Large-scale dose evaluation of deep learning organ contours in head-and-neck radiotherapy by leveraging existing plans.

Background and purpose: Retrospective dose evaluation for organ-at-risk auto-contours has previously used small cohorts due to additional manual effort required for treatment planning on auto-contours. We aimed to do this at large scale, by a) proposing and assessing an automated plan optimization workflow that used existing clinical plan parameters and b) using it for head-and-neck auto-contour dose evaluation. Materials and methods: Our automated workflow emulated our clinic's treatment planning protocol and reused existing clinical plan optimization parameters. This workflow recreated the original clinical plan (POG) with manual contours (PMC) and evaluated the dose effect (POG-PMC) on 70 photon and 30 proton plans of head-and-neck patients. As a use-case, the same workflow (and parameters) created a plan using auto-contours (PAC) of eight head-and-neck organs-at-risk from a commercial tool and evaluated their dose effect (PMC-PAC). Results: For plan recreation (POG-PMC), our workflow had a median impact of 1.0% and 1.5% across dose metrics of auto-contours, for photon and proton respectively. Computer time of automated planning was 25% (photon) and 42% (proton) of manual planning time. For auto-contour evaluation (PMC-PAC), we noticed an impact of 2.0% and 2.6% for photon and proton radiotherapy. All evaluations had a median ΔNTCP (Normal Tissue Complication Probability) less than 0.3%. Conclusions: The plan replication capability of our automated program provides a blueprint for other clinics to perform auto-contour dose evaluation with large patient cohorts. Finally, despite geometric differences, auto-contours had a minimal median dose impact, hence inspiring confidence in their utility and facilitating their clinical adoption.

Non-invasive imaging prediction of tumor hypoxia: A novel developed and externally validated CT and FDG-PET-based radiomic signatures.

BACKGROUND: Tumor hypoxia increases resistance to radiotherapy and systemic therapy. Our aim was to develop and validate a disease-agnostic and disease-specific CT (+FDG-PET) based radiomics hypoxia classification signature. MATERIAL AND METHODS: A total of 808 patients with imaging data were included: N = 100 training/N = 183 external validation cases for a disease-agnostic CT hypoxia classification signature, N = 76 training/N = 39 validation cases for the H&N CT signature and N = 62 training/N = 36 validation cases for the Lung CT signature. The primary gross tumor volumes (GTV) were manually defined by experts on CT. In order to dichotomize between hypoxic/well-oxygenated tumors a threshold of 20% was used for the [18F]-HX4-derived hypoxic fractions (HF). A random forest (RF)-based machine-learning classifier/regressor was trained to classify patients as hypoxia-positive/ negative based on radiomic features. RESULTS: A 11 feature "disease-agnostic CT model" reached AUC's of respectively 0.78 (95% confidence interval [CI], 0.62-0.94), 0.82 (95% CI, 0.67-0.96) and 0.78 (95% CI, 0.67-0.89) in three external validation datasets. A "disease-agnostic FDG-PET model" reached an AUC of 0.73 (0.95% CI, 0.49-0.97) in validation by combining 5 features. The highest "lung-specific CT model" reached an AUC of 0.80 (0.95% CI, 0.65-0.95) in validation with 4 CT features, while the "H&N-specific CT model" reached an AUC of 0.84 (0.95% CI, 0.64-1.00) in validation with 15 CT features. A tumor volume-alone model was unable to significantly classify patients as hypoxia-positive/ negative. A significant survival split (P = 0.037) was found between CT-classified hypoxia strata in an external H&N cohort (n = 517), while 117 significant hypoxia gene-CT signature feature associations were found in an external lung cohort (n = 80). CONCLUSION: The disease-specific radiomics signatures perform better than the disease agnostic ones. By identifying hypoxic patients our signatures have the potential to enrich interventional hypoxia-targeting trials.

Distributed learning on 20 000+ lung cancer patients - The Personal Health Train.

