A neural network model to predict lung radiation-induced pneumonitis

Optimizing lung cancer radiation therapy: A systematic review of multifactorial risk assessment for radiation-induced lung toxicity

BACKGROUND

Radiation therapy (RT) is essential in treating advanced lung cancer, but may lead to radiation pneumonitis (RP). This systematic review investigates the use of pulmonary function tests (PFT) and other parameters to predict and mitigate RP, thereby improving RT planning.

METHODS

A systematic review sifted through PubMed and on BioMed Central, targeting articles from September 2005 to December 2022 containing the keywords: Lung Cancer, Radiotherapy, and pulmonary function test.

RESULTS

From 1153 articles, 80 were included. RP was assessed using CTCAEv.4 in 30 % of these. Six studies evaluated post-RT quality of life in lung cancer patients, reporting no decline. Patients with RP and chronic obstructive pulmonary disease (COPD) generally exhibited poorer overall survival. Notably, forced expiratory volume in one second (FEV1) and diffusing capacity of the lung for carbon monoxide (DLCO) declined 24 months post-RT, while forced vital capacity (FVC) stayed stable. In the majority of studies, age over 60, tumors located in the lower part of the lung, and low FEV1 before RT were associated with a higher risk of RP. Dosimetric factors (V5, V20, MLD) and metabolic imaging emerged as significant predictors of RP risk. A clinical checklist blending patient and tumor characteristics, PFT results, and dosimetric criteria was proposed for assessing RP risk before RT.

CONCLUSION

The review reveals the multifactorial nature of RP development following RT in lung cancer. This approach should guide individualized management and calls for a prospective study to validate these findings and enhance RP prevention strategies.

Prediction of toxicity outcomes following radiotherapy using deep learning-based models: A systematic review

PURPOSE

This study aims to perform a comprehensive systematic review of deep learning (DL) models in predicting RT-induced toxicity.

MATERIALS AND METHODS

A literature review was performed following the PRISMA guidelines. Systematic searches were performed in PubMed, Scopus, Cochrane and Embase databases from the earliest record up to September 2022. Related studies on deep learning models for radiotherapy toxicity prediction were selected based on predefined PICOS criteria.

RESULTS

Fourteen studies of radiotherapy-treated patients on different types of cancer [prostate (n=2), HNC (n=4), liver (n=2), lung (n=4), cervical (n=1), and oesophagus (n=1)] were eligible for inclusion in the systematic review. Information regarding patient characteristics and model development was summarized. Several approaches, such as ensemble learning, data augmentation, and transfer learning, that were utilized by selected studies were discussed.

CONCLUSION

Deep learning techniques are able to produce a consistent performance for toxicity prediction. Future research using large and diverse datasets and standardization of the study methodologies are required to improve the consistency of the research output.

Predicting pulmonary ventilation damage after radiation therapy for nonsmall cell lung cancer using a ResNet generative adversarial network

BACKGROUND

Functional lung avoidance radiation therapy (RT) is a technique being investigated to preferentially avoid specific regions of the lung that are predicted to be more susceptible to radiation-induced damage. Reducing the dose delivered to high functioning regions may reduce the occurrence radiation-induced lung injuries (RILIs) and toxicities. However, in order to develop effective lung function-sparing plans, accurate predictions of post-RT ventilation change are needed to determine which regions of the lung should be spared.

PURPOSE

To predict pulmonary ventilation change following RT for nonsmall cell lung cancer using machine learning.

METHODS

A conditional generative adversarial network (cGAN) was developed with data from 82 human subjects enrolled in a randomized clinical trial approved by the institution's IRB to predict post-RT pulmonary ventilation change. The inputs to the network were the pre-RT pulmonary ventilation map and radiation dose distribution. The loss function was a combination of the binary cross-entropy loss and an asymmetrical structural similarity index measure (aSSIM) function designed to increase penalization of under-prediction of ventilation damage. Network performance was evaluated against a previously developed polynomial regression model using a paired sample t-test for comparison. Evaluation was performed using eight-fold cross-validation.

RESULTS

From the eight-fold cross-validation, we found that relative to the polynomial model, the cGAN model significantly improved predicting regions of ventilation damage following radiotherapy based on true positive rate (TPR), 0.14±0.15 to 0.72±0.21, and Dice similarity coefficient (DSC), 0.19±0.16 to 0.46±0.14, but significantly declined in true negative rate, 0.97±0.05 to 0.62±0.21, and accuracy, 0.79±0.08 to 0.65±0.14. Additionally, the average true positive volume increased from 104±119 cc in the POLY model to 565±332 cc in the cGAN model, and the average false negative volume decreased from 654±361 cc in the POLY model to 193±163 cc in the cGAN model.

CONCLUSIONS

The proposed cGAN model demonstrated significant improvement in TPR and DSC. The higher sensitivity of the cGAN model can improve the clinical utility of functional lung avoidance RT by identifying larger volumes of functional lung that can be spared and thus decrease the probability of the patient developing RILIs.

Mechanistic Understanding of Lung Inflammation: Recent Advances and Emerging Techniques

Acute respiratory distress syndrome (ARDS) is a life-threatening lung injury characterized by an acute inflammatory response in the lung parenchyma. Hence, it is considered as the most appropriate clinical syndrome to study pathogenic mechanisms of lung inflammation. ARDS is associated with increased morbidity and mortality in the intensive care unit (ICU), while no effective pharmacological treatment exists. It is very important therefore to fully characterize the underlying pathobiology and the related mechanisms, in order to develop novel therapeutic approaches. In vivo and in vitro models are important pre-clinical tools in biological and medical research in the mechanistic and pathological understanding of the majority of diseases. In this review, we will present data from selected experimental models of lung injury/acute lung inflammation, which have been based on clinical disorders that can lead to the development of ARDS and related inflammatory lung processes in humans, including ventilation-induced lung injury (VILI), sepsis, ischemia/reperfusion, smoke, acid aspiration, radiation, transfusion-related acute lung injury (TRALI), influenza, Streptococcus (S.) pneumoniae and coronaviruses infection. Data from the corresponding clinical conditions will also be presented. The mechanisms related to lung inflammation that will be covered are oxidative stress, neutrophil extracellular traps, mitogen-activated protein kinase (MAPK) pathways, surfactant, and water and ion channels. Finally, we will present a brief overview of emerging techniques in the field of omics research that have been applied to ARDS research, encompassing genomics, transcriptomics, proteomics, and metabolomics, which may recognize factors to help stratify ICU patients at risk, predict their prognosis, and possibly, serve as more specific therapeutic targets.

