Tumor Region Research Articles

DeepGlioSeg: advanced glioma MRI data segmentation with integrated local-global representation architecture

IntroductionGlioma segmentation is vital for diagnostic decision-making, monitoring disease progression, and surgical planning. However, this task is hindered by substantial heterogeneity within gliomas and imbalanced region distributions, posing challenges to existing segmentation methods.MethodsTo address these challenges, we propose the DeepGlioSeg network, a U-shaped architecture with skip connections for continuous contextual feature integration. The model includes two primary components. First, a CTPC (CNN-Transformer Parallel Combination) module leverages parallel branches of CNN and Transformer networks to fuse local and global features of glioma images, enhancing feature representation. Second, the model computes a region-based probability by comparing the number of pixels in tumor and background regions and assigns greater weight to regions with lower probabilities, thereby focusing on the tumor segment. Test-time augmentation (TTA) and volume-constrained (VC) post-processing are subsequently applied to refine the final segmentation outputs.ResultsExtensive experiments were conducted on three publicly available glioma MRI datasets and one privately owned clinical dataset. The quantitative and qualitative findings consistently show that DeepGlioSeg achieves superior segmentation performance over other state-of-the-art methods.DiscussionBy integrating CNN- and Transformer-based features in parallel and adaptively emphasizing underrepresented tumor regions, DeepGlioSeg effectively addresses the challenges associated with glioma heterogeneity and imbalanced region distributions. The final pipeline, augmented with TTA and VC post-processing, demonstrates robust segmentation capabilities. The source code for this work is publicly available at https://github.com/smallboy-code/Brain-tumor-segmentation.

Quantifying the tumour vasculature environment from CD-31 immunohistochemistry images of breast cancer using deep learning based semantic segmentation.

Tumour vascular density assessed from CD-31 immunohistochemistry (IHC) images has previously been shown to have prognostic value in breast cancer. Current methods to measure vascular density, however, are time-consuming, suffer from high inter-observer variability and are limited in describing the complex tumour vasculature morphometry. We propose a method for automatically measuring a range of vascular parameters from CD-31 IHC images, which together provide a detailed description of the vasculature morphology. We first used a U-Net based convolutional neural network, trained and validated using 36 partially annotated whole slide images from 27 patients, to segment vessel structures and tumour regions from which the measurements are taken. The model also segments the vascular smooth muscle, benign epithelium, adipose tissue, stroma, lymphocyte clusters, nerves and CD-31 positive leukocytes, and we applied it to an additional 21 images from 15 patients. Using these segmentations, we investigated the relationship between the various tissue types and the vasculature and studied the relationship of various vascular parameters with clinical parameters. We also performed a 3D histology analysis on a separate tumour sample as a proof of principle, providing a more comprehensive visualization of vasculature morphology compared to the standard 2D cross-section of a tissue sample. Using two-way cross-validation, we show that vessels were accurately segmented, with Dice scores of 0.875 and 0.856, and were accurately identified, with F1 scores of 0.777 and 0.748. All vascular parameters exhibit strong ( ) and significant (p<0.001) correlations with measurements taken from the manual ground truth vessel segmentations. A significant relationship between the major/minor axis ratio, a measure of elongation, and the tumour grade was found. Our proposed method shows promise as a tool for studying the tumour vasculature and its relationship with surrounding cells and tissue types. Furthermore, the correlation with tumour grade highlights the clinical relevance of our approach. These findings suggest that our method could have substantial implications for improving prognostic assessments and personalizing therapeutic strategies in breast cancer treatment.

Immune microenvironment spatial landscapes of tertiary lymphoid structures in gastric cancer

