Investigation of clinical and dosimetric factors associated with postoperative pulmonary complications in esophageal cancer patients treated with concurrent chemoradiotherapy followed by surgery

Discussion

In this study, the volume of lung spared from doses of 5 Gy or higher (VS5) was found to be the only independent predictive factor associated with postoperative pulmonary complications for esophageal cancer patients treated with concurrent chemoradiotherapy followed by surgery. The smaller the volume of lung receiving doses <5 Gy, the higher the incidence of postoperative pulmonary complications.

Findings from this study suggest that the volume of remaining undamaged lung, rather than the volume of damaged functional lung, determines postoperative pulmonary function. In other words, patients who have a small lung volume to start might be at high risk of experiencing pulmonary complications even if their relative V5 is low. Strong evidence for this is that, in our study, gender was associated with the incidence of postoperative pulmonary complications on univariate analysis but not on multivariate analysis. As noted, this might have been because the total lung volume was smaller in women than in men, thus making the volume spared in female patients smaller than in male patients, despite similar lung volumes exposed to radiation. This in turn suggests that patients with a smaller lung volume and less functional reserve might be more susceptible to postoperative pulmonary complications. Thus, to reduce the risk of postoperative pulmonary complications, more attention might have to be paid not only to the DVH of the lung but also to the total lung volume and nonirradiated lung volume during treatment planning.

The advent of 3-D dosimetry and treatment planning has provided new opportunities to assess potential radiation risk. In the present study, numerous parameters generated from the lung DVH (mean dose, relative V5, and VS5–35) were found to be significantly related to the incidence of postoperative pulmonary complications. This suggests that the radiation dose distribution to lung might have a greater impact on postoperative pulmonary complications than do other clinical factors. However, the various dosimetric factors are highly correlated with one another, which makes it difficult to determine which aspect of the dose distribution to lung has the greatest impact on complication risk. Nevertheless, multivariate analysis of our data set suggests that low-dose lung volume has more to do with the incidence of postoperative pulmonary complications than does high-dose lung volume, which agrees with our previous analysis. At the same time, it must be kept in mind that our current findings are based on data from 110 patients, of whom only 18 had postoperative pulmonary complications. Our conclusions need to be further substantiated with additional data; in the meantime, other dosimetric parameters should continue to be taken into consideration during treatment planning.

Published data on the association of DVH parameters and radiation-induced lung injury in patients receiving concurrent chemoradiation are scarce, and most of this information has been derived from studies of lung cancer patients. To date, relative V20 and mean lung dose are the only factors that have been shown to be predictive of radiation pneumonitis in patients receiving concurrent chemoradiation. Our present results suggest that mean lung dose might be an important predictive factor for esophageal cancer patients treated with chemoradiotherapy. However, our study also suggests that doses much lower than 20 Gy might be critical in causing lung injury. This difference is perhaps due to the use of a different endpoint in our study. The use of surgical resection of esophageal cancer after neoadjuvant treatment with drugs and radiation is believed to trigger the postoperative complication in lungs that otherwise might have suffered only subclinical damage after chemoradiotherapy, the so-called “two-hit” theory . Of note, in a study of 26 lung cancer patients, 80% of whom were treated with concurrent chemoradiotherapy, using a more sensitive endpoint than overt radiation pneumonitis, namely DLCO measured with pulmonary function tests, there was a pronounced decrease in DLCO when the local radiation dose exceeded 13 Gy. This study also indicates that low radiation doses (<20 Gy) result in pulmonary damage.

That we found no other clinical factors that we investigated in this study to be associated with the incidence of postoperative pulmonary complications might be owing to our very strict selection criteria, which allowed a relatively homogeneous population, and to the small number of patients we analyzed in this study. The majority of our patients had good performance status, no COPD, and adequate pulmonary function before treatment, which might have improved their chances of avoiding postoperative pulmonary complications. In addition, studies have shown genes related to inflammation to be a disposing factor for radiation-induced pneumonitis. For example, Kong reported that loss of heterozygosity at the mannose 6-phosphate insulin-like growth factor 2 receptor (M6P/IGF2R) locus gene strongly correlates with the development of radiation pneumonitis after thoracic radiotherapy, and that patients with loss of heterozygosity were much more likely to have elevated plasma-transforming growth factor ß, suggesting an inability to normally process this cytokine. The investigators concluded that loss of the M6P/IGF2R gene might predispose patients to the development of radiation-induced lung injury. Pharmacogenetic studies on the population included in the current study have been planned.

In conclusion, we found that dosimetric factors were strongly associated with postoperative pulmonary complications in esophageal cancer patients treated with concurrent chemoradiation followed by surgery. Of these factors, the volume of lung spared exposure to doses of =5 Gy (VS5) was the factor most strongly associated with postoperative pulmonary complications in this cohort. No single cut-off value of VS5 was found to guarantee freedom from pulmonary complications; the risk of pulmonary complications seems to continue to decrease with larger VS5 but is predicted by our modeling to be <5% when VS5 >3000 cm3. In light of our results, it seems prudent to limit the lung volume irradiated to even quite low doses, particularly in patients who have a smaller lung volume to begin with. In addition, further study is needed to generate a more reliable predictive model that includes other potentially influential dosimetric, biologic, and clinical factors.