Dataset Information

Comparison of Machine-Learning and Deep-Learning Methods for the Prediction of Osteoradionecrosis Resulting From Head and Neck Cancer Radiation Therapy.

ABSTRACT:

Purpose

Deep-learning (DL) techniques have been successful in disease-prediction tasks and could improve the prediction of mandible osteoradionecrosis (ORN) resulting from head and neck cancer (HNC) radiation therapy. In this study, we retrospectively compared the performance of DL algorithms and traditional machine-learning (ML) techniques to predict mandible ORN binary outcome in an extensive cohort of patients with HNC.

Methods and materials

Patients who received HNC radiation therapy at the University of Texas MD Anderson Cancer Center from 2005 to 2015 were identified for the ML (n = 1259) and DL (n = 1236) studies. The subjects were followed for ORN development for at least 12 months, with 173 developing ORN and 1086 having no evidence of ORN. The ML models used dose-volume histogram parameters to predict ORN development. These models included logistic regression, random forest, support vector machine, and a random classifier reference. The DL models were based on ResNet, DenseNet, and autoencoder-based architectures. The DL models used each participant's dose cropped to the mandible. The effect of increasing the amount of available training data on the DL models' prediction performance was evaluated by training the DL models using increasing ratios of the original training data.

Results

The F1 score for the logistic regression model, the best-performing ML model, was 0.3. The best-performing ResNet, DenseNet, and autoencoder-based models had F1 scores of 0.07, 0.14, and 0.23, respectively, whereas the random classifier's F1 score was 0.17. No performance increase was apparent when we increased the amount of training data available for DL model training.

Conclusions

The ML models had superior performance to their DL counterparts. The lack of improvement in DL performance with increased training data suggests that either more data are needed for appropriate DL model construction or that the image features used in DL models are not suitable for this task.

SUBMITTER: Reber B

PROVIDER: S-EPMC9926206 | biostudies-literature | 2023 Jul-Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Comparison of Machine-Learning and Deep-Learning Methods for the Prediction of Osteoradionecrosis Resulting From Head and Neck Cancer Radiation Therapy.

Reber Brandon B Van Dijk Lisanne L Anderson Brian B Mohamed Abdallah Sherif Radwan ASR Fuller Clifton C Lai Stephen S Brock Kristy K

Advances in radiation oncology 20221227 4

<h4>Purpose</h4>Deep-learning (DL) techniques have been successful in disease-prediction tasks and could improve the prediction of mandible osteoradionecrosis (ORN) resulting from head and neck cancer (HNC) radiation therapy. In this study, we retrospectively compared the performance of DL algorithms and traditional machine-learning (ML) techniques to predict mandible ORN binary outcome in an extensive cohort of patients with HNC.<h4>Methods and materials</h4>Patients who received HNC radiation ...[more]

PMID: 36798732

Similar Datasets

Project description:BackgroundRadiomics can provide in-depth characterization of cancers for treatment outcome prediction. Conventional radiomics rely on extraction of image features within a pre-defined image region of interest (ROI) which are typically fed to a classification algorithm for prediction of a clinical endpoint. Deep learning radiomics allows for a simpler workflow where images can be used directly as input to a convolutional neural network (CNN) with or without a pre-defined ROI.PurposeThe purpose of this study was to evaluate (i) conventional radiomics and (ii) deep learning radiomics for predicting overall survival (OS) and disease-free survival (DFS) for patients with head and neck squamous cell carcinoma (HNSCC) using pre-treatment 18F-fluorodeoxuglucose positron emission tomography (FDG PET) and computed tomography (CT) images.Materials and methodsFDG PET/CT images and clinical data of patients with HNSCC treated with radio(chemo)therapy at Oslo University Hospital (OUS; n = 139) and Maastricht University Medical Center (MAASTRO; n = 99) were collected retrospectively. OUS data was used for model training and initial evaluation. MAASTRO data was used for external testing to assess cross-institutional generalizability. Models trained on clinical and/or conventional radiomics features, with or without feature selection, were compared to CNNs trained on PET/CT images without or with the gross tumor volume (GTV) included. Model performance was measured using accuracy, area under the receiver operating characteristic curve (AUC), Matthew's correlation coefficient (MCC), and the F1 score calculated for both classes separately.ResultsCNNs trained directly on images achieved the highest performance on external data for both endpoints. Adding both clinical and radiomics features to these image-based models increased performance further. Conventional radiomics including clinical data could achieve competitive performance. However, feature selection on clinical and radiomics data lead to overfitting and poor cross-institutional generalizability. CNNs without tumor and node contours achieved close to on-par performance with CNNs including contours.ConclusionHigh performance and cross-institutional generalizability can be achieved by combining clinical data, radiomics features and medical images together with deep learning models. However, deep learning models trained on images without contours can achieve competitive performance and could see potential use as an initial screening tool for high-risk patients.

