascular invasion in 37%. Excluding patients receiving neo-adjuvant chemotherapy, lymph node status was negative in 41%, and positive in 52% (1 to 3, 32%, more than 4N+, 20%). By immunohistochemistry 97.4% were PR-and 58.72% overexpressed HER2.

Surgery was conservative in 46%, and 48% patients underwent mastectomy. Sixteen patients received neoadjuvant chemotherapy (27%), 41 received adjuvant chemotherapy (68%), 16 received hormonal therapy (27%), and

16 received trastuzumab (27%). With a median follow up of 60 months, 34 events (local recurrence n=9, contralateral n=3, distant metastasis n=22), and 13 deaths occurred. Median disease free survival was 48 months.

Discussion:

In this cohort of apocrine molecular carcinomas, tumor phenotypes appears to be rather aggressive, with a high proportion of poor prognosis factors (grade SBR3, lymphovascular invasion, node involvement), and are generally well-correlated to a poor clinical outcome in this population that received heterogeneous treatments. Further data are needed to precisely characterise this particular breast cancer subtype, notably patients who are not eligible to Herceptin-based regimen.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P6-05-06.

Abstract P1-11-14: Neoadjuvant Chemotherapy in Lobular and Ductal Carcinoma: Comparison of Clinical, Pathological Response Rates and Survival

Introduction: Neoadjuvant chemotherapy is a standard of care in locally advanced breast carcinomas. The major purpose of neoadjuvant chemotherapy is breast conserving surgery. Advantages in survival have only been shown in patients undergoing achieving pathological complete response (pCR). Lobular carcinomas seem less chemosensitive than ductal carcinomas to neoadjuvant chemotherapy (NCT).

Purpose: To compare the clinical and pathological response rate and the outcome of lobular (ILC) versus ductal (IDC) invasive breast carcinomas after NCT.

Patients and methods: Between 1985 and 2010, 385 patients with locally advanced and/or inflammatory breast carcinomas from Saint Louis hospital received neoadjuvant chemotherapy. Forty-four (11, 4%) ILC and 341 (88, 6%) IDC were diagnosed by surgical or core needle biopsy before CT. All patients received anthracyclines based CT,181 (47%) additionally received a taxane (four cycles of epirubicin 75 mg/m2 and cyclophosphamide 750 mg/m2 then four cycles of taxotere 100mg/m2 or six cycles of a dose dense regimen of 75 mg/m2 epirubicin and 1200mg/m2 cyclophosphamide, every 14 days),and underwent breast surgical excision (lumpectomy or mastectomy) and axillary node dissection.Clinical response was defined by the lack of palpable tumor in the breast before surgery.Pathological complete response was defined by no residual invasive tumor in breast. Radiotherapy and hormonotherapy were delivered to patients, when appliable.

Results: Clinical response to NCT was higher for ILC (27, 3%) than IDC (13,4%). ILC with clinical response tended to have higher histological grade 2 (91,6 % versus 30,4% for IDC),more estrogens receptors (RE) positivity (91,6% vs 45,6%), p53 wild type (66,6% vs 30,4%) and HER2- negative tumors (91,6% vs 76%) Pathologic response (pCR) rate was lower for ILC than for IDC (2, 7% vs 9, 3%). Only one ILC (grade 2, RE+, HER2 negative, p53 wild type) underwent pCR. Thirty two (9, 3%) IDC haved pCR: 15,6% haved histological grade 2, 15,6% RE positivity,9,3% p53 wild type, 62,5% HER2 negative. Thirteen percent of ILD and 9 % of ILC haved breast conservating surgery .At a median follow up of 60 months, ILC patients tended to have longer overall survival (55% vs 48%) and recurrence free survival (44% vs 36,8%) than IDC.

Conclusions: ILC was characterized by better clinical response rates but lower pathologic response and breast conservating surgery rates .Despite the low pCR rate, patients with ILC tended to have better outcomes than did patients with IDC. Pathological complete response to NCT in ILC did not seem to have prognostic significance. Further data are warranted to help clarify the characteristics genomics and proteomics of ILC which explains this better outcome.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P1-11-14.