BackgroundTertiary lymphoid structures (TLS) correlate with tumour prognosis and immunotherapy responses in gastric cancer (GC) studies. However, understanding the complex and diverse immune microenvironment within TLS requires comprehensive analysis.MethodsWe examined the prognostic impact of TLS within the tumour core (TC) of 59 GC patients undergoing immunotherapy. Multispectral fluorescence imaging was employed to evaluate variations in immune cell infiltration across different TLS sites among 110 GC patients, by quantifying immune cell density and spatial characteristics. We also generated a single-cell transcriptomic atlas of TLS-positive (n = 4) and TLS-negative (n = 8) microenvironments and performed spatial transcriptomics (ST) analysis on two samples.ResultsTLS presence in the TC significantly correlated with improved immune-related overall survival (P = 0.049). CD8+LAG-3−PD-1+TIM-3−, CD4+PD-L1+, and CD4+FoxP3− T cell densities were significantly higher in the TLS within TC compared to tumour and stromal regions. Immune cells within TLS exhibited closer intercellular proximity than those outside TLS. Five key density and spatial characteristics of immune cells within TLS in the TC were selected to develop the Density and Spatial Score risk model. Single-cell RNA sequencing revealed strong intercellular interactions in the presence of TLS within the microenvironment. However, TLS-absent environment facilitated tumour cell interactions with immune cells through MIF- and galectin-dependent pathways, recruiting immunosuppressive cells. ST analysis confirmed that T and B cells co-localise within TLS, enhancing immune response activation compared to cancer nests and exerting a strong anti-tumour effect.ConclusionsTLS presence facilitates frequent cell-to-cell communication, forming an active immune microenvironment, highlighting the prognostic value of TLS.

Quantifying cell divisions along evolutionary lineages in cancer.

Cell division drives somatic evolution but is challenging to quantify. We developed a framework to count cell divisions with DNA replication-related mutations in polyguanine homopolymers. Analyzing 505 samples from 37 patients, we studied the milestones of colorectal cancer evolution. Primary tumors diversify at ~250 divisions from the founder cell, while distant metastasis divergence occurs significantly later, at ~500 divisions. Notably, distant but not lymph node metastases originate from primary tumor regions that have undergone surplus divisions, tying subclonal expansion to metastatic capacity. Then, we analyzed a cohort of 73 multifocal lung cancers and showed that the cell division burden of the tumors' common ancestor distinguishes independent primary tumors from intrapulmonary metastases and correlates with patient survival. In lung cancer too, metastatic capacity is tied to more extensive proliferation. The cell division history of human cancers is easily accessible using our simple framework and contains valuable biological and clinical information.

TFOPT-WEQN: OPTIMIZED WEIGHTED EUCLIDEAN LOSS-ENABLED QUANTUM CONVOLUTIONAL NEURAL NETWORK FOR BRAIN TUMOR DETECTION AND CLASSIFICATION

Brain tumor is a debilitating disease causing irregular growth of brain cells insisting on early tumor diagnosis for treatment planning to enhance the patients’ survival rate. Due to the miniature tumor areas, brain tumor segmentation has proven to be a challenging issue, even for neural networks. Further, the experienced radiologist delineates the tumor regions due to noise present in MRI images. Specifically, the tiny size of the tumor locations and the significant differences in area occupancy between different tumors make it impossible to identify the tumor regions. To avoid such constraints, this research proposes a trail and flee optimization-based weighted Euclidean loss-enabled quantum convolutional neural network (TFOpt-WEQN) model for effective brain tumor detection. Here, the WEQN model is integrated with hybrid weighted categorical cross entropy and Euclidean distance-based loss function to determine the significant tumor regions using featured datasets. Moreover, the TFOpt algorithm is developed by combining Harris Hawks and Red Kite optimizations to tune the model while training. In this research, the segmentation using visual geometry group-16 enabled eigen-cam-full grad-based hybrid U-Network ([Formula: see text]) model plays a key role in separating tumor regions from the MRI images. The proposed model ultimately overcomes the major limitations of previous models and accurately detects the tumor regions utilizing the BraTS dataset consisting of four MRI modalities including the FLAIR, T1, T2, and T1C for diagnosis. The experimental results demonstrate that the TFOpt-WEQN model greatly achieves maximum accuracy, positive predicted value, negative predicted value, and F1-score of 95.66%, 0.96, 0.95, and 95.72% for the BraTS-2020 dataset exceeding the other state-of-the-art methods.

Automated radiomics model for preoperative pancreatic neuroendocrine tumor grade prediction.

