Comprehensive toxicity risk profiling in radiation therapy for head and neck cancer: A new concept for individually optimised treatment

Clinical and Dosimetric Predictors of Early Onset Postradiation Hypothyroidism in Patients with Head and Neck Malignancies: A Logistic Regression Analysis

INTRODUCTION

Hypothyroidism commonly occurs as a side effect following radiotherapy for head and neck malignancies, yet limited information exists to predict the risk of postradiation hypothyroidism. This study aims to investigate the clinical and dosimetric factors that may predict early onset postradiation hypothyroidism (EO-PRH).

METHODS

A retrospective study was conducted on patients with head and neck cancer treated between 2018 and 2021, with a minimum follow-up duration of 12 months. The thyroid gland was contoured on computed tomography (CT) scans, and dose-volume histograms were analyzed, incorporating thyroid volume and V5-60 into the analysis. Logistic regression and receiver operating characteristic (ROC) analysis were performed to identify predictors and assess the model's predictive value.

RESULTS

Among the 84 eligible patients, 17 (20.2%) developed hypothyroidism within 1 year. The percentage of thyroid volume receiving 30 Gy (V30) emerged as the sole significant dosimetric predictor of EO-PRH (odds ratio [OR] 1.02, 95% confidence interval [95% CI] 1.005-1.05, p = 0.03). Univariable analysis revealed significant differences in cancer histopathology, primary tumor site, V15,30, and VS15,30 (the volume of the thyroid spared from radiation doses 15 Gy and 30 Gy) between the hypothyroid and euthyroid groups (p < 0.10). Multivariable analysis identified the primary cancer site (OR 9.09, 95% CI 1.59-100) and V30 (OR 1.26, 95% CI 1.007-1.76) as independent significant variables predicting EO-PRH. The predictive model incorporating cancer histopathology, primary tumor site, V15,30, and VS15,30 effectively predicted postradiation thyroid dysfunction (area under the receiver operating characteristic curve [AUC-ROC] 0.84, 95% CI 0.73-0.95, p < 0.001).

CONCLUSIONS

V30 could serve as a dosimetric predictor of hypothyroidism following neck radiotherapy. This study underscores that a predictive model encompassing cancer type and site, along with V15,30 and VS15,30, can effectively predict EO-PRH.

Normal tissue complication probability model predicting taste impairment in head and neck cancer patients: Evaluating the taste bud bearing tongue mucosa as a predictor

BACKGROUND/PURPOSE

Taste impairment is a common yet complex toxicity of head and neck cancer (HNC) radiotherapy treatment that may affect quality of life of survivors. This study aimed to predict acute and late taste impairment using taste bud bearing tongue mucosa as a new taste-specific organ-at-risk compared to full oral cavity as identified in previous studies.

MATERIALS/METHODS

Included HNC patients were treated with curative radiotherapy between 2007 and 2022. The endpoint was patient-rated moderate-to-severe taste loss scored with the EORTC QLQ-H&N35. The new tongue mucosa structure was derived from the existing oral cavity structure in accordance with published guidelines. An auto-segmentation tool was developed and verified by comparison to manually delineated structures. The performance of the mean dose admitted to this new structure was evaluated with both univariable analysis and a refit of a reference NTCP model substituting the oral cavity with the tongue mucosa.

RESULTS

A total of 691 HNC patients were included. Good conformity between manually delineated and auto-segmented structures was observed with no significant differences in mean dose (22.2 Gy vs. 22.1 Gy) or volume (20.7 cm3 vs. 20.3 cm3). Full oral cavity mean dose showed comparable effect size in univariable analysis compared to tongue mucosa mean dose. The NTCP model with tongue mucosa did not outperform the reference model with oral cavity for any evaluated time points.

CONCLUSION

The tongue mucosa mean dose did not outperform the oral cavity mean dose in the logistic regression NTCP model predicting acute and late taste impairment.

Evaluation of a comprehensive set of normal tissue complication probability models for patients with head and neck cancer in an international cohort

BACKGROUND/PURPOSE

Normal tissue complication probability (NTCP) models can be used to guide radiation therapy (RT) decisions by estimating side-effect risks pretreatment to minimize (late) side-effects. Recently, a comprehensive individual toxicity risk (CITOR) profile of NTCP models addressing common side-effects in head and neck cancer (HNC) patients was developed. This study investigates the generalizability of these models in an international setting, with different treatment approaches and side-effect assessments, promoting their integration into more widespread clinical practice.

MATERIALS/METHODS

From a prospective registry study, 407 HNC patients were included who were treated with definitive RT with or without systemic therapy between 2015 and 2022. NTCP models predicting dysphagia, aspiration, xerostomia, sticky saliva, taste loss, speech problems, oral pain, and fatigue at 6 and 12 months after RT were evaluated. All side-effects were patient-rated using the MDASI-HN, except dysphagia which was reported by clinicians using the PSS-HN diet normalcy score. Model performance was appraised by discrimination (area under the curve [AUC]) and calibration.

RESULTS

CITOR models showed moderate-to-high performance in this cohort (mean AUC = 0.67[range = 0.55-0.80], moderate-to-good calibration). NTCP models for dysphagia, xerostomia, sticky saliva, and fatigue were the top performing models. Models for aspiration, taste loss and speech problems performed moderately well, which was partly explained by lower incidences.

CONCLUSION

Despite differences between the CITOR development and this evaluation cohort, including use of different side-effect scoring systems, most models exhibited moderate-to-high performance. This demonstrated that the dose-effect relations were generalizable. Therefore, this study supports further integration of these NTCP models in clinical practice.

Improved NTCP model for late radiation-induced aspiration based on dose delivered to specific aspiration-related OARs

BACKGROUND AND PURPOSE

Radiation-induced aspiration is a serious complication following (chemo)radiation for head and neck cancer. The standard set of swallowing organs at risk (SWOARs) does not include all aspiration-related organs (OARs). An alternative proposed in earlier work includes a definition and delineation atlas for additional OARs, called Functional Swallowing Units (FSU). The purpose of this study was to compare two NTCP models for late aspiration, based on either SWOARs only or the FSU concept.

METHODS AND MATERIALS

Data from 189 patients were analysed. Aspiration at baseline (Asp_T0) and 6 months after treatment (Asp_T6) were scored according to Penetration Aspiration Scale (PAS). All SWOARs and FSUs were delineated and the DVH was recorded. Clinical factors and average dose (Dmean) to all aspiration-related OARs were included in multivariable analysis. Two models were built, model 1: including clinical factors and SWOARs only and model 2: including clinical factors, SWOARs and FSUs.

RESULTS

Both final models included Asp_T0 and Dmean to the supraglottic larynx as predictors. Model 2 included the dose to three additional OARs as a predictor: 1) Anterior Segment (floor of mouth/ thyrohyoid muscles) 2) hyoglossus/styloglossus muscles complex (HSG) 3) upper oesophageal sphincter (UES). Adding FSUs to model 1 resulted in significant model updates and model 2 performed better than model 1 (AUC 0.79 vs. 0.75).

CONCLUSION

NTCP models for late aspiration may be improved by including the dose to aspiration-related OARs, defined by the FSU concept. In addition to the supraglottic larynx, sparing of the Anterior Segment, HSG and UES could further decrease the risk of radiation-induced aspiration, but this remains to be confirmed in clinical studies.

Assessing the Impact of Nutritional Status on the Quality of Life in Head and Neck Cancer Patients-The Need for Comprehensive Digital Tools

Background/Objectives: Malnutrition is a key determinant of quality of life (QoL) in patients with head and neck cancers (HNCs), influencing treatment outcomes and the occurrence of adverse events (AEs). Despite there being numerous studies on nutritional status and QoL, there is no standardized risk or prognostic model integrating clinical and demographic factors. Methods: A literature search was conducted in September 2024 in Scopus, PubMed, and Web of Science, covering studies published between 2013 and 2024. Articles were selected based on their relevance to AEs, nutritional interventions, and QoL assessments in HNC patients. Results: The key factors influencing QoL in HNC patients include age, sex, weight, BMI, educational level, and tumor features. Mucositis was identified as the most significant food intake-impairing AE, contributing to malnutrition and reduced QoL. Current QoL assessments rely on descriptive questionnaires, which lack personalization and predictive capabilities. Digital tools, including machine learning models and digital twins, offer potential solutions for risk prediction and personalized nutritional interventions. Conclusions: Despite significant research efforts, QoL assessment in HNC patients remains non-uniform, and risk models integrating nutritional status are lacking. A comprehensive, personalized approach is needed, leveraging digital tools to improve nutritional intervention strategies.

Geometric and Dosimetric Evaluation of a RayStation Deep Learning Model for Auto-Segmentation of Organs at Risk in a Real-World Head and Neck Cancer Dataset

AIMS

To assess geometric accuracy and dosimetric impact of a deep learning segmentation (DLS) model on a large, diverse dataset of head and neck cancer (HNC) patients treated with intensity-modulated proton therapy (IMPT).

MATERIALS AND METHODS

A 3D U-Net-based DLS model was applied to CT datasets of 124 HNC patients treated with IMPT at 50.4-70.0 GyRBE. Thirty organs-at-risk (OARs), delineated manually (GT-OARs) were analysed for similarity metrics with auto-segmented OARs, without (DLS-nonedited) and with (DLS-edited) manual correction, using volume, Dice similarity coefficient (DSC), and Hausdorff distance (HD). Dosimetric impact of auto-segmentation error was assessed as absolute dose difference of mean (ΔDmean) and maximum (ΔDmax).