BACKGROUND AND PURPOSE: Access to healthcare data is indispensable for scientific progress and innovation. Sharing healthcare data is time-consuming and notoriously difficult due to privacy and regulatory concerns. The Personal Health Train (PHT) provides a privacy-by-design infrastructure connecting FAIR (Findable, Accessible, Interoperable, Reusable) data sources and allows distributed data analysis and machine learning. Patient data never leaves a healthcare institute. MATERIALS AND METHODS: Lung cancer patient-specific databases (tumor staging and post-treatment survival information) of oncology departments were translated according to a FAIR data model and stored locally in a graph database. Software was installed locally to enable deployment of distributed machine learning algorithms via a central server. Algorithms (MATLAB, code and documentation publicly available) are patient privacy-preserving as only summary statistics and regression coefficients are exchanged with the central server. A logistic regression model to predict post-treatment two-year survival was trained and evaluated by receiver operating characteristic curves (ROC), root mean square prediction error (RMSE) and calibration plots. RESULTS: In 4 months, we connected databases with 23 203 patient cases across 8 healthcare institutes in 5 countries (Amsterdam, Cardiff, Maastricht, Manchester, Nijmegen, Rome, Rotterdam, Shanghai) using the PHT. Summary statistics were computed across databases. A distributed logistic regression model predicting post-treatment two-year survival was trained on 14 810 patients treated between 1978 and 2011 and validated on 8 393 patients treated between 2012 and 2015. CONCLUSION: The PHT infrastructure demonstrably overcomes patient privacy barriers to healthcare data sharing and enables fast data analyses across multiple institutes from different countries with different regulatory regimens. This infrastructure promotes global evidence-based medicine while prioritizing patient privacy.

Multifactorial risk factors for mortality after chemotherapy and radiotherapy for non-small cell lung cancer.

BACKGROUND AND PURPOSE: A higher radiation dose to the heart is known to be associated with increased mortality in non-small cell lung cancer (NSCLC) patients. It is however unknown what the contribution of the heart dose is when other risk factors for mortality are also accounted for. MATERIALS AND METHODS: We constructed and externally validated prediction models of mortality after definitive chemoradiotherapy for NSCLC. Models were developed in 145 stage I-IIIB NSCLC patients. Clinical (performance status, age, gross tumour volume (GTV) combining primary tumour and involved lymph nodes, current smoker) and dosimetric (mean lung (MLD) and heart (MHD) dose) variables were considered. Multivariable logistic regression models predicting 12 and 24 month mortality were built in 5-fold cross-validation. Discrimination and calibration was assessed in 3 external validation datasets containing 878 (via distributed learning), 127 and 96 NSCLC patients. RESULTS: The best discriminating prediction models combined GTV, smoker and/or MHD: bootstrapping AUC (95% CI) of 0.74 (0.66-0.78) and 0.69 (0.55-0.74) at 12 and 24 months. At external validation, the 24 month mortality GTV-smoker-MHD model robustly showed moderate discrimination (AUC = 0.61-0.64 before and 0.64-0.65 after model update) with limited 0.01-0.07 improvement over a GTV-only model, and calibration slope (0.64-0.65). This model can identify patients for whom a MHD reduction may be useful (e.g. PPV = 77%, NPV = 52% (60% cut-off)). CONCLUSIONS: Tumour volume is strongly related to mortality risk in the first 2 years after chemoradiotherapy for NSCLC. Modelling indicates that efforts to reduce cardiac dose may be relevant for small tumours and that smoking has an important negative association with survival.

Distributed radiomics as a signature validation study using the Personal Health Train infrastructure.

