Preoperative concurrent chemoradiotherapy (CCRT) and surgery are treatment options available for locally advanced non-small-cell lung cancer (NSCLC) (1). However, radiation pneumonitis (RP) is a serious adverse event of CCRT (2). Prediction and management of grade 2 or more RP are therefore important (3).

Skeletal muscle mass depletion is related to poor cancer-related prognosis (4). In addition, obesity is closely associated with cancer incidence (5), and visceral adipose expansion is related to the secretion of inflammatory cytokines such as interleukin-6 (IL6) (6), which is involved in RP development (7-9). Watanabe et al. found that patients with a larger amount of visceral adipose tissue had a higher rate of admission to an intensive care unit owing to coronavirus disease (COVID-19) (10). However, to the best of our knowledge, no studies have focused on the relationship between RP and soft tissue-related parameters. In addition to body mass index (BMI), computed tomography (CT)-determined evaluation has recently been performed for measuring muscle and adipose mass (4,10-13). Therefore, we aimed to examine the relationship between muscle and adipose mass and RP using CT.

Patients and treatment. We reviewed the medical records of patients who underwent preoperative CCRT and subsequent surgery between January 2004-March 2018 at the Okayama University Hospital and for whom data on plain CT at the third lumbar vertebra (L3) level performed before starting CCRT were available. The staging of NSCLC was determined using the seventh edition of the tumour-node-metastasis classification (14). Three-dimensional radiation therapy was performed as previously described, and resection was performed after CCRT (15). This study was approved by the Ethics Committee of our Institution (approval number 1809-018). All procedures implemented were in accordance with the ethical standards of the 1964 Declaration of Helsinki and subsequent amendments. We provided the right to opt out by posting relevant information on the Institutional website and in the Outpatient Ward.

CT assessment of muscle and adipose masses. Using plain CT scans at the L3 level, contours were extracted automatically by setting the Hounsfield unit (HU) thresholds within a specific range while manually correcting evident errors on the workstation computer (SYNAPSE VINCENT version 5.5; FUJIFILM Medical Co., Ltd., Tokyo, Japan). Skeletal muscles included all muscles at the L3 level with HU thresholds ranging from −29 to +150 (4,11). Psoas muscle contours were created by manually deleting other muscles (12). The HU thresholds for automatic extraction ranged from −150 to −50 and from −190 to −30 for visceral and subcutaneous adipose tissue, respectively (Figure 1) (13). Body composition variables were expressed as cm²/m². Each parameter was used to construct their respective indexes: skeletal muscle index (SMI), psoas muscle index (PMI), visceral adipose tissue index (VAI), and subcutaneous adipose tissue index (SAI) (12,13). Additionally, the visceral-to-subcutaneous adipose tissue area ratio was calculated to investigate the abdominal adipose tissue distribution (13).

Evaluation and statistics. RP grade was determined using the Common Toxicity Criteria for Adverse Events version 4.0.3 (16). We analyzed dose-volume histogram (DVH) parameters related to the lung volume percentage receiving >5 Gy (V5) and >20 Gy (V20), and the mean lung dose (MLD). Cut-off values for continuous variables were calculated from individual time-dependent receiver operating characteristic curves (17). We examined the relationship between these parameters and grade two or more RP using a log-rank test for univariate analysis and Cox proportional hazard model for multivariate analysis comprising factors that had significant effects (p<0.05) in the univariate analysis. The Wilcoxon–Mann–Whitney test was used to assess differences between the two groups. A p-value of less than 0.05 was considered statistically significant (two-tailed). R version 3.5.1 and an additional package of survival ROC version 1.0.3 (R Foundation for Statistical Computing, Vienna, Austria) were used for all analyses.

The study included 94 eligible patients. Table I shows their demographic and clinical characteristics. In the 28 patients who developed grade two or more RP, the median period from starting CCRT to RP onset was 12.43 (range=7.71-60.86) weeks.

Table II shows the results of the cut-off values, univariate and multivariate analyses of factors associated with grade 2 or more RP. The absolute Pearson correlation coefficients between V5/V20, V5/MLD, and V20/MLD were 0.862, 0.848, and 0.926, respectively. On multivariate analysis, only VAI was associated with grade 2 or more RP (p=0.026).