655 Background: Pancreatic neuroendocrine tumors (PNETs) have varying biologic behavior based on factors including tumor grade. Many prognostic factors can only be determined post-pancreatectomy; thus, are unable to guide clinical decision making preoperatively. This is particularly relevant in small, non-functional (NF)-PNETs. Imaging analysis with radiomics may be able to elucidate biologic data such as tumor grade, without the need for tissue sampling. Our study aims to create an automatic pipeline from segmentation to radiomics-signature building to preoperatively predict tumor grade. Methods: Patients that underwent resection between 2003-21 with adequate preoperative arterial phase CT scans were divided into training and validation cohorts. In the training cohort, the pancreas and tumor region were manually segmented and used to train an auto-segmentation model. The validation cohort then underwent automatic segmentation. A total of 255 radiomics features were extracted from the tumor region. Correlated features were removed, and the minimum redundancy maximum relevance (mRMR) method was used to select the final feature set. These features were used to train a Support Vector Machine (SVM)-based classifier through the training cohort to predict grade, which was dichotomized into grade I versus II/III. The radiomic based prediction model was then evaluated in the automatically segmented validation cohort. Results: A total of 186 patients were identified and divided into training (n = 140) and validation (n = 46) cohorts. Median age was 57 (27, 85) and 52% were male. 113 (62%) and 70 (38%) patients were grades I and II/III, respectively. Median tumor size was 24 mm (6, 200) and 43 (26%) patients had positive nodal disease. The auto-segmentation model was able to accurately segment the tumor region in 33 (72%) patients of the validation cohort. In the training cohort (n = 140), the radiomics-signature produced an area under the curve (AUC) of 0.87 (0.81, 0.93). Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were 0.94 (0.9, 0.98), 0.74 (0.66, 0.81), 0.72 (0.65, 0.80) and 0.97 (0.94, 1.00) respectively. In the auto-segmented validation cohort (n = 33), the radiomics model produced an AUC of 0.75 (0.61, 0.90). Sensitivity, specificity, PPV and NPV were 0.88 (0.77, 0.99), 0.62 (0.46, 0.79), 0.71 (0.56, 0.87) and 0.83 (0.71, 0.96), respectively. Conclusions: The proposed auto-segmentation model was able to automatically identify tumor regions with a high degree of accuracy. Similarly, the subsequent radiomics-signature was also able to strongly discern PNET grade. This automatic pipeline bypasses manual segmentation and could help incorporate a radiomics-signature into preoperative clinical decision-making for PNETs.

Normalization of tumor vasculature by imiquimod: proposal for a new anticancer therapeutic indication for a TLR7 agonist

Imiquimod (IMQ), a drug from aminoquinoline group, is the toll-like receptor 7 (TLR7) agonist. It acts as an immunostimulant and radio-sensitizing agent. IMQ stimulates both innate and adaptive immune response. Despite studies conducted, there are no unambiguous data showing how IMQ affects the condition of tumor blood vessels. Tumor vasculature plays the main role in tumor progression. Formation of abnormal blood vessels increases area of hypoxia which recruits suppressor cells, blocks tumor infiltration by CD8+ T lymphocytes, inhibits efficacy of chemoterapeutic drug and leads to cancer relapse. Normalization is a type of therapy targeted at abnormal tumor blood vessels. Here, we demonstrated that 50 µg of IMQ inhibits the growth of melanoma tumors more efficiently, compared to other tested doses and the control group. Dose escalation did not improve the therapeutic antitumor potential of TLR7 agonist. A dose of 50 µg of IMQ most effectively reduced tumor blood vessel density. Imiquimod normalized tumor vasculature both structurally (by reducing vessel tortuosity and increasing pericyte coverage) and functionally (by improving tumor perfusion) in a dose-dependent manner. Hypoxia regions in tumors of treated mice were significantly reduced after IMQ administration. A dose of 50 µg of IMQ had also the greatest impact on the changes in tumor-infiltrating T lymphocytes levels. TLR7 agonist inhibited angiogenesis in treated mice. Functional vascular normalization by IMQ increases the effectiveness of low dose of doxorubicin. Higher dose of IMQ showed worse effects than lower doses including decreased tumor perfusion, increased tumor hypoxia and immunosuppression. This knowledge may help to optimize the combination of the selected IMQ dose with e.g. chemotherapy or radiotherapy to elicit synergistic effect in cancer treatment. To conclude, we outline IMQ repurposing as a vascular normalizing agent. Our research results may contribute to expanding the therapeutic indications for the use of IMQ in anticancer therapy in the future.

