Normal Tissue Complication Probability Models Research Articles

External validation of a multimodality deep-learning normal tissue complication probability model for mandibular osteoradionecrosis trained on 3D radiation distribution maps and clinical variables

Background and purposeWhile the inclusion of spatial dose information in deep learning (DL)-based normal-tissue complication probability (NTCP) models has been the focus of recent research studies, external validation is still lacking. This study aimed to externally validate a DL-based NTCP model for mandibular osteoradionecrosis (ORN) trained on 3D radiation dose distribution maps and clinical variables. Methods and materialsA 3D DenseNet-40 convolutional neural network (3D-mDN40) was trained on clinical and radiation dose distribution maps on a retrospective class-balanced matched cohort of 184 subjects. A second model (3D-DN40) was trained on dose maps only and both DL models were compared to a logistic regression (LR) model trained on DVH metrics and clinical variables. All models were externally validated by means of their discriminative ability and calibration on an independent dataset of 82 subjects. ResultsNo significant difference in performance was observed between models. In internal validation, these exhibited similar Brier scores around 0.2, Log Loss values of 0.6–0.7 and ROC AUC values around 0.7 (internal) and 0.6 (external). Differences in clinical variable distributions and their effect sizes were observed between internal and external cohorts, such as smoking status (0.6 vs. 0.1) and chemotherapy (0.1 vs. −0.5), respectively. ConclusionTo our knowledge, this is the first study to externally validate a multimodality DL-based ORN NTCP model. Utilising mandible dose distribution maps, these models show promise for enhancing spatial risk assessment and guiding dental and oncological decision-making, though further research is essential to address overfitting and domain shift for reliable clinical use.

Normal tissue complication probability modeling for late rectal bleeding after conventional or hypofractionated radiotherapy for prostate cancer

PurposeTo develop a single NTCP model for grade ≥ 2 late rectal bleeding (G2 LRB) after conventional or hypofractionated radiotherapy for prostate cancer. Methods and MaterialsThe development dataset consisted of prostate cancer patients (n = 656) previously randomized to conventional (39 x 2 Gy) or hypofractionated (19 x 3.4 Gy) external beam radiotherapy with N = 89 G2 LRB cases. Candidate predictors were obtained from literature. We fitted five separate logistic regression models to the data, each with one of the following dose parameters as candidate predictors in biological effective dose (BED), assuming α/β = 3 Gy: Equivalent uniform dose (EUD) with n = 0.1, EUD with n = 0.2, the relative volume receiving ≥ 111.9 Gy in BED (V111.9, the equivalent of physical V70 for a conventional schedule), minimum BED to the hottest 0.1 cm3 (D0.1cm3) or 2 cm3 (D2cm3). Previous abdominal surgery was included in every model and fractionation schedule was tested as predictor in each model. A sensitivity analysis was performed by varying the α/β-ratio, n and dose-volume cutoff. ResultsThe pre-selected candidate dosimetric predictor and previous abdominal surgery were significantly associated with the outcome in all five models. Fractionation schedule was eliminated by the backward scheme in only the EUD (n = 0.1), D0.1cm3 and D2cm3-based models. In internal validation these models showed AUC’s of 0.64, 0.60 & 0.62, respectively. The sensitivity analyses showed that EUD models with n ≥ 0.15 and / or α/β ≥ 4 Gy failed, and EUD models based on α/β = 2 Gy with n = 0.05–0.2 showed good fits as well. ConclusionsOur trial data set with different fractionation schedules offered the unique possibility to generate unbiased BED-based models. EUD (n = 0.1), D0.1cm3 and D2cm3 performed overall best in predicting G2 LRB; with α/β = 2 Gy equally good models were obtained. External validation is required to confirm our results.

Variable-RBE-induced NTCP predictions for various side-effects following proton therapy for brain tumors – Identification of high-risk patients and risk mitigation

