Tumor Region Research Articles

Review of deep learning-based pathological image classification: From task-specific models to foundation models

Pathological diagnosis is considered the gold standard in cancer diagnosis, playing a crucial role in guiding treatment decisions and prognosis assessment for patients. However, achieving accurate diagnosis of pathology images poses several challenges, including the scarcity of pathologists and the inherent subjective variability in their interpretations. The advancements in whole-slide imaging technology and deep learning methods provide new opportunities for digital pathology, especially in low-resource settings, by enabling effective pathological image classification. In this article, we begin by introducing the datasets, which include both unimodal and multimodal types, as essential resources for advancing pathological image classification. We then provide a comprehensive overview of deep learning-based pathological image classification models, covering task-specific models such as supervised, unsupervised, weakly supervised, and semi-supervised learning methods, as well as unimodal and multimodal foundation models. Next, we review tumor-related indicators that can be predicted from pathological images, focusing on two main categories: indicators that can be recognized by pathologists, such as tumor classification, grading, and region recognition; and those that cannot be recognized by pathologists, including molecular subtype prediction, tumor origin prediction, biomarker prediction, and survival prediction. Finally, we summarize the key challenges in digital pathology and propose potential future directions.

Development of a Nomogram-Integrated Model Incorporating Intra-tumoral and Peri-tumoral Ultrasound Radiomics Alongside Clinical Parameters for the Prediction of Histological Grading in Invasive Breast Cancer

To develop a comprehensive nomogram to predict the histological grading of breast cancer and further examine its clinical significance by integrating both intra-tumoral and peri-tumoral ultrasound radiomics features. In a retrospective study 468 female breast cancer patients were analyzed from 2017 to 2020 at the Second Affiliated Hospital of Harbin Medical University. Patients were grouped into high-grade (n = 215) and low-grade (n = 253) categories based on pathological evaluation. Tumor regions of interest were defined and expanded automatically to peri-tumor regions of interest. Ultrasound radiomics features were extracted independently. To ensure rigor, cases were randomly divided into 80% training and 20% test sets. Optimal features were selected using statistical and machine learning methods. Intra-tumor, peri-tumor, and combined radiomics models were constructed. To determine the best predictors of breast cancer histological grading, we screened the features using single- and multi-factor logistic regression analyses. Finally, a nomogram was developed and evaluated for its predictive value in this context. By applying logistic regression, we integrated ultrasound, clinicopathologic, and radiomics features to generate a nomogram. The combined model outperformed others, achieving areas under the curve of 0.934 and 0.812 in training and test sets. Calibration curves also showed high accuracy and reliability. A nomogram constructed through the integration of combined intra-tumor-peri-tumor ultrasound radiomics features along with clinicopathologic characteristics exhibited remarkable performance in distinguishing the histologic grades of invasive breast cancer.

Targeting Hypoxia-Inducible Factor-1α in Pancreatic Cancer: siRNA Delivery Using Hyaluronic Acid-Displaying Nanoparticles

Background/Objectives: Conventional anticancer therapies often lack specificity, targeting both cancerous and normal cells, which reduces efficacy and leads to undesired off-target effects. An additional challenge is the presence of hypoxic regions in tumors, where the Hypoxia Inducible Factor (HIF) transcriptional system drives the expression of pro-survival and drug resistance genes, leading to radio- and chemo-resistance. This study aims to explore the efficacy of targeted nanoparticle (NP)-based small interfering RNA (siRNA) therapies in downregulating these genes to enhance treatment outcomes in pancreatic cancer, a tumor type characterized by high CD44 expression and hypoxia. Methods: We utilized hyaluronic acid (HA)-displaying nanoparticles composed of positively charged chitosan (CS) complexed with siRNA to target and knock down HIF-1α in pancreatic cancer cells. Two NP formulations were prepared using either low molecular weight (LMW) or high molecular weight (HMW) CS. These formulations were evaluated for their internalization by cells and their effectiveness in gene silencing, both in vitro and in vivo. Results: The study found that the molecular weight (MW) of CS influenced the interaction between HA and CD44, as well as the release of siRNA upon internalization. The LMW CS formulation shows faster uptake kinetics, while HMW CS is more effective in gene knockdown across different cell lines in vitro. In vivo, both were able to significantly knockdown HIF-1α and some of its downstream genes. Conclusions: The results suggest that HMW and LMW CS-based NPs exhibit distinct characteristics, showing that both MWs have potential for targeted pancreatic cancer therapy by influencing different aspects of delivery and gene silencing, particularly in the hypoxic tumor microenvironment.

