RatioLogo
Back

Breast Cancer Screening: The Mathematics of Shadows

In the high-stakes world of breast cancer screening, the difference between a life-saving catch and a missed lesion often comes down to the math of shadows. To identify a tumor, software must first understand the baseline of the healthy fat and glands surrounding it.

The Data Gap in Medical Physics

For years, medical physics has grappled with a surprising inconsistency: scientists couldn't quite agree on the exact X-ray attenuation of breast fat.

This data gap matters to every patient because accurate spectral imaging depends on these values to differentiate between dense tissue and potential threats. It ensures the clearest possible image while keeping radiation doses optimized.

Resolving the Fog: A New Study

A new study utilizing photon-counting spectral mammography has set out to resolve this fog. It compared:

  • 6 fixed in-vitro tissue samples from the Cambridge Breast Unit
  • 94 in-vivo screening mammograms from Stockholm

By measuring how X-rays lose energy as they pass through these samples, researchers are refining the digital map used to navigate the human breast.

Key Findings

In-Vivo Measurements

The team found that for every 10 mm of tissue, the in-vivo adipose measurements equated to:

  • 8.15 mm of PMMA (a standard reference plastic)
  • -0.123 mm of Aluminum

These results stayed remarkably consistent, showing a coefficient of variation of just 1.4% for the PMMA equivalence.

The In-Vitro Discrepancy

The study also uncovered a "statistical signature" of the challenges involved in lab work. The in-vitro samples showed a significantly lower PMMA equivalence of 7.95 mm.

This discrepancy yielded a p-value of 0.03, suggesting that when tissue is removed and fixed in formalin for a lab setting, its chemical composition changes enough to alter how it interacts with X-rays.

The Future of Research

"The in-vivo method provides a new non-invasive way of finding values of tissue attenuation," the authors noted.

This suggests that future research can now skip the lab bench and use large, anonymized clinical datasets to reduce errors.

Remaining Mysteries & Next Steps

Despite the precision of the photon-counting system, some mysteries remain:

  • The researchers had to assume a fixed 3 mm skin thickness for all patients
  • The small sample size of just 6 lab specimens increased the "noise" in the in-vitro data (variability for aluminum measurements jumped to 15%)

Moving forward, the team aims to:

  1. Refine skin-thickness assumptions
  2. Account for calibration drifts
  3. Ensure the next generation of mammograms is even more personalized to the patient’s specific anatomy

Based on: "X-ray attenuation of adipose breast tissue: in-vitro and in-vivo measurements using spectral imaging," by Fredenberg, E., et al. (2015), published in Proc. SPIE 9412.