- Email: [email protected]
P237 The impact of body composition change on neo-adjuvant chemotherapy for breast cancer patients
S106
Poster Abstracts II / The Breast 24S1 (2015) S87–S150
identified gene was structurally validated by RT-PCR using the remaining RNA of the same samples previously examined by microarray. Pathway analysis was carried out to predict the major functional pathways. Results: Among a number of differentially expressed genes, BRIP1 showing a 5-fold up-regulation was identified as a potential gene that might underpin the transition to the malignant phenotype. The differential expression of BRIP-1 was structurally validated by RT-PCR. Conclusion: Ongoing sequencing of these genes using DNA extracted from the same samples will ultimately identify any genetic alteration that can affect the normal function of these genes. Cell migration/invasion assays will validate further the physiological relevance of BRIP-1 in tumor malignancy, and perhaps other novel gene(s) specific to BC in the Omani population. This study discovered, BRIP-1 as a novel potential marker for BC malignancy, and might be considered as a biomarker for early diagnosis of BC and/or a target to pave the way towards the design of anti-BC therapeutic strategies. Disclosure of Interest: No significant relationships. P235 Prognostic and predictive value of tumor-infiltrating lymphocytes in triple negative breast cancer E. Kim1 *, J. Kwon1 , W. Byon1 , B.J. Song2 , C. Park1 . 1 Breast-thyroid Cancer Center, Department of Surgery, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, Korea, 2 Department of Surgery, Seoul St. Mary’s Hospital, The Catholic University of Korea, Seoul, Korea Goals: Previous preclinical and clinical data suggest that increased lymphocytic infiltration would be associated with good prognosis and benefit from immunogenic chemotherapy especially in triple negative breast cancer (TNBC). We investigated a single-center experience of TNBC and relationship with lymphocytic infiltration. Methods: From January 2004 to December 2012, at department of surgery, Kangbuk Samsung Hospital, Sungkyunkwan University, School of Medicine, we retrospectively reviewed 897 breast cancer patients-clinical outcomes, clinicopathological characteristics, breast cancer subtypes. And we reviewed lymphocytic infiltration of TNBC specimens by two pathologists. Statistical analysis of risk factors associated with recurrence was performed. Results: A total of 897 patients, 76 were TNBC (8.47%). Mean age of TNBC patients were 50.95 (SD10.42) years, mean followup periods was 40.06 months. We reviewed 49 slides, and there were 8 recurrent breast cancer patients (16.32%), and 4 patients were expired (8.16%). There were 9 lymphocytic predominant breast cancers (LPBC) – carcinomas with either intratumoral lymphocytes in >60% of tumor cell nests. 1 patient of LPBC was recurred and 8 were not. In multivariate logistic regression, the odds ratio of lymphocytic infiltration was 0.59 (p = 0.643). Conclusion: In a single-center experience of TNBC, the lymphocytic infiltration in tumor cell nest might be good trend on the prognosis but there was not statistically significant. Further study with more patients will be needed. Disclosure of Interest: No significant relationships.