Liquid Biopsy in Cancer Diagnosis-Treatment Methodology and Applications

The comprehensive review of liquid biopsy as a cutting-edge, non-invasive technique for cancer detection, diagnosis, and monitoring may be found in the document “Liquid Biopsy in Cancer Diagnosis – Treatment Methodology and Applications.” Via liquid biopsy, tumor-related indicators present in physiological fluids such as blood, urine, and saliva are examined, including exosomes, circulating tumor cells (CTCs), and circulating tumor DNA (ctDNA). Liquid biopsy enables regular, real-time sampling of tumor dynamics, in contrast to standard tissue biopsies, which are invasive and restricted to particular tumor regions. This renders it a potent instrument for the identification of cancer mutations and the tracking of tumor progression, particularly in cases of metastatic or elusive cancers. Technological developments like digital droplet PCR (ddPCR) and next-generation sequencing (NGS) have greatly improved the sensitivity and precision of ctDNA detection, making it possible to identify unusualexceptionally rare genetic mutations. This enhances the efficacy of tailored cancer treatment and monitoring. The genetic heterogeneity of cancers is captured by liquid biopsies, providing information on several tumor subclones that may be missed by traditional biopsies. The document also describes the use of different body fluids for different forms of cancer, such as head and neck and urological tumors, such as saliva, urine, and blood. Blood is the most often utilized medium for tracking genetic changes due to its ease of accessibility. Saliva and urine are useful in the detection of head and neck malignancies and bladder cancer, respectively. Liquid biopsy has several advantages, but it also has drawbacks, such as the requirement for standardized procedures, high prices, and differences in sensitivity and specificity between platforms.In order to overcome these obstacles and integrate liquid biopsy into standard clinical practice, the paper emphasizes the significance of more research and development. This will improve patient outcomes, early cancer detection, and personalized treatment.

The spatial and cellular portrait of transposable element expression during gastric cancer.

Gastric Cancer (GC) is a lethal malignancy, with urgent need for the discovery of novel biomarkers for its early detection. I previously showed that Transposable Elements (TEs) become activated in early GC (EGC), suggesting a role in gene expression. Here, I follow-up on that evidence using single-cell data from gastritis to EGC, and show that TEs are expressed and follow the disease progression, with 2,430 of them being cell populations markers. Pseudotemporal trajectory modeling revealed 111 TEs associated with the origination of cancer cells. Analysis of spatial data from GC also confirms TE expression, with 204 TEs being spatially enriched in the tumor regions and the tumor microenvironment, hinting at a role of TEs in tumorigenesis. Finally, a network of TE-mediated gene regulation was modeled, indicating that ~ 2,000 genes could be modulated by TEs, with ~ 500 of them already implicated in cancer. These results suggest that TEs might play a functional role in GC progression, and highlights them as potential biomarker for its early detection.