Prediction modelling with radiomics is a rapidly developing research topic that requires access to vast amounts of imaging data. Methods that work on decentralized data are urgently needed, because of concerns about patient privacy. Previously published computed tomography medical image sets with gross tumour volume (GTV) outlines for non-small cell lung cancer have been updated with extended follow-up. In a previous study, these were referred to as Lung1 (n = 421) and Lung2 (n = 221). The Lung1 dataset is made publicly accessible via The Cancer Imaging Archive (TCIA; https://www.cancerimagingarchive.net ). We performed a decentralized multi-centre study to develop a radiomic signature (hereafter "ZS2019") in one institution and validated the performance in an independent institution, without the need for data exchange and compared this to an analysis where all data was centralized. The performance of ZS2019 for 2-year overall survival validated in distributed radiomics was not statistically different from the centralized validation (AUC 0.61 vs 0.61; p = 0.52). Although slightly different in terms of data and methods, no statistically significant difference in performance was observed between the new signature and previous work (c-index 0.58 vs 0.65; p = 0.37). Our objective was not the development of a new signature with the best performance, but to suggest an approach for distributed radiomics. Therefore, we used a similar method as an earlier study. We foresee that the Lung1 dataset can be further re-used for testing radiomic models and investigating feature reproducibility.

External validation of an NTCP model for acute esophageal toxicity in locally advanced NSCLC patients treated with intensity-modulated (chemo-)radiotherapy.

BACKGROUND AND PURPOSE: We externally validated a previously established multivariable normal-tissue complication probability (NTCP) model for Grade ≥2 acute esophageal toxicity (AET) after intensity-modulated (chemo-)radiotherapy or volumetric-modulated arc therapy for locally advanced non-small cell lung cancer. MATERIALS AND METHODS: A total of 603 patients from five cohorts (A-E) within four different Dutch institutes were included. Using the NTCP model, containing predictors concurrent chemoradiotherapy, mean esophageal dose, gender and clinical tumor stage, the risk of Grade ≥2 AET was estimated per patient and model discrimination and (re)calibration performance were evaluated. RESULTS: Four validation cohorts (A, B, D, E) experienced higher incidence of Grade ≥2 AET compared to the training cohort (49.3-70.2% vs 35.6%; borderline significant for one cohort, highly significant for three cohorts). Cohort C experienced lower Grade ≥2 AET incidence (21.7%, p < 0.001). For three cohorts (A-C), discriminative performance was similar to the training cohort (area under the curve (AUC) 0.81-0.89 vs 0.84). In the two remaining cohorts (D-E) the model showed poor discriminative power (AUC 0.64 and 0.63). Reasonable calibration performance was observed in two cohorts (A-B), and recalibration further improved performance in all three cohorts with good discrimination (A-C). Recalibration for the two poorly discriminating cohorts (D-E) did not improve performance. CONCLUSIONS: The NTCP model for AET prediction was successfully validated in three out of five patient cohorts (AUC ≥0.80). The model did not perform well in two cohorts, which included patients receiving substantially different treatment. Before applying the model in clinical practice, validation of discrimination and (re)calibration performance in a local cohort is recommended.

Machine learning algorithms for outcome prediction in (chemo)radiotherapy: An empirical comparison of classifiers.

PURPOSE: Machine learning classification algorithms (classifiers) for prediction of treatment response are becoming more popular in radiotherapy literature. General Machine learning literature provides evidence in favor of some classifier families (random forest, support vector machine, gradient boosting) in terms of classification performance. The purpose of this study is to compare such classifiers specifically for (chemo)radiotherapy datasets and to estimate their average discriminative performance for radiation treatment outcome prediction. METHODS: We collected 12 datasets (3496 patients) from prior studies on post-(chemo)radiotherapy toxicity, survival, or tumor control with clinical, dosimetric, or blood biomarker features from multiple institutions and for different tumor sites, that is, (non-)small-cell lung cancer, head and neck cancer, and meningioma. Six common classification algorithms with built-in feature selection (decision tree, random forest, neural network, support vector machine, elastic net logistic regression, LogitBoost) were applied on each dataset using the popular open-source R package caret. The R code and documentation for the analysis are available online (https://github.com/timodeist/classifier_selection_code). All classifiers were run on each dataset in a 100-repeated nested fivefold cross-validation with hyperparameter tuning. Performance metrics (AUC, calibration slope and intercept, accuracy, Cohen's kappa, and Brier score) were computed. We ranked classifiers by AUC to determine which classifier is likely to also perform well in future studies. We simulated the benefit for potential investigators to select a certain classifier for a new dataset based on our study (pre-selection based on other datasets) or estimating the best classifier for a dataset (set-specific selection based on information from the new dataset) compared with uninformed classifier selection (random selection). RESULTS: Random forest (best in 6/12 datasets) and elastic net logistic regression (best in 4/12 datasets) showed the overall best discrimination, but there was no single best classifier across datasets. Both classifiers had a median AUC rank of 2. Preselection and set-specific selection yielded a significant average AUC improvement of 0.02 and 0.02 over random selection with an average AUC rank improvement of 0.42 and 0.66, respectively. CONCLUSION: Random forest and elastic net logistic regression yield higher discriminative performance in (chemo)radiotherapy outcome and toxicity prediction than other studied classifiers. Thus, one of these two classifiers should be the first choice for investigators when building classification models or to benchmark one's own modeling results against. Our results also show that an informed preselection of classifiers based on existing datasets can improve discrimination over random selection.

