Biology subjects

Soni, N. D.

Publications and source records attributed to Soni, N. D..

3 recordsLinked to original sources

bioRxiv2026

Contrast-induced changes in chemical exchange saturation transfer MRI differentiate tumor progression from pseudoprogression

Tumor pseudo-progression (PsP) refers to an initial increase in tumor size or the appearance of new lesions. These pseudo-progressive lesions are predominantly composed of infiltrative inflammatory cells, such as macrophages. This phenomenon commonly occurs in patients undergoing radiation therapy or immunotherapy and typically indicates a positive treatment response. However, it often leads to premature treatment cessation due to misinterpretation as disease progression. Non-invasive imaging biomarkers capable of distinguishing pseudo-progression from true progression would greatly aid in treatment decision-making. In our preliminary study, we explored the potential of gadoterate meglumine (Gd-DOTA, a macrocyclic Gd-contrast) in combination with amine chemical-exchange saturation transfer (amine-CEST) imaging to differentiate tumor from radiation necrosis by assessing Gd-DOTA uptake by infiltrating immune cells, such as macrophages. To evaluate whether amine-CEST, in combination with Gd-DOTA, can differentiate macrophages from cancer cells, we incubated them with Gd-DOTA for 30 minutes. Subsequently, the cells were processed, and amine-CEST imaging was performed on a 9.4 Tesla preclinical scanner. Upon treatment with Gd-DOTA, we did not observe a significant change in amine-CEST contrast in F98 cells compared with untreated cells, whereas treated macrophages exhibited a marked decrease ([~]40%) in amine-CEST signal compared with untreated macrophages. This reduction in signal was attributed to the uptake of Gd-DOTA by macrophages, which notably shortened water T1 relaxation, thereby quenching the amine-CEST signal. Conversely, cancer cells showed no appreciable change in the amine-CEST signal, indicating no Gd-DOTA uptake. Furthermore, to validate that T1 shortening influences amine-CEST signal, cancer cells were also treated with manganese chloride (MnCl2) for 30 minutes. The uptake of MnCl2 by cancer cells similarly induced T1 shortening, as observed in macrophages, resulting in a decrease in the amine-CEST signal from these cells. Next, we performed the amin-CEST imaging on F98 tumor-bearing rats and radiation necrotic rats. Post-injection with Gd-DOTA showed no appreciable change in the amine-CEST contrast in the tumor-bearing rat, whereas a significant decrease in contrast was observed in the radiation necrotic rat. This further demonstrates that no change in the amine-CEST contrast in tumor-bearing rats is due to cancer cells failing to take up Gd-DOTA. The decrease in amine-CEST contrast in radiation-treated rats reflects the uptake of Gd-DOTA by macrophages infiltrating the radiation-necrotic regions. This straightforward imaging approach holds promise for clinical translation. It offers a novel method for characterizing pseudo-progressive lesions and monitoring diverse treatment responses in cancer patients using standard clinical scanners.

cancer biology↗

bioRxiv2026

Unsupervised anomaly detection for tumor delineation in a preclinical model of glioblastoma using CEST MRI