Background and purposeDisregarding the increase of relative biological effectiveness (RBE) may raise the risk of acute and late adverse events after proton beam therapy (PBT). This study aims to explore the relationship between variable RBE (above 1.1)-induced normal tissue complication probabilities (NTCP) and patient-specific factors, identify patients at high risk of RBE-induced NTCP increase, and assess risk mitigation by incorporating RBE variability into treatment planning. Materials and methodsWe retrospectively analyzed 105 primary brain tumor patients treated with PBT (RBE = 1.1). We calculated differences in estimated NTCP (ΔNTCP) using a variable RBE-weighted dose (DRBE, Wedenberg model) and a constant RBE-weighted dose (DRBE=1.1), across 16 NTCP models. These differences were correlated with patient-specific characteristics. Based on ΔNTCP, patients were classified as high risk (32 %) or low risk (68 %) for adverse events due to RBE-induced NTCP. This classification was compared with alternative classifications based on (a) relevant patient-specific characteristics, (b) DRBE=1.1, and (c) the difference between DRBE and DRBE=1.1 (ΔD), assessing the balanced accuracy. The potential to reduce RBE-induced NTCP through track-end and linear energy transfer (LET) optimization was evaluated in six example patients. ResultsUsing a variable RBE instead of a constant one resulted in NTCP increases (up to 32 percentage points). Variable-RBE-induced NTCP increases were strongly negatively correlated with the distance between the clinical target volume (CTV) and the organ at risk (OAR) for most side-effects, and positively correlated with CTV volume for certain side-effects. High increases were associated with (a) specific patient factors, particularly the proximity of the CTV to OARs, (b) DRBE=1.1, and (c) ΔD, with a balanced accuracy of 0.88, 0.94, and 0.86, respectively. Optimization of track-ends and LET considerably reduced NTCP values, achieving a mean reduction of 31 % for optimized OARs. ConclusionThe risk of variable-RBE-induced NTCP strongly depends on patient-specific factors and the considered side-effect. A small distance between the tumor and OARs notably increases the risk. Integrating biologically-guided objectives into treatment planning can effectively mitigate the risk.

Comparison of online adaptive and non-adaptive magnetic resonance image-guided radiation therapy in prostate cancer using dose accumulation

Background and purposeConventional image-guided radiotherapy (conv-IGRT) is standard in prostate cancer (PC) but does not account for inter-fraction anatomical changes. Online-adaptive magnetic resonance-guided RT (OA-MRgRT) may improve organ-at-risk (OARs) sparing and clinical target volume (CTV) coverage. The aim of this study was to analyze accumulated OAR and target doses in PC after OA-MRgRT and conv-IGRT in comparison to pre-treatment reference planning (refPlan). Material and methodsTen patients with PC, previously treated with OA-MRgRT at the 1.5 T MR-Linac (20x3Gy), were included. Accumulated OA-MRgRT doses were determined by deformably registering all fraction’s MR-images. Conv-IGRT was simulated through rigid registration of the planning computed tomography with each fraction’s MR-image for dose mapping/accumulation. Dose-volume parameters (DVPs), including CTV D50% and D98%, rectum, bladder, urethra, Dmax and V56Gy for OA-MRgRT, conv-IGRT and refPlan were compared using the Wilcoxon signed-rank test. Clinical relevance of accumulated dose differences was analyzed using a normal-tissue complication-probability model. ResultsCTV-DVPs were comparable, whereas OA-MRgRT yielded decreased median OAR-DVPs compared to conv-IGRT, except for bladder V56Gy. OA-MRgRT demonstrated significantly lower median rectum Dmax over conv-IGRT (59.1/59.9 Gy, p = 0.006) and refPlan (60.1 Gy, p = 0.012). Similarly, OA-MRgRT yielded reduced median bladder Dmax compared to conv-IGRT (60.0/60.4 Gy, p = 0.006), and refPlan (61.2 Gy, p = 0.002). Overall, accumulated dose differences were small and did not translate into clinically relevant effects. ConclusionDeformably accumulated OA-MRgRT using 20x3Gy in PC showed significant but small dosimetric differences comparted to conv-IGRT. Feasibility of a dose accumulation methodology was demonstrated, which may be relevant for evaluating future hypo-fractionated OA-MRgRT approaches.

Brachial plexopathy following stereotactic body radiation therapy in apical lung malignancies: A dosimetric pooled analysis of individual patient data

