Tumor Enhancement Research Articles

Computed tomographic imaging features to differentiate gastric schwannomas from gastrointestinal stromal tumours: a matched case–control study

To investigate clinical data and computed tomographic (CT) imaging features in differentiating gastric schwannomas (GSs) from gastric stromal tumours (GISTs) in matched patients, 31 patients with GSs were matched with 62 patients with GISTs (1:2) in sex, age, and tumour site. The clinical and imaging data were analysed. A significant (P < 0.05) difference was found in the tumour margin, enhancement pattern, growth pattern, and LD values between the 31 patients with GSs and 62 matched patients with GISTs. The GS lesions were mostly (93.5%) well defined while only 61.3% GIST lesions were well defined.The GS lesions were significantly (P = 0.036) smaller than the GIST lesions, with the LD ranging 1.5–7.4 (mean 3.67 cm) cm for the GSs and 1.0–15.30 (mean 5.09) cm for GIST lesions. The GS lesions were more significantly (P = 0.001) homogeneously enhanced (83.9% vs. 41.9%) than the GIST lesions. The GS lesions were mainly of the mixed growth pattern both within and outside the gastric wall (74.2% vs. 22.6%, P < 0.05) compared with that of GISTs. No metastasis or invasion of adjacent organs was present in any of the GS lesions, however, 1.6% of GISTs experienced metastasis and 3.2% of GISTs presented with invasion of adjacent organs. Heterogeneous enhancement and mixed growth pattern were two significant (P < 0.05) independent factors for distinguishing GS from GIST lesions. In conclusion: GS and GIST lesions may have significantly different features for differentiation in lesion margin, heterogeneous enhancement, mixed growth pattern, and longest lesion diameter, especially heterogeneous enhancement and mixed growth pattern.

Time Efficiency and Performance of Single-Use vs Reusable Cystoscopes: A Randomized Benchtop and Simulated Clinical Assessment.

Introduction: A flexible cystoscope is an indispensable tool for urologists, facilitating a variety of procedures in both the operating room and at bedside. Single-use cystoscopes offer benefits including accessibility and decreased burden for reprocessing. The aims of this study were to compare time efficiency and performance of single-use and reusable cystoscopes. Methods: Ten new Ambu® aScope™ 4 Cysto single-use and two Olympus CYF-5 reusable cystoscopes were compared in simulated bedside cystoscopy and benchtop testing. Ten urologists performed simulated cystoscopy using both cystoscopes in a randomized order. Times for supply-gathering, setup, cystoscopy, cleanup, and cumulative time were recorded, followed by a Likert feedback survey. For benchtop assessment, physical, optical, and functional specifications were assessed and compared between cystoscopes. Results: The single-use cystoscope demonstrated shorter supply-gathering, setup, cleanup, and cumulative times (824 vs 1231 seconds; p < 0.05) but a comparable cystoscopy time to the reusable cystoscope (202 vs 212 seconds; p = 0.32). The single-use cystoscope had a higher image resolution, but a narrower field of view. Upward deflection was greater for the single-use cystoscope (214.50° vs 199.45°; p < 0.01) but required greater force (2.5 × ). The working channel diameter and irrigation rate were greater in the reusable cystoscope. While the single-use cystoscope lacked tumor enhancing optical features, it had higher Likert scale scores for Time Efficiency and Overall Satisfaction. Conclusion: The single-use cystoscope demonstrates comparable benchtop performance and superior time efficiency compared to reusable cystoscopes. However, the reusable cystoscope has superior optical versatility and flow rate. Knowledge of these differences allows for optimal cystoscope selection based on procedure indication.

α-SMA+ cancer-associated fibroblasts increased tumor enhancement ratio on contrast-enhanced multidetector-row computed tomography in stages I-III colon cancer.

Our prior research revealed that the tumor enhancement ratio (TER) on triphasic abdominal contrast-enhanced MDCT (CE-MDCT) scans was a prognostic factor for patients with stages I-III colon cancer. Building upon this finding, the present study aims to investigate the proteomic changes in colon cancer patients with varying TER values. TER was analyzed on preoperative triphasic CE-MDCT scans of 160 stages I-III colon cancer patients. The survival outcomes of those in the low-TER and high-TER groups were compared. Proteomic analysis on colon cancer tissues was performed by mass spectrometry (MS) and verified by immune-histological chemistry (IHC) assays. In vivo, mouse xenograft models were employed to test the function of target proteins identified through the MS. CE-MDCT scans were conducted on mice xenografts, and the TER values were compared. Patients in the high-TER group had a significantly worse prognosis than those in the low-TER group. Proteomic analysis of colon cancer tissues revealed 153 differentially expressed proteins between the two groups. A correlation between TER and the abundance of α-SMA protein in tumor tissue was observed. IHC assays further confirmed that α-SMA protein expression was significantly increased in high-TER colon cancer, predominantly in cancer-associated fibroblasts (CAFs) within the cancer stroma. Moreover, CAFs promoted the growth of CRC xenografts in vivo and increased TER. Our study identified the distinct protein changes in colon cancer with low and high TER for the first time. The presence of CAFs may promote the growth of colon cancer and contribute to an increased TER.

