Quantitative early decision making metric for identifying irregular breathing in 4DCT

Area of Science:

• Medical Imaging
• Radiation Oncology
• Pulmonary Function

Background:

• Four-dimensional computed tomography (4DCT) is crucial for radiation therapy planning, but image quality can be compromised by irregular breathing patterns.
• Accurate assessment of breathing patterns is essential for ensuring the efficacy and safety of radiation treatments.
• Existing methods for assessing breathing patterns during 4DCT may be cumbersome or lack quantitative predictive capabilities.

Purpose of the Study:

• To develop and validate a quantitative metric for early prediction of breathing patterns and irregular breathing during 4DCT.
• To assess the metric's performance in a large patient cohort undergoing clinical 4DCT.

Main Methods:

• A novel metric, kappa-relative (κrel), was developed based on the amplitude of a single abdominal surrogate, adapted from a tidal volume-based method.
• Three patient cohorts (N=47, 256, and 86) were analyzed, with breathing patterns classified as Type 1 (apneic at end-exhalation) or Type 2 (forced respiration).
• Statistical analyses, including Ansari-Bradley and Kruskal-Wallis tests, were used to compare breathing classifications and validate the κrel metric against expert-defined regular/irregular breathing.

Main Results:

• The κrel metric demonstrated significant consistency in breathing pattern classification across different cohorts.
• κrel values >1.72 indicated substantial irregular breathing, while values <1.45 indicated minimal irregularity, impacting 4DCT image quality.
• The metric provided a reliable early warning system, with physician discretion for re-acquisition needed in only 11.9% of cases.

Conclusions:

• The κrel metric accurately classifies regular and irregular breathing patterns in a large patient population during 4DCT.
• κrel, using a single surrogate, effectively classifies breathing types comparable to methods using multiple surrogates.
• This quantitative metric offers a rapid and accurate tool for assessing breathing patterns and guiding decisions during 4DCT acquisition.