Project description:ImportanceProton radiation therapy (PRT) has reduced radiation-induced toxic effects, such as mucositis and xerostomia, over conventional photon radiation therapy, leading to significantly improved quality of life in patients with head and neck cancers. However, the prevalence of osteoradionecrosis (ORN) of the jaw following PRT in these patients is less clear.ObjectiveTo report the prevalence and clinical characteristics of ORN in patients with oral and oropharyngeal cancer (OOPC) treated with PRT.Design, setting, and participantsThis case series reports a single-institution experience (Memorial Sloan Kettering Cancer Center, New York, New York) between November 2013 and September 2019 and included 122 radiation therapy-naive patients with OOPC treated with PRT. Data were analyzed from 2013 to 2019.Main outcomes and measuresClinical parameters, including sex, age, comorbidities, tumor histology, concurrent chemotherapy, smoking, comorbidities, and preradiation dental evaluation, were obtained from the medical record. Patients with clinical or radiographic signs of ORN were identified and graded using the adopted modified Glanzmann and Grätz grading system. Characteristics of ORN, such as location, clinical presentation, initial stage at diagnosis, etiology, time to diagnosis, management, and clinical outcome at the last follow-up, were also collected.ResultsOf the 122 patients (mean [SD] age, 63 [13] years; 45 [36.9%] women and 77 [63.1%] men) included in this study, 13 (10.6%) developed ORN following PRT during a median (range) follow-up time of 40.6 (<1-101) months. All patients had spontaneous development of ORN. At the time of initial diagnosis, grade 0, grade 1, grade 2, and grade 3 ORN were seen in 2, 1, 9, and 1 patient, respectively. The posterior ipsilateral mandible within the radiation field that received the full planned PRT dose was the most involved ORN site. At a median (range) follow-up of 13.5 (0.2-58.0) months from the time of ORN diagnosis, complete resolution, stable condition, and progression of ORN were seen in 3, 6, and 4 patients, respectively. The 3-year rates of ORN and death in the total cohort were 5.2% and 21.5%, while the 5-year rates of ORN and death were 11.5% and 34.4%, respectively.Conclusions and relevanceIn this case series, the prevalence of ORN following PRT was found to be 10.6%, indicating that ORN remains a clinical challenge even in the era of highly conformal PRT. Clinicians treating patients with OOPC with PRT should be mindful of this complication.

Project description: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.

Project description:PurposeOsteoradionecrosis (ORN) of the mandible represents a severe, debilitating complication of radiation therapy (RT) for head and neck cancer (HNC). At present, no normal tissue complication probability (NTCP) models for risk of ORN exist. The aim of this study was to develop a multivariable clinical/dose-based NTCP model for the prediction of ORN any grade (ORNI-IV) and grade IV (ORNIV) after RT (±chemotherapy) in patients with HNC.Methods and materialsIncluded patients with HNC were treated with (chemo-)RT between 2005 and 2015. Mandible bone radiation dose-volume parameters and clinical variables (ie, age, sex, tumor site, pre-RT dental extractions, chemotherapy history, postoperative RT, and smoking status) were considered as potential predictors. The patient cohort was randomly divided into a training (70%) and independent test (30%) cohort. Bootstrapped forward variable selection was performed in the training cohort to select the predictors for the NTCP models. Final NTCP model(s) were validated on the holdback test subset.ResultsOf 1259 included patients with HNC, 13.7% (n = 173 patients) developed any grade ORN (ORNI-IV primary endpoint) and 5% (n = 65) ORNIV (secondary endpoint). All dose and volume parameters of the mandible bone were significantly associated with the development of ORN in univariable models. Multivariable analyses identified D30% and pre-RT dental extraction as independent predictors for both ORNI-IV and ORNIV best-performing NTCP models with an area under the curve (AUC) of 0.78 (AUCvalidation = 0.75 [0.69-0.82]) and 0.81 (AUCvalidation = 0.82 [0.74-0.89]), respectively.ConclusionsThis study presented NTCP models based on mandible bone D30% and pre-RT dental extraction that predict ORNI-IV and ORNIV (ie, needing invasive surgical intervention) after HNC RT. Our results suggest that less than 30% of the mandible should receive a dose of 35 Gy or more for an ORNI-IV risk lower than 5%. These NTCP models can improve ORN prevention and management by identifying patients at risk of ORN.

Dataset Information

Comparison of Machine-Learning and Deep-Learning Methods for the Prediction of Osteoradionecrosis Resulting From Head and Neck Cancer Radiation Therapy.

Purpose

Methods and materials

Results

Conclusions

Publications

Comparison of Machine-Learning and Deep-Learning Methods for the Prediction of Osteoradionecrosis Resulting From Head and Neck Cancer Radiation Therapy.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets