Yihui Zhu

All Publications

• A Data-Driven Normalization Framework for Subject-Specific Cerebrovascular Reactivity Assessment in Cerebrovascular Disease. Journal of magnetic resonance imaging : JMRI Zhu, Y., Dogra, S., Wang, X., Polimeni, J. R., Dehkharghani, S. 2026

  Abstract

  Assessment of cerebrovascular reactivity (CVR) has been reported using acetazolamide-augmented blood oxygenation level-dependent (BOLD) MRI; however, wide intersubject baseline variability can complicate interpretation.To develop an integrated normalization framework combining machine learning-guided identification of healthy voxels for subject-specific baseline rescaling with atlas-based assessment.Retrospective.26 subjects (age 51 ± 13 at first scan, 50% women, 34 exams) with unilateral steno-occlusive disease (SOD) and 9 subjects with bilateral disease (age 55 ± 12, 4/9 women) underwent acetazolamide-augmented BOLD-MRI.T1-weighted MPRAGE (1 mm isotropic) and gradient-echo EPI at 3 T.Machine learning models (random forest, XGBoost, LightGBM, and neural networks) were trained to identify healthy candidate voxels using a previously reported pipeline based on baseline resting-state BOLD temporal shift features derived from 32 arterial territories. A healthy reference CVR atlas was constructed from unaffected hemispheres. Individual CVR maps were rescaled using predicted healthy voxels and converted to Z-scores and risk index maps using the reference atlas.Coefficient of variation (CV) within normal hemispheres, lesion contrast-to-noise ratio (CNR), Lesion ROC AUC, asymmetry index (AI) in unilateral patients, and chi-square distance between thresholded normal and abnormal CVR distributions across all patients were compared between atlas-based group normalization alone and the proposed integrated normalization using paired t-tests. (cutoff p = 0.05).CV within normal hemispheres significantly decreased (1.07 ± 0.15 vs. 0.45 ± 0.09), indicating improved inter-subject stability. Lesion CNR and ROC AUC improved across all impairment thresholds (CNR: 0.73-1.12 vs. 0.79-1.19; AUC: 0.76-0.89 vs. 0.81-0.90). AI increased significantly at higher risk thresholds (≥ 0.78), with no significant difference at the lowest threshold (p = 0.181). Chi-square analysis demonstrated significantly increased separation between CVR distributions at intermediate thresholds.The proposed integrated normalization improves the stability, discriminability, and interpretability of acetazolamide-augmented BOLD-CVR for detection of cerebrovascular impairment.3.Stage 2.

  View details for DOI 10.1002/jmri.70388

  View details for PubMedID 42233513

• Automated calculation of background parenchymal enhancement as a biomarker of treatment responses and recurrence-free survival in breast cancer BREAST CANCER RESEARCH AND TREATMENT Zhu, Y., Hadidchi, R., Nguyen, H., Hariharan, S., Weiss, J., Hou, W., Chung, C., Luu, H., Ragupathi, S., Maldjian, T., Duong, T. Q. 2026; 216 (3)

  Abstract

  To determine whether automated quantification of background parenchymal enhancement (BPE) from dynamic contrast-enhanced MRI (DCE-MRI) can serve as an imaging biomarker for clinical outcomes including overall survival (OS), recurrence-free survival (RFS), and pathological complete response (pCR) in breast cancer.The multi-institutional data consisted of 922 biopsy-confirmed invasive breast cancer patients from the Duke-Breast-Cancer-MRI dataset and 152 patients with whole-breast pre- (T0) and/or post (T3) DCE-MRI from the I-SPY2 dataset for validation. Automated fibroglandular tissue (FGT) segmentation and BPE quantification were performed on DCE-MRI. The optimal intensity enhancement threshold by volume-based method was established against four radiologist-defined BPE categories. The area under the curve (AUC) was obtained for classification of BPE categories. Cox proportional hazards models were used to predict OS and RFS. Logistic regression was used to predict pCR.Peak-contrast BPE showed strong correlation with radiologist-defined BPE, achieving the best performance at a 55% signal enhancement threshold (AUC 0.70-0.86). The calculated BPE decreased after neoadjuvant chemotherapy. A reduction in calculated BPE grade after neoadjuvant chemotherapy was predictive of pCR for the high baseline BPE group (adjusted odds ratio = 5.88 [1.03, 33.33]) and for the low baseline BPE group (adjusted odds ratio = 6.54 [1.26, 33.94]). Baseline BPE was independently associated with improved OS (adjusted hazard ratio 0.58 [0.34, 0.99]) but not associated with RFS.Automated quantification of BPE from DCE-MRI provides an objective and reproducible imaging biomarker associated with treatment response and overall survival in breast cancer. These results support its potential utility for individualized risk stratification and therapeutic decision-making.

