Author Correction: Imaging and clinical data archive for head and neck squamous cell carcinoma patients treated with radiotherapy.

In the original version of the Data Descriptor the surname of author Hesham Elhalawani was misspelled. This has now been corrected in the HTML and PDF versions.

Machine Learning Applications in Head and Neck Radiation Oncology: Lessons From Open-Source Radiomics Challenges.

Radiomics leverages existing image datasets to provide non-visible data extraction via image post-processing, with the aim of identifying prognostic, and predictive imaging features at a sub-region of interest level. However, the application of radiomics is hampered by several challenges such as lack of image acquisition/analysis method standardization, impeding generalizability. As of yet, radiomics remains intriguing, but not clinically validated. We aimed to test the feasibility of a non-custom-constructed platform for disseminating existing large, standardized databases across institutions for promoting radiomics studies. Hence, University of Texas MD Anderson Cancer Center organized two public radiomics challenges in head and neck radiation oncology domain. This was done in conjunction with MICCAI 2016 satellite symposium using Kaggle-in-Class, a machine-learning and predictive analytics platform. We drew on clinical data matched to radiomics data derived from diagnostic contrast-enhanced computed tomography (CECT) images in a dataset of 315 patients with oropharyngeal cancer. Contestants were tasked to develop models for (i) classifying patients according to their human papillomavirus status, or (ii) predicting local tumor recurrence, following radiotherapy. Data were split into training, and test sets. Seventeen teams from various professional domains participated in one or both of the challenges. This review paper was based on the contestants' feedback; provided by 8 contestants only (47%). Six contestants (75%) incorporated extracted radiomics features into their predictive model building, either alone (n = 5; 62.5%), as was the case with the winner of the "HPV" challenge, or in conjunction with matched clinical attributes (n = 2; 25%). Only 23% of contestants, notably, including the winner of the "local recurrence" challenge, built their model relying solely on clinical data. In addition to the value of the integration of machine learning into clinical decision-making, our experience sheds light on challenges in sharing and directing existing datasets toward clinical applications of radiomics, including hyper-dimensionality of the clinical/imaging data attributes. Our experience may help guide researchers to create a framework for sharing and reuse of already published data that we believe will ultimately accelerate the pace of clinical applications of radiomics; both in challenge or clinical settings.

Imaging and clinical data archive for head and neck squamous cell carcinoma patients treated with radiotherapy.

Cross sectional imaging is essential for the patient-specific planning and delivery of radiotherapy, a primary determinant of head and neck cancer outcomes. Due to challenges ensuring data quality and patient de-identification, publicly available datasets including diagnostic and radiation treatment planning imaging are scarce. In this data descriptor, we detail the collection and processing of computed tomography based imaging in 215 patients with head and neck squamous cell carcinoma that were treated with radiotherapy. Using cross sectional imaging, we calculated total body skeletal muscle and adipose content before and after treatment. We detail techniques for validating the high quality of these data and describe the processes of data de-identification and transfer. All imaging data are subject- and date-matched to clinical data from each patient, including demographics, risk factors, grade, stage, recurrence, and survival. These data are a valuable resource for studying the association between patient-specific anatomic and metabolic features, treatment planning, and oncologic outcomes, and the first that allows for the integration of body composition as a risk factor or study outcome.

Deep Learning Algorithm for Auto-Delineation of High-Risk Oropharyngeal Clinical Target Volumes With Built-In Dice Similarity Coefficient Parameter Optimization Function.

PURPOSE: Automating and standardizing the contouring of clinical target volumes (CTVs) can reduce interphysician variability, which is one of the largest sources of uncertainty in head and neck radiation therapy. In addition to using uniform margin expansions to auto-delineate high-risk CTVs, very little work has been performed to provide patient- and disease-specific high-risk CTVs. The aim of the present study was to develop a deep neural network for the auto-delineation of high-risk CTVs. METHODS AND MATERIALS: Fifty-two oropharyngeal cancer patients were selected for the present study. All patients were treated at The University of Texas MD Anderson Cancer Center from January 2006 to August 2010 and had previously contoured gross tumor volumes and CTVs. We developed a deep learning algorithm using deep auto-encoders to identify physician contouring patterns at our institution. These models use distance map information from surrounding anatomic structures and the gross tumor volume as input parameters and conduct voxel-based classification to identify voxels that are part of the high-risk CTV. In addition, we developed a novel probability threshold selection function, based on the Dice similarity coefficient (DSC), to improve the generalization of the predicted volumes. The DSC-based function is implemented during an inner cross-validation loop, and probability thresholds are selected a priori during model parameter optimization. We performed a volumetric comparison between the predicted and manually contoured volumes to assess our model. RESULTS: The predicted volumes had a median DSC value of 0.81 (range 0.62-0.90), median mean surface distance of 2.8 mm (range 1.6-5.5), and median 95th Hausdorff distance of 7.5 mm (range 4.7-17.9) when comparing our predicted high-risk CTVs with the physician manual contours. CONCLUSIONS: These predicted high-risk CTVs provided close agreement to the ground-truth compared with current interobserver variability. The predicted contours could be implemented clinically, with only minor or no changes.