Predicting radiation pneumonitis with fuzzy clustering neural network using 4DCT ventilation image based dosimetric parameters

Background

To develop a fuzzy clustering neural network to predict radiation-induced pneumonitis (RP) using four-dimensional computed tomography (4DCT) ventilation image (VI) based dosimetric parameters for thoracic cancer patients.

Methods

The VI were retrospectively calculated from pre-treatment 4DCT data using a deformable image registration (DIR) and an improved VI algorithm. Similar to dose-volume histogram (DVH) of intensity modulated radiotherapy (IMRT), dose-function histogram (DFH) was derived from dose distribution and VI. Then, the dose-function metrics were calculated from DFH. For comparison, the dose-volume metrics were calculated from DVH. Correspondingly, two sets of feature vectors were formed from the dose-volume metrics and the dose-function metrics, respectively. For the feature vectors of each set, they were first pre-processed by principal component analysis (PCA) to reduce feature dimensions. Then, they were grouped to few clusters determined by fuzzy c-means (FCM) algorithm. Next, the neural network was trained to correlate the dosimetric parameters with RP based on the feature vectors of each cluster. Finally, the occurrence of RP was predicted by the neural network on the test data.

Results

Through PCA analysis, the top 5 principal components were selected. Their contribution is more than 98%, which is adequate to represent the original feature space of input data. Based on the clustering validity indexes, the optimal number of clusters is 4 and used for subsequent fuzzy clustering of the input data. After network training, the areas under the curve (AUC) of the prediction model is 0.77 using VI-based dosimetric parameters and 0.67 using structure-based dosimetric parameters.

Conclusions

Compared to the structure-based dosimetric features, the VI-based dosimetric features are more relevant to lung function and presented higher prediction accuracy of RP. The fuzzy clustering neural network improved the prediction accuracy of RP compared to the conventional neural network. The combination of VI-based dose-function metrics and fuzzy clustering neural network provides an effective predictive model for assessing lung toxicity risk after radiotherapy.

A review of deep learning based methods for medical image multi-organ segmentation

Deep learning has revolutionized image processing and achieved the-state-of-art performance in many medical image segmentation tasks. Many deep learning-based methods have been published to segment different parts of the body for different medical applications. It is necessary to summarize the current state of development for deep learning in the field of medical image segmentation. In this paper, we aim to provide a comprehensive review with a focus on multi-organ image segmentation, which is crucial for radiotherapy where the tumor and organs-at-risk need to be contoured for treatment planning. We grouped the surveyed methods into two broad categories which are 'pixel-wise classification' and 'end-to-end segmentation'. Each category was divided into subgroups according to their network design. For each type, we listed the surveyed works, highlighted important contributions and identified specific challenges. Following the detailed review, we discussed the achievements, shortcomings and future potentials of each category. To enable direct comparison, we listed the performance of the surveyed works that used thoracic and head-and-neck benchmark datasets.

Radiomic and radiogenomic modeling for radiotherapy: strategies, pitfalls, and challenges

The power of predictive modeling for radiotherapy outcomes has historically been limited by an inability to adequately capture patient-specific variabilities; however, next-generation platforms together with imaging technologies and powerful bioinformatic tools have facilitated strategies and provided optimism. Integrating clinical, biological, imaging, and treatment-specific data for more accurate prediction of tumor control probabilities or risk of radiation-induced side effects are high-dimensional problems whose solutions could have widespread benefits to a diverse patient population-we discuss technical approaches toward this objective. Increasing interest in the above is specifically reflected by the emergence of two nascent fields, which are distinct but complementary: radiogenomics, which broadly seeks to integrate biological risk factors together with treatment and diagnostic information to generate individualized patient risk profiles, and radiomics, which further leverages large-scale imaging correlates and extracted features for the same purpose. We review classical analytical and data-driven approaches for outcomes prediction that serve as antecedents to both radiomic and radiogenomic strategies. Discussion then focuses on uses of conventional and deep machine learning in radiomics. We further consider promising strategies for the harmonization of high-dimensional, heterogeneous multiomics datasets (panomics) and techniques for nonparametric validation of best-fit models. Strategies to overcome common pitfalls that are unique to data-intensive radiomics are also discussed.

Acute severe radiation pneumonitis among non-small cell lung cancer (NSCLC) patients with moderate pulmonary dysfunction receiving definitive concurrent chemoradiotherapy: Impact of pre-treatment pulmonary function parameters

PURPOSE

Severe acute radiation pneumonitis (SARP) is a life-threatening complication of thoracic radiotherapy. Pre-treatment pulmonary function (PF) may influence its incidence. We have previously reported on the incidence of SARP among patients with moderate pulmonary dysfunction who received definitive concurrent chemoradiotherapy (dCCRT) for non-small cell lung cancer (NSCLC).

METHODS

The clinical outcomes, dose-volume histograms (DVH), and PF parameters of 122 patients (forced expiratory volume in 1 s [FEV1%]: 60-69%) receiving dCCRT between 2013 and 2019 were recorded. SARP was defined as grade ≥3 RP occurring during or within 3 months after CCRT. Logistic regression, receiver operating characteristics curves (ROC), and hazard ratio (HR) analyses were performed to evaluate the predictive value of each factor for SARP.