  View details for DOI 10.1007/s10549-026-07941-5

  View details for Web of Science ID 001720779000001

  View details for PubMedID 41872564

  View details for PubMedCentralID PMC13009020

• Automated In-line Normalization Procedure for BOLD-CVR Using the Resting-State Temporal Shift with Machine Learning. AJNR. American journal of neuroradiology Zhu, Y., Dogra, S., Wang, X., Polimeni, J. R., Dehkharghani, S. 2026

  Abstract

  Cerebrovascular reactivity (CVR) is commonly used to estimate hemodynamic impairment. Conventional use is best suited to unilateral vascular disease, such that CVR can be normalized to reference values from the contralateral hemisphere or to posterior circulation territories; however, major confounds have been identified that leave implementation difficult in more common cases of bilateral disease, even despite common cerebellar normalization. Recently, we reported data-driven identification of candidate healthy voxel signatures learned from contemporaneous imaging data. Here, we introduce an entirely inline, automated approach exploiting the dynamics of resting-state BOLD (rs-BOLD) signal from the BOLD baseline, hypothesizing prediction to within ten percent error relative to ground truth healthy-voxel CVR values.22 subjects with strictly unilateral intracranial steno-occlusive disease (SOD) underwent 28 CVR studies under pharmacologic provocation using acetazolamide with BOLD-MRI (ACZ-BOLD). Separate affected and unaffected hemispheric masks were segmented to train machine learning models to learn signatures of the unaffected hemisphere using the rs-BOLD baseline, as well as anatomic and vascular parameters. Twenty additional healthy subjects from the Human Connectome Project supplemented training, wherein all voxels were classified normals. 32 distinct time-delays were computed voxelwise, with 32 maximum correlation values constrained to each of 32 paired arterial territories. Performance in prediction of ground-truth reference CVR was computed and compared.The ensembled model achieved AUC of 0.81 in predicting candidate unaffected voxels, demonstrating strong performance in estimation of normal-hemisphere CVR (median absolute percent error [95%CIs] 7.28[3.48-10.34] and 5.61[2.90-9.86] to predict median and mean reference CVR), exhibiting significant improvements over naïve whole-brain voxel selection (P=0.005 and P=0.048, respectively) or conventional cerebellar normalization (26.4[10.1-40.3] median and 27.6[23.7-33.2] mean). In nine bilateral cases assessed to illustrate potential use, the proportion of candidate voxels and corresponding volumes predicted by the ensembled model was significantly lower than in most healthy hemispheres, but yielded subjectively improved delineation of putatively abnormal regions.We demonstrate feasibility of learning unaffected reference voxel CVR signatures for BOLD-CVR MRI. The approach facilitates extension of brain CVR beyond existing constraints in subjects with bilateral disease.

