Analysis of related factors of radiation pneumonia caused by precise radiotherapy of esophageal cancer based on random forest algorithm

A prognostic nomogram for T3N0M0 esophageal squamous cell carcinoma patients undergoing radical surgery based on computed tomography radiomics and inflammatory nutritional biomarkers

BACKGROUND

This study explores the significance of computed tomography (CT) radiomic features, along with inflammation and nutrition biomarkers, in the prognosis of postoperative patients with T3N0M0 esophageal squamous cell carcinoma (ESCC). The study aims to construct a related nomogram.

METHODS

A total of 114 patients were enrolled and randomly assigned to training and validation cohorts in a 7:3 ratio. Radiomic features were extracted from their preoperative chest-enhanced CT arterial images of the primary tumor, and inflammatory and nutritional indices, including neutrophil-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR), platelet-to-lymphocyte ratio (PLR), systemic immune-inflammation index (SII), and prognostic nutritional index (PNI), were calculated based on laboratory data from the 3 days before surgery. Intra-class correlations coefficient (ICC) and least absolute shrinkage and selection operator (Lasso) were applied to screen valuable radiomics features predicting overall survival (OS), and the Rad-score was calculated. In the training cohort, univariate and multivariate Cox regression analyses identified independent prognostic factors, which were adopted to establish the nomogram.

RESULTS

Eight radiomic features were selected for Rad-score calculation. Multivariate Cox regression revealed Rad-score, PNI, NLR, and PLR as independent prognostic factors for ESCC patients (p < 0.05). A nomogram was constructed based on these variables. The concordance index (C-index) for the nomogram was 0.797 (95% CI: 0.726-0.868) in the training cohort and 0.796 (95% CI: 0.702-0.890) in the validation cohort. Calibration curves indicated good calibration ability, and the receiver operating characteristic (ROC) analysis demonstrated superior discriminative ability for the nomogram in comparison to the Rad-score alone. Decision curve analysis (DCA) confirmed the clinical utility of the nomogram.

CONCLUSION

We developed and validated a nomogram for predicting the OS of postoperative T3N0M0 ESCC patients, integrating nutritional, inflammatory markers, and radiomic signature. The combined nomogram can serve as a robust tool for risk stratification and clinical management.

Multi-omics deep learning for radiation pneumonitis prediction in lung cancer patients underwent volumetric modulated arc therapy

BACKGROUND AND OBJECTIVE

To evaluate the feasibility and accuracy of radiomics, dosiomics, and deep learning (DL) in predicting Radiation Pneumonitis (RP) in lung cancer patients underwent volumetric modulated arc therapy (VMAT) to improve radiotherapy safety and management.

METHODS

Total of 318 and 31 lung cancer patients underwent VMAT from First Affiliated Hospital of Wenzhou Medical University (WMU) and Quzhou Affiliated Hospital of WMU were enrolled for training and external validation, respectively. Models based on radiomics (R), dosiomics (D), and combined radiomics and dosiomics features (R+D) were constructed and validated using three machine learning (ML) methods. DL models trained with CT (DLR), dose distribution (DLD), and combined CT and dose distribution (DL(R+D)) images were constructed. DL features were then extracted from the fully connected layers of the best-performing DL model to combine with features of the ML model with the best performance to construct models of R+DLR, D+DLD, R+D+DL(R+D)) for RP prediction.

RESULTS

The R+D model achieved a best area under curve (AUC) of 0.84, 0.73, and 0.73 in the internal validation cohorts with Support Vector Machine (SVM), XGBoost, and Logistic Regression (LR), respectively. The DL(R+D) model achieved a best AUC of 0.89 and 0.86 using ResNet-34 in training and internal validation cohorts, respectively. The R+D+DL(R+D) model achieved a best performance in the external validation cohorts with an AUC, accuracy, sensitivity, and specificity of 0.81(0.62-0.99), 0.81, 0.84, and 0.67, respectively.

CONCLUSIONS

The integration of radiomics, dosiomics, and DL features is feasible and accurate for the RP prediction to improve the management of lung cancer patients underwent VMAT.

Does Neoadjuvant Radiation Therapy Contribute to the Incidence of Pulmonary Complications Following Esophagectomy for Malignant Neoplasm?

