ABSTRACT
We describe a method to determine the phase III slope for the purpose of calculating indexes of ventilation heterogeneity, S(acin) and S(cond), from the multiple breath nitrogen washout test (MBNW). Our automated method applies a recursive, segmented linear regression technique to each breath of the MBNW test and determines the best point of transition, or breakpoint, between each phase of the washout. A sample set of 50 MBNW tests (controls, asthma, and COPD) was used to establish the conditions in which the phase III slope obtained from the automated technique best matched that obtained by two manual interpreters. We then applied our technique to a test set of 30 subjects (with an even number of subjects in each of the above groups) and compared these results against the manual analysis of a third independent manual interpreter. Indexes of ventilation heterogeneity were determined using both methods and compared. The phase III slopes determined by the automatic technique best matched the manual interpreter when the phase III slope was calculated from the phase II-III transition point plus the addition of 50% of the phase II volume to the end of the expiration. Calculation of the indexes S(acin) and S(cond) showed no overall difference between analysis methods in either S(acin) (P = 0.14) or S(cond) (P = 0.59) when the set threshold was applied to our automated analysis. Our analysis method provides an alternate means for rapid quantification of the MBNW test, removing operator dependence without alteration in either S(acin) or S(cond).