Background and objectivesThe aim of this study is to establish dosimetric constraints for the brachial plexus at risk of developing grade ≥ 2 brachial plexopathy in the context of stereotactic body radiation therapy (SBRT). Patients and MethodsIndividual patient data from 349 patients with 356 apical lung malignancies who underwent SBRT were extracted from 5 articles. The anatomical brachial plexus was delineated following the guidelines provided in the atlases developed by Hall, et al. and Kong, et al.. Patient characteristics, pertinent SBRT dosimetric parameters, and brachial plexopathy grades (according to CTCAE 4.0 or 5.0) were obtained. Normal tissue complication probability (NTCP) models were used to estimate the risk of developing grade ≥ 2 brachial plexopathy through maximum likelihood parameter fitting. ResultsThe prescription dose/fractionation schedules for SBRT ranged from 27 to 60 Gy in 1 to 8 fractions. During a follow-up period spanning from 6 to 113 months, 22 patients (6.3 %) developed grade ≥2 brachial plexopathy (4.3 % grade 2, 2.0 % grade 3); the median time to symptoms onset after SBRT was 8 months (ranged, 3–54 months). NTCP models estimated a 10 % risk of grade ≥2 brachial plexopathy with an anatomic brachial plexus maximum dose (Dmax) of 20.7 Gy, 34.2 Gy, and 42.7 Gy in one, three, and five fractions, respectively. Similarly, the NTCP model estimates the risks of grade ≥2 brachial plexopathy as 10 % for BED Dmax at 192.3 Gy and EQD2 Dmax at 115.4 Gy with an α/β ratio of 3, respectively. Symptom persisted after treatment in nearly half of patients diagnosed with grade ≥2 brachial plexopathy (11/22, 50 %). ConclusionsThis study establishes dosimetric constraints ranging from 20.7 to 42.7 Gy across 1–5 fractions, aimed at mitigating the risk of developing grade ≥2 brachial plexopathy following SBRT. These findings provide valuable guidance for future ablative SBRT in apical lung malignancies.

Mandibular dose-volume predicts time-to-osteoradionecrosis in an actuarial normal-tissue complication probability (NTCP) model: External validation of right-censored clinico-dosimetric and competing risk application across international multi-institutional observational cohorts and online graphical user interface clinical support tool assessment.

Existing studies on osteoradionecrosis of the jaw (ORNJ) have primarily used cross-sectional data, assessing risk factors at a single time point. Determining the time-to-event profile of ORNJ has important implications to monitor oral health in head and neck cancer (HNC) long-term survivors. Demographic, clinical and dosimetric data were retrospectively obtained for a clinical observational cohort of 1129 patients with HNC treated with radiotherapy (RT) at The University of Texas MD Anderson Cancer Center. ORNJ was diagnosed in 198 patients (18%). A multivariable logistic regression analysis with forward stepwise variable selection identified significant predictors for ORNJ. These predictors were then used to train a Weibull Accelerated Failure Time (AFT) model, which was externally validated using an independent cohort of 265 patients (92 ORNJ cases and 173 controls) treated at Guy's and St. Thomas' Hospitals. Our model identified that each unit increase in D25% is significantly associated with a 12% shorter time to ORNJ (Adjusted Time Ratio [ATR] 0·88, p<0·005); pre-RT dental extractions was associated to a 27% faster (ATR 0·73, p=0·13) onset of ORNJ; male patients experienced a 38% shorter time to ORNJ (ATR 0·62, p = 0·11). The model demonstrated strong internal calibration (integrated Brier score of 0·133, D-calibration p-value 0.998) and optimal discrimination at 72 months (Harrell's C-index of 0·72). The model also showed good generalization to the independent cohort, despite a slight drop in performance. This study is the first to demonstrate a direct relationship between radiation dose and the time to ORNJ onset, providing a novel characterization of the impact of delivered dose not only on the probability of a late effect (ORNJ), but the conditional risk during survivorship. This work was supported by various funding sources including NIH, NIDCR, NCI, NAPT, NASA, BCM, Affirmed Pharma, CRUK, KWF Dutch Cancer Society, NWO ZonMw, and the Apache Corporation.

A deep-learning-based surrogate model for Monte-Carlo simulations of the linear energy transfer in primary brain tumor patients treated with proton-beam radiotherapy

Objective. This study explores the use of neural networks (NNs) as surrogate models for Monte-Carlo (MC) simulations in predicting the dose-averaged linear energy transfer (LET d ) of protons in proton-beam therapy based on the planned dose distribution and patient anatomy in the form of computed tomography (CT) images. As LET d is associated with variability in the relative biological effectiveness (RBE) of protons, we also evaluate the implications of using NN predictions for normal tissue complication probability (NTCP) models within a variable-RBE context. Approach. The predictive performance of three-dimensional NN architectures was evaluated using five-fold cross-validation on a cohort of brain tumor patients (n = 151). The best-performing model was identified and externally validated on patients from a different center (n = 107). LET d predictions were compared to MC-simulated results in clinically relevant regions of interest. We assessed the impact on NTCP models by leveraging LET d predictions to derive RBE-weighted doses, using the Wedenberg RBE model. Main results. We found NNs based solely on the planned dose distribution, i.e. without additional usage of CT images, can approximate MC-based LET d distributions. Root mean squared errors (RMSE) for the median LET d within the brain, brainstem, CTV, chiasm, lacrimal glands (ipsilateral/contralateral) and optic nerves (ipsilateral/contralateral) were 0.36, 0.87, 0.31, 0.73, 0.68, 1.04, 0.69 and 1.24 keV µm−1, respectively. Although model predictions showed statistically significant differences from MC outputs, these did not result in substantial changes in NTCP predictions, with RMSEs of at most 3.2 percentage points. Significance. The ability of NNs to predict LET d based solely on planned dose distributions suggests a viable alternative to compute-intensive MC simulations in a variable-RBE setting. This is particularly useful in scenarios where MC simulation data are unavailable, facilitating resource-constrained proton therapy treatment planning, retrospective patient data analysis and further investigations on the variability of proton RBE.

