Intratumoral Regions Research Articles

A radiomics strategy based on CT intra-tumoral and peritumoral regions for preoperative prediction of neoadjuvant chemoradiotherapy for esophageal cancer

ObjectiveDevelop a method for selecting esophageal cancer patients achieving pathological complete response with pre-neoadjuvant therapy chest-enhanced CT scans. MethodsTwo hundred and one patients from center 1 were enrolled, split into training and testing sets (7:3 ratio), with an external validation set of 30 patients from center 2. Radiomics features from intra-tumoral and peritumoral images were extracted and dimensionally reduced using Student's t-test and least absolute shrinkage and selection operator. Four machine learning classifiers were employed to build models, with the best-performing models selected based on accuracy and stability. ROC curves were utilized to determine the top prediction model, and its generalizability was evaluated on the external validation set. ResultsAmong 16 models, the integrated-XGBoost and integrated-random forest models performed the best, with average ROC AUCs of 0.906 and 0.918, respectively, and RSDs of 6.26 and 6.89 in the training set. In the testing set, AUCs were 0.845 and 0.871, showing no significant difference in ROC curves. External validation set AUCs for integrated-XGBoost and integrated-random forest models were 0.650 and 0.749. ConclusionIncorporating peritumoral radiomics features into the analysis enhances predictive performance for esophageal cancer patients undergoing neoadjuvant chemoradiotherapy, paving the way for improved treatment outcomes.

Constructing the Optimal Classification Model for Benign and Malignant Breast Tumors Based on Multifeature Analysis from Multimodal Images.

The purpose of this study was to fuse conventional radiomic and deep features from digital breast tomosynthesis craniocaudal projection (DBT-CC) and ultrasound (US) images to establish a multimodal benign-malignant classification model and evaluate its clinical value. Data were obtained from a total of 487 patients at three centers, each of whom underwent DBT-CC and US examinations. A total of 322 patients from dataset 1 were used to construct the model, while 165 patients from datasets 2 and 3 formed the prospective testing cohort. Two radiologists with 10-20years of work experience and three sonographers with 12-20years of work experience semiautomatically segmented the lesions using ITK-SNAP software while considering the surrounding tissue. For the experiments, we extracted conventional radiomic and deep features from tumors from DBT-CCs and US images using PyRadiomics and Inception-v3. Additionally, we extracted conventional radiomic features from four peritumoral layers around the tumors via DBT-CC and US images. Features were fused separately from the intratumoral and peritumoral regions. For the models, we tested the SVM, KNN, decision tree, RF, XGBoost, and LightGBM classifiers. Early fusion and late fusion (ensemble and stacking) strategies were employed for feature fusion. Using the SVM classifier, stacking fusion of deep features and three peritumoral radiomic features from tumors in DBT-CC and US images achieved the optimal performance, with an accuracy and AUC of 0.953 and 0.959 [CI: 0.886-0.996], a sensitivity and specificity of 0.952 [CI: 0.888-0.992] and 0.955 [0.868-0.985], and a precision of 0.976. The experimental results indicate that the fusion model of deep features and peritumoral radiomic features from tumors in DBT-CC and US images shows promise in differentiating benign and malignant breast tumors.

Removal of schwannoma from the psoas muscle with intraoperative neurophysiological monitoring: A case report.

The incidence of a schwannoma within the psoas muscle is rare, and only a few cases have been reported. The surgical approach to removing schwannomas present in the psoas muscle is challenging because of its anatomical proximity to the lumbar plexus. A 31-year-old man experienced right lower back pain and anterolateral thigh numbness for 2 months. Magnetic resonance imaging of the patient's lumbar spine revealed a mass lesion, which was radiologically diagnosed as a well-demarcated schwannoma. The patient underwent surgery for excision of the schwannoma in the right psoas muscle at the second to fourth lumbar vertebrae levels. During surgery, intraoperative neurophysiological monitoring modalities, free-running and triggered electromyography and evoked potentials, from the target muscles were recorded. There was no neurotonic discharge corresponding to neuronal injury. Compound motor nerve action potential was detected in the triggered electromyography of muscles around the medial margin of the tumor. However, direct integration of the motor nerve was not observed in the intra-tumor region. We report that schwannoma removal in the psoas muscle, which is adjacent to the lumbar plexus, can be safely performed using intraoperative neurophysiological monitoring.

