Unsupervised HCL of an external dataset of breast cancer metastases (GSE14018) using the 358 "breast cancer liver metastasis-selective genes" derived from SAM analyses of metastases in the test dataset.

A. Association between the 17-gene signature and IntClust subgroups. P-value is from anova test. B. Relapse free survival for the subset of patients in IntClust subgroups 3, 7 and 8, which are enriched for Luminal A tumors. 73% of tumors with high expression belonged to IntClust subgroup 3, while 41% of tumors with low expression were of subgroup 7 and 8. C. Associations between the 17 liver metastasis-selective genes and RFS and D. OS for all patients included in GOBO irrespective of ER status.

Gene ontology analysis showing significantly enriched GO-terms for biological process among the significantly altered genes in breast cancer liver metastases compared to metastases from other sites.

Abstract Purpose: The complete molecular basis of the organ-specificity of metastasis is elusive. This study aimed to provide an independent characterization of the transcriptional landscape of breast cancer metastases with the specific objective to identify liver metastasis–selective genes of prognostic importance following primary tumor diagnosis. Experimental Design: A cohort of 304 women with advanced breast cancer was studied. Associations between the site of recurrence and clinicopathologic features were investigated. Fine-needle aspirates of metastases (n = 91) were subjected to whole-genome transcriptional profiling. Liver metastasis–selective genes were identified by significance analysis of microarray (SAM) analyses and independently validated in external datasets. Finally, the prognostic relevance of the liver metastasis–selective genes in primary breast cancer was tested. Results: Liver relapse was associated with estrogen receptor (ER) expression (P = 0.002), luminal B subtype (P = 0.01), and was prognostic for an inferior postrelapse survival (P = 0.01). The major variation in the transcriptional landscape of metastases was also associated with ER expression and molecular subtype. However, liver metastases displayed unique transcriptional fingerprints, characterized by downregulation of extracellular matrix (i.e., stromal) genes. Importantly, we identified a 17-gene liver metastasis–selective signature, which was significantly and independently prognostic for shorter relapse-free (P < 0.001) and overall (P = 0.001) survival in ER-positive tumors. Remarkably, this signature remained independently prognostic for shorter relapse-free survival (P = 0.001) among luminal A tumors. Conclusions: Extracellular matrix (stromal) genes can be used to partition breast cancer by site of relapse and may be used to further refine prognostication in ER positive primary breast cancer. Clin Cancer Res; 22(1); 146–57. ©2015 AACR.

Abstract Background: Several randomized trials have proven that preoperative chemotherapy is equivalent to adjuvant treatment, and allows for clinical and radiological assessment of efficacy in vivo. Recent results show that response-guided neoadjuvant therapy is a favorable policy for hormone receptor positive tumors, while pCR predicts prognosis for triple-negative (TNBC) and HER2 positive breast cancer (von Minckwitz 2013; Cortazar 2014). Since 2012 the FDA accepts pCR (ypT0 ypN0) as endpoint for accelerated drug approval. Methods: In total, 150 women 18 years or older with verified HER2 negative breast cancer suitable for primary medical treatment were included in the trial between September 2008 and December 2011. The patients received two courses of epirubicin and docetaxel (Taxotere®), both 75mg/m2 for the 1st two courses, followed by the same treatment with addition of bevacizumab (Avastin®) 15 mg/kg for 4 additional courses. Clinical and radiological evaluations with mammography and ultrasound were performed before start and after courses 2, 4 and 6. Core biopsies were taken before start, after 2 courses, and at time of surgery. Blood samples were drawn before and 24 hours after the first 4 courses. Results: Median age was 49 years, range: 27 to 70 years; 73% were reported as ductal, 15% as lobular, and 12% as rare histological types. Mean tumor size was 59 mm, median 55 mm, range: 20-180 mm; 3 tumors (2.0%) were reported as inflammatory, and 13 (8.7%) presented with skin involvement (T4b). Enlarged axillary nodes were found in 102 patients (68%) before start of treatment, 77 of these (64%) verified as metastatic. SNB in cases of normal axillary status was performed in 16 cases, in 9 cases (8%) with positive finding. Supra- or infraclavicular node involvement was verified in 20 cases (13%). 25% of all tumors were ER- and/or PR-negative, tumor grade based on a diagnostic biopsy was evaluable in only 83 cases. Of these, 4 (5%) were grade I, 46 (55%) grade II, and 33 (40%) grade III. Mean proliferation count (Ki67) was 37%, median 30%, range 1-90%. Breakdown into intrinsic subtypes based on immunohistochemistry defined 68 (46%) as luminal A-like, 36 (24%) as luminal B-like, and 44 (30%) as TNBC. pCR was achieved in 20 cases, 3 (2%) luminal A-like, 5 (3.4%) luminal B-like and 12 (8.2%) TNBC. After 2.2 years of follow-up, 35 patients (23.3%) have experienced recurrence and 18 of these (12%) have deceased due to breast cancer, among these 6 despite pCR, 2 classified as luminal B-like, and 4 as TNBC. The molecular subtype of the tumor predicted outcome, but pCR was not in our material, even after adjustment for tumor size at diagnosis, a predictor of favorable outcome. The number of events in relation to molecular subtypes is however limited. Updated outcome data will be presented. Conclusions: The present trial does not confirm previously reported observations that pCR is a marker of favorable prognosis. One possible explanation is that unfavorable biological characteristics, particularly heterogeneity, may increase with tumor burden. Genomic and proteomic analyses are currently ongoing. Citation Format: Thomas Hatschek, Judith Bjöhle, Elisabet Lidbrink, Tobias Lekberg, Niklas Loman, Anna von Wachenfeldt Väppling, Martin Söderberg, Zakaria Einbeigi, Lena Carlsson, Henrik Lindman, Irma Fredriksson, Jan Frisell, Lars Löfgren, Lisa Rydén, Mats Hellström, Mårten Fernö, Jonas Bergh. Is pathologic complete response (pCR) a valid marker of outcome even in large breast cancer? Clinical results from a neoadjuvant trial using a combination of epirubicin, docetaxel and bevacizumab (PROMIX) [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 P3-11-14.