RESULTS

Univariate and multivariate analysis indicated that the ratio of carbon monoxide diffusing capacity (DLCO%; odds ratio [OR]: 0.934, 95% confidence interval [CI] 0.896-0.974, p = 0.001) and mean lung dose (MLD; OR: 1.002, 95% CI 1.001-1.003, p = 0.002) were independent predictors of SARP. The ROC AUC of combined DLCO%/MLD was 0.775 (95% confidence interval [CI]: 0.688-0.861, p = 0.001), with a sensitivity and specificity of 0.871 and 0.637, respectively; this was superior to DLCO% (0.656) or MLD (0.667) alone. Compared to the MLD-low/DLCO%-high group, the MLD-high/DLCO%-low group had the highest risk for SARP, with an HR of 9.346 (95% CI: 2.133-40.941, p = 0.003).

CONCLUSION

The DLCO% and MLD may predict the risk for SARP among patients with pre-treatment moderate pulmonary dysfunction who receive dCCRT for NSCLC. Prospective studies are needed to validate our findings.

Deep Learning: A Review for the Radiation Oncologist

Introduction: Deep Learning (DL) is a machine learning technique that uses deep neural networks to create a model. The application areas of deep learning in radiation oncology include image segmentation and detection, image phenotyping, and radiomic signature discovery, clinical outcome prediction, image dose quantification, dose-response modeling, radiation adaptation, and image generation. In this review, we explain the methods used in DL and perform a literature review using the Medline database to identify studies using deep learning in radiation oncology. The search was conducted in April 2018, and identified studies published between 1997 and 2018, strongly skewed toward 2015 and later. Methods: A literature review was performed using PubMed/Medline in order to identify important recent publications to be synthesized into a review of the current status of Deep Learning in radiation oncology, directed at a clinically-oriented reader. The search strategy included the search terms "radiotherapy" and "deep learning." In addition, reference lists of selected articles were hand searched for further potential hits of relevance to this review. The search was conducted in April 2018, and identified studies published between 1997 and 2018, strongly skewed toward 2015 and later. Results: Studies using DL for image segmentation were identified in Brain (n = 2), Head and Neck (n = 3), Lung (n = 6), Abdominal (n = 2), and Pelvic (n = 6) cancers. Use of Deep Learning has also been reported for outcome prediction, such as toxicity modeling (n = 3), treatment response and survival (n = 2), or treatment planning (n = 5). Conclusion: Over the past few years, there has been a significant number of studies assessing the performance of DL techniques in radiation oncology. They demonstrate how DL-based systems can aid clinicians in their daily work, be it by reducing the time required for or the variability in segmentation, or by helping to predict treatment outcomes and toxicities. It still remains to be seen when these techniques will be employed in routine clinical practice.

Use of artificial neural network for pretreatment verification of intensity modulation radiation therapy fields

OBJECTIVE

The accuracy of dose delivery for intensity modulated radiotherapy (IMRT) treatments should be determined by an accurate quality assurance procedure. In this work, we used artificial neural networks (ANNs) as an application for the pre-treatment dose verification of IMRT fields based two-dimensional-fluence maps acquired by an electronic portal imaging device (EPID).

METHODS

The ANN must be trained and validated before use for the pretreatment dose verification. Hence, 60 EPID fluence maps of the anteroposterior prostate and nasopharynx IMRT fields were used as an input for the ANN (feed forward type), and a dose map of those fluence maps that were acquired by two-dimensional Array Seven29TM as an output for the ANN.

RESULTS

After the training and validation of the neural network, the analysis of 20 IMRT anteroposterior fields showed excellent agreement between the ANN output and the dose map predicted by the treatment planning system. The average overall global and local γ field pass rate was greater than 90% for the prostate and nasopharynx fields, with the 2 mm/3% criteria.

CONCLUSION

The results indicated that the ANN can be used as a fast and powerful tool for pretreatment dose verification, based on an EPID fluence map.

ADVANCES IN KNOWLEDGE

In this study, ANN is proposed for EPID based dose validation of IMRT fields. The proposed method has good accuracy and high speed in response to problems. Neural network show to be low price and precise method for IMRT fields verification.

Radiation-Induced Lung Injury (RILI)

Radiation pneumonitis (RP) and radiation fibrosis (RF) are two dose-limiting toxicities of radiotherapy (RT), especially for lung, and esophageal cancer. It occurs in 5-20% of patients and limits the maximum dose that can be delivered, reducing tumor control probability (TCP) and may lead to dyspnea, lung fibrosis, and impaired quality of life. Both physical and biological factors determine the normal tissue complication probability (NTCP) by Radiotherapy. A better understanding of the pathophysiological sequence of radiation-induced lung injury (RILI) and the intrinsic, environmental and treatment-related factors may aid in the prevention, and better management of radiation-induced lung damage. In this review, we summarize our current understanding of the pathological and molecular consequences of lung exposure to ionizing radiation, and pharmaceutical interventions that may be beneficial in the prevention or curtailment of RILI, and therefore enable a more durable therapeutic tumor response.

Forecasting the seasonality and trend of pulmonary tuberculosis in Jiangsu Province of China using advanced statistical time-series analyses

Objective

Forecasting the seasonality and trend of pulmonary tuberculosis is important for the rational allocation of health resources; however, this foresting is often hampered by inappropriate prediction methods. In this study, we performed validation research by comparing the accuracy of the autoregressive integrated moving average (ARIMA) model and the back-propagation neural network (BPNN) model in a southeastern province of China.

Methods

We applied the data from 462,214 notified pulmonary tuberculosis cases registered from January 2005 to December 2015 in Jiangsu Province to modulate and construct the ARIMA and BPNN models. Cases registered in 2016 were used to assess the prediction accuracy of the models. The root mean square error (RMSE), mean absolute percentage error (MAPE), mean absolute error (MAE) and mean error rate (MER) were used to evaluate the model fitting and forecasting effect.

