Explor MedEMExploration of Medicine2692-3106Open Exploration Publishing10.37349/emed.2026.10013941001394Original ArticlePrognostic prediction of head and neck cancer through radiomics: a stacking ensemble approach with machine learning and deep learning modelshttps://orcid.org/0009-0001-1520-2540WongHeylie YTConceptualizationSupervisionProject administration12†https://orcid.org/0000-0001-8848-596XXueChengMethodologyValidationFormal analysis3†https://orcid.org/0000-0001-6530-7730TANGFuk-hay1*ChanCynthia CYInvestigationWriting—original draft1LiVictoria TYInvestigationResourcesWriting—review & editing1LeeSarah WYInvestigationResources1LiNingAcademic EditorChinese Academy of Medical Sciences and Peking Union Medical College, China1School of Medical and Health Sciences, Tung Wah College, Hong Kong, China2Department of Pathology, United Christian Hospital, Hong Kong, China3School of Computer Science and Engineering, Southeast University, Nanjing 211189, Jiangsu, China
This study aimed to develop and evaluate a stacking ensemble machine learning (SEML) model that integrates deep learning (DL) algorithms to improve the accuracy of prognostic predictions for patients with head and neck squamous cell carcinoma (HNSCC).
Methods:
A cohort of 215 HNSCC patients’ CT images, featuring gross tumor volume (GTV) and planning target volume (PTV) contours, was analyzed. Radiomics features were extracted and converted into quantitative data. These features were then used to train and compare a novel SEML model against standard DL algorithms to predict patient prognosis.
Results:
The proposed SEML model demonstrated superior predictive performance compared to the DL model, achieving 93% accuracy, 100% sensitivity, and 83% specificity. Statistical analysis using the chi-square test indicated no substantial difference in prediction performance between features derived from GTV and PTV contours (p > 0.05).
Conclusions:
The SEML model effectively enhances the prognostic prediction accuracy for HNSCC based on radiomic features. This approach shows significant potential to inform clinical decision-making and support the development of customized treatment strategies for improved patient care.
head and neck cancerartificial intelligenceradiomicsstacking ensemble learningIntroductionImpact of HNSCC in healthcare
Head and neck squamous cell carcinoma (HNSCC) rank as the 6th most common cancer worldwide, with approximately 890,000 new cases and 450,000 deaths each year recently [1]. In Hong Kong, HNSCC incidence has risen up to 16,000 new cases reported each year, mostly due to tobacco use, alcohol consumption, betel quid chewing, and HPV infection [2]. Furthermore, 5-year survival rate for late-stage HNSCC patients is less than 50%, primarily due to tumor recurrence, metastasis, and therapy resistance [3]. Radiotherapy combined with surgery, chemotherapy, or immunotherapy remain as main treatment methods, yet heterogeneous treatment responses demand more comprehensive personalized approaches [4]. In our study, we attempted to refocused the study on oropharyngeal squamous cell carcinoma (OPSCC) as a particular example of HNSCC, acknowledging this specificity as a strength for a homogeneous analysis.
Consideration of current prognostic methods
The Tumor, Node, Metastasis (TNM) staging system (AJCC 8th Edition) is the gold standard for HNSCC prognosis, it relates to tumor size (T), lymph node involvement (N), and distant metastasis (M) [5]. However, this system suffers from the following weakness:
Biopsies may promote patient discomfort and risk [6], while it has irreplaceable role in the confirmation of diagnosis.
Time delays in histopathological analysis deter immediate treatment decisions [7].
Difficult to detect tumor heterogeneity, molecular subtypes, or dynamic treatment responses [8].
It appears that, other than biomarkers (e.g., PD-L1, HPV status), non-invasive, AI-enhanced radiomics may offer a promising alternative, yet predictive models, like other AI models, can only augment rather than replace AJCC staging.
