Deep-Learning Model Prediction of Radiation Pneumonitis Using Pretreatment Chest Computed Tomography and Clinical Factors.

Objectives: This study aimed to build a comprehensive deep-learning model for the prediction of radiation pneumonitis using chest computed tomography (CT), clinical, dosimetric, and laboratory data. Introduction: Radiation therapy is an effective tool for treating patients with lung cancer. Despite its effectiveness, the risk of radiation pneumonitis limits its application. Although several studies have demonstrated models to predict radiation pneumonitis, no reliable model has been developed yet. Herein, we developed prediction models using pretreatment chest CT and various clinical data to assess the likelihood of radiation pneumonitis in lung cancer patients. Methods: This retrospective study analyzed 3-dimensional (3D) lung volume data from chest CT scans and 27 features including dosimetric, clinical, and laboratory data from 548 patients who were treated at our institution between 2010 and 2021. We developed a neural network, named MergeNet, which processes lung 3D CT, clinical, dosimetric, and laboratory data. The MergeNet integrates a convolutional neural network with subsequent fully connected layers. A support vector machine (SVM) and light gradient boosting machine (LGBM) model were also implemented for comparison. For comparison, the convolution-only neural network was implemented as well. Three-dimensional Resnet-10 network and 4-fold cross-validation were used. Results: Classification performance was quantified by using the area under the receiver operative characteristic curve (AUC) metrics. MergeNet showed the AUC of 0.689. SVM, LGBM, and convolution-only networks showed AUCs of 0.525, 0.541, and 0.550, respectively. Application of DeLong test to pairs of receiver operating characteristic curves respectively yielded P values of .001 for the MergeNet-SVM pair and 0.001 for the MergeNet-LGBM pair. Conclusion: The MergeNet model, which incorporates chest CT, clinical, dosimetric, and laboratory data, demonstrated superior performance compared to other models. However, since its prediction performance has not yet reached an efficient level for clinical application, further research is required. Contribution: This study showed that MergeNet may be an effective means to predict radiation pneumonitis. Various predictive factors can be used together for the radiation pneumonitis prediction task via the MergeNet.

Comparing deep learning and handcrafted radiomics to predict chemoradiotherapy response for locally advanced cervical cancer using pretreatment MRI.

Concurrent chemoradiotherapy (CRT) is the standard treatment for locally advanced cervical cancer (LACC), but its responsiveness varies among patients. A reliable tool for predicting CRT responses is necessary for personalized cancer treatment. In this study, we constructed prediction models using handcrafted radiomics (HCR) and deep learning radiomics (DLR) based on pretreatment MRI data to predict CRT response in LACC. Furthermore, we investigated the potential improvement in prediction performance by incorporating clinical factors. A total of 252 LACC patients undergoing curative chemoradiotherapy are included. The patients are randomly divided into two independent groups for the training (167 patients) and test datasets (85 patients). Contrast-enhanced T1- and T2-weighted MR scans are obtained. For HCR analysis, 1890 imaging features are extracted and a support vector machine classifier with a five-fold cross-validation is trained on training dataset to predict CRT response and subsequently validated on test dataset. For DLR analysis, a 3-dimensional convolutional neural network was trained on training dataset and validated on test dataset. In conclusion, both HCR and DLR models could predict CRT responses in patients with LACC. The integration of clinical factors into radiomics prediction models tended to improve performance in HCR analysis. Our findings may contribute to the development of personalized treatment strategies for LACC patients.

Predictive factors for severe radiation-induced lung injury in patients with lung cancer and coexisting interstitial lung disease.

