Short-term Endpoints for Cancer Screening Trials: Does Tumor Subtype Matter?

Area of Science:

• Oncology
• Biostatistics
• Cancer Screening Research

Background:

• Multicancer early detection tests necessitate reevaluation of short-term endpoints in cancer screening trials.
• Reduction in advanced stage cancer incidence is a key candidate endpoint.
• Stage-shift models are used to predict mortality reduction from screening by substituting early-stage for late-stage survival.

Purpose of the Study:

• To compare mortality predictions from stage-shift models that ignore versus preserve prognostic subtypes.
• To illustrate differences using ovarian cancer (histologic subtype) and prostate cancer (grade).
• To examine implications for short-term endpoints in cancer screening trials.

Main Methods:

• Analysis of stage-shift models, differentiating between those that ignore and those that preserve prognostic subtypes.
• Utilized ovarian cancer partitioned by histologic subtype and prostate cancer partitioned by grade for illustration.
• Inferred conditions under which subtype-preserving models yield different mortality predictions.

Main Results:

• Standard stage-shift models may inaccurately imply a subtype shift when only stage shift is considered.
• Models ignoring prognostic subtypes can lead to different mortality reduction predictions compared to subtype-preserving models.
• The impact of ignoring subtypes varies depending on the cancer's specific prognostic subtypes.

Conclusions:

• Prognostic subtype information is critical for accurate mortality reduction predictions in cancer screening.
• Stage-shift models that ignore prognostic subtypes may not reliably predict true mortality benefits.
• Refining short-term endpoints in cancer screening trials requires careful consideration of prognostic subtypes.