RESULTS

The cohort includes patients with postoperative (47.6%), flap reconstruction (12.1%), mouth bites (79.8%), dental implants (54.8%), and surgical implants (3.2%). DLS failed in 11 patients with significant anatomical challenges and artifact. Compared with GT-OARs, DLS-nonedited under-segmented 11/12 Gr-A (central nervous system, arteries, bone) (p < 0.05) and over-segmented 13/18 Gr-B (glandular, digestive, airways) OARs. DSC scores were good (>0.8), intermediate (0.6-0.8), intermediate-poor (0.5-0.6), and poor (<0.5) in 12, 6, 4, and 8 OARs. HD were good (<4mm), intermediate (4-6mm), poor (6-8mm), and very poor (>8mm) in 5, 7, 4, and 14 OARs. Compared with manually corrected, DLS-edited OARs, all DLS-nonedited OARs demonstrated excellent similarity with DSC>0.8 and HD<4mm. On average, auto-segmentation took 2.51 minutes, while correction took 6.24 minutes. The mean values of ΔDmean and ΔDmax were within ±300 and ±3 cGyRBE, except for oesophagus and larynx, where the mean ΔDmean increases up to 837.14 cGyRBE.

CONCLUSION

Patient posture, nonbiological materials, and anatomical deformities influence DLS accuracy. The model's overall performance is adequate and efficient with skilled manual editing needed for few OARs.

Evaluation of machine learning models for predicting xerostomia in adults with head and neck cancer during proton and heavy ion radiotherapy

BACKGROUND AND PURPOSE

Few studies have examined the factors associated with xerostomia during proton and carbon ion radiotherapy for head and neck cancer (HNC), which are reported to have fewer toxic effects compared to traditional photon-based radiotherapy. This study aims to evaluate the performance of machine learning approaches in predicting grade 2 + xerostomia in adults with HNC receiving proton and carbon ion radiotherapy.

MATERIALS AND METHODS

A retrospective study involving 1,769 adults with HNC who completed proton or carbon ion radiotherapy was conducted. Xerostomia was graded using the Radiation Therapy Oncology Group criteria. Eight machine learning models with different combinations sampling methods and class weights were compared to identify the model with the highest balanced accuracy.

RESULTS

The mean age of patients was 47.8 years (range 18-80), with 33.5 % female. The average total radiation dose was 71.0 GyE (SD = 5.7). Grade 1 xerostomia was recorded in 572 patients (32.3 %) and grade 2 in 103 patients (5.8 %). No cases of grade 3 or higher xerostomia were reported. A support vector machine with a linear kernel, a 1:2 positive-to-negative class weight, and SMOTE oversampling achieved the highest balanced accuracy (0.66) and AUC-ROC (0.69) for predicting grade 2 xerostomia, outperforming the logistic regression model (balanced accuracy:0.50, AUC-ROC. 0.67).

CONCLUSION

The prevalence of grade 2 radiation-induced xerostomia during proton and carbon ion radiotherapy was low in adults with HNC, posing challenges for accurate prediction. Further research is needed to develop improved methods for predicting xerostomia during proton and carbon ion radiotherapy.

In Silico Interim Adaptation of Proton Therapy in Head and Neck Cancer by Simultaneous Dose and Linear Energy Transfer Escalation

PURPOSE

The outcome of proton therapy for head and neck cancer (HNC) varies considerably. We investigated the feasibility of adapting proton therapy plans based on 18F-fluorodeoxyglucose-positron emission tomography-defined biologic tumor volumes (BTVs) reflecting remaining aggressive tumor subvolumes 2 weeks into treatment (interim). Recognizing the potential to improve proton therapy response with increasing linear energy transfer (LET), we simulated a combined dose-LET escalation to the BTVs and compared it to pure dose escalation. In addition, the impact of relative biological effectiveness (RBE) was evaluated by comparing the constant RBE of 1.1 (RBE1.1) with a variable-RBE model.

METHODS AND MATERIALS

A semiautomated method was used to segment the BTV from 18F-fluorodeoxyglucose-positron emission tomography-defined for 9 patients with HNC, assuming high standardized uptake value at interim to reflect tumor radioresistance. An in-house Monte Carlo-based recalculation and reoptimization tool simulated proton therapy plans with both constant RBE1.1 and variable-RBE, aimed to deliver 68 Gy (RBE) to high-risk target volumes, 10% dose escalation to the BTV, and a LET boost to the BTV. Dose distributions were prioritized over LET optimization goals. Results were quantified by dose and LET distributions to target volumes and organs at risk, as well as normal tissue complication probabilities (NTCPs) for xerostomia and dysphagia.

RESULTS

Dose-LET adapted proton therapy plans achieved 10% dose escalation and mean dose-averaged LET (LETd) increases to the BTV above 1.0 keV/μm, with no significant LET increases to organs at risk. NTCP for xerostomia and dysphagia from dose-LET and dose-only escalation were similar. However, NTCPs increased 6% to 10% when variable-RBE was used instead of the constant RBE1.1.

CONCLUSIONS

Our in silico study showed that dose-LET escalation in proton therapy integrating a variable-RBE model may improve proton therapy for patients with HNC. Clinical evaluation of such a biological image-based dose-LET escalation in proton therapy of HNC remains to be investigated.

An automatic pipeline for temporal monitoring of radiotherapy-induced toxicities in head and neck cancer patients

Radiotherapy for head and neck cancer often causes a spectrum of toxicities. Such toxicities are usually unavailable as structured data and are reported within textual clinical reports. To reduce the burden of manual assessment of toxicities, we propose a language processing model for the automatic extraction of toxicities. The cohort consists of 384 patients with head and neck cancer who underwent radiotherapy, either as monotherapy or in combination with chemotherapy or surgery. A total of 3510 notes were extracted. The toxicities were then manually annotated. Two tasks of toxicity mention detection and toxicity extraction were defined. Pre-trained language models such as BERT, Clinical BioBERT, and Clinical Longformer were fine-tuned. Our best model achieves an F1 score of 90% for automatic extraction of toxicity mentions. An automatic system enables real-time extraction of toxicities and insights into their temporal patterns, offering actionable data to support dose optimization and minimize toxicities in personalized treatments.

Predictive biomarkers of radiotherapy- related dermatitis, xerostomia, mucositis and dysphagia in head and neck cancer: A systematic review

BACKGROUND

Radiotherapy is essential for treating head and neck cancer but often leads to severe toxicity. Traditional predictors include anatomical location, tumor extent, and dosimetric data. Recently, biomarkers have been explored to better predict and understand toxicity. This review aims to summarize the current literature, assess data quality, and guide future research.

METHODS

Two reviewers independently screened EMBASE and PubMed for studies published between 2010 and 2023. Endpoints were dermatitis, mucositis, sticky saliva/xerostomia, and dysphagia. Statistical analysis was performed using R, and bias assessed via a modified QUIPS questionnaire. Pathway analysis was conducted using gProfiler. The study adhered to PRISMA and COSMOS-E guidelines and was registered in the PROSPERO database (#CRD42023361245).

RESULTS

Of 2,550 abstracts, 69 publications met the inclusion criteria. These studies involved a median of 81 patients, primarily male (75 %), with common primary tumors in the nasopharynx (32 %) and oropharynx (27 %). Most patients (84 %) had advanced disease (stage III/IV). The most frequently studied biomarkers were DNA-based single-nucleotide polymorphisms (SNPs, 59 %), salivary proteins (13 %), and bacteria (10 %). Ten statistically-significant biomarkers (all SNPs) in low-bias publications were identified, particularly in DNA repair and cell detoxification pathways. Data quality was often poor and few validation studies were present in the dataset.

CONCLUSION

This review provides an overview of the research landscape, highlights research gaps and provides recommendations for future research directions. We identified several potential biomarkers, particularly in DNA repair pathways, that align with current understanding of radiation-induced cell damage. However, the overall data quality was poor, with key clinical variables often missing. Overall, rigorous standardization of reporting, validation studies and multi-center collaborations to increase study power and sample sizes are necessary to build high-level evidence for clinical application.

Performance Comparison of 10 State-of-the-Art Machine Learning Algorithms for Outcome Prediction Modeling of Radiation-Induced Toxicity

Purpose

To evaluate the efficacy of prominent machine learning algorithms in predicting normal tissue complication probability using clinical data obtained from 2 distinct disease sites and to create a software tool that facilitates the automatic determination of the optimal algorithm to model any given labeled data set.

Methods and Materials

We obtained 3 sets of radiation toxicity data (478 patients) from our clinic: gastrointestinal toxicity, radiation pneumonitis, and radiation esophagitis. These data comprised clinicopathological and dosimetric information for patients diagnosed with non-small cell lung cancer and anal squamous cell carcinoma. Each data set was modeled using 11 commonly employed machine learning algorithms (elastic net, least absolute shrinkage and selection operator [LASSO], random forest, random forest regression, support vector machine, extreme gradient boosting, light gradient boosting machine, k-nearest neighbors, neural network, Bayesian-LASSO, and Bayesian neural network) by randomly dividing the data set into a training and test set. The training set was used to create and tune the model, and the test set served to assess it by calculating performance metrics. This process was repeated 100 times by each algorithm for each data set. Figures were generated to visually compare the performance of the algorithms. A graphical user interface was developed to automate this whole process.

Results

LASSO achieved the highest area under the precision-recall curve (0.807 ± 0.067) for radiation esophagitis, random forest for gastrointestinal toxicity (0.726 ± 0.096), and the neural network for radiation pneumonitis (0.878 ± 0.060). The area under the curve was 0.754 ± 0.069, 0.889 ± 0.043, and 0.905 ± 0.045, respectively. The graphical user interface was used to compare all algorithms for each data set automatically. When averaging the area under the precision-recall curve across all toxicities, Bayesian-LASSO was the best model.

Conclusions

Our results show that there is no best algorithm for all data sets. Therefore, it is important to compare multiple algorithms when training an outcome prediction model on a new data set. The graphical user interface created for this study automatically compares the performance of these 11 algorithms for any data set.

Risk reduction of radiation-induced aspiration by sparing specific aspiration-related-organs at risk; an in silico feasibility study

PURPOSE

To assess the feasibility and benefit of NTCP optimized aspiration-prevention treatment planning by sparing specific aspiration related organs at risk, and to assess the impact of baseline complaints on the planning results.