MALEXNET-EUNET: MULTI-CLASS LIVER CANCER CLASSIFICATION AND SEGMENTATION USING A HYBRID DEEP LEARNING SYSTEM

Background and objective: One of the most prevalent and significant causes of cancer-related mortality worldwide is considered to be liver cancer. Techniques for automatically segmenting and classifying liver tumors are crucial for supporting medical professionals during the tumor diagnostic process. Classifying liver tumors is challenging due to noise, nonhomogeneity, and the significant appearance diversity observed in tumor tissue. Also, in recent years, most of the research has performed binary classification, but there is still a lack of research on multi-class liver cancer classification. Therefore, we perform a multi-class liver cancer classification and segmentation in this research. Methods: We propose a hybrid deep learning-based multi-class liver cancer classification and segmentation system in this research. The collected CT images are pre-processed in four stages: contrast enhancement, noise filtering, smoothing and sharpening, and liver region segmentation. Next, the binary, texture, histogram, and rotational, scalability, and translational (RST) features are extracted from the pre-processed images. Then, the average correlation (AC) and probability of error (POE) approaches are applied to selecting relevant features and excluding less significant features. After feature selection, a modified AlexNet (MAlexNet) model is used to classify the multi-class classification of liver tumors. Finally, the identified liver tumor regions are segmented using the enhanced U-Net (EUNet) model. Results: Identified liver tumors are accurately classified into three different categories using the proposed classification model: hemangioma (HEM), hepatocellular carcinoma (HCC), and metastatic carcinoma (MET), with an average accuracy of 99.19%. We have collected the multi-class liver cancer CT images from real time patients. The results of the experiments show that proposed hybrid system provides adequate overall accuracy, is less noise-sensitive, and outperforms other state-of-the-art techniques such as SVM, Faster RCNN, Mask RCNN, and SAR-U-Net on a wide range of CT images. Conclusion: Radiologists and doctors can identify liver tumors more accurately using the suggested innovative framework.

CDSG-SAM: A cross-domain self-generating prompt few-shot brain tumor segmentation pipeline based on SAM

In clinical practice, accurate segmentation of brain tumor regions is essential for patient treatment and survival. Accurate brain tumor segmentation is an important task and it is based on a large amount of labeled data. Clinical data are from different cities and device protocols, which leads to the poor generalization ability of the model. In addition, the small amount of clinical data also leads to the decline of the segmentation performance. In this paper, a cross-domain self-generating prompt few-shot brain tumor segmentation pipeline called CDSG-SAM is developed. This pipeline is based on the Segment Anything Model (SAM), which integrates a new Cross-domain self-attention (CDS) Adapter module and Self-Generating (SG) Prompt module. The SAM-based segmentation model is more suitable for few-shot medical image segmentation. A dynamic fuzzy support mask decoder module (DFSMD) is proposed in the decoder stage to enhance the accuracy of brain tumor segmentation edges. Furthermore, a Joint Loss function (JL) is designed to comprehensively consider multiple aspects of performance indicators to improve the accuracy and robustness of the model. Experimental results evaluated on the BraTS2021 glioma dataset, Clinical glioma brain tumor data (CBTD), and BraTS2023 metastatic tumor dataset show that the proposed pipeline has achieved excellent performance with an average dice coefficient of 0.918, 0.828, and 0.868 which outperforms than existing methods.

Excessive autophagy-inducing and highly penetrable biomineralized bacteria for multimodal imaging-guided and mild hyperthermia-enhanced immunogenic cell death

The tumor microenvironment, characterized by hypoxia, supports the efficacy of anaerobic bacteria like attenuated S. typhimurium in cancer therapies. These bacteria target and penetrate deep tumor regions, significantly reducing tumor size but often lead to tumor regrowth due to limited long-term efficacy. To enhance the therapeutic impact, a novel biohybrid system, S@UIL, has been developed by coating S. typhimurium with a zirconium-based nanoscale metal–organic framework (UiO-66-NH2) loaded with indocyanine green (ICG) and luteolin (LUT). This system maintains the bacteria’s tumor-targeting ability while increasing the penetration and therapeutic effectiveness through excessive autophagy and mild hyperthermia. In a subcutaneous colon cancer model, the integration of LUT and ICG promotes autophagic cell death and photothermal sensitization, leading to the release of damage-associated molecular patterns (DAMPs). These DAMPs activate immune responses through dendritic cells and T-cells, enhancing immunogenic cell death (ICD) and potentially reducing immune evasion by tumors. This single-administration approach also integrates multimodal imaging capabilities, providing a promising strategy for improved tumor ICD induction and cancer progression inhibition.

