Breast and Lung Cancer Biomarker Research at ASCO:
Changing Treatment Patterns

By Julie Katz, MPH, MPhil

Biomarkers played a prominent role in the research presented in a number of tumor types at the 2013 American Society of Clinical Oncology (ASCO) annual meeting. New data presented at ASCO 2013 showed novel clinical implications of biomarkers, including estrogen receptor (ER), progesterone receptor (PR), and HER2 in breast cancer, and EGFR and EML4-ALK in lung cancer. In our post-ASCO discussion, we illustrate the influence of biomarkers on testing and treatment, of genetic changes throughout the course of disease, acquired resistance and treatment patterns, and biopsy practices and plasma DNA testing methods. All of these issues have potentially profound implications on a tumor’s behavior and sensitivity throughout the course of the disease. The biomarker research presented represents different approaches to cancer treatment.

The Biomarker France Study

One of the most exciting and promising presentations on biomarkers was the Biomarkers France study (Barlesi F. J Clin Oncol. 2013;31 [suppl]; abstr 8000)—the largest biomolecular study ever conducted. Investigators analyzed the mutation status of 10,000 advanced non-small cell lung cancer (NSCLC) patients and provided molecular profiling of EGFR, HER2, KRAS, BRAF and PIK3CA mutations as well as EML4-ALK translocation. This study demonstrated the feasibility of NSCLC tumor profiling and indicated the feasibility of a similar profiling study for other tumors. Moreover, tumor profiling in this study identified a known therapeutic target in 46% of the samples. Tumor profiling helped manage NSCLC patients in 57% of the cases and in 70% of the EGFR-mutant patients. Upon study completion, 19,000 patients will be included in the final analysis. The authors conclude that the study provides solid data on the value of a nationwide biomarker screening policy for NSCLC patients.

Breast Cancer Biomarkers

Three abstracts at ASCO reported investigations of new clinical uses of estrogen-receptor (ER) status, progesterone-receptor (PR) status and HER2 amplification status. ER, PR and HER2 were the first three predictive biomarkers identified historically, representing three of 11 unique biomarkers in oncology that are currently recommended to guide choice of drug regimen.  Beyond the predictive and prognostic role of ER, PR and HER2 in breast cancer, these studies demonstrate important considerations about how biomarkers may change throughout the course of disease and through different disease states, and the impact this may have in guiding treatment decisions and informing prognosis. 

Concordance between the Primary Tumor and Recurrence

A Japanese study of 117 consecutive cases investigated changes in biomarkers ER, PR, HER2 and Ki-67 between the primary tumor and a recurrence after breast-conserving surgery (Okumura Y et al. J Clin Oncol. 2013;31 [suppl]; abstr 1116). Patients in this study with an ipsilateral breast cancer tumor recurrence (IBTR) without distant metastases underwent surgery for IBTR between 1989 and 2008. The percentage of cases with concordance in breast cancer subtype (luminal A, luminal B/HER2-negative, luminal B/HER2-positive, HER2, triple negative) between primary and recurrent lesions was 62.1%. The PR-positive rate decreased significantly (p=0.01) and the mean Ki-67 index increased significantly (p=0.047) from the primary tumor to the recurrence. Discordance in the Ki-67 index between the primary tumor and IBTR was associated with a lower distant disease-free survival (DDFS), and this association was statistically significant among patients with a recurrence in the same quadrant of the ipsilateral breast.

A Korean study also examined biomarker agreement, comparing the primary breast tumor and a metastatic recurrence (Shin HC et al. J Clin Oncol. 2013;31 [suppl]; abstr 1039)). This study of 193 patients looked at the triple receptor status (ER, PR and HER2) and divided the patients into three groups: concordant non-triple-negative (TN), concordant TN, and a discordant group (Figure 1).

Source: Shin HC et al. J Clin Oncol. 2013;31 [suppl]; abstr 1039.

The study compared long-term outcomes, systemic recurrence-free survival (SRFS), overall survival (OS) from primary and OS from metastasis. The highest OS rate from primary was among the concordant non-TN group, followed by the discordant group; the concordant TN group had the lowest OS rate from primary. Similarly, the OS rate from metastasis was highest among the concordant non-TN group. However, the OS rate from metastasis was not significantly different between the discordant group and the concordant TN group. The relative risk of 2.5 (95% CL: 1.2 to 5.3) among the discordant group showed it to be an independent prognostic factor for death after metastasis.

These studies raise important questions about biomarker testing during the course of disease and treatment. Biopsy and biomarker testing at recurrence may be necessary to detect a change in the biomarker status of the recurrence compared to the primary tumor and could influence patient prognosis and treatment decisions. These data provide further support for the National Comprehensive Cancer Network (NCCN) recommendation to retest patients who were negative or unknown at initial biomarker testing (http://www.nccn.org/professionals/physician_gls/f_guidelines.asp#breast, Accessed June 25, 2013).

Breast and Lung Cancer Biomarker Research at ASCO:
Changing Treatment Patterns (cont.)

