The Norwegian University of Science and Technology discusses biomarker signatures that predict long-term survival for breast cancer patients.

The heterogeneous nature of breast cancer is reflected in its wide prognostic variation. On the one extreme, highly malignant tumours require aggressive treatment, whereas on the other extreme, some tumours may not have the malignant potential to influence the patient’s life expectancy, even if left untreated. Between these extremes, there are moderately malignant tumours that will recur many years after treatment is completed.

For most tumours, their aggressiveness cannot be reliably determined at diagnosis, and it is not yet realistic to streamline treatment that is targeted to the malignant potential of the disease. Therefore, most patients typically receive chemotherapy and adjuvant treatment in addition to surgery and radiotherapy. Whereas many patients will benefit from the treatment, its effects may be mostly harmful for others, both in the short and long term.

In an attempt to improve the situation, we are currently in the process of identifying biomarker signatures in tumour tissue that can reliably predict which tumours are likely to have an excellent prognosis, and to distinguish these tumours from those that may eventually recur. The ultimate purpose is to develop personalised treatment where the tumour’s underlying biology is the critical target and simultaneously prevent unnecessary and harmful treatment of tumours with low malignant potential.

Molecular subtypes of breast cancer

Using stored diagnostic tumour material, we have constructed tissue microarrays (TMAs) from formalin fixed, paraffin-embedded breast cancer tissue. Using immunohistochemical and in situ hybridisation analyses, molecular markers (abbreviated as ER, PR, HER2, Ki67, CK5 and EGFR) are used to reclassify tumours into molecular subtypes (in Fig. 1). We have used these methods as surrogates for gene expression analyses and determined six breast cancer subtypes, entitled Luminal A, Luminal B (HER2 negative), Luminal B (HER2 positive), HER2 subtype, Basal-like phenotype, and Five-negative phenotype breast cancers.

Fig. 2 shows patient survival according to molecular subtype in 909 reclassified breast cancer patients. The Kaplan-Meier plots show that five-year survival was best for the Luminal A subtype, followed by the Luminal B (HER2 negative) subtype. After five years of follow-up the survival curves conversed, and in the long-term survival differences between subtypes were less apparent, suggesting that a certain proportion of patients within each subtype category may have an excellent prognosis. This begs the question of whether the long-term prognosis of all tumours is determined by some common underlying features, or whether uniquely different factors that are specific for each subtype determine which tumours have an excellent prognosis.

Therefore, we want to characterise tumours beyond the molecular markers that define each subtype, where the aim is to identify molecular biomarkers that are predictive for the long-term survival of breast cancer. The ultimate aim is to develop molecular methods that can reliably predict which tumours are not likely to be clinically significant, and to distinguish these tumours from tumours that are likely to recur, sometimes many years after the initial treatment.

Biomarkers that predict long-term prognosis

To characterise tumours in more detail, we have combined the biomarkers that are used to define each subtype and other biomarkers that are associated with tumour growth, proliferation and metastasis. These biomarkers include indicators of angiogenesis, tumour infiltration, proliferation markers, hormone and growth factor receptors, and markers of apoptosis regulation and mitosis.

We expect that different combinations of these biomarkers may indicate distinct differences in patient survival and provide reliable information related to the prediction of long-term survival. The testing of prediction models requires long-term follow-up of a large group of breast cancer patients, and it would be an advantage to study patients who have not received modern oncological treatment. We have therefore included a large number of patients who were diagnosed prior to the introduction of modern therapy (before 1985), by using stored diagnostic tissue from the 1970s and early 1980s. Another advantage for the analysis is that most of these patients can be followed from diagnosis until death.

In total, stored tumour tissue from 2,327 patients is being used for tissue microarray (TMA) construction and determination of the six molecular subtypes of breast cancer described previously. The two specific aims of the project are to: 1. Identify biomarker signatures in breast cancer tissue that predict long-term survival in approximately 2,300 breast cancer patients who have been followed from diagnosis and, for a majority, until death. Hypothesis: One or more biomarker signatures will reliably identify patients whose tumours have limited malignant potential, as reflected in long-term survival without recurrence. 2. Identify biomarker signatures in breast cancer tissue that predict recurrence of breast cancer up to many years after completed treatment, and distinguish these tumours from those that do not recur. Hypothesis: One or more biomarker signatures will reliably identify patients with long-term survival but whose disease is likely to recur many years after completed treatment. These biomarker signatures differ from those that can identify patients whose disease does not recur. The recording of cancer is mandatory in Norway, and the reporting of breast cancer, including breast cancer mortality, is virtually complete and highly reliable. The underlying data infrastructure and data quality, including archival diagnostic breast tissue material, provide a powerful and extremely cost-effective approach in this project.

The patients of this project are derived from three separate populations of women who have been followed up for breast cancer through the Norwegian Cancer Registry. The first group includes 909 patients diagnosed among 25,897 women born between 1886 and 1928, who participated in a breast screening study in the late 1950s in Norway. They lived through a time period with no exposure to exogenous hormones (no oral contraceptives and no hormone treatment around menopause), and they never participated in organised mammography screening. Those who developed breast cancer were diagnosed before modern oncological treatment was introduced, and treatment was mostly restricted to surgery (mastectomy). Some patients were diagnosed after modern treatment was introduced (after 1985), but were too old to receive the new treatment.

The second group consists of 870 patients diagnosed among 22,931 women born in Trondheim, Norway, between 1920 and 1966. The third group includes 556 patients who were diagnosed during the follow-up of 34,500 women who participated in the second wave of the HUNT Study in Norway. The baseline collection of data in that study took place between 1995 and 1997.

Future directions

Our overriding goal is to identify one or more sets of biomarker signatures that may predict which breast cancer tumours have very limited malignant potential. Therefore, our findings may be important for the prevention of unnecessary treatment and for the development of appropriate clinical care for these patients. Thus, apart from minimal surgery (lumpectomy) and careful long-term surveillance, these patients may benefit from alternative, as yet undeveloped, interventions.

We also want to identify biomarker signatures that can reliably predict which tumours are likely to recur many years after primary treatment. By anticipating recurrence of these cancers, new management strategies may be developed that have the potential to improve clinical care for patients with recurring disease.

Thus, this research has the potential to influence clinical practice and to improve the differentiation of breast cancer treatment. In particular, this work may contribute to preventing unnecessary treatment and to reducing the harmful side effects of modern interventions.

Lars Vatten, MD, PhD, MPH
Professor in Epidemiology
Department of Public Health, Faculty of Medicine
The Norwegian University of Science and Technology
+47 73598787
Lars.Vatten@ntnu.no
http://www.ntnu.no/

The post Heterogeneity of breast cancer appeared first on Horizon 2020 Projects.