Remodeling of tumor microenvironments by EGFR tyrosine kinase inhibitors in EGFR-mutant non-small cell lung cancer.

Patients with EGFR mutations exhibit immunosuppressive microenvironments, limiting responsiveness to immunotherapy. We used digital spatial profiling to analyze non-small cell lung carcinomas in 25 patients before and after EGFR tyrosine kinase inhibitor (TKI) treatment, including 14 patients treated with first-line osimertinib, focusing on CD45-positive immune regions and pan-cytokeratin-positive tumor regions. Osimertinib treatment resulted in altered angiogenic pathways and immune cell proportions, with reduced plasma cells (22.2%-11.7%; p= 0.025) and increased macrophage infiltration (p= 0.145). The most predominant immune subtypes before and after treatment was the interferon-γ (IFN-γ)-dominant C2 subtype and the lymphocyte-depleted C4 subtype. Two patients who showed the opposite pattern, transiting from C4 to C2, had durable responses to subsequent atezolizumab/bevacizumab/carboplatin/paclitaxel treatment. Our results shed light on the immunomodulatory effects of osimertinib treatment and suggest that co-targeting angiogenesis and anti-programmed death (ligand) 1 might be effective in EGFR-TKI-resistant non-small cell lung cancer.

A feature fusion method based on radiomic features and revised deep features for improving tumor prediction in ultrasound images.

Radiomic features and deep features are both vitally helpful for the accurate prediction of tumor information in breast ultrasound. However, whether integrating radiomic features and deep features can improve the prediction performance of tumor information is unclear. A feature fusion method based on radiomic features and revised deep features was proposed to predict tumor information. Radiomic features were extracted from the tumor region on ultrasound images, and the optimal radiomic features were subsequently selected based on Gini score. Revised deep features, which were extracted using the revised CNN models integrating prior information, were combined with radiomic features to build a logistic regression classifier for tumor prediction. The performance was evaluated using area under the receiver operating characteristic (ROC) curve (AUC). The results showed that the proposed feature fusion method (AUC=0.9845) obtained better prediction performance than that based on radiomic features (AUC=0.9796) or deep features (AUC=0.9342). Our results demonstrate that the proposed feature fusion framework integrating the radiomic features and revised deep features is an efficient method to improve the prediction performance of tumor information.

Strategies for specific multimodal imaging of cancer-associated fibroblasts and applications in theranostics of cancer.

Fibroblast activating protein (FAP) is up-regulated in cancer-associated fibroblasts (CAFs) of more than 90% of tumor microenvironment and also highly expressed on the surface of multiple tumor cells like glioblastoma, which can be used as a specific target for tumor diagnosis and treatment. At present, small-molecule radiotracer targeting FAP with high specificity exhibit limited functionality, which hinders the integration of theranostics as well as multifunctionality. In this work, we have engineered a multifunctional nanoplatform utilizing organic melanin nanoparticles that specifically targets FAP, facilitating both multimodal imaging and synergistic therapeutic applications. This nanoplatform can perform positron emission tomography (PET), magnetic resonance imaging (MRI) and photoacoustic imaging (PAI) with strong near infrared absorption and metal chelating ability, achieving efficiently targeting accumulation and display long retention in the tumor region. Meanwhile, 131I-labeled nanoplatform for targeted radioisotope therapy (TRT) and photothermal therapy (PTT) were significantly suppressed tumor growth in glioblastoma xenograft models without obvious side effects. These results demonstrated that this novel nanoparticles-based theranostics nanoplatform can effectively enhance multimodal imaging and targeted radionuclide-photothermal synergistic therapy for solid tumors with FAP expression.