P236 Intraoperative predictive model for non-SLN metastasis using total tumor load assessed by OSNA 1 2 ´ J.I. Sanchez-M ´ endez ´ *, C. Mart´ı Alvarez , A. Roman ´ Guindo2 , L. Paz 3 4 Ram´ırez , L. Yebenes ´ Gregorio , G. Steinberg Contreras2 , S. Serrano Velayos2 , A. Rychlik2 , J. De Santiago Garc´ıa1 , D. Hardisson4 . 1 Breast Unit – Gynaecologist, Hospital Universitario La Paz – IdiPaz, Madrid, Spain, 2 Breast Unit – Gynaecologist, Hospital Universitario La Paz, Madrid, Spain, 3 Universidad Aut´ onoma de Madrid, Madrid, Spain, 4 Breast Unit – Pathologist, Hospital Universitario La Paz – IdiPaz, Madrid, Spain
Goals: The aim of this study is to develope an intraoperative prediction model of non sentinel lymph node (SLN) axillary metastasis in breast cancer patients with macrometastasis in SLN. Methods: The study included patients with clinically and ultrasonographically node-negative, cT1–3 invasive breast cancer, who had undergone intraoperative sentinel lymph node evaluation by one step nucleic acid amplification (OSNA) with a result of macrometastasis. A logistic regression model, “La Paz Score”, modified from previously published “Teramoto Score”, including tumor size, number of affected SLN, total tumor load (TTL) of cytokeratin 19 (CK19) mRNA, histopathological and molecular phenotype of the tumor was developed to predict intraoperatively the likehood of non-SLN axillary metastasis. The discriminating ability of some variables and logistic regression models, was assessed by plotting the ROC curve and computing the area under the curve (AUC) and corresponding 95% confidence interval (CI). All statistical analyses were performed with the statistical language R version 3.0.1. Results: Ninety patients were recruited. The size of tumors ranged from 5 to 100 mm, with median (IQR) of 19.5 (12.75). TTL values (expressed as 103 copies/uL) ranged from 5.34 to 8401, with median (IQR) of 44.5 (219.5). The number of positive SLN were 1 (64.4%), 2 (27.8%) and 3 (7.8%). Sixty-five cases (72.2%) were ductal invasive carcinomas, 23 (25.6%) cases were lobular invasive, and 2 (2.2%) other histologic subtypes. Regarding the phenotypic molecular type 57.8% were luminal A, 38.9% luminal B, and 3.3% basal type. There is no Her2 positive tumors. In 60 patients (66.7%) there were no more axillary metastasis, and in 30 patients (33.3%) there were at least one non-SLN axillary metastasis. The negative predictive value of our model is 90.24%, with a sensitivity of 86.67%, a specificity of 61.67% and a predictive positive value of 53.06%. Model
AUC
[CI]
Teramoto Score La Paz Score
0.56 0.79
[0.44 to 0.68] [0.69 to 0.89]
Conclusion: We showed that a predictive model with five variables may significantly improve the discrimination ability of the score proposed by Teramoto (p = 0.001). This logistic regression model including the (log10) TTL, tumor size, number of positive SLN, histology and molecular subtype provided a reasonable discrimination ability (AUC = 0.79). However, this should be validated in a new sample of larger size. Disclosure of Interest: No significant relationships. P237 The impact of body composition change on neo-adjuvant chemotherapy for breast cancer patients T. Iwase *, T. Sangai, E. Ishigami, J. Sakakibara, K. Fujisaki, N. Shiina, H. Fujimoto, M. Sakakibara, T. Nagashima, M. Miyazaki. General Surgery, Chiba University Hospital, Chiba, Japan Goals: Studies have reported that obesity decreases the chemotherapeutic efficacy of neo-adjuvant chemotherapy (NAC)
14th St.Gallen International Breast Cancer Conference / The Breast 24S1 (2015) S87–S150
S107
on breast cancer. However, the relationship between actual body composition and therapy outcome is still unknown. Therefore, we set out to clarify the effect of body composition on NAC. Methods: A total of 172 cases with advanced breast cancer who underwent surgery after NAC between January 2004 and December 2012 were retrospectively analyzed. Abdominal circumference (AC), Subcutaneous fat area (SFA), visceral fat area (VFA), and skeletal muscle area (SMA) were calculated using computed tomography volume analyzing software, SYNAPSE VINCENT® (FUJIFILM Co., Ltd. Tokyo, Japan). VFA/SFA ratio was used for evaluating visceral obesity. The relationships between those body composition parameters, pathological complete remission (pCR) and survival prognosis were analyzed. Results: Firstly, AC, SFA, and VFA were significantly correlated to body mass index (BMI) in body composition analysis (p < 0.05, AC; r = 0.82. SFA; r = 0.71, VFA; r = 0.78). AC, SFA, and VFA were significantly increased after menopause, while SMA was significantly decreased (p < 0.05). In addition, VFA/SFA ratio was significantly increased after menopause, despite no significant change in BMI. Secondly, a total of 46 cases achieved pCR. However, body composition parameters were not associated with pCR. Lastly, survival analysis demonstrated high VFA group (>100 cm2 ) was significantly had worse DDFS compared to the low VFA group (p < 0.05). Moreover, that difference was more significant in postmenopausal group. Multivariate analysis demonstrated VFA pCR, and subtypes were independent prognostic factors for DDFS. Conclusion: Present study suggests that increased visceral fat after menopause significantly relates to the worse outcome after NAC. It is highly probable to give an appropriate intervention to those high-risk cases improve the outcome. Disclosure of Interest: No significant relationships.