Radiomics and AI in HNSCC prognostic prediction
Radiomics extracts high-dimensional quantitative features (e.g., texture, shape, wavelet transforms1) from computed tomography (CT), MRI, and PET scans to identify tumour characteristics [9, 10].
The main advantages are:
Prognostic prediction modeling: Correlates imaging features with treatment response, survival, and recurrence [11, 12].
Recent studies have noted that machine learning (ML) algorithms have good performance (Table 1). However, single-model approaches suffer from overfitting and may lack generalizability [16].
Performance of machine learning algorithms.
Algorithm
Application
Accuracy
Reference
SVM (support vector machine): A supervised learning algorithm that finds the optimal hyperplane to separate data into classes with maximum margin.
Oral SCC classification
100%
(Kumar et al., 2021) [13]
Random forest model with six decision trees and seven splits.
Rectal cancer prognosis
95.3%
(Shen et al., 2020) [14]
Multiregional spatial interaction (MSI) matrix with 22 image features. A network strategy was used to integrate all image features and classify patients into different risk groups.
Breast cancer prediction
97.8%, to 98.6%
(Wu et al., 2018) [15]
Ensemble learning: an improvement
Ensemble methods combine multiple models to enhance prediction performance:
Bagging [e.g., Random forest (RF)]: Reduces variance via bootstrap aggregation [17].
Voting ensembles: Achieves better accuracy in HNSCC prognosis [19, 20].
Research gaps and objectives
Despite progress, critical gaps persist, such as:
No stacking-based radiomics models for HNSCC prognosis exist.
Most studies use single-institution data, limiting generalizability [21].
Multi-modal data integration, that is, imaging with genomics, has more room for investigation [22].
Materials and methodsResearch workflow
The study followed the following steps (Figure 1):
Research workflow.
Performance was evaluated via ROC-AUC, accuracy, sensitivity, and specificity.
Data acquisition
Publicly available planning CT images and clinical data from The Cancer Imaging Archive (TCIA) were retrieved, mainly for patients with HNSCC (2003–2013).
Data preprocessing
Quality control: 164 eligible cases were selected from 215 initial entries after excluding incomplete/inconsistent records.
Image processing: gross tumor volume (GTV) and planning target volume (PTV) structures were segmented using 3D Slicer (version 4.10.2). The images were first processed with filtering (e.g., Laplacian of Gaussian for edge enhancement) or wavelet transforms to highlight texture features.
For GTV, it is the demonstrable extent and location of the malignant tumor, including any macroscopically visible or palpable tumors, masses, or nodal involvement.
PTV is a geometric expansion of the clinical target volume (CTV) that accounts for internal motion (e.g., breathing, organ movement) and setup variability (e.g., patient positioning errors during treatment).
Feature extraction
Radiomics features were extracted from the GTV and PTV structures using 3D Slicer (version 4.10.2) with the PyRadiomics extension. The features were categorized into various groups, including shape, first-order statistics, gray-level zone matrix (GLSZM), gray-level dependence matrix (GLDM), gray-level run-length matrix (GLRLM), gray-level co-occurrence matrix (GLCM), and neighborhood gray-tone difference matrix (GLTDM). A total of 107 radiomics features were extracted for analysis.
Model development
Stacking ensemble model (SEML) and deep learning (DL) approaches were implemented.
Patient data
TCIA serves as the data source for the current study. This publicly available repository is operated by the National Cancer Institute (NCI). Planning CT images of patients who received radical radiotherapy for HNSCC during the period from 2003 to 2013, with “HNSCC” labeled dataset was used. The research included GTV and PTV radiotherapy structures and clinical information about patient age, sex, and diagnosis, smoking habits, staging progression, and three- and five-year survival statistics. After quality assessment, finally, 164 valid cases out of the 215 confirmed collections were obtained.
ML models
Cancer prognosis is typically assessed based on specific time points, such as the five-year survival rate [23]. This allows for objective comparisons across different cancer studies, as patients who survive five years post-treatment are generally classified as "cancer survivors."