Sorafenib plus memory-like natural killer cell immunochemotherapy boosts treatment response in liver cancer

BackgroundHeterogeneity of hepatocellular carcinoma (HCC) presents significant challenges for therapeutic strategies and necessitates combinatorial treatment approaches to counteract suppressive behavior of tumor microenvironment and achieve improved outcomes. Here, we employed cytokines to induce memory-like behavior in natural killer (NK) cells, thereby enhancing their cytotoxicity against HCC. Additionally, we evaluated the potential benefits of combining sorafenib with this newly developed memory-like NK cell (pNK) immunochemotherapy in a preclinical model.MethodsHCC tumors were grown in SD rats using subcapsular implantation. Interleukin 12/18 cytokines were supplemented to NK cells to enhance cytotoxicity through memory activation. Tumors were diagnosed using MRI, and animals were randomly assigned to control, pNK immunotherapy, sorafenib chemotherapy, or combination therapy groups. NK cells were delivered locally via the gastrointestinal tract, while sorafenib was administered systemically. Therapeutic responses were monitored with weekly multi-parametric MRI scans over three weeks. Afterward, tumor tissues were harvested for histopathological analysis. Structural and functional changes in tumors were evaluated by analyzing MRI and histopathology data using ANOVA and pairwise T-test analyses.ResultsThe tumors were allowed to grow for six days post-cell implantation before treatment commenced. At baseline, tumor diameter averaged 5.27 mm without significant difference between groups (p = 0.16). Both sorafenib and combination therapy imposed greater burden on tumor dimensions compared to immunotherapy alone in the first week. By the second week of treatment, combination therapy had markedly expanded its therapeutic efficacy, resulting in the most significant tumor regression observed (6.05 ± 1.99 vs. 13.99 ± 8.01 mm). Histological analysis demonstrated significantly improved cell destruction in the tumor microenvironment associated with combination treatment (63.79%). Interestingly, we observed fewer viable tumor regions in the sorafenib group (38.9%) compared to the immunotherapy group (45.6%). Notably, there was a significantly higher presence of NK cells in the tumor microenvironment with combination therapy (34.79%) compared to other groups (ranging from 2.21 to 26.50%). Although the tumor sizes in the monotherapy groups were similar, histological analysis revealed a stronger response in pNK cell immunotherapy group compared to the sorafenib group.ConclusionsExperimental results indicated that combination therapy significantly enhanced treatment response, resulting in substantial tumor growth reduction in alignment with histological analysis.

HIPI: Spatially resolved multiplexed protein expression inferred from H&E WSIs.

Solid tumors are characterized by complex interactions between the tumor, the immune system and the microenvironment. These interactions and intra-tumor variations have both diagnostic and prognostic significance and implications. However, quantifying the underlying processes in patient samples requires expensive and complicated molecular experiments. In contrast, H&E staining is typically performed as part of the routine standard process, and is very cheap. Here we present HIPI (H&E Image Interpretation and Protein Expression Inference) for predicting cell marker expression from tumor H&E images. We process paired H&E and CyCIF images taken from serial sections of colorectal cancers to train our model. We show that our model accurately predicts the spatial distribution of several important cell markers, on both held-out tumor regions as well as new tumor samples taken from different patients. Moreover, using only the tissue image morphology, HIPI is able to colocalize the interactions between different cell types, further demonstrating its potential clinical significance.

Hypoxia induces ROS-resistant memory upon reoxygenation in vivo promoting metastasis in part via MUC1-C

Hypoxia occurs in 90% of solid tumors and is associated with metastasis and mortality. Breast cancer cells that experience intratumoral hypoxia are 5x more likely to develop lung metastasis in animal models. Using spatial transcriptomics, we determine that hypoxic cells localized in more oxygenated tumor regions (termed ‘post-hypoxic’) retain expression of hypoxia-inducible and NF-kB-regulated genes, even in the oxygen-rich bloodstream. This cellular response is reproduced in vitro under chronic hypoxic conditions followed by reoxygenation. A subset of genes remains increased in reoxygenated cells. MUC1/MUC1-C is upregulated by both HIF-1α and NF-kB-p65 during chronic hypoxia. Abrogating MUC1 decreases the expression of superoxide dismutase enzymes, causing reactive oxygen species (ROS) production and cell death. A hypoxia-dependent genetic deletion of MUC1, or MUC1-C inhibition by GO-203, increases ROS levels in circulating tumor cells (CTCs), reducing the extent of metastasis. High MUC1 expression in tumor biopsies is associated with recurrence, and MUC1+ CTCs have lower ROS levels than MUC1- CTCs in patient-derived xenograft models. This study demonstrates that therapeutically targeting MUC1-C reduces hypoxia-driven metastasis.