BACKGROUND

Post-operative pulmonary complications (POPC) are common in patients undergoing esophagectomy and neoadjuvant radiotherapy may exacerbate POPC. This study assessed whether neoadjuvant radiation increases the incidence of POPC in patients undergoing esophagectomy for malignancy.

METHODS

The American College of Surgeons-National Surgical Quality Improvement Program database files from 2016 to 2018 were queried for patients undergoing esophagectomy for malignancy. Inverse probability treatment weighting (IPTW) was used to create balanced cohorts in which the control group received neoadjuvant chemotherapy (nCT) and the treatment cohort received neoadjuvant chemoradiotherapy (nCRT). A subset analysis was performed on patients with pre-existing pulmonary disease (PEPD). Primary outcomes were POPC and 30-day mortality.

RESULTS

The all-patient analysis did not demonstrate a consistent association between neoadjuvant radiation and POPC. However, in patients with PEPD, POPC occurred more often in the nCRT cohort. Comparing nCRT to nCT and after IPTW adjustment for confounders, there was higher odds of pneumonia (aOR = 3.0, P = .002), unplanned intubation (aOR = 2.0, P = .03), and extended mechanical ventilation (aOR = 3.6, P = .002).

DISCUSSION

In esophageal cancer patients with PEPD that undergo nCRT vs nCT prior to esophagectomy, the greater risk of POPCs must be weighed against the potential for improved oncologic outcomes.

Radiation pneumonia predictive model for radiotherapy in esophageal carcinoma patients

BACKGROUND

The machine learning models with dose factors and the deep learning models with dose distribution matrix have been used to building lung toxics models for radiotherapy and achieve promising results. However, few studies have integrated clinical features into deep learning models. This study aimed to explore the role of three-dimension dose distribution and clinical features in predicting radiation pneumonitis (RP) in esophageal cancer patients after radiotherapy and designed a new hybrid deep learning network to predict the incidence of RP.

METHODS

A total of 105 esophageal cancer patients previously treated with radiotherapy were enrolled in this study. The three-dimension (3D) dose distributions within the lung were extracted from the treatment planning system, converted into 3D matrixes and used as inputs to predict RP with ResNet. In total, 15 clinical factors were normalized and converted into one-dimension (1D) matrixes. A new prediction model (HybridNet) was then built based on a hybrid deep learning network, which combined 3D ResNet18 and 1D convolution layers. Machine learning-based prediction models, which use the traditional dosiomic factors with and without the clinical factors as inputs, were also constructed and their predictive performance compared with that of HybridNet using tenfold cross validation. Accuracy and area under the receiver operator characteristic curve (AUC) were used to evaluate the model effect. DeLong test was used to compare the prediction results of the models.

RESULTS

The deep learning-based model achieved superior prediction results compared with machine learning-based models. ResNet performed best in the group that only considered dose factors (accuracy, 0.78 ± 0.05; AUC, 0.82 ± 0.25), whereas HybridNet performed best in the group that considered both dose factors and clinical factors (accuracy, 0.85 ± 0.13; AUC, 0.91 ± 0.09). HybridNet had higher accuracy than that of Resnet (p = 0.009).

CONCLUSION

Based on prediction results, the proposed HybridNet model could predict RP in esophageal cancer patients after radiotherapy with significantly higher accuracy, suggesting its potential as a useful tool for clinical decision-making. This study demonstrated that the information in dose distribution is worth further exploration, and combining multiple types of features contributes to predict radiotherapy response.

Predictors of lung injury during durvalumab maintenance therapy following concurrent chemoradiotherapy in unresectable locally advanced non-small cell lung carcinoma

BACKGROUND

Based on the results of the PACIFIC trial, maintenance with durvalumab has emerged as the standard treatment following concurrent chemoradiotherapy in patients with unresectable locally advanced non-small cell lung carcinoma (NSCLC). However, adverse events attributed to durvalumab, especially lung injuries, including immune-related adverse events, and radiation pneumonitis, are concerning. This study retrospectively investigated the factors related to lung injury in patients receiving the PACIFIC regimen.