Genetic Heterogeneity

In addition to the breast cancer studies that examine the concordance and discordance between primary breast tumors and recurrence and survival, an abstract from the National University Cancer Institute of Singapore examined the prevalence of genetic heterogeneity of HER2, concordance or discordance within the primary tumor, and its impact on survival (Ho J et al. J Clin Oncol. 2013;31 [suppl]; abstr 586). The 158 primary breast tumor specimens in this study had similar distribution of HER2 status to SEER data.  

This study used immunohistochemistry (IHC) and the HER2:CEP17 ratio to classify tumors as HER2 genetically heterogeneous, according to the ASCO-CAP (College of American Pathologists) guidelines. The definition of genetic heterogeneity was 5% to 50% of tumor cells with a ratio of more than 2.2. 

Among the HER2-negative cases (n=106), 43.4% were genetically heterogeneous for HER2. All of the HER2 equivocal cases (FISH ratio 1.8-2.2) were genetically heterogeneous for HER2. In total, 36.1% of all 158 cases exhibited genetic heterogeneity for HER2 (Figure 2). HER2 genetic heterogeneity was associated with significantly higher stage (p=0.028), poorer overall survival (p=0.034) and higher grade, although this association was not statistically significant. 

Source: Ho J et al. J Clin Oncol. 2013;31 [suppl]; abstr 586.

This study calls for further prospective studies on HER2 genetic heterogeneity. It raises the importance of testing HER2-negative breast tumors for genetic heterogeneity because of the potential impact on treatment decisions. It also shows the importance of further investigation into the possible benefit of anti-HER2 therapy in HER2-negative breast cancers that have genetic heterogeneity.

Acquired Resistance in NSCLC

Acquired resistance in NSCLC provides an analogous example of the utility of predictive biomarkers and targeted therapies in oncology. Investigating acquired resistance raises the issues of testing, genetic heterogeneity and treatment decisions based on mutation status throughout the course of disease. In the presentation “The Clinical Perspective: Defining and Overcoming Barriers” in the session, “Beyond Progression: Treating EGFR or ALK-Positive Non-Small Cell Lung Cancer (NSCLC) after First-Line Therapy,” Dr. Howard J. West discussed acquired resistance to targeted lung cancer therapies and heterogeneous molecular patterns. The diverse molecular mechanisms of resistance led to diverse clinical patterns of progression. The central question of the presentation was whether clinicians should make changes in therapy based on progression. The data provided many examples of targeted therapy for EGFR and ALK-positive NSCLC, which continued to increase survival after progression and in disease with heterogeneous molecular profiles.

In NSCLC patients with an EML4-ALK translocation who developed resistance to an ALK inhibitor, recent data have shown that while one-half of the resistant patients continue to have tumors driven by ALK mutations, the other half have developed alternative mutations that may render them insensitive to further ALK inhibition. For example, 12% of ALK inhibitor-resistant tumor samples were shown to now possess an EGFR mutation, which may be best managed with an EGFR tyrosine-kinase inhibitor (TKI) (Doebele R. J Clin Oncol. 2012;30 [suppl]; abstr 7504).

Another study (Chen et al. Oncologist. 2012) demonstrated a variability of molecular markers over time and across anatomical sites of disease. They found that acquired resistance might be anatomically isolated, therefore illustrating a heterogeneity of tumor markers in one patient. Dr. West said that “a rogue satellite of disease” that is not responding to a TKI can be treated with local therapy or chemotherapy while the patient continues on the TKI. Oxnard et al. found that a combination of resistant and responsive disease usually exists throughout a patient’s course of disease (Clin Cancer Res. 2011).

The ASCO 2013 presentation on acquired resistance in EGFR- and ALK-positive NSCLC points out that biomarker analysis affects treatment throughout the course of disease. For cancers with a known driver mutation, continued inhibition of the target may be beneficial after progression if a population of cells that are sensitive to the targeted therapy remains. Treating with anti-HER2 therapy in breast cancer post-progression, for example, has been shown to increase the response rate and have a survival benefit. There is also evidence to suggest that NSCLC tumors may remain sensitive to TKI or may be resensitized following a course of chemotherapy (Heon S. J Clin Oncol. 2012;30 [suppl]; abstr 7525).

The resensitizing approach has been shown to be effective both in the preclinical and clinical settings, and it avoids the introduction of a new drug therapy. A study at the Dana Farber Cancer Institute and Massachusetts General Hospital followed 24 EGFR-mutant patients for approximately 10 years. The patients were treated with a TKI, resensitized to the TKI and then retreated with the same drug. Two-thirds of the patients had stable disease with a good median progression-free survival, supporting the practice of restarting the targeted therapy compared with introducing a new drug. Randomized clinical trials are currently ongoing to compare the relative efficacies of these different treatment approaches in NSCLC patients.

In order to determine the best course of action for a recurrent patient, it is critical to understand the molecular profile that is driving the recurrence. This may dictate repeat biomarker testing upon relapse. However, repeat biopsies are not yet commonly performed in NSCLC, in part due to the difficulty in obtaining sufficient lung cancer tissue. 

Breast and Lung Cancer Biomarker Research at ASCO:
Changing Treatment Patterns (cont.)