MRI Directed Magnevist Effective to Study Toxicity of Gd-Doped Mesoporous Carbon Nanoparticles in Mice Model.

Magnetic resonance imaging (MRI) has been a valuable and widely used examination technique in clinical diagnosis and prognostic efficacy evaluation. The introduction of MRI contrast agent (CA) improves its sensitivity obviously, particularly with the development of nano-CA, which presents higher contrast enhancement ability. However, systematical evaluation of their toxicity is still limited, hampering their further translation in clinics. In this paper, to systematically evaluate the toxicity of nano-CA, Gd-doped mesoporous carbon nanoparticles (Gd-MCNs) prepared by a one-step hard template method were introduced as a model and clinically used MRI CA, Magnevist (Gd-DTPA) as control. Their in vitro blood compatibility, cellular toxicity, DNA damage, oxidative stress, inflammation response as well as in vivo toxicity and MR imaging behaviors were studied and compared. The experimental results showed that compared with Gd-DTPA, Gd-MCNs displayed negligible influence on the red blood cell shape, aggregation, BSA structure, macrophage morphology and mitochondrial function. Meanwhile, limited ROS and inflammatory cytokine production also illustrated the cellular compatibility of Gd-MCNs. For in vivo toxicity evaluation, Gd-MCNs presented acceptable in vivo biosafety even under 12 times injection for 12 weeks. More importantly, at the same concentration of Gd, Gd-MCNs displayed better contrast enhancement of tumor than Gd-DTPA, mainly coming from its high MRI relaxation rate which is nearly 9 times that of Gd-DTPA. In this paper, we focus on the toxicity evaluation of MRI nano-CA, Gd-MCNs from different angles. With Gd-DTPA as control, Gd-MCNs appeared to be highly biocompatible and safe nanoparticles that possessed promising potentials for the use of MRI nano-CA. In the future, more research on the long-term genotoxicity and the fate of nanoparticles after being swallowed should be performed.

A Phase 0 Study Assessing the Intracranial Activity of a Metabolic Radiosensitizer in Patients with Glioblastoma

Efforts to overcome treatment resistance in glioblastoma (GBM) have been unsuccessful due to tumor heterogeneity and poor intracranial drug penetration. Targeting altered metabolism is a promising approach to improve GBM therapy despite this heterogeneity. Mycophenolate mofetil (MMF) is an inhibitor of purine synthesis that sensitizes GBM to radiation and temozolomide (TMZ) in vitro and in vivo, but its ability to cross the blood brain barrier and inhibit GBM metabolism in patients is unknown. NCT04477200 is a phase 0/1 dose escalation study of MMF combined with radiation and temozolomide in GBM. Here we report the phase 0 results of this study assessing the intracranial activity of MMF. Purine (GTP and IMP) and mycophenolic acid (MPA, the active metabolite of MMF) concentrations were determined using mass spectrometry in flash-frozen tumor (enhancing and non-enhancing) and normal cortex obtained from 8 patients with recurrent GBM who received MMF (500, 1000, 1500 and 2000 mg BID, N = 2 patients each dose level) for 1 week prior to re-resection and 5 control patients who did not receive MMF prior to re-resection. Plasma MPA concentration was similarly quantified to calculate the enhancing tumor, non-enhancing tumor and normal cortex to plasma MPA ratios. Patients who received MMF had a mean MPA concentration of 2.2 ± 0.7 µM in the enhancing tumor samples, 1.2 ± 0.5 µM in the non-enhancing tumor samples and 1.3 ± 0.5 µM in normal cortex. MPA concentration was negligible in control patients. This corresponded to tissue/plasma MPA ratios of 0.31, 0.17 and 0.10 for enhancing tumor, non-enhancing tumor and normal cortex, respectively. The GTP/IMP ratio was decreased by 75% in enhancing tumor in MMF-treated patients compared to untreated controls (p = 0.009), indicating effective target engagement and inhibition of purine synthesis. The GTP/IMP ratio was also decreased in cortex and non-enhancing tumor, though a paucity of control samples prevented statistical analysis. Twice daily MMF treatment yields intracranial drug concentrations above 1 µM and lowers the GTP/IMP ratio in GBMs, consistent with target engagement. As we have previously observed radiosensitization in vitro with MPA concentrations of 1 µM, these data suggest that MMF may achieve adequate CNS penetration for therapeutic benefit. The Phase 1 component of this study to determine the dose limiting toxicity and maximally tolerated dose of MMF when combined with reirradiation in recurrent GBM and radiation and TMZ in newly diagnosed GBM is ongoing.