METHODS

Patients with unresectable locally advanced NSCLC who received durvalumab maintenance therapy following concurrent chemoradiotherapy at Yokohama City University Medical Centre between July 2018 and March 2022 were included. Clinical data, volume of normal lung receiving 20 or 5 Gy or more (V20 or V5), planning target volume (PTV), and relative lung parenchyma volume in emphysematous lung receiving 20 or 5 Gy or more (RLPV20 or 5; V20 or V5/100-percentage of low-attenuation volume) were evaluated.

RESULTS

Performance status (PS), V20, V5, PTV, RLPV20, and RLPV5 were significantly higher in the lung injury group in the univariate analysis. Furthermore, RLPV20 was the most significant factor in the lung injury group in the multivariate analysis comprising PS, PTV, V20, and RLPV20.

CONCLUSION

RLPV20 and RLPV5 are useful in estimating lung inflammation. RLPV20 could be considered the most reliable risk factor for maintenance therapy with durvalumab following concurrent chemoradiotherapy in patients with unresectable locally advanced NSCLC.

Influencing factors and prediction methods of radiotherapy and chemotherapy in patients with lung cancer based on logistic regression analysis

Logistic regression analysis has widespread applications in clinical disease diagnosis, but it has not yet been applied to assess the acceptance of radiotherapy and chemotherapy in patients with lung cancer. A prediction model was established to investigate the influencing factors of radiotherapy and chemotherapy in lung cancer patients in order to provide useful information for clinicians to develop targeted and effective treatment. A sample was admitted of lung cancer patients to Binzhou Medical University Hospital stays from January 2020 to June 2021. After investigating doctors, nurses, patients, managers and conducting expert demonstration, the questionnaire was formed. The questionnaire was filled out by the patient or the patient's family members. The factors in the questionnaire data of patients accepting and not accepting radiotherapy and chemotherapy were compared for univariate analysis, and the significantly different single factor were analyzed by multifactor logistic regression analysis, explored the influencing factors of radiotherapy and chemotherapy in lung cancer patients established a predictive model and drew the receiver operating characteristic curve (ROC curve). The factors of two groups had statistically significant differences or no statistically significant differences. After multifactor logistic regression analysis was conducted, own personality, self-care ability, disease course classification, own attitude towards disease treatment, and family attitude towards disease treatment were included in the influencing factors of radiotherapy and chemotherapy in patients with lung cancer. Then, a predictive model was established. The area under the ROC curve of the predicted model was 0.973, the 95% confidence interval was 0.952-0.995, the optimal critical value was 0.832, the sensitivity was 91.84%, the specificity was 89.09%, and the accuracy was 90.85%. Based on logistic regression analysis, the prediction model could predict the extent of accepting radiotherapy and chemotherapy in patients with lung cancer. Understanding the factors related to patients with lung cancer accepting radiotherapy and chemotherapy could provide useful information for the targeted and effective treatment by clinicians.

Construction and validation of a novel prognostic model for thyroid cancer based on N7-methylguanosine modification-related lncRNAs

BACKGROUND

To construct and verify a novel prognostic model for thyroid cancer (THCA) based on N7-methylguanosine modification-related lncRNAs (m7G-lncRNAs) and their association with immune cell infiltration.

METHODS

In this study, we identified m7G-lncRNAs using co-expression analysis and performed differential expression analysis of m7G-lncRNAs between groups. We then constructed a THCA prognostic model, performed survival analysis and risk assessment for the THCA prognostic model, and performed independent prognostic analysis and receiver operating characteristic curve analyses to evaluate and validate the prognostic value of the model. Furthermore, analysis of the regulatory relationship between prognostic differentially expressed m7G-related lncRNAs (PDEm7G-lncRNAs) and mRNAs and correlation analysis of immune cells and risk scores in THCA patients were carried out.

RESULTS

We identified 29 N7-methylguanosine modification-related mRNAs and 116 differentially expressed m7G-related lncRNAs, including 87 downregulated and 29 upregulated lncRNAs. Next, we obtained 8 PDEm7G-lncRNAs. A final optimized model was constructed consisting of 5 PDEm7G-lncRNAs (DOCK9-DT, DPP4-DT, TMEM105, SMG7-AS1 and HMGA2-AS1). Six PDEm7G-lncRNAs (DOCK9-DT, DPP4-DT, HMGA2-AS1, LINC01976, MID1IP1-AS1, and SMG7-AS1) had positive regulatory relationships with 10 PDEm7G-mRNAs, while 2 PDEm7G-lncRNAs (LINC02026 and TMEM105) had negative regulatory relationships with 2 PDEm7G-mRNAs. Survival curves and risk assessment predicted the prognostic risk in both groups of patients with THCA. Forest maps and receiver operating characteristic curves were used to evaluate and validate the prognostic value of the model. Finally, we demonstrated a correlation between different immune cells and risk scores.