MATERIALS AND METHODS

This in silico planning study included 30 HNC patients who were previously treated with definitive radiotherapy. New fully automated plans, allowing for sparing specific aspiration related organs at risk, were optimised directly on normal tissue complication probability (NTCP) models for common toxicities: xerostomia and dysphagia. Optimisation was performed with and without aspiration-prevention, i.e., with and without specific sparing of recently identified aspiration-related muscles, and with and without the assumption of existing baseline complaints.

RESULTS

All plans complied with the pre-defined treatment planning quality criteria and were successful in limiting the risk of xerostomia and dysphagia. Aspiration-prevention VMAT, optimized using the additional NTCP model for aspiration, significantly reduced the estimated risk of late aspiration (p < 0.001) in all 30 patients when compared to plans without NTCP optimisation for late aspiration. The predicted risk of late aspiration was reduced even further when baseline aspiration was assumed present during optimisation, resulting in an average risk reduction of 13.3 % versus 8.3 % in plans assuming no aspiration at baseline. Aspiration-prevention did not reduce overall plan quality and maintained NTCP values obtained for various other toxicities.

CONCLUSION

Sparing specific aspiration-related organs at risk has the potential to significantly reduce the risk of late RT-induced aspiration, especially in patients who experience aspiration already at baseline.

Prediction of the need of enteral nutrition during radiation therapy for head and neck cancers

INTRODUCTION

Patients with a head and neck (HN) cancer undergoing radiotherapy risk critical weight loss and oral intake reduction leading to enteral nutrition. We developed a predictive model for the need for enteral nutrition during radiotherapy in this setting. Its performances were reported on a real-world multicentric cohort.

MATERIAL AND METHODS

Two models were trained on a prospective monocentric cohort of 230 patients. The first model predicted an outcome combining severe or early fast weight loss, or severe oral intake impairment (grade 3 anorexia or dysphagia or the prescription of enteral nutrition). The second outcome only combined oral intake impairment criteria. We trained a gradient boosted tree with a nested cross validation for Bayesian optimization on a prospective cohort and predictive performances were reported on the external multicentric real-world cohort of 410 patients from 3 centres. Predictions were explainable for each patient using Shapley values.

RESULTS

For the first and second outcome, the model yielded a ROC curve AUC of 81 % and 80%, an accuracy of 77 % and 77 %, a positive predictive value of 77 % and 72 %, a specificity of 78 % and 79 % and a sensitivity of 75 % and 73 %. The negative predictive value was 80 % and 80 %. For each patient, the underlying Shapley values of each clinical predictor to the prediction could be displayed. Overall, the most contributing predictor was concomitant chemotherapy.

CONCLUSION

Our predictive model yielded good performance on a real life multicentric validation cohort to predict the need for enteral nutrition during radiotherapy for HN cancers.

Toxicity and Quality of Life After Locoregional Radiotherapy in Patients With Thyroid Cancer

BACKGROUND

Locoregional external beam radiotherapy (EBRT) is selectively used in thyroid cancer patients to induce locoregional control. However, despite technological advances, EBRT remains associated with toxicities. We evaluated thyroid-cancer specific toxicities and long-term Quality of Life (QoL) post-EBRT.

METHODS

EBRT-treated thyroid cancer patients at Universal Medical Centre Groningen (2007-2023) were retrospectively evaluated (n = 66). Acute (< 6 weeks) and late (≥ 3 months) toxicities and QLQ-H&N35 results, prospectively collected as standard patient care, were analyzed (available in 24/66). Additionally, 17/66 living patients cross-sectionally completed the QLQ-H&N43 [renewed QLQ-H&N35] and SF-36-RAND-36.

RESULTS

In 24/66 patients who completed questionnaires during EBRT treatment, most severe acute toxicities occurred around week 6 (91% dermatitis, 74% pain, 70% hoarseness, 65% dysphagia). Late toxicities included persisting acute toxicities and fibrosis. Six months post-treatment, only QLQ-H&N35 domains "social eating" (p = 0.031) and "dry mouth/sticky saliva" (p = 0.025) were affected, in comparison to pre-radiation. In the 10/17 patients who completed the QLQ-H&N35 6 months post-radiation and the cross-sectionally performed QLQ-H&N43, no long-term mitigation of assessed domains was identified in a longitudinal analysis. The most advanced EBRT technique was associated with better QLQ-H&N43 scores (p = 0.047).

CONCLUSIONS

EBRT causes acute and late toxicities in most thyroid cancer patients and may be associated with a decreased QoL. As these patients generally survive for multiple years, there is a compelling need to minimize toxicities with more refined radiation techniques, such as proton therapy.

Radiomorphometric indices for measuring mandibular bone quality in oncologic patients

OBJECTIVE

This retrospective study compared the thickness and degree of resorption of the mandibular cortex in patients with head and neck cancer (AG), patients with cancer at sites other than the head and neck (BG), and patients with no cancer (CG) to describe and compare the changes in the mandible after antineoplastic therapy and their possible clinical implications.

MATERIALS AND METHODS

A total of 287 panoramic radiographs were examined. The following radiomorphometric indices were analyzed: mental index (MI), panoramic mandibular index (PMI), and mandibular cortical index (MCI). Analysis of variance (ANOVA) and the Kruskal‒Wallis test, with p < 0.05 considered significant, were performed.

RESULTS

Males predominated in the AG (83%), while females predominated in the BG and CG (78.6 and 62%, respectively). In the AG, tongue carcinoma (22.1%) was prevalent, while in the BG, breast carcinoma was predominant (53.8%). All parameters measured in the AG and BG patients were significantly lower than those in the CG patients: MI (p < 0.001), right PMIc/a (p < 0.001), left PMIc/a (p < 0.001), right PMIc/b (p = 0.004), left PMIc/b (p < 0.001), and MCI (p < 0.001).

CONCLUSIONS

Radiomorphometric indices MI, PMI, and MCI were significantly lower in panoramic radiographs of patients with head and neck cancer and patients with cancer in other regions of the body than in those of nononcological patients.

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

PURPOSE

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).

METHODS AND MATERIALS

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.

RESULTS

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).

CONCLUSION

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.

Clinical Introduction of Stem Cell Sparing Radiotherapy to Reduce the Risk of Xerostomia in Patients with Head and Neck Cancer

BACKGROUND/OBJECTIVES

Studies have shown that dose to the parotid gland stem cell rich (SCR) regions should be reduced to lower the risk of xerostomia after radiotherapy (RT). This study aimed to assess whether stem cell sparing (SCS)-RT can be adopted in routine clinical practice.

METHODS

Multiple planning studies were performed to compare SCS-RT with standard (ST)-RT using 30 head and neck cancer patients. Shifts in mean dose to the SCR regions (Dmean,SCR) and other organs at risk and their estimated impact on normal tissue complication probability (NTCP) for side-effects were compared using Wilcoxon signed-rank test. A multicenter study was performed (eight institutions, three patients) to test the generalizability of SCS-RT using the Friedman test.

RESULTS

Using photons, Dmean,SCR was reduced with median 4.1/3.5 Gy for ipsilateral/contralateral (p < 0.001). The largest reductions were when the SCR regions overlapped less with target volumes. Subsequently, NTCPs for xerostomia decreased (p < 0.001). Using protons, Dmean,SCR was also reduced (2.2/1.9 Gy for ipsilateral/contralateral, p < 0.002). Nevertheless, SCS-RT did not further decrease NTCPs for xerostomia (p > 0.17). Target coverage and prevention of other side-effects were not compromised. However, increased mean oral cavity dose was observed in some patients. Lastly, in the multicenter study Dmean,SCR could be reduced by slightly adjusting the standard optimization. Contralateral Dmean,SCR reductions differed between centers (p = 0.01), which was attributed to differences in ST-RT plans.

CONCLUSIONS

Stem cell sparing radiotherapy can be clinically introduced by making small adjustments to the optimization strategy and can reduce the risk of xerostomia.

Malnutrition and radiation therapy in head and neck cancers, a systematic review on reported definitions and associated factors

Radiation therapy is a major treatment in head and neck cancers that can induce mucositis, pain, and dysgeusia that could impair oral intake and lead to weight loss and malnutrition. Intensity modulation has diminished toxicity of radiation therapy. We performed a review to assess the rate of malnutrition and how malnutrition was defined across cohorts of patients undergoing modern curative radiation therapy. We performed a systematic review of prospective cohorts to assess how was defined malnutrition and severe malnutrition as binary outcomes and to report their rates depending on their definition. We screened 250 papers and included 27 papers in the review. Only two studies reported malnutrition using the Global Leadership Initiative on Malnutrition (GLIM) criteria, and the most reported criteria were PG SGA (patient-graded subjective global assessment) B or C and weight loss above 10%. The most frequently reported factors associated with malnutrition were concomitant chemotherapy, tumoral stage, and tumor site. Our review highlighted the major heterogeneity of the reporting of patient's nutritional state across cohorts of head and neck cancers treated by modern curative radiotherapy. Using consensual definition of malnutrition would help to provide stronger evidence on preventive enteral nutrition as well as the causes and consequences of malnutrition in head and neck cancers.

A bioconvergence study on platinum-free concurrent chemoradiotherapy for the treatment of HPV-negative head and neck carcinoma

Locally advanced head and neck squamous cell carcinoma (LA-HNSCC) is characterized by high rate of recurrence, resulting in a poor survival. Standard treatments are associated with significant toxicities that impact the patient's quality of life, highlighting the urgent need for novel therapies to improve patient outcomes. On this regard, noble metal nanoparticles (NPs) are emerging as promising agents as both drug carriers and radiosensitizers. On the other hand, co-treatments based on NPs are still at the preclinical stage because of the associated metal-persistence.In this bioconvergence study, we introduce a novel strategy to exploit tumour chorioallantoic membrane models (CAMs) in radio-investigations within clinical equipment and evaluate the performance of non-persistent nanoarchitectures (NAs) in combination with radiotherapy with respect to the standard concurrent chemoradiotherapy for the treatment of HPV-negative HNSCCs. A comparable effect has been observed between the tested approaches, suggesting NAs as a potential platinum-free agent in concurrent chemoradiotherapy for HNSCCs. On a broader basis, our bioconvergence approach provides an advance for the translation of Pt-free radiosensitizer to the clinical practice, positively shifting the therapeutic vs. side effects equilibrium for the management of HNSCCs.