Differentiation of glioblastoma and primary central nervous system lymphomas using multiparametric diffusion and perfusion magnetic resonance imaging.

The present study aimed to determine whether combining diffusion-weighted (DWI) and dynamic susceptibility contrast-enhanced perfusion-weighted (DSC-PWI) magnetic resonance imaging (MRI) could differentiate between primary central nervous system lymphoma (PCNSL) and glioblastoma (GBM). The present retrospective study evaluated 45 patients with histologically confirmed brain tumors, of which 18 had PCNSLs and 27 had GBMs. All patients underwent conventional, DWI, and DSC-PWI MRIs before the surgical removal of the lesion or stereotactic biopsy. The solid tumor component, peritumoral edema, and abnormal white matter were measured in three regions of interest to evaluate relative cerebral blood volume (rCBV), apparent diffusion coefficient (ADC) and DWI. In conventional MRI, there were significant differences in tumor numbers, tumor enhancement type, tumor necrosis, hemorrhage and open-ring sign between GBM and PCNSL. Solid tumor ADC and rCBV values (ADCt and rCBVt, respectively) and their ratios with abnormal white matter amounts were significantly higher in GBM cases than in PCNSL cases (P<0.05). The rCBV value for peritumoral edema (rCBVe) and its ratio with abnormal white matter amount (rCBVe/n) were significantly higher in GBM cases than in PCNSL cases (P<0.05). However, ADC values did not differ significantly for peritumoral edema. DWI values did not differ significantly. Combining rCBVt and rCBVe/n provided a perfect area under the receiver operating characteristic curve of 1.00, with 100% sensitivity and 100% specificity for distinguishing GBM from PCNSL. In the results of the present study, the major criterion in the decision-making process distinguishing PCNSL from GBM was the combined rCBVt and rCBVe/n parameter. A minor criterion was the ADCt value of the lesion.

A radiomics-incorporated deep ensemble learning model for multi-parametric MRI-based glioma segmentation

Objective. To develop a deep ensemble learning (DEL) model with radiomics spatial encoding execution for improved glioma segmentation accuracy using multi-parametric magnetic resonance imaging (mp-MRI). Approach. This model was developed using 369 glioma patients with a four-modality mp-MRI protocol: T1, contrast-enhanced T1 (T1-Ce), T2, and FLAIR. In each modality volume, a 3D sliding kernel was implemented across the brain to capture image heterogeneity: 56 radiomic features were extracted within the kernel, resulting in a fourth-order tensor. Each radiomic feature can then be encoded as a 3D image volume, namely a radiomic feature map (RFM). For each patient, all RFMs extracted from all four modalities were processed using principal component analysis for dimension reduction, and the first four principal components (PCs) were selected. Next, a DEL model comprised of four U-Net sub-models was trained for the segmentation of a region-of-interest: each sub-model utilizes the mp-MRI and one of the four PCs as a five-channel input for 2D execution. Last, four softmax probability results given by the DEL model were superimposed and binarized using Otsu’s method as the segmentation results. Three DEL models were trained to segment the enhancing tumor (ET), tumor core (TC), and whole tumor (WT), respectively. The segmentation results given by the proposed ensemble were compared to the mp-MRI-only U-Net results. Main Results. All three radiomics-incorporated DEL models were successfully implemented: compared to the mp-MRI-only U-net results, the dice coefficients of ET (0.777 → 0.817), TC (0.742 → 0.757), and WT (0.823 → 0.854) demonstrated improvement. The accuracy, sensitivity, and specificity results demonstrated similar patterns. Significance. The adopted radiomics spatial encoding execution enriches the image heterogeneity information that leads to the successful demonstration of the proposed DEL model, which offers a new tool for mp-MRI-based medical image segmentation.

Quantifying U-Net uncertainty in multi-parametric MRI-based glioma segmentation by spherical image projection.