Results

During 2005–2015, the annual pulmonary tuberculosis notification rate in Jiangsu Province was 56.35/100,000, ranging from 40.85/100,000 to 79.36/100,000. Through screening and comparison, the ARIMA (0, 1, 2) (0, 1, 1)₁₂ and BPNN (3-9-1) were defined as the optimal fitting models. In the fitting dataset, the RMSE, MAPE, MAE and MER were 0.3901, 6.0498, 0.2740 and 0.0608, respectively, for the ARIMA (0, 1, 2) (0, 1, 1)₁₂ model, 0.3236, 6.0113, 0.2508 and 0.0587, respectively, for the BPNN model. In the forecasting dataset, the RMSE, MAPE, MAE and MER were 0.1758, 4.6041, 0.1368 and 0.0444, respectively, for the ARIMA (0, 1, 2) (0, 1, 1)₁₂ model, and 0.1382, 3.2172, 0.1018 and 0.0330, respectively, for the BPNN model.

Conclusion

Both the ARIMA and BPNN models can be used to predict the seasonality and trend of pulmonary tuberculosis in the Chinese population, but the BPNN model shows better performance. Applying statistical techniques by considering local characteristics may enable more accurate mathematical modeling.

Combining handcrafted features with latent variables in machine learning for prediction of radiation-induced lung damage

PURPOSE

There has been burgeoning interest in applying machine learning methods for predicting radiotherapy outcomes. However, the imbalanced ratio of a large number of variables to a limited sample size in radiation oncology constitutes a major challenge. Therefore, dimensionality reduction methods can be a key to success. The study investigates and contrasts the application of traditional machine learning methods and deep learning approaches for outcome modeling in radiotherapy. In particular, new joint architectures based on variational autoencoder (VAE) for dimensionality reduction are presented and their application is demonstrated for the prediction of lung radiation pneumonitis (RP) from a large-scale heterogeneous dataset.

METHODS

A large-scale heterogeneous dataset containing a pool of 230 variables including clinical factors (e.g., dose, KPS, stage) and biomarkers (e.g., single nucleotide polymorphisms (SNPs), cytokines, and micro-RNAs) in a population of 106 nonsmall cell lung cancer (NSCLC) patients who received radiotherapy was used for modeling RP. Twenty-two patients had grade 2 or higher RP. Four methods were investigated, including feature selection (case A) and feature extraction (case B) with traditional machine learning methods, a VAE-MLP joint architecture (case C) with deep learning and lastly, the combination of feature selection and joint architecture (case D). For feature selection, Random forest (RF), Support Vector Machine (SVM), and multilayer perceptron (MLP) were implemented to select relevant features. Specifically, each method was run for multiple times to rank features within several cross-validated (CV) resampled sets. A collection of ranking lists were then aggregated by top 5% and Kemeny graph methods to identify the final ranking for prediction. A synthetic minority oversampling technique was applied to correct for class imbalance during this process. For deep learning, a VAE-MLP joint architecture where a VAE aimed for dimensionality reduction and an MLP aimed for classification was developed. In this architecture, reconstruction loss and prediction loss were combined into a single loss function to realize simultaneous training and weights were assigned to different classes to mitigate class imbalance. To evaluate the prediction performance and conduct comparisons, the area under receiver operating characteristic curves (AUCs) were performed for nested CVs for both handcrafted feature selections and the deep learning approach. The significance of differences in AUCs was assessed using the DeLong test of U-statistics.

RESULTS

An MLP-based method using weight pruning (WP) feature selection yielded the best performance among the different hand-crafted feature selection methods (case A), reaching an AUC of 0.804 (95% CI: 0.761-0.823) with 29 top features. A VAE-MLP joint architecture (case C) achieved a comparable but slightly lower AUC of 0.781 (95% CI: 0.737-0.808) with the size of latent dimension being 2. The combination of handcrafted features (case A) and latent representation (case D) achieved a significant AUC improvement of 0.831 (95% CI: 0.805-0.863) with 22 features (P-value = 0.000642 compared with handcrafted features only (Case A) and P-value = 0.000453 compared to VAE alone (Case C)) with an MLP classifier.

CONCLUSION

The potential for combination of traditional machine learning methods and deep learning VAE techniques has been demonstrated for dealing with limited datasets in modeling radiotherapy toxicities. Specifically, latent variables from a VAE-MLP joint architecture are able to complement handcrafted features for the prediction of RP and improve prediction over either method alone.

Artificial Neural Network with Composite Architectures for Prediction of Local Control in Radiotherapy

In this study, we investigated the application of artificial neural networks (ANNs) with composite architectures into the prediction of local control (LC) of lung cancer patients after radiotherapy. The motivation of this study was to take advantage of the temporal associations among longitudinal (sequential) data to improve the predictive performance of outcome models under the circumstance of limited sample sizes. Two composite architectures: (1) a one dimension (1D) convolutional + fully connected and (2) a locally-connected+ fully connected architectures were implemented for this purpose. Compared with the fully-connected architecture (multi-layer perceptron [MLP]), our composite architectures yielded better predictive performance of LC in lung cancer patients who received radiotherapy. Specifically, in a cohort of 98 patients (29 patients failed locally), the composite architecture of 1D convolutional layers and fully-connected layers achieved an AUC (area under receiver operating characteristic curve) of 0.83 (95% confidence interval (CI): 0.807~0.841) with 18 features (14 features are longitudinal data). Whereas, the composite architecture of locally- connected layers and fully-connected layers achieved an AUC of 0.80 (95%CI: 0.775~0.811). Both outperformed an MLP in the prediction performance with the same set of features, which achieved an AUC of 0.78 (95%CI: 0.751~0.790); (P-values for differences in AUC using the DeLong tests were 1.609 × 10^-14and 1.407 × 10^-4, respectively).

