Multiple regression analysis of risk factors related to radiation pneumonitis

Clinical predictors of severe radiation pneumonitis in patients undergoing thoracic radiotherapy for lung cancer

Background

Severe radiation pneumonitis (RP), one of adverse events in patients with lung cancer receiving thoracic radiotherapy, is more likely to lead to more mortality and poor quality of life, which could be predicted by clinical information and treatment scheme. In this study, we aimed to explore the clinical predict model for severe RP.

Methods

We collected information on lung cancer patients who received radiotherapy from August 2020 to August 2022. Clinical features were obtained from 690 patients, including baseline and treatment data as well as radiation dose measurement parameters, including lung volume exceeding 5 Gy (V5), lung volume exceeding 20 Gy (V20), lung volume exceeding 30 Gy (V30), mean lung dose (MLD), etc. Among them, 621 patients were in the training cohort, and 69 patients were in the test cohort. Three models were built using different screening methods, including multivariate logistics regression (MLR), backward stepwise regression (BSR), and random forest regression (RFR), to evaluate their predictive power. Overoptimism in the training cohorts was evaluated by four validation methods, including hold-out, 10-fold, leave-one-out, and bootstrap methods, and test cohort was used to evaluate the predictive performance of the model. Model calibration, decision curve analysis (DCA), and evaluation of the nomograms for the three models were completed.

Results

Severe RP was up to 9.4%. The results of multivariate analysis of logistics regression in all patients showed that patients with subclinical (untreated and asymptomatic) interstitial lung disease (ILD) could increase the risk of severe RP, and patients with a better lung diffusion function and received standardized steroids treatment could decrease the risk of severe RP. The three models built by MLR, BSR, and RFR all had good accuracy (>0.850) and moderate κ value (>0.4), and the model 2 built by BSR had the highest area under the receiver operating characteristic (ROC) curve (AUC) in three models, which was 0.958 [95% confidence interval (CI): 0.932-0.985]. The calibration curve showed good agreement between the predicted and actual values, and the DCA showed a positive net benefit for the model 2 which drew the nomogram. The model 2 included subclinical ILD, diffusing capacity of the lung for carbon monoxide (DLCO), ipsilateral lung V20, and standardized steroid treatment, which could affect the incidence of severe RP.

Conclusions

Subclinical ILD, DLCO, ipsilateral lung V20, and with or not standardized steroid treatment could affect the incidence of severe RP. Strict lung dose limitation and standardized steroid treatment could contribute to a decrease in severe RP.

Hydrogen-rich solution alleviates acute radiation pneumonitis by regulating oxidative stress and macrophages polarization

This study was aimed to investigate the effect of hydrogen-rich solution (HRS) on acute radiation pneumonitis (ARP) in rats. The ARP model was induced by X-ray irradiation. Histopathological changes were assessed using HE and Masson stains. Inflammatory cytokines were detected by ELISA. Immunohistochemistry and flow cytometry were performed to quantify macrophage (CD68) levels and the M2/M1 ratio. Western blot analysis, RT-qPCR, ELISA and flow cytometry were used to evaluate mitochondrial oxidative stress injury indicators. Immunofluorescence double staining was performed to colocalize CD68/LC3B and p-AMPK-α/CD68. The relative expression of proteins associated with autophagy activation and the adenosine 5'-monophosphate-activated protein kinase/mammalian target of rapamycin/Unc-51-like kinase 1 (AMPK/mTOR/ULK1) signaling pathway were detected by western blotting. ARP decreased body weight, increased the lung coefficient, collagen deposition and macrophage infiltration and promoted M1 polarization in rats. After HRS treatment, pathological damage was alleviated, and M1 polarization was inhibited. Furthermore, HRS treatment reversed the ARP-induced high levels of mitochondrial oxidative stress injury and autophagy inhibition. Importantly, the phosphorylation of AMPK-α was inhibited, the phosphorylation of mTOR and ULK1 was activated in ARP rats and this effect was reversed by HRS treatment. HRS inhibited M1 polarization and alleviated oxidative stress to activate autophagy in ARP rats by regulating the AMPK/mTOR/ULK1 signaling pathway.

Integration of dosimetric parameters, clinical factors, and radiomics to predict symptomatic radiation pneumonitis in lung cancer patients undergoing combined immunotherapy and radiotherapy

PURPOSE

This study aimed to combine clinical/dosimetric factors and handcrafted/deep learning radiomic features to establish a predictive model for symptomatic (grade ≥ 2) radiation pneumonitis (RP) in lung cancer patients who received immunotherapy followed by radiotherapy.

MATERIALS AND METHODS

This study retrospectively collected data of 73 lung cancer patients with prior receipt of ICIs who underwent thoracic radiotherapy (TRT). Of these 73 patients, 41 (56.2 %) developed symptomatic grade ≥ 2 RP. RP was defined per multidisciplinary clinician consensus using CTCAE v5.0. Regions of interest (ROIs) (from radiotherapy planning CT images) utilized herein were gross tumor volume (GTV), planning tumor volume (PTV), and PTV-GTV. Clinical/dosimetric (mean lung dose and V5-V30) parameters were collected, and 107 handcrafted radiomic (HCR) features were extracted from each ROI. Deep learning-based radiomic (DLR) features were also extracted based on pre-trained 3D residual network models. HCR models, Fusion HCR model, Fusion HCR + ResNet models, and Fusion HCR + ResNet + Clinical models were built and compared using the receiver operating characteristic (ROC) curve with measurement of the area under the curve (AUC). Five-fold cross-validation was performed to avoid model overfitting.

RESULTS

HCR models across various ROIs and the Fusion HCR model showed good predictive ability with AUCs from 0.740 to 0.808 and 0.740-0.802 in the training and testing cohorts, respectively. The addition of DLR features improved the effectiveness of HCR models (AUCs from 0.826 to 0.898 and 0.821-0.898 in both respective cohorts). The best performing prediction model (HCR + ResNet + Clinical) combined HCR & DLR features with 7 clinical/dosimetric characteristics and achieved an average AUC of 0.936 and 0.946 in both respective cohorts.

CONCLUSIONS

In patients undergoing combined immunotherapy/RT for lung cancer, integrating clinical/dosimetric factors and handcrafted/deep learning radiomic features can offer a high predictive capacity for RP, and merits further prospective validation.