Plasma DNA Testing

Two abstracts at ASCO focused on an alternative to tumor samples for molecular testing. The first study was conducted in breast cancer patients and assessed mutations in the PIK3CA gene.  PIK3CA is mutated in approximately 30% of all breast cancers and 45% of luminal A breast cancers. This study (Beaver J. J Clin Oncol. 2013;31 [suppl]; abstr 11019), conducted at Johns Hopkins, evaluated PIK3CA mutation rates in both formalin fixed paraffin-embedded (FFPE) tumor specimens and pre-surgery plasma samples using polymerase chain reaction (PCR). The results showed both a high sensitivity (92.3%) and specificity (100%) between the FFPE samples and the pre-surgery plasma.  

Similar to the aforementioned study of PIK3CA detection in plasma DNA in breast cancer patients, Mok et al., which reported on the detection of EGFR mutations from plasma DNA of NSCLC patients, showed that when using tissue as a comparator the sensitivity of plasma test was 76% (68 of 89 patients) and the specificity of plasma test was 96% (130 of 135 patients) (J Clin Oncol. 2013:31 [suppl]; abstr 8021).

The study concluded that an EGFR blood test can be used to reliably detect EGFR mutations in plasma and is a potent predictor of survival outcomes. These studies demonstrate that the ability to identify solid tumor mutations in plasma DNA by digital PCR could indicate improvements in diagnostics and advancements in therapeutic decision-making. They also raise the question of how to account for genetic heterogeneity in the tumor or multiple tumor sites when analyzing plasma DNA, as well as whether the plasma DNA test would be as effective if there were discordance between the primary tumor and a recurrence. 

Conclusion

Established biomarkers including ER, PR, HER2, EGFR and ALK hold great potential for the continued development of new targeted therapies and improved treatment decision-making throughout the course of disease. The studies presented at ASCO raise important issues that have repercussions for future treatment decisions: concordance between the primary tumor and a recurrence, acquired resistance, and genetic heterogeneity. All of these issues have the potential to change a tumor’s behavior and sensitivity throughout the course of the disease.

Changes over time may dictate vastly different approaches to treatment that should be considered.

These studies also raise important questions about companion diagnostics – when and how often should the testing occur? Are we using the appropriate testing cut points for HER2, for instance? What is the acceptable and clinically meaningful level of agreement between plasma DNA testing and tumor DNA testing in order to establish plasma DNA as a surrogate tissue for biomarker testing in the case of PIK3CA? When re-biopsying a patient’s tumor, should multiple disease sites be biopsied to best understand discordance and heterogeneity? In many cases no evidence-based treatment approach for patients resistant to initial targeted therapy exists. However, it is clear that the impact of biomarkers on treatment will continue to grow and improve the capacity to provide targeted therapy to cancer patients safely and effectively.

________________________________________________________________________

SIDEBAR: Evolution of Current Predictive Biomarkers

While over 300 biomarkers are currently recommended for use in oncology across all indications, only 25 of them directly influence treatment decisions (predictive and prognostic – guide decision to treat) Predictive biomarkers in many tumor types―including KRAS in colorectal cancer, BRAF in melanoma and EGFR in NSCLC―are now influencing treatment decisions.

Biomarkers for diagnosis and predicting risk of developing disease are common in several disease states, including autoimmune disease, diabetes and cancer. However, biomarkers that are predictive of response to a specific drug therapy are most developed within oncology. Heterogeneity of molecular mechanisms to drive tumor growth establishes the rationale for individualizing treatment decisions and utilizing targeted therapies. The extent and speed with which predictive biomarkers are growing in use in oncology makes this an especially significant therapeutic area.

The five types of oncology biomarkers, their function in oncology and the number (n) of biomarkers in each category are shown in Table 1.

Figure 3a illustrates the rapid growth of the development of predictive biomarkers in oncology in recent years. Figure 3b shows that an increasing percentage of cancer lives are affected by biomarkers. Kantar Health analysis estimates that approximately 800,000 cancer lives are affected by predictive biomarkers, and this number is expected to rise in the coming years as new biomarkers and targeted therapeutics are discovered and established into practice. 

Source: Kantar Health, CancerMPact® Patient Metrics, accessed April 26, 2013

Source: Kantar Health, CancerMPact® Patient Metrics, accessed April 26, 2013

________________________________________________________________________

About the Contributor

Julie Katz, MPH, MPhil, is an Associate Consultant, Clinical & Scientific Assessment at Kantar Health.

Kantar Health is a leading global healthcare advisory firm and trusted advisor to the world’s largest pharmaceutical, biotech, and medical device and diagnostic companies. It combines evidence-based research capabilities with deep scientific, therapeutic and clinical knowledge, commercial development know-how, and marketing expertise to help clients launch products and differentiate their brands in the marketplace.

Kantar Health’s oncology-related offers include CancerMPact® Biomarker Analysis, a global resource based on a thorough review of literature and recently published data that discusses the current and evolving oncology landscape with regard to biomarker segmentation and differences.

If you would like us to act as catalysts for you, contact us at www.kantarhealth.com/contactus.

© Caribou Publishing. All rights reserved. Reproduction in whole or in part is prohibited.