In this study, the five-year survival rate was selected as the treatment outcome to ensure comparability with other studies. Of the 164 cases, 118 patients survived, while 46 did not. An overfitting test was performed by randomly selecting 46 cases from the 117 patients who survived to balance the sample outcomes.
ML process
Two ML models were employed:
Two-layer stacking ensemble machine learning (SEML) model
In this model, the radiomics data were divided into two sets for data splitting: training (70%) and testing (30%). Four classifiers were utilized: decision trees (DT), RF, support vector machine (SVM), and generalized linear model (GLM). The training set was used to train the models, which were then validated with the validation set to generate predictions. Predicted outcomes were quantified numerically, with values of 0 representing survival beyond five years or death from other causes, and 1 indicating death within five years of diagnosis.
The SEML consists of heterogeneous classifiers organized in a two-layer structure. The base classifiers were initially trained with the radiomics data. The meta-learner was trained using the prediction results from the base classifiers, with XGBoost selected as the meta-classifier. The selection of an appropriate meta-classifier is crucial for model performance. Previous studies have indicated that XGBoost is optimal for recurrent HNSCC prognosis [24–26]. The final prediction outcome was the training result of the XGboost. Details of protocol were illustrated in Figure 2.
Workflow for the adoption of a stacking ensemble machine learning model. DT: deep learning; RF: random forest; SVM: support vector machine; GLM: generalized linear model.
Overfitting test
As the dataset exhibited an unbalanced distribution of outcomes, an overfitting test was conducted using a balanced sample with an equal number of cases for each treatment outcome. The sample with the same number and unequal were conducted to address class imbalance (118 vs. 46), and a balanced subset was created via random undersampling of the majority class. Overfitting was evaluated by comparing performance metrics (AUC, accuracy) and observing the convergence of learning curves (training vs. validation loss) across epochs/folds.
DL methods
To evaluate the performance of the SEML model against DL-based approaches, a deep neural network (DNN) was implemented and trained using radiomics features. The DL model was structured as a fully connected feed-forward neural network, consisting of three hidden layers designed to capture hierarchical patterns in the data.
Architecture details
The architecture of the DL model was carefully designed to ensure effective learning while avoiding overfitting:
Input layer: The model used 107 radiomics features extracted from medical imaging data, serving as the initial input representation.
Hidden layers: Three fully connected (dense) layers were employed with progressively decreasing units to facilitate feature abstraction and dimensionality reduction:
First hidden layer: 256 neurons
Second hidden layer: 128 neurons
Third hidden layer: 64 neurons
Each of these layers utilized the Rectified Linear Unit (ReLU) activation function, which introduces non-linearity while mitigating the vanishing gradient problem.
Output layer: A single neuron with a sigmoid activation function was used to produce a probabilistic output between “0” and “1”, enabling binary classification (that is, survive and die).
Training protocol
The model was trained using the following optimization and validation strategies:
Optimizer: The “Adam optimizer” was employed with a learning rate of 0.001, chosen for its adaptive momentum properties, which help in efficient convergence.
Batch size: A batch size of 16 was selected to balance computational efficiency and gradient stability.
Epochs: Training was conducted for a maximum of 100 epochs, with an early stopping mechanism (patience = 10) monitoring the validation loss to prevent overfitting. If no improvement was observed for 10 consecutive epochs, training was halted.
Data splitting: The dataset was partitioned into a “70/30 ratio” for training and testing, respectively, ensuring sufficient data for model generalization while retaining an independent test set for unbiased evaluation.
This comparative framework allowed for a systematic assessment of the SEML model's performance relative to traditional DL methods, highlighting its potential advantages in interpretability, computational efficiency, and robustness in medical diagnostics.