2D Dose Reconstruction by Artificial Neural Network for Pretreatment Verification of IMRT Fields

The use of intensity-modulated radiation therapy (IMRT) is developing rapidly in clinical routines. Because of the high complexity and uniqueness of IMRT treatment plans, patient-specific pretreatment quality assurance is generally considered a necessary prerequisite for patient treatment. In this work, we proposed a modified methodology of electronic portal imaging device (EPID)-based dose validation for pretreatment verification of IMRT fields by applying artificial neural networks (ANNs). The ANN must be trained and validated before use for pretreatment dose verification. For this purpose, 20 EPID fluence maps of IMRT prostate anterior-posterior fields were used as an input for ANN (feed forward type) and a dose map of those fluence maps that were predicted by treatment planning system as an output for ANN. After the training and validation of the neural network, the analysis of 10 IMRT prostate anterior-posterior fields showed excellent agreement between ANN output and dose map predicted by the treatment planning system. The average overall fields pass rate was 96.0% ± 0.1% with 3 mm/3% criteria. The results indicated that the ANN can be used as a low-cost, fast, and powerful tool for pretreatment dose verification, based on an EPID fluence map.

Abnormal pulmonary function tests predict the development of radiation-induced pneumonitis in advanced non-small cell lung Cancer

BACKGROUND

Radiation pneumonitis (RP) is a frequent complication of concurrent chemoradiotherapy (CCRT) and is associated with severe symptoms that decrease quality of life and might result in pulmonary fibrosis or death. The aim of this study is to identify whether pulmonary function test (PFT) abnormalities may predict RP in non-small cell lung cancer (NSCLC) patients.

METHODS

A prospective multi-institutional study was conducted with locally advanced and oligometastatic NSCLC patients. All participants were evaluated at baseline, end of CCRT, week 6, 12, 24, and 48 post-CCRT. They completed forced spirometry with a bronchodilator, body plethysmography, impulse oscillometry, carbon monoxide diffusing capacity (DLCO), molar mass of CO2, six-minute walk test and exhaled fraction of nitric oxide (FeNO). Radiation pneumonitis was assessed with RTOG and CTCAE. The protocol was registered in www.clinicaltrials.gov (NCT01580579), registered April 19, 2012.

RESULTS

Fifty-two patients were enrolled; 37 completed one-year follow-up. RP ≥ Grade 2 was present in 11/37 (29%) for RTOG and 15/37 (40%) for CTCAE. Factors associated with RP were age over 60 years and hypofractionated dose. PFT abnormalities at baseline that correlated with the development of RP included lower forced expiratory volume in one second after bronchodilator (p = 0.02), DLCO (p = 0.02) and FeNO (p = 0.04). All PFT results decreased after CCRT and did not return to basal values at follow-up.

CONCLUSIONS

FEV1, DLCO and FeNO prior to CCRT predict the development of RP in NSCLC. This study suggests that all patients under CCRT should be assessed by PFT to identify high-risk patients for close follow-up and early treatment.

Statistical-learning strategies generate only modestly performing predictive models for urinary symptoms following external beam radiotherapy of the prostate: A comparison of conventional and machine-learning methods

PURPOSE

Given the paucity of available data concerning radiotherapy-induced urinary toxicity, it is important to ensure derivation of the most robust models with superior predictive performance. This work explores multiple statistical-learning strategies for prediction of urinary symptoms following external beam radiotherapy of the prostate.

METHODS

The performance of logistic regression, elastic-net, support-vector machine, random forest, neural network, and multivariate adaptive regression splines (MARS) to predict urinary symptoms was analyzed using data from 754 participants accrued by TROG03.04-RADAR. Predictive features included dose-surface data, comorbidities, and medication-intake. Four symptoms were analyzed: dysuria, haematuria, incontinence, and frequency, each with three definitions (grade ≥ 1, grade ≥ 2 and longitudinal) with event rate between 2.3% and 76.1%. Repeated cross-validations producing matched models were implemented. A synthetic minority oversampling technique was utilized in endpoints with rare events. Parameter optimization was performed on the training data. Area under the receiver operating characteristic curve (AUROC) was used to compare performance using sample size to detect differences of ≥0.05 at the 95% confidence level.

RESULTS

Logistic regression, elastic-net, random forest, MARS, and support-vector machine were the highest-performing statistical-learning strategies in 3, 3, 3, 2, and 1 endpoints, respectively. Logistic regression, MARS, elastic-net, random forest, neural network, and support-vector machine were the best, or were not significantly worse than the best, in 7, 7, 5, 5, 3, and 1 endpoints. The best-performing statistical model was for dysuria grade ≥ 1 with AUROC ± standard deviation of 0.649 ± 0.074 using MARS. For longitudinal frequency and dysuria grade ≥ 1, all strategies produced AUROC>0.6 while all haematuria endpoints and longitudinal incontinence models produced AUROC<0.6.

CONCLUSIONS

Logistic regression and MARS were most likely to be the best-performing strategy for the prediction of urinary symptoms with elastic-net and random forest producing competitive results. The predictive power of the models was modest and endpoint-dependent. New features, including spatial dose maps, may be necessary to achieve better models.

Big Data Analytics for Prostate Radiotherapy

Radiation therapy is a first-line treatment option for localized prostate cancer and radiation-induced normal tissue damage are often the main limiting factor for modern radiotherapy regimens. Conversely, under-dosing of target volumes in an attempt to spare adjacent healthy tissues limits the likelihood of achieving local, long-term control. Thus, the ability to generate personalized data-driven risk profiles for radiotherapy outcomes would provide valuable prognostic information to help guide both clinicians and patients alike. Big data applied to radiation oncology promises to deliver better understanding of outcomes by harvesting and integrating heterogeneous data types, including patient-specific clinical parameters, treatment-related dose-volume metrics, and biological risk factors. When taken together, such variables make up the basis for a multi-dimensional space (the "RadoncSpace") in which the presented modeling techniques search in order to identify significant predictors. Herein, we review outcome modeling and big data-mining techniques for both tumor control and radiotherapy-induced normal tissue effects. We apply many of the presented modeling approaches onto a cohort of hypofractionated prostate cancer patients taking into account different data types and a large heterogeneous mix of physical and biological parameters. Cross-validation techniques are also reviewed for the refinement of the proposed framework architecture and checking individual model performance. We conclude by considering advanced modeling techniques that borrow concepts from big data analytics, such as machine learning and artificial intelligence, before discussing the potential future impact of systems radiobiology approaches.