low and 32% as high risk. Therefore, treatment according to the EP result would have reduced total adjuvant chemotherapy usage by 18%. 27% of patients who had received chemotherapy would not receive it if treated according to EP, 9% had not received chemotherapy and were classified as high risk by EP. In 64% of patients no change of risk classification was observed. Interestingly, re-classification by EP was particularly found in patients with small tumors: 34 of the 36 patients with change of risk classification had T1a/b or T1c tumors. These results suggest that EP would not only have reduced over-treatment but also under-treatment, especially of small tumors. Assuming direct costs for chemotherapy ranging from € 11,296 (Blank et al., 2014) to € 19,263 (Hornberger et al., 2012) a treatment following the EP result would have saved € 255–1,713 per patient in this single center cohort. The cost of EP was assumed to be € 1,800. Conclusion: Our study suggests that EP applied in patients with intermediate risk according to classical guidelines can help reducing chemotherapy usage without impairment of outcome. Disclosure of Interest: GS and WS have received honoraria from Sividon Diagnostics GmbH for lectures and advisory boards concerning EP. GS and WS have received research support from Sividon Diagnostics. RK and CP are shareholders and employees of Sividon Diagnostics GmbH.
P238 EndoPredict-based treatment decision can reduce chemotherapy usage in ER+, HER2− breast cancer
Goals: The most widely used tumour markers in breast cancer are CA 15-3 and CEA. Current literature does not recommend the use of both tumour markers in detecting disease recurrence mainly due to a lack of specificity. This cohort study therefore aims to investigate the value of CA 15-3 and CEA in breast cancer surveillance and detection of disease recurrence. Methods: Consecutive patients with available serum CA 15-3 and CEA level at initial diagnosis of recurrence in a prospectively collected database over 15 years (1998–2013) were reviewed. A second group of patients with no recurrence who were matched in terms of demographic and tumour characteristics were included as a control. Patients with metastatic disease at presentation or those without surveillance tumour markers performed were excluded from the study. Statistical analysis was performed using SPSS Version 20. Results: A total of 137 patients (67 patients with recurrence and control group of 70 patients) were selected from a prospectively collected database over a 15-year period. ROC curve were plotted for both CA 15-3 and CEA which showed high discriminatory power in detecting recurrence with an AUC of 0.856 and 0.901 for CA 15-3 and CEA respectively (p = 0.001). Elevation of CA 15-3 and CEA above the normal upper limit of 25 U/ml and 5 ug/L respectively have been shown to be highly predictive of recurrence. CA 15-3 has a specificity of 97.1% and sensitivity of 71.6% (p = 0.001) whereas CEA was also statistically significant (p = 0.001) with specificity of 95.7% and sensitivity of 67.2%. Combined measurement of CA 15-3 and CEA level were shown to be statistically significant (p = 0.001) with improved specificity of 91.8%, sensitivity of 94.2% as well as PPV and NPV at 0.93. Logistic regression analysis was performed to investigate the interaction between tumour characteristics and disease recurrence. In univariate analysis, elevation of tumour markers above normal upper limit (p = 0.01), size of tumour (p = 0.05), axillary lymph node involvement (p = 0.004) as well as histological grade (p = 0.003) have been shown to be highly significant factors in predicting recurrence.