Berkeley wavelet transform and improved YOLOv7-based classification technique for brain tumor severity prediction

Abnormality in brain tissues is a life-threatening illness in humans Un-bias to gender and age if it is unrecognized and untreated within time, will lead to severe complications and extreme conditions. The brain tumor is mainly influenced by a variety of unpredicted and unavoidable reasons. Its evaluation, spread pattern, and identification involves complex assignment. Its early grading and the proper classification ensure effective treatment. The proposed work attempts to extract and classify the tumor region using an automatic classification system for magnetic resonance imaging (MRI) brain tumors. A deep learning convolutional neural network-based architecture YOLO is employed to classify and detect the tumor from brain MR images. The proposed method resulted in superior segmentation, and classification performance in terms of subjective visualization and objective metrics as compared to state of art approaches. The proposed YOLO-based method collectively achieved 98.89% classification accuracy on the BRAINIX and Kaggle datasets.

Classification and Pixel Change Detection of Brain Tumor Using Adam Kookaburra Optimization-Based Shepard Convolutional Neural Network.

The uncommon growth of cells in the brain is termed as brain tumor. To identify chronic nerve problems, like strokes, brain tumors, multiple sclerosis, and dementia, brain magnetic resonance imaging (MRI) is normally utilized. Identifying the tumor on early stage can improve the patient's survival rate. However, it is difficult to identify the exact tumor region with less computational complexity. Also, the tumors can vary significantly in shape, size, and appearance, which complicates the task of correctly classifying tumor types and detecting subtle pixel changes over time. Hence, an Adam kookaburra optimization-based Shepard convolutional neural network (AKO-based Shepard CNN) is established in this study for the classification and pixel change detection of brain tumor. The Adam kookaburra optimization (AKO) is established by integrating the kookaburra optimization algorithm (KOA) and Adam. Here, the pre- and post-operative MRIs are pre-processed and then segmented by U-Net++. The tuning of U-Net++ is done by the bald Border collie firefly optimization algorithm (BBCFO). The bald eagle search (BES), firefly algorithm (FA), and Border collie optimization (BCO) are combined to form the BBCFO. The next operation is the feature extraction and the classification is conducted at last using ShCNN. The AKO is utilized to tune the ShCNN for obtaining effective classification results. Unlike conventional optimization algorithms, AKO offers faster convergence and higher accuracy in classification. The highest negative predictive value (NPV), true negative rate (TNR), true positive rate (TPR), positive predictive value (PPV), and accuracy produced by the AKO-based ShCNN are 89.91%, 92.26%, 93.78%, and 93.60%, respectively, using Brain Images of Tumors for Evaluation database (BITE).

Generating 3D brain tumor regions in MRI using vector-quantization Generative Adversarial Networks.

Medical image analysis has significantly benefited from advancements in deep learning, particularly in the application of Generative Adversarial Networks (GANs) for generating realistic and diverse images that can augment training datasets. The common GAN-based approach is to generate entire image volumes, rather than the region of interest (ROI). Research on deep learning-based brain tumor classification using MRI has shown that it is easier to classify the tumor ROIs compared to the entire image volumes. In this work, we present a novel framework that uses vector-quantization GAN and a transformer incorporating masked token modeling to generate high-resolution and diverse 3D brain tumor ROIs that can be used as additional data for tumor ROI classification. We apply our method to two imbalanced datasets where we augment the minority class: (1) low-grade glioma (LGG) ROIs from the Multimodal Brain Tumor Segmentation Challenge (BraTS) 2019 dataset; (2) BRAF V600E Mutation genetic marker tumor ROIs from the internal pediatric LGG (pLGG) dataset. We show that the proposed method outperforms various baseline models qualitatively and quantitatively. The generated data was used to balance the data to classify brain tumor types. Our approach demonstrates superior performance, surpassing baseline models by 6.4% in the area under the ROC curve (AUC) on the BraTS 2019 dataset and 4.3% in the AUC on the internal pLGG dataset. The results indicate the generated tumor ROIs can effectively address the imbalanced data problem. Our proposed method has the potential to facilitate an accurate diagnosis of rare brain tumors using MRI scans.

Spatially resolved protein profiling in pancreatic adenosquamous carcinoma.