Cluster-based radiomics reveal spatial heterogeneity of bevacizumab response for treatment of radiotherapy-induced cerebral necrosis

Background

Bevacizumab is used in the treatment of radiation necrosis (RN), which is a debilitating toxicity following head and neck radiotherapy. However, there is no biomarker to predict if a patient would respond to bevacizumab.

Purpose

We aimed to develop a cluster-based radiomics approach to characterize the spatial heterogeneity of RN and map their responses to bevacizumab.

Methods

118 consecutive nasopharyngeal carcinoma patients diagnosed with RN were enrolled. We divided 152 lesions from the patients into 101 for training, and 51 for validation. We extracted voxel-level radiomics features from each lesion segmented on T1-weighted+contrast and T2 FLAIR sequences of pre- and post-bevacizumab magnetic resonance images, followed by a three-step analysis involving individual- and population-level clustering, before delta-radiomics to derive five radiomics clusters within the lesions. We tested the association of each cluster with response to bevacizumab and developed a clinico-radiomics model using clinical predictors and cluster-specific features.

Results

71 (70.3%) and 34 (66.7%) lesions had responded to bevacizumab in the training and validation datasets, respectively. Two radiomics clusters were spatially mapped to the edema region, and the volume changes were significantly associated with bevacizumab response (OR:11.12 [95% CI: 2.54-73.47], P = 0.004; and 1.63[1.07-2.78], P = 0.042). The combined clinico-radiomics model based on textural features extracted from the most significant cluster improved the prediction of bevacizumab response, compared with a clinical-only model (AUC:0.755 [0.645-0.865] to 0.852 [0.764-0.940], training; 0.708 [0.554-0.861] to 0.816 [0.699-0.933], validation).

Conclusion

Our radiomics approach yielded intralesional resolution, enabling a more refined feature selection for predicting bevacizumab efficacy in the treatment of RN.

Clinical implementation of deep learning robust IMPT planning in oropharyngeal cancer patients: A blinded clinical study

BACKGROUND AND PURPOSE

This study aimed to evaluate the plan quality of our deep learning-based automated treatment planning method for robustly optimized intensity-modulated proton therapy (IMPT) plans in patients with oropharyngeal carcinoma (OPC). The assessment was conducted through a retrospective and prospective study, blindly comparing manual plans with deep learning plans.

MATERIALS AND METHODS

A set of 95 OPC patients was split into training (n = 60), configuration (n = 10), test retrospective study (n = 10), and test prospective study (n = 15). Our deep learning optimization (DLO) method combines IMPT dose prediction using a deep learning model with a robust mimicking optimization algorithm. Dosimetrists manually adjusted the DLO plan for individual patients. In both studies, manual plans and manually adjusted deep learning (mDLO) plans were blindly assessed by a radiation oncologist, a dosimetrist, and a physicist, through visual inspection, clinical goal evaluation, and comparison of normal tissue complication probability values. mDLO plans were completed within an average time of 2.5 h. In comparison, the manual planning process typically took around 2 days.

RESULTS

In the retrospective study, in 10/10 (100%) patients, the mDLO plans were preferred, while in the prospective study, 9 out of 15 (60%) mDLO plans were preferred. In 4 out of the remaining 6 cases, the manual and mDLO plans were considered comparable in quality. Differences between manual and mDLO plans were limited.

CONCLUSION

This study showed a high preference for mDLO plans over manual IMPT plans, with 92% of cases considering mDLO plans comparable or superior in quality for OPC patients.

Radiation-induced Xerostomia is Related to Stem Cell Dose-dependent Reduction of Saliva Production

PURPOSE

Previous studies have shown that the mean dose to the parotid gland stem cell rich regions (Dmean,SCR) is the strongest dosimetric predictor for the risk of patient-reported daytime xerostomia. This study aimed to test whether the relationship between patient-reported xerostomia and Dmean,SCR is explained by a dose-dependent reduction of saliva production.

METHODS AND MATERIALS

In 570 patients with head and neck cancer treated with definitive radiation therapy (RT), flow from the parotid (FLOWPAR) and submandibular/sublingual (FLOWSMSL) glands, and patient-reported daytime (XERDAY) and nighttime (XERNIGHT) xerostomia were prospectively measured before, at 6 months, and 12 months after RT. Using linear mixed effect models, the relationship of the mean dose to the parotid glands (Dmean,par), Dmean,SCR, non-SCR parotid gland tissue (Dmean,non-SCR), submandibular glands (Dmean,sub), and oral cavity (Dmean,oral) with salivary flow and xerostomia was analyzed while correcting for known confounders.

RESULTS

Dmean,SCR proved to be responsible for the effect of Dmean,par on FLOWPAR (P ≤ .03), while Dmean,non-SCR did not affect FLOWPAR (P ≥ .11). To illustrate, increasing Dmean,SCR by 10 Gy at a fixed Dmean,non-SCR reduced FLOWPAR by 0.02 mL/min (25%) after RT. However, if the opposite happened, no change in FLOWPAR was observed (0.00 mL/min [4%]). As expected, Dmean,sub was significantly associated with FLOWSMSL (P < .001). For example, increasing Dmean,sub by 10 Gy reduced FLOWSMSL by 0.07 mL/min (26%) after RT. Xerostomia scores were also affected by dose to the salivary glands. Dmean,SCR and Dmean,oral were associated with higher XERDAY scores (P ≤ .05), while Dmean,sub increased XERNIGHT scores (P = .01). For example, an increase of 10 Gy in Dmean,SCR raised XERDAY scores by 2.13 points (5%) after RT, while an additional 10 Gy in Dmean,subs increased XERNIGHT scores by 2.20 points (6%) after RT. Salivary flow was not only associated with radiation dose, but also with xerostomia scores in line with the salivary glands' functions; ie, FLOWPAR only influenced XERDAY (P < .001, 10.92 points lower XERDAY per 1 mL/min saliva), while FLOWSMSL affected XERDAY and XERNIGHT (P ≤ .004, 6.69 and 5.74 points lower XERDAY and XERNIGHT, respectively, per 1 mL/min saliva). Therefore, the observed relationships between dose and xerostomia were corrected for salivary flow. As hypothesized, Dmean,SCR only increased XERDAY scores via reducing FLOWPAR, whereas the effects of Dmean,oral on XERDAY and Dmean,sub on XERNIGHT were independent of salivary flow.

CONCLUSIONS

Higher SCR region dose reduced parotid gland saliva production, subsequently resulting in higher daytime xerostomia scores. Consequently, this study supports the clinical implementation of stem cell sparing RT to preserve salivary flow with the aim of reducing the risk of xerostomia.

Radiation Therapy for HPV-Positive Oropharyngeal Squamous Cell Carcinoma: An ASTRO Clinical Practice Guideline

PURPOSE

Human Papilloma Virus (HPV)-associated oropharyngeal squamous cell carcinoma (OPSCC) is a distinct disease from other head and neck tumors. This guideline provides evidence-based recommendations on the critical decisions in its curative treatment, including both definitive and postoperative radiation therapy (RT) management.

METHODS

ASTRO convened a task force to address 5 key questions on the use of RT for management of HPV-associated OPSCC. These questions included indications for definitive and postoperative RT and chemoradiation; dose-fractionation regimens and treatment volumes; preferred RT techniques and normal tissue considerations; and posttreatment management decisions. The task force did not address indications for primary surgery versus RT. Recommendations were based on a systematic literature review and created using a predefined consensus-building methodology and system for grading evidence quality and recommendation strength.

RESULTS

Concurrent cisplatin is recommended for patients receiving definitive RT with T3-4 disease and/or 1 node >3 cm, or multiple nodes. For similar patients who are ineligible for cisplatin, concurrent cetuximab, carboplatin/5-fluorouracil, or taxane-based systemic therapy are conditionally recommended. In the postoperative setting, RT with concurrent cisplatin (either schedule) is recommended for positive surgical margins or extranodal extension. Postoperative RT alone is recommended for pT3-4 disease, >2 nodes, or a single node >3 cm. Observation is conditionally recommended for pT1-2 disease and a single node ≤3 cm without other risk factors. For patients treated with definitive RT with concurrent systemic therapy, 7000 cGy in 33 to 35 fractions is recommended, and for patients receiving postoperative RT without positive surgical margins and extranodal extension, 5600 to 6000 cGy is recommended. For all patients receiving RT, intensity modulated RT over 3-dimensional techniques with reduction in dose to critical organs at risk (including salivary and swallowing structures) is recommended. Reassessment with positron emission tomography-computed tomography is recommended approximately 3 months after definitive RT/chemoradiation, and neck dissection is recommended for convincing evidence of residual disease; for equivocal positron emission tomography-computed tomography findings, either neck dissection or repeat imaging is recommended.

CONCLUSIONS

The role and practice of RT continues to evolve for HPV-associated OPSCC, and these guidelines inform best clinical practice based on the available evidence.

Radiation Dose-Induced Carotid Artery Stenosis and Brain Necrosis in Head and Neck Cancer-A Real World Cohort Study

Objective: This study aims to examine whether radiation therapy doses are related to incidences of carotid artery stenosis and brain necrosis in a large-scale real-world database. Methods: We identified a cohort of HNC patients from the catastrophic illness patient dataset using ICD-9 or ICD-10 to compare the incidence and risks of carotid artery stenosis (CAS) and brain necrosis (RIBN) in patients who received a radiation therapy dose of ≥5400 cGy/30 fractions (group A) with those who received a radiation therapy dose of <5400 cGy/30 fractions (group B). The incidence and hazard ratios were quantified using Cox proportional hazards models. Results: A total of 19,964 patients were identified in group A and group B. Among them, 965 and 863 cases of CAS and 435 and 359 cases of RIBN were identified in group A and group B, respectively. There was no statistically significant association between the two groups for CAS risk, whereas there was a statistically significant association between the two groups for RIBN risk. The most common primary site of head and neck cancers was the nasopharynx (1144 of 19,964, 5.73%). Conclusions: Our study suggests that RT may increase the risk of carotid stenosis and brain necrosis in patients with NPC. To ensure patient safety during treatment, the optimal balance between tumor control and toxicity prevention in individual patients through minimization of the radiation dose to all relevant OARs must be properly understood.