Data analysis
The predicted outcomes from the two-layer stacking model were compared with those from the base classifiers. Model performance was evaluated using the receiver operating characteristic (ROC) curve, with metrics including the area under the ROC curve (AUC), accuracy, specificity, sensitivity, and chi-square test, which were calculated using ROCkit from the University of Chicago (1995). SEML and DL models were tested on the same holdout dataset (30%) using Delong test by Matlab calculation.
ResultsDemographic cohort
The dataset comprised 215 patients, of which 51 were excluded due to missing data, resulting in 164 cases for analysis. Both PTV and GTV CT datasets were collected. Demographic details are summarized in Table 2.
Patient demographics, staging, and clinical data.
Patient and tumour characteristics(All n = 164)
Data
Age range (years) (mean ± SD)
24–9160 ± 13
Female (mean ± SD)Male (mean ± SD)
55 ± 262 ± 13
Sex
Female
25 (15%)
Male
139 (85%)
Staging
Stage I
3
Stage II
3
Stage III
23
Stage IV
135
Diagnosis
Ca Base of Tongue
60
Ca Tonsil
58
Ca others
46
Smoking status
Smoker
54
Non-smoker
110
Comparison of performance between individual model and stacked model
SEML models using SEML as classifier, consistently outperformed any of the individual ML models in ROC analysis.
Among 4 of the individual ML models using PTV features, SVM attained the best performance in prognostic prediction for HNC. Compared with the SVM model, the SEML model further enhanced the performance in both accuracy from 73.3% to 93.3% (AUC = 0.723 vs. AUC = 0.982) and sensitivity from 66.7% to 100%. The details refer to Table 3.
Performance of each machine learning model and SEML.
Performance
DL
RF
SVM
GLM
SEML
AUC
GTV
0.472
0.672
0.652
0.643
0.820
PTV
0.550
0.716
0.723
0.444
0.982
Accuracy
GTV
0.467
0.733
0.467
0.667
0.733
PTV
0.400
0.600
0.733
0.400
0.933
Sensitivity
GTV
0.500
0.875
0.625
0.500
0.625
PTV
0.444
0.556
0.667
0.333
1.000
Specificity
GTV
0.429
0.571
0.286
0.857
0.857
PTV
0.333
0.667
0.833
0.500
0.833
DL: deep learning; RF: random forest; SVM: support vector machine; GLM: Generalized Linear Model; SEML: stacking ensemble machine learning.
Prediction performance of SEML and DL models
The SEML model demonstrated exceptional performance, with AUC ranging from 0.82 to 0.982 across all target volumes. For PTV radiomic features, the AUC reached 0.982. The model exhibited sensitivity of 100%, specificity of 83%, and accuracy of 93%. In contrast, the AUC for GTV features was slightly lower at 0.82, with sensitivity, specificity, and accuracy at 62.5%, 85.7%, and 73.3%, respectively. In comparison, the DL model showed an AUC ranging from 0.605 to 0.774, with accuracy between 0.655 and 0.724, indicating superior performance of the SEML model (Table 4).
A summary of prognosis prediction performance with both PTV and GTV for SEML and DL.
Despite the PTV radiomics features yielding better predictions, ROC analysis indicated no significant difference between PTV and GTV features (chi-square test, p > 0.05) (Figure 3).
ROC curve of 5-year survival using GTV and PTV radiomics features in the SEML model. GTV: gross tumor volume; PTV: planning target volume; SEML: stacking ensemble machine learning.
Comparison of SEML and DL models
The SEML model consistently outperformed the DL model in predicting 5-year survival for HNSCC patients. Sensitivity, specificity, accuracy, and AUC of the SEML model surpassed those of the DL model, although ROC analysis indicated no statistically significant differences between PTV and GTV radiomic features (chi-square test, p > 0.05, Figure 4).
The ROC curve of GTV and PTV using deep learning model. GTV: gross tumor volume; PTV: planning target volume.
Test for SEML and DL performance
A DeLong test for AUCs is conducted to compare the performance of SEML and DL. It was noted that there is no significant difference for GTV, but there is a significant difference for PTV (Table 5).
