Limitations of “Classic” Risk Factors to Predict Neoadjuvant Chemotherapy Response for Operable Breast Cancer
- First Online:
- Cite this article as:
- Zografos, G.C. & Ziogas, D. World J Surg (2009) 33: 1545. doi:10.1007/s00268-009-9933-5
- 35 Views
To the Editor,
Preoperative or neoadjuvant chemotherapy (NC) has been the standard of clinical practice in locally advanced breast cancer. This approach offers the potential of tumor downstaging, allowing breast-conserving surgery (BCS) for women with large tumors at diagnosis. It is also used more and more frequently for patients with early-stage breast cancer. Despite several advantages, however, NC also has limitations, including incomplete surgical resection in BCS after NC and a delay in definitive surgical treatment that may harm patients whose disease does not respond to the treatment.
The study reported by Shien et al.  in a recent issue of the World Journal of Surgery provides data supporting the limitations of current clinicopathologic factors in predicting pathological complete response (pCR) to NC. We congratulate the authors for the accuracy and objective interpretation of their data. We wish to comment on this study precisely because the information provided may influence surgeons in their clinical practice. We also wish to increase awareness of the potential ability of genetics, gene expression-profiling, and genome-wide association studies to improve pCR prediction when neoadjuvant chemotherapy and targeted therapy are applied. Predicting pCR may help breast surgeons with the selection of patients for neoadjuvant treatment.
Breast-conserving surgery is the preferred approach to the treatment of cancer when certain clinicopathologic criteria have been established: chiefly, microscopically negative surgical margins and absence of multicentric disease. Patients with familial susceptibility to breast cancer and BRCA1 or BRCA2 positive testing may gain greater benefit from an aggressive surgery like bilateral mastectomy, rather than BCS .
Shien and colleagues  report on 368 patients who underwent curative surgical treatment after NC (anthracycline and/or taxane without trastuzumab). The pCR rate in their series was only 17%. In multivariate analysis, a scirrhous-type tumor was an independent predictor of reduced pCR rate. Beyond that finding, this small retrospective study has two further limitations. First, the patients in this study were not treated according to recent guidelines with the standard agent trastuzumab for patients with human epidermal growth factor receptor 2 (HER2)-positive disease. By contrast, the data from this study may be informative in HER2-negative disease. Even in these cases, however, confirmatory studies are required to exclude patients with scirrhous cancer from NC. Second, the authors conclude that computerized tomographic (CT) imaging and histological grading are effective in predicting tumor reduction in response to NC. However, there is international consensus that conventional clinicopathologic factors and “classic” imaging have strong limitations in discriminating “true” responders, who will benefit from neoadjuvant treatment, from nonresponders, who will benefit from primary surgery immediately.
Current research is focusing on the development of genetic and molecular markers to predict response to treatment. The aim of such research is to establish a patient-tailored therapeutic approach. Rapid progress in genetics, and in the clinical use of genetic testing in the pretreatment diagnostic work-up, can allow identification of patients with an increased risk of local failures after BCS. Patients with heritable mutations, particularly in the BRCA1 or BRCA2 gene , and perhaps in the CDH1 gene as well [3– 5], face a very high risk of developing hereditary breast–ovarian, and diffuse gastric-lobular breast cancer, respectively. These patients are likely at high risk of developing ipsilateral breast cancer and contralateral breast cancer . Therefore, they may benefit more from a primary aggressive surgery like bilateral mastectomy than from neoadjuvant treatment followed by BCS. More data on this topic are needed from new studies to guide treatment of patients with familial breast cancer.
The recent technological revolution, characterized by microarrays and next-generation sequencing techniques, has already provided exciting results leading to personalized genomics-based therapeutic approaches in the treatment of various cancers. New models and algorithmic approaches have been developed and proposed for solid cancers, including breast cancer [7, 8]. It is clear, however, that more extensive analyses and new, more appropriate, and unbiased studies are required to allow better understanding of gene–gene and gene–environment interactions. Despite the multiple challenges that remain, a rational avenue has opened: Gene expression profiling studies, genome-wide association studies, and computational biology promise to reach the goal of personalized medicine and oncology [9–11].