Estimating the risk of major adverse cardiac events following radiotherapy for left breast cancer using a modified generalized Lyman normal-tissue complication probability model

BackgroundWe introduced an adapted Lyman normal-tissue complication probability (NTCP) model, incorporating clinical risk factors and censored time-to-event data, to estimate the risk of major adverse cardiac events (MACE) following left breast cancer radiotherapy (RT). Materials and methodsClinical characteristics and MACE data of 1100 women with left-side breast cancer receiving postoperative RT from 2005 to 2017 were retrospectively collected. A modified generalized Lyman NTCP model based on the individual left ventricle (LV) equivalent uniform dose (EUD), accounting for clinical risk factors and censored data, was developed using maximum likelihood estimation. Subgroup analysis was performed for low-comorbidity and high-comorbidity groups. ResultsOver a median follow-up 7.8 years, 64 patients experienced MACE, with higher mean LV dose in affected individuals (4.1 Gy vs. 2.9 Gy). The full model accounting for clinical factors identified D50 = 43.3 Gy, m = 0.59, and n = 0.78 as the best-fit parameters. The threshold dose causing a 50 % probability of MACE was lower in the high-comorbidity group (D50 = 30 Gy) compared to the low-comorbidity group (D50 = 45 Gy). Predictions indicated that restricting LV EUD below 5 Gy yielded a 10-year relative MACE risk less than 1.3 and 1.5 for high-comorbidity and low-comorbidity groups, respectively. ConclusionPatients with comorbidities are more susceptible to cardiac events following breast RT. The proposed modified generalized Lyman model considers nondosimetric risk factors and addresses incomplete follow-up for late complications, offering comprehensive and individualized MACE risk estimates post-RT.

A novel predictor for dosimetry data of lung and the radiation pneumonitis incidence prior to SBRT in lung cancer patients

Normal tissue complication probability (NTCP) models for radiation pneumonitis (RP) in lung cancer patients with stereotactic body radiation therapy (SBRT), which based on dosimetric data from treatment planning, are limited to patients who have already received radiation therapy (RT). This study aims to identify a novel predictive factor for lung dose distribution and RP probability before devising actionable SBRT plans for lung cancer patients. A comprehensive correlation analysis was performed on the clinical and dose parameters of lung cancer patients who underwent SBRT. Linear regression models were utilized to analyze the dosimetric data of lungs. The performance of the regression models was evaluated using mean squared error (MSE) and the coefficient of determination (R2). Correlational analysis revealed that most clinical data exhibited weak correlations with dosimetric data. However, nearly all dosimetric variables showed "strong" or "very strong" correlations with each other, particularly concerning the mean dose of the ipsilateral lung (MI) and the other dosimetric parameters. Further study verified that the lung tumor ratio (LTR) was a significant predictor for MI, which could predict the incidence of RP. As a result, LTR can predict the probability of RP without the need to design an elaborate treatment plan. This study, as the first to offer a comprehensive correlation analysis of dose parameters, explored the specific relationships among them. Significantly, it identified LTR as a novel predictor for both dose parameters and the incidence of RP, without the need to design an elaborate treatment plan.