IntroductionGlioblastoma is characterized by heterogeneous tumor characteristics and infiltrative tumor boundaries, making accurate delineation difficult with extensive manual annotations. Chemical exchange saturation transfer (CEST) is a non-invasive MRI technique used for in vivo assessment of metabolic and macromolecular information through a Z-spectrum. CEST may provide insight into metabolic changes present in early-stage disease that are not visible in routine clinical imaging, thereby improving tumor delineation. In this work, we use an unsupervised anomaly detection (UAD) strategy to learn the distribution of features present in Z-spectra of healthy tissue and capture their deviations in pathology, foregoing the need for extensive labels. The approach leverages the metabolic information provided by CEST to improve the detection and delineation of glioblastoma and inform further treatment planning. MethodsA 1D convolutional autoencoder (CAE) was implemented to reconstruct Z-spectra from individual tissue voxels. The network was trained on Z-spectra acquired at 9.4T from healthy Sprague-Dawley rats and tested on data acquired from F98 glioma-bearing rats post Gd-administration. For baseline comparisons, Isolation Forest and Local Outlier Factor, which have shown success in anomaly detection, were implemented. For the CAE, our anomaly score was determined to be the mean squared reconstruction error. To facilitate clinical translation and evaluate the robustness of our model for under sampled Z-spectra, acceleration factors of 2x and 7x were performed with two sampling schemes: uniformly skipping frequency offsets and selecting offsets based on feature importance identified by Shapley value analysis and Integrated Gradients (IG). Binarization was performed by determining an optimal anomaly threshold, followed by comparison to ground truth tumor masks. Metrics related to model performance were assessed for baseline anomaly detectors on the fully sampled dataset and for the CAE on fully and under sampled datasets. ResultsThe best baseline anomaly detector was Isolation Forest, with an ROC-AUC of 0.967 and an F1-score of 0.584. Our method, the CAE, accurately reconstructed Z-spectral features, achieving Dice scores of up to 0.72 and outperforming the baseline model with an ROC-AUC of 0.968 and F1-score of 0.642. This model performance remained robust across sampling schemes and acceleration factors, with ROC-AUCs of [~]0.96 and similar Dice scores (up to 0.7). Feature importance analysis indicated that offsets in the range of {+/-}3.0 to 5.0ppm contributed most to the anomaly score. DiscussionThis study successfully demonstrated a UAD pipeline utilizing the Z-spectrum from CEST MRI for metabolically informed tumor delineation. The framework captures biochemical deviations that may precede or extend beyond morphologic abnormalities, enabling sensitive detection of tumor regions and intra-tumoral heterogeneity that previous methods may fail to capture. The offsets from the feature analysis indicated a strong contribution from the magnetization transfer (MT) pool to the spectral deviations captured by the model, with additional contributions from relayed nuclear Overhauser effect (rNOE) and amide proton transfer (APT). Model robustness with under sampling further highlights the pipelines potential in accelerated acquisitions, thus improving clinical practicality. While there is a need for validation on larger cohorts and clinical datasets, the current results demonstrate that this label-free, Z-spectral anomaly mapping can serve as an interpretable and scalable tool for monitoring tumor heterogeneity and progression, with potential applicability to other diffuse or metabolically subtle pathologies.

cancer biology↗

bioRxiv2025

Multimodal MR Imaging for quantification of brain lipid in mice at 9.4T

BackgroundAdvanced MR imaging techniques like steady state Nuclear Overhauser enhancement (ssNOE), transient NOE (tNOE), and myelin water fraction (MWF) provide a non-invasive way to assess the biochemical and structural integrity of brain tissue. Their sensitivity to endogenous lipids and macromolecules allows for the early detection of neuropathological changes, making them valuable tools in studying brain health and disease progression. In this study, we systematically evaluate the repeatability and sensitivity of NOEMTR, tNOE, and MWF for quantifying lipid and myelin content in the brains of wild-type (WT) mice, correlating the results with immunohistochemistry (IHC). MethodsFive 6-month-old C57BL6/J mice were imaged using 3D-NOE, and four mice underwent imaging with 2D tNOE and MWF across four repeated sessions using a 9.4T Scanner. For ssNOE imaging, CEST-weighted images at 56 frequency offsets were acquired using B1rms of 1.0 T and 3s saturation duration. For tNOE, 52 offsets were acquired with a hyperbolic secant inversion pulse (bandwidth = 400Hz, duration = 44ms) and a mixing time of 200ms. For MWF, a multi-echo spin-echo (MESE) sequence was acquired with 40 evenly spaced echoes from 5.5ms to 200ms. For both ssNOE and tNOE, B0 correction was performed using WASSR. Repeatability was quantified using intra- and inter-subject coefficients of variation (COV%). Pearson correlation was performed to see the association between imaging matrices and IHC measures, Luxol fast blue (LFB) stained sections, and myelin basic protein (MBP). ResultsAll techniques demonstrated high repeatability across the whole brain (WB) and selected regions of interest (ROIs). Whole-brain intra-subject COV% for NOEMTR ranged from 1.92% to 3.40%, with corresponding inter-subject COVs of 1.50%. tNOE exhibited improved intra-subject repeatability with COVs ranging from 0.75% to 5.57%, but a reduced inter-subject COV of 2.97%. MWF imaging showed the highest stability overall, with an intra-subject COV ranging from 0.47% to 2.03% and an inter-subject COV of 0.75%. Visually, tNOE offers superior contrast in myelin-rich areas compared to NOEMTR and MWF imaging, showing greater sensitivity to myelinated regions. tNOE strongly correlates with histological markers: r = 0.83 with MBP staining and r = 0.72 with LFB staining (both p < 0.001). MWF and NOEMTR showed correlations with MBP (r = 0.63 and r = 0.57, respectively). ConclusionNOEMTR, tNOE, and MWF imaging are reliable and repeatable methods for quantifying macromolecules in the brain. Among these, tNOE emerges as the most sensitive for detecting myelin lipids as confirmed by histological validation. These findings highlight the translational potential of tNOE for studying demyelinating disorders and neurodegenerative diseases.

neuroscience↗