Radiation pneumonitis(RP) is a common and potentially fatal dose-limiting toxicity that occurs in lung cancer patients undergoing lung and/or mediastinal radiotherapy or synchronous radiochemotherapy, and its incidence is 14.6–37.2%. Most cases occur around 6 months after radiotherapy[14, 15]. RTOG classified RP into grades 0 to 5 based on the clinical symptoms, lung radiologic presentation, and corticosteroid treatment results. Grade 3 and above RP is classified as severe acute RP(SARP)[8, 16]. With advances in radiotherapy, the indications for chest radiotherapy in lung cancer patients are currently expanding, and the incidence of radiotherapy-induced severe lung complications has decreased. However, SARP still occurs after lung cancer radiotherapy with an incidence rate of 3–6%[14, 17]. In clinical practice, the clinical outcomes of such patients were different even though they were promptly discovered and aggressive clinical treatment was carried out. Some patients recovered, while symptoms gradually worsened in some patients and they died. There have been no related studies that compared the radiation dosimetric parameters between lung cancer patients who developed SARP with different outcomes and evaluated their role in treatment-related toxicity. This study was the first to reveal the heterogeneity in clinical factors, physical dosimetric factors, and time of SARP onset in SARP patients. This could provide a theoretical research foundation for further optimization of lung radiotherapy plan and decrease the incidence of fatal RP in lung cancer patients in clinical practice.
In this study, the data of 31 patients who had undergone lung and/or mediastinal radiotherapy and died of SARP from January 2013 to December 2019 were collected. In addition, data were collected from 35 patients who developed SARP and survived during the same period. It is worth noting that we tried as much as possible to collect patients who had undergone chest and/or mediastinal radiotherapy and died of SARP. However, we did not include all patients who developed SARP and survived because equal sampling was employed in this study and the surviving patients that were included to the analysis could represent the overall characteristics of all patients who developed SARP and survived. Therefore, the number of patients in this analysis does not represent the mortality rate of SARP in our center. In the existing studies on lung cancer patients who develop RP after radiotherapy, some clinical characteristics have been considered important risk factors for RP progression. Previous studies have shown that performance status, history of smoking, nutritional status, past lung disease, tumor stage, pathology and site, concurrent chemotherapy, pre-radiotherapy lung function, and surgery are associated with the occurrence and severity of RP[7, 18–24]. According to Giuranno et al., older age (> 65 years) is a factor for decreased tolerability to radiotherapy[16, 25], and women may have a higher risk of RP due to their lower lung volume[26]. A retrospective study in a cancer center in China showed that the incidence of fatal RP was higher in lung cancer patients with chronic silicosis. In lung cancer patients, symptomatic interstitial lung disease (ILD), even asymptomatic subclinical ILD, is a risk factor for RP[27, 28]. There is still debate over the roles of COPD and lung infection. A prospective study by Zhou et al. showed that emphysema was a risk factor for RP after radiotherapy in non-Small Cell Lung Cancer patients[29], but this correlation was not found in other studies[30]. A large number of studies evaluated smoking. Surprisingly, the risk of RP in smokers is decreased. This may be due to decreased inflammatory responses, which do not respond to DNA damage[15, 25, 28]. The risk of RP in patients whose radiotherapy target region is the lung field is three times higher than that in patients whose radiotherapy target region is the mediastinum[14, 18]. At present, there are related papers on clinical factor prediction model for SARP[31–33]. In this study, we did not observe significant differences in these clinical factors between the death and the survivor cohorts (all P > 0.05). In future studies, the sample size should be expanded, and more well-designed prospective randomized controlled studies should be carried out to further analyze and validate our conclusions.
It has been widely observed and reported that DVH dose parameters are associated with RP. Multiple doses, including mean lung dose and the percentage volume that received a fixed dose, are associated with an increased risk of SARP. Previous studies have found that MLD shows a significant positive correlation with RP[34]. Hernando found that the risk of RP in patients who received an MLD of more than 8.5 Gy was 3.8 times higher than that in patients who received ≤ 8.5 Gy[35]. In our SARP patients, the MLD of the death group and the survivor group was 10.69 Gy (9.475 Gy–12.18 Gy) and 7.134 Gy (4.34 Gy–11.13 Gy), respectively, and the mean MLD of the death group was slightly higher than that of the survivor group. This result is consistent with previous data. It is exciting to observe significant differences in MLD between the two groups (P < 0.01). Univariate analysis showed that OR of MLD was 0.998 (95% CI: 0.997–1; P < 0.05), which means that MLD can be used as a potential evaluation marker for the final outcome of SARP patients. However, MLD did not reach significance in multivariate analysis, showing that it may not play a critical role in SARP outcome. Therefore, we will carry out more prospective studies in the future for validation. We will also confirm the optimal cutoff point for dosimetric standard to more accurately predict the final outcome of SARP patients.