In vivo dosimetry in gynecological applications-A feasibility study.

PURPOSE: To investigate the feasibility of in vivo dosimetry using microMOSFET dosimeters in patients treated with brachytherapy using two types of gynecological applicators. METHODS AND MATERIALS: In this study, a microMOSFET was placed in an empty needle of an Utrecht Interstitial Fletcher applicator or MUPIT (Martinez Universal Perineal Interstitial Template) applicator for independent verification of treatment delivery. Measurements were performed in 10 patients, with one to three microMOSFETs per applicator and repeated for one to four fractions, resulting in 50 in vivo measurements. Phantom measurements were used to determine characteristics of the microMOSFETs. RESULTS: Phantom measurements showed a linear relationship between dose and microMOSFET threshold voltage, and a calibration coefficient (mV/cGy) was determined. Reproducibility of repeated 50 cGy irradiations was 2% (1 standard deviation). Distance and angle dependencies were measured and correction factors were determined. Subsequently, three microMOSFETs were placed in a phantom to measure a validation plan. The difference between predicted and measured dose was less than the measurement uncertainty (±9%, 2 standard deviations). In vivo measurements were corrected for distance and angle dependencies. Differences between predicted and measured dose in the patients were smaller than the measurement uncertainty for the majority of the measurements. CONCLUSIONS: In vivo dosimetry using microMOSFETs in MUPIT and Utrecht Interstitial Fletcher applicators has proved to be feasible. Reimaging should be performed after detection of differences larger than 10% between predicted and measured dose to verify the applicator configuration. Movement of the applicator relative to the target or organs at risk is undetectable with this method.

Inclusion of Incidental Radiation Dose to the Cardiac Atria and Ventricles Does Not Improve the Prediction of Radiation Pneumonitis in Advanced-Stage Non-Small Cell Lung Cancer Patients Treated With Intensity Modulated Radiation Therapy.

PURPOSE: To evaluate whether inclusion of incidental radiation dose to the cardiac atria and ventricles improves the prediction of grade ≥3 radiation pneumonitis (RP) in advanced-stage non-small cell lung cancer (AS-NSCLC) patients treated with intensity modulated radiation therapy (IMRT) or volumetric modulated arc therapy (VMAT). METHODS AND MATERIALS: Using a bootstrap modeling approach, clinical parameters and dose-volume histogram (DVH) parameters of lungs and heart (assessing atria and ventricles separately and combined) were evaluated for RP prediction in 188 AS-NSCLC patients. RESULTS: After a median follow-up of 18.4 months, 26 patients (13.8%) developed RP. Only the median mean lung dose (MLD) differed between groups (15.3 Gy vs 13.7 Gy for the RP and non-RP group, respectively; P=.004). The MLD showed the highest Spearman correlation coefficient (Rs) for RP (Rs = 0.21; P<.01). Most Rs of the lung DVH parameters exceeded those of the heart DVH parameters. After predictive modeling using a bootstrap procedure, the MLD was always included in the predictive model for grade ≥3 RP, whereas the heart DVH parameters were seldom included in the model. CONCLUSION: Incidental dose to the cardiac atria and ventricles did not improve RP risk prediction in our cohort of 188 AS-NSCLC patients treated with IMRT or VMAT.