DeLong’s test for AUCs.
n = 164
SEML (AUC)
DL (AUC)
p-value
Any statistical significance?
GTV
0.82
0.788
p = 0.2996
p > 0.05
PTV
0.982
0.712
p = 0
p < 0.001
Discussion
Using radiomics characteristics, this study assesses the performance of the SEML and DL models in predicting five-year survival in patients with HNSCC. Although there was no clear difference between the models, the SEML showed a considerable improvement in prognostic prediction, especially using PTV (P < 0.001 for PTV).
The SEML model employed in this research is the first to quantitatively explore a stacking ensemble approach for enhancing cancer prognosis predictions based on CT radiomics. Prior studies have highlighted the utility of ML in cancer prognosis, with notable successes such as an AUC of 0.61 for head and neck cancer using RF models [27] and a C-index of 0.782 for laryngeal squamous cell carcinoma prognosis [28].
The SEML model’s superior performance compared to the DL model underscores its potential in integrating various algorithms for enhanced prognostic accuracy. This approach leverages the strengths of individual classifiers, leading to significant improvements in predictive performance.
The findings align with previous research on stacking ensemble learning, which has demonstrated improved accuracy and higher AUC compared to single ML models [29, 30]. The limitations of current applications in cancer prognosis highlight the need for further exploration of the stacking ensemble approach.
It should be noted that tumor heterogeneity is not an aspect that can be evaluated through predictive models; these are cellular characteristics that must be assessed using molecular techniques and histopathology.
There is a potential link between smoking status (available in our data) and HPV-positivity as a confounding prognostic factor [19]. We propose this as a key variable for future multi-modal models.
In our study, it is noted that classical ML/ensemble methods often outperform DNNs on smaller datasets due to DNNs’ higher parameter complexity, which aligns with our findings.
While the DeLong test showed no “statistical” significance (p > 0.05) for the PTV vs. GTV AUC difference, the consistent, large-magnitude superiority of PTV features across all metrics (e.g., AUC difference of 0.162) suggests a “clinically relevant” improvement, potentially underpowered due to sample size.
Future research direction
Incorporating clinical and genomic data alongside radiomics features could enhance predictive capabilities. Recent studies have indicated that radiomics-clinical (RC) models yield higher accuracy compared to radiomics-only models [31, 32]. Similarly, radiomics-genomics (RG) models have shown promise in improving survival predictions [33]. This opens avenues for developing integrated models that utilize multi-modal data sources, enhancing the overall predictive power, such as:
Improvement for a more advanced AI model:
Hybrid frameworks that integrate DL with ensemble approaches will help overcome limitations in DL models seen in this study. While we agree that extended timeframe prediction could offer theoretical insights, we proposed multi-modal integration: adding genomic markers to improve long-term robustness of the AI model [31–33].
More sample size and feature diversity:
Enhance the SEML model using larger and more diverse cohorts, including different demographic and tumor characteristics, to ensure broad clinical applicability and robustness of the model.
External validation/Generalizability
To increase SEML model reliability and ability to apply wider, a multi-step validation approach is proposed. (1) Partnership with international institutions [e.g., TCIA, International Cancer Genome Consortium (ICGC)] to test performance based on a different demographic, and the imaging techniques; (2) Prospective studies undertaken in not less than 3 hospitals using a standard feature extraction of images; (3) Comparison with previous approaches like TNM staging, single-model radiomics; and (4) Enrichment of prediction model incorporating genomic and clinical data. It is expected that outcomes will include generalization confirmation, biases detection, and ways of FDA clearance. The timeline is 6 months for setup, 1 year for retrospective validation, and 2–3 years for prospective trials. This approach ensures rigorous, scalable validation for clinical adoption.