Intratumoral and peritumoral CT-based radiomics for predicting the microsatellite instability in gastric cancer.

To investigate the value of intratumoral and peritumoral radiomics based on contrast-enhanced computer tomography (CECT) to preoperatively predict microsatellite instability (MSI) status in gastric cancer (GC) patients. A total of 189 GC patients, including 63 patients with MSI-high (MSI-H) and 126 patients with MSI-low/stable (MSI-L/S), were randomly divided into the training cohort and validation cohort. Intratumoral and 5-mm peritumoral regions' radiomics features were extracted from CECT images. The features were standardized by Z-score, and the Inter-andintraclass correlation coefficient, univariate logistic regression analysis, and least absolute shrinkage and selection operator (LASSO) were applied to select the optimal radiomics features. Radiomics scores (Rad-score) based on intratumoral regions, peritumoral regions, and intratumoral + 5-mm peritumoral regions were calculated by weighting the linear combination of the selected features with their respective coefficients to construct the intratumoral model, peritumoral model, and intratumoral + peritumoral model. Logistic regression was used to establish a combined model by combining clinical characteristics, CT semantic features, and Rad-score of intratumoral and peritumoral regions. Eleven radiomics features were selected to establish a radiomics intratumoral + peritumoral model. CT-measured tumor length and tumor location were independent risk factors for MSI status. The established combined model obtained the highest area under the receiver operating characteristic (ROC) curve (AUC) of 0.830 (95% CI, 0.727-0.906) in the validation cohort. The calibration curve and decision curve demonstrated its good model fitness and clinical application value. The combined model based on intratumoral and peritumoral CECT radiomics features and clinical factors can predict the MSI status of GS with moderate accuracy before surgery, which helps formulate personalized treatment strategies.

Ultrasound-based radiomics nomogram for predicting axillary lymph node metastasis in early-stage breast cancer.

We aimed at assessing the predictive ability of ultrasound-based radiomics combined with clinical characteristics for axillary lymph node (ALN) status in early-stage breast cancer patients and to compare performance in different peritumoral regions. A total of 755 patients (527 in the primary cohort and 228 in the external validation cohort) were enrolled in this study. Ultrasound images for all patients were acquired and radiomics analysis performed for intratumoral and different peritumoral regions. The MRMR and LASSO regression analyses were performed on extracted features from the primary cohort to construct a radiomics signature formula combined with clinical characteristics. Pearson's coefficient and the variance inflation factor (VIF) were performed to check the correlation and the multicollinearity among the final predictors. The best performing model was selected to develop a nomogram, which was established by performing binary logistic regression and acquiring cut-off values based on the corresponding nomogram scores of the masses. Among all the radiomics models, the "Mass + Margin3mm" model exhibited the best performance. The areas under the curves (AUC) of the nomogram in the primary and external validation cohorts were 0.906 (95% confidence intervals [CI] 0.882-0.930) and 0.922 (95% CI 0.894-0.960), respectively. They both showed good calibrations. The nomogram exhibited a good ability to discriminate between positive and negative lymph nodes (AUC: 0.853 (95% CI 0.816-0.889) in primary cohort, 0.870 (95% CI 0.818-0.922) in validation cohort), and between low-volume and high-volume lymph nodes (AUC: 0.832 (95% CI 0.781-0.884) in primary cohort, 0.911 (95% CI 0.858-0.964) in validation cohort). The established nomogram is a prospective clinical prediction tool for non-invasive assessment of ALN status. It has the ability to enhance the accuracy of early-stage breast cancer treatment.