Cluster-Based Toxicity Estimation of Osteoradionecrosis Via Unsupervised Machine Learning: Moving Beyond Single Dose-Parameter Normal Tissue Complication Probability by Using Whole Dose-Volume Histograms for Cohort Risk Stratification

PURPOSE

Given the limitations of extant models for normal tissue complication probability estimation for osteoradionecrosis (ORN) of the mandible, the purpose of this study was to enrich statistical inference by exploiting structural properties of data and provide a clinically reliable model for ORN risk evaluation through an unsupervised-learning analysis that incorporates the whole radiation dose distribution on the mandible.

METHODS AND MATERIALS

The analysis was conducted on retrospective data of 1259 patients with head and neck cancer treated at The University of Texas MD Anderson Cancer Center between 2005 and 2015. During a minimum 12-month posttherapy follow-up period, 173 patients in this cohort (13.7%) developed ORN (grades I to IV). The (structural) clusters of mandibular dose-volume histograms (DVHs) for these patients were identified using the K-means clustering method. A soft-margin support vector machine was used to determine the cluster borders and partition the dose-volume space. The risk of ORN for each dose-volume region was calculated based on incidence rates and other clinical risk factors.

RESULTS

The K-means clustering method identified 6 clusters among the DVHs. Based on the first 5 clusters, the dose-volume space was partitioned by the soft-margin support vector machine into distinct regions with different risk indices. The sixth cluster entirely overlapped with the others; the region of this cluster was determined by its envelopes. For each region, the ORN incidence rate per preradiation dental extraction status (a statistically significant, nondose related risk factor for ORN) was reported as the corresponding risk index.

CONCLUSIONS

This study presents an unsupervised-learning analysis of a large-scale data set to evaluate the risk of mandibular ORN among patients with head and neck cancer. The results provide a visual risk-assessment tool for ORN (based on the whole DVH and preradiation dental extraction status) as well as a range of constraints for dose optimization under different risk levels.

Predict nutrition-related adverse outcomes in head and neck cancer patients undergoing radiotherapy: A systematic review

BACKGROUND

Acute nutrition-related adverse outcomes are common in head and neck cancer patients undergoing radiotherapy. Predictive models can assist in identifying high-risk patients to enable targeted intervention. We aimed to systematically evaluate predictive models for predicting severe acute nutritional symptoms, insufficient intake, tube feeding, sarcopenia, and weight loss.

METHODS

We searched PubMed, Web of Science, EBSCO, Embase, WanFang, CNKI, and SinoMed. We selected studies developing predictive models for the aforementioned outcomes. Data were extracted using a predefined checklist. Risk of bias and applicability assessment were assessed using the Prediction model Risk of Bias Assessment Tool. A narrative synthesis was conducted to summarize the model characteristics, risk of bias, and performance.

RESULTS

A total of 2941 studies were retrieved and 19 were included. Study outcome measure were different symptoms (n = 11), weight loss (n = 5), tube feeding (n = 3), and symptom or tube feeding (n = 1). Predictive factors mainly encompassed sociodemographic data, disease-related data, and treatment-related data. Seventeen studies reported area under the curve or C-index values ranging from 0.610 to 0.96, indicating moderate to good predictive performance. However, candidate predictors were incomplete, outcome measures were diverse, and the risk of bias was high. Most of them used traditional model development methods, and only two used machine learning.

CONCLUSIONS

Most current models showed moderate to good predictive performance. However, predictors are incomplete, outcome are inconsistent, and the risk of bias is high. Clinicians could carefully select the models with better model performance from the available models according to their actual conditions. Future research should include comprehensive and modifiable indicators and prioritize well-designed and reported studies for model development.

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

BACKGROUND

One 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.

METHODS

In 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).

DISCUSSION

This 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.

Validation of quantitative magnetic resonance imaging techniques in head and neck healthy structures involved in the salivary and swallowing function: Accuracy and repeatability

Background and Purpose

Radiation-induced damage to the organs at risk (OARs) in head-and-neck cancer (HNC) patient can result in long-term complications. Quantitative magnetic resonance imaging (qMRI) techniques such as diffusion-weighted imaging (DWI), DIXON for fat fraction (FF) estimation and T2 mapping could potentially provide a spatial assessment of such damage. The goal of this study is to validate these qMRI techniques in terms of accuracy in phantoms and repeatability in-vivo across a broad selection of healthy OARs in the HN region.

Materials and Methods

Scanning was performed at a 3 T diagnostic MRI scanner, including the calculation of apparent diffusion coefficient (ADC) from DWI, FF and T2 maps. Phantoms were scanned to estimate the qMRI techniques bias using Bland-Altman statistics. Twenty-six healthy subjects were scanned twice in a test-retest study to determine repeatability. Repeatability coefficients (RC) were calculated for the parotid, submandibular, sublingual and tubarial salivary glands, oral cavity, pharyngeal constrictor muscle and brainstem. Additionally, a linear mixed-effect model analysis was used to evaluate the effect of subject-specific characteristics on the qMRI values.

Results

Bias was 0.009x10-3 mm2/s for ADC, -0.7 % for FF and -7.9 ms for T2. RCs ranged 0.11-0.25x10-3 mm2/s for ADC, 1.2-6.3 % for FF and 2.5-6.3 ms for T2. A significant positive linear relationship between age and the FF and T2 for some of the OARs was found.

Conclusion

These qMRI techniques are feasible, accurate and repeatable, which is promising for treatment response monitoring and/or differentiating between healthy and unhealthy tissues due to radiation-induced damage in HNC patients.

Late-xerostomia prediction model based on 18F-FDG PET image biomarkers of the main salivary glands

BACKGROUND AND PURPOSE

Recently, a comprehensive xerostomia prediction model was published, based on baseline xerostomia, mean dose to parotid glands (PG) and submandibular glands (SMG). Previously, PET imaging biomarkers (IBMs) of PG were shown to improve xerostomia prediction. Therefore, this study aimed to explore the potential improvement of the additional PET-IBMs from both PG and SMG to the recent comprehensive xerostomia prediction model (i.e., the reference model).

MATERIALS AND METHODS

Totally, 540 head and neck cancer patients were split into training and validation cohorts. PET-IBMs from the PG and SMG, were selected using bootstrapped forward selection based on the reference model. The IBMs from both the PG and SMG with the highest selection frequency were added to the reference model, resulting in a PG-IBM model and a SMG-IBM model which were combined into a composite model. Model performance was assessed using the area under the curve (AUC). Likelihood ratio test compared the predictive performance between the reference model and models including IBMs.

RESULTS

The final selected PET-IBMs were 90th percentile of the PG SUV and total energy of the SMG SUV. The additional two PET-IBMs in the composite model improved the predictive performance of the reference model significantly. The AUC of the reference model and the composite model were 0.67 and 0.69 in the training cohort, and 0.71 and 0.73 in the validation cohort, respectively.

CONCLUSION

The composite model including two additional PET-IBMs from PG and SMG improved the predictive performance of the reference xerostomia model significantly, facilitating a more personalized prediction approach.

Unravelling Quality of Life for Head and Neck Cancer Patients after VMAT Radiation Therapy: Insights from Toxicity, Dosimetry and Symptoms Correlation

(1) Background: Head and neck cancer treatment, including advanced techniques like Volumetric Modulated Arc Therapy (VMAT), presents challenges for maintaining patient quality of life (QoL). Thus, thoroughly investigating how radiation therapy (RT) affects patients has been proved essential. Derived by that, this study aims to understand the complex interactions between not only RT and QoL but also symptom severity, and treatment-related toxicities in three distinct time points of patient's treatment; (2) Methods: To achieve that, EORTC-QLQ-C30 and EORTC QLQ-H&N35 questionnaires were used in combination with EORTC_RTOG scoring criteria and Spearman's rho statistical analysis for 74 patients with cancer undergoing VMAT radiation therapy; (3) Results: The results revealed a significant improvement in the Overall Health Index post-treatment, indicating a temporary decline during therapy followed by subsequent recovery, often surpassing pre-treatment QoL levels. Concurrently a reduction in symptomatology was observed, notably in pain, swallowing difficulties, and dry mouth, aligning with prior research indicating decreased symptom burden post-treatment. However, Spearman's correlation coefficient analysis at two distinct time points during therapy uncovered varying degrees of correlation between dosimetric data at Organs at Risk (OARs) and reported symptoms, highlighting potential limitations in using QoL questionnaires as sole indicators of treatment efficacy. Our investigation into the correlation between dosimetric data, toxicity, and symptoms focused on the relationship between radiation doses and oral mucositis levels, a common toxicity in head and neck cancer patients. Significant associations were identified between toxicity levels and dosimetric parameters, particularly with OARs such as the parotid glands, oral cavity, and swallowing muscles, underlining the utility of the EORTC method as a reliable toxicity assessment tool; (4) Conclusions: To summarize, current research attempts to underscore the importance of refining QoL assessments for enhanced patient care. The integration of dosimetric data, symptom severity, and treatment-related toxicities in the QoL outcomes of head and neck cancer patients undergoing VMAT radiation therapy, can lead towards the optimization of treatment strategies and the improvement of patient outcomes in future patient-centered radiation therapy practices.

Chronic fatigue in long-term survivors of head and neck cancer treated with radiotherapy

BACKGROUND

There is lack of evidence on chronic fatigue (CF) following radiotherapy (RT) in survivors of head and neck cancer (HNC). We aimed to compare CF in HNC survivors > 5 years post-RT with a reference population and investigate factors associated with CF and the possible impact of CF on health-related quality of life (HRQoL).