Clinical Deployment Pathway
Integration of the SEML model in clinical practice is likely to launch in multi-centers or institutions, followed by modifications and final implementation in the form of operations in hospital systems to support real-time prognoses. There are challenges that may involve Imaging protocol standardization, clinician training, and practices. One of them is obtaining regulatory approval (e.g., FDA/CE marking). Data privacy (e.g., GDPR/HIPAA), and more efficient computational resource use will also help make deployment easier. The pilot studies and implementation in phases are suggested to fine-tune the model to integrate it with the real-world workflows.
Hyperparameter tuning: In our future study, we adopt a systematic approach to model development that includes an initial exploration of hyperparameter settings for our base learners and the meta-learner. While the specific tuning methods, such as grid search or random search, were not detailed in the manuscript, we will conduct preliminary experiments to identify optimal parameter ranges for each model. For each base learner (DT, RF, SVM, and GLM), we will utilize default settings that are commonly accepted in previous studies, ensuring a strong baseline performance while allowing for model interpretability.
Additionally, for the meta-learner (XGBoost), the parameters were chosen based on best practices and general guidelines from the literature, focusing on performance metrics rather than exhaustive tuning, given the constraints of our dataset size.
To ensure reproducibility, we plan to provide detailed hyperparameter settings and tuning methodologies. We believe that this balance between rigor in model setup and practical application provides a solid foundation for our research findings.
Potential applications
The SEML model can pave the way for personalized treatment strategies by:
Predicting patient-specific survival outcomes to optimize treatment plans.
Integrating into clinical decision support systems for better, but not limited to, head-and-neck cancer management.
Enabling earlier interventions in high-risk cases based on model predictions.
Integrating clinical/genomic data. A study [34] indicated that the integration of pre-treatment CT-derived radiomic biomarkers and TNM stage was predictive of 5-year progression-free survival post-chemoradiation in locally advanced HNSCC (LA-HNSCC) patients, suggesting its utility for clinical risk stratification.
Conclusions
This study is the first to employ a stacking ensemble learning approach in a predictive model for estimating cancer prognosis in HNSCC patients. The SEML model demonstrated high accuracy (93%), sensitivity (100%), and specificity (83%) in predicting five-year survival. The results affirm the effectiveness of the stacking ensemble approach in enhancing prognostic accuracy, laying a foundation for its clinical application and potential to facilitate personalized treatment for cancer patients.
AbbreviationsCT
computed tomography
DL
deep learning
DNN
deep neural network
GLM
generalized linear model
GTV
gross tumor volume
HNSCC
head and neck squamous cell carcinoma
ML
machine learning
PTV
planning target volume
RF
Random forest
ROC
receiver operating characteristic
SEML
stacking ensemble machine learning
SVM
support vector machine
TCIA
The Cancer Imaging Archive
TNM
Tumor, Node, Metastasis
1Wavelet Transform is a mathematical tool that breaks down signals (like sound or images) into small wave-like components called wavelets. Since the wavelet transform captures both frequency and timing information, making it great for analyzing signals with sudden changes.
DeclarationsAuthor contributions
HYTW: Conceptualization, Supervision, Project administration. CX: Methodology, Validation, Formal analysis. FT: Conceptualization, Methodology, Software, Formal analysis, Writing—original draft, Supervision, Funding acquisition. CCYC: Investigation, Writing—original draft. VTYL: Investigation, Resources, Writing—review & editing. SWYL: Investigation, Resources. All authors have read and agreed to the published version of the manuscript.
Conflicts of interest
The authors declare no conflict of interest.
Ethical approval
This is a retrospective study using public database. No IRB review is needed.
Consent to participate
The data is obtained from public database, no informed consent is needed.
Consent to publication
Not applicable.
Availability of data and materials
The image data is available from: https://www.cancerimagingarchive.net/collection/hnscc/.
Funding
UGC Research Matching Grant: [2021-02-75 RMGS210201], TWC College Research Grant: [2023-00-51 CRG230204], TWC School Research Grant: [2023-02-52 SRG230203]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Open Exploration maintains a neutral stance on jurisdictional claims in published institutional affiliations and maps. All opinions expressed in this article are the personal views of the author(s) and do not represent the stance of the editorial team or the publisher.