Single Circulating Tumor Cell Detection and Overall Survival in Non Metastatic Breast Cancer

Background: Circulation of cancer cells in the blood is a necessary step of hematogeneous metastasis while circulating tumor cells (CTC) have been reported to have a low metastatic efficiency in preclinical animal models. After a median follow-up of 18 months, we previously reported that CTC detection influences the distant metastasis-free survival (DMFS) in non-metastatic breast cancer (BC) patients (pts) treated by neoadjuvant chemotherapy (NACT) in a multicenter prospective trial. Updated results are presented here, focusing on overall survival (OS) and predictors of metastatic relapse.Methods: In 115 localized BC pts, CTC were prospectively screened (CellSearch) before and after NACT (REMAGUS02). We analyzed their outcome after a median follow-up of 36 months.Results: At baseline, 23% of pts were CTC-positive, but only 10% had more than 1 CTC per 7.5ml of blood. At an individual level, CTC detection before chemotherapy, used as a test to predict metastatic relapse, exhibited a global accuracy of 77%, higher than that of tumor grade (54%), tumor size (57%), lymph node invasion (40%), triple negative phenotype (76%) and pathological complete response (27%). Multivariate analyses for OS and DMFS showed that CTC detection before chemotherapy was a strong independent prognostic factor for both DMFS (p=0.01, RR=5.0, 95%CI[1.4-17]) and OS (p=0.007, RR=9, 95%CI[1.8-45]), along with tumor size and triple-negative phenotype, while post-chemotherapy CTC detection had a lower significance for both endpoints (p=0.07 and p=0.09 respectively).Conclusion: Biologically, the metastatic efficiency of CTC could be higher than previously thought. Clinically, besides confirming our previously reported results, this study shows that CTC detection may become the main prognostic factor in BC pts treated with NACT. Implementing this technique in everyday management might help to identify high-risk pts in whom innovative strategies should be investigated.Supported by PHRC AOM/2OO2/02117, Pfizer inc., Roche, sanofi-aventis.ISRCTN10059974

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 3017.

Correlation between 18F Fluorodeoxyglucose (FDG) Uptake and Tumor Characteristics in Locally Advanced Breast Cancers

Background: This study analysed the correlation between [18F]fluorodeoxyglucose (FDG) uptake, assessed by positron emission tomography (PET), and prognostic factors in locally advanced breast cancersMaterial and Methods: All locally advanced breast cancers seen at Saint Louis hospital and treated with neo-adjuvant chemotherapy (NAC) have a FDG PET after core needle biopsy and before chemotherapy. We correlated the tumor characteristics: T-stage, histological grade, estrogen and progesterone receptors, c-erbB2 over-expression (immuno-histochemistry determination) and P53 (determinate on frozen biopsies by the FASAYmethod) to FDG standardized uptake value (SUV max). The statistics tests used are student test (comparison of 2 means) and kendall correlation.Results: From June 2006 to April 2009, 91 patients with locally advanced breast tumors have both PET scan and frozen tissue before NAC. Median age at diagnosis is 48 (26-81) and 45 % are post menopausal.Conclusion: This study indicates that FDG-PET uptake is correlated with the phenotype of breast tumours. Over-expression of c-erbB2 does not influence FDG uptake. Triple negative tumours and p53 mutated tumors have a high initial SUV which can reflect their aggressiveness and their chemotherapy sensitivity. The knowledge of SUV uptake according to tumor characteristics allows a better understanding of the role of FDG-PET in the prediction of neoadjuvant chemotherapy response.Patients characteristicsPtes characteristicsNumber of patients (%)Tumor size: T2/T3/T438 (41)/32 (35)/21(23)Lymph nodes N0/N1/N232 (35)/46 (50.5)/13 (14 )Invasive ductal carcinoma/invasive lobular carcinoma/miscellaneous82 (90 )/5 (5 )/3 (3)Histological grade/ 1/2/37 (7 )/49 (54)/35 (38)ER + / ER-58 (64 )/33 (36)PR+ /PR-35 (38) /56 (61)C-erbB2 +++15 (16 )Triple negative22 (24 )P53 ( 64 ptes) mutated/wild type32 (50)/32 (50) 

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 5010.