760 Background: Pancreatic adenosquamous carcinoma (PASC) is a rare histopathological type of pancreatic cancer that is more aggressive than pancreatic ductal adenocarcinoma (PDAC). Although PASC has a unique tumor biology and poorer prognosis, there are no guidelines for differentially treating PASC patients, and its rarity makes it significantly understudied. We used Nanostring GeoMX digital spatial profiling (DSP) to provide a targeted proteomic characterization of key tumor-promoting proteins in PASC and PDAC. Methods: FFPE tissue from PASC (n=10 primary, n= 4 met) and PDAC (n=21 primary, n=5 met) cases were analyzed using the DSP platform. Sections were stained with an 82-protein cocktail of antibodies conjugated to UV-photocleavable DNA barcodes. A pan-cytokeratin morphology marker was used to select tumor regions of interest (ROI) and inferred surrounding stromal (i.e., pan-CK negative) ROIs. UV excitation of the defined ROIs was used to release the DNA barcodes and were quantified on the NanoString nCounter system. For PASC, squamous enriched ROI were inferred from corresponding p40 IHC. Corresponding clinical and mutational status for all cases were abstracted from medical records. Results: TP53 (PDAC, 80%; PASC, 83%) and KRAS (PDAC, 76%; PASC, 100%) were the most prevalent mutations . In comparing the tumor ROIs of PASC and PDAC (all cases), DSP-assessed global expression and activation of EGFR-MAPK signaling is significantly increased in PASC, whereas PI3K-AKT signaling is significantly decreased. Compared to PDAC, EGFR expression and phosphorylation (p-EGFR), as well as p-MEK1, are significantly increased in PASC. PASC showed significantly lower levels of PLCG1 and INPP4B, while the majority of other AKT signaling effectors were unchanged. For DNA damage response signaling, PASC showed significantly higher levels of p-ATM, p-CHK2, and p21, but not p-ATR. Likewise, expression of oncoproteins c-MYC and BCL6 were significantly increased in PASC (vs. PDAC). Conclusions: Given that the normal pancreas lacks squamous cells, their histological presence is a uniquely aberrant feature that can be used to guide deep spatially-resolved interrogation of complex cellular and molecular programs. Our preliminary findings show that regions of squamous cell differentiation within PASC have intrinsically higher levels of active EGFR signaling. With emerging RAS-targeted agents poised to alter the treatment landscape, high EGFR signaling in PASC suggest a possible compensatory resistance mechanism that may require additional targeting. Likewise, reliance on ATM signaling highlights this as preferential target for DDR-directed therapies. Building on these results with additional multiomic assays will inform strategies for improving the management of patients with PASC.

An explainable transformer model integrating PET and tabular data for histologic grading and prognosis of follicular lymphoma: a multi-institutional digital biopsy study.

Pathological grade is a critical determinant of clinical outcomes and decision-making of follicular lymphoma (FL). This study aimed to develop a deep learning model as a digital biopsy for the non-invasive identification of FL grade. This study retrospectively included 513 FL patients from five independent hospital centers, randomly divided into training, internal validation, and external validation cohorts. A multimodal fusion Transformer model was developed integrating 3D PET tumor images with tabular data to predict FL grade. Additionally, the model is equipped with explainable modules, including Gradient-weighted Class Activation Mapping (Grad-CAM) for PET images, SHapley Additive exPlanations analysis for tabular data, and the calculation of predictive contribution ratios for both modalities, to enhance clinical interpretability and reliability. The predictive performance was evaluated using the area under the receiver operating characteristic curve (AUC) and accuracy, and its prognostic value was also assessed. The Transformer model demonstrated high accuracy in grading FL, with AUCs of 0.964-0.985 and accuracies of 90.2-96.7% in the training cohort, and similar performance in the validation cohorts (AUCs: 0.936-0.971, accuracies: 86.4-97.0%). Ablation studies confirmed that the fusion model outperformed single-modality models (AUCs: 0.974 - 0.956, accuracies: 89.8%-85.8%). Interpretability analysis revealed that PET images contributed 81-89% of the predictive value. Grad-CAM highlighted the tumor and peri-tumor regions. The model also effectively stratified patients by survival risk (P < 0.05), highlighting its prognostic value. Our study developed an explainable multimodal fusion Transformer model for accurate grading and prognosis of FL, with the potential to aid clinical decision-making.