Three-Dimensional Deep Learning Normal Tissue Complication Probability Model to Predict Late Xerostomia in Patients With Head and Neck Cancer

Conventional normal tissue complication probability (NTCP) models for patients with head and neck cancer are typically based on single-value variables, which, for radiation-induced xerostomia, are baseline xerostomia and mean salivary gland doses. This study aimed to improve the prediction of late xerostomia by using 3-dimensional information from radiation dose distributions, computed tomography imaging, organ-at-risk segmentations, and clinical variables with deep learning (DL). An international cohort of 1208 patients with head and neck cancer from 2 institutes was used to train and twice validate DL models (deep convolutional neural network, EfficientNet-v2, and ResNet) with 3-dimensional dose distribution, computed tomography scan, organ-at-risk segmentations, baseline xerostomia score, sex, and age as input. The NTCP endpoint was moderate-to-severe xerostomia 12 months postradiation therapy. The DL models' prediction performance was compared with a reference model: a recently published xerostomia NTCP model that used baseline xerostomia score and mean salivary gland doses as input. Attention maps were created to visualize the focus regions of the DL predictions. Transfer learning was conducted to improve the DL model performance on the external validation set. All DL-based NTCP models showed better performance (area under the receiver operating characteristic curve [AUC]test, 0.78-0.79) than the reference NTCP model (AUCtest, 0.74) in the independent test. Attention maps showed that the DL model focused on the major salivary glands, particularly the stem cell-rich region of the parotid glands. DL models obtained lower external validation performance (AUCexternal, 0.63) than the reference model (AUCexternal, 0.66). After transfer learning on a small external subset, the DL model (AUCtl, external, 0.66) performed better than the reference model (AUCtl, external, 0.64). DL-based NTCP models performed better than the reference model when validated in data from the same institute. Improved performance in the external data set was achieved with transfer learning, demonstrating the need for multicenter training data to realize generalizable DL-based NTCP models.

Therapy-Associated Saliva and Taste change Evaluation (TASTE) in head & neck cancer patients undergoing radiotherapy: a study protocol

BackgroundOne of the main side effects of radiation therapy to the head and neck region is altered taste sensation. This causes significant morbidity and has profound effects on the quality of life (QoL) of patients. While radiation-associated toxicities like xerostomia and dysphagia are part of large investigations, data on taste impairment is sparse. Small cohort sizes in the majority of studies and a variety of analysis methods limit our current understanding of the underlying processes. None of the studies published to date used a taste-specific QoL questionnaire with differentiation of the different taste qualities (e.g. sour, bitter). Furthermore, data regarding the correlation of taste impairment with radiation-associated change in saliva composition is currently not available. The aim of the TASTE study is to fill this gap. Based on the acquired data, a normal tissue complication probability (NTCP) model for late radiation-associated taste impairment will be developed.MethodsIn this prospective, observational multicenter study 150 head and neck cancer patients undergoing radiation therapy will be recruited and undergo repetitive (semi-) objective and subjective assessment of their taste, smell and salivary function (questionnaires, taste and smell assessment, saliva analysis). Primary endpoint will be patient-reported taste impairment 12 months post radiation therapy using a standardized questionnaire. Secondary endpoints will include taste impairment measured using taste strips at 12 months and 2 years post radiation therapy. Differences between subgroups (radiation side, chemotherapy, etc.) and changes over time will be assessed while adjusting for confounding factors (e.g. age, sex, smoking history).DiscussionThis study sets out to further our understanding of taste impairment in patients undergoing radiation therapy to the head and neck region with the goal to prevent this common side effect in future patients. The results of the study may be used to evaluate taste-preserving radiotherapy for patients with head and neck cancer, which may significantly reduce the long-term burden in this patient cohort.

A marginal structural model for normal tissue complication probability.

The goal of radiation therapy for cancer is to deliver prescribed radiation dose to the tumor while minimizing dose to the surrounding healthy tissues. To evaluate treatment plans, the dose distribution to healthy organs is commonly summarized as dose-volume histograms (DVHs). Normal tissue complication probability (NTCP) modeling has centered around making patient-level risk predictions with features extracted from the DVHs, but few have considered adapting a causal framework to evaluate the safety of alternative treatment plans. We propose causal estimands for NTCP based on deterministic and stochastic interventions, as well as propose estimators based on marginal structural models that impose bivariable monotonicity between dose, volume, and toxicity risk. The properties of these estimators are studied through simulations, and their use is illustrated in the context of radiotherapy treatment of anal canal cancer patients.

Integrating local and distant radiation-induced lung injury: Development and validation of a predictive model for ventilation loss.

