Discriminating Breast Cancer Patients by Spectral Analysis of Arterial Pulse Waveforms

Area of Science:

• Cardiovascular physiology
• Biomedical engineering
• Oncology

Background:

• Breast cancer (BC) poses a significant health challenge, necessitating accessible screening methods.
• Cardiovascular system alterations are observed in BC patients.
• This study investigates noninvasive arterial pulse-wave analysis for BC detection.

Purpose of the Study:

• To compare arterial pulse-wave characteristics between breast cancer patients and healthy controls.
• To evaluate a novel pulse-distribution analysis (PDA) method for breast cancer screening.
• To identify distinct pulse-waveform signatures indicative of breast cancer.

Main Methods:

• Derived 40 harmonic indices from one-minute radial arterial waveform recordings.
• Compared machine learning classifiers with a novel pulse-distribution analysis (PDA).
• Analyzed differences in pulse-wave transmission conditions between BC patients and controls.

Main Results:

• PDA achieved an AUC of 0.74, outperforming machine learning (AUC up to 0.58).
• Subgroup analysis showed improved discrimination with disease progression (AUC 0.86 for stage IV).
• Identified significant differences in pulse-waveform indices between groups.

Conclusions:

• Distinct pulse-waveform signatures were identified in breast cancer patients.
• The proposed low-cost, noninvasive PDA method shows potential for breast cancer screening.
• Further validation in larger cohorts is warranted.