Large blood vessel segmentation in quantitative DCE-MRI of brain tumors: A Swin UNETR approach.

Brain tumor growth is associated with angiogenesis, wherein the density of newly developed blood vessels indicates tumor progression and correlates with the tumor grade. Quantitative dynamic contrast enhanced-magnetic resonance imaging (DCE-MRI) has shown potential in brain tumor grading and treatment response assessment. Segmentation of large-blood-vessels is crucial for automatic and accurate tumor grading using quantitative DCE-MRI. Traditional manual and semi-manual rule-based large-blood-vessel segmentation methods are time-intensive and prone to errors. This study proposes a novel deep learning-based technique for automatic large-blood-vessel segmentation using Swin UNETR architectures and comparing it with U-Net and Attention U-Net architectures. The study employed MRI data from 187 brain tumor patients, with training, validation, and testing datasets sourced from two centers, two vendors, and two field-strength magnetic resonance scanners. To test the generalizability of the developed model, testing was also carried out on different brain tumor types, including lymphoma and metastasis. Performance evaluation demonstrated that Swin UNETR outperformed other models in segmenting large-blood-vessel regions (achieving Dice scores of 0.979, and 0.973 on training and validation sets, respectively, with test set performance ranging from 0.835 to 0.982). Moreover, most quantitative parameters showed significant differences (p < 0.05) between with and without large-blood-vessel. After large-blood-vessel removal, using both ground truth and predicted masks, the values of parameters in non-vascular tumoral regions were statistically similar (p > 0.05). The proposed approach has potential applications in improving the accuracy of automatic grading of tumors as well as in treatment planning.

Inverse Problem of Parameter Estimation in Natural Convection of Iron Oxide – Distilled Water Nanofluids

Abstract Nanofluids have been used to facilitate the transport of nanoparticles to tumor regions for different purposes, such as drug delivery, promotion of antioxidant effects, and selective absorption of energy from external sources for thermal treatments. The characterization of nanofluids by solving an inverse parameter estimation model was the main objective of this work. A nanofluid of Fe2O3 nanoparticles dissolved in distilled water was heated by a diode-laser, causing natural convection currents during the experiment. The parameter estimation problem was solved within the Bayesian framework of statistics by applying the Metropolis-Hastings algorithm of the Markov Chain Monte Carlo method, thus demanding large computational times associated with stochastic simulations of a natural convection problem. A multivariate linear regression model was then trained with the high-fidelity natural convection model, in order to speed up calculations during the solution of the inverse problem. It is shown that the multivariate linear regression low-fidelity model can be used as an accurate representation of the temperatures at the heated surface of the nanofluid, thus resulting in estimated parameters with small uncertainties.

Functional connectivity between tumor region and resting-state networks as imaging biomarker for overall survival in recurrent gliomas diagnosed by FET PET

Abstract Background Amino acid PET using the tracer O-(2-[18F]fluoroethyl)-L-tyrosine (FET) is one of the most reliable imaging methods for detecting glioma recurrence. Here, we hypothesized that functional MR connectivity between the metabolic active recurrent tumor region and resting-state networks of the brain could serve as a prognostic imaging biomarker for overall survival (OS). Methods The study included 82 patients (26-81 years; median ECOG performance score, 0) with recurrent gliomas following therapy (WHO-CNS 2021 grade 4 glioblastoma, n=57; grade 3 or 4 astrocytoma, n=12; grade 2 or 3 oligodendroglioma, n=13) diagnosed by FET PET simultaneously acquired with functional resting-state MR. Functional connectivity (FC) was assessed between tumor regions and seven canonical resting-state networks. Results WHO tumor grade and IDH mutation status were strong predictors of OS after recurrence (p&amp;lt;0.001). Overall FC between tumor regions and networks was highest in oligodendrogliomas and was inversely related to tumor grade (p=0.031). FC between the tumor region and the dorsal attention network was associated with longer OS (HR, 0.88; 95%CI, 0.80-0.97; p=0.007), and showed an independent association with OS (HR, 0.90; 95%CI, 0.81-0.99; p=0.033) in a model including clinical factors, tumor volume and MGMT. In the glioblastoma subgroup, tumor volume and FC between tumor and the visual network (HR, 0.90; 95%CI, 0.82-0.99, p=0.031) were independent predictors of survival. Conclusion Recurrent gliomas exhibit significant FC to resting-state networks of the brain. Besides tumor type and grade, high FC between the tumor and distinct networks could serve as independent prognostic factors for improved OS in these patients.