Investigations on radiation-induced lung injury (RILI) have predominantly focused on local effects, primarily those associated with radiation damage to lung parenchyma. However, recent studies from our group and others have revealed that radiation-induced damage to branching serial structures such as airways and vessels may also have a substantial impact on post-radiotherapy (RT) lung function. Furthermore, recent results from multiple functional lung avoidance RT trials, although promising, have demonstrated only modest toxicity reduction, likely because they were primarily focused on dose avoidance to lung parenchyma. These observations emphasize the critical need for predictive dose-response models that effectively incorporate both local and distant RILI effects. We develop and validate a predictive model for ventilation loss after lung RT. This model, referred to as P+A, integrates local (parenchyma [P]) and distant (central and peripheral airways [A]) radiation-induced damage, modeling partial (narrowing) and complete (collapse) obstruction of airways. In an IRB-approved prospective study, pre-RT breath-hold CTs (BHCTs) and pre- and one-year post-RT 4DCTs were acquired from lung cancer patients treated with definitive RT. Up to 13 generations of airways were automatically segmented on the BHCTs using a research virtual bronchoscopy software. Ventilation maps derived from the 4DCT scans were utilized to quantify pre- and post-RT ventilation, serving, respectively, as input data and reference standard (RS) in model validation. To predict ventilation loss solely due to parenchymal damage (referred to as P model), we used a normal tissue complication probability (NTCP) model. Our model used this NTCP-based estimate and predicted additional loss due radiation-induced partial or complete occlusion of individual airways, applying fluid dynamics principles and a refined version of our previously developed airway radiosensitivity model. Predictions of post-RT ventilation were estimated in the sublobar volumes (SLVs) connected to the terminal airways. To validate the model, we conducted a k-fold cross-validation. Model parameters were optimized as the values that provided the lowest root mean square error (RMSE) between predicted post-RT ventilation and the RS for all SLVs in the training data. The performance of the P+A and the P models was evaluated by comparing their respective post-RT ventilation values with the RS predictions. Additional evaluation using various receiver operating characteristic (ROC) metrics was also performed. We extracted a dataset of 560 SLVs from four enrolled patients. Our results demonstrated that the P+A model consistently outperformed the P model, exhibiting RMSEs that were nearly half as low across all patients (13±3 percentile for the P+A model vs. 24±3 percentile for the P model on average). Notably, the P+A model aligned closely with the RS in ventilation loss distributions per lobe, particularly in regions exposed to doses ≥13.5Gy. The ROC analysis further supported the superior performance of the P+A model compared to the P model in sensitivity (0.98vs. 0.07), accuracy (0.87vs. 0.25), and balanced predictions. These early findings indicate that airway damage is a crucial factor in RILI that should be included in dose-response modeling to enhance predictions of post-RT lung function.

Evaluation of an automated clinical decision system with deep learning dose prediction and NTCP model for prostate cancer proton therapy

Objective. To evaluate the feasibility of using a deep learning dose prediction approach to identify patients who could benefit most from proton therapy based on the normal tissue complication probability (NTCP) model. Approach. Two 3D UNets were established to predict photon and proton doses. A dataset of 95 patients with localized prostate cancer was randomly partitioned into 55, 10, and 30 for training, validation, and testing, respectively. We selected NTCP models for late rectum bleeding and acute urinary urgency of grade 2 or higher to quantify the benefit of proton therapy. Propagated uncertainties of predicted ΔNTCPs resulting from the dose prediction errors were calculated. Patient selection accuracies for a single endpoint and a composite evaluation were assessed under different ΔNTCP thresholds. Main results. Our deep learning-based dose prediction technique can reduce the time spent on plan comparison from approximately 2 days to as little as 5 seconds. The expanded uncertainty of predicted ΔNTCPs for rectum and bladder endpoints propagated from the dose prediction error were 0.0042 and 0.0016, respectively, which is less than one-third of the acceptable tolerance. The averaged selection accuracies for rectum bleeding, urinary urgency, and composite evaluation were 90%, 93.5%, and 93.5%, respectively. Significance. Our study demonstrates that deep learning dose prediction and NTCP evaluation scheme could distinguish the NTCP differences between photon and proton treatment modalities. In addition, the dose prediction uncertainty does not significantly influence the decision accuracy of NTCP-based patient selection for proton therapy. Therefore, automated deep learning dose prediction and NTCP evaluation schemes can potentially be used to screen large patient populations and to avoid unnecessary delays in the start of prostate cancer radiotherapy in the future.

Dosimetric and NTCP analyses for selecting parotid gland cancer patients for proton therapy.

