Molecular Subtype Profile Reveals Therapy Predictive Power

BackgroundClassification of breast cancers into molecular subtypes may be important for accurate selection of therapy for patients. Here we report the respective chemotherapy responsiveness of the molecular subtypes profile defined Luminal, ERBB2 and Basal type.MethodsAn 80-gene subtype profile was developed based on a series of 200 samples with concordant ER, PR and Her2 receptor IHC and single gene readout status. Previously we reported the excellent validation of the profile classification using 784 samples. Here we report a second in silico validation consisting of 133 samples (Hess et al, JCO, 2006) which tested the profile as a predictor of pathological Complete Response (pCR) in these patients treated with T/FAC neoadjuvant chemotherapy.ResultsIn the 133 publicly available samples the profile classified 62% (82) as Luminal-type, 18% (24) as ERBB2-type and 20% (27) as Basal-type. These results were consistent with percentages found in the training and validation cohorts (n=1079; 295 training and 784 validation samples); 66% (712) Luminal-type, 18% (194) as ERBB2-type and 16% (173) Basal-type. Chemotherapy response was measured by pathological Complete Response (pCR) at the time of surgery In the Luminal-type subgroup 9% (7) of patients showed pCR, in the ERBB2-type subgroup 50% (12) of patients had a pCR and in the Basal-type subgroup 56% (15) of patients had a pCR.ConclusionsThe developed multi-gene profile can classify breast cancer tumors into Luminal-, ERBB2- and Basal-type subgroups. Within the subgroups, a significant difference in chemotherapy response, as measured by pCR, is observed. Implementation of this knowledge may improve the clinical management of breast cancer patients, by enabling the physician to decide who is most likely to benefit from chemotherapy or endocrine therapy prior to surgery.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 6131.

Basal-, Luminal-, and HER2- Molecular Subtype, and the MammaPrint 70-Gene Signature as Predictors of Response to Neoadjuvant Chemotherapy (NCT) with Docetaxel, Doxorubicin, Cyclophosphamide (TAC), or AC and Nab-Paclitaxel and Carboplatin +/- Trastuzumab in Patients (Pts) with Stage II-III and Inflammatory Breast Cancer (BC)

Background: Pathologic complete response (pCR) and minimal residual cancer burden (RCB scores of 0 [pCR]-1[near CR]) after NCT may predict for improved survival (Symmans et al. J Clin Oncol 25:4414-22, 2007). Hence, improved NCT regimens in conjunction with molecular markers that predict for both response and/or resistance are needed. Materials and Methods: 115 pts with stages II-III BC were to be prospectively randomized to receive 6 cycles of docetaxel 75 mg/m2, doxorubicin 50 mg/m2, cyclophosphamide 500 mg/m2 with filgrastim support (TAC, arm A) versus a novel regimen of A 60 mg/m2 and C 600 mg/m2 given every 2 weeks x 4, followed by 3 weekly doses of carboplatin (AUC 2) and nab-paclitaxel 100 mg/m2 repeated as 28 day cycles x 3 (arm B). Pts with HER2 + BC received NCT similar to arm B, but with the addition of 12 weekly doses of trastuzumab given together with carboplatin and nab-paclitaxel (arm C). Core biopsies were performed prior to NCT and were preserved fresh frozen. 70-gene (MammaPrint™) profiling and 80-gene profiling (van de Vijver et al. NEJM 347:1999-2009, 2002) to categorize all tumors for basal-, HER2-, and luminal subtypes were carried out. We set out to assess the predictive value of Mammaprint scores (poor vs. good), as well as basal, vs. luminal, vs. HER2 molecular subtype profiling, for response to treatment on arms A vs. B vs. C. Responses were dichotomized as complete or near complete response (Symmans RCB scores of 0-1) vs. suboptimal response (RCB score > 1). Results: Sufficient amount of BC tissue and good quality RNA for gene array assessment were procured in 64% of the first 90 patients who have undergone pre-treatment core biopsies, and then proceeded to NCT, followed by definitive surgery. Here we report on the first 50 pts with complete set of data analyzed. The median age was 50 years (range:31-69). Pts were treated for stage II (49%) and III locally advanced (41%), and inflammatory BC (10%). By gene profiling, 28% of the tumors were HER2-type (vs. 38% by IHC 3+, or FISH, representing all pts treated on arm C), 26% basal-type, 42% luminal-type, and 4% borderline luminal-type. Poor-prognosis signature by the 70-gene (MammaPrint) assay was observed in 74% of pts: 92% of HER2-type, 100% of basal-type, and 52% of luminal-type tumors were characterized as poor-risk by the 70-gene assay. Following NCT, Symmans RCB scores of 0-1 were observed in 71% of pts with HER2-type, in 38% with basal-type, and 28% of pts with luminal-type molecular subtype characteristics. Conclusion: BC with HER2- and basal-molecular subtypes are more likely to respond to NCT and is frequently associated with poor-risk characteristics as determined by the 70-gene assay. The complete analysis of correlations among response to specific sets of NCT, molecular subtype, and 70-gene assay results in the entire pt population will be presented.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 2026.