The lungs are “parallel organs” as they consist of multiple parallel functional units. Other functional units are not affected when one functional unit is damaged[36]. Therefore, some studies have found that the severity of RP is intimately associated with the lung volume that received a radiation dose that exceeded the radiation tolerability of the lungs. The specific marker reflecting this relationship is PTV/LV[23, 37]. Many previous studies have shown that PTV/LV plays a critical role in RP progression. According to Jinming Yu, et al., PTV/LV is independent from other dosimetric factors and is considered a novel and special marker for SARP after radiotherapy in esophageal cancer patients[31]. In our study, no significant difference in the PTV/LV ratio was observed between the SARP death and survivor cohorts. Univariate analysis showed that OR for PTV/LV was 0.714 (95% CI: 0.358–3.801; P = 0.862), meaning that PTV/LV cannot predict the final outcome of SARP patients.
Many studies have shown that the mean V20 and V30 of patients with RP were 42.0% and 38.0%, respectively, which were higher compared with patients who did not develop RP[38]. Currently, there is widespread acceptance of V20 evaluation of treatment plans and prediction of RP incidence[39]. Hanania AN et al. reported that V20 was not only associated with the incidence of RP but also significantly correlated with the severity of RP[16]. Some researchers have found that the V20 of patients who died of RP was ≥ 32%, and the V20 of patients who developed grade 3 and above RP was ≥ 30%. V20 should be less than 25% in order to prevent SARP[14, 40, 41]. Our results showed that the V20 of ipsilateral lungs in deceased SARP patients was 32.34 (26.5–36.5), and the V20 of ipsilateral lungs in surviving patients was 29.27 (26.69–33.74), which revalidates the above findings. However, no significant differences in V20 were observed in the ipsilateral lung, contralateral lung, or total lung between these two cohorts. Hence, V20 is not a decisive factor for the final outcome of SARP patients. The study by Marks LB et al. showed that the incidence of RP was 6% and 24% when lung V30 was 17.7% and 17.8–24.5%, respectively. Therefore, the lung V30 should be controlled at below 18%[14, 41, 42]. In our study, the mean ipsilateral lung V30 of the two groups was greater than 18%; specifically, the ipsilateral lung V30 of the death group and the survivor group was 28.3 (23.41–33.58) and 20.36 (16.56–22.53), respectively. The univariate analysis of V30 in these two groups showed significant differences, but multivariate analysis did not confirm the role of ipsilateral lung V30 in the final outcome of SARP patients. Although significant differences in contralateral lung and total lung V10, mean heart dose, and maximum spinal cord dose were found in univariate analysis, they were not significant in multivariate analysis.
With advances and developments in the IMRT era, researchers have paid more attention to the potential negative effects of low-dose lung DVH parameters on RP, particularly V5[43]. Many researchers have found that limiting V5 to 60% and below can greatly decrease the incidence of RP and radiation fibrosis in patients[44]. However, some studies on the occurrence and progression of RP have not observed significant differences in V5. The multivariate analysis reported by Lu Wang et al. showed that V5 did not play an important role in SARP progression. Our study showed that the ipsilateral lung, contralateral lung, and total lung V5 were all below 60%, and there was a significant difference between the death and the survivor groups in univariate analysis. Surprisingly, multivariate analysis showed significant differences in ipsilateral lung V5 (OR, 0.225; 95% CI: 0.051–0.993; P < 0.05) and total lung V5 (OR, 40.976; 95% CI: 1.029–1632.127; P < 0.05). Finally, we believe that these two factors are independent predictors and reliable parameters of the outcome of SARP patients and jointly determine the final outcome of SARP patients. Higher ipsilateral lung and total lung V5 in the DVH are associated with greater susceptibility to SARP and poorer final outcome or death. However, we hope to increase our sample size and conduct prospective studies to decrease selection bias and further increase the reliability of the data.
Our study showed that the median time to SARP onset in the survivor group and the death group was 64 days and 98 days, respectively. The difference in the time of SARP onset between the two groups was statistically significant, and the median time of SARP onset in the survivor cohort was significantly earlier than that in the death cohort. This may be due to the pathophysiological developmental trends of RP, and patients with earlier SARP onset have significant treatment advantages as they receive treatment earlier. Therefore, these patients have better outcomes. We also found that the median time from SARP onset to death was 159.5 days in the death group, and the median time from SARP onset to outcome was 93.5 days in the survivor group. Although this study was the first to reveal the heterogeneity in clinical factors, physical dosimetric factors, and time of SARP onset in SARP patients, there are several limitations that need to be urgently solved. First, the sample size of this study was small, and the sample size should be expanded in future studies to further validate our prediction results. Furthermore, this was a retrospective study, and selection bias may be present. Lastly, we only analyzed basic clinical factors and physical parameters. In subsequent studies, bioinformatics, genome information, and radiomics information will be added to improve the prognosis determination criteria for SARP patients.