G. Schlake1 *, R. Kronenwett2 , F. Tiecke1 , K. Kastrup3 , N. Bleuel1 , 4 C. Petry2 , S. Wilhelms3 , P. Tonnies ¨ , W. Schlake1 . 1 Pathologie, Institut f¨ ur Pathologie und Molekularpathologie, Gelsenkirchen, Germany, 2 Research and Development, Sividon Diagnostics GmbH, Cologne, Germany, 3 Klinik f¨ ur Frauenheilkunde und GeburtshilfeBrustzentrum am St. Elisabeth-Krankenhaus, Dorsten, Germany, 4 Frauenklinik/Brustzentrum am Krankenhaus Bethanien, Moers, Germany Goals: EndoPredict (EP) is a prognostic gene expression test for patients with estrogen receptor (ER) positive, HER2 negative primary breast cancer. It can be performed locally by the pathologist. The aim of this retrospective study was to evaluate whether treatment according to the EP result instead of the classical risk parameters alone would have an impact on chemotherapy usage in a German breast center. Methods: All patients with ER+, HER2− primary breast cancer with up to 3 positive lymph nodes admitted to one hospital between January 2008 and June 2011 were screened. All intermediate-risk patients according to German S3 Guideline (2008) who were eligible for chemotherapy were included in this study. Archived tumor tissue was retrospectively analyzed with EP by the local pathologist and the molecular-clinico-pathological EPclin score was calculated using gene expression values, tumor size and nodal status. Risk classification into low or high risk by the EPclin score was compared with the original therapy decision that the tumor board had made before EP became available. Results: 82 patients were included in this study (age: 37–75, median: 62). 58 were node negative, 24 had 1 to 3 positive lymph nodes. 50% had actually received adjuvant chemotherapy in addition to endocrine therapy. Using EP, 68% of patients were classified as
P239 Clinical significance of CA 15-3 and CEA in detection of breast cancer recurrence C.W. Mok1 *, S.K. Lim1 , J. Seah1 , J.W. Kam2 , S.M. Tan1 . 1 Breast Service, General Surgery, Changi General Hospital, Singapore, Singapore, 2 Clinical Trials and Research Unit, Changi General Hospital, Singapore, Singapore
Poster Abstracts II / The Breast 24S1 (2015) S87–S150
identified gene was structurally validated by RT-PCR using the remaining RNA of the same samples previously examined by microarray. Pathway analysis was carried out to predict the major functional pathways. Results: Among a number of differentially expressed genes, BRIP1 showing a 5-fold up-regulation was identified as a potential gene that might underpin the transition to the malignant phenotype. The differential expression of BRIP-1 was structurally validated by RT-PCR. Conclusion: Ongoing sequencing of these genes using DNA extracted from the same samples will ultimately identify any genetic alteration that can affect the normal function of these genes. Cell migration/invasion assays will validate further the physiological relevance of BRIP-1 in tumor malignancy, and perhaps other novel gene(s) specific to BC in the Omani population. This study discovered, BRIP-1 as a novel potential marker for BC malignancy, and might be considered as a biomarker for early diagnosis of BC and/or a target to pave the way towards the design of anti-BC therapeutic strategies. Disclosure of Interest: No significant relationships. P235 Prognostic and predictive value of tumor-infiltrating lymphocytes in triple negative breast cancer E. Kim1 *, J. Kwon1 , W. Byon1 , B.J. Song2 , C. Park1 . 1 Breast-thyroid Cancer Center, Department of Surgery, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, Korea, 2 Department of Surgery, Seoul St. Mary’s Hospital, The Catholic University of Korea, Seoul, Korea Goals: Previous preclinical and clinical data suggest that increased lymphocytic infiltration would be associated with good prognosis and benefit from immunogenic chemotherapy especially in triple negative breast cancer (TNBC). We investigated a single-center experience of TNBC and relationship with lymphocytic infiltration. Methods: From January 2004 to December 2012, at department of surgery, Kangbuk Samsung Hospital, Sungkyunkwan University, School of Medicine, we retrospectively reviewed 897 breast cancer patients-clinical outcomes, clinicopathological characteristics, breast cancer subtypes. And we reviewed lymphocytic infiltration of TNBC specimens by two pathologists. Statistical analysis of risk factors associated with recurrence was performed. Results: A total of 897 patients, 76 were TNBC (8.47%). Mean age of TNBC patients were 50.95 (SD10.42) years, mean followup periods was 40.06 months. We reviewed 49 slides, and there were 8 recurrent breast cancer patients (16.32%), and 4 patients were expired (8.16%). There were 9 lymphocytic predominant breast cancers (LPBC) – carcinomas with either intratumoral lymphocytes in >60% of tumor cell nests. 1 patient of LPBC was recurred and 8 were not. In multivariate logistic regression, the odds ratio of lymphocytic infiltration was 0.59 (p = 0.643). Conclusion: In a single-center experience of TNBC, the lymphocytic infiltration in tumor cell nest might be good trend on the prognosis but there was not statistically significant. Further study with more patients will be needed. Disclosure of Interest: No significant relationships.