Risk of malignancies in patients with diabetes treated with human insulin or insulin analogues: a cohort study

AIMS/HYPOTHESIS

The aim of this cohort study was to investigate the risk of malignant neoplasms and mortality in patients with diabetes treated either with human insulin or with one of three insulin analogues.

METHODS

Data were provided by the largest German statutory health insurance fund (time-frame: January 1998 to June 2005 inclusive), on patients without known malignant disease who had received first-time therapy for diabetes mellitus exclusively with human insulin, aspart, lispro or glargine. The primary outcome was the diagnosis of a malignant neoplasm. Data were analysed by multiple Cox regression models adjusting for potential confounders.

RESULTS

A total of 127,031 patients were included, with a mean follow-up time of 1.63 (median 1.41, maximum 4.41) years. A positive association between cancer incidence and insulin dose was found for all insulin types. Because patients receiving combined therapy with insulin analogues and human insulin were excluded, the mean daily dose was much lower for glargine than for human insulin, and a slightly lower cancer incidence in the glargine group was found. After adjusting for dose, a dose-dependent increase in cancer risk was found for treatment with glargine compared with human insulin (p < 0.0001): the adjusted HR was 1.09 (95% CI 1.00 to 1.19) for a daily dose of 10 IU, 1.19 (95% CI 1.10 to 1.30) for a daily dose of 30 IU, and 1.31 (95% CI 1.20 to 1.42) for a daily dose of 50 IU. No increased risk was found for aspart (p = 0.30) or lispro (p = 0.96) compared with human insulin.

CONCLUSIONS/INTERPRETATION

Considering the overall relationship between insulin dose and cancer, and the lower dose with glargine, the cancer incidence with glargine was higher than expected compared with human insulin. Our results based on observational data support safety concerns surrounding the mitogenic properties of glargine in diabetic patients. Prospective long-term studies are needed to further evaluate the safety of insulin analogues, especially glargine.

Cost-effectiveness of targeting chemotherapy with the 70-gene prognostic signature in early-stage breast cancer (ESBC) patients

6570

Background: The 70-gene microarray test (MammaPrint) has been shown to provide additional prognostic information to clinicopathologic risk assessment for women ESBC; however, the cost-effectiveness of this strategy is not well understood. Methods: The objective of this analysis was to estimate the incremental benefits, costs, and cost-effectiveness of the treatments guided by the 70-gene signature versus Adjuvant! Software (AS) to decide on the use of adjuvant chemotherapy for women ≤61 years with lymph node negative, HER-2 negative ESBC with estrogen receptor (ER) positive or negative disease. A Markov model with a lifetime horizon and three health states (alive without recurrence, death from cancer and death from other causes) was constructed using TreeAge Pro software. Risk classification and patient outcomes data were based on a multi-center 70-gene signature validation study. Efficacy of chemotherapy derived from published meta-analysis of clinical trials. Costs and health utilities were obtained from the literature. Costs and benefits were discounted 3%/year. Results: Compared to AS, the 70-gene signature strategy resulted in 35% of patients being reassigned to a different risk classification and avoided chemotherapy in 9% of patients. In the base case, the 70-gene signature strategy was cost neutral (lifetime costs per patient: $178,811 versus $178,893 for the 70-gene signature and AS strategy). Moreover the 70-gene signature strategy was associated with an increase of 0.13 life years (LYs) and 0.16 quality adjusted life years (QALYs). The model results were sensitive to the cost of 70-gene signature test, cost of adjuvant chemotherapy, and relative risk reduction associated with chemotherapy; however, the 70-gene strategy remained cost-effective across a wide range of assumptions. Conclusions: In this analysis, the 70-gene signature was associated with a reduction in chemotherapy use and an increase in life expectancy. The 70-gene signature appears to be a cost-effective strategy for obtaining additional information to guide the decision to use adjuvant chemotherapy in patients with lymph node negative ESBC.

