Distinct tumor signatures using deep learning-based characterization of the peritumoral microenvironment in glioblastomas and brain metastases.
Ontology highlight
ABSTRACT:
SUBMITTER: Samani ZR
PROVIDER: S-EPMC8280204 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Background and purposePrevious studies have suggested that increased mortality and disability in patients with brain tumor are associated with peritumoral brain edema. However, the mechanism of peritumoral brain edema in brain tumors is unknown. This study aimed to investigate the effect of Piezo1 overexpression on peritumoral brain edema in glioblastomas.Materials and methodsThe Piezo1 expression in cell lines and paired samples was detected by quantitative reverse transcription polymerase chain reaction, Western blot, and immunohistochemistry. Sixty-four patients with glioblastomas were analyzed retrospectively. The Piezo1 expression of tumor tissue was detected by immunohistochemistry. The diameters of tumor and edema were measured by preoperative MR imaging, and the edema index value was calculated.ResultsWestern blot and quantitative reverse transcription polymerase chain reaction showed that Piezo1 expression was higher in 6 glioma cell lines than in the normal astrocyte cell line. Compared with peritumoral tissues, Piezo1 was up-regulated in tumor tissues. Sixty-four patients with glioblastomas were enrolled in further study. Piezo1 was higher in the moderate edema group than in the mild edema group (P < .001), higher in the severe edema group than in the moderate edema group (P < .001), and correlated with the edema index (r = 0.73; P < .001). Receiver operating characteristic curve analysis showed that the edema index yielded an area under the curve of 0.867 (95% CI, 0.76-0.97; P < .001), with a sensitivity of 100% and a specificity of 70%.ConclusionsPiezo1 overexpression is positively correlated with the degree of peritumoral brain edema in glioblastomas. Predicting high Piezo1 expression in tumor tissues based on the edema extent shows good sensitivity and specificity.
| S-EPMC7658862 | biostudies-literature
Project description:BackgroundHigh tumor-infiltrating lymphocytes (TILs) and hemorrhage are important prognostic factors in patients who have undergone craniotomy for melanoma brain metastases (MBM) before 2011 at the University of Pittsburgh Medical Center (UPMC). We have investigated the prognostic or predictive role of these histopathologic factors in a more contemporary craniotomy cohort from the University of North Carolina at Chapel Hill (UNC-CH). We have also sought to understand better how various immune cell subsets, angiogenic factors, and blood vessels may be associated with clinical and radiographic features in MBM.MethodsBrain tumors from the UPMC and UNC-CH patient cohorts were (re)analyzed by standard histopathology, tumor tissue imaging, and gene expression profiling. Variables were associated with overall survival (OS) and radiographic features.ResultsThe patient subgroup with high TILs in craniotomy specimens and subsequent treatment with immune checkpoint inhibitors (ICIs, n=7) trended to have longer OS compared to the subgroup with high TILs and no treatment with ICIs (n=11, p=0.059). Bleeding was significantly associated with tumor volume before craniotomy, high melanoma-specific expression of basic fibroblast growth factor (bFGF), and high density of CD31+αSMA- blood vessels. Brain tumors with high versus low peritumoral edema before craniotomy had low (17%) versus high (41%) incidence of brisk TILs. Melanoma-specific expression of the vascular endothelial growth factor (VEGF) was comparable to VEGF expression by TILs and was not associated with any particular prognostic, radiographic, or histopathologic features. A gene signature associated with gamma delta (gd) T cells was significantly higher in intracranial than same-patient extracranial metastases and primary melanoma. However, gdT cell density in MBM was not prognostic.ConclusionsICIs may provide greater clinical benefit in patients with brisk TILs in MBM. Intratumoral hemorrhage in brain metastases, a significant clinical problem, is not merely associated with tumor volume but also with underlying biology. bFGF may be an essential pathway to target. VEGF, a factor principally associated with peritumoral edema, is not only produced by melanoma cells but also by TILs. Therefore, suppressing low-grade peritumoral edema using corticosteroids may harm TIL function in 41% of cases. Ongoing clinical trials targeting VEGF in MBM may predict a lack of unfavorable impacts on TIL density and/or intratumoral hemorrhage.