  View details for DOI 10.3174/ajnr.A9267

  View details for PubMedID 41741218

• Review of Prognostic Significance of Quantitative BPE Measurements. Diagnostics (Basel, Switzerland) Weiss, J., Hunt, E., Zhu, Y., Duong, T. Q., Maldjian, T. 2026; 16 (3)

  Abstract

  Background/Objectives: Background parenchymal enhancement (BPE) on breast magnetic resonance imaging reflects hormonal and vascular activity of fibroglandular tissue and is studied as a prognostic marker for breast cancer. This paper serves as a review that evaluates quantitative methods for BPE measurements for predicting treatment outcomes. Methods: PubMed was searched for papers on evaluating BPE with outcomes to compare, such as pathologic complete response, recurrence-free survival, disease-free survival, and overall survival, from 2015 to 2025. In total, eleven papers using quantitative methods to measure BPE were selected. Results: Quantitative results showed that BPE reduction during neoadjuvant chemotherapy and high pre-treatment/baseline BPE are linked to improved treatment response and reduced risk of recurrence. Conclusions: Quantitative assessment methods yield objective and reproducible prognostic information. Incorporating quantitative BPE measurements alongside tumor-focused imaging features may further improve predictive accuracy in clinical settings.

  View details for DOI 10.3390/diagnostics16030495

  View details for PubMedID 41681813

• Aggregate Multi-tiered Normalization to Enhance Detection of Impairment in Dynamic BOLD-CVR for Assessment of Hemodynamic Impairment Zhu, Y., Wang, X., Dogra, S., Voorst, H., Heylen, E., Konduri, P. R., Schiphorst, A., Heit, J., Lansberg, M., Albers, G., Dehkharghani, S. LIPPINCOTT WILLIAMS & WILKINS. 2026

  View details for DOI 10.1161/str.57.suppl_1.WP265

  View details for Web of Science ID 001690949600028

• Using Data-Driven Methods to Improve Brain Blood Flow Measurements in Cerebrovascular Disease with Dynamic Imaging. AJNR. American journal of neuroradiology Dogra, S., Wang, X., Gee, J. M., Zhu, Y., Ishida, K., Dehkharghani, S. 2025

  Abstract

  Cerebrovascular reactivity (CVR) is a widely studied biomarker of cerebral hemodynamics, commonly used in risk stratification and treatment planning in patients with steno-occlusive disease (SOD). Conventional use relies on normalization of estimates to contralateral hemisphere reference values, which is unsuitable for bilateral or indeterminate distributions of disease. We report upon a custom data-driven approach leveraging random forest classifiers (RFc) to identify candidate voxels for normalization in order to facilitate interrogation outside conditions of known unilateral SOD MATERIALS AND METHODS: We retrospectively analyzed 16 patients with unilateral SOD who underwent acetazolamide-augmented BOLD-MRI and DSC perfusion. Three RFc models were trained using leave-one-out cross-validation (LOOCV) to identify candidate voxels brain-wide whose CVR were within 10% of the normal hemispheric median: i. all voxels; ii. gray matter only; and iii. white matter only. Model input features included time-to-maximum (Tmax), mean transit time (MTT), cerebral blood flow (CBF), and cerebral blood volume (CBV) from contemporaneous DSC. The median model-predicted reference CVR (CVRref) was compared to ground-truth medians in LOOCV, and its impact on threshold-based volumetric classification of CVR reduction assessed.RFc models effectively predicted ground-truth CVR voxels, achieving median absolute percent differences of 12.8% (IQR: 5.0%-18.9%) using all voxels, 11.3% (IQR: 9.3%-16.1%) for gray matter, and 9.8% (IQR: 4.4%-16.9%) for white matter. Volumetric estimates of CVR reduction across thresholds for the models revealed excellent agreement between ground-truth and model estimates without statistically significant differences (p>0.01), excepting lowest white matter CVR thresholds. Model use in a small pilot deployment of bilateral SOD cases demonstrated the potential utility, enabling voxel-wise CVR assessment without reliance on contralateral reference.We present a novel data-driven approach for normalizing CVR maps in patients with bilateral or indeterminate SOD. Using an RFc, our method provides an individualized, brain-wide reference CVR, expanding the utility of CVR estimates beyond the typical constraints of unilateral disease, and with potential application to other, similarly constrained scenarios such as for SPECT or PET hemodynamic studies.CVR = cerebrovascular reactivity; RFc = random forest classifier; SOD = steno-occlusive disease.

  View details for DOI 10.3174/ajnr.A8813

  View details for PubMedID 40262947