Esophageal wall dose-surface maps do not improve the predictive performance of a multivariable NTCP model for acute esophageal toxicity in advanced stage NSCLC patients treated with intensity-modulated (chemo-)radiotherapy.

In our previous work, a multivariable normal-tissue complication probability (NTCP) model for acute esophageal toxicity (AET) Grade ⩾2 after highly conformal (chemo-)radiotherapy for non-small cell lung cancer (NSCLC) was developed using multivariable logistic regression analysis incorporating clinical parameters and mean esophageal dose (MED). Since the esophagus is a tubular organ, spatial information of the esophageal wall dose distribution may be important in predicting AET. We investigated whether the incorporation of esophageal wall dose-surface data with spatial information improves the predictive power of our established NTCP model. For 149 NSCLC patients treated with highly conformal radiation therapy esophageal wall dose-surface histograms (DSHs) and polar dose-surface maps (DSMs) were generated. DSMs were used to generate new DSHs and dose-length-histograms that incorporate spatial information of the dose-surface distribution. From these histograms dose parameters were derived and univariate logistic regression analysis showed that they correlated significantly with AET. Following our previous work, new multivariable NTCP models were developed using the most significant dose histogram parameters based on univariate analysis (19 in total). However, the 19 new models incorporating esophageal wall dose-surface data with spatial information did not show improved predictive performance (area under the curve, AUC range 0.79-0.84) over the established multivariable NTCP model based on conventional dose-volume data (AUC = 0.84). For prediction of AET, based on the proposed multivariable statistical approach, spatial information of the esophageal wall dose distribution is of no added value and it is sufficient to only consider MED as a predictive dosimetric parameter.

Comparison of toxicity and outcome in advanced stage non-small cell lung cancer patients treated with intensity-modulated (chemo-)radiotherapy using IMRT or VMAT.

Retrospective evaluation of 188 advanced stage non-small cell lung cancer patients treated with IMRT or VMAT revealed a limited increase of moderate to severe acute esophageal toxicity after VMAT. Acute pulmonary toxicity and severe late toxicity were low. Overall survival did not differ between the IMRT and VMAT groups.

Multivariable normal-tissue complication modeling of acute esophageal toxicity in advanced stage non-small cell lung cancer patients treated with intensity-modulated (chemo-)radiotherapy.

BACKGROUND AND PURPOSE: The majority of normal-tissue complication probability (NTCP) models for acute esophageal toxicity (AET) in advanced stage non-small cell lung cancer (AS-NSCLC) patients treated with (chemo-)radiotherapy are based on three-dimensional conformal radiotherapy (3D-CRT). Due to distinct dosimetric characteristics of intensity-modulated radiation therapy (IMRT), 3D-CRT based models need revision. We established a multivariable NTCP model for AET in 149 AS-NSCLC patients undergoing IMRT. MATERIALS AND METHODS: An established model selection procedure was used to develop an NTCP model for Grade ⩾2 AET (53 patients) including clinical and esophageal dose-volume histogram parameters. RESULTS: The NTCP model predicted an increased risk of Grade ⩾2 AET in case of: concurrent chemoradiotherapy (CCR) [adjusted odds ratio (OR) 14.08, 95% confidence interval (CI) 4.70-42.19; p<0.001], increasing mean esophageal dose [Dmean; OR 1.12 per Gy increase, 95% CI 1.06-1.19; p<0.001], female patients (OR 3.33, 95% CI 1.36-8.17; p=0.008), and ⩾cT3 (OR 2.7, 95% CI 1.12-6.50; p=0.026). The AUC was 0.82 and the model showed good calibration. CONCLUSIONS: A multivariable NTCP model including CCR, Dmean, clinical tumor stage and gender predicts Grade ⩾2 AET after IMRT for AS-NSCLC. Prior to clinical introduction, the model needs validation in an independent patient cohort.

EEG source localization in full-term newborns with hypoxic-ischemia.

The aim of this study was to evaluate EEG source localization by standardized weighted low-resolution brain electromagnetic tomography (swLORETA) for monitoring of fullterm newborns with hypoxic-ischemic encephalopathy, using a standard anatomic head model. Three representative examples of neonatal hypoxic-ischemia were included. The method was validated with MRI data. Hypoxic-ischemic areas, visible on MRI, correlated well with swLORETA current density distributions. In addition, neonatal seizure activity may be localized. The calculated current density distributions provide easy-to-interpret localized information about neonatal brain function, which may enable detailed longitudinal monitoring and potential assessment of treatment efficacy.

Immobilization precision of a modified GTC frame.

The purpose of this study was to evaluate and quantify the interfraction reproducibility and intrafraction immobilization precision of a modified GTC frame. The error of the patient alignment and imaging systems were measured using a cranial skull phantom, with simulated, predetermined shifts. The kV setup images were acquired with a room-mounted set of kV sources and panels. Calculated translations and rotations provided by the computer alignment software relying upon three implanted fiducials were compared to the known shifts, and the accuracy of the imaging and positioning systems was calculated. Orthogonal kV setup images for 45 proton SRT patients and 1002 fractions (average 22.3 fractions/patient) were analyzed for interfraction and intrafraction immobilization precision using a modified GTC frame. The modified frame employs a radiotransparent carbon cup and molded pillow to allow for more treatment angles from posterior directions for cranial lesions. Patients and the phantom were aligned with three 1.5 mm stainless steel fiducials implanted into the skull. The accuracy and variance of the patient positioning and imaging systems were measured to be 0.10 ± 0.06 mm, with the maximum uncertainty of rotation being ±0.07°. 957 pairs of interfraction image sets and 974 intrafraction image sets were analyzed. 3D translations and rotations were recorded. The 3D vector interfraction setup reproducibility was 0.13 mm ± 1.8 mm for translations and the largest uncertainty of ± 1.07º for rotations. The intrafraction immobilization efficacy was 0.19 mm ± 0.66 mm for translations and the largest uncertainty of ± 0.50º for rotations. The modified GTC frame provides reproducible setup and effective intrafraction immobilization, while allowing for the complete range of entrance angles from the posterior direction.

Effects of midazolam and lidocaine on spectral properties of the EEG in full-term neonates with stroke.

Assessment of the neonatal EEG may be hampered by drug-specific changes in electrocortical activity. To quantify effects of a loading dose of midazolam and lidocaine on the EEG frequency spectrum of full-term neonates with perinatal arterial ischemic stroke (PAIS), 11 full-term infants underwent multi-channel amplitude-integrated EEG (aEEG) and EEG recordings. During recording, midazolam and/or lidocaine were administered as anti-epileptic drug. Retrospectively, we performed spectral analysis on 4-h EEG segments around the loading dose. The frequency spectrum was divided in δ (1-4 Hz), θ (4-8 Hz), α (8-13 Hz) and ß (13-30 Hz) bands. Midazolam induced immediate suppression of the aEEG background pattern for 30-60 min. Spectral EEG analysis showed decreased total and absolute frequency band powers. Relative δ power decreased, θ power increased while α and ß powers remained constant. Lidocaine induced no aEEG background pattern suppression. Total and absolute EEG band powers were unchanged. Relative δ power decreased, θ and α power increased and ß power remained constant. Effects of lidocaine were more pronounced in the stroke-affected hemisphere. In conclusions, both drugs induced a shift from low to higher frequency electrocortical activity. Additionally, midazolam reduced total EEG power. These spectral changes differ from those seen in adult studies.