Role of diffusion-weighted imaging in the diagnosis of pituitary region tumors.

This study aimed to assess the role of Diffusion-Weighted Imaging (DWI) in routine pituitary Magnetic Resonance Imaging (MRI) protocols for distinguishing sellar and parasellar tumors, addressing the lack of clear guidelines in contemporary literature. A retrospective analysis of 242 pituitary MRI scans with DWI sequences was conducted in a single-center study using a 1.5T scanner and standard DWI sequence parameters. Measurements of both absolute and relative mean apparent diffusion coefficient (ADC) values, along with minimal ADC values within tumors, were performed. The adopted region of interest (ROI) based method used for these measurements was validated. Invasive pituitary adenomas exhibited significantly lower min ADC and min rADC than meningiomas, with optimal cut-off points of 0.64 (sensitivity 73%, specificity 82%) and 0.78 (sensitivity 73%, specificity 89%), respectively. Post-hemorrhagic pituitary adenomas demonstrated lower ADC values than adamantinomatous craniopharyngiomas, with an AUC of 0.893 for min rADC = 1.07, and Rathke's Cleft Cysts with mucous content, AUC 0.8 for min rADC = 1.01. Specific differentiation with high sensitivity and specificity based on diffusion parameters was observed for these tumor groups. Cystic pituitary non-functional adenomas obtained significantly lower ADC values compared to the adamantinomatous type of craniopharyngiomas and serous Rathke's Cleft Cysts (AUC up to 0.942). The study concludes that integrating DWI into routine pituitary MRI protocols enhances diagnostic accuracy in distinguishing sellar and parasellar tumors. The short scan time of one minute makes DWI a valuable and precise tool, supporting its recommendation as a standard component of pituitary MRI examinations.

Valence-Transforming O2-Depleting Nano-Assembly Enable In Situ Tumor Depositional Embolization.

Abnormal metabolism and blood supply/O2 imbalance in tumor cells affect drug transport delivery and increase the difficulty of tumor treatment. Controlling tumor growth by inhibiting tumor cell metabolism and regulating progressive embolization in the tumor region provides an innovative basis for constructing tumor therapeutic models. A highly biocompatible and efficient O2-depleting agent has been investigated to enable in situ precipitation and embolization within the tumor microenvironment. In situ deformation embolizer, Fe-GA@CaCO3 nano-assembly (GA: gallic acid), can convert into the large granular size embolization components of Fe(III) precipitates and affluent Ca2+ within the tumor microenvironment. In situ progressive O2 depletion produces Fe(III) precipitates that embolize tumor regions, isolating O2 and nutrients by blocking supply. Meanwhile, affluent Ca2+ acts on the intracellular, causing mitochondrial dysfunction through calcium overload and contributing to irreversible tumor cell damage. Both internal and external routes work synergistically to produce precise functional inhibition of tumors from the inside out, simultaneously responding to both intracellular and the corresponding tumor regions, providing an innovative solution for anti-tumor therapy.

Nanoparticle Delivery of an Oligonucleotide Payload in a Glioblastoma Multiforme Animal Model.