CONCLUSION

Our results will help identify high-risk or low-risk patients with THCA and facilitate early prediction and clinical intervention in patients with high risk and poor prognosis.

Nomogram based on circulating lymphocyte subsets for predicting radiation pneumonia in esophageal squamous cell carcinoma

Purpose

Currently, the relationship between radiation pneumonia (RP) and circulating immune cell in patients with esophageal squamous cell carcinoma (ESCC) remains unclear. This study aimed to explore the relationship between RP and circulating lymphocyte subsets in patients with ESCC receiving chemoradiotherapy (CRT), and develop a nomogram model to predict RP. Since we should implement clinical intervention to ≥ grade 2 RP, a nomogram model for ≥ grade 2 RP was also established to provide an early warning.

Patients and methods

This study retrospectively included 121 patients with ESCC receiving CRT from Guangxi Medical University Cancer Hospital from 2013 to 2021. Independent factors associated with occurrence of RP and ≥ grade 2 RP were identified by univariate and multivariate logistic regression analysis in the training cohort, and incorporated into nomograms. The predictive accuracy and discrimination of the model was assessed using Concordance Index (C-index), calibration curve and decision curve analysis (DCA). And each model was internally validated. Additionally, to verify the optimized predictive performance of the nomograms, the area under the ROC curve (AUC) of each nomogram was compared to that of single independent risk factors, lung V10 and lung V20, respectively. Moreover, each model was further evaluated for risk stratification to identify populations at high risk of RP and ≥ grade 2 RP.

Results

Multivariate analysis suggested that TNM stage, post-RT percentage of CD8+ T cell, and lung V15 were independent predictive factors of RP. Besides, pre- and post-RT percentage of CD8+ T cell, and V15 were independent factors of ≥ grade 2 RP. The C-indexes of RP and ≥ grade 2 RP nomograms were 0.809 (95% CI: 0.715-0.903) and 0.787 (95% CI: 0.685-0.889) in the training cohort, respectively. And the C-indexes of RP and ≥ grade 2 RP nomograms were 0.718 (95% CI: 0.544-0.892) and 0.621 (95% CI: 0.404-0.837) in the validation cohort, respectively. The calibration curves showed that the predicted values of model agreed well with actual observations. Moreover, DCA results indicated the applicability and accuracy of the models to predict RP and ≥ grade 2 RP. After stratification, the incidence of the high-risk group was significantly higher than that of the low-risk group with respect to either RP or ≥ grade 2 RP.

Conclusion

TNM stage, post-RT percentage of CD8+ T cell, and lung V15 were the independent predictors of RP toxicity. Pre- and post-RT percentage of CD8+ T cell, and lung V15 were the independent factors of ≥ grade 2 RP toxicity. The nomograms based on circulating lymphocyte subsets can robustly predict RP and ≥ grade 2 RP, guiding clinicians in risk stratification and early intervention.

Construction of Diagnosis Model of Moyamoya Disease Based on Convolution Neural Network Algorithm

Objective

The convolutional neural network (CNN) was used to improve the accuracy of digital subtraction angiography (DSA) in diagnosing moyamoya disease (MMD), providing a new method for clinical diagnosis of MMD.

Methods

A total of 40 diagnosed with MMD by DSA in the neurosurgery department of our hospital were included. At the same time, 40 age-matched and sex-matched patients were selected as the control group. The 80 included patients were divided into training set (n = 56) and validation set (n = 24). The DSA image was preprocessed, and the CNN was used to extract features from the preprocessed image. The precision and accuracy of the preprocessed image results were evaluated.

Results

There was no significant difference in baseline data between the training set and validation set (P > 0.05). The precision and accuracy of the images before processing were 79.68% and 81.45%, respectively. After image processing, the precision and accuracy of the model are 96.38% and 97.59%, respectively. The area under the curve of the CNN algorithm model was 0.813 (95% CI: 0.718-0.826).