The Role of FDG PET for Early Prediction of Response after Two Cycles of Epirubicin + Cyclophosphamide Neoadjuvant Chemotherapy in Breast Cancer

Background: Previous studies showed a possible role of Fluorodeoxyglucose Positron Emission Tomography (FDG PET) in monitoring response to neoadjuvant chemotherapy (NACT) in breast cancer patients. Most studies, however, mixed various chemotherapy protocols. We assessed the ability of FDG PET to predict response after two cycles of epirubicin + cyclophosphamide (EC) and compared it with histopathological response as determined by the Sataloff scale after completion of chemotherapy.Material and Methods: From 07/2007 to 05/2009, 54 patients seen at Saint Louis hospital underwent FDG PET at baseline and after the second cycle of NACT. We present data for the first 22 consecutive patients for whom pathology data are available. Standard chemotherapy treatment was four cycles of epirubicin + cyclophosphamide followed by four cycles of docetaxel. The study was performed according to the guidelines of the institutional ethical committee. The standardized uptake value (SUVmax) of FDG was measured with a PET-CT instrument at baseline and after the second cycle of chemotherapy. The change in SUV was expressed as Δ SUVmax (%) = 100 X (2nd cycle SUVmax - baseline SUVmax)/baseline SUVmax. A Δ SUVmax cutoff value of -45% was used to differentiate metabolic responders and non-responders. Histopathological response was assessed on fresh surgical specimens (mastectomy or lumpectomy) by an experienced pathologist and graded according to the scale established by Sataloff: total or near-total therapeutic effect (grade A), more than 50% therapeutic effect but less than total or near-total effect (grade B), less than 50% therapeutic effect but visible effect (grade C), or no therapeutic effect (grade D). For the analysis, grades A and B were considered as histopathological responders and grades C and D were as non-responders.Results: Initial T-stage was T2 in 12 cases, T3 in 6 cases and T4 in 4 cases. There were 21 cases of invasive ductal carcinoma and 1 case of invasive lobular carcinoma. Mean SUVmax in initial PET was 7.22 (ranges from 2.7 to 18.5) and mean SUVmax after 2 cycles of chemotherapy was 4.77 (1.4-15.3). Breast-conserving surgery was performed in 12 patients and mastectomy in 10 patients. Nine (41%) of 22 patients were classified as histopathologic responders and 13 (59%) as non-responders. PET after two cycles of NACT revealed 9 patients (41%) as responders and 13 (59%) as non-responders. Among 9 metabolic responders, 7 were true positive, and 2 were false positive. Among 13 metabolic non-responders, 11 were true negative, and 2 were false negative. The sensitivity, specificity, positive predictive value, negative predictive value and accuracy of FDG PET after two cycles of NACT were 78%, 85%, 78%, 85% and 82% respectively.Discussion: NACT has proved useful in stage II and III breast cancer, to reduce tumour volume, increasing the chance of breast-conserving surgery. In order to minimize adverse effects of NACT, non-responders must be identified as early as possible. Our preliminary results on a small series of patients show that FDG PET can differentiate responders from non-responders with good accuracy after two cycles of neoadjuvant chemotherapy with EC.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 5009.

A Multicenter Randomized Phase II Study of Sequential Epirubicin/Cyclophosphamide Followed by Docetaxel with or without Celecoxib or Trastuzumab According to HER2 Status, as Primary Chemotherapy for Localized Invasive Breast Cancer Patient

Purpose: To assess anti tumor activity of sequential epirubicin/cyclophosphamide followed by docetaxel with the randomized addition of celecoxib In Her2-ve patients or trastuzumab in Her2+ve patients versus no additional treatment, in terms of pathological complete response (pCR), defined as absence of residual invasive breast carcinoma and of nodal involvement.Patients and methods: 340 patients with stage II and III breast adenocarcinoma, with tumors ineligible for breast conservative surgery received 8 sequential 3 weekly cycles with epirubicin (75mg/m2)-cyclophosphamide (750mg/m2) for 4 cycles followed by docetaxel (100 mg/m2) for 4 cycles. According to HER2 status (IHC + FISH) they were randomized to receive together with docetaxel, in HER2 negative group (220 patients) celecoxib 800 mg/day during cycles 5-8 or no additional treatment and in HER2 positive group (120 patients) trastuzumab (8mg/kg then 6mg/kg) infused together with docetaxel or no additional treatment. All patients except eight with HER2+ tumor received adjuvant trastuzumab for a cumulative duration of 12 months.Results: In the HER2 negative group, pCR (grade 1 and 2 of Chevallier's classification) was observed in 11.5% and 13% of patients treated without and with neoadjuvant Celecoxib respectively. In the HER2+ group, pCR rate reached 26% in those having received neoadjuvant trastuzumab versus 19% in the others. There was no cardiac toxicity and no toxic death. Triple negative breast cancers experience the highest pCR rate of 30%.Conclusion: It is the first report on the effect of the addition of celecoxib to chemotherapy in neoadjuvant setting in breast cancer, showing that celecoxib does not improve the pCR rate. Addition of trastuzumab does, yet not to the extent reported with a protracted co-administration before surgery. Expression of hormonal receptors appears to be the major prognosticator for pCR. Molecular studies of gene expression profiling should allow improving such prediction.Supported by PHRC AOM/2OO2/02117, Pfizer inc., Roche, sanofi-aventis.ISRCTN10059974

