Validation of radiofrequency determined lung fluid using thoracic CT: Findings in acute decompensated heart failure patients

Abstract

Background: Noninvasive outpatient monitoring for heart failure (HF) has significant opportunity to reduce patient morbidity and the costs associated with recurrent hospitalization. The purpose of this study was to validate the ability of radiofrequency (RF) to assess lung fluid via a wearable patch device compared to thoracic CT in order to characterize volume overload. Methods: 120 subjects were studied: 66 acute heart failure (AHF) inpatients and 54 subjects without AHF (Control – 44 healthy and 10 stable HF). All underwent supine thoracic CT scans and supine RF readings from the wearable patch device placed on the left mid-axillary line (age = 74 ± 16 vs. 57 ± 15 yrs.; female = 38 vs. 44%; BMI = 33.2 ± 9.0 vs. 27.3 ± 5.1, AHF vs. Control respectively). Reflected RF signals and subject-specific anthropometric data were used to calculate the RF-determined lung fluid content. CT Lung fluid was reported as percentage of lung volume. Classification analyses were used to compare RF and CT performance. Results: AHF presented with higher lung fluid than controls by both CT and RF (CT: 20.1 ± 4.2% vs. 15.4 ± 2.4%; RF: 20.7 ± 5.6% vs. 15.6 ± 3.3%; p < 0.05 for all). The correlation between lung fluid measured by CT vs. RF was r = 0.7 (p < 0.001). RF determined lung fluid performed as well as CT in distinguishing AHF from control subjects: Sensitivity: 70% vs. 86%; Specificity: 82% vs. 83%; Positive Predictive Value: 82% vs. 86%; Negative Predictive Value: 69% vs. 83%, CT vs. RF respectively. Conclusions: Noninvasive nonionizing RF determined lung fluid provides a potential alternative to other measures for diagnosing and monitoring pulmonary fluid overload.

Keywords

• Pulmonary congestion
• Lung fluid
• chest CT
• Classification analyses