Big Data and machine learning in radiation oncology: State of the art and future prospects

Precision medicine relies on an increasing amount of heterogeneous data. Advances in radiation oncology, through the use of CT Scan, dosimetry and imaging performed before each fraction, have generated a considerable flow of data that needs to be integrated. In the same time, Electronic Health Records now provide phenotypic profiles of large cohorts of patients that could be correlated to this information. In this review, we describe methods that could be used to create integrative predictive models in radiation oncology. Potential uses of machine learning methods such as support vector machine, artificial neural networks, and deep learning are also discussed.

Outcome modeling techniques for prostate cancer radiotherapy: Data, models, and validation

Prostate cancer is a frequently diagnosed malignancy worldwide and radiation therapy is a first-line approach in treating localized as well as locally advanced cases. The limiting factor in modern radiotherapy regimens is dose to normal structures, an excess of which can lead to aberrant radiation-induced toxicities. Conversely, dose reduction to spare adjacent normal structures risks underdosing target volumes and compromising local control. As a result, efforts aimed at predicting the effects of radiotherapy could invaluably optimize patient treatments by mitigating such toxicities and simultaneously maximizing biochemical control. In this work, we review the types of data, frameworks and techniques used for prostate radiotherapy outcome modeling. Consideration is given to clinical and dose-volume metrics, such as those amassed by the QUANTEC initiative, and also to newer methods for the integration of biological and genetic factors to improve prediction performance. We furthermore highlight trends in machine learning that may help to elucidate the complex pathophysiological mechanisms of tumor control and radiation-induced normal tissue side effects.

Machine Learning Approaches for Predicting Radiation Therapy Outcomes: A Clinician's Perspective

Radiation oncology has always been deeply rooted in modeling, from the early days of isoeffect curves to the contemporary Quantitative Analysis of Normal Tissue Effects in the Clinic (QUANTEC) initiative. In recent years, medical modeling for both prognostic and therapeutic purposes has exploded thanks to increasing availability of electronic data and genomics. One promising direction that medical modeling is moving toward is adopting the same machine learning methods used by companies such as Google and Facebook to combat disease. Broadly defined, machine learning is a branch of computer science that deals with making predictions from complex data through statistical models. These methods serve to uncover patterns in data and are actively used in areas such as speech recognition, handwriting recognition, face recognition, "spam" filtering (junk email), and targeted advertising. Although multiple radiation oncology research groups have shown the value of applied machine learning (ML), clinical adoption has been slow due to the high barrier to understanding these complex models by clinicians. Here, we present a review of the use of ML to predict radiation therapy outcomes from the clinician's point of view with the hope that it lowers the "barrier to entry" for those without formal training in ML. We begin by describing 7 principles that one should consider when evaluating (or creating) an ML model in radiation oncology. We next introduce 3 popular ML methods--logistic regression (LR), support vector machine (SVM), and artificial neural network (ANN)--and critique 3 seminal papers in the context of these principles. Although current studies are in exploratory stages, the overall methodology has progressively matured, and the field is ready for larger-scale further investigation.

Quantitative comparison of automatic and manual IMRT optimization for prostate cancer: the benefits of DVH prediction

A recent publication indicated that the patient anatomical feature (PAF) model was capable of predicting optimal objectives based on past experience. In this study, the benefits of IMRT optimization using PAF-predicted objectives as guidance for prostate were evaluated. Three different optimization methods were compared.1) Expert Plan: Ten prostate cases (16 plans) were planned by an expert planner using conventional trial-and-error approach started with institutional modified OAR and PTV constraints. Optimization was stopped at 150 iterations and that plan was saved as Expert Plan. 2) Clinical Plan: The planner would keep working on the Expert Plan till he was satisfied with the dosimetric quality and the final plan was referred to as Clinical Plan. 3) PAF Plan: A third sets of plans for the same ten patients were generated fully automatically using predicted DVHs as guidance. The optimization was based on PAF-based predicted objectives, and was continued to 150 iterations without human interaction. DMAX and D98% for PTV, DMAX for femoral heads, DMAX, D10cc, D25%/D17%, and D40% for bladder/rectum were compared. Clinical Plans are further optimized with more iterations and adjustments, but in general provided limited dosimetric benefits over Expert Plans. PTV D98% agreed within 2.31% among Expert, Clinical, and PAF plans. Between Clinical and PAF Plans, differences for DMAX of PTV, bladder, and rectum were within 2.65%, 2.46%, and 2.20%, respectively. Bladder D10cc was higher for PAF but < 1.54% in general. Bladder D25% and D40% were lower for PAF, by up to 7.71% and 6.81%, respectively. Rectum D10cc, D17%, and D40% were 2.11%, 2.72%, and 0.27% lower for PAF, respectively. DMAX for femoral heads were comparable (< 35 Gy on average). Compared to Clinical Plan (Primary + Boost), the average optimization time for PAF plan was reduced by 5.2 min on average, with a maximum reduction of 7.1min. Total numbers of MUs per plan for PAF Plans were lower than Clinical Plans, indicating better delivery efficiency. The PAF-guided planning process is capable of generating clinical-quality prostate IMRT plans with no human intervention. Compared to manual optimization, this automatic optimization increases planning and delivery efficiency, while maintainingplan quality.