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 5054.

Nonpalpable Breast Lesions in a Breast Care Unit: Prospective Analysis on 2708 Consecutive Cases

Background: Breast cancer screening increases the detection of nonpalpable breast lesions, These lesions raise specific concerns, involving radiological imaging, biopsy techniques, and pathological analysis. The objective of the study is to evaluate the management of nonpalpable breast lesions in a breast disease unit.Material and Methods: From 2001 to 2007, 2708 nonpalpable breast lesions were prospectively evaluated by a multidisciplinary team. Radiologic lesions were detected by mammography alone (71,5%), ultrasonography (27,90%), MRI (0,20%). All lesions were classified according to the BI RADS classification. Three hundred and nine (309) core needle biopsies were performed, 807 vacuum assisted biopsies, and 521 open breast biopsies. The pathologic results were correlated with surgery, rebiopsy or long-term imaging follow up.Results: The pathologic results showed 33 % of malignant lesions (DCIS, invasive carcinoma), 9 % of high risk lesions (atypical ductal or lobular hyperplasia, lobular carcinoma in situ) and 58 % of benign lesions. The prevalence of cancer as a function of BI-RADS was: BI-RADS 0 : 2,6% (4/152), BI-RADS 2: 0% (0/55), BI-RADS 3: 2,3% (17/740), BI-RADS 4: 23,4% (352/1502) et BI-RADS 5 : 78,7% (185/235). Twelve of 152 (7,9 %) high risk lesions were upgraded to malignant lesions and 34/211 DCIS (16,1%) were upgraded to invasive carcinoma after surgery. Diagnostic performance rates exhibited the following results: agreement rate=96,6%, sensibility=96,2%, overall underestimation rate=12,6%, and false-negative rate=1,6%.Table 1: Diagnostic performance of core needle biopsy, vacuum assisted biopsy, open breast biopsy core needle biopsy%vacuum assisted biopsy%open breast biopsy%Totaln=309 807 521 1637Pathologic results       benign21970,9%48960,6%23444,9%942 (57,5%)High risk82,6%728,9%6813,1%148 (9,0%)malignant8226,5%24630,5%21942,0%547 (33,4%)Diagnostic performance       sensibility80/8297,6%228/24593,1%217/21999,1%525/546 (96,2%)agreement304/30998,4%762/80794,4%515/52198,8%1581/1637 (96,6%)high risk underestimate1/616,7%10/7713,0%1/691,4%12/152 (7,9%)DCIS underestimate3/650,0%27/12421,8%4/814,9%34/211 (16,1%)overall underestimate rate4/1330,8%37/20118,4%5/1503,3%46/364 (12,6%)false negative rate1/821,2%7/2462,8%1/2190,5%9/547 (1,6%) After vacuum assisted biopsy, one-step surgery was performed in 82,9% and after core needle biopsy in 68,4%.Conclusion: This kind of quality evaluation in community practice should be encouraged. Management of those lesions continuously evolves with the widespread of RMI and new biopsy techniques. Efforts should be made in exploring imaging-pathologic discrepancies, and in identifying predictive factors of invasion on biopsies. We currently perform a focused analysis on lesions that required two surgical steps despite a prior biopsy, in order to point out new ways to improve our practices.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 6020.