[Table: see text]

Microarray-based determination of ER, PR and HER2 receptor status: validation and comparison with IHC assessments

Abstract #3007

Background&#x2028; In breast cancer patients the level of expression of estrogen receptor (ER), progesterone receptor (PR) and HER2 is predictive for prognosis and/or treatment response. However, differences in assessment methods and interpretation can substantially affect the accuracy and reproducibility of the results. Previously, we have determined the association between immunohistochemistry (IHC) and mRNA levels for ER, PR and HER2, and have confirmed the accuracy of microarray readout on &gt;400 samples. In the current study we describe the use of this microarray based readout on prospectively collected samples. We compared these readouts with multiple IHC and fluorescent in situ hybridization (FISH) assessments generated in various hospitals and a CLIA-certified reference laboratory and developed a microarray based test called TargetPrint™.&#x2028; Methods&#x2028; Gene expression data for ER, PR and HER2 were obtained by analysis of 100 breast carcinomas that have been collected prospectively within the RASTER study. Samples were stratified as receptor positive or negative using thresholds for ER, PR and HER2 mRNA levels. IHC assessment was performed (1) according to local standards of the hospital from where the sample originated, (2) by the central laboratory of the Netherlands Cancer Institute, and (3) at an independent reference laboratory using FDA-approved procedures and ASCO/CAP guidelines. A tumor was classified positive for ER and PR when ≥10% of tumor cells showed positive staining. HER2 IHC status was scored as 0, 1+, 2+ or 3+; a score of 3+ was considered positive. In case of 2+ samples, a FISH was performed to assess final HER2 amplification status. The cohort used in this study was pre-selected to include about two-third ER and PR positive samples and one-third HER2 positive samples.&#x2028; Results&#x2028; Multiple microarray readouts were highly reproducible (Pearson correlation 0.991) and resulted in 67, 61 and 39 percent positive samples for ER, PR and HER2, respectively. Comparison of microarray results with IHC (including FISH for HER2) performed at the three centers indicated highly similar results for receptor readout with a concordance of 92, 93 and 92% for ER; 84, 81 and 86% for PR; and 93, 95 and 94% for HER2. Overall misclassification rates between microarray and IHC readout were low for ER (0.08) and HER2 (0.06) and quite low for PR (0.14), and were comparable to the misclassification rates between the three IHC methods.&#x2028; Conclusion&#x2028; A microarray-based assessment of ER, PR and HER2 in relation to mRNA levels gives results comparable to multiple IHC methods and FISH and provides an objective and more quantitative assessment of tumor receptor status than IHC alone. Using TargetPrint™ for microarray readouts for hormone and HER2 receptor in addition to standard IHC will improve molecular characterization of breast cancer tissue.

Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 3007.

Können Mindestmengen für Knieprothesen anhand von Routinedaten errechnet werden?

AIM OF THE STUDY

The aim of this study was to calculate, for the first time, minimum provider volumes in total knee replacement using routine German data.

MATERIALS AND METHODS

In patients with primary total knee replacement (TKR), the relationship between hospital volume per year and risk of "insufficient mobility" (primary quality indicator) and "wound infection" (secondary quality indicator) was calculated by means of logistic regression models.