| S-EPMC7860978 | biostudies-literature
Project description:Complete resection is an indispensable treatment option in the management of brain metastases (BM). 5-aminolevulinic acid (5-ALA) fluorescence is used for improved intraoperative visualization of tumor tissue in gliomas and was recently observed in BM. We investigated the potential of 5-ALA fluorescence to visualize the infiltrative growth of BM in the peritumoral brain tissue and its histopathological correlate. Patients with BM resection after 5-ALA administration and collection of tissue samples from peritumoral brain tissue were included. Each tissue sample was histopathologically investigated for tumor cell infiltration and angiogenesis. Altogether, 88 samples were collected from the peritumoral brain tissue in 58 BM of 55 patients. Visible 5-ALA fluorescence was found in 61 (69%) of the samples, tumor infiltration in 19 (22%) and angiogenesis in 13 (15%) of samples. Angiogenesis showed a significant correlation with presence of fluorescence (p = 0.008). Moreover, angiogenesis was related to visible 5-ALA fluorescence and showed an association with patient prognosis since it was significantly correlated to shorter time to local progression/recurrence (p = 0.001) and lower one-year survival (p = 0.031). Consequently, angiogenesis in the peritumoral brain tissue of BM might be a novel prognostic marker for individualized perioperative treatment concepts in the future.
| S-EPMC7913757 | biostudies-literature
Project description:BackgroundThree-dimensional T1 magnetization prepared rapid acquisition gradient echo (3D-T1-MPRAGE) is preferred in detecting brain metastases (BM) among MRI. We developed an automatic deep learning-based detection and segmentation method for BM (named BMDS net) on 3D-T1-MPRAGE images and evaluated its performance.MethodsThe BMDS net is a cascaded 3D fully convolution network (FCN) to automatically detect and segment BM. In total, 1652 patients with 3D-T1-MPRAGE images from 3 hospitals (n = 1201, 231, and 220, respectively) were retrospectively included. Manual segmentations were obtained by a neuroradiologist and a radiation oncologist in a consensus reading in 3D-T1-MPRAGE images. Sensitivity, specificity, and dice ratio of the segmentation were evaluated. Specificity and sensitivity measure the fractions of relevant segmented voxels. Dice ratio was used to quantitatively measure the overlap between automatic and manual segmentation results. Paired samples t-tests and analysis of variance were employed for statistical analysis.ResultsThe BMDS net can detect all BM, providing a detection result with an accuracy of 100%. Automatic segmentations correlated strongly with manual segmentations through 4-fold cross-validation of the dataset with 1201 patients: the sensitivity was 0.96 ± 0.03 (range, 0.84-0.99), the specificity was 0.99 ± 0.0002 (range, 0.99-1.00), and the dice ratio was 0.85 ± 0.08 (range, 0.62-0.95) for total tumor volume. Similar performances on the other 2 datasets also demonstrate the robustness of BMDS net in correctly detecting and segmenting BM in various settings.ConclusionsThe BMDS net yields accurate detection and segmentation of BM automatically and could assist stereotactic radiotherapy management for diagnosis, therapy planning, and follow-up.
| S-EPMC7158643 | biostudies-literature
Project description:BackgroundApproximately one-fourth of all cancer metastases are found in the brain. MRI is the primary technique for detection of brain metastasis, planning of radiotherapy, and the monitoring of treatment response. Progress in tumor treatment now requires detection of new or growing metastases at the small subcentimeter size, when these therapies are most effective.PurposeTo develop a deep-learning-based approach for finding brain metastasis on MRI.Study typeRetrospective.SequenceAxial postcontrast 3D T1 -weighted imaging.Field strength1.5T and 3T.PopulationA total of 361 scans of 121 patients were used to train and test the Faster region-based convolutional neural network (Faster R-CNN): 1565 lesions in 270 scans of 73 patients for training; 488 lesions in 91 scans of 48 patients for testing. From the 48 outputs of Faster R-CNN, 212 lesions in 46 scans of 18 patients were used for training the RUSBoost algorithm (MatLab) and 276 lesions in 45 scans of 30 patients for testing.AssessmentTwo radiologists diagnosed and supervised annotation of metastases on brain MRI as ground truth. This data were used to produce a 2-step pipeline consisting of a Faster R-CNN for detecting abnormal hyperintensity that may represent brain metastasis and a RUSBoost classifier to reduce the number of false-positive foci detected.Statistical testsThe performance of the algorithm was evaluated by using sensitivity, false-positive rate, and receiver's operating characteristic (ROC) curves. The detection performance was assessed both per-metastases and per-slice.ResultsTesting on held-out brain MRI data demonstrated 96% sensitivity and 20 false-positive metastases per scan. The results showed an 87.1% sensitivity and 0.24 false-positive metastases per slice. The area under the ROC curve was 0.79.ConclusionOur results showed that deep-learning-based computer-aided detection (CAD) had the potential of detecting brain metastases with high sensitivity and reasonable specificity.Level of evidence3 TECHNICAL EFFICACY STAGE: 2 J. Magn. Reson. Imaging 2020;52:1227-1236.