Glioblastoma multiforme (GBM) is the most common and aggressive form of primary brain malignancy for which there is no cure. The blood-brain barrier is a significant hurdle in the delivery of therapies to GBM. Reported here is an image-guided, iron oxide-based therapeutic delivery nano platform capable of bypassing this physiological barrier by virtue of size and accumulating in the tumor region, delivering its payload. This 25 nm nano platform consists of crosslinked dextran-coated iron oxide nanoparticles labeled with Cy5.5 fluorescent dye and containing antisense oligonucleotide as a payload. The magnetic iron oxide core enables tracking of the nanoparticles through in vivo magnetic resonance imaging, while Cy5.5 dye allows tracking by optical imaging. This report details the monitoring of the accumulation of this nanoparticle platform (termed MN-anti-miR10b) in orthotopically implanted glioblastoma tumors following intravenous injection. In addition, it provides insight into the in vivo delivery of RNA oligonucleotides, a problem that has hampered the translation of RNA therapeutics into the clinic.

Lymphocyte Function in Tertiary Lymphoid Structures Predicts Hepatocellular Carcinoma Outcome

An increasing number of studies have revealed a correlation between tertiary lymphoid structures (TLSs) and the outcome of hepatocellular carcinoma (HCC). Nevertheless, the associations between the heterogeneity of cellular composition and the overall survival (OS) in HCC remain unexplored. Here, we evaluated the cancer tissues from 150 HCC individuals treated at the Tumor Hospital Affiliated with Nantong University using multiplex immunofluorescence to determine the presence and characteristics of TLS and to investigate the relationship between intra-TLS immunologic activity, TLS maturation, and intratumoral immune cell infiltration. Prognostic factors influencing the outcome were identified through both univariate and multivariate analyses. Additionally, the levels of cytotoxic T-lymphocyte antigen-4 (CTLA-4), programmed death 1, programmed death-ligand 1, and lymphocyte activation gene-3 were determined, as well as their relationship with TLS features were determined. TLS was detected in 71 (47.3%) of the 150 HCC cases and was related to higher intratumoral infiltration levels of lymphocytes. Additionally, intra-TLS lymphocyte proliferation correlated with that of intratumoral lymphocytes, and the presence of TLS and a high proportion of mature TLS demonstrated a significant correlation with better prognosis (P = .013 and P = .03, respectively). Among TLS-positive tumors, a high proportion of B cells expressing activation-induced cytidine deaminase and a high proportion of CD8+ T cells expressing CD45RO were significantly related to improved OS (P = .01 and P < .001, respectively). Comparatively, a high proportion of CD21+CD20+ B cells demonstrated a significant correlation with poorer OS (P < .002). A markedly reduced number of CTLA-4+ cells in the stromal regions in TLS-negative tumors was observed compared with TLS-positive tumors (P = .01). These findings reveal a correlation between TLS presence and improved OS in HCC patients. However, TLS exhibited significant variation in maturation state, T- and B-cell proliferation, and expression of markers related to B- and T-cell function. Notably, these characteristics were also found to possess prognostic significance, indicating that certain TLS might hinder tumor immunity by inhibiting immune cells, whereas others may foster antigen-driven immune responses, likely influenced by the composition and functional status of intra-TLS lymphocytes.

An "AND" logic gate-based supramolecular therapeutic nanoplatform for combatting drug-resistant non-small cell lung cancer.

Despite targeted therapies like epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs), non-small cell lung cancer (NSCLC) remains a clinical challenge due to drug resistance hampering their efficacy. Here, we designed an "AND" logic gate-based supramolecular therapeutic platform (HA-BPY-GEF-NPs) for the treatment of EGFR-TKI resistant NSCLC. This system integrates both internal and external stimuli-responsive mechanisms that need to be activated in a preset sequence, enabling it to precisely control drug release behavior for enhancing therapeutic precision. By programming the system to respond to sequential near-infrared (NIR) irradiation and enzyme (cathepsin B) inputs, the release of gefitinib is effectively confined to the tumor region. Moreover, the NIR irradiation induces reactive oxygen species production, suppressing tumor growth and inhibiting bypass signaling pathways. The designed drug delivery system offers a highly controlled and targeted therapeutic approach, effectively inhibiting tumor growth, suppressing bypass signaling pathways, and overcoming EGFR-TKI resistance, thus offering a potential solution for maximizing therapeutic benefits.