Uncertainty quantification in deep learning is an important research topic. For medical image segmentation, the uncertainty measurements are usually reported as the likelihood that each pixel belongs to the predicted segmentation region. In potential clinical applications, the uncertainty result reflects the algorithm's robustness and supports the confidence and trust of the segmentation result when the ground-truth result is absent. For commonly studied deep learning models, novel methods for quantifying segmentation uncertainty are in demand. To develop a U-Net segmentation uncertainty quantification method based on spherical image projection of multi-parametric MRI (MP-MRI) in glioma segmentation. The projection of planar MRI data onto a spherical surface is equivalent to a nonlinear image transformation that retains global anatomical information. By incorporating this image transformation process in our proposed spherical projection-based U-Net (SPU-Net) segmentation model design, multiple independent segmentation predictions can be obtained from a single MRI. The final segmentation is the average of all available results, and the variation can be visualized as a pixel-wise uncertainty map. An uncertainty score was introduced to evaluate and compare the performance of uncertainty measurements. The proposed SPU-Net model was implemented on the basis of 369 glioma patients with MP-MRI scans (T1, T1-Ce, T2, and FLAIR). Three SPU-Net models were trained to segment enhancing tumor (ET), tumor core (TC), and whole tumor (WT), respectively. The SPU-Net model was compared with (1) the classic U-Net model with test-time augmentation (TTA) and (2) linear scaling-based U-Net (LSU-Net) segmentation models in terms of both segmentation accuracy (Dice coefficient, sensitivity, specificity, and accuracy) and segmentation uncertainty (uncertainty map and uncertainty score). The developed SPU-Net model successfully achieved low uncertainty for correct segmentation predictions (e.g., tumor interior or healthy tissue interior) and high uncertainty for incorrect results (e.g., tumor boundaries). This model could allow the identification of missed tumor targets or segmentation errors in U-Net. Quantitatively, the SPU-Net model achieved the highest uncertainty scores for three segmentation targets (ET/TC/WT): 0.826/0.848/0.936, compared to 0.784/0.643/0.872 using the U-Net with TTA and 0.743/0.702/0.876 with the LSU-Net (scaling factor=2). The SPU-Net also achieved statistically significantly higher Dice coefficients, underscoring the improved segmentation accuracy. The SPU-Net model offers a powerful tool to quantify glioma segmentation uncertainty while improving segmentation accuracy. The proposed method can be generalized to other medical image-related deep-learning applications for uncertainty evaluation.

P20.13.A AUTOGRAFTED HUMAN CORTICAL SLICE MODEL REVEALS MESENCHYMAL-LIKE TRANSITION IN BRAF-MUTATED GLIOBLASTOMA UNDER TARGETED THERAPY

Abstract BACKGROUND The emergence of personalized therapy, which targets oncogenic drivers, has provided a novel adjunct to the standard-of-care therapy for glioblastoma, the most common central nervous system malignancy. Among the less frequently observed oncogenic drivers in glioblastoma are mutations in the proto-oncogene B-Raf. Although treatment options for BRAF-mutated malignancies have been extensively investigated, the implications of BRAF and MEK inhibition in gliomas remain ambiguous. In this study, we present a model system designed to predict cellular temporal trajectories, employing human neocortical slice models and patient-derived tumor cells, coupled with spatially resolved and single-cell transcriptomic analyses for comprehensive computational modeling. MATERIAL AND METHODS Non-neoplastic cortical tissue was procured from a BRAF(V600E) mutant glioblastoma patient during surgical intervention. The tissue was sectioned and inoculated with autologous glioblastoma cells to establish an ex-vivo personalized brain slice model. To facilitate adequate tumor growth, slices were cultured for seven days with clinically approved BRAF/MEK inhibitors and respective controls. For high-dimensional analysis, the tumor specimen and slices in all treatment conditions were profiled using live cell imaging, spatially resolved transcriptomics (Visium 10X) and single nucleus RNA-sequencing (10X). Advanced computational modeling techniques, encompassing graph neural networks and single-cell deconvolution, were employed to scrutinize the spatio-temporal evolution of the tumor and its microenvironment. RESULTS Live cell imaging showed varied treatment responses in tumor growth during culture. Immunofluorescence and immunohistochemistry analyses (using NeuN, GFAP, and Iba-1 cell markers) indicated no toxic effects from drug combinations on the tumor microenvironment. Single nucleus profiling revealed a tumoral trajectory towards induction of mesenchymal-like transcriptional programs under BRAF/MEK inhibition along with up-regulation of alternative receptor tyrosine kinases such as HER2/3 and EGFR. Spatial projection of single cell trajectories using cytospace confirmed the escape mechanism of BRAF/MEK inhibition due to acquisition of anti-inflammatory myeloid cells in close proximity to the tumor niches and enhancement of mesenchymal transcriptional programs. In line with the experimental set-up, we observed a rapid progression of the tumor und BRAF/MEK inhibition in the patient. CONCLUSION Our reported model system, employing high-dimensional readouts and advanced computational modeling, showcases a promising avenue to address uncertainties in personalized medicine, emphasizing that conventional oncological drivers may necessitate alternative treatment strategies in central nervous system malignancies.

Imaging features to distinguish posterior fossa ependymoma subgroups.