The aim of this was to determine whether the change of size observed at the first response evaluation after initiation of first-line combination chemotherapy correlates with overall survival (OS) in patients with metastatic breast cancer (MBC). The change in size of tumors derived from measurements according to Response Evaluation Criteria In Solid Tumors (RECIST) at the first evaluation on computed tomography (CT) was obtained from a multicenter, randomized phase III trial (“TEX trial,” n = 287) comparing treatment with a combination of epirubicin and paclitaxel alone or with capecitabine (TEX). Cox regression and Kaplan–Meier analyses were performed to evaluate the correlations between the first change in tumor size, response according to RECIST and OS. Data from CT evaluations of 233 patients were available. Appearance of new lesions or progression of non-target lesions (new/non-target) indicated short OS by univariable regression analysis (HR 3.76, 95 % CI 1.90–7.42, p < 0.001). A decrease by >30 % at this early time point was prognostic favorable (HR 0.69, 95 % CI 0.49–0.98, p = 0.04) and not significantly less than the best overall response according to RECIST. After adjustment for previous adjuvant treatment and the treatment given within the frame of the randomized trial, OS was still significantly shorter in patients with new/non-target lesions after a median 8 weeks of treatment (HR 4.41, 95 % CI 2.74–7.11, p < 0.001). Disease progression at the first evaluation correlates with OS in patients with MBC treated with first-line combination chemotherapy. The main reason for early disease progression was the appearance of new lesions or progression of non-target lesions. These patients had poor OS even though more lines of treatment were available. Thus, these factors should be focused on in the response evaluations besides tumor size changes.

Abstract Background: Breast cancer molecular subtypes preferentially metastasize to specific organs, but the underlying molecular mechanisms are not well understood. While several studies have attempted to delineate these mechanisms by making predictions from primary tumours, the likelihood of missing further relevant biological differences which may be enriched or distinctively present in the metastases cannot be overlooked. In addition, survival after relapse is influenced by the anatomical location of the metastasis, with visceral involvement associated with the poorest survival, suggesting that differences in biology and response to treatment may exist. However, the scarcity of clinical material from metastatic lesions has largely prevented detailed molecular characterization of breast cancer site-specific metastases. Material and Methods: We performed global gene expression profiling on 122 fine needle aspirates of metastatic lesions from different anatomical sites obtained from breast cancer patients treated within the Swedish randomized trial (TEX) of chemotherapy for metastatic breast cancer. Sub-classification was performed using unsupervised hierarchical clustering, while differentially expressed genes and modulated pathways in metastases from different sites and sub-classes were identified using Significance Analysis of Microarrays (SAM) analysis. Breast cancer survival after relapse was also compared between patients. Results: We confirmed the observation that breast cancer metastasis to visceral organs is associated with the poorest outcome. Molecularly, samples were sub-classified into groups with strong correlations with the previously described breast cancer intrinsic subtypes and also striking overlap with pathological characteristics of the primary tumour. Metastasis pairs from the same individual were more similar to each other than to independent metastatic lesions from other patients. More interestingly, liver metastases formed a unique sub-cluster which could be distinguished from the other sites by the expression of several genes (and pathways) involved in focal adhesion, matrix remodelling, and wnt signalling. Conclusions: Our data suggest that breast cancer molecular subtypes and primary tumour characteristics are to a large extent maintained in recurrences and further that liver metastases may be notably different and represent a distinct entity with a separate underlying biology which may hence require specific targeted therapeutic management. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 3386. doi:1538-7445.AM2012-3386

