The Truth About Proxy Measures in Assessing Cancer Progression

Table of Contents

1. Introduction
2. Background of the Study
3. Research Questions
4. Study Population
5. Data Collection and Analysis
6. Findings of the Study
7. Implications of the Findings
8. Limitations of the Study
9. Future Research Directions
10. Conclusion

Introduction

In this article, we will discuss the findings of a research study conducted at Concerto Health AI on the performance of proxy measures compared with direct observation of progression-free survival (PFS) in metastatic breast cancer patients. The study aimed to evaluate the use of alternative endpoints such as time to discontinuation and time to next treatment as proxy indicators of PFS. We will explore the background of the study, the research questions addressed, the study population, data collection and analysis methods, the findings of the study, implications of the findings, limitations of the study, and potential future research directions.

Background of the Study

Real-world research often focuses on examining effectiveness outcomes that are not readily available in structured form in electronic medical records (EMRs). Endpoints like PFS and tumor response require human review and curation of unstructured content. As a result, alternative endpoints or proxy indicators such as time to discontinuation and time to next treatment are often used to assess the underlying value of interest. However, the performance of these proxy measures compared to direct observation of PFS is not well understood.

Research Questions

The study aimed to address the following research questions:

1. How do time to discontinuation and time to next treatment compare to PFS as proxy indicators in metastatic breast cancer patients?
2. Do the treatment groups based on aromatase inhibitor (AI) receipt show differences in these endpoints?
3. What are the implications of using proxy measures for comparative effectiveness research?

Study Population

The study included a sample of 378 female patients diagnosed with metastatic breast cancer. Eligible patients were at least 18 years of age and diagnosed in 2008 or later. All patients had hormone receptor-positive, HER2-negative breast cancer. The sample consisted of patients who received first-line therapy with or without an aromatase inhibitor.

Data Collection and Analysis

The research utilized structured and unstructured EMR data obtained through Concerto Health AI. Unstructured content, such as physician and nurse progress notes, lab reports, pathology reports, and radiological scan reports, underwent human review and curation. Descriptive methods were used to examine the study population, while Kaplan-Meier methods and Cox regression analysis were employed to analyze the outcomes of interest (time to discontinuation, time to next treatment, and PFS) in the first-line metastatic setting.

Findings of the Study

The study found significant differences in the three endpoints examined. The median PFS was approximately 9.1 months, while the median time to discontinuation and time to next treatment were around 5.5 and 6 months, respectively. The Kaplan-Meier figures showed clear separation between the survival curves for time to discontinuation and time to next treatment compared to PFS. This indicated that using these proxy measures would underestimate PFS.

Treatment groups based on aromatase inhibitor receipt also exhibited differences in the proxy indicators. Patients receiving aromatase inhibitor therapy had longer time to discontinuation and time to next treatment, suggesting fewer patients in this group discontinued or switched treatment before disease progression. These differences in outcome and treatment effect highlight the limitations of using proxy measures in comparative effectiveness research.

Implications of the Findings

The study's findings have several implications. Firstly, in certain cancer populations, proxy measures such as time to discontinuation and time to next treatment may yield different values than PFS and may not be reliable estimators of PFS. Secondly, the treatment effect observed in group comparisons using proxy measures can differ from the effect observed for PFS. This suggests that proxy indicators may not accurately capture the true effectiveness of treatments, especially if their safety profiles differ.

Limitations of the Study

It is important to acknowledge the limitations of this research. The study focused specifically on metastatic breast cancer patients and may not generalize to other populations or cancer types. The patient sample was drawn from community oncology practices, and Patterns of care may vary in other settings. Additionally, there may be alternative definitions of endpoints that could influence the findings.

Future Research Directions

Further research is needed to understand the underlying mechanisms driving the differences observed between proxy indicators and PFS. Investigating whether alternative definitions of endpoints alter the pattern of findings would provide valuable insights. Additionally, future studies should explore the use of proxy measures in comparative effectiveness research, particularly when comparing treatments with different safety profiles.

Conclusion

In conclusion, this research study sheds light on the performance of proxy measures compared to direct observation of PFS in metastatic breast cancer patients. The findings suggest that time to discontinuation and time to next treatment are not reliable estimators of PFS and can yield different outcome values and treatment effects. Understanding the limitations and implications of using proxy measures in real-world research is crucial for accurate assessment of treatment effectiveness. Further research in diverse cancer populations is necessary to validate these findings and explore potential solutions.

Highlights:

• Proxy measures such as time to discontinuation and time to next treatment may not accurately reflect the true disease progression in metastatic breast cancer patients.
• Treatment groups based on aromatase inhibitor receipt showed differences in the proxy indicators, indicating the influence of treatment safety profiles.
• The performance of proxy measures in comparative effectiveness research may yield incorrect estimations of treatment effects.

FAQs:

Q: Can proxy measures completely replace direct observation of progression-free survival (PFS)? A: No, proxy measures like time to discontinuation and time to next treatment cannot fully replace direct observation of PFS as they may underestimate disease progression.

Q: Why do patients switch treatments before experiencing disease progression? A: The most common reason for early treatment switching is treatment toxicity or tolerability. Patients may switch treatments if they experience side effects and hope for a more tolerable alternative, even if there is no evidence of disease progression.

Q: Are the findings of this study applicable to other types of cancer? A: The study focused specifically on metastatic breast cancer patients, and further research is needed to determine if the findings hold true for other cancer populations.

Q: What are the potential implications of using proxy measures in comparative effectiveness research? A: Using proxy measures with different safety profiles can lead to different treatment effect estimates, potentially influencing the conclusions drawn from comparative effectiveness research.

Q: What are the limitations of this study? A: The study was retrospective and based on data from community oncology practices, limiting its generalizability. Alternative definitions of endpoints were not extensively explored, and further investigation is required.

Q: What are the future research directions in this field? A: Future research should focus on understanding the underlying mechanisms behind the differences observed between proxy measures and direct observation of PFS. Exploring the use of alternative endpoint definitions and expanding the research to other cancer populations would also be valuable.