Parametric Response Mapping as a Diagnostic Indicator of Bronchiolitis Obliterans Syndrome

Bronchiolitis obliterans syndrome (BOS) is associated with significant morbidity and mortality following allogeneic stem cell transplant (SCT). Defined by NIH consensus criteria, the diagnosis is often confounded by the presence of concurrent infectious complications in affected patients, making interpretation of spirometry and imaging studies difficult. We now report a novel imaging biomarker that identifies obstructive small airway disease, even in the presence of infectious pneumonitis. A computed tomography (CT)-based image processing technique termed parametric response mapping (PRM) has been generated at our center to capture parenchymal alterations consequent to lung disease. PRM is performed from a biphasic (inspiratory and expiratory) high-resolution-CT (HRCT), with subsequent image co-registration of the paired HRCT data, and classification of individual voxels within the lung parenchyma as either (a) normal, (b) severe emphysematous disease, (b) functional small airway disease (fSAD) characteristic of BOS or (d) parenchymal disease (PD) characteristic of infection (Figure). Spatial information is retained allowing for visualization and quantification of disease within individual lobes. The relative volumes for each classification are determined by normalizing the sum of all voxels within a particular class to the total lung volume. Using HRCT, broncho-alveolar lavage (BAL) and spirometry data, 55 SCT recipients were classified into one of 3 groups: A) Infection, no BOS (n=11); B) BOS, no infection (n=34); and C) BOS + Infection (n=9). BAL were performed within 1 week of the HRCT, in all 55 patients. Mean values (±SE) for functional small airway disease (fSAD), parenchymal (infectious) disease (PD), and normal lung parenchyma were determined for each group (Figure). RESULTS: Distinct imaging profiles were identified for patients with BOS and for patients with an acute infectious pneumonitis. In particular, the %fSAD was significantly greater in patients with BOS when compared to those with infection alone, 38±2% vs 17±4%, p=0.05. There was no difference in the %fSAD for subjects with BOS, whether a concurrent infection was present or not, 35±3% (BOS) vs 38±2% (BOS + Infection), p=31.  Patients with an acute infectious pneumonitis had significantly higher levels of PD than patients with BOS, 30±4% vs 17±1.5%. In 7 of the 34 (21%) BOS cases, significant increases in fSAD (>30%) were present when radiographic features characteristic for BOS (air trapping, bronchiectasis, septal lines) were either absent or minimal on HRCT. There were no differences in PRM imaging by the type of infection. PRM provides a major advance in our ability to diagnose small airway obstruction that characterizes BOS, even in the presence of an active pulmonary infection. A prospective trial comparing PRM with spirometry and standard HRCT as an early indicator of BOS is planned.