Meta-analysis of gender effect for first-line chronomodulated 5-fluorouracil-leucovorin-oxaliplatin (ChronoFLO) compared with FOLFOX or constant infusion (conventional delivery, CONV) against metastatic colorectal cancer (MCC) in three international controlled phase III randomized trials (RT)

4112

Background: Gender predicted for the most effective schedule in a RT of ChronoFLO vs CONV against MCC: overall survival (OS) was significantly increased in men on chronoFLO vs FOLFOX, whereas the reverse was found in women (Giacchetti, JCO 2006). Methods: To assess the relevance of gender for patient (pt) outcome, meta-analysis was performed on individual pt data (IPD) from 3 RT in 845 MCC pts treated with chronoFLO vs CONV (346 F, 499 M at 36 centers in 1990–2002)(Lévi, JNCI 1994; Lancet 1997). Data bases were merged and updated at 9 y after inclusion of the 1st pt. Main prognostic factors were comparable in each RT according to gender and treatment arm (median age: 61y; PS=0, 46% pts; liver M, 85% pts; liver involvement >25%, 41% pts; lung M, 37% pts; CEA>10, 56% pts). Results: No significant difference was found according to delivery schedule or gender in the whole population for Response Rate (RR), Progression-Free Survival (PFS) and OS. However, men on chronoFLO had highest RR, longest PFS and OS. PFS and OS were highest in women on CONV ( Table ). The rate of complete macroscopic resections of liver metastases (R0+R1) was 12.5% in men on chronoFLO vs 7.8–8.5% in men on CONV or in women on either schedule. A complete histologic response of liver metastases was documented in 2.1% of the men on chronoFLO vs 0–1.1% in the other groups. The relative risk of an earlier death in men vs women was 0.76 [95% CL, 0.91 to 0.94] on chronoFLO and 1.24 [0.99 to 1.56] on CONV. Conclusions: This IPD meta-analysis of 3 RT in MCC with a minimum follow up of 5 years confirms that men benefit from chronoFLO as compared to CONV delivery, with regard to long term outcome and medico-surgical strategy. ChronoFLO should be preferred to conventional oxaliplatin-5-FU-LV schedules in men with MCC. Support: ARTBC Internationale, P. Brousse Hospital, Villejuif, France.

[Table: see text]

No significant financial relationships to disclose.

p53 mutations to predict efficacy of alkylating-containing regimen: a metaanalysis of four different clinical trials

Abstract #6064

Background. The predictive value of p53 mutations for efficacy of anthracycline-based chemotherapy is matter of controversy. Inconsistencies among studies could be related to the heterogeneous use of alkylating agents in combination with anthracyclines in different studies. We examined the predictive value of p53 mutations in four different series of breast cancer patients treated with preoperative anthracycline-based chemotherapy including different doses of cyclophosphamide (C).&#x2028; Patients and Methods. All patients had stage II-III breast cancer and received anthracycline-containing chemotherapies. A total of 352 patients were included in four different clinical studies : 65 patients with estrogen receptor (ER)-negative cancers treated with single agent epirubicine (E) (100 mg/m2/3w x 4) in the TOP trial (R1), 52 patients treated with FAC (500 mg/m2 C, 5-FU and 50 mg/m2 doxorubicin/3w x 6) (R2), 96 patients treated with EC-T (75mg/m2 E and 750 mg/m2 C/3w x 4 followed by 100mg/m2 docetaxel/3w x 4) (R3) and 139 patients treated with dose-dense EC (1200mg/m2 C and 75 mg/m2 E/2w x 6) (R4). Before therapy, p53 status was determined in all tumors by yeast functional complementation (FASAY) assay. After chemotherapy, all patients underwent surgery. Pathologic complete response (pCR) was defined as no residual invasive tumour cells in breast and lymph nodes.&#x2028; Results. P53 mutations were more frequently observed in the 128 ER-neg compared to ER-pos cancers (78% vs 29.4%). In p53 mutated tumours, the pCR rate increased with the dose-intensity of C. Conversely, in p53 wild type tumours, pCR rates decreased.&#x2028; &#x2028; Focusing on ER-neg, p53 mutated tumors, the pCR rates rose from 11%, 6%, 32% to 52% in R1, R2, R3 and R4 groups. The R1 and R4 regimen, differing only in C dose intensity, showed marked differences in pCR in ERneg tumors.&#x2028; &#x2028; Conclusions: Increasing doses of C do not seem to improve pCR rates in P53 wild type tumours, raising the possibility of antagonism with anthracycline in this group. On the other hand, in ER-neg, p53 mutant tumors, inclusion of dose-intense C seems to significantly increase pCR rates.

Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 6064.