| S-EPMC7487020 | biostudies-literature
Project description:BackgroundThe differential diagnosis of glioblastomas (GBM) from solitary brain metastases (SBM) is essential because the surgical strategy varies according to the histopathological diagnosis. Intraoperative ultrasound elastography (IOUS-E) is a relatively novel technique implemented in the surgical management of brain tumors that provides additional information about the elasticity of tissues. This study compares the discriminative capacity of intraoperative ultrasound B-mode and strain elastography to differentiate GBM from SBM.MethodsWe performed a retrospective analysis of patients who underwent craniotomy between March 2018 to June 2020 with glioblastoma (GBM) and solitary brain metastases (SBM) diagnoses. Cases with an intraoperative ultrasound study were included. Images were acquired before dural opening, first in B-mode, and then using the strain elastography module. After image pre-processing, an analysis based on deep learning was conducted using the open-source software Orange. We have trained an existing neural network to classify tumors into GBM and SBM via the transfer learning method using Inception V3. Then, logistic regression (LR) with LASSO (least absolute shrinkage and selection operator) regularization, support vector machine (SVM), random forest (RF), neural network (NN), and k-nearest neighbor (kNN) were used as classification algorithms. After the models' training, ten-fold stratified cross-validation was performed. The models were evaluated using the area under the curve (AUC), classification accuracy, and precision.ResultsA total of 36 patients were included in the analysis, 26 GBM and 10 SBM. Models were built using a total of 812 ultrasound images, 435 of B-mode, 265 (60.92%) corresponded to GBM and 170 (39.8%) to metastases. In addition, 377 elastograms, 232 (61.54%) GBM and 145 (38.46%) metastases were analyzed. For B-mode, AUC and accuracy values of the classification algorithms ranged from 0.790 to 0.943 and from 72 to 89%, respectively. For elastography, AUC and accuracy values ranged from 0.847 to 0.985 and from 79% to 95%, respectively.ConclusionAutomated processing of ultrasound images through deep learning can generate high-precision classification algorithms that differentiate glioblastomas from metastases using intraoperative ultrasound. The best performance regarding AUC was achieved by the elastography-based model supporting the additional diagnostic value that this technique provides.
| S-EPMC7884775 | biostudies-literature
Project description:Brain is the most frequent site for distant metastases of non-small cell lung cancer (NSCLC). Brain metastasis (BM) is also the leading cause of disabilities and death in advanced NSCLC. In recent years, the application and effectiveness of small-molecule tyrosine kinase inhibitors has formed the basis for the treatment of NSCLC brain metastases with driver gene mutations. With the development of programmed cell death protein 1 (PD-1)/programmed cell death protein ligand 1 (PD-L1) inhibitors and relevant combination therapies, immunotherapy has become an important choice for non-classic oncogene addicted NSCLC BM patients. Also, the roles of relevant biomarkers are increasingly standing out. By reason of the particular immunopathological features of NSCLC brain metastases and its microenvironment, the aim of this review is to summarize relevant research progresses and provide more references for combination strategies of different therapeutic methods as well as the development of novel immunotherapies.?.
