Assessment of biomarker expression in predicting pathologic response to neoadjuvant chemotherapy in patients with locally advanced breast cancer

Issues in collecting, processing and storing human tissues and associated information to support biomedical research

The availability of human tissues to support biomedical research is critical to advance translational research focused on identifying and characterizing approaches to individualized (personalized) medical care. Providing such tissues relies on three acceptable models - a tissue banking model, a prospective collection model and a combination of these two models. An unacceptable model is the "catch as catch can" model in which tissues are collected, processed and stored without goals or a plan or without standard operating procedures, i.e., portions of tissues are collected as available and processed and stored when time permits. In the tissue banking model, aliquots of tissues are collected according to SOPs. Usually specific sizes and types of tissues are collected and processed (e.g., 0.1 gm of breast cancer frozen in OCT). Using the banking model, tissues may be collected that may not be used and/or do not meet specific needs of investigators; however, at the time of an investigator request, tissues are readily available as is clinical information including clinical outcomes. In the model of prospective collection, tissues are collected based upon investigator requests including specific requirements of investigators. For example, the investigator may request that two 0.15 gm matching aliquots of breast cancer be minced while fresh, put in RPMI media with and without fetal calf serum, cooled to 4°C and shipped to the investigator on wet ice. Thus, the tissues collected prospectively meet investigator needs, all collected specimens are utilized and storage of specimens is minimized; however, investigators must wait until specimens are collected, and if needed, for clinical outcome. The operation of any tissue repository requires well trained and dedicated personnel. A quality assurance program is required which provides quality control information on the diagnosis of a specimen that is matched specifically to the specimen provided to an investigator instead of an overall diagnosis of the specimen via a surgical pathology report. This is necessary because a specific specimen may not match the diagnosis of the case due to many factors such as necrosis, unsuspected tumor invasion of apparently normal tissue, and areas of fibrosis which are mistaken grossly for tumor. Aliquots for quality control (QC) may or may not be collected at the time of collection and in some cases, QC may not occur until specimens are distributed to investigators. In establishing a tumor repository, multiple issues need to be considered. These include the available resources, long term support, space and equipment. The needs of the potential users need to be identified as to the types of tissues and services needed and the annotation expected. Other specific issues to be considered include collection of specimens potentially infected with blood borne pathogens (e.g., hepatitis B), charge back mechanisms, informatics needs and support, and investigator requirements (e.g., recognition of repository contributions in publications). In general, the repository should not perform the research of the investigators, but should provide the infrastructure necessary to support the research of the investigator. Thus, the goals of the repository must be established. Similarly, ethical and regulatory issues must be evaluated. In general, tissue repositories need ethical (e.g., IRB) and privacy (e.g., HIPAA) review. Also, safety issues need to be considered as well as how biohazards will be addressed by investigator-users. Considerations involving the transfer of specimens to other organization usually require a material transfer agreement (MTA). A MTA should address biohazards as well as indemnification. Thus, many issues must be considered and addressed in order to establish and operate successfully a biorepository.

The potential biomarkers in predicting pathologic response of breast cancer to three different chemotherapy regimens: a case control study

Background

Preoperative chemotherapy (PCT) has become the standard of care in locally advanced breast cancer. The identification of patient-specific tumor characteristics that can improve the ability to predict response to therapy would help optimize treatment, improve treatment outcomes, and avoid unnecessary exposure to potential toxicities. This study is to determine whether selected biomarkers could predict pathologic response (PR) of breast tumors to three different PCT regimens, and to identify a subset of patients who would benefit from a given type of treatment.

Methods

118 patients with primary breast tumor were identified and three PCT regimens including DEC (docetaxel+epirubicin+cyclophosphamide), VFC (vinorelbine/vincristine+5-fluorouracil+cyclophosphamide) and EFC (epirubicin+5-fluorouracil+cyclophosphamide) were investigated. Expression of steroid receptors, HER2, P-gp, MRP, GST-pi and Topo-II was evaluated by immunohistochemical scoring on tumor tissues obtained before and after PCT. The PR of breast carcinoma was graded according to Sataloff's classification. Chi square test, logistic regression and Cochran-Mantel-Haenszel assay were performed to determine the association between biomarkers and PR, as well as the effectiveness of each regimen on induction of PR.

Results

There was a clear-cut correlation between the expression of ER and decreased PR to PCT in all three different regimens (p < 0.05). HER2 expression is significantly associated with increased PR in DEC regimen (p < 0.05), but not predictive for PR in EFC and VFC groups. No significant correlation was found between biomarkers PgR, Topo-II, P-gp, MRP or GST-pi and PR to any tested PCT regimen. After adjusted by a stratification variable of ER or HER2, DEC regimen was more effective in inducing PR in comparison with VFC and EFC regimens.

Conclusion

ER is an independent predictive factor for PR to PCT regimens including DEC, VFC and EFC in primary breast tumors, while HER2 is only predictive for DEC regimen. Expression of PgR, Topo-II, P-gp, MRP and GST-pi are not predictive for PR to any PCT regimens investigated. Results obtained in this clinical study may be helpful for the selection of appropriate treatments for breast cancer patients.

Ki-67 as a prognostic molecular marker in routine clinical use in breast cancer patients

INTRODUCTION

The proliferation biomarker Ki-67 is a prognostic factor for breast cancer that has been investigated in several retrospective studies and a few prospective ones. The aims of the present study were to examine interactions between Ki-67 and other biomarkers in breast cancer patients and to assess the relationship of Ki-67 to histological grading.

PATIENTS AND METHODS

Patients with uniform immunohistochemical staining of Ki-67 by MIB-1 were identified from the database of the University Breast Center for Franconia. Data were available for 1232 of 2523 patients with invasive breast cancer who had been treated between 1998 and 2005. Ki-67 index was determined during routine work-up of the breast cancers by several surgical pathologists according to a standardized procedure. The Ki-67 proliferation index was correlated with hormone receptor status, HER2/neu status, age, tumor staging, and prognosis. In routine clinical practice, the grading was assessed according to Elston and Ellis, along with all other parameters.

RESULTS

Ki-67 proliferation index>or=20% was found to be associated with all of the prognostic factors that were tested. However, it also maintained statistical significance relative to poor overall survival in a multivariate Cox proportional hazards model (hazards ratio 1.81; 95% CI, 1.17-2.78). The hazards ratio for disease-free survival did not reach statistical significance (HR 1.41; 95% CI, 0.95-2.09; P=0.086). However, in both models the grade was not an independent prognostic factor.

CONCLUSIONS

For routine clinical purposes, grading appears to add only limited information about the prognosis in comparison with Ki-67 expression. Further investigation of quality assurance for grading and of Ki-67 as a prognostic and predictive factor is warranted.