The aim of the study was to evaluate the feasibility of tailored and dose-dense epirubicin and cyclophosphamide followed by docetaxel as adjuvant breast cancer therapy. Material and methods. Patients with node-positive breast cancer received either four cycles of biweekly and tailored EC (epirubicin 38-60-75-90-105-120 mg/m(2), cyclophosphamide 450-600-900-1200 mg/m(2)) followed by four cycles of docetaxel (60-75-85-100 mg/m(2)) (arm A) or the same regimen with fixed doses (E(90)C(600) + 4 → T(75) + 4) (arm B) or docetaxel, doxorubicin and cyclophosphamide (T(75)A(50)C(500)) every three weeks for six cycles (arm C). All patients received G-CSF support and prophylactic ciprofloxacin. Results. One-hundred and twenty-four patients were randomised in the study. In the A, B and C arm, 17% 19% and 3% of the patients had one or more cycles delayed due to side-effects whereas 24%, 5% and 15% experienced a grade 3 infection or febrile neutropenia. After the introduction of an extra week between the EC and T parts in the A and B arms, grade 3 hand-foot-skin reactions were reduced from 5 to 0.2%. Twenty-nine percent (A and B) and 20% (C) of the patients were hospitalised due to side-effects. Discussion. Dose-dense and tailored EC/T can be given with manageable toxicity and is after adjustment presently studied in the phase III Panther trial.

The purpose of this study was to evaluate amplification of topoisomerase IIalpha (TOP2A) and HER2 genes as predictors of response to chemotherapy in advanced breast cancer. Gene copy number of TOP2A and HER2 were analysed with chromogenic in situ hybridization (CISH) on paraffin-embedded tissue sections from the primary tumour of 85 patients treated with anthracycline containing chemotherapy. TOP2A gene amplification was present in 14 (16%) and HER2 gene amplification in 38 (45%) of the primary tumours. Two of the 14 cases with TOP2A amplification were amplified without concurrent HER2 amplification. Neither TOP2A nor HER2 gene amplification were significantly associated with response to chemotherapy (p = 0.35 and p = 0.49, respectively).

An open, non-comparative, Nordic multicenter study was performed during 1991-1992 to evaluate the new 5-HT3 receptor antagonist tropisetron, as an antiemetic agent in various types of cancer chemotherapy. More than 600 patients were recruited from 16 cancer centers in Sweden, Finland and Denmark. In this report an interim analysis on 231 patients is presented. Gynecological cancers (61%), lung cancer (14%) and breast cancer (7%), were the main diagnoses. In 118 of 231 patients (51%) prior experience of chemotherapy was documented. In 91 patients (39%) cisplatin was part of the cytostatic regimen. Carboplatin (27%), doxorubicin (32%), epidoxorubicin (18%) were also frequently included. In all, 18 cytostatic agents were studied. The median number of courses studied was 3.3 (range 1-15). Overall 153 of 231 patients (67%) were completely protected from acute nausea and vomiting during the first course of chemotherapy. Delayed nausea and vomiting (Days 2-6) were completely controlled in 45%-72%. Treatment efficacy remained stable (57%-89%) over 10 consecutive courses of chemotherapy. For non-cisplatin regimens complete protection was achieved in 78% compared with 51% for cisplatin-regimens (p < 0.0001). Patients with no prior experience of chemotherapy had greater control of acute nausea and vomiting (73%) than patients treated before (61%) in the first course, but not in subsequent courses. There were no such differences in control of delayed nausea and vomiting between chemotherapy-naive and previously treated patients. Sex and age were significant prognostic factors with regard to antiemetic response. Adverse events were recorded in 19%-36% of the cases during long-term follow-up. Headache (16%) and constipation (5%) were most frequent.(ABSTRACT TRUNCATED AT 250 WORDS)

Tauromustine was administered orally in weekly doses with interindividual dose escalation to patients with disseminated malignant melanoma. The dose in the first cohort of 6 patients was 20 mg/m2/week. The dose escalation was 5 mg/m2/week. The limit of tolerance was 55 mg/m2/week. 99 patients completed at least 8 weeks of treatment and eight dose levels were evaluated for toxicity. Reversible thrombocytopenia, and to a lesser degree leukopenia, were dose limiting. From a starting dose of 40 mg/m2/week, the long-term tolerated dose was 35 mg/m2/week, thus achieving a considerable increase of dose intensity without a significant increase of toxicity by employing this weekly schedule of tauromustine.