Enhancing Ki-67 Prediction in Breast Cancer: Integrating Intratumoral and Peritumoral Radiomics From Automated Breast Ultrasound via Machine Learning

Rationale and ObjectivesTraditional Ki-67 evaluation in breast cancer (BC) via core needle biopsy is limited by repeatability and heterogeneity. The automated breast ultrasound system (ABUS) offers reproducibility but is constrained to morphological and echoic assessments. Radiomics and machine learning (ML) offer solutions, but their integration for improving Ki-67 predictive accuracy in BC remains unexplored. This study aims to enhance ABUS by integrating ML-assisted radiomics for Ki-67 prediction in BC, with a focus on both intratumoral and peritumoral regions. Materials and MethodsA retrospective analysis was conducted on 936 BC patients, split into training (n=655) and testing (n=281) cohorts. Radiomics features were extracted from intra- and peritumoral regions via ABUS. Feature selection involved Z-score normalization, intraclass correlation, Wilcoxon rank sum tests, minimum redundancy maximum relevance, and least absolute shrinkage and selection operator logistic regression. ML classifiers were trained and optimized for enhanced predictive accuracy. The interpretability of the optimized model was further augmented by employing Shapley additive explanations (SHAP). ResultsOf the 2632 radiomics features in each patient, 15 were significantly associated with Ki-67 levels. The support vector machine (SVM) was identified as the optimal classifier, with area under the receiver operating characteristic curve values of 0.868 (training) and 0.822 (testing). SHAP analysis indicated that five peritumoral and two intratumoral features, along with age and lymph node status, were key determinants in the predictive model. ConclusionIntegrating ML with ABUS-based radiomics effectively enhances Ki-67 prediction in BC, demonstrating the SVM model's strong performance with both radiomics and clinical factors.

Radiomic Machine Learning in Invasive Ductal Breast Cancer: Prediction of Ki-67 Expression Level Based on Radiomics of DCE-MRI.

Our study aimed to investigate the potential of radiomics with DCE-MRI for predicting Ki-67 expression in invasive ductal breast cancer. We conducted a retrospective study including 223 patients diagnosed with invasive ductal breast cancer. Radiomics features were extracted from DCE-MRI using 3D-Slicer software. Two Ki-67 expression cutoff values (20% and 29%) were examined. Patients were divided into training (70%) and test (30%) sets. The Elastic Net method selected relevant features, and five machine-learning models were established. Radiomics models were created from intratumoral, peritumoral, and combined regions. Performance was assessed using ROC curves, accuracy, sensitivity, and specificity. For a Ki-67 cutoff value of 20%, the combined model exhibited the highest performance, with area under the curve (AUC) values of 0.838 (95% confidence interval (CI): 0.774-0.897) for the training set and 0.863 (95% CI: 0.764-0.949) for the test set. The AUC values for the tumor model were 0.816 (95% CI: 0.745-0.880) and 0.830 (95% CI: 0.724-0.916), and for the peritumor model were 0.790 (95% CI: 0.711-0.857) and 0.808 (95% CI: 0.682-0.910). When the Ki-67 cutoff value was set at 29%, the combined model also demonstrated superior predictive ability in both training set (AUC: 0.796; 95% CI: 0.724-0.862) and the test set (AUC: 0.823; 95% CI: 0.723-0.911). The AUC values for the tumor model were 0.785 (95% CI: 0.708-0.861) and 0.784 (95% CI: 0.663-0.882), and for the peritumor model were 0.773 (95% CI: 0.690-0.844) and 0.729 (95% CI: 0.603-0.847). Radiomics with DCE-MRI can predict Ki-67 expression in invasive ductal breast cancer. Integrating radiomics features from intratumoral and peritumoral regions yields a dependable prognostic model, facilitating pre-surgical detection and treatment decisions. This holds potential for commercial diagnostic tools.

Factors affecting heterogeneity in breast cancer microenvironment: A narrative mini review.

Breast cancer (BC) heterogeneity is a key trait of BC tumors with crucial implications on tumorigenesis, diagnosis, and therapeutic modalities. It is influenced by tumor intrinsic features and by the tumor microenvironment (TME) composition of different intra-tumoral regions, which in turn affect cancer progression within patients. In this mini review, we will highlight the mechanisms that generate cancer heterogeneity in BC and how they affect the responses to cancer therapies.