To perform a dosimetric and a normal tissue complication probability (NTCP) comparison between intensity modulated proton therapy and photon volumetric modulated arc therapy in a cohort of patients with parotid gland cancers in a post-operative or radical setting. From May 2011 to September 2021, 37 parotid gland cancers patients treated at two institutions were eligible. Inclusion criteria were as follows: patients aged ⩾ 18 years, diagnosis of parotid gland cancers candidate for postoperative radiotherapy or definitive radiotherapy, presence of written informed consent for the use of anonymous data for research purposes. Organs at risk (OARs) were retrospectively contoured. Target coverage goal was defined as D95 > 98%. Six NTCP models were selected. NTCP profiles were calculated for each patient using an internally-developed Python script in RayStation TPS. Average differences in NTCP between photon and proton plans were tested for significance with a two-sided Wilcoxon signed-rank test. Seventy-four plans were generated. A lower Dmean to the majority of organs at risk (inner ear, cochlea, oral cavity, pharyngeal constrictor muscles, contralateral parotid and submandibular gland) was obtained with intensity modulated proton therapy vs volumetric modulated arc therapy with statistical significance (p < .05). Ten (27%) patients had a difference in NTCP (photon vs proton plans) greater than 10% for hearing loss and tinnitus: among them, seven qualified for both endpoints, two patients for hearing loss only, and one for tinnitus. In the current study, nearly one-third of patients resulted eligible for proton therapy and they were the most likely to benefit in terms of prevention of hearing loss and tinnitus.

The role of the soft palate dose regarding normal tissue toxicities in older adults with head and neck cancer undergoing definitive radiotherapy.

The number of older adults with head and neck squamous cell carcinoma (HNSCC) is continuously increasing. Older HNSCC patients may be more vulnerable to radiotherapy-related toxicities, so that extrapolation of available normal tissue complication probability (NTCP) models to this population may not be appropriate. Hence, we aimed to investigate the correlation between organ at risk (OAR) doses and chronic toxicities in older patients with HNSCC undergoing definitive radiotherapy. Patients treated with definitive radiotherapy, either alone or with concomitant systemic treatment, between 2009 and 2019 in a large tertiary cancer center were eligible for this analysis. OARs were contoured based on international consensus guidelines, and EQD2 doses using α/ß values of 3Gy for late effects were calculated based on the radiation treatment plans. Treatment-related toxicities were graded according to Common Terminology Criteria for Adverse Events version 5.0. Logistic regression analyses were carried out, and NTCP models were developed and internally validated using the bootstrapping method. A total of 180 patients with a median age of 73 years fulfilled the inclusion criteria and were analyzed. Seventy-three patients developed chronic moderate xerostomia (grade 2), 34 moderate dysgeusia (grade 2), and 59 moderate-to-severe (grade 2-3) dysphagia after definitive radiotherapy. The soft palate dose was significantly associated with all analyzed toxicities (xerostomia: OR = 1.028, dysgeusia: OR = 1.022, dysphagia: OR = 1.027) in the multivariable regression. The superior pharyngeal constrictor muscle was also significantly related to chronic dysphagia (OR = 1.030). Consecutively developed and internally validated NTCP models were predictive for the analyzed toxicities (optimism-corrected AUCs after bootstrapping: AUCxerostomia=0.64, AUCdysgeusia=0.60, AUCdysphagia=0.64). Our data suggest that the dose to the soft palate is associated with chronic moderate xerostomia, moderate dysgeusia and moderate-to-severe dysphagia in older HNSCC patients undergoing definitive radiotherapy. If validated in external studies, efforts should be undertaken to reduce the soft palate dose in these patients.

Predicting the risk of neurocognitive decline after brain irradiation in adult patients with a primary brain tumor.

Deterioration of neurocognitive function in adult patients with a primary brain tumor is the most concerning side effect of radiotherapy. This study aimed to develop and evaluate normal-tissue complication probability (NTCP) models using clinical and dose-volume measures for 6-month, 1-year, and 2-year Neurocognitive Decline (ND) postradiotherapy. A total of 219 patients with a primary brain tumor treated with radical photon and/or proton radiotherapy (RT) between 2019 and 2022 were included. Controlled oral word association test, Hopkins verbal learning test-revised, and trail making test were used to objectively measure ND. A comprehensive set of potential clinical and dose-volume measures on several brain structures were considered for statistical modeling. Clinical, dose-volume and combined models were constructed and internally tested in terms of discrimination (area under the curve, AUC), calibration (mean absolute error, MAE), and net benefit. Fifty percent, 44.5%, and 42.7% of the patients developed ND at 6-month, 1-year, and 2-year time points, respectively. The following predictors were included in the combined model for 6-month ND: age at radiotherapy > 56 years (OR = 5.71), overweight (OR = 0.49), obesity (OR = 0.35), chemotherapy (OR = 2.23), brain V20 Gy ≥ 20% (OR = 3.53), brainstem volume ≥ 26 cc (OR = 0.39), and hypothalamus volume ≥ 0.5 cc (OR = 0.4). Decision curve analysis showed that the combined models had the highest net benefits at 6-month (AUC = 0.79, MAE = 0.021), 1-year (AUC = 0.72, MAE = 0.027), and 2-year (AUC = 0.69, MAE = 0.038) time points. The proposed NTCP models use easy-to-obtain predictors to identify patients at high risk of ND after brain RT. These models can potentially provide a base for RT-related decisions and post-therapy neurocognitive rehabilitation interventions.

