CLINICAL RELEVANCE OF GENOMIC ABERRATIONS IN HOMOGENEOUSLY TREATED HIGH-RISK STAGE II/III BREAST CANCER PATIENTS Alexander SEUTE 1 , Hans-Peter SINN 2 , Richard F. SCHLENK 1 , Robert EMIG 3 , Diethelm WALLWIENER 3 , Eva-Maria GRISCHKE 4 , Stefan HOHAUS 5 , Hartmut D¨ OHNER 1 , Rainer HAAS 6 and Martin BENTZ 1 * 1 Abt. Innere Medizin III, Universita ¨t Ulm, Ulm, Germany 2 Pathologisches Institut, Universita ¨t Heidelberg, Heidelberg, Germany 3 Frauenklinik, Universita ¨t Tu ¨bingen, Tu ¨bingen, Germany 4 Frauenklinik, Universita ¨t Heidelberg, Heidelberg, Germany 5 Istituto di Semeiotica Medica, Universita ´ Cattolica del Sacro Cuore, Rome, Italy 6 Klinik fu ¨r Ha ¨matologie, Onkologie und klinische Immunologie, Universita ¨t Du ¨sseldorf, Germany Little is known about the prognostic impact of chromo- some aberrations in breast cancer. The aim of our study was to determine whether genomic aberrations of prognostic relevance can be identified in the context of a clinical study using molecular cytogenetics. Paraffin-embedded tumor samples of 44 patients with high-risk stage II/III breast cancer were analyzed by comparative genomic hybridization. All patients received identical therapy including dose-escalated chemotherapy followed by peripheral blood stem cell trans- plantation. The most frequent chromosomal aberrations were gains on chromosome arms 17q (24 cases), 1q (21 cases), 8q (17 cases), 20q (13 cases), 6p (9 cases) as well as losses on chromosome arms 13q (25 cases), 11q (20 cases), 5q (11 cases), 6q (11 cases), 9p (10 cases), 18q (10 cases), 8p (9 cases) and 16q (9 cases). In univariate analysis, the corre- lation with the clinical outcome revealed a higher risk for patients with tumors exhibiting 13q losses and a reduced risk for tumors exhibiting 16q losses (p  0.020), 6q losses (p  0.041) and estrogen-receptor positivity (0.051). In multivari- ate analysis using the Cox model, only the loss of 16q exhib- ited borderline significance (p  0.065). These data show that comparative genomic hybridization can be performed in the context of a clinical trial. In our subgroup of high-risk breast cancer patients, chromosomal aberrations were valuable prognostic parameters. © 2001 Wiley-Liss, Inc. Key words: breast cancer; stem cell transplantation; chromosomes; prognosis Risk-adapted treatment strategies have become the standard therapeutic approach for patients with breast cancer. The single most relevant prognostic factor in primary breast cancer still is the axillary lymph node status. Other prognostic factors include tumor size, histologic subtype, tumor grade and estrogen-receptor and progesterone-receptor status. 1 Despite the proven value of these clinicopathological risk factors, there is a considerable proportion of patients for whom the clinical course is not reliably predicted. Therefore, there is a clear need for the definition of additional prognostic variables. In leukemias, genomic aberrations have proven to be closely associated with the clinical outcome, 2 and treatment trials have been stratified according to genetic risk categories. Genetic abnor- malities in breast cancer were also related to the clinical course. Both HER2/NEU protein overexpression and gene amplification were associated with an inferior outcome, 3–6 as was the overex- pression of the p53 protein and mutations of the TP53 tumor- suppressor gene. 7,8 However, still there are no comprehensive genetic studies in homogeneous cohorts of patients receiving iden- tical therapies. We therefore analyzed tumor samples of stage II and stage III patients who received the same adjuvant high-dose chemotherapy following surgery. In all samples, comparative genomic hybrid- ization was performed. With this technique, a comprehensive overview of chromosomal gains and losses can be obtained within a single experiment. For all patients in this homogeneous group, clinical data were correlated with the most frequent genetic find- ings. MATERIAL AND METHODS Patients Paraffin-embedded tumor samples of 44 women with UICC stage II and III high-risk breast cancer were analyzed by CGH. These patients were treated at the University of Heidelberg and are a subgroup of a larger study, the clinical data of which have been published previously. 9 This subgroup was defined based on the availability of appropriate tumor material. Forty of these patients had tumor involvement of more than 10 axillary lymph nodes, and 4 patients had less than 10 positive lymph nodes but additional high-risk factors such as hormone receptor-negative tumor, high S phase or tumor cells in the bone marrow. For all cases DNA was isolated from paraffin-embedded samples of the primary carci- noma. The patients were enrolled within 4 weeks after operation for the adjuvant treatment protocol. Table I shows the patient characteristics including stage of the disease and menopausal and hormone receptor status. Cytotoxic chemotherapy Between 1993 and 1997 the patients received adjuvant treatment including high-dose chemotherapy followed by autologous stem cell transplantation as previously described. 9 Briefly, initial cyto- toxic chemotherapy consisted of 2 cycles of ifosfamide (2,500 mg/m 2 , 3 days) and epirubicin (40 mg/m 2 , 3 days). The chemo- therapy was continued with 2 cycles of peripheral blood stem cell (PBSC)-supported high-dose ifosfamide (total dose of 12,000 mg/ m 2 ), epirubicin (180 mg/m 2 ) and carboplatin (900 mg/m 2 ). The dose of all drugs was delivered over a period of 5 days. In the first 4 patients treated in 1993, carboplatin was not included in the high-dose regimen. PBSC were reinfused following 1 day after the end of cytotoxic chemotherapy. Premenopausal women received antihormonal therapy with goserelin (3.6 mg) administered subcu- taneously once per month. Postmenopausal women received ta- moxifen orally at a daily dose of 30 mg. Comparative genomic hybridization (CGH) Genomic DNA was prepared from paraffin-embedded tumor tissue as described using proteinase K digestion and phenol-chlo- Grant sponsor: Deutsche Krebshilfe; Grant numbers: 47/95 Be I, 70- 2312 Be2. *Correspondence to: Abt. Innere Medizin III, Robert-Koch-Str. 8, 89081 Ulm, Germany. Fax: +49-731-500 24492. E-mail: martin.bentz@medizin.uni-ulm.de Received 19 September 2000; Revised 15 January 2001; Accepted 29 January 2001 Published online 6 April 2001 Int. J. Cancer: 93, 80 – 84 (2001) © 2001 Wiley-Liss, Inc. Publication of the International Union Against Cancer