Morpho-phenotypic characterization of melanoma brain metastases immune microenvironment: A multicentre retrospective study

IntroductionThe recent rise of immunotherapy for melanoma brain metastases (MBM) has enhanced the interest in characterizing the composition of local immune microenvironment. Despite the central nervous system (CNS) is well known for harbouring a peculiar immunological milieu, its role in MBM is only partially understood. The aim of our study was to characterize tumour-infiltrating lymphocytes (TILs) and tumour-associated macrophages (TAMs) distribution and density in a cohort of MBM and their potential relevance as prognostic biomarkers. Materials and methodsNinety-four MBM patients were retrospectively evaluated for morphological, molecular features and follow-up data. Formalin-fixed and paraffin-embedded (FFPE) tissue sections were immunostained with the following antibodies: CD4, CD8, FoxP3 CD68, CD163, PD-L1, HLA-ABC. Semiquantitative assessment of TILs and TAMs density and spatial distribution was performed and statistical analysis for prognostic correlations was carried out. ResultsThe distribution of CD4+ TILs was higher in the intratumoral region (p=0.012) while CD163+ TAMs in the peritumoral (p=0.016). An association was also found between tumour-associated astrogliosis (TAA) and CD163+ TAMs (p=0.021). Increased CD68+ TAMs correlated with BRAF V600 mutation (p=0.038). CD4+ TILs were significantly higher in PD-L1 <1 % group (p=0.030). CD68+ high expression group significantly correlated with PD-L1 ≥1 % (p=0.029). Median HLA-ABC H-score was significantly associated with intratumoral CD4+ (p=0.005) and peritumoral CD8+ infiltration (p=0.048). CD68+ TAMs intratumoral high expression showed a favorable association with OS in patients with multiple MBM (p=0.017), whereas a high HLA-ABC H-score with a prolonged OS in those with single MBM (p=0.007). ConclusionsOur findings suggest a peculiar topographical distribution of immune cells in both intratumoral and peritumoral areas, and correlations with both BRAF V600 mutation and PD-L1 expression in a large cohort of MBM. Moreover, CD68+ TAMs distribution and HLA-ABC favourably impacted prognosis in patients with multiple and single brain metastases, respectively.

A Multi-Task Transformer With Local-Global Feature Interaction and Multiple Tumoral Region Guidance for Breast Cancer Diagnosis.

Breast cancer, as a malignant tumor disease, has maintained high incidence and mortality rates over the years. Ultrasonography is one of the primary methods for diagnosing early-stage breast cancer. However, correctly interpreting breast ultrasound images requires massive time from physicians with specialized knowledge and extensive experience. Recently, deep learning-based method have made significant advancements in breast tumor segmentation and classification due to their powerful fitting capabilities. However, most existing methods focus on performing one of these tasks separately, and often failing to effectively leverage information from specific tumor-related areas that hold considerable diagnostic value. In this study, we propose a multi-task network with local-global feature interaction and multiple tumoral region guidance for breast ultrasound-based tumor segmentation and classification. Specifically, we construct a dual-stream encoder, paralleling CNN and Transformer, to facilitate hierarchical interaction and fusion of local and global features. This architecture enables each stream to capitalize on the strengths of the other while preserving its unique characteristics. Moreover, we design a multi-tumoral region guidance module to explicitly learn long-range non-local dependencies within intra-tumoral and peri-tumoral regions from spatial domain, thus providing interpretable cues beneficial for classification. Experimental results on two breast ultrasound datasets show that our network outperforms state-of-the-art methods in tumor segmentation and classification tasks. Compared with the second-best competitive method, our network improves the diagnosis accuracy from 73.64% to 80.21% on a large external validation dataset, which demonstrates its superior generalization capability.

Concomitant [18F]F-FAZA and [18F]F-FDG Imaging of Gynecological Cancer Xenografts: Insight into Tumor Hypoxia.

