Xinhua Zhang

Xinhua Zhang

Auto-generated profile, based on publicly available information

Xinhua Zhang

Department of Computer Science, University of Illinois at Chicago, Chicago, IL, USA

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
20 April 2022 in Journal of Medical Internet Research
Verified
Journal Article
Open access
ETElisa Tardini
Elisa Tardini
XZXinhua Zhang
Xinhua Zhang
GCGuadalupe Canahuate
Guadalupe Canahuate
AWAndrew Wentzel
Andrew Wentzel
Abdallah Sherif Radwan MohamedAbdallah Sherif Radwan Mohamed
Abdallah Sherif Radwan Mohamed
LMLiz Marai
Liz Marai
Background: Currently, selection of patients for sequential versus concurrent chemotherapy and radiation regimens lacks evidentiary support and it is based on locally optimal decisions for each step. Objective: We aim to optimize the multistep treatment of patients with head and neck cancer and predict multiple patient survival and toxicity outcomes, and we develop, apply, and evaluate a first application of deep Q-learning (DQL) and simulation to this problem. Methods: The treatment decision DQL digital twin and the patient’s digital twin were created, trained, and evaluated on a data set of 536 patients with oropharyngeal squamous cell carcinoma with the goal of, respectively, determining the optimal treatment decisions with respect to survival and toxicity metrics and predicting the outcomes of the optimal treatment on the patient. Of the data set of 536 patients, the models were trained on a subset of 402 (75%) patients (split randomly) and evaluated on a separate set of 134 (25%) patients. Training and evaluation of the digital twin dyad was completed in August 2020. The data set includes 3-step sequential treatment decisions and complete relevant history of the patient cohort treated at MD Anderson Cancer Center between 2005 and 2013, with radiomics analysis performed for the segmented primary tumor volumes. Results: On the test set, we found mean 87.35% (SD 11.15%) and median 90.85% (IQR 13.56%) accuracies in treatment outcome prediction, matching the clinicians’ outcomes and improving the (predicted) survival rate by +3.73% (95% CI –0.75% to 8.96%) and the dysphagia rate by +0.75% (95% CI –4.48% to 6.72%) when following DQL treatment decisions. Conclusions: Given the prediction accuracy and predicted improvement regarding the medically relevant outcomes yielded by this approach, this digital twin dyad of the patient-physician dynamic treatment problem has the potential of aiding physicians in determining the optimal course of treatment and in assessing its outcomes.

10 September 2024 in IEEE Transactions on Visualization and Computer Graphics
Verified
Journal Article
Closed access
AWAndrew Wentzel
Andrew Wentzel
SASerageldin Attia
Serageldin Attia
XZXinhua Zhang
Xinhua Zhang
GCGuadalupe Canahuate
Guadalupe Canahuate
CFClifton David Fuller
Clifton David Fuller
LMLiz Marai
Liz Marai
Digital twin models are of high interest to Head and Neck Cancer (HNC) oncologists, who have to navigate a series of complex treatment decisions that weigh the efficacy of tumor control against toxicity and mortality risks. Evaluating individual risk profiles necessitates a deeper understanding of the interplay between different factors such as patient health, spatial tumor location and spread, and risk of subsequent toxicities that can not be adequately captured through simple heuristics. To support clinicians in better understanding tradeoffs when deciding on treatment courses, we developed DITTO, a digital-twin and visual computing system that allows clinicians to analyze detailed risk profiles for each patient, and decide on a treatment plan. DITTO relies on a sequential Deep Reinforcement Learning digital twin (DT) to deliver personalized risk of both long-term and short-term disease outcome and toxicity risk for HNC patients. Based on a participatory collaborative design alongside oncologists, we also implement several visual explainability methods to promote clinical trust and encourage healthy skepticism when using our system. We evaluate the efficacy of DITTO through quantitative evaluation of performance and case studies with qualitative feedback. Finally, we discuss design lessons for developing clinical visual XAI applications for clinical end users.