RESULTS

For both indicators, a statistically significant relationship between hospital volume and outcome could be demonstrated. Other risk factors such as age and ASA status also had a significant influence, but did not appear as important confounders. The risk for the secondary quality indicator "infection" decreased constantly with increasing hospital volume, thus the curve was very flat. This supports the hypothesis that high volume hospitals have a higher quality level than low volume hospitals. A threshold value could be calculated. However, the explanation value for hospital volume was too low to derive a threshold level that clearly discriminates between good and bad quality of care. The relationship between the primary quality indicator "insufficient mobility" and hospital volume unexpectedly showed a U-shaped distribution. This questions the concept of a minimum provider volume regulation for primary total knee replacement for the quality indicator "insufficient mobility". Therefore, in this case no quantitative threshold values were calculated.

CONCLUSION

This analysis supports the hypothesis of a volume-outcome relationship in primary total knee replacement. However, a minimum provider volume that clearly discriminates between good and bad quality of care could not be calculated on the basis of these German quality assurance data.

Aspekte der Fallzahlkalkulation und Powerberechnung anhand von Beispielen aus der rehabilitationswissenschaftlichen Forschung

Often it is reported in medical studies that an expected effect could not be detected. This may be the case if the sample size had been too small to detect an effect which actually exists. This often is due to the fact that sound sample size estimation had been omitted prior to the study outset. As a result, it is not known how many persons should have been involved in the study to detect this effect if present. On the other hand, if sample size estimation has not been realized, more persons than needed might be included in the study. This is problematic for economic and in particular for ethical reasons. The aim of this paper is to point out the principles of sample size estimation as well as to emphasize its importance not only in general but also in medical rehabilitation research.

Identification of cancer of unknown primary with gene expression profiling

10052

Background: Patients with metastatic cancer of unknown primary (CUP) generally have a poor prognosis, with a median survival of 2–10 months. Conventional diagnostic approaches for identifying the primary tumor site are successful in only 20%-30% of cases; however, such identification provides prognostic information and helps with selection of tumor-specific therapy, leading to improved survival. Recent studies indicate that gene expression-based classification of CUP is highly successful in predicting the site of origin. We report herein development and validation of a method that determines the site of tumor origin by comparing the gene expression profiles of CUP cases to those in a database created from known tumor types. Methods: RNA extracted from frozen and formalin-fixed, paraffin-embedded (FFPE) tissue wasis purified and amplified using the Paradise Reagent System System (Arcturus, Mountain View, CA). Following reverse-transcription, cDNA products wereare used in a semi-quantitative real-time PCR to detect 87 tumor-associated genes and 5 reference genes in an ABI PRISM 7900HT Detection System (Applied Biosystems, Foster City, CA). Gene expression data wereare then compared to those in a database, composed of gene expression profiles of 571 samples from 39 different tumor types, using k-nearest neighbor analysis to predict the most likely site of tumor origin. Intra- and interassay reproducibility was determined. Frozen and FFPE tissues (n=57) from a well-characterized, independent sample set were also tested in a blinded manner to further validate the method. Results: Based on the real-time PCR cycle threshold, the intra- and interassay reproducibility ranged from 0.1%-4.3% and 0.5%-8.2%, respectively. The primary tumor type was identified in 77% of cases. The assay determined the correct tumor type in 88% (44/50) of the samples. Seven samples were not reported: 3 failed to amplify adequately and 4 had an unacceptably low confidence level. Conclusions: We have shown that gene expression profiling can determine the most likely site of tumor origin. Our data suggest that this new method is able tomay identify the primary site of tumor origin in 77% of CUP cases.

No significant financial relationships to disclose.

Factors of non-compliance with the therapeutic regimen among hypertensive men and women: a case-control study to investigate risk factors of stroke

