Can We See Clearly Now? Computed Tomography for Bronchiolitis Obliterans Syndrome Prognostication

Chronic lung allograft dysfunction (CLAD) or “chronic rejection” is the main limitation to survival in lung transplant recipients; half of patients develop CLAD within 3–5 y of transplant.¹ Unfortunately, there are no curative options for CLAD. Therapies are targeted at prevention or stabilization of lung function after diagnosis.² Although patients view CLAD as a death knell, the reality is that the prognosis can be less dire than anticipated, especially for the bronchiolitis obliterans syndrome (BOS) phenotype.^3,4 There is, however, considerable interpatient variability in outcomes with BOS, and prognostication tools are lacking. The ability to accurately estimate survival at the time a patient is newly diagnosed with CLAD is important for preserving their peace of mind and quality of life.^5,6

In this issue of Transplantation, Van Herck and colleagues offer a radiographic prognostication tool for BOS, which can used by pulmonologists at the bedside to provide timely counseling.⁷ The group utilizes a simplified Brody II scoring system, which was originally conceived to objectify severity of pulmonary abnormalities on chest computed tomography (CT) for patients with cystic fibrosis.⁸ The Brody II score aims to quantify aspects of lung airway pathology including severity and extent of (a) central and peripheral bronchiectasis; (b) central and peripheral mucus plugging; (c) central and peripheral peri-bronchial wall thickening; (d) parenchymal changes (such as ground glass opacification, consolidation, blebs/bullae); and (e) air-trapping on expiratory images.⁹ The main benefit of the Brody II score is that it enables quantification of morphologic disease progression that may be inconsistently captured by spirometry. Other notable benefits are its low interobserver variability and validation in non-CF populations.^8,10 Although the Brody II scoring system tabulates parameters of disease severity in each lobe of the lung (counting the lingula separately), Van Herck and colleagues chose to simplify the process and assess the lungs as a singular entity. BOS shares some pathologic features with cystic fibrosis, which favors the extrapolation of the Brody score to that population; BOS is characterized by small airway inflammation and fibrosis with subsequent occlusion and development of proximal airway bronchiectasis.¹¹ The authors compared high versus low Brody II scores (dichotomized based on the median value) to duration of allograft survival (death or retransplantation) following a diagnosis of definite CLAD/BOS.¹² In addition, the authors performed a comparison between Brody II scores and differences in CLAD risk factors, including acute cellular rejection, antibody-mediated rejection, de novo donor-specific antibody formation, gastroesophageal reflux disease, CARV, and CMV viremia. A separate comparison of incidence of bacterial and fungal infections preceding CLAD/BOS diagnosis and Brody II scores was also performed. The single-center study included 106 bilateral lung transplant recipients between 2004 and 2016 who subsequently developed BOS.

This study showed that a high modified Brody II score was directly associated with severity of BOS and greater decrements in both FEV1 and FVC. Time to diagnosis of BOS was shorter in patients with high Brody II scores by 331 d. Additionally, there was a greater rate of decline in FEV1 (~25 mL/mo) in the 6-mo preceding BOS diagnosis in patients with higher scores, although this was not statistically significant. Finally, Brody II scores above 4.5 were associated with a median survival of 5 y versus a median survival of 10 y for those with lower scores (P = 0.046). A Brody II score >6.3 was associated with an even more dire prognosis, with a median survival of <2 y. High subscores for mucus plugging, peri-bronchial thickening, and parenchymal abnormalities were associated with a significant decrease in median survival, whereas features of bronchiectasis and hyperinflation were not. In this cohort, there was no association between modified Brody II scores at BOS diagnosis and BOS risk factors such as acute cellular rejection, antibody-mediated rejection, de novo donor-specific antibody, and gastroesophageal reflux disease. Greater Brody II scores were associated with increased frequency of prior bacterial infections, especially due to Pseudomonas species. Moreover, Brody II scores were highly associated with increased bronchoalveolar fluid neutrophilia at the time of BOS diagnosis (despite 80% of the cohort being maintained on azithromycin prophylaxis).

This work demonstrates that in all-comers with CLAD/BOS, features on CT imaging can identify patients with a poor survival prognosis. There are 2 additional noteworthy points that this study highlights: (1) infections due to Pseudomonas are associated with increased radiographic derangements and greater severity of BOS. This finding corroborates the work of several other groups that have identified Pseudomonas infection and colonization as a risk factor for CLAD.^13,14 (2) Van Herck and colleagues demonstrate that patients with higher Brady scores at BOS diagnosis had CT chest evidence of lung abnormalities as early as 3 mo posttransplant; 2 of 3 of patients with Brody score >4.5 at the time of BOS diagnosis had median Brody II scores of >3 on baseline 3-mo CT scans. This finding is congruent with what is known about baseline lung allograft dysfunction (BLAD), defined as an FEV1/FVC< 70% and FEV1< 80% predicted by 12 mo posttransplant. BLAD has been shown to be a risk factor for CLAD. Until now, BLAD has been defined by spirometric criteria, but the findings of this study suggest that radiographic criteria may be a useful adjunct for refining the definition of BLAD.^15,16 A modified Brody II score may prove invaluable to risk stratifying lung transplant recipients with early CT chest lung abnormalities.

There are several strengths of this study. First, the cohort is well-defined and includes patients with the BOS phenotype of CLAD only. Despite a 12-y duration of this study, all CT scan images were of similar quality because high-resolution inspiratory and expiratory imaging was an inclusion criterion. Finally, a subset of patients had imaging that was reviewed by 2 attending radiologists and showed high interobserver agreement, similar to the original Brody II score validation in patients with cystic fibrosis. As with all studies, there were aspects that could be improved upon. The single-center, retrospective design limits generalizability to other cohorts and increases the likelihood for biases. In particular, this cohort had a greater population of patients with COPD (53%) and a higher population of women (58%); both female gender and COPD as the underlying disease type are associated with improved long-term survival.^17,18 In addition, the authors utilized a modified version of the Brody II score, which has not been previously validated. Although the modified scoring system is simpler to use, it may minimize the importance of regional variation of disease.¹⁹ The scoring is somewhat subjective, which may limit interinstitutional comparisons, although the scoring guide available in the supplementary section should be helpful in attempting to standardize the scoring process as best as possible.

The use of a simple algorithm to objectify lung abnormalities on CT provides clinicians with the ability to counsel patients regarding prognosis at the time they receive a diagnosis of BOS. From a patient’s viewpoint, this can provide comfort and perspective when receiving an anxiety-provoking diagnosis. For physicians, such a scoring system allows for the identification of higher-risk patients who may benefit from aggressive therapies to stabilize lung function, including enrollment in clinical trials. It will be interesting to determine whether Brody II scores are modifiable with therapies for CLAD/BOS, as this has potential as a marker of treatment response. Furthermore, utilizing Brody II scoring prior to when patients meet spirometric criteria for BOS might identify candidates that would benefit from early interventions to mitigate disease progression.

Van Herck and colleagues provide a paradigm shift by harnessing the potential for CT imaging to prognosticate and risk stratify; in doing so, they will help the transplant community gain further clarity regarding the management of BOS.