Scope

Introducing Bayescores: A Pioneering Platform for Comprehensive Oncological Evaluation

Welcome to a new chapter in oncological assessment, where the art of medicine meets the precision of Bayesian analysis. In the intricate world of cancer treatment, quantification of a therapy's success is crucial. This is where Bayescores enters the picture - as a groundbreaking platform designed to address a critical gap in oncological evaluation.

The inception of Bayescores stems from the need to address the multifaceted challenges faced by clinicians, researchers and patients with cancer. Traditional metrics, while invaluable, often fall short in capturing the full spectrum of a therapy's impact. Recognizing this, we developed user-friendly tools that quantify the efficacy of cancer treatments with increased clarity and relevance.

Why Bayescores, and why now? Our efforts began with an acknowledgment of the evolving landscape of cancer care and the ever-present quest for improved patient outcomes. As we navigate through the nuances of oncology treatments, it becomes evident that a more sophisticated approach is needed - one that harmoniously blends clinical insight with advanced statistical analysis.

Join us as we delve into the medical and statistical foundations of Bayescores, and explore how this tool can improve how we assess the efficacy of cancer treatments.

Foundational challenges that motivated the development of Bayescores:

1. What is statistically significant is not always clinically meaningful::

  • Delta value in clinical trials is key for sample size calculation to achieve adequate statistical power.
  • Delta represents the minimal detectable difference between treatments. However, this difference may not be clinically meaningful even if “statistically significant”.
  • Increasing sample size may make any small difference statistically significant, underscoring the need for clinical interpretation of statistical results.

2. Growing Demand for Drug Value Assessment Tools (Post-2005):

  • Increasing healthcare costs and economic burden reflected in GDP spending.
  • Need for methods to comprehensively classify the clinical benefits of drugs accounting for cure fractions and not only survival extension.

3. Expert groups evaluating drug value beyond survival:

  • Introduction of multidimensional frameworks in clinical guidelines such as the Value Framework proposed by the American Society of Clinical Oncology (ASCO), the Evidence Blocks used by the National Comprehensive Cancer Network (NCCN), and the Magnitude of Clinical Benefit Scale (MCBS) recommended by the European Society for Medical Oncology (ESMO).
  • These frameworks consider efficacy, safety, evidence quality, data consistency, and treatment cost.
  • ESMO-MCBS in particular provides a multidimensional evaluation that incorporates multiple endpoints and metrics. It also emphasizes the need to adjust evaluation criteria based on the baseline risk of each tumor type.

4. Limitations of existing frameworks:

  • Dependence on hazard ratios, which may not accurately depict absolute clinical benefits.
  • Hazard ratios are not intuitive and can be misleading in terms of real patient health impact.
  • Complexity and potential fragmentation in understanding clinical benefits.
  • Discontinuities in scoring and inconsistency in the quantification of clinical benefit.

5. ESMO-MCBS approach:

  • Importance of data maturity and potential misinterpretation of plateau effects.
  • Variability in long-term benefits highlighted by the evolution of ESMO MCBS scores.

6. Challenges with multidimensional value assessment:

  • Embrace probability language to quantify uncertainty.
  • Limit discontinuities in value weighting to avoid inconsistencies.
  • Focus on clinically interpretable measures like differences in cured fractions or time ratios.
  • Take advantage of Bayesian methodology towards a more comprehensive and transparent assessment.

7. Data maturity and uncertainty at end of KM curves:

  • Overcoming the limitations of commonly used metrics such as hazard ratios is crucial for accurate clinical benefit interpretation.
  • Time ratios are informative metrics for non-curable scenarios.
  • For treatments that can produce cures, differences in cure fractions are a useful comparator.
  • Bayesian modeling intuitively quantifies and represent the uncertainty in the data.
  • Diverse perspectives can be incorporating by emphasizing different components of treatment efficacy in Bayescores.