Comparison of the accumulation manner of a macromolecular drug between two mouse tumour models: study with magnetic resonance imaging and the model macromolecular drug, gadolinium-conjugated dextran

A knowledge of the difference of spatio-temporal behaviour of nanomedicine in different type of tumour models is important to develop well-targeted nanomedicine for tumour. In this study, intratumoral accumulation of the model nanomedicine, gadolinium-conjugated dextran (Gd-Dex), was examined with magnetic resonance imaging in two tumour models; mouse sarcoma S180 and radiation-induced mouse fibrosarcoma RIF-1. From time-course of the distribution images, the plasma-to-tumour interstitial tissue transfer constant (Ktrans ) and fractional plasma volume (Vp ) were calculated and mapped. Gd-Dex preferentially distributed to the marginal region of S180 tumours immediately after its injection, and then started to accumulate in some parts of the central region. Ktrans and Vp values were large in the marginal region, while only Ktrans was large in some parts of the central region. In contrast, the distribution of Gd-Dex in RIF-1 tumours was fairly homogeneous, and may have resulted from the homogeneous distributions of Ktrans and Vp . The amounts of Gd-Dex that accumulated in entire tumours in both tumour models correlated with the volume of tumours; however, accumulation in large S180 tumours deviated from the correlation in the early phase. The differences in the manner and pharmacokinetics of nanomedicine among tumour models may affect the accumulation of the medicine.

Development of a Specifically Labeled 89Zr Antibody for the Noninvasive Imaging of Tumors Overexpressing B7-H3.

B7-H3 has emerged as a promising target and potential biomarker for diagnosing tumors, evaluating treatment efficacy, and determining patient prognosis. Hu4G4 is a recombinant humanized antibody that selectively targets the extracellular domain of human B7-H3. In this study, we describe the radiolabeling of hu4G4 with the positron emission tomography (PET) emitter radionuclide zirconium 89 (89Zr) and evaluate its potency as an immuno-PET tracer for B7-H3-targeted imaging by comparing it in vitro and in vivo to [89Zr]Zr-DFO-DS-5573a using various models. The radiolabeled compound, [89Zr]Zr-desferrioxamine-hu4G4 ([89Zr]Zr-DFO-hu4G4), demonstrated a high radiochemical purity (RCP) of greater than 99% and a specific activity of 74 MBq/mg following purification. Additionally, it maintained stability in human serum albumin (HSA) and acetate buffer, preserving over 90% of its RCP after 7 days. Three cell lines targeting human B7-H3(U87/CT26-CD276/GL261-CD276) were used. Flow cytometry analysis indicated that the B7-H3-positive cells (U87/CT26-CD276/GL261-CD276) had a higher B7-H3 protein level with no expression in the B7-H3-negative cells (CT26-wt/GL261-wt) (P < 0.001). Moreover, the cellular uptake was 45.71 ± 3.78% for [89Zr]Zr-DFO-hu4G4 in CT26-CD276 cells versus only 0.93 ± 0.47% in CT26-wt cells and 30.26 ± 0.70% when [89Zr]Zr-DFO-hu4G4 in CT26-CD276 cells were blocked with 100× 8H9. The cellular uptake of [89Zr]Zr-DFO-hu4G4 was akin to that observed with [89Zr]Zr-DFO-DS-5573a with no significant differences (45.71 ± 3.78 % vs 47.07 ± 0.86 %) in CT26-CD276 cells. Similarly, the CT26-CD276 mouse model demonstrated markedly low organ uptake and elevated tumor uptake 48 h after [89Zr]Zr-DFO-hu4G4 injection. PET/CT analysis showed that the tumor-to-muscle (T/M) ratios were substantially higher compared to other imaging groups: 27.65 ± 3.17 in CT26-CD276 mice versus 11.68 ± 4.19 in CT26-wt mice (P < 0.001) and 16.40 ± 0.78 when 100× 8H9 was used to block [89Zr]Zr-DFO-hu4G4 in CT26-CD276 mice (P < 0.01) at 48 h post-injection. Additionally, the tracer showed markedly high accumulation in the tumor region (22.57 ± 3.03% ID/g), comparable to the uptake of [89Zr]Zr-DFO-DS-5573a (24.76 ± 5.36% ID/g). A dosimetry estimation study revealed that the effective dose for [89Zr]Zr-DFO-hu4G4 was 2.96 × 10-01 mSv/MBq, which falls within the acceptable range for further research in nuclear medicine. Collectively, these results indicated that [89Zr]Zr-DFO-hu4G4 was successfully fabricated and applied in B7-H3-targeted tumor PET/CT imaging, which showed excellent imaging quality and tumor detection efficacy in tumor-bearing mice. It is a promising imaging agent for identifying tumors that overexpress B7-H3 for future clinical applications.