Posterior fossa ependymoma group A (EPN_PFA) and group B (EPN_PFB) can be distinguished by their DNA methylation and give rise to different prognoses. We compared the MRI characteristics of EPN_PFA and EPN_PFB at presentation. Preoperative imaging of 68 patients with posterior fossa ependymoma from two centers was reviewed by three independent readers, blinded for histomolecular grouping. Location, tumor extension, tumor volume, hydrocephalus, calcifications, tissue component, enhancement or diffusion signal, and histopathological data (cellular density, calcifications, necrosis, mitoses, vascularization, and microvascular proliferation) were compared between the groups. Categorical data were compared between groups using Fisher's exact tests, and quantitative data using Mann-Whitney tests. We performed a Benjamini-Hochberg correction of the p values to account for multiple tests. Fifty-six patients were categorized as EPN_PFA and 12 as EPN_PFB, with median ages of 2 and 20years, respectively (p = 0.0008). The median EPN_PFA tumoral volume was larger (57 vs 29 cm3, p = 0.003), with more pronounced hydrocephalus (p = 0.002). EPN_PFA showed an exclusive central position within the 4th ventricle in 61% of patients vs 92% for EPN_PFB (p = 0.01). Intratumor calcifications were found in 93% of EPN_PFA vs 40% of EPN_PFB (p = 0.001). Invasion of the posterior fossa foramina was mostly found for EPN_PFA, particularly the foramina of Luschka (p = 0.0008). EPN_PFA showed whole and homogeneous tumor enhancement in 5% vs 75% of EPN_PFB (p = 0.0008). All mainly cystic tumors were EPN_PFB (p = 0.002). The minimal and maximal relative ADC was slightly lower in EPN_PFA (p = 0.02 and p = 0.01, respectively). Morphological characteristics from imaging differ between posterior fossa ependymoma subtypes and may help to distinguish them preoperatively. This study provides a tool to differentiate between group A and group B ependymomas, which will ultimately allow the therapeutic strategy to be adapted in the early stages of patient management. • Posterior fossa ependymoma subtypes often have different imaging characteristics. • Posterior fossa ependymomas group A are commonly median or lateral tissular calcified masses, with incomplete enhancement, affecting young children and responsible for pronounced hydrocephalus and invasion of the posterior fossa foramina. • Posterior fossa ependymomas group B are commonly median non-calcified masses of adolescents and adults, predominantly cystic, and minimally invasive, with total and homogeneous enhancement.

Novel insight into histological and molecular astrocytoma, IDH-mutant, Grade 4 by the updated WHO classification of central nervous system tumors.

The latest fifth edition of the World Health Organization (WHO) classification of the central nervous system (CNS) tumors (WHO CNS 5 classification) released in 2021 defined astrocytoma, IDH-mutant, Grade 4. However, the understanding of this subtype is still limited. We conducted this study to describe the features of astrocytoma, IDH-mutant, Grade 4 and explored the similarities and differences between histological and molecular subtypes. Patients who underwent surgery from January 2011 to January 2022, classified as astrocytoma, IDH-mutant, Grade 4 were included in this study. Clinical, radiological, histopathological, molecular pathological, and survival data were collected for analysis. Altogether 33 patients with astrocytoma, IDH-mutant, Grade 4 were selected, including 20 with histological and 13 with molecular WHO Grade 4 astrocytoma. Tumor enhancement, intratumoral-necrosis like presentation, larger peritumoral edema, and more explicit tumor margins were frequently observed in histological WHO Grade 4 astrocytoma. Additionally, molecular WHO Grade 4 astrocytoma showed a tendency for relatively longer overall survival, while a statistical significance was not reached (47 vs. 25 months, p = 0.22). TP53, CDK6, and PIK3CA alteration was commonly observed, while PIK3R1 (p = 0.033), Notch1 (p = 0.027), and Mycn (p = 0.027) alterations may affect the overall survival of molecular WHO Grade 4 astrocytomas. Our study scrutinized IDH-mutant, Grade 4 astrocytoma. Therefore, further classification should be considered as the prognosis varied between histological and molecular WHO Grade 4 astrocytomas. Notably, therapies aiming at PIK3R1, Notch 1, and Mycn may be beneficial.

Correlation of tumor-associated macrophage infiltration in glioblastoma with magnetic resonance imaging characteristics: a retrospective cross-sectional study.