Herein we assessed the feasibility of imaging protocols using both hypoxia-specific [18F]F-FAZA and [18F]F-FDG in bypassing the limitations derived from the non-specific findings of [18F]F-FDG PET imaging of tumor-related hypoxia. CoCl2-generated hypoxia was induced in multidrug resistant (Pgp+) or sensitive (Pgp-) human ovarian (Pgp- A2780, Pgp+ A2780AD), and cervix carcinoma (Pgp- KB-3-1, Pgp+ KB-V-1) cell lines to establish corresponding tumor-bearing mouse models. Prior to [18F]F-FDG/[18F]F-FAZA-based MiniPET imaging, in vitro [18F]F-FDG uptake measurements and western blotting were used to verify the presence of hypoxia. Elevated GLUT-1, and hexokinase enzyme-II expression driven by CoCl2-induced activation of hypoxia-inducible factor-1α explains enhanced cellular [18F]F-FDG accumulation. No difference was observed in the [18F]F-FAZA accretion of Pgp+ and Pgp- tumors. Tumor-to-muscle ratios for [18F]F-FAZA measured at 110-120 min postinjection (6.2±0.1) provided the best contrasted images for the delineation of PET-oxic and PET-hypoxic intratumor regions. Although all tumors exhibited heterogenous uptake of both radiopharmaceuticals, greater differences for [18F]F-FAZA between the tracer avid and non-accumulating regions indicate its superiority over [18F]F-FDG. Spatial correlation between [18F]F-FGD and [18F]F-FAZA scans confirms that hypoxia mostly occurs in regions with highly active glucose metabolism. The addition of [18F]F-FAZA PET to [18F]F-FGD imaging may add clinical value in determining hypoxic sub-regions.

Radiotherapy dose escalation using pre-treatment diffusion-weighted imaging in locally advanced rectal cancer: a planning study.

Diffusion-weighted MRI (DWI) may provide biologically relevant target volumes for dose-escalated radiotherapy in locally advanced rectal cancer (LARC). This planning study assessed the dosimetric feasibility of delivering hypofractionated boost treatment to intra-tumoural regions of restricted diffusion prior to conventional long-course radiotherapy. Ten patients previously treated with curative-intent standard long-course radiotherapy (50 Gy/25#) were re-planned. Boost target volumes (BTVs) were delineated semi-automatically using 40th centile intra-tumoural apparent diffusion coefficient value with expansions (anteroposterior 11 mm, transverse 7 mm, craniocaudal 13 mm). Biased-dosed combined plans consisted of a single-fraction volumetric modulated arc therapy flattening-filter-free (VMAT-FFF) boost (phase 1) of 5, 7, or 10 Gy before long-course VMAT (phase 2). Phase 1 plans were assessed with reference to stereotactic conformality and deliverability measures. Combined plans were evaluated with reference to standard long-course therapy dose constraints. Phase 1 BTV dose targets at 5/7/10 Gy were met in all instances. Conformality constraints were met with only 1 minor violation at 5 and 7 Gy. All phase 1 and combined phase 1 + 2 plans passed patient-specific quality assurance. Combined phase 1 + 2 plans generally met organ-at-risk dose constraints. Exceptions included high-dose spillage to bladder and large bowel, predominantly in cases where previously administered, clinically acceptable non-boosted plans also could not meet constraints. Targeted upfront LARC radiotherapy dose escalation to DWI-defined is feasible with appropriate patient selection and preparation. This is the first study to evaluate the feasibility of DWI-targeted upfront radiotherapy boost in LARC. This work will inform an upcoming clinical feasibility study.

Radiogenomics for predicting microsatellite instability status and PD-L1 expression with machine learning in endometrial cancers: A multicenter study

PurposeTo evaluate the effectiveness of machine learning model based on magnetic resonance imaging (MRI) in identifying microsatellite instability (MSI) status and PD-L1 expression in endometrial cancer (EC). MethodsThis retrospective study included 82 EC patients from 2 independent centers. Radiomics features from the intratumoral and peritumoral regions, obtained from four conventional MRI sequences (T2-weighted images; contrast-enhanced T1-weighted images; diffusion-weighted images; apparent diffusion coefficient), were combined with clinicopathologic characteristics to develop machine learning model for predicting MSI status and PD-L1 expression. 60 patients from center 1 were used as the training set for model construction, while 22 patients from center 2 were used as an external validation set for model evaluation. ResultsFor predicting MSI status, the clinicopathologic model, radscore model, and combination model achieved area under the curves (AUCs) of 0.728, 0.833, and 0.889 in the training set, respectively, and 0.595, 0.790, and 0.848 in the validation set, respectively. For predicting PD-L1 expression, the clinicopathologic model, radscore model, and combination model achieved AUCs of 0.648, 0.814, and 0.834 in the training set, respectively, and 0.660, 0.708, and 0.764 in the validation set, respectively. Calibration curve analysis and decision curve analysis demonstrated good calibration and clinical utility of the combination model. ConclusionThe machine learning model incorporating MRI-based radiomics features and clinicopathologic characteristics could be a potential tool for predicting MSI status and PD-L1 expression in EC. This approach may contribute to precision medicine for EC patients.