To identify potential risk factors among the therapeutic regimen and life style which may increase the risk for stroke, a pair matched case-control study was conducted in Gaza Strip among 112 patients, who had been hospitalized for acute stroke and history of hypertension, and 224 controls with history of hypertension. Conditional logistic regression models show significant associations between stroke and medication not taking as prescribed (OR = 6.07; 95% CI: 1.53, 24.07), using excessive salt at meals (OR = 4.51; 95% CI: 2.05, 9.90), eating diet high in fat (OR = 4.67; 95% CI: 2.09, 10.40), and high level of stress (OR = 2.77; 95% CI: 1.43, 5.38). No significant association between smoking and the development of stroke (OR = 2.12; 95% CI: 0.82, 5.51) was found. Regular physical exercise was a protective factor (OR=0.26; 95% CI: 0.12, 0.57). Using excessive salt at meals was a significant risk factor (OR = 16.61; 95% CI: 4.40, 62.80) in people having low level of stress, whereas it was not significant in people having high level of stress. (OR = 1.76; 95% CI: 0.58, 5.33). Smoking in combination with low level of stress was a significant risk factor for stroke (OR = 9.88; 95% CI: 2.52, 38.78), but a non-significant protective factor in combination with high level of stress (OR=0.52; 95% CI: 0.14, 1.99). An increase in compliance with the pharmacological and non-pharmacological therapeutic regimen might be a key to a reduction of stroke incidence and prevalence among hypertensive patients.

Glycaemic control and severe hypoglycaemia following training in flexible, intensive insulin therapy to enable dietary freedom in people with type 1 diabetes: a prospective implementation study

AIMS/HYPOTHESIS

The aim of this study was to evaluate the implementation of a course teaching flexible, intensive insulin therapy on glycaemic control and severe hypoglycaemia in routine care.

METHODS

This is a continuous quality-assurance project involving hospital diabetes centres. Every third year each centre re-examines 50 consecutive patients (evaluation sample) 1 year after participation in the course. Ninety-six diabetes centres in Germany participated and 9,583 patients with type 1 diabetes (190 evaluation samples) were re-examined between 1992 and 2004. The intervention was a 5-day inpatient course for groups of up to ten patients with a fixed curriculum of education and training for dietary flexibility and insulin adjustment. The main outcome measures were HbA1c and severe hypoglycaemia.

RESULTS

Mean baseline HbA1c was 8.1%, and had decreased to 7.3% at follow-up; incidence of severe hypoglycaemia was 0.37 events per patient per year prior to intervention and 0.14 after intervention. In mixed-effects models adjusted for effects of centres, age and diabetes duration, the mean difference was -0.7% (95% CI -0.9 to -0.6%, p<0.0001) for HbA1c and -0.21 events per patient per year (95% CI -0.32 to -0.11, p=0.0001) for severe hypoglycaemia, with similar results for evaluation samples, with a maximum of 10% of patients lost to follow-up. Before intervention, the incidence of severe hypoglycaemia was three-fold higher in the lowest quartile than in the highest quartile of HbA1c, whereas the risk was comparable across the range of HbA1c values after intervention.

CONCLUSIONS/INTERPRETATION

Implemented as part of a continuous quality-assurance programme the self-management programme is effective and safe in routine care. Improvement of glycaemic control can be achieved without increasing the risk of severe hypoglycaemia.

Effect on hip fractures of increased use of hip protectors in nursing homes: cluster randomised controlled trial

OBJECTIVE

To assess the effects of an intervention programme designed to increase use of hip protectors in elderly people in nursing homes.

DESIGN

Cluster randomised controlled trial with 18 months of follow up.

SETTING

Nursing homes in Hamburg (25 clusters in intervention group; 24 in control group).

PARTICIPANTS

Residents with a high risk of falling (459 in intervention group; 483 in control group).

INTERVENTION

Single education session for nursing staff, who then educated residents; provision of three hip protectors per resident in intervention group. Usual care optimised by brief information to nursing staff about hip protectors and provision of two hip protectors per cluster for demonstration purposes.

MAIN OUTCOME MEASURE

Incidence of hip fractures.

RESULTS

Mean follow up was 15 months for the intervention group and 14 months for the control group. In total 167 residents in the intervention group and 207 in the control group died or moved away. There were 21 hip fractures in 21 (4.6%) residents in the intervention group and 42 hip fractures in 39 (8.1%) residents in the control group (relative risk 0.57, absolute risk difference -3.5%, 95% confidence interval -7.3% to 0.3%, P=0.072). After adjustment for the cluster randomisation the proportions of fallers who used a hip protector were 68% and 15% respectively (mean difference 53%, 38% to 67%, P=0.0001). There were 39 other fractures in the intervention group and 38 in the control group.

CONCLUSION

The introduction of a structured education programme and the provision of free hip protectors in nursing homes increases the use of protectors and may reduce the number of hip fractures.