P236 Intraoperative predictive model for non-SLN metastasis using total tumor load assessed by OSNA 1 2 ´ J.I. Sanchez-M ´ endez ´ *, C. Mart´ı Alvarez , A. Roman ´ Guindo2 , L. Paz 3 4 Ram´ırez , L. Yebenes ´ Gregorio , G. Steinberg Contreras2 , S. Serrano Velayos2 , A. Rychlik2 , J. De Santiago Garc´ıa1 , D. Hardisson4 . 1 Breast Unit – Gynaecologist, Hospital Universitario La Paz – IdiPaz, Madrid, Spain, 2 Breast Unit – Gynaecologist, Hospital Universitario La Paz, Madrid, Spain, 3 Universidad Aut´ onoma de Madrid, Madrid, Spain, 4 Breast Unit – Pathologist, Hospital Universitario La Paz – IdiPaz, Madrid, Spain
Goals: The aim of this study is to develope an intraoperative prediction model of non sentinel lymph node (SLN) axillary metastasis in breast cancer patients with macrometastasis in SLN. Methods: The study included patients with clinically and ultrasonographically node-negative, cT1–3 invasive breast cancer, who had undergone intraoperative sentinel lymph node evaluation by one step nucleic acid amplification (OSNA) with a result of macrometastasis. A logistic regression model, “La Paz Score”, modified from previously published “Teramoto Score”, including tumor size, number of affected SLN, total tumor load (TTL) of cytokeratin 19 (CK19) mRNA, histopathological and molecular phenotype of the tumor was developed to predict intraoperatively the likehood of non-SLN axillary metastasis. The discriminating ability of some variables and logistic regression models, was assessed by plotting the ROC curve and computing the area under the curve (AUC) and corresponding 95% confidence interval (CI). All statistical analyses were performed with the statistical language R version 3.0.1. Results: Ninety patients were recruited. The size of tumors ranged from 5 to 100 mm, with median (IQR) of 19.5 (12.75). TTL values (expressed as 103 copies/uL) ranged from 5.34 to 8401, with median (IQR) of 44.5 (219.5). The number of positive SLN were 1 (64.4%), 2 (27.8%) and 3 (7.8%). Sixty-five cases (72.2%) were ductal invasive carcinomas, 23 (25.6%) cases were lobular invasive, and 2 (2.2%) other histologic subtypes. Regarding the phenotypic molecular type 57.8% were luminal A, 38.9% luminal B, and 3.3% basal type. There is no Her2 positive tumors. In 60 patients (66.7%) there were no more axillary metastasis, and in 30 patients (33.3%) there were at least one non-SLN axillary metastasis. The negative predictive value of our model is 90.24%, with a sensitivity of 86.67%, a specificity of 61.67% and a predictive positive value of 53.06%. Model
AUC
[CI]
Teramoto Score La Paz Score
0.56 0.79
[0.44 to 0.68] [0.69 to 0.89]
Conclusion: We showed that a predictive model with five variables may significantly improve the discrimination ability of the score proposed by Teramoto (p = 0.001). This logistic regression model including the (log10) TTL, tumor size, number of positive SLN, histology and molecular subtype provided a reasonable discrimination ability (AUC = 0.79). However, this should be validated in a new sample of larger size. Disclosure of Interest: No significant relationships. P237 The impact of body composition change on neo-adjuvant chemotherapy for breast cancer patients T. Iwase *, T. Sangai, E. Ishigami, J. Sakakibara, K. Fujisaki, N. Shiina, H. Fujimoto, M. Sakakibara, T. Nagashima, M. Miyazaki. General Surgery, Chiba University Hospital, Chiba, Japan Goals: Studies have reported that obesity decreases the chemotherapeutic efficacy of neo-adjuvant chemotherapy (NAC)