Glioblastoma (Gb) is the most common primary malignant tumor of brain with poor prognosis. Immune cells are the main factors affecting the prognosis of Gb, tumor-associated macrophages (TAMs) are the predominant infiltrating immune cell population in the immune microenvironment of Gb. Analyzing the relationship between magnetic resonance imaging (MRI) features and TAMs of Gb, and using imaging features to characterize the infiltration level of TAMs in tumor tissue may provide indicators for clinical decision-making and prognosis evaluation of Gb. Data from 140 in patients with isocitrate dehydrogenase (IDH) wild-type Gb diagnosed via histopathology and molecular diagnosis in the Second Hospital of Lanzhou University from January 2018 to April 2022 were collected in this retrospective, cross-sectional study. MRI images were reviewed for lesion location, cyst, necrosis, hemorrhage, contrast-enhanced T1-weighted MRI signal intensity, average apparent diffusion coefficient (ADCmean), and minimum apparent diffusion coefficient (ADCmin). Immunohistochemical staining with anti-CD163 and anti-CD68 antibodies was employed for macrophage detection. The positive cell percentage was estimated in 9 microscopic fields at 400× magnification per whole-slide image with ImageJ software (National Institutes of Health). Additionally, the relationship between MRI features, molecular, states and the positive CD68 and CD163 expression was analyzed. Our study discovered that the mean or median values of CD68+ and CD163+ TAMs were 7.39% and 14.98%, respectively. There was an obvious correlation between CD163+ TAMs and CD68+ TAMs (r=0.497; P=0.000). CD68+ and CD163+ macrophage infiltration correlated with age at diagnosis in patients with Gb (CD68+: r=0.230, P=0.006; CD163+: r=0.172, P=0.042). The levels of Gb TAM infiltration in different tumor locations varied, with the temporal lobe having the highest CD163+ macrophage and CD68+ macrophage infiltration (18.58% and 9.46%, respectively). CD163+ macrophage infiltration was positively correlated with ADCmean (r=0.208; P=0.014). The infiltration of CD68+ macrophages differed significantly between groups with varying degrees of tumor enhancement (H =4.228; P=0.017). There was a significant difference in CD68+ TAMs and CD163+ TAMs between the wild-type and mutant-type telomerase reverse transcriptase (TERT) types (P=0.004 and P=0.031, respectively). Age, location of the tumor, degree of tumor enhancement, ADC value, and TERT mutation status were associated with macrophage infiltration. These findings may serve as an effective tool for characterizing the tumor microenvironment in patients with Gb.

Deep learning based clinico-radiological model for paediatric brain tumor detection and subtype prediction.

Early diagnosis of paediatric brain tumors significantly improves the outcome. The aim is to study magnetic resonance imaging (MRI) features of paediatric brain tumors and to develop an automated segmentation (AS) tool which could segment and classify tumors using deep learning methods and compare with radiologist assessment. This study included 94 cases, of which 75 were diagnosed cases of ependymoma, medulloblastoma, brainstem glioma, and pilocytic astrocytoma and 19 were normal MRI brain cases. The data was randomized into training data, 64 cases; test data, 21 cases and validation data, 9 cases to devise a deep learning algorithm to segment the paediatric brain tumor. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of the deep learning model were compared with radiologist's findings. Performance evaluation of AS was done based on Dice score and Hausdorff95 distance. Analysis of MRI semantic features was done with necrosis and haemorrhage as predicting features for ependymoma, diffusion restriction and cystic changes were predictors for medulloblastoma. The accuracy of detecting abnormalities was 90%, with a specificity of 100%. Further segmentation of the tumor into enhancing and non-enhancing components was done. The segmentation results for whole tumor (WT), enhancing tumor (ET), and non-enhancing tumor (NET) have been analyzed by Dice score and Hausdorff95 distance. The accuracy of prediction of all MRI features was compared with experienced radiologist's findings. Substantial agreement observed between the classification by model and the radiologist's given classification [K-0.695 (K is Cohen's kappa score for interrater reliability)]. The deep learning model had very high accuracy and specificity for predicting the magnetic resonance (MR) characteristics and close to 80% accuracy in predicting tumor type. This model can serve as a potential tool to make a timely and accurate diagnosis for radiologists not trained in neuroradiology.

AI-based classification of three common malignant tumors in neuro-oncology: A multi-institutional comparison of machine learning and deep learning methods

PurposeTo determine if machine learning (ML) or deep learning (DL) pipelines perform better in AI-based three-class classification of glioblastoma (GBM), intracranial metastatic disease (IMD) and primary CNS lymphoma (PCNSL). MethodologyRetrospective analysis included 502 cases for training (208 GBM, 67 PCNSL and 227 IMD), with external validation on 86 cases (27:27:32). Multiparametric MRI images (T1W, T2W, FLAIR, DWI and T1-CE) were co-registered, resampled, denoised and intensity normalized, followed by semiautomatic 3D segmentation of the enhancing tumor (ET) and peritumoral region (PTR). Model performance was assessed using several ML pipelines and 3D-convolutional neural networks (3D-CNN) using sequence specific masks, as well as combination of masks. All pipelines were trained and evaluated with 5-fold nested cross-validation on internal data followed by external validation using multi-class AUC. ResultsTwo ML models achieved similar performance on test set, one using T2-ET and T2-PTR masks (AUC: 0.885, 95% CI: [0.816, 0.935] and another using T1-CE-ET and FLAIR-PTR mask (AUC: 0.878, CI: [0.804, 0.930]). The best performing DL models achieved an AUC of 0.854, (CI [0.774, 0.914]) on external data using T1-CE-ET and T2-PTR masks, followed by model derived from T1-CE-ET, ADC-ET and FLAIR-PTR masks (AUC: 0.851, CI [0.772, 0.909]). ConclusionBoth ML and DL derived pipelines achieved similar performance. T1-CE mask was used in three of the top four overall models. Additionally, all four models had some mask derived from PTR, either T2WI or FLAIR.