Association between high expression of intratumoral fibroblast activation protein and survival in patients with intrahepatic cholangiocarcinoma

BackgroundCancer-associated fibroblasts (CAFs) have been reported to exhibit protumorigenic effects. Among the well-known CAF markers such as smooth muscle actin (SMA) and fibroblast activation protein (FAP), high expression of SMA in the peritumoral stroma has been reported to be a prognostic factor in various cancers. However, the effect of high FAP expression in intrahepatic cholangiocarcinoma (IHCC) has not been fully clarified. We evaluated the expression of CAF markers, focusing on FAP expression in the peripheral and intratumoral regions, to clarify the association with survival in patients with IHCC.MethodsThe study cohort comprised 37 patients who underwent curative resection for IHCC. The FAP expressions were evaluated in the peripheral and intratumoral regions of the resected tissues. Clinicopathological factors and survival outcomes were investigated between patients with high versus low FAP expression. Uni- and multivariate analyses were performed to identify the prognostic factors for overall survival and relapse-free survival.ResultsThe median area percentages of FAP expression in the peripheral and intratumoral regions were 15.5% and 17.8%, respectively. High FAP expression in the intratumoral region was significantly associated with worse overall survival and disease-free survival than low FAP expression in the intratumoral region. Multivariate analysis identified high intratumoral FAP expression as a risk factor for worse overall survival (hazard ratio, 2.450; p = 0.049) and relapse-free survival (hazard ratio, 2.743; p = 0.034).ConclusionsHigh intratumoral FAP expression was associated with worse survival, suggesting that intratumoral FAP expression represents malignant progression in patients with IHCC.

Spatial multimodal analysis revealed tertiary lymphoid structures as a risk stratification indicator in combined hepatocellular-cholangiocarcinoma

The microenvironment created by tertiary lymphoid structures (TLSs) can support and regulate immune responses, affecting the prognosis and immune treatment of patients. Nevertheless, the actual importance of TLSs for predicting the prognosis of combined hepatocellular-cholangiocarcinoma (cHCC-CCA) patients remains unclear. Herein, using spatial transcriptomic analysis, we revealed that a gene signature of TLSs specific to cHCC-CCA was associated with high-intensity immune infiltration. Then, a novel scoring system was developed to evaluate the distribution and frequency of TLSs in intra-tumoral and extra-tumoral regions (iTLS and eTLS scores) in 146 cHCC-CCA patients. iTLS score was positively associated with promising prognosis, likely due to the decreased frequency of suppressive immune cell like Tregs, and the ratio of CD163+ macrophages to macrophages in intra-tumoral TLSs via imaging mass cytometry, while improved prognosis is not necessarily indicated by a higher eTLS score. Overall, this study highlights the potential of TLSs as a prognostic factor and an indicator of immune therapy in cHCC-CCA.

Prediction of recurrence risk factors in patients with early-stage cervical cancers by nomogram based on MRI handcrafted radiomics features and deep learning features: a dual-center study.

To establish and validate a deep learning radiomics nomogram (DLRN) based on intratumoral and peritumoral regions of MR images and clinical characteristics to predict recurrence risk factors in early-stage cervical cancer and to clarify whether DLRN could be applied for risk stratification. Two hundred and twenty five pathologically confirmed early-stage cervical cancers were enrolled and made up the training cohort and internal validation cohort, and 40 patients from another center were enrolled into the external validation cohort. On the basis of region of interest (ROI) of intratumoral and different peritumoral regions, two sets of features representing deep learning and handcrafted radiomics features were created using combined images of T2-weighted MRI (T2WI) and diffusion-weighted imaging (DWI). The signature subset with the best discriminant features was chosen, and deep learning and handcrafted signatures were created using logistic regression. Integrated with independent clinical factors, a DLRN was built. The discrimination and calibration of DLNR were applied to assess its therapeutic utility. The DLRN demonstrated satisfactory performance for predicting recurrence risk factors, with AUCs of 0.944 (95% confidence interval 0.896-0.992) and 0.885 (95% confidence interval 0.834-0.937) in the internal and external validation cohorts. Furthermore, decision curve analysis revealed that the DLRN outperformed the clinical model, deep learning signature, and radiomics signature in terms of net benefit. A DLRN based on intratumoral and peritumoral regions had the potential to predict and stratify recurrence risk factors for early-stage cervical cancers and enhance the value of individualized precision treatment.