14th St.Gallen International Breast Cancer Conference / The Breast 24S1 (2015) S87–S150
S107
on breast cancer. However, the relationship between actual body composition and therapy outcome is still unknown. Therefore, we set out to clarify the effect of body composition on NAC. Methods: A total of 172 cases with advanced breast cancer who underwent surgery after NAC between January 2004 and December 2012 were retrospectively analyzed. Abdominal circumference (AC), Subcutaneous fat area (SFA), visceral fat area (VFA), and skeletal muscle area (SMA) were calculated using computed tomography volume analyzing software, SYNAPSE VINCENT® (FUJIFILM Co., Ltd. Tokyo, Japan). VFA/SFA ratio was used for evaluating visceral obesity. The relationships between those body composition parameters, pathological complete remission (pCR) and survival prognosis were analyzed. Results: Firstly, AC, SFA, and VFA were significantly correlated to body mass index (BMI) in body composition analysis (p < 0.05, AC; r = 0.82. SFA; r = 0.71, VFA; r = 0.78). AC, SFA, and VFA were significantly increased after menopause, while SMA was significantly decreased (p < 0.05). In addition, VFA/SFA ratio was significantly increased after menopause, despite no significant change in BMI. Secondly, a total of 46 cases achieved pCR. However, body composition parameters were not associated with pCR. Lastly, survival analysis demonstrated high VFA group (>100 cm2 ) was significantly had worse DDFS compared to the low VFA group (p < 0.05). Moreover, that difference was more significant in postmenopausal group. Multivariate analysis demonstrated VFA pCR, and subtypes were independent prognostic factors for DDFS. Conclusion: Present study suggests that increased visceral fat after menopause significantly relates to the worse outcome after NAC. It is highly probable to give an appropriate intervention to those high-risk cases improve the outcome. Disclosure of Interest: No significant relationships.
low and 32% as high risk. Therefore, treatment according to the EP result would have reduced total adjuvant chemotherapy usage by 18%. 27% of patients who had received chemotherapy would not receive it if treated according to EP, 9% had not received chemotherapy and were classified as high risk by EP. In 64% of patients no change of risk classification was observed. Interestingly, re-classification by EP was particularly found in patients with small tumors: 34 of the 36 patients with change of risk classification had T1a/b or T1c tumors. These results suggest that EP would not only have reduced over-treatment but also under-treatment, especially of small tumors. Assuming direct costs for chemotherapy ranging from € 11,296 (Blank et al., 2014) to € 19,263 (Hornberger et al., 2012) a treatment following the EP result would have saved € 255–1,713 per patient in this single center cohort. The cost of EP was assumed to be € 1,800. Conclusion: Our study suggests that EP applied in patients with intermediate risk according to classical guidelines can help reducing chemotherapy usage without impairment of outcome. Disclosure of Interest: GS and WS have received honoraria from Sividon Diagnostics GmbH for lectures and advisory boards concerning EP. GS and WS have received research support from Sividon Diagnostics. RK and CP are shareholders and employees of Sividon Diagnostics GmbH.