Model observer performance in contrast-enhanced lesions in breast CT: The influence of contrast concentration on detectability.

The use of iodine-based contrast agent for better delineation of tumors in breast CT (bCT) has been shown to be compelling, similar to the tumor enhancement in contrast-enhanced breast MRI. Contrast-enhanced bCT (CE-bCT) is a relatively new tool, and a structured evaluation of different imaging parameters at play has yet to be conducted. In this investigation, data sets of acquired bCT images from 253 patients imaged at our institution were used in concert with simulated mathematically inserted spherical contrast-enhanced lesions to study the role of contrast enhancement on detectability. To quantitatively evaluate the improvement in lesion detectability due to contrast enhancement across lesion diameter, section thickness, view plane, and breast density using a pre-whitened matched filter (PWMF) model observer. The relationship between iodine concentration and Hounsfield units (HU) was measured using spectral modeling. The lesion enhancement from clinical CE-bCT images in 22 patients was evaluated, and the average contrast enhancement (ΔHU) was determined. Mathematically generated spherical mass lesions of varying diameters (1, 3, 5, 9, 11, 15mm) and contrast enhancement levels (0, 0.25, 0.50, 0.75, 1) were inserted at random locations in 253 actual patient bCT datasets. Images with varying thicknesses (0.4-19.8mm) were generated by slice averaging, and the role of view plane (coronal and axial planes) was studied. A PWMF was used to generate receiver operating characteristic (ROC) curves across parameters of lesion diameter, contrast enhancement, section thickness, view plane, and breast density. The area under the ROC curve (AUC) was used as the primary performance metric, generated from over 90,000 simulated lesions. An average 20% improvement (ΔAUC=0.1) in lesion detectability due to contrast enhancement was observed across lesion diameter, section thickness, breast density, and view plane. A larger improvement was observed when stratifying patients based on breast density. For patients with VGF≤40%, detection performance improved up to 20% (until AUC →1), and for patients with denser breasts (VGF>40%), detection performance improved more drastically, ranging from 20% to 80% for 1- and 5-mm lesions. For the 1mm lesion, detection performance raised slightly at the 1.2mm section thickness before falling off as thickness increased. For larger lesions, detection performance was generally unaffected as section thickness increased up until it reached 5.8mm, where performance began to decline. Detection performance was higher in the axial plane compared to the coronal plane for smaller lesions and thicker sections. For emerging diagnostic tools like CE-bCT, it is important to optimize imaging protocols for lesion detection. In this study, we found that intravenous contrast can be used to detect small lesions in dense breasts. Optimal section thickness for detectability has dependencies on breast density and lesion size, therefore, display thickness should be adjusted in real-time using display software. These findings may be useful for the development of CE-bCT as well as other x-ray-based breast imaging modalities.

Prognostic Factors for Survival in Adult Patients with Cerebral Low-Grade Glioma

Background Low-grade gliomas are considered a therapeutic dilemma because of the heterogeneity of their clinical behavior. Patients affected by these tumors are often young and management decisions should put into consideration the potential indolent nature of these tumors and the potential long-term side effects of available therapies. Aim of the Work to analyze the collective data from studies to define prognostic factors for overall survival in adult patients with cerebral low-grade gliomas. Materials and Methods We prepared this systematic review with a careful following of the Cochrane Handbook for Systematic Reviews of Interventions guidelines. We conducted a literature search till December 2020 using PubMed, Scopus, Web of Science, and Cochrane Library. We performed a search for all published articles that evaluated the impact of several factors (patient’s age, KPS, neurological symptoms, histological subtype, pre-op tumor size, tumor crossing the midline, tumor enhancement, extent of resection, radiotherapy, and chemotherapy) on the overall survival (OS) and progression-free survival (PFS) of adult patients with cerebral low-grade glioma (LGG), Studies were independently assessed for risk of bias using (Cochrane Risk of Bias 2.0) for randomized controlled trials, and (Newcastle–Ottawa Scale) for retrospective cohort studies. Results Patient-related factors (age was found significant according to 14 of total 19 studies, KPS according to 7 of total 11 studies, and presence of seizures according to 5 of total 7 studies). Tumor-related factors (tumor subtype was significant according to 10 of total 13 studies, IDH-mutation according to 4 of total 5 studies, 1p/19q codeletion according to 5 of total 5 studies, and tumor enhancement according to 4 of total 7 studies), and treatmentrelated factors (extent of resection was significant according to 18 of total 20 studies, while timing and dose of radiation, combined chemotherapy and radiotherapy failed to show major statistical significance, with better PFS in delayed radiotherapy “till progression”, and slightly better OS in combined chemo, radiotherapy). Conclusion Highly significant factors are extent of tumor resection, patient’s age, pre-op tumor volume and molecular subtype.