20 August 2024 in medRxiv
Verified
Preprint
Open access
PDF file
AHAnderson Head
Anderson Head
NGNeck Symptom Working Group
Neck Symptom Working Group
LHLaia Humbert-Vidan
Laia Humbert-Vidan
SKSerageldin Kamel
Serageldin Kamel
AWAndrew Wentzel
Andrew Wentzel
CFClifton David Fuller
Clifton David Fuller
Background Existing studies on osteoradionecrosis of the jaw (ORNJ) have primarily used cross-sectional data, assessing risk factors at a single time point. Determining the time-to-event profile of ORNJ has important implications to monitor oral health in head and neck cancer (HNC) long-term survivors. Methods Demographic, clinical and dosimetric data were retrospectively obtained for a clinical observational cohort of 1129 patients with HNC treated with radiotherapy (RT) at The University of Texas MD Anderson Cancer Center. ORNJ was diagnosed in 198 patients (18%). A multivariable logistic regression analysis with forward stepwise variable selection identified significant predictors for ORNJ. These predictors were then used to train a Weibull Accelerated Failure Time (AFT) model, which was externally validated using an independent cohort of 265 patients (92 ORNJ cases and 173 controls) treated at Guy’s and St. Thomas’ Hospitals. Findings Our model identified that each unit increase in D25% is significantly associated with a 12% shorter time to ORNJ (Adjusted Time Ratio [ATR] 0·88, p<0·005); pre-RT dental extractions was associated to a 27% faster (ATR 0·73, p=0·13) onset of ORNJ; male patients experienced a 38% shorter time to ORNJ (ATR 0·62, p = 0·11). The model demonstrated strong internal calibration (integrated Brier score of 0·133, D-calibration p-value 0.998) and optimal discrimination at 72 months (Harrell’s C-index of 0·72). The model also showed good generalization to the independent cohort, despite a slight drop in performance. Interpretation This study is the first to demonstrate a direct relationship between radiation dose and the time to ORNJ onset, providing a novel characterization of the impact of delivered dose not only on the probability of a late effect (ORNJ), but the conditional risk during survivorship. Funding This work was supported by various funding sources including NIH, NIDCR, NCI, NAPT, NASA, BCM, Affirmed Pharma, CRUK, KWF Dutch Cancer Society, NWO ZonMw, and the Apache Corporation.

23 March 2026
Verified
Conference Paper
Open access
EAEric Ababio Anyimadu
Eric Ababio Anyimadu
XZXinhua Zhang
Xinhua Zhang
CFClifton David Fuller
Clifton David Fuller
LMLiz Marai
Liz Marai
GCGuadalupe Canahuate
Guadalupe Canahuate
Normal tissue complication probability (NTCP) modeling using dose–volume histograms (DVHs) is fundamentally challenged by high dimensionality, severe multicollinearity, and substantial overlap between DVH profiles of patients with differing toxicity outcomes, limiting the effectiveness of classical classification approaches. We introduce SC–NTCP, a supervised contrastive learning framework that transforms DVH data into a compact, separable latent representation optimized for predicting osteoradionecrosis (ORN) in head and neck cancer patients. Rather than relying on raw, high-dimensional DVH features, SC–NTCP explicitly maximizes intra-class similarity and inter-class separability within the embedding space, enabling more accurate downstream classification. Using a cohort of head and neck cancer patients, we benchmarked SC–NTCP against logistic regression, support vector machines, multilayer perceptrons, and convolutional neural networks. SC–NTCP demonstrated superior discrimination (AUC = 0.77), improved calibration, and enhanced interpretability via gradient-based feature attribution, while the integration of clinical covariates further augmented predictive performance. By addressing the inherent limitations of DVH data, SC–NTCP offers a principled and interpretable approach for robust radiation toxicity prediction, with the potential to inform personalized treatment planning and improve clinical outcomes.