The cumulative 6-month incidence rate of grade 2 or more RP was 27.7% (95% confidence intervaI=18.0-36.2%). Grade 2 or more RP incidence rates at 6-months were 21.4% (range=9.9-31.5%) and 36.8% (range=19.5-50.5%) in patients with VAI <39 cm²/m² and ≥39 cm²/m², respectively (Figure 2). Of 28 patients with grade 2 or more RP, four experienced RP before surgery. The median VAI was 34.03 cm²/m² and 51.26 cm²/m² in those who developed grade 2 or more RP prior to and after surgery, respectively. There was no significant difference in parameters between the two groups (p=0.8). The median V5, V20, and MLD values were 29.0%, 25.1% and 12.0 Gy, respectively, in the group that developed grade 2 or more RP prior to surgery and 38.4%, 24.5% and 11.7 Gy, respectively, in those who developed it after surgery. No significant difference was observed between these two groups (p=0.47, 0.68, and 0.82, respectively).

We demonstrated that only VAI is positively related to RP in patients with preoperative CCRT. To the best of our knowledge, this is the first study to show an association between visceral adipose mass and RP. In obese patients, the adipose tissue becomes dysfunctional, thereby promoting the formation of an inflammatory environment. The excessive production of inflammatory adipocytokines, such as tumour necrosis factor-alpha (TNFα) and IL6, leads to a mild state of chronic inflammation (18). IL6 is a multifaceted pro-inflammatory cytokine with a wide range of biological activity (19). Some studies have reported a relationship between RP and IL6. High pre- and post-treatment IL6 levels were correlated with RP development (7). A meta-analysis suggested that the serum IL6 levels in patients with RP before radiotherapy were significantly higher than those in patients without RP (8). Early IL6 variations during radiotherapy are associated with RP risk (9). TNFα may also be involved in RP development. Furthermore, radiotherapy was shown to increase the TNFα level, and blockage of TNFα activity reduced the number of fibrous lesions in the lungs of mice that received radiotherapy (20). Patients with large amounts of visceral adipose tissue may experience excessive IL6 and TNFα production from these tissues before CCRT or IL6 and TNFα levels may easily be increased by stimulation, which may lower the threshold for inflammation and facilitate RP development owing to stimulation by CCRT.

Regarding adipose tissue, the SAI was not associated with RP in our study. For visceral adipose tissue, omental fat produces IL6 at levels that are two- to three-fold higher than those produced by subcutaneous adipose tissue (21); this may be why SAI was not associated with RP. Additionally, SMI and PMI were not associated with RP. Although exercise training was found to reduce the degree of adipose tissue inflammation by inhibiting the infiltration of inflammatory macrophages and CD8⁺ T-cells (22), muscle mass itself did not affect the promotion and suppression of RP development.

Fujiwara et al. examined the relationship between visceral and subcutaneous adipose mass measured using CT, in addition to BMI, and the prognosis of patients with hepatocellular carcinoma (13). In their study, the visceral-to-subcutaneous adipose tissue area ratio was a prognostic factor, while BMI was not. Similarly, in our study, grade 2 or more RP was associated with visceral adipose mass but not with BMI. Two studies have examined obesity and RP in patients with breast cancer. Lee et al. showed that a high BMI tends to be associated with a lower risk of RP (23), whereas Allen et al. showed a high BMI to be the only risk factor for RP development (24). The assessment of obesity by BMI alone may be inadequate.

On univariate analysis, V20 in the DVH is a predictive factor for RP (25,26), which our findings are consistent with. As in our previous study (15), the lower lobe as the disease location was a factor predictive of grade 2 or more RP in this study.

This study has limitations. Firstly, it had a retrospective design. Secondly, inflammatory adipocytokine levels were not measured.

In conclusion, we demonstrated that visceral adipose mass is associated with grade 2 or more RP in patients undergoing preoperative CCRT. Measuring visceral adipose mass may help predict RP development. Further studies on the relationship between visceral adipose mass and RP are needed.

The Authors declare that they have no competing interests in regard to this study.

KuK collected the data, and drafted the article; TO performed statistical analysis and drafted the article; SS collected the data. All Authors participated in the design of the study, and read and approved the final article.

The Authors would like to thank Dr Kenta Watanabe (Department of Radiology, Okayama University Hospital, Okayama, Japan) for data collection.

This study was supported by a donation from Tsuyama Chuo Hospital. The study sponsor did not participate in any procedure of this study.