Development of an interpretable machine learning model for Ki-67 prediction in breast cancer using intratumoral and peritumoral ultrasound radiomics features.

Traditional immunohistochemistry assessment of Ki-67 in breast cancer (BC) via core needle biopsy is invasive, inaccurate, and nonrepeatable. While machine learning (ML) provides a promising alternative, its effectiveness depends on extensive data. Although the current mainstream MRI-centered radiomics offers sufficient data, its unsuitability for repeated examinations, along with limited accessibility and an intratumoral focus, constrain the application of predictive models in evaluating Ki-67 levels. This study aims to explore ultrasound (US) image-based radiomics, incorporating both intra- and peritumoral features, to develop an interpretable ML model for predicting Ki-67 expression in BC patients. A retrospective analysis was conducted on 263 BC patients, divided into training and external validation cohorts. From intratumoral and peritumoral regions of interest (ROIs) in US images, 849 distinctive radiomics features per ROI were derived. These features underwent systematic selection to analyze Ki-67 expression relationships. Four ML models-logistic regression, random forests, support vector machine (SVM), and extreme gradient boosting-were formulated and internally validated to identify the optimal predictive model. External validation was executed to ascertain the robustness of the optimal model, followed by employing Shapley Additive Explanations (SHAP) to reveal the significant features of the model. Among 231 selected BC patients, 67.5% exhibited high Ki-67 expression, with consistency observed across both training and validation cohorts as well as other clinical characteristics. Of the 1698 radiomics features identified, 15 were significantly correlated with Ki-67 expression. The SVM model, utilizing combined ROI, demonstrated the highest accuracy [area under the receiver operating characteristic curve (AUROC): 0.88], making it the most suitable for predicting Ki-67 expression. External validation sustained an AUROC of 0.82, affirming the model's robustness above a 40% threshold. SHAP analysis identified five influential features from intra- and peritumoral ROIs, offering insight into individual prediction. This study emphasized the potential of SVM model using radiomics features from both intra- and peritumoral US images, for predicting elevated Ki-67 levels in BC patients. The model exhibited strong performance in validations, indicating its promise as a noninvasive tool to enable personalized decision-making in BC care.

CNSC-26. GLIOBLASTOMA-NEURONAL CIRCUIT INTEGRATION IS MODULATED BY INTERLEUKIN-6

Abstract The pleiotropic cytokine interleukin-6 (IL-6) is known to be involved in both pro- and anti-inflammatory signaling cascades. In glioblastoma studies using animal and human models, IL-6 promotes immunosuppression and tumor progression and is negatively associated with survival. Intriguingly, IL-6 also contributes to increased synaptogenesis during development and following focal brain injury. Glioblastoma remodeling of neuronal circuits influences patient survival; therefore, it is possible that IL-6 has mechanistic significance. Here, we investigated IL-6 as a driver of activity-dependent glioblastoma proliferation using patient-derived xenograft (PDX) mouse and human glioblastoma models. We assessed neuronal activity by microelectrode arrays and in vitro calcium imaging using weighted mean firing rate (WMFR), network burst frequency (NBF), and synchrony index. Next, we used bulk and single-cell RNA sequencing on 13,731 cells from ten tumors including pair-matched samples to identify differentially expressed genes in intratumoral regions with elevated functional connectivity and found IL6 to be highly upregulated. Mouse embryonic cortical neurons and cerebral organoids co-cultured with primary patient-derived glioblastoma cells demonstrated concentration-dependent neuronal hyperexcitability (increased WMFR and NBF) in IL-6-overexpressing conditions compared to the neuron-only condition. Pharmacological inhibition using the humanized IL-6 receptor antibody tocilizumab reduced network synchrony across models. These findings suggest that IL-6-induced glioma-neuronal hyperexcitability may be inhibited by the FDA-approved tocilizumab, thereby providing preclinical support for its use to treat activity-dependent mechanisms of glioblastoma proliferation. Future studies are needed to uncover mechanisms and clinical efficacy.