P238 EndoPredict-based treatment decision can reduce chemotherapy usage in ER+, HER2− breast cancer
Goals: The most widely used tumour markers in breast cancer are CA 15-3 and CEA. Current literature does not recommend the use of both tumour markers in detecting disease recurrence mainly due to a lack of specificity. This cohort study therefore aims to investigate the value of CA 15-3 and CEA in breast cancer surveillance and detection of disease recurrence. Methods: Consecutive patients with available serum CA 15-3 and CEA level at initial diagnosis of recurrence in a prospectively collected database over 15 years (1998–2013) were reviewed. A second group of patients with no recurrence who were matched in terms of demographic and tumour characteristics were included as a control. Patients with metastatic disease at presentation or those without surveillance tumour markers performed were excluded from the study. Statistical analysis was performed using SPSS Version 20. Results: A total of 137 patients (67 patients with recurrence and control group of 70 patients) were selected from a prospectively collected database over a 15-year period. ROC curve were plotted for both CA 15-3 and CEA which showed high discriminatory power in detecting recurrence with an AUC of 0.856 and 0.901 for CA 15-3 and CEA respectively (p = 0.001). Elevation of CA 15-3 and CEA above the normal upper limit of 25 U/ml and 5 ug/L respectively have been shown to be highly predictive of recurrence. CA 15-3 has a specificity of 97.1% and sensitivity of 71.6% (p = 0.001) whereas CEA was also statistically significant (p = 0.001) with specificity of 95.7% and sensitivity of 67.2%. Combined measurement of CA 15-3 and CEA level were shown to be statistically significant (p = 0.001) with improved specificity of 91.8%, sensitivity of 94.2% as well as PPV and NPV at 0.93. Logistic regression analysis was performed to investigate the interaction between tumour characteristics and disease recurrence. In univariate analysis, elevation of tumour markers above normal upper limit (p = 0.01), size of tumour (p = 0.05), axillary lymph node involvement (p = 0.004) as well as histological grade (p = 0.003) have been shown to be highly significant factors in predicting recurrence.
G. Schlake1 *, R. Kronenwett2 , F. Tiecke1 , K. Kastrup3 , N. Bleuel1 , 4 C. Petry2 , S. Wilhelms3 , P. Tonnies ¨ , W. Schlake1 . 1 Pathologie, Institut f¨ ur Pathologie und Molekularpathologie, Gelsenkirchen, Germany, 2 Research and Development, Sividon Diagnostics GmbH, Cologne, Germany, 3 Klinik f¨ ur Frauenheilkunde und GeburtshilfeBrustzentrum am St. Elisabeth-Krankenhaus, Dorsten, Germany, 4 Frauenklinik/Brustzentrum am Krankenhaus Bethanien, Moers, Germany Goals: EndoPredict (EP) is a prognostic gene expression test for patients with estrogen receptor (ER) positive, HER2 negative primary breast cancer. It can be performed locally by the pathologist. The aim of this retrospective study was to evaluate whether treatment according to the EP result instead of the classical risk parameters alone would have an impact on chemotherapy usage in a German breast center. Methods: All patients with ER+, HER2− primary breast cancer with up to 3 positive lymph nodes admitted to one hospital between January 2008 and June 2011 were screened. All intermediate-risk patients according to German S3 Guideline (2008) who were eligible for chemotherapy were included in this study. Archived tumor tissue was retrospectively analyzed with EP by the local pathologist and the molecular-clinico-pathological EPclin score was calculated using gene expression values, tumor size and nodal status. Risk classification into low or high risk by the EPclin score was compared with the original therapy decision that the tumor board had made before EP became available. Results: 82 patients were included in this study (age: 37–75, median: 62). 58 were node negative, 24 had 1 to 3 positive lymph nodes. 50% had actually received adjuvant chemotherapy in addition to endocrine therapy. Using EP, 68% of patients were classified as
P239 Clinical significance of CA 15-3 and CEA in detection of breast cancer recurrence C.W. Mok1 *, S.K. Lim1 , J. Seah1 , J.W. Kam2 , S.M. Tan1 . 1 Breast Service, General Surgery, Changi General Hospital, Singapore, Singapore, 2 Clinical Trials and Research Unit, Changi General Hospital, Singapore, Singapore