MATERIAL AND METHODS

In this cross-sectional study we included HNC survivors treated in 2007-2013. Participants filled in patient-reported outcome measures and attended a one-day examination. CF was measured with the Fatigue Questionnaire and compared with a matched reference population using t-tests and Cohen's effect size. Associations between CF, clinical and RT-related factors were investigated using logistic regression. HRQoL was measured with the EORTC Quality of Life core questionnaire.

RESULTS

The median age of the 227 HNC survivors was 65 years and median time to follow-up was 8.5 years post-RT. CF was twice more prevalent in HNC survivors compared to a reference population. In multivariable analyses, female sex (OR 3.39, 95 % CI 1.82-6.31), comorbidity (OR 2.17, 95 % CI 1.20-3.94) and treatment with intensity-modulated RT (OR 2.13, 95 % CI 1.16-3.91) were associated with CF, while RT dose parameters were not. Survivors with CF compared to those without, had significantly worse HRQoL.

CONCLUSIONS

CF in HNC survivors is particularly important for female patients, while specific factors associated with RT appear not to play a role. The high CF prevalence in long-term HNC survivors associated with impaired HRQoL is important information beneficial for clinicians and patients to improve patient follow-up.

Organs-at-risk dose and normal tissue complication probability with dynamic trajectory radiotherapy (DTRT) for head and neck cancer

We compared dynamic trajectory radiotherapy (DTRT) to state-of-the-art volumetric modulated arc therapy (VMAT) for 46 head and neck cancer cases. DTRT had lower dose to salivary glands and swallowing structure, resulting in lower predicted xerostomia and dysphagia compared to VMAT. DTRT is deliverable on C-arm linacs with high dosimetric accuracy.

Three-Dimensional-Printed Template-Guided Radioactive Seed Brachytherapy via a Submental Approach for Recurrent Base of Tongue and Floor of Mouth Cancer

Background

This study assessed clinical outcomes of three-dimensional-printed template (3DPT)-guided radioactive seed brachytherapy (RSBT) via a submental approach for recurrent base of tongue and floor of mouth cancer.

Methods

Thirty-one patients with recurrent lingual and floor of mouth squamous cell carcinoma after surgery and radiotherapy were treated with 3DPT-guided RSBT from 2015 to 2022. Seeds were implanted through a submental approach guided by 3DPTs. Local control (LC), overall survival (OS), disease control (DC) and quality of life (QOL) were evaluated.

Results

The median follow-up was 13.7 months. The 1-, 3- and 5-year LC rates were 66.1%, 66.1%, and 55.1% respectively. The 1-, 3- and 5-year OS rates were 63.4%, 33.4%, and 8.3%. The 1-, 3- and 5-year DC rates were 37.8%, 26.5%, and 21.2%. Univariate analysis showed tumor size significantly affected LC (P = 0.031). The presence of extraterritorial lesions affected DC and OS on multivariate analysis (P < 0.01). QOL improved significantly in domains of pain, swallowing, chewing, taste, and emotion after treatment compared to baseline. Four patients (13%) developed necrosis and osteoradionecrosis.

Conclusions

3DPT-guided submental RSBT provided favorable LC and QOL for recurrent tongue/floor of mouth cancer with minimal toxicity; moreover, severe toxicity should be noted.

Optimal timing of re-planning for head and neck adaptive radiotherapy

BACKGROUND AND PURPOSE

Adaptive radiotherapy (ART) relies on re-planning to correct treatment variations, but the optimal timing of re-planning to account for dose changes in head and neck organs at risk (OARs) is still under investigation. We aimed to find out the optimal timing of re-planning in head and neck ART.

MATERIALS AND METHODS

A total of 110 head and neck cancer patients were retrospectively enrolled. A semi auto-segmentation method was applied to obtain the weekly mean dose (Dmean) to OARs. The K-nearest-neighbour method was used for missing data imputation of weekly Dmean. A dose deviation map was built using the planning Dmean and weekly Dmean values and then used to simulate different ART scenarios consisting of 1 to 6 re-plannings. The difference between accumulated Dmean and planning Dmean before re-planning (ΔDmean_acc_noART) and after re-planning (ΔDmean_acc_ART) were evaluated and compared.

RESULTS

Among all the OARs, supraglottic showed the largest ΔDmean_acc_noART (1.23 ± 3.13 Gy) and most cases of ΔDmean_acc_noART > 3 Gy (26 patients). The 3rd week is suggested in the optimal timing of re-planning for 10 OARs. For all the organs except arytenoid, 2 re-plannings were able to guarantee the ΔDmean_acc_ART below 3 Gy while the average |ΔDmean_acc_ART| was below 1 Gy. ART scenarios of 2_4, 3_4, 3_5 (week of re-planning separated with "_") were able to guarantee ΔDmean_acc_ART of 99 % of patients below 3 Gy simultaneously for 19 OARs.

CONCLUSIONS

The optimal timing of re-planning was suggested for different organs at risk in head and neck adaptive radiotherapy. Generic scenarios of timing and frequency for re-planning can be applied to guarantee the increase of accumulated mean dose within 3 Gy simultaneously for multiple organs.

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

PURPOSE

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.

METHODS

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 3 Gy 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.

RESULTS

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).

CONCLUSIONS

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.

Development and Validation of a 3D Resnet Model for Prediction of Lymph Node Metastasis in Head and Neck Cancer Patients

The accurate diagnosis and staging of lymph node metastasis (LNM) are crucial for determining the optimal treatment strategy for head and neck cancer patients. We aimed to develop a 3D Resnet model and investigate its prediction value in detecting LNM. This study enrolled 156 head and neck cancer patients and analyzed 342 lymph nodes segmented from surgical pathologic reports. The patients' clinical and pathological data related to the primary tumor site and clinical and pathology T and N stages were collected. To predict LNM, we developed a dual-pathway 3D Resnet model incorporating two Resnet models with different depths to extract features from the input data. To assess the model's performance, we compared its predictions with those of radiologists in a test dataset comprising 38 patients. The study found that the dimensions and volume of LNM + were significantly larger than those of LNM-. Specifically, the Y and Z dimensions showed the highest sensitivity of 84.6% and specificity of 72.2%, respectively, in predicting LNM + . The analysis of various variations of the proposed 3D Resnet model demonstrated that Dual-3D-Resnet models with a depth of 34 achieved the highest AUC values of 0.9294. In the validation test of 38 patients and 86 lymph nodes dataset, the 3D Resnet model outperformed both physical examination and radiologists in terms of sensitivity (80.8% compared to 50.0% and 91.7%, respectively), specificity(90.0% compared to 88.5% and 65.4%, respectively), and positive predictive value (77.8% compared to 66.7% and 55.0%, respectively) in detecting individual LNM + . These results suggest that the 3D Resnet model can be valuable for accurately identifying LNM + in head and neck cancer patients. A prospective trial is needed to evaluate further the role of the 3D Resnet model in determining LNM + in head and neck cancer patients and its impact on treatment strategies and patient outcomes.

Oncologic outcome with versus without target volume compartmentalization in postoperative radiotherapy for oral cavity squamous cell carcinoma

Background and purpose

The volume treated with postoperative radiation therapy (PORT) in patients with oral cavity squamous cell carcinoma (OCSCC) is a mediator of toxicity affecting quality of life. Current guidelines only allow for very limited reduction of PORT volumes. This study investigated the safety and efficacy of de-intensified PORT for patients with OCSCC by refined compartmentalization of the treatment volume.

Materials and methods

This retrospective cohort study identified 103 OCSCC patients treated surgically from 2014 to 2019 with a loco-regional risk profile qualifying for PORT according to guidelines. PORT was administered only to the at-risk compartment and according to a refined compartmentalization concept (CC). Oncological outcome of this CC cohort was compared to a historical cohort (HC) of 98 patients treated before the CC was implemented.

Results

Median follow-up time was 4.5 and 4.8 years in the CC and HC cohorts, respectively. In the CC cohort, a total of 72 of 103 patients (70%) had a pathological risk profile that allowed for further compartmentalization and, hence, received a reduced treatment volume or omission of PORT altogether. Loco-regional control at 3 and 5 years was 77% and 73% in the CC cohort versus 78% and 73% in the HC (p = 0.93), progression-free survival was 72% and 64% versus75% and 68% (p = 0.58), respectively. Similarly, no statistically significant difference was seen in other outcome measures.

Conclusions

De-intensified PORT limiting the treatment volume to the at-risk compartment or avoiding PORT altogether for low-risk patients with OCSCC does not seem to compromise disease control in this retrospective comparison. Based on these hypothesis-generating findings, a prospective study is being planned.

Predicting severe radiation-induced oral mucositis in head and neck cancer patients using integrated baseline CT radiomic, dosimetry, and clinical features: A machine learning approach

Purpose

To establish the early prediction models of radiation-induced oral mucositis (RIOM) based on baseline CT-based radiomic features (RFs), dosimetric data, and clinical features by machine learning models for head and neck cancer (HNC) patients.

Methods

In this single-center prospective study, 49 HNCs treated with curative intensity modulated radiotherapy (IMRT) were enrolled. Baseline CT images (i.e., CT simulation), dosimetric, and clinical features were collected. RIOM was assessed using CTCAE v.5.0. RFs were extracted from manually-contoured oral mucosa structures. Minimum-redundancy-maximum-relevance (mRMR) method was applied to select the most informative radiomics, dosimetric, and clinical features. Then, binary prediction models were constructed for predicting acute RIOM based on the top mRMR-ranked radiomics, dosimetric, and clinical features alone or in combination, using random forest classifier algorithm. The predictive performance of models was assessed using the area under the receiver operating curve (AUC), accuracy, weighted-average based sensitivity, precision, and F1-measure.