Nondosimetric risk factors for radiation-induced lung toxicity

The decision to administer a radical course of radiotherapy (RT) is largely influenced by the dose-volume metrics of the treatment plan, but what are the patient-related and other factors that may independently increase the risk of radiation lung toxicity? Poor pulmonary function has been regarded as a risk factor and a relative contraindication for patients undergoing radical RT, but recent evidence suggests that patients with poor spirometry results may tolerate conventional or high-dose RT as well as, if not better than, patients with normal function. However, caution may need to be exercised in patients with underlying interstitial pulmonary fibrosis. Furthermore, there is emerging evidence of molecular markers of increased risk of toxicity. This review discusses patient-related risk factors other than dosimetry for radiation lung toxicity.

Identifying progression related disease risk modules based on the human subcellular signaling networks

Many studies have shown that the structure and dynamics of the human signaling network are disturbed in complex diseases such as coronary artery disease, and gene expression profiles can distinguish variations in diseases since they can accurately reflect the status of cells. Integration of subcellular localization and the human signaling network holds promise for providing insight into human diseases. In this study, we performed a novel algorithm to identify progression-related-disease-risk modules (PRDRMs) among patients of different disease states within eleven subcellular sub-networks from a human signaling network. The functional annotation and literature retrieval showed that the PRDRMs were strongly associated with disease pathogenesis. The results indicated that the PRDRM expression values as classification features had a good classification performance to distinguish patients of different disease states. Our approach compared with the method PageRank had a better classification performance. The identification of the PRDRMs in response to the dynamic gene expression change could facilitate our understanding of the pathological basis of complex diseases. Our strategy could provide new insights into the potential use of prognostic biomarkers and the effective guidance of clinical therapy from the human subcellular signaling network perspective.

Predicting 2-year survival for radiation regimens in advanced non-small cell lung cancer

AIMS

Total dose, dose per fraction, number of fractions and treatment time are important determinants of the biological effect of a radiation regimen. Several randomised clinical trials (RCTs) have tested a variety of dosing regimens in advanced unresected non-small cell lung cancer, but survival remains poor. This work used past RCT data to develop and validate a predictive model that could help in designing new radiation regimens for successful testing in RCTs.

MATERIALS AND METHODS

Eleven RCTs that compared radiation regimens alone were used to define the relationship between radiation regimens and 2-year survival. On the basis of this relationship, predictive models were developed. Predicted values were internally and externally validated against observed values from the same 11 RCTs and 21 other RCTs. Scatter plots and Pearson's correlation coefficient (r) were used for validation. Finally, regimens were explored that could improve survival.

RESULTS

Increments in the total dose, dose per day and the number of treatment days were associated with improved survival; increments in dose-squared and treatment weeks were associated with reduced survival. The observed and predicted values were similar on internal (r = 0.96) and external validation (r = 0.76). Regimens that delivered a higher total dose over a shorter time had higher survival rates compared with the standard (60 Gy, 30 fractions, 6 weeks); survival may be improved by delivering the standard treatment in 5 weeks rather than 6 weeks.

CONCLUSION

The developed model can predict the effect of thoracic radiation on survival in advanced non-small cell lung cancer patients. It is a useful tool for designing new radiation regimens for clinical trials.

Poor baseline pulmonary function may not increase the risk of radiation-induced lung toxicity

PURPOSE

Poor pulmonary function (PF) is often considered a contraindication to definitive radiation therapy for lung cancer. This study investigated whether baseline PF was associated with radiation-induced lung toxicity (RILT) in patients with non-small cell lung cancer (NSCLC) receiving conformal radiation therapy (CRT).

METHODS AND MATERIALS

NSCLC patients treated with CRT and tested for PF at baseline were eligible. Baseline predicted values of forced expiratory volume in 1 sec (FEV1), forced vital capacity (FVC), and diffusion capacity of lung for carbon monoxide (DLCO) were analyzed. Additional factors included age, gender, smoking status, Karnofsky performance status, coexisting chronic obstructive pulmonary disease (COPD), tumor location, histology, concurrent chemotherapy, radiation dose, and mean lung dose (MLD) were evaluated for RILT. The primary endpoint was symptomatic RILT (SRILT), including grade ≥2 radiation pneumonitis and fibrosis.

RESULTS

There was a total of 260 patients, and SRILT occurred in 58 (22.3%) of them. Mean FEV1 values for SRILT and non-SRILT patients were 71.7% and 65.9% (P=.077). Under univariate analysis, risk of SRILT increased with MLD (P=.008), the absence of COPD (P=.047), and FEV1 (P=.077). Age (65 split) and MLD were significantly associated with SRILT in multivariate analysis. The addition of FEV1 and age with the MLD-based model slightly improved the predictability of SRILT (area under curve from 0.63-0.70, P=.088).

CONCLUSIONS

Poor baseline PF does not increase the risk of SRILT, and combining FEV1, age, and MLD may improve the predictive ability.

Toward IMRT 2D dose modeling using artificial neural networks: a feasibility study

PURPOSE

To investigate the feasibility of artificial neural networks (ANN) to reconstruct dose maps for intensity modulated radiation treatment (IMRT) fields compared with those of the treatment planning system (TPS).

METHODS

An artificial feed forward neural network and the back-propagation learning algorithm have been used to replicate dose calculations of IMRT fields obtained from PINNACLE(3) v9.0. The ANN was trained with fluence and dose maps of IMRT fields for 6 MV x-rays, which were obtained from the amorphous silicon (a-Si) electronic portal imaging device of Novalis TX. Those fluence distributions were imported to the TPS and the dose maps were calculated on the horizontal midpoint plane of a water equivalent homogeneous cylindrical virtual phantom. Each exported 2D dose distribution from the TPS was classified into two clusters of high and low dose regions, respectively, based on the K-means algorithm and the Euclidian metric in the fluence-dose domain. The data of each cluster were divided into two sets for the training and validation phase of the ANN, respectively. After the completion of the ANN training phase, 2D dose maps were reconstructed by the ANN and isodose distributions were created. The dose maps reconstructed by ANN were evaluated and compared with the TPS, where the mean absolute deviation of the dose and the γ-index were used.