Conclusion

This diagnostic method based on CNN performs well in MMD detection.

Prediction Model of Bone Marrow Infiltration in Patients with Malignant Lymphoma Based on Logistic Regression and XGBoost Algorithm

Objective

The prediction model of bone marrow infiltration (BMI) in patients with malignant lymphoma (ML) was established based on the logistic regression and the XGBoost algorithm. The model's prediction efficiency was evaluated.

Methods

A total of 120 patients diagnosed with ML in the department of hematology from January 2018 to January 2021 were retrospectively selected. The training set (n = 84) and test set (n = 36) were randomly divided into 7 : 3, and logistic regression and XGBoost algorithm models were constructed using the training set data. Predictors of BMI were screened based on laboratory indicators, and the model's efficacy was evaluated using test set data.

Results

The prediction algorithm model's top three essential characteristics are the blood platelet count, soluble interleukin-2 receptor, and non-Hodgkin's lymphoma. The area under the curve of the logistic regression model for predicting the BMI of patients with ML was 0.843 (95% CI: 0.761~0.926). The area under the curve of the XGBoost model is 0.844 (95% CI: 0.765~0.937).

Conclusion

The prediction model constructed in this study based on logistic regression and XGBoost algorithm has a good prediction model. The results showed that blood platelet count and soluble interleukin-2 receptor were good predictors of BMI in ML patients.

Survival Risk Prediction of Esophageal Cancer Based on the Kohonen Network Clustering Algorithm and Kernel Extreme Learning Machine

Accurate prediction of the survival risk level of patients with esophageal cancer is significant for the selection of appropriate treatment methods. It contributes to improving the living quality and survival chance of patients. However, considering that the characteristics of blood index vary with individuals on the basis of their ages, personal habits and living environment etc., a unified artificial intelligence prediction model is not precisely adequate. In order to enhance the precision of the model on the prediction of esophageal cancer survival risk, this study proposes a different model based on the Kohonen network clustering algorithm and the kernel extreme learning machine (KELM), aiming to classifying the tested population into five catergories and provide better efficiency with the use of machine learning. Firstly, the Kohonen network clustering method was used to cluster the patient samples and five types of samples were obtained. Secondly, patients were divided into two risk levels based on 5-year net survival. Then, the Taylor formula was used to expand the theory to analyze the influence of different activation functions on the KELM modeling effect, and conduct experimental verification. RBF was selected as the activation function of the KELM. Finally, the adaptive mutation sparrow search algorithm (AMSSA) was used to optimize the model parameters. The experimental results were compared with the methods of the artificial bee colony optimized support vector machine (ABC-SVM), the three layers of random forest (TLRF), the gray relational analysis–particle swarm optimization support vector machine (GP-SVM) and the mixed-effects Cox model (Cox-LMM). The results showed that the prediction model proposed in this study had certain advantages in terms of prediction accuracy and running time, and could provide support for medical personnel to choose the treatment mode of esophageal cancer patients.

Construction of Prediction Model of Deep Vein Thrombosis Risk after Total Knee Arthroplasty Based on XGBoost Algorithm

Objective

Based on the XGBoost algorithm, the prediction model of the risk of deep vein thrombosis (DVT) in patients after total knee arthroplasty (TKA) was established, and the prediction performance was compared.

Methods

A total of 100 patients with TKA from January 2019 to December 2020 were retrospectively selected as the study subjects and randomly divided into a training set (n = 60) and a test set (n = 40). The training set data was used to construct the XGBoost algorithm prediction model and to screen the predictive factors of postoperative DVT in TKA patients. The prediction effect of the model was evaluated by using the test set data. An independent sample T-test was used for comparison between groups, and the χ² test was used for comparison between counting data groups.

Results

The top five items were combined with multiple injuries (35 points), time from injury to operation (28 points), age (24 points), combined with coronary heart disease (21 points), and D-dimer 1 day after operation (16 points). In the training set, the area under the curve of the XGBoost algorithm model was 0.832 (95% CI: 0.748-0.916).

Conclusion

The model based on the XGBoost algorithm can predict the incidence of DVT in patients after TKA with good performance.