Results

Among extracted features, the top 10 RFs, the top 5 dose-volume features, and the top 5 clinical features were selected using mRMR method. The model exploiting the integrated features (10-radiomics + 5-dosimetric + 5-clinical) achieved the best prediction with AUC, accuracy, sensitivity, precision, and F1-measure values of 91.7 %, 90.0 %, 83.0 % 100.0 %, and 91.0 %, respectively. The model developed using baseline CT RFs alone provided the best performance compared to dose-volume features or clinical features alone, with an AUC of 87.0 %.

Conclusion

Our results suggest that the integration of baseline CT radiomic features with dosimetric and clinical features showed promising potential to improve the performance of machine learning models in early prediction of RIOM. The ultimate goal is to personalize radiotherapy for HNC patients.

Radiation-induced long-term dysphagia in survivors of head and neck cancer and association with dose-volume parameters

BACKGROUND

Although dysphagia is a common side effect after radiotherapy (RT) of head and neck cancer (HNC), data on long-term dysphagia is scarce. We aimed to 1) compare radiation dose parameters in HNC survivors with and without dysphagia, 2) investigate factors associated with long-term dysphagia and its possible impact on health-related quality of life (HRQoL), and 3) investigate how our data agree with existing NTCP models.

METHODS

This cross-sectional study conducted in 2018-2020, included HNC survivors treated in 2007-2013. Participants attended a one-day examination in hospital and filled in patient questionnaires. Dysphagia was measured with the EORTC QLQ-H&N35 swallowing scale. Toxicity was scored with CTCAE v.4. We contoured swallowing organs at risk (SWOAR) on RT plans, calculated dose-volume histograms (DVHs), performed logistic regression analyses and tested our data in established NTCP models.

RESULTS

Of the 239 participants, 75 (31%) reported dysphagia. Compared to survivors without dysphagia, this group had reduced HRQoL and the DVHs for infrahyoid SWOAR were significantly shifted to the right. Long-term dysphagia was associated with age (OR 1.07, 95% CI 1.03-1.10), female sex (OR 2.75, 95% CI 1.45-5.21), and mean dose to middle pharyngeal constrictor muscle (MD-MPCM) (OR 1.06, 95% CI 1.03-1.09). NTCP models overall underestimated the risk of long-term dysphagia.

CONCLUSIONS

Long-term dysphagia was associated with higher age, being female, and high MD-MPCM. Doses to distally located SWOAR seemed to be risk factors. Existing NTCP models do not sufficiently predict long-term dysphagia. Further efforts are needed to reduce the prevalence and consequences of this late effect.

Evaluation of decentralised model-based selection of head and neck cancer patients for a proton treatment study. DAHANCA 35

INTRODUCTION

Proton treatment can potentially spare patients with H&N cancer for substantial treatment-related toxicities. The current study investigated the reproducibility of a decentralised model-based selection of patients for a proton treatment study when the selection plans were compared to the clinical treatment plans performed at the proton centre.

METHODS

Sixty-three patients were selected for proton treatment in the six Danish Head and Neck Cancer (DAHANCA) centres. The patients were selected based on normal tissue complication probability (NTCP) estimated from local photon and proton treatment plans, which showed a ΔNTCP greater than 5%-point for either grade 2 + dysphagia or grade 2 + xerostomia at six months. The selection plans were compared to the clinical treatment plans performed at the proton centre.

RESULTS

Of the 63 patients, 49 and 25 were selected based on an estimated benefit in risk of dysphagia and xerostomia, respectively. Eleven patients had a potential gain in both toxicities. The mean ΔNTCP changed from the local selection plan comparison to the clinical comparison from 6.9 to 5.3 %-points (p = 0.01) and 7.3 to 4.9 %-points (p = 0.03) for dysphagia and xerostomia, respectively. Volume differences in both CTV and OAR could add to the loss in ΔNTCP. 61 of the 63 clinical plans had a positive ΔNTCP, and 38 had a ΔNTCP of 5%-points for at least one of the two endpoints.

CONCLUSION

A local treatment plan comparison can be used to select candidates for proton treatment. The local comparative proton plan overestimates the potential benefit of the clinical proton plan. Continuous quality assurance of the delineation procedures and planning is crucial in the subsequent randomised clinical trial setting.

Considerations for study design in the DAHANCA 35 trial of protons versus photons for head and neck cancer

Proton radiotherapy offers a dosimetric advantage compared to photon therapy in sparing normal tissue, but the clinical evidence for toxicity reductions in the treatment of head and neck cancer is limited. The Danish Head and Neck Cancer Group (DAHANCA) has initiated the DAHANCA 35 randomised trial to clarify the value of proton therapy (NCT04607694). The DAHANCA 35 trial is performed in an enriched population of patients selected by an anticipated benefit of proton therapy to reduce the risk of late dysphagia or xerostomia based on normal tissue complication probability (NTCP) modelling. We present our considerations on the trial design and a test of the selection procedure conducted before initiating the randomised study.

Using machine learning algorithm to analyse the hypothyroidism complications caused by radiotherapy in patients with head and neck cancer

Machine learning algorithms were used to analyze the odds and predictors of complications of thyroid damage after radiation therapy in patients with head and neck cancer. This study used decision tree (DT), random forest (RF), and support vector machine (SVM) algorithms to evaluate predictors for the data of 137 head and neck cancer patients. Candidate factors included gender, age, thyroid volume, minimum dose, average dose, maximum dose, number of treatments, and relative volume of the organ receiving X dose (X: 10, 20, 30, 40, 50, 60 Gy). The algorithm was optimized according to these factors and tenfold cross-validation to analyze the state of thyroid damage and select the predictors of thyroid dysfunction. The importance of the predictors identified by the three machine learning algorithms was ranked: the top five predictors were age, thyroid volume, average dose, V50 and V60. Of these, age and volume were negatively correlated with thyroid damage, indicating that the greater the age and thyroid volume, the lower the risk of thyroid damage; the average dose, V50 and V60 were positively correlated with thyroid damage, indicating that the larger the average dose, V50 and V60, the higher the risk of thyroid damage. The RF algorithm was most accurate in predicting the probability of thyroid damage among the three algorithms optimized using the above factors. The Area under the receiver operating characteristic curve (AUC) was 0.827 and the accuracy (ACC) was 0.824. This study found that five predictors (age, thyroid volume, mean dose, V50 and V60) are important factors affecting the chance that patients with head and neck cancer who received radiation therapy will develop hypothyroidism. Using these factors as the prediction basis of the algorithm and using RF to predict the occurrence of hypothyroidism had the highest ACC, which was 82.4%. This algorithm is quite helpful in predicting the probability of radiotherapy complications. It also provides references for assisting medical decision-making in the future.

Prediction of Radiation-Induced Parotid Gland-Related Xerostomia in Patients With Head and Neck Cancer: Regeneration-Weighted Dose

PURPOSE

Despite improvements to treatment, patients with head and neck cancer (HNC) still experience radiation-induced xerostomia due to salivary gland damage. The stem cells of the parotid gland (PG), concentrated in the gland's main ducts (stem cell rich [SCR] region), play a critical role in the PG's response to radiation. Treatment optimization requires a dose metric that properly accounts for the relative contributions of dose to this SCR region and the PG's remainder (non-SCR region) to the risk of xerostomia in normal tissue complication probability (NTCP) models for xerostomia.

MATERIALS AND METHODS

Treatment and toxicity data of 1013 prospectively followed patients with HNC treated with definitive radiation therapy (RT) were used. The regeneration-weighted dose, enabling accounting for the hypothesized different effects of dose to the SCR and non-SCR region on the risk of xerostomia, was defined as Dreg PG = Dmean SCR region + r × Dmean non-SCR region, where Dreg is the regeneration-weighted dose, Dmean is the mean dose, and r is the weighting factor. Considering the different volumes of these regions, r > 3.6 in Dreg PG demonstrates an enhanced effect of the SCR region. The most predictive value of r was estimated in 102 patients of a previously published trial testing stem cell sparing RT. For each endpoint, Dreg PG, dose to other organs, and clinical factors were used to develop NTCP models using multivariable logistic regression analysis in 663 patients. The models were validated in 350 patients.

RESULTS

Dose to the contralateral PG was associated with daytime, eating-related, and physician-rated grade ≥2 xerostomia. Consequently, r was estimated and found to be smaller than 3.6 for most PG function-related endpoints. Therefore, the contribution of Dmean SCR region to the risk of xerostomia was larger than predicted by Dmean PG. Other frequently selected predictors were pretreatment xerostomia and Dmean oral cavity. The validation showed good discrimination and calibration.

CONCLUSIONS

Tools for clinical implementation of stem cell sparing RT were developed: regeneration-weighted dose to the parotid gland that accounted for regional differences in radiosensitivity within the gland and NTCP models that included this new dose metric and other prognostic factors.

Dose planning study of proton versus photon radiotherapy for head and neck squamous cell carcinoma of unknown primary in the primary and recurrent setting

BACKGROUND

Patients with head and neck squamous cell carcinoma of unknown primary (HNCUP) are often treated with extensive radiotherapy (RT). Frequently, the bilateral nodal clinical target volume (nCTV) and the volumes of suspected mucosal primary sites (mCTV) of the pharynx and larynx is irradiated. This treatment is effective but toxic. New data suggest that omission of the contralateral nCTV and mCTV, results in few recurrences. The present study explores photon versus proton therapy, in the primary and recurrent setting.

MATERIAL AND METHODS

An analysis of twelve patients previously treated for HNCUP was performed. A fictitious recurrence was defined in patients treated for unilateral disease. Independently a volumetric arc photon plan and an intensity-modulated proton plan was made for all cases and scenarios.