SungHFerlayJSiegelRLLaversanneMSoerjomataramIJemalAet al.Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 CountriesCA Cancer J Clin2021712094910.3322/caac.2166033538338NgWTWongECYLeeVHFChanJYWLeeAWMHead and neck cancer in Hong KongJpn J Clin Oncol201848132110.1093/jjco/hyx15129145620JohnsonDEBurtnessBLeemansCRLuiVWYBaumanJEGrandisJRHead and neck squamous cell carcinomaNat Rev Dis Primers202069210.1038/s41572-020-00224-333243986PMC7944998BrayFFerlayJSoerjomataramISiegelRLTorreLAJemalAGlobal cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countriesCA Cancer J Clin20186839442430207593AminMBEdgeSBGreeneFLByrdDRBrooklandRKWashingtonMKet al.AJCC Cancer Staging Manual8th edNew YorkSpringer2017WadeJRosarioDJMacefieldRCAveryKNSalterCEGoodwinMLet al.Psychological impact of prostate biopsy: physical symptoms, anxiety, and depressionJ Clin Oncol20133142354110.1200/JCO.2012.45.480124145343Salto-TellezMJamesJAHamiltonPWMolecular pathology - the value of an integrative approachMol Oncol201481163810.1016/j.molonc.2014.07.02125160635PMC5528577LeemansCRSnijdersPJFBrakenhoffRHThe molecular landscape of head and neck cancerNat Rev Cancer2018182698229497144GilliesRJKinahanPEHricakHRadiomics: Images Are More than Pictures, They Are DataRadiology20162785637710.1148/radiol.201515116926579733PMC4734157AertsHJVelazquezERLeijenaarRTParmarCGrossmannPCarvalhoSet al.Decoding Tumour Phenotype by Noninvasive Imaging Using a Quantitative Radiomics ApproachNat Commun20145400610.1038/ncomms500624892406PMC4059926LambinPLeijenaarRTHDeistTMPeerlingsJde JongEECvan TimmerenJet al.Radiomics: the bridge between medical imaging and personalized medicineNat Rev Clin Oncol2017147496210.1038/nrclinonc.2017.14128975929TangFHChuCYWCheungEYWRadiomics AI prediction for head and neck squamous cell carcinoma (HNSCC) prognosis and recurrence with target volume approachBJR Open202132020007310.1259/bjro.2020007334381946PMC8320130KumarVGuYBasuSBerglundAEschrichSASchabathMBet al.Radiomics: The Process and the ChallengesMagn Reson Imaging20213012344810.1016/j.mri.2012.06.01022898692PMC3563280ShenWCChenSWWuKCLeePYFengCLHsiehTCet al.Predicting pathological complete response in rectal cancer after chemoradiotherapy with a random forest using 18F-fluorodeoxyglucose positron emission tomography and computed tomography radiomicsAnn Transl Med2020820710.21037/atm.2020.01.10732309354PMC7154452WuJCaoGSunXLeeJRubinDLNapelSet al.Intratumoral Spatial Heterogeneity at Perfusion MR Imaging Predicts Recurrence-free Survival in Locally Advanced Breast Cancer Treated with Neoadjuvant ChemotherapyRadiology2018288263510.1148/radiol.201817246229714680PMC6029132Haibe-KainsBAdamGAHosnyAKhodakaramiFMassive Analysis Quality Control (MAQC) Society Board of DirectorsWaldronLWangBMcIntoshCGoldenbergAKundajeAGreeneCSet al.Transparency and reproducibility in artificial intelligenceNature2020586E14610.1038/s41586-020-2766-y33057217PMC8144864BreimanLRandom ForestsMach Learn20014553210.1023/a:1010933404324ChenTGuestrinCXGBoost: A Scalable Tree Boosting System. In: Krishnapuram B, Shah M, Smola A, Aggarwal C, Shen D, Rastogi R, editors. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining; 2016 Aug 13–17; San Francisco, CA, USA. New York: Association for Computing Machinery; 2016. pp. 785–94.10.1145/2939672.2939785TamSYTangFHChanMYLaiHCCheungSPrognosis Prediction in Head and Neck Squamous Cell Carcinoma by Radiomics and Clinical InformationBiomedicines202412164610.3390/biomedicines1208164639200111PMC11352052TangFHCheungEYWongHLYuenCMYuMHHoPCRadiomics from Various Tumour Volume Sizes for Prognosis Prediction of Head and Neck Squamous Cell Carcinoma: A Voted Ensemble Machine Learning ApproachLife (Basel)202212138010.3390/life1209138036143416PMC9505304ZhangJLamSTengXMaZHanXZhangYet al.Radiomic feature repeatability and its impact on prognostic model generalizability: A multi-institutional study on nasopharyngeal carcinoma patientsRadiother Oncol202318310957836822357VallièresMFreemanCRSkameneSRElNaqa IA radiomics model from joint FDG-PET and MRI texture features for the prediction of lung metastases in soft-tissue sarcomas of the extremitiesPhys Med Biol20156054719610.1088/0031-9155/60/14/547126119045National Research CouncilFrom Cancer Patient to Cancer SurvivorLost in Transition. Washington: The National Academies Press200610.17226/11468KwonHParkJLeeYStacking Ensemble Technique for Classifying Breast CancerHealthc Inform Res201925283810.4258/hir.2019.25.4.28331777671PMC6859259AgarwalABreast Cancer Prognosis Using Stacking Ensemble [dissertation]State University of New York at Binghamton; 2020.OwusuDKNyarkoPKStacked ensemble model for recurrent head and neck squamous cell carcinoma prognosis based on clinicopathologic and genomic markersJ Math Probl Equ Stat2023412134ParmarCGrossmannPRietveldDRietbergenMMLambinPAertsHJRadiomic Machine-Learning Classifiers for Prognostic Biomarkers of Head and Neck CancerFront Oncol2015527210.3389/fonc.2015.0027226697407PMC4668290ChenLWangHZengHZhangYMaXEvaluation of CT-based radiomics signature and nomogram as prognostic markers in patients with laryngeal squamous cell carcinomaCancer Imaging2020202810.1186/s40644-020-00310-532321585PMC7178759YanFFengYA two-stage stacked-based heterogeneous ensemble learning for cancer survival predictionComplex Intell Syst2022846193910.1007/s40747-022-00791-wKumarMSinghaiSShekharSSharmaBSrivaatavaGOptimized stacking ensemble learning model for breast cancer detection and classification using machine learningSustainability2022141399810.3390/su142113998ChingJCFLamSLamCCHLuiAOYKwongJCKLoAYHet al.Integrating CT-based radiomic model with clinical features improves long-term prognostication in high-risk prostate cancerFront Oncol202313106068710.3389/fonc.2023.106068737205204PMC10186349TangFHFongYWYungSHWongCKTuCLChanMTRadiomics-Clinical AI Model with Probability Weighted Strategy for Prognosis Prediction in Non-Small Cell Lung CancerBiomedicines202311209310.3390/biomedicines1108209337626590PMC10452490SanchezIRahmanRRadiogenomics as an Integrated Approach to Glioblastoma Precision MedicineCurr Oncol Rep20242612132210.1007/s11912-024-01580-z39009914PMC11480134BruixolaGDualdeDNogueAAgustíVAnaJiménez PastorBellvísFet al.Development of CT-based radiomic model to predict 5-year progression-free survival (PFS) in locally advanced head and neck squamous cell carcinoma (LAHNSCC) treated with definitive chemoradiationJ Clin Oncol202341607610.1200/jco.2023.41.16_suppl.6076