Patient reported dry mouth: Instrument comparison and model performance for correlation with quality of life in head and neck cancer survivors.

PURPOSE: To identify a clinically meaningful cut-point for the single item dry mouth question of the MD Anderson Symptom Inventory-Head and Neck module (MDASI-HN). METHODS: Head and neck cancer survivors who had received radiation therapy (RT) completed the MDASI-HN, the University of Michigan Hospital Xerostomia Questionnaire (XQ), and the health visual analog scale (VAS) of the EuroQol Five Dimension Questionnaire (EQ-5D). The Bayesian information criteria (BIC) were used to test the prediction power of each tool for EQ-5D VAS. The modified Breiman recursive partitioning analysis (RPA) was used to identify a cut point of the MDASI-HN dry mouth score (MDASI-HN-DM) with EQ-5D VAS, using a ROC-based approach; regression analysis was used to confirm the threshold effect size. RESULTS: Two-hundred seven respondents formed the cohort. Median follow-up from the end of RT to questionnaire completion was 88 months. The single item MDASI-HN-DM score showed a linear relationship with the XQ composite score (ρ = 0.80, p < 0.001). The MDASI-HN-DM displayed improved model performance for association with EQ-5D VAS as compared to XQ (BIC of 1803.7 vs. 2016.9, respectively). RPA showed that an MDASI-HN-DM score of ≥6 correlated with EQ-5D VAS decline (LogWorth 5.5). CONCLUSION: The single item MDASI-HN-DM correlated with the multi-item XQ and performed favorably in the prediction of QOL. A MDASI-HN-DM cut point of ≥6 correlated with decline in QOL.

Correction to: Long-term patient reported outcomes following radiation therapy for oropharyngeal cancer: cross-sectional assessment of a prospective symptom survey in patients ≥65 years old.

In the original publication [1] the name of author Jeremy M. Aymard was spelled wrong. The original article was updated to rectify this error.

Long-term patient reported outcomes following radiation therapy for oropharyngeal cancer: cross-sectional assessment of a prospective symptom survey in patients ≥65 years old.

BACKGROUND: Given the potential for older patients to experience exaggerated toxicity and symptoms, this study was performed to characterize patient reported outcomes in older patients following definitive radiation therapy (RT) for oropharyngeal cancer (OPC). METHODS: Cancer-free head and neck cancer survivors (>6 months since treatment completion) were eligible for participation in a questionnaire-based study. Participants completed the MD Anderson Symptom Inventory-Head and Neck module (MDASI-HN). Those patients ≥65 years old at treatment for OPC with definitive RT were included. Individual and overall symptom severity and clinical variables were analyzed. RESULTS: Of the 79 participants analyzed, 82% were male, 95% white, 41% T3/4 disease, 39% RT alone, 27% induction chemotherapy, 52% concurrent, and 18% both, and 96% IMRT. Median age at RT was 71 yrs. (range: 65-85); median time from RT to MDASI-HN was 46 mos. (2/3 > 24 mos.). The top 5 MDASI-HN items rated most severe in terms of mean (±SD) ratings (0-10 scale) were dry mouth (3.48 ± 2.95), taste (2.81 ± 3.29), swallowing (2.59 ± 2.96), mucus in mouth/throat (2.04 ± 2.68), and choking (1.30 ± 2.38) reported at moderate-severe levels (≥5) by 35, 29, 29, 18, and 13%, respectively. Thirty-nine % reported none (0) or no more than mild (1-4) symptoms across all 22 MDASI-HN symptoms items, and 38% had at least one item rated as severe (≥7). Hierarchical cluster analysis resulted in 3 patient groups: 1) ~65% with ranging from none to moderate symptom burden, 2) ~35% with moderate-severe ratings for a subset of classically RT-related symptoms (e.g. dry mouth, mucus, swallowing) and 3) 2 pts. with severe ratings of most items. CONCLUSIONS: The overall long-term symptom burden seen in this older OPC cohort treated with modern standard therapy was largely favorable, yet a higher symptom group (~35%) with a distinct pattern of mostly local and classically RT-related symptoms was identified.