RESULTS

A good agreement between the doses calculated from the TPS and the trained ANN was achieved. In particular, an average relative dosimetric difference of 4.6% and an average γ-index passing rate of 93% were obtained for low dose regions, and a dosimetric difference of 2.3% and an average γ-index passing rate of 97% for high dose region.

CONCLUSIONS

An artificial neural network has been developed to convert fluence maps to corresponding dose maps. The feasibility and potential of an artificial neural network to replicate complex convolution kernels in the TPS for IMRT dose calculations have been demonstrated.

Modeling plan-related clinical complications using machine learning tools in a multiplan IMRT framework

PURPOSE

To predict organ-at-risk (OAR) complications as a function of dose-volume (DV) constraint settings without explicit plan computation in a multiplan intensity-modulated radiotherapy (IMRT) framework.

METHODS AND MATERIALS

Several plans were generated by varying the DV constraints (input features) on the OARs (multiplan framework), and the DV levels achieved by the OARs in the plans (plan properties) were modeled as a function of the imposed DV constraint settings. OAR complications were then predicted for each of the plans by using the imposed DV constraints alone (features) or in combination with modeled DV levels (plan properties) as input to machine learning (ML) algorithms. These ML approaches were used to model two OAR complications after head-and-neck and prostate IMRT: xerostomia, and Grade 2 rectal bleeding. Two-fold cross-validation was used for model verification and mean errors are reported.

RESULTS

Errors for modeling the achieved DV values as a function of constraint settings were 0-6%. In the head-and-neck case, the mean absolute prediction error of the saliva flow rate normalized to the pretreatment saliva flow rate was 0.42% with a 95% confidence interval of (0.41-0.43%). In the prostate case, an average prediction accuracy of 97.04% with a 95% confidence interval of (96.67-97.41%) was achieved for Grade 2 rectal bleeding complications.

CONCLUSIONS

ML can be used for predicting OAR complications during treatment planning allowing for alternative DV constraint settings to be assessed within the planning framework.

A genetic algorithm for variable selection in logistic regression analysis of radiotherapy treatment outcomes

A given outcome of radiotherapy treatment can be modeled by analyzing its correlation with a combination of dosimetric, physiological, biological, and clinical factors, through a logistic regression fit of a large patient population. The quality of the fit is measured by the combination of the predictive power of this particular set of factors and the statistical significance of the individual factors in the model. We developed a genetic algorithm (GA), in which a small sample of all the possible combinations of variables are fitted to the patient data. New models are derived from the best models, through crossover and mutation operations, and are in turn fitted. The process is repeated until the sample converges to the combination of factors that best predicts the outcome. The GA was tested on a data set that investigated the incidence of lung injury in NSCLC patients treated with 3DCRT. The GA identified a model with two variables as the best predictor of radiation pneumonitis: the V30 (p=0.048) and the ongoing use of tobacco at the time of referral (p=0.074). This two-variable model was confirmed as the best model by analyzing all possible combinations of factors. In conclusion, genetic algorithms provide a reliable and fast way to select significant factors in logistic regression analysis of large clinical studies.

The importance of patient characteristics for the prediction of radiation-induced lung toxicity

PURPOSE

Extensive research has led to the identification of numerous dosimetric parameters as well as patient characteristics, associated with lung toxicity, but their clinical usefulness remains largely unknown. We investigated the predictive value of patient characteristics in combination with established dosimetric parameters.

PATIENTS AND METHODS

Data from 438 lung cancer patients treated with (chemo)radiation were used. Lung toxicity was scored using the Common Toxicity Criteria version 3.0. A multivariate model as well as two single parameter models, including either V(20) or MLD, was built. Performance of the models was expressed as the AUC (Area Under the Curve).

RESULTS

The mean MLD was 13.5 Gy (SD 4.5 Gy), while the mean V(20) was 21.0% (SD 7.3%). Univariate models with V(20) or MLD both yielded an AUC of 0.47. The final multivariate model, which included WHO-performance status, smoking status, forced expiratory volume (FEV(1)), age and MLD, yielded an AUC of 0.62 (95% CI: 0.55-0.69).

CONCLUSIONS

Within the range of radiation doses used in our clinic, dosimetric parameters play a less important role than patient characteristics for the prediction of lung toxicity. Future research should focus more on patient-related factors, as opposed to dosimetric parameters, in order to identify patients at high risk for developing radiation-induced lung toxicity more accurately.

Combining multiple models to generate consensus: application to radiation-induced pneumonitis prediction

The fusion of predictions from disparate models has been used in several fields to obtain a more realistic and robust estimate of the "ground truth" by allowing the models to reinforce each other when consensus exists, or, conversely, negate each other when there is no consensus. Fusion has been shown to be most effective when the models have some complementary strengths arising from different approaches. In this work, we fuse the results from four common but methodologically different nonlinear multivariate models (Decision Trees, Neural Networks, Support Vector Machines, Self-Organizing Maps) that were trained to predict radiation-induced pneumonitis risk on a database of 219 lung cancer patients treated with radiotherapy (34 with Grade 2+ postradiotherapy pneumonitis). Each model independently incorporated a small number of features from the available set of dose and nondose patient variables to predict pneumonitis; no two models had all features in common. Fusion was achieved by simple averaging of the predictions for each patient from all four models. Since a model's prediction for a patient can be dependent on the patient training set used to build the model, the average of several different predictions from each model was used in the fusion (predictions were made by repeatedly testing each patient with a model built from different cross-validation training sets that excluded the patient being tested). The area under the receiver operating characteristics curve for the fused cross-validated results was 0.79, with lower variance than the individual component models. From the fusion, five features were extracted as the consensus among all four models in predicting radiation pneumonitis. Arranged in order of importance, the features are (1) chemotherapy; (2) equivalent uniform dose (EUD) for exponent a=1.2 to 3; (3) EUD for a=0.5 to 1.2, lung volume receiving >20-30 Gy; (4) female sex; and (5) squamous cell histology. To facilitate ease of interpretation and prospective use, the fused outcome results for the patients were fitted to a logistic probability function.