7 April 2021 in JMIR Preprints
Verified
Preprint
Open access
ETElisa Tardini
Elisa Tardini
XZXinhua Zhang
Xinhua Zhang
GCGuadalupe Canahuate
Guadalupe Canahuate
AWAndrew Wentzel
Andrew Wentzel
Abdallah Sherif Radwan MohamedAbdallah Sherif Radwan Mohamed
Abdallah Sherif Radwan Mohamed
LMLiz Marai
Liz Marai
BACKGROUND Currently, selection of patients for sequential versus concurrent chemotherapy and radiation regimens lacks evidentiary support and it is based on locally optimal decisions for each step. OBJECTIVE We aim to optimize the multistep treatment of patients with head and neck cancer and predict multiple patient survival and toxicity outcomes, and we develop, apply, and evaluate a first application of deep Q-learning (DQL) and simulation to this problem. METHODS The treatment decision DQL digital twin and the patient’s digital twin were created, trained, and evaluated on a data set of 536 patients with oropharyngeal squamous cell carcinoma with the goal of, respectively, determining the optimal treatment decisions with respect to survival and toxicity metrics and predicting the outcomes of the optimal treatment on the patient. Of the data set of 536 patients, the models were trained on a subset of 402 (75%) patients (split randomly) and evaluated on a separate set of 134 (25%) patients. Training and evaluation of the digital twin dyad was completed in August 2020. The data set includes 3-step sequential treatment decisions and complete relevant history of the patient cohort treated at MD Anderson Cancer Center between 2005 and 2013, with radiomics analysis performed for the segmented primary tumor volumes. RESULTS On the test set, we found mean 87.35% (SD 11.15%) and median 90.85% (IQR 13.56%) accuracies in treatment outcome prediction, matching the clinicians’ outcomes and improving the (predicted) survival rate by +3.73% (95% CI –0.75% to 8.96%) and the dysphagia rate by +0.75% (95% CI –4.48% to 6.72%) when following DQL treatment decisions. CONCLUSIONS Given the prediction accuracy and predicted improvement regarding the medically relevant outcomes yielded by this approach, this digital twin dyad of the patient-physician dynamic treatment problem has the potential of aiding physicians in determining the optimal course of treatment and in assessing its outcomes.

30 April 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
EAEric Ababio Anyimadu
Eric Ababio Anyimadu
XZXinhua Zhang
Xinhua Zhang
AMAmy Moreno
Amy Moreno
CFClifton David Fuller
Clifton David Fuller
LMLiz Marai
Liz Marai
GCGuadalupe Canahuate
Guadalupe Canahuate
Objective:Modeling normal tissue complication probability (NTCP) from dose–volume histogram (DVH) data, a 3D planning graph used in radiation therapy to summarize the radiation dose distribution within specific anatomical structures, remains challenging due to the high dimensionality of DVHs and the complex, multi-organ interactions underlying radiation-induced toxicities. Conventional approaches often rely on simplified DVH summaries or single-organ features, limiting their ability to capture full dose–volume relationships and inter-organ effects. This study proposes a self-supervised framework to learn discriminative representations from multi-organ DVH data for improved toxicity prediction. Methods:We developed a self-supervised contrastive clustering framework that learns compact, discriminative representations directly from full multi-organ DVH data. Structure-specific convolutional encoders independently process each organ’s DVH, and the resulting embeddings are jointly refined through a deep embedded clustering objective that promotes well-separated latent representations. The framework was applied to a cohort of head and neck cancer (HNC) patients treated at The University of Texas MD Anderson Cancer Center to predict patient-reported xerostomia severity at six months post-radiotherapy. Results:The framework identified three distinct dosimetric subgroups with a silhouette score of 0.60. Gradient-based attribution analysis revealed that the parotid glands and hard palate, particularly within low-to-intermediate dose ranges, were the primary contributors to cluster differentiation, consistent with established dose–response relationships for xerostomia. Incorporating cluster assignments into a logistic regression model improved prediction performance with an AUC of 0.74 compared to 0.66 using clinical variables alone, with cluster membership emerging as the strongest predictor.Conclusion:Self-supervised dosimetric representation learning can uncover clinically meaningful subgroups and enhance toxicity prediction. The proposed framework provides a principled and scalable alternative to handcrafted feature engineering for NTCP modeling.