RESULTS

Compared to the standard bilateral treatment this study shows that limiting the target to unilateral nCTV leads to a significant decrease in dysphagia of 18% and 17% and xerostomia of 4.0% and 5% for photon and protons, respectively. Comparing photon RT directly to proton RT shows a small and often insignificant gain, using protons for both bilateral and unilateral targets. Focusing on re-irradiation, benefits from using protons in both the primary setting and at re-irradiation were limited. However, using protons for re-irradiation only leads to a decrease in the tissue volume receiving a specific dose outside the target overlapping region, e.g., V90Gymean was 31, 25, and 22 cm3 for photons-photons, photons-protons, and protons-protons, respectively. For V100Gy of the ipsilateral carotid artery, no differences were observed.

CONCLUSION

Omitting contralateral nCTV irradiation and mCTV irradiation will significantly reduce toxicity. The accumulated high dose volumes can be minimised using protons for re-irradiation. However, the use of protons for primary treatment provides limited benefit in most patients.

Comparing supervised and semi-supervised machine learning approaches in NTCP modeling to predict complications in head and neck cancer patients

Background and purpose

Head and neck cancer (HNC) patients treated with radiotherapy often suffer from radiation-induced toxicities. Normal Tissue Complication Probability (NTCP) modeling can be used to determine the probability to develop these toxicities based on patient, tumor, treatment and dose characteristics. Since the currently used NTCP models are developed using supervised methods that discard unlabeled patient data, we assessed whether the addition of unlabeled patient data by using semi-supervised modeling would gain predictive performance.

Materials and methods

The semi-supervised method of self-training was compared to supervised regression methods with and without prior multiple imputation by chained equation (MICE). The models were developed for the most common toxicity outcomes in HNC patients, xerostomia (dry mouth) and dysphagia (difficulty swallowing), measured at six months after treatment, in a development cohort of 750 HNC patients. The models were externally validated in a validation cohort of 395 HNC patients. Model performance was assessed by discrimination and calibration.

Results

MICE and self-training did not improve performance in terms of discrimination or calibration at external validation compared to current regression models. In addition, the relative performance of the different models did not change upon a decrease in the amount of (labeled) data available for model development. Models using ridge regression outperformed the logistic models for the dysphagia outcome.

Conclusion

Since there was no apparent gain in the addition of unlabeled patient data by using the semi-supervised method of self-training or MICE, the supervised regression models would still be preferred in current NTCP modeling for HNC patients.

Interventional Radiotherapy (Brachytherapy) for Nasal Vestibule: Novel Strategies to Prevent Side Effects

Interventional radiotherapy (brachytherapy) has become the new therapeutic standard in the management of early stages nasal vestibule tumors; in fact it allows for high local control rates and low toxicity profiles. However, since more and more patients will receive interventional radiotherapy (brachytherapy) as primary treatment, it is desirable to implement novel strategies to reduce the dose to organs at risk with the future aim to result in further lowering long-term side effects.

MATERIALS AND METHODS

We were able to identify two different strategies to reduce dose to the treatment volume, including the implantation technique (the implant can be interstitial, endocavitary or mixed and the catheters may be placed either using the Paris system rules or the anatomical approach) and the dose distribution within the implant (the most commonly used parameter to consider is the dose non-uniformity ratio). We subsequently propose two novel strategies to reduce dose to organs at risk, including the use of metal shields for fixed organs as in the case of the eyes and the use of a mouth swab to push away mobile organs, such in the case of the mandible. We used two different algorithms to verify the values namely the TG-43 and the TG-186.

RESULTS

We provided an accurate literature review regarding strategies to reduce toxicity to the treatment volume, underlining the pros and cons of all implantation techniques and about the use dose non-uniformity ratio. Regarding the innovative strategies to reduce the dose to organs at risk, we investigated the use of eye shielding and the use of swabs to push away the mandible by performing an innovative calculation using two different algorithms in a series of three consecutive patients. Our results show that the dose reduction, both in the case of the mandible and in the case of eye shielding, was statistically significant.

CONCLUSION

Proper knowledge of the best implantation technique and dose non-uniformity ratio as highlighted by existing literature is mandatory in order to reduce toxicity within the treatment volume. With regard to the dose reduction to the organs at risk we have demonstrated that the use of eye shielding and mouth swab could play a pivotal role in clinical practice; in fact, they are effective at lowering the doses to the surrounding organs and do not require any change to the current clinical workflow.

EviGUIDE - a tool for evidence-based decision making in image-guided adaptive brachytherapy for cervical cancer

PURPOSE

To develop a novel decision-support system for radiation oncology that incorporates clinical, treatment and outcome data, as well as outcome models from a large clinical trial on magnetic resonance image-guided adaptive brachytherapy (MR-IGABT) for locally advanced cervical cancer (LACC).

METHODS

A system, called EviGUIDE, was developed that combines dosimetric information from the treatment planning system, patient and treatment characteristics, and established tumor control probability (TCP), and normal tissue complication probability (NTCP) models, to predict clinical outcome of radiotherapy treatment of LACC. Six Cox Proportional Hazards models based on data from 1341 patients of the EMBRACE-I study have been integrated. One TCP model for local tumor control, and five NTCP models for OAR morbidities.

RESULTS

EviGUIDE incorporates TCP-NTCP graphs to help users visualize the clinical impact of different treatment plans and provides feedback on achievable doses based on a large reference population. It enables holistic assessment of the interplay between multiple clinical endpoints and tumour and treatment variables. Retrospective analysis of 45 patients treated with MR-IGABT showed that there exists a sub-cohort of patients (20%) with increased risk factors, that could greatly benefit from the quantitative and visual feedback.

CONCLUSION

A novel digital concept was developed that can enhance clinical decision- making and facilitate personalized treatment. It serves as a proof of concept for a new generation of decision support systems in radiation oncology, which incorporate outcome models and high-quality reference data, and aids the dissemination of evidence-based knowledge about optimal treatment and serve as a blueprint for other sites in radiation oncology.

An efficient strategy to select head and neck cancer patients for adaptive radiotherapy

BACKGROUND AND PURPOSE

Adaptive radiotherapy (ART) is workload intensive but only benefits a subgroup of patients. We aimed to develop an efficient strategy to select candidates for ART in the first two weeks of head and neck cancer (HNC) radiotherapy.

MATERIALS AND METHODS

This study retrospectively enrolled 110 HNC patients who underwent modern photon radiotherapy with at least 5 weekly in-treatment re-scan CTs. A semi auto-segmentation method was applied to obtain the weekly mean dose (Dmean) to OARs. A comprehensive NTCP-profile was applied to obtain NTCP's. The difference between planning and actual values of Dmean (ΔDmean) and dichotomized difference of clinical relevance (BIOΔNTCP) were used for modelling to determine the cut-off maximum ΔDmean of OARs in week 1 and 2 (maxΔDmean_1 and maxΔDmean_2). Four strategies to select candidates for ART, using cut-off maxΔDmean were compared.

RESULTS

The Spearman's rank correlation test showed significant positive correlation between maxΔDmean and BIOΔNTCP (p-value <0.001). For major BIOΔNTCP (>5%) of acute and late toxicity, 10.9% and 4.5% of the patients were true candidates for ART. Strategy C using both cut-off maxΔDmean_1 (3.01 and 5.14 Gy) and cut-off maxΔDmean_2 (3.41 and 5.30 Gy) showed the best sensitivity, specificity, positive and negative predictive values (0.92, 0.82, 0.38, 0.99 for acute toxicity and 1.00, 0.92, 0.38, 1.00 for late toxicity, respectively).

CONCLUSIONS

We propose an efficient selection strategy for ART that is able to classify the subgroup of patients with >5% BIOΔNTCP for late toxicity using imaging in the first two treatment weeks.

Development and Validation of an Automated Image-Based Deep Learning Platform for Sarcopenia Assessment in Head and Neck Cancer

Importance

Sarcopenia is an established prognostic factor in patients with head and neck squamous cell carcinoma (HNSCC); the quantification of sarcopenia assessed by imaging is typically achieved through the skeletal muscle index (SMI), which can be derived from cervical skeletal muscle segmentation and cross-sectional area. However, manual muscle segmentation is labor intensive, prone to interobserver variability, and impractical for large-scale clinical use.

Objective

To develop and externally validate a fully automated image-based deep learning platform for cervical vertebral muscle segmentation and SMI calculation and evaluate associations with survival and treatment toxicity outcomes.

Design, Setting, and Participants

For this prognostic study, a model development data set was curated from publicly available and deidentified data from patients with HNSCC treated at MD Anderson Cancer Center between January 1, 2003, and December 31, 2013. A total of 899 patients undergoing primary radiation for HNSCC with abdominal computed tomography scans and complete clinical information were selected. An external validation data set was retrospectively collected from patients undergoing primary radiation therapy between January 1, 1996, and December 31, 2013, at Brigham and Women's Hospital. The data analysis was performed between May 1, 2022, and March 31, 2023.

Exposure

C3 vertebral skeletal muscle segmentation during radiation therapy for HNSCC.

Main Outcomes and Measures

Overall survival and treatment toxicity outcomes of HNSCC.

Results

The total patient cohort comprised 899 patients with HNSCC (median [range] age, 58 [24-90] years; 140 female [15.6%] and 755 male [84.0%]). Dice similarity coefficients for the validation set (n = 96) and internal test set (n = 48) were 0.90 (95% CI, 0.90-0.91) and 0.90 (95% CI, 0.89-0.91), respectively, with a mean 96.2% acceptable rate between 2 reviewers on external clinical testing (n = 377). Estimated cross-sectional area and SMI values were associated with manually annotated values (Pearson r = 0.99; P < .001) across data sets. On multivariable Cox proportional hazards regression, SMI-derived sarcopenia was associated with worse overall survival (hazard ratio, 2.05; 95% CI, 1.04-4.04; P = .04) and longer feeding tube duration (median [range], 162 [6-1477] vs 134 [15-1255] days; hazard ratio, 0.66; 95% CI, 0.48-0.89; P = .006) than no sarcopenia.

Conclusions and Relevance

This prognostic study's findings show external validation of a fully automated deep learning pipeline to accurately measure sarcopenia in HNSCC and an association with important disease outcomes. The pipeline could enable the integration of sarcopenia assessment into clinical decision making for individuals with HNSCC.