Hepatic artery triplet chemotherapy for liver metastases from colorectal cancer

14554

Background: Chronomodulated (Chrono) irinotecan (I), 5-fluorouracil (F), leucovorin and oxaliplatin (O) offered sustained tumor control in CRC patients (pts) failing several chemo (Gholam et al. The Oncologist 2006). The relevance of Chrono IFO as hepatic artery infusion (HAI) for long term outcome was investigated in CRC pts with liver metastases. Methods: 32 heavily pretreated non hospitalized CRC pts received 5-day (d) q21 d courses (c) with d1 I (160 mg/m2 from 2 to 8 am, peak at 5 am) and d2–5 F (600 mg/m2/d from 10 pm to 10 am, peak at 4 am) and O (20 mg/m2/d from 10 am to 10 pm, peak at 4 pm). Intravenous cetuximab was also given to 3 pts. 172 courses (c) were given (median, 5; 1–15) using a multichannel pump (Mélodie, Aguettant, F). Toxicity was assessed q21 d and response q3 c with CT scan. Results: Prior chemo lines 1/2/3/4+: 3/5/10/14 pts; WHO Performance Status 0/1/2/3: 14/11/6/1 pts; median age: 63 years (32–73); liver only: 25 pts; liver and lung: 7 pts. Treatment was withdrawn for thrombosis (7 pts, 21%) and/or Grade (G) 3 abdominal pain (4 pts ). G3–4 diarrhea occurred in 6 pts (12%). G3–4 leucopenia, anemia and thrombocytopenia were respectively found in 5, 2 and 1 pt. G3 sensory neuropathy occurred in 3 pts with similar G at baseline. 3 pts displayed alopecia. Of 29 pts with measurable lesions, disease progressed in 12 pts (exclusively outside the liver for 4 pts) and was controlled in 17 pts (58%), including 10 objective responses - 34% [95% C.L. 13.4 to 50.6]. Partial hepatectomy was performed in 4 pts with measurable disease (14%): R0 (3 PR) and R1 (1 SD). Median Progression free survival (months, m) is 5 m [1.9 to 8.0] and median survival is 18.4 m [9.9 to 26.9], with 6 pts alive at 2.9 to 63 m. Conclusions: Triplet chronoHAI is safe in heavily pretreated pts and achieves consistent activity against CRC liver metastases despite prior failure on the same 3 drugs. The combination of systemic cetuximab with triplet HAI is feasible and could prevent extra hepatic dissemination, a hypothesis to be soon tested in a prospective European trial. Supported by ARTBC, Hôpital P. Brousse, Villejuif, France.

No significant financial relationships to disclose.

Triple-negative phenotype is a strong predictor of sensitivity to epirubicin-cyclophosphamide (EC) then docetaxel (D) (ECD) primary chemotherapy (PCT) for localized breast cancer

21128

Background: Molecular markers (GEP, p53 mutations,) could overcome usual predictors (size, pathology, Hormone receptors, HER2) in identifying patients (pts) experiencing complete pathological response (pCR) with anthracyclin based chemotherapy (Clin.Cancer Res., 2004, 10 6789). We aimed at validating and refining these finding in pts treated with ECD. Methods: From 05/2004 to 04/2006 170 pts not amenable to Breast Conserving Therapy and/or with high evolutive potential were randomly allocated to EC (75/750mg/sqm)x4 then D (100 mg/sqm)x 4 (with or without celecoxib in HER2-ve or trastuzumab (T) in HER2+ve. The primary endpoint - absence of residual invasive breast carcinoma and of nodal involvement (pCR)- was to be correlated with usual predictors , phenotype, GEP and p53 mutations assessed from core biopsies. pCR ranged from 13 to 14% in the arms without T thus without suggestion of a difference between these arms. pCR in the 30 HER2+ve pts having received ECD + T was 30% (NS). Results in 135 fully evaluable pts not allocated to T and having undergone secondary surgery are analyzed. Results: Main predictors and related pCR are shown in the table below Results of ongoing molecular analysis will be reported. Conclusions: Expression of ER appears to be the major prognosticator for ECD induced pCR. Triple negative breast cancers experience the highest pCR rate (p< 0.0001) (chi2 test with Yates correction). Molecular studies to be presented will show if GEP and/or p53 mutations could allow to improve such prediction.

[Table: see text]

No significant financial relationships to disclose.