Deep bone oncology Diagnostics: Computed tomography based Machine learning for detection of bone tumors from breast cancer metastasis

PurposeThe objective of this study is to develop a novel diagnostic tool using deep learning and radiomics to distinguish bone tumors on CT images as metastases from breast cancer. By providing a more accurate and reliable method for identifying metastatic bone tumors, this approach aims to significantly improve clinical decision-making and patient management in the context of breast cancer. MethodsThis study utilized CT images of bone tumors from 178 patients, including 78 cases of breast cancer bone metastases and 100 cases of non-breast cancer bone metastases. The dataset was processed using the Medical Image Segmentation via Self-distilling TransUNet (MISSU) model for automated segmentation. Radiomics features were extracted from the segmented tumor regions using the Pyradiomics library, capturing various aspects of tumor phenotype. Feature selection was conducted using LASSO regression to identify the most predictive features. The model’s performance was evaluated using ten-fold cross-validation, with metrics including accuracy, sensitivity, specificity, and the Dice similarity coefficient. ResultsThe developed radiomics model using the SVM algorithm achieved high discriminatory power, with an AUC of 0.936 on the training set and 0.953 on the test set. The model’s performance metrics demonstrated strong accuracy, sensitivity, and specificity. Specifically, the accuracy was 0.864 for the training set and 0.853 for the test set. Sensitivity values were 0.838 and 0.789 for the training and test sets, respectively, while specificity values were 0.896 and 0.933 for the training and test sets, respectively. These results indicate that the SVM model effectively distinguishes between bone metastases originating from breast cancer and other origins. Additionally, the average Dice similarity coefficient for the automated segmentation was 0.915, demonstrating a high level of agreement with manual segmentations. ConclusionThis study demonstrates the potential of combining CT-based radiomics and deep learning for the accurate detection of bone metastases from breast cancer. The high-performance metrics indicate that this approach can significantly enhance diagnostic accuracy, aiding in early detection and improving patient outcomes. Future research should focus on validating these findings on larger datasets, integrating the model into clinical workflows, and exploring its use in personalized treatment planning.

Statistical analysis of multiple regions-of-interest in multiplexed spatial proteomics data.

Multiplexed spatial proteomics reveals the spatial organization of cells in tumors, which is associated with important clinical outcomes such as survival and treatment response. This spatial organization is often summarized using spatial summary statistics, including Ripley's K and Besag's L. However, if multiple regions of the same tumor are imaged, it is unclear how to synthesize the relationship with a single patient-level endpoint. We evaluate extant approaches for accommodating multiple images within the context of associating summary statistics with outcomes. First, we consider averaging-based approaches wherein multiple summaries for a single sample are combined in a weighted mean. We then propose a novel class of ensemble testing approaches in which we simulate random weights used to aggregate summaries, test for an association with outcomes, and combine the $P$-values. We systematically evaluate the performance of these approaches via simulation and application to data from non-small cell lung cancer, colorectal cancer, and triple negative breast cancer. We find that the optimal strategy varies, but a simple weighted average of the summary statistics based on the number of cells in each image often offers the highest power and controls type I error effectively. When the size of the imaged regions varies, incorporating this variation into the weighted aggregation may yield additional power in cases where the varying size is informative. Ensemble testing (but not resampling) offered high power and type I error control across conditions in our simulated data sets.