BACKGROUND AND PURPOSE: This study aimed to investigate the predictive factors of severe radiation-induced lung injury (RILI) in patients with lung cancer and coexisting interstitial lung disease (ILD) undergoing conventionally fractionated thoracic radiotherapy. MATERIALS AND METHODS: The study includes consecutive patients treated with thoracic radiotherapy for lung cancer at two tertiary centers between 2010 and 2021. RILI severity was graded using the National Cancer Institute Common Terminology Criteria version 5.0, with severe RILI defined as toxicity grade ≥4, and symptomatic RILI as grade ≥2. The absolute neutrophil count (ANC), absolute lymphocyte count (ALC), and C-reactive protein were collected within 4 weeks before starting radiotherapy. Neutrophil-lymphocyte ratios (NLR) were calculated as ANC/ALC. The median follow-up was 9 (range, 6-114) months. RESULTS: Among 54 patients, 22 (40.7 %) had severe RILI. On multivariate logistic regression analysis, high pretreatment ANC (p = 0.030, OR = 4.313), pretreatment NLR (p = 0.007, OR = 5.784), and ILD severity (p = 0.027, OR = 2.416) were significant predictors of severe RILI. Dosimetric factors were not associated with severe RP. Overall survival was significantly worse for patients with severe RILI than those without, with 1-year cumulative overall survival rates of 7.4 % and 62.8 %, respectively. CONCLUSION: Pretreatment blood NLR, ANC, and ILD severity were associated with severe RILI. Overall survival was dismal for patients with severe RILI.

Development of an experimental model for radiation-induced inhibition of cranial bone regeneration.

BACKGROUND: Radiation therapy is widely employed in the treatment of head and neck cancer. Adverse effects of therapeutic irradiation include delayed bone healing after dental extraction or impaired bone regeneration at the irradiated bony defect. Development of a reliable experimental model may be beneficial to study tissue regeneration in the irradiated field. The current study aimed to develop a relevant animal model of post-radiation cranial bone defect. METHODS: A lead shielding block was designed for selective external irradiation of the mouse calvaria. Critical-size calvarial defect was created 2 weeks after the irradiation. The defect was filled with a collagen scaffold, with or without incorporation of bone morphogenetic protein 2 (BMP-2) (1 µg/ml). The non-irradiated mice treated with or without BMP-2-included scaffold served as control. Four weeks after the surgery, the specimens were harvested and the degree of bone formation was evaluated by histological and radiographical examinations. RESULTS: BMP-2-treated scaffold yielded significant bone regeneration in the mice calvarial defects. However, a single fraction of external irradiation was observed to eliminate the bone regeneration capacity of the BMP-2-incorporated scaffold without influencing the survival of the animals. CONCLUSION: The current study established an efficient model for post-radiation cranial bone regeneration and can be applied for evaluating the robust bone formation system using various chemokines or agents in unfavorable, demanding radiation-related bone defect models.

Evaluation of the hybrid-dynamic conformal arc therapy technique for radiotherapy of lung cancer.

PURPOSE: A hybrid-dynamic conformal arc therapy (HDCAT) technique consisting of a single half-rotated dynamic conformal arc beam and static field-in-field beams in two directions was designed and evaluated in terms of dosimetric benefits for radiotherapy of lung cancer. MATERIALS AND METHODS: This planning study was performed in 20 lung cancer cases treated with the VERO system (BrainLAB AG, Feldkirchen, Germany). Dosimetric parameters of HDCAT plans were compared with those of three-dimensional conformal radiotherapy (3D-CRT) plans in terms of target volume coverage, dose conformity, and sparing of organs at risk. RESULTS: HDCAT showed better dose conformity compared with 3D-CRT (conformity index: 0.74 ± 0.06 vs. 0.62 ± 0.06, p < 0.001). HDCAT significantly reduced the lung volume receiving more than 20 Gy (V20: 21.4% ± 8.2% vs. 24.5% ± 8.8%, p < 0.001; V30: 14.2% ± 6.1% vs. 15.1% ± 6.4%, p = 0.02; V40: 8.8% ± 3.9% vs. 10.3% ± 4.5%, p < 0.001; and V50: 5.7% ± 2.7% vs. 7.1% ± 3.2%, p < 0.001), V40 and V50 of the heart (V40: 5.2 ± 3.9 Gy vs. 7.6 ± 5.5 Gy, p < 0.001; V50: 1.8 ± 1.6 Gy vs. 3.1 ± 2.8 Gy, p = 0.001), and the maximum spinal cord dose (34.8 ± 9.4 Gy vs. 42.5 ± 7.8 Gy, p < 0.001) compared with 3D-CRT. CONCLUSION: HDCAT could achieve highly conformal target coverage and reduce the doses to critical organs such as the lung, heart, and spinal cord compared to 3D-CRT for the treatment of lung cancer patients.