430. Tumor microbiota and tumor-infiltrating T cells in the tumor microenvironment of colorectal cancer

Abstract Background Recent studies indicate that the gut microbiota influences the immune cell infiltrates in colorectal cancer. For example, a study has shown that the amount of Fusobacterium nucleatum in colorectal carcinoma has been inversely associated with CD3+ cell density in tumor. Evidence also indicates that Bacteroides fragilis may suppress the adaptive T cell response. Based on such biological evidence, we hypothesized that T cell infiltrates might differ by the colorectal tumor tissue microbiota. Methods We conducted 16s rRNA sequencing to identify the abundance of 1,518 microbiome species in 861 colorectal cancer patients in the Nurses’ Health Study and the Health Professionals Follow-up Study. Multiplex immunofluorescence combined with digital image analysis and machine learning algorithms was used to measure various T-cell subsets, such as CD3, CD4, and CD8. Spearman correlation test was performed to investigate the association between the abundance of gut microbiota and CD3+, CD3+CD4+, and CD3+CD8+ T-cell densities in the overall tumor region. For the validation, we used tumor microbial data by qPCR. Results The strongest negative association was observed between the genus Clostridium and CD3+CD4+ T-cell density (rho = -0.13, p=8.27E-05). Bacteroides fragilis had a statistically significant negative association with CD3+ T-cell density (rho = -0.10, p = 0.0026) and CD3+CD4+ T-cell density (rho = -0.12, p &amp;lt; 0.001). Fusobacterium nucleatum also had a significant negative association with CD3+ T-cell density (rho = - 0.078, p = 0.021) and CD3+CD4+ T-cell density (rho = -0.11, p = 0.0011). These findings were validated by analyses using qPCR data. Conclusion We identified tumor microbiota, the abundance of which was inversely associated with tumor-infiltrating T cells in the tumor microenvironment of colorectal cancer. The strongest negative association was observed between Clostridium and CD3+CD4+ cell density. The abundance of Bacteroides fragilis and Fusobacterium nucleatum correlated negatively with the density of CD3+ and CD3+CD4+ cells in colorectal carcinoma tissue. Our findings provide new evidence for microbial-immune interactions in the tumor microenvironment. Disclosures All Authors: No reported disclosures

The effectiveness of nurse-led visual rehabilitation on visual outcome among post-operative patients of sellar/suprasellar tumours-protocol of a randomized controlled trial

Background: Tumour or mass of the suprasellar region are common pathologies and are also important due to their adjacency to the vital anterior visual apparatus. They may cause serious ocular signs and symptoms in addition to neurological complications resulting from increased intracranial pressure, cranial nerve impairment, or brain compression. Increased ICP may lead to headache, double vision and loss of peripheral vision, can lead to sinus pain or ear pain, drooping eyelid and seizures. After surgery, patients may continue to experience neuropsychological symptoms, physical symptoms and develop complications. Methods: Randomized controlled trial. A computer random table with allocation concealment will be used to recruit and assign patients with visual complaints resulting from sellar/suprasellar tumors operated on at PGIMER to the experimental and control groups. Enrolment will be done on 5th to 7th day of surgery and intervention will be provided on the same day until discharge and booklet with nurse-led visual rehabilitation will be provided for continuous practice at home. The post-op follow-up will be done at 1 month and 3rd month of surgery. Conclusions: This aim of the study is to determine the effectiveness of “nurse-led visual rehabilitation” on visual outcome among post-op patient with sellar/suprasellar tumours at 1 month and 3 months. The comprehensive nurse-led visual rehabilitation will be designed with the inputs from all nursing and clinical specialist in neurosurgery and ophthalmology, has the ability to work well with the dynamic care of postoperative patients experiencing visual problems from suprasellar and sellar tumors. CTRI registration number: CTRI/2023/08/056030 ON 2/8/23