Discussion
This study provides a first insight into the association between extent of structural damage in the lungs and airways in people with CF and the FEV1 response to treatment with CFTR modulators during real life follow up. There is no previous literature about this association.
Unsurprisingly, our data confirm the correlation between baseline ppFEV1 and Brody score derived from chest CT, as has been established before by Helbich et al. for an earlier CT score in adult and pediatric CF patients, and by Brody et al. for Brody score in children with CF. 12,13
This study shows different findings concerning the correlation between Brody score and lung function change for treatment with ivacaftor versus lumacaftor/ivacaftor: we find a positive correlation between Brody score and ppFEV1 response to ivacaftor treatment, whereas we do not find any correlation between these same parameters for treatment with lumacaftor/ivacaftor.
This contrast in findings could be explained by the differences in treatment efficacy of the CFTR modulating drug; lumacaftor/ivacaftor for F508del homozygous patients is known to be far less effective compared to ivacaftor for patients with a gating mutation5,14. This difference adds to the difficulties in the assessment of factors correlated with magnitude of response.
Notably, the ivacaftor group consisted of patients with a relatively well-preserved lung function and low Brody scores, reflecting only minimal structural changes in the lungs. This may explain why a positive correlation between Brody scores and lung function change on treatment was found; in patients whose lungs are not severely affected by the disease, an increase in lung function cannot be expected after the start of treatment; the goal of treatment would be to preserve lung function and to prevent lung damage rather than to reverse it. In a group of patients with more severely affected lungs, it is possible that the correlation would be absent or even negative, if extensive structural damage proves to be irreversible by CFTR modulating therapy.
Also the fact that the timing of the CT scans was just before start of therapy in the ivacaftor group, while this was anywhere in the 18 months before or after start of therapy for the lumacaftor/ivacaftor group, could in part account for the difference in findings. The decision to include scans within the time span of 18 months before or after start, was based on literature about the progression of structural damage visible on CT scan in CF, showing no relevant progression in 18 months on most parameters15.
This study faced several limitations, most of these concerning the group of patients who had started lumacaftor/ivacaftor, as we use clinically indicated CT scans in this group. Firstly, this implicates that differences in scanning protocol occurred with variation of the clinical indications that warranted the chest CT. This resulted in missing air trapping scores in 5 patients in this group. We therefore used Brody score excluding air trapping score for this analysis. As air trapping (together with mucus plugging) would be the aberrance most likely to improve on treatment, also influencing FEV1, this could decrease the potential of the Brody score to pick up a possible correlation between baseline CT anomalies and FEV1change on therapy. Another consequence of using indicated CT scans, is that the timing of the scans is imperfect in comparison with CT scans in the ivacaftor group (where CT was timed just before start of the treatment as part of the regular follow up protocol). This might lead to differences in the findings, as theoretically, some anomalies (such as air trapping, mucus plugging) might improve quickly after initiation of therapy. No information about short term effects of CFTR modulators on CT findings is available in literature however.
A third limitation that the use of clinically indicated scans brings along, is the fact that these scans may show more (possibly reversible) abnormalities than would be present in a stable situation, as the scan is carried out due to respiratory complaints. Also purely based on CT scan it is not easy to discriminate between reversible and irreversible structural abnormalities. For example, a CT scan could show consolidations or increased bronchial thickening, reflecting an exacerbation, in a patient undergoing CT scan because of hemoptysis. The ivacaftor group on the other hand is small, making the data more prone to be influenced by outliers, troubling calculation of partial correlations, and making it impossible to tell if the correlation found is in fact linear, or curved. An advantage was that in this group CT scans were timed optimally and included expiration slides for air trapping analysis.
Verification and closer analysis of this correlation in a larger patient cohort would be favorable. Another additional approach could be to evaluate if Brody scores from follow up CT scans improve in patients while they are using CFTR modulators, and whether or not this improvement is in line with the change in ppFEV1.