AI-Techniques Loss-Based Algorithm for Severity Classification (ATLAS): a novel approach for continuous...

Area of Science:

• Cardiopulmonary exercise testing (CPET)
• Clinical informatics
• Symptom quantification

Background:

• Muscular effort and breathlessness (dyspnea) are significant, disabling symptoms.
• Current assessment methods often focus only on maximal effort during CPET.
• There is a need for continuous symptom quantification throughout CPET.

Purpose of the Study:

• To develop new paradigms for assessing effort-related symptoms continuously during CPET.
• To improve the quantification and analysis of dyspnea and muscular effort.
• To create a novel algorithm for classifying symptom severity.

Main Methods:

• Established sex- and age-adjusted reference centiles for the Borg scale (0-10).
• Developed a novel algorithm, AI-Techniques Loss-Based Algorithm for Severity Classification (ATLAS), using reciprocal exponential loss.
• Applied the ATLAS algorithm to CPET data from patients with chronic obstructive lung disease (COPD).

Main Results:

• The ATLAS algorithm's dyspnea intensity categories, unlike peak exercise dyspnea, accurately discriminated between patients with increasing levels of resting and exercise impairment (P < .05).
• Continuous quantification of symptoms throughout CPET provided a more nuanced assessment than peak-effort measures.
• The developed algorithm demonstrated effectiveness in classifying symptom severity in COPD patients.

Conclusions:

• The novel AI-techniques approach, ATLAS, offers a paradigm shift in analyzing effort-related symptoms during CPET.
• This method is poised for translation into clinical practice for managing patients with activity-related symptoms.
• Innovative informatics research has enabled a new way to display, quantify, and analyze symptoms in patient populations.