Prediction of auditory brain stem responses damage in patients with head-and-neck cancers receiving radiotherapy using the functional assays of normal tissue complication probability models.

The purpose of this study was to set four NTCP models on clinical data and develop a model that calculates the possibility of hearing damage due to irradiation of healthy and at-risk brainstem tissue. ABR tests were performed on 50 head-and-neck cancer patients three years after radiotherapy for evaluation of lesions in a part of the auditory nerve or the auditory pathway in the brainstem. It indicated a significant difference in the latency of the waves assessed by the ABR test between the two groups. The paired sample t-test indicated the latency time of waves I, III, V, I-III, and I-V (P < 0.001) in the right ear, and in the left ear latency time of waves III, V, I-III, I-V, and III-V (P < 0.001) were significantly higher in the case group's ear than those in the control group. The confidence interval of the fitted parameters was 95% for NTCP models. ABR test's binary outcome with differential dose-volume histograms (dDVHs) was calculated and imported as input to the NTCP modeling. The values of the parameters n = 2.3-2.9 and the value s = 1 were obtained, which indicated that the brainstem organ is seriality. The best model ranked for the prediction of brainstem hearing damage was the logit model, which had the lowest Akaike value. The nervousness of the auditory organ of the brainstem (VIII nerve) can be declared as one of the reasons for being independent of the received dose.

A deep learning model for translating CT to ventilation imaging: analysis of accuracy and impact on functional avoidance radiotherapy planning.

Radiotherapy planning incorporating functional lung images has the potential to reduce pulmonary toxicity. Free-breathing 4DCT-derived ventilation image (CTVI) may help quantify lung function. This study introduces a novel deep-learning model directly translating planning CT images into CTVI. We investigated the accuracy of generated images and the impact on functional avoidance planning. Paired planning CT and 4DCT scans from 48 patients with NSCLC were randomized to training (n = 41) and testing (n = 7) data sets. The ventilation maps were generated from 4DCT using a Jacobian-based algorithm to provide ground truth labels (CTVI4DCT). A 3D U-Net-based model was trained to map CT to synthetic CTVI (CTVISyn) and validated using fivefold cross-validation. The highest-performing model was applied to the testing set. Spearman's correlation (rs) and Dice similarity coefficient (DSC) determined voxel-wise and functional-wise concordance between CTVI4DCT and CTVISyn. Three plans were designed per patient in the testing set: one clinical plan without CTVI and two functional avoidance plans combined with CTVI4DCT or CTVISyn, aimed at sparing high-functional lungs defined as the top 50% of the percentile ventilation ranges. Dose-volume (DVH) parameters regarding the planning target volume (PTV) and organs at risk (OARs) were recorded. Radiation pneumonitis (RP) risk was estimated using a dose-function (DFH)-based normal tissue complication probability (NTCP) model. CTVISyn showed a mean rs value of 0.65 ± 0.04 compared to CTVI4DCT. Mean DSC values over the top 50% and 60% of ventilation ranges were 0.41 ± 0.07 and 0.52 ± 0.10, respectively. In the test set (n = 7), all patients' RP-risk benefited from CTVI4DCT-guided plans (Riskmean_4DCT_vs_Clinical: 29.24% vs. 49.12%, P = 0.016), and six patients benefited from CTVISyn-guided plans (Riskmean_Syn_vs_Clinical: 31.13% vs. 49.12%, P = 0.022). There were no significant differences in DVH and DFH metrics between CTVISyn and CTVI4DCT-guided plan (P > 0.05). Using deep-learning techniques, CTVISyn generated from planning CT exhibited a moderate-to-high correlation with CTVI4DCT. The CTVISyn-guided plans were comparable to the CTVI4DCT-guided plans, effectively reducing pulmonary toxicity in patients while maintaining acceptable plan quality. Further prospective trials are needed to validate these findings.

49 Development of a Normal Tissue Complication Probability Model for Dysphagia in PATHOS trial patients

49 Development of a Normal Tissue Complication Probability Model for Dysphagia in PATHOS trial patients