Mammographic sensitivity as a function of tumor size: A novel estimation based on population-based screening data

Area of Science:

• Radiology
• Oncology
• Biostatistics

Background:

• Mammographic sensitivity is better represented by a function of tumor size than a single value.
• Previous models may have overestimated size-specific sensitivity.
• A novel approach is needed to accurately estimate mammography sensitivity as a function of tumor size.

Purpose of the Study:

• To develop a novel model for estimating mammographic sensitivity as a function of tumor size.
• To improve the accuracy of sensitivity estimation in breast cancer screening.
• To provide a more realistic reflection of mammography's detection capabilities.

Main Methods:

• Utilized aggregated data from 22,915 screen-detected and 10,670 interval breast cancers.
• Employed an exponential tumor growth model and a 4-year follow-up to back-calculate tumor sizes.
• Determined sensitivity as a function of tumor size by estimating false-negative cancers and conducted univariate sensitivity analysis and systematic review for validation.

Main Results:

• The developed model demonstrated increasing sensitivity from 0% at 2 mm to 85% at 20 mm tumor size.
• Sensitivity for a 20-mm tumor ranged from 74% to 93% based on tumor volume doubling time confidence intervals.
• The model's estimated sensitivity aligned with findings from two out of three studies in a systematic review.

Conclusions:

• The novel model, derived from aggregated breast screening data, offers a more accurate representation of observed screening program outcomes.
• This size-specific sensitivity function provides a better understanding of mammography's detection performance.
• The findings suggest this model can enhance the interpretation of mammographic screening results.