NIMG-12. SPATIALLY MAPPING TREATMENT-EFFECT IN RECURRENT GLIOMA WITH PHYSIOLOGIC MRI AND DEEP LEARNING

Abstract INTRODUCTION Discriminating between glioma tumor recurrence (rTumor) and treatment-induced effects (TxE) from chemo and radiotherapy using MRI is an ongoing challenge in neuro-oncology. Although deep learning has been used to characterize suspicious lesions, these models typically do not account for lesions that are comprised of mixtures of rTumor and TxE components. We hypothesize that using an approach that includes using tissue samples with known location on MRI, incorporating perfusion-weighted and diffusion-weighted with anatomical MRI, self-supervised pretraining, and a novel ensembling strategy will improve performance and attain spatial maps of rTumor and TxE for determining subsequent treatment. METHODS Using 218 pathology confirmed tissue samples with coordinates on multi-parametric MRI from 111 patients with suspected tumor progression, we developed a deep learning approach to spatially map intratumoral regions of rTumor and TxE. Our model leveraged 10mm isotropic cubes centered on the tissue sample coordinates that includes Diffusion-Weighted and Dynamic Susceptibility-Contrast Perfusion MRI and large amounts of pathologically unlabeled data from the lesion in self-supervised pretraining. The multiple networks of the ensemble were generated through repeated constrained cross validation (RC-CV) which enabled unique repetitions of training with more variation in class proportions. RESULTS Our best classifier performed with a RC-CV/test AUROC of 0.85/0.81 and was able to generate maps of tumor probability as well as ensenble variance and uncertainty. Ensemble variance was significantly higher in prediction failures of rTumor (P=.006) and epistemic uncertainty was significantly higher (P=.002) in prediction failures of TxE in training. Test Performance dropped by 14.7% when removing self-supervised pretraining, 21.3% when removing physiological imaging contrasts, and 1.2% when ensembling across standard 4-fold CV. CONCLUSION The AUROC of .81 supports the feasibility of discriminating between TxE and rTumor within a single lesion non-invasively. Physiologic MRI, pretraining on pathology unlabeled data, and ensembling networks, improve performance and help characterize model failures.

An integrated model combined intra- and peritumoral regions for predicting chemoradiation response of non small cell lung cancers based on radiomics and deep learning

PurposeThe purpose of this study was to develop a model for predicting chemoradiation response in non-small cell lung cancer (NSCLC) patients by integrating radiomics and deep-learning features and combined intra- and peritumoral regions with pre-treated CT images. Materials and methodsThis study enrolled 462 patients with NSCLC who received chemoradiation. On the basis of pretreated CT images, we developed three models to compare the prediction of chemoradiation: intratumoral, peritumoral and combined regions. To further illustrate each model, we established different feature integration methods: a) radiomics model with 1500 features; b) deep learning model with a multiple instance learning algorithm; c) integrated model by integrating radiomic and deep learning features. For radiomics and integrated models, support vector machine and the least absolute shrinkage and selection operator were used to extract and select features. Transfer learning and max pooling algorithms were used to identify high informative features in deep learning models. We applied ten-fold cross validation in model training and testing. ResultsThe best area under the curve (AUC) of intratumoral, peritumoral and combined models were 0.89 (95% CI, 0.74–0.93), 0.86 (95% CI, 0.75–0.92) and 0.92 (95% CI, 0.81–0.95), respectively. It indicated the importance of the peritumoral region for treatment response prediction and should be used in combination with the intratumoral region. Integrated models gave better results than models with radiomics and deep learning features alone in all regions of interest and radiomics models outperformed deep learning models in any comparative models. ConclusionsThe model that integrate radiomic and deep learning features and combined intra- and peritumoral regions provide valuable information in predicting treatment response of chemoradiation. It can help oncologists customize personalized clinical treatment plans for NSCLC patients.