The present study demonstrates that RT induces RILI in all patients who undergo external RT with variable clinical manifestations and different degrees of lung damage. This variability in the degree of RILI suggests that severe RP may be associated with differences in individual predisposition. The early decrease of DLCO(%) was a predicting factor of severe RP development and thus could serve as a marker for early diagnosis and treatment modification.
The reported prevalence of RP in lung cancer patients ranges from 0–58%. 4 This variation is likely due to differences in diagnostic scales, non-specific symptoms, and the lack of standardized assessment protocols.4 In the present study, standard follow-up protocols detected RILI in all of the patients but with different severity presentation. This finding suggests that all patients treated with RT are likely to develop RILI to a greater or lesser extent, which may depend on individual biological characteristics.
Previous studies have shown that subtle, local changes in PFT after RT can be used as indicators of acute and chronic lung damage, although published results are not always consistent.3,9 The largest and most consistent changes in PFT values after RT are observed in DLCO, which has been directly associated with respiratory morbidity.31 In the present study, we evaluated the mean differences in FEV1(%) and DLCO(%) between low-grade RP patients [G1-G2] and high-grade RP patients [G3-G5]. We found no statistically significant differences between the two groups in FEV1(%) values at the different time points, thus leading us to conclude that FEV1(%) is not a predictor of RP, a finding that is consistent with other reports.9,31 By contrast, we found that a decrease in DLCO(%) 2 month post-RT was predictive of RP severity, in line with some previous reports.9,31 Importantly, the decline in DLCO(%) after three weeks of RT was an early predictor of severe RP. The explanatory power of this variable could be that histopathological alterations present during the latency phase (i.e., without clinical manifestations) for RILI may alter gas exchange.3 Clearly, the ability to early predict severe RILI would be helpful to optimize therapeutic options.
Classically, RP grade has been associated with the radiation dose.3,8 In our study, the high-grade group received a higher mean radiation dose but lower V20 than the low-grade group. This finding is contrast with the meta-analysis published by Palma et al.32 The absence of a significant association in this study between RP grade and radiation parameters could be due to the limited sample size, the small differences among patients with regards to the radiation dose and biological predisposition.9
RP increases cellular metabolism. 18F-FDG-PET/CT quantify this metabolic increase; indeed, 18F-FDG uptake during RT and post-treatment is a marker of symptomatic RP.13–17 Our results suggest that a high pre-treatment SUVmax is associated with high grade RP. Accordingly, Castillo et al demonstrated the predictive value of pre-treatment 18F-FDG lung uptake in the subsequent development of RP symptoms.16–17
RILI produce an imbalance between type 1 and type 2 helper T-cells and abnormal fibroproliferative wound healing33. Variations among patients in terms of RP severity and lung repair capacity could be related to individual pre-treatment lung biomolecular conditions and genetic factors34. In the present study, increased expression of some mediators were obtained in BAL of tumor lungs in both patients (RP-G1 and RP-G3), although these proteins were also expressed in the tumor-free lung of the RP-G3 patient before RT. Previous studies indicate that IL-1ra is involved in acute inflammation35; MIF modulates RILI36; and CXCL1 promotes angiogenesis and thus contribute to the pathogenesis of PF.33 Interestingly, CD154, IFN-γ, IL-23 and PAI-1 were expressed in both lungs (tumor and tumor-free lung) before RT only in the RP-G3 patient. These four cytokines have been described in animal models of lung fibrosis.37, 38 Furthermore, a recent study reported that a truncated PAI-1 protein protects against RILI in a murine model.39 Finally, Liu et al. found that rs7242 GT/GG genotypes located in the 3ÚTR of PAI-1 were associated with a significantly increased risk of RP.40 Overall, our findings suggest a potential biological predisposition to lung damage and altered wound healing in RILI development, which would deserve a depth study to better understand pathogenesis.
Study strengths and limitations.
We have to recognize some limitations: First, the small sample size which was calculated using “observed versus a reference mean” and even although we have enrolled sixteen patients, instead of including seventeen, we have not had any loss of follow up. Secondly, the low power of the biological lung analysis (only two patients). In this sense, this is a pilot study to identify if there are differences in biological features associated with different grades of RP. Our findings, warrant further investigation in a larger sample. The main strength of the study is that it is the first prospective study to evaluate patients with NSCLC through a longitudinal clinical and biological follow-up that demonstrate RT induces RILI in all cases but in some of them with a high-grade of lung injury and consequent altered wound repair.