| S-EPMC6717868 | biostudies-literature
Project description:Brain metastases signify a deleterious milestone in the progression of several advanced cancers, predominantly originating from lung, breast and melanoma malignancies, with a median survival timeframe nearing six months. Existing therapeutic regimens yield suboptimal outcomes; however, burgeoning insights into the tumor microenvironment, particularly the immunosuppressive milieu engendered by tumor-brain interplay, posit immunotherapy as a promising avenue for ameliorating brain metastases. In this review, we meticulously delineate the research advancements concerning the microenvironment of brain metastases, striving to elucidate the panorama of their onset and evolution. We encapsulate three emergent immunotherapeutic strategies, namely immune checkpoint inhibition, chimeric antigen receptor (CAR) T cell transplantation and glial cell-targeted immunoenhancement. We underscore the imperative of aligning immunotherapy development with in-depth understanding of the tumor microenvironment and engendering innovative delivery platforms. Moreover, the integration with established or avant-garde physical methodologies and localized applications warrants consideration in the prevailing therapeutic schema.
| S-EPMC10725587 | biostudies-literature
Project description:BackgroundBrain metastasis is a common outcome in non-small cell lung cancer, and despite aggressive treatment, its clinical outcome is still frustrating. In recent years, immunotherapy has been developing rapidly, however, its therapeutic outcomes for primary lung cancer and brain metastases are not the same, suggesting that there may be differences in the immune microenvironment of primary lung cancer and brain metastases, however, we currently know little about these differences.MethodsSeventeen paired samples of NSCLC and their brain metastases and 45 other unpaired brain metastases samples were collected for the current study. Immunohistochemical staining was performed on all samples for the following markers: immune checkpoints CTLA-4, PD-1, PD-L1, B7-H3, B7-H4, IDO1, and EphA2; tumor-infiltrating lymphocytes (TILs) CD3, CD4, CD8, and CD20; tumor-associated microglia/macrophages (TAMs) CD68 and CD163; and tumor proliferation index Ki-67. The differences in expression of these markers were compared in 17 paired samples, and the effect of the expression level of these markers on the prognosis of patients was analyzed in lung adenocarcinoma brain metastases samples. Subsequently, multiplex immunofluorescence staining was performed in a typical lung-brain paired sample based on the aforementioned results. The multiplex immunofluorescence staining results revealed the difference in tumor immune microenvironment between primary NSCLC and brain metastases.ResultsIn 17 paired lesions, the infiltration of CTLA-4+ (P = 0.461), PD-1+ (P = 0.106), CD3+ (P = 0.045), CD4+ (P = 0.037), CD8+ (P = 0.008), and CD20+ (P = 0.029) TILs in brain metastases were significantly decreased compared with primary tumors. No statistically significant difference was observed in the CD68 (P = 0.954) and CD163 (P = 0.654) TAM infiltration between primary NSCLC and paired brain metastases. In all the brain metastases lesions, the expression of PD-L1 is related to the time interval of brain metastases in NSCLC. In addition, the Cox proportional hazards regression models showed high expression of B7-H4 (hazard ratio [HR] = 3.276, 95% confidence interval [CI] 1.335-8.041, P = 0.010) and CD68 TAM infiltration (HR = 3.775, 95% CI 1.419-10.044, P = 0.008) were independent prognosis factors for lung adenocarcinoma brain metastases patients.ConclusionsBoth temporal and spatial heterogeneity is present between the primary tumor and brain metastases of NCSLC. Brain metastases lesions exhibit a more immunosuppressive tumor immune microenvironment. B7-H4 and CD68+ TAMs may have potential therapeutic value for lung adenocarcinoma brain metastases patients.
| S-EPMC10809508 | biostudies-literature
Project description:Image registration and segmentation are the two most studied problems in medical image analysis. Deep learning algorithms have recently gained a lot of attention due to their success and state-of-the-art results in variety of problems and communities. In this paper, we propose a novel, efficient, and multi-task algorithm that addresses the problems of image registration and brain tumor segmentation jointly. Our method exploits the dependencies between these tasks through a natural coupling of their interdependencies during inference. In particular, the similarity constraints are relaxed within the tumor regions using an efficient and relatively simple formulation. We evaluated the performance of our formulation both quantitatively and qualitatively for registration and segmentation problems on two publicly available datasets (BraTS 2018 and OASIS 3), reporting competitive results with other recent state-of-the-art methods. Moreover, our proposed framework reports significant amelioration (p < 0.005) for the registration performance inside the tumor locations, providing a generic method that does not need any predefined conditions (e.g., absence of abnormalities) about the volumes to be registered. Our implementation is publicly available online at https://github.com/TheoEst/joint_registration_tumor_segmentation.
| S-EPMC7100603 | biostudies-literature