ABSTRACT
RATIONALE AND OBJECTIVES: Partial liquid ventilation using Perfluorcarbon (PFC) is an innovative treatment of acute respiratory distress syndrome. However, the underlying mechanisms are not totally clear. The aim was to investigate the distribution of oxygen partial pressure within the PFC-filled lung (ppO2). METHODS: Nine pigs underwent partial liquid ventilation, receiving 20 mL PFC/kg bodyweight (bw). Measurements were obtained by a chemical shift selective TurboFLASH sequence at different axial lung levels. ppO2 was calculated from 19F-MRI by nonlinear curve T1-fitting technique after noise correction. RESULTS: Quantification and distribution of ppO2 was performed successfully. A narrow relationship of the inspiratory O2 fraction and ppO2, as well as a significant ventral-to-dorsal gradient of ppO2 (ventral:dependent lung = 1.9:1) were detected in all subjects and slice positions. CONCLUSIONS: In vivo measurement of local ppO2 gains new and clinical important insights into the physiology of PLV. The previously unknown ppO2 gradient within PFC fits to distribution of perfusion. Dependent lung regions appear to have limited access to O2 from central airways.
Subject(s)
Liquid Ventilation/methods , Magnetic Resonance Imaging/methods , Oxygen/analysis , Pulmonary Alveoli/physiology , Respiratory Distress Syndrome/physiopathology , Respiratory Distress Syndrome/therapy , Animals , Fluorocarbon Polymers , Partial Pressure , Pulmonary Ventilation/physiology , Respiratory Distress Syndrome/bloodABSTRACT
A method for in vivo measurement of oxygen partial pressure (pO2) in porcine lungs during partial liquid ventilation (PLV) with perflubron (PFOB) was developed. A pulse sequence for high-resolution MRI of the distribution of PFOB in the lung after intratracheal administration was developed as well. Moreover, quantitative measurements of longitudinal relaxation time T(1) of (19)F resonances for assessment of regional pO2 are described. Due to the need to acquire data during a single expiratory breathhold, only low SNRs were achieved in vivo. Therefore, simulations were performed to investigate the influence of background noise on T(1) values calculated from data with low SNR. Based on these simulations, a postprocessing strategy was developed to correct for systematic errors by background noise prior to quantitative analysis. Results of a pilot study in pigs under conditions of PLV are presented.