Modality preserving U-Net for segmentation of multimodal medical images.

Recent advances in artificial intelligence and digital image processing have inspired the use of deep neural networks for segmentation tasks in multimodal medical imaging. Unlike natural images, multimodal medical images contain much richer information regarding different modal properties and therefore present more challenges for semantic segmentation. However, there is no report on systematic research that integrates multi-scaled and structured analysis of single-modal and multimodal medical images. We propose a deep neural network, named as Modality Preserving U-Net (MPU-Net), for modality-preserving analysis and segmentation of medical targets from multimodal medical images. The proposed MPU-Net consists of a modality preservation encoder (MPE) module that preserves the feature independency among the modalities and a modality fusion decoder (MFD) module that performs a multiscale feature fusion analysis for each modality in order to provide a rich feature representation for the final task. The effectiveness of such a single-modal preservation and multimodal fusion feature extraction approach is verified by multimodal segmentation experiments and an ablation study using brain tumor and prostate datasets from Medical Segmentation Decathlon (MSD). The segmentation experiments demonstrated the superiority of MPU-Net over other methods in the segmentation tasks for multimodal medical images. In the brain tumor segmentation tasks, the Dice scores (DSCs) for the whole tumor (WT), the tumor core (TC) and the enhancing tumor (ET) regions were 89.42%, 86.92%, and 84.59%, respectively. In the meanwhile, the 95% Hausdorff distance (HD95) results were 3.530, 4.899 and 2.555, respectively. In the prostate segmentation tasks, the DSCs for the peripheral zone (PZ) and the transitional zone (TZ) of the prostate were 71.20% and 90.38%, respectively. In the meanwhile, the 95% HD95 results were 6.367 and 4.766, respectively. The ablation study showed that the combination of single-modal preservation and multimodal fusion methods improved the performance of multimodal medical image feature analysis. In the segmentation tasks using brain tumor and prostate datasets, the MPU-Net method has achieved the improved performance in comparison with the conventional methods, indicating its potential application for other segmentation tasks in multimodal medical images.

Preoperative planning and intraoperative real-time navigation with indocyanine green fluorescence in robotic liver surgery

PurposeWe aimed at exploring indocyanine green (ICG) fluorescence wide spectrum of applications in hepatobiliary surgery as can result particularly useful in robotic liver resections (RLR) in order to overcome some technical limitations, increasing safety, and efficacy.MethodsWe describe our experience of 76 RLR performed between March 2020 and December 2022 exploring all the possible applications of pre- and intraoperative ICG administration.ResultsHepatocellular carcinoma and colorectal liver metastases were the most common indications for RLR (34.2% and 26.7% of patients, respectively), and 51.3% of cases were complex resections with high IWATE difficulty scores. ICG was administered preoperatively in 61 patients (80.3%), intraoperatively in 42 patients (55.3%) and in both contexts in 25 patients (32.9%), with no observed adverse events. The most frequent ICG goal was to achieve tumor enhancement (59 patients, 77.6%), with a success rate of 94.9% and the detection of 3 additional malignant lesions. ICG facilitated evaluation of the resection margin for residual tumor and perfusion adequacy in 33.9% and 32.9% of cases, respectively, mandating a resection enlargement in 7.9% of patients. ICG fluorescence allowed the identification of the transection plane through negative staining in the 25% of cases. Vascular and biliary structures were visualized in 21.1% and 9.2% of patients, with a success rate of 81.3% and 85.7%, respectively.ConclusionRLR can benefit from the routine integration of ICG fluoresce evaluation according to each individual patient and condition-specific goals and issues, allowing liver functional assessment, anatomical and vascular evaluation, tumor detection, and resection margins assessment.

Survival prediction of hepatocellular carcinoma by measuring the extracellular volume fraction with single-phase contrast-enhanced dual-energy CT imaging.

This study aimed to investigate the value of quantified extracellular volume fraction (fECV) derived from dual-energy CT (DECT) for predicting the survival outcomes of patients with hepatocellular carcinoma (HCC) after transarterial chemoembolization (TACE). A total of 63 patients with HCC who underwent DECT before treatment were retrospectively included. Virtual monochromatic images (VMI) (70 keV) and iodine density images (IDI) during the equilibrium phase (EP) were generated. The tumor VMI-fECV and IDI-fECV were measured and calculated on the whole tumor (Whole) and maximum enhancement of the tumor (Maximum), respectively. Univariate and multivariate Cox models were used to evaluate the effects of clinical and imaging predictors on overall survival (OS) and progression-free survival (PFS). The correlation between tumor VMI-fECV and IDI-fECV was strong (both p< 0.001). The Bland-Altman plot between VMI-fECV and IDI-fECV showed a bias of 5.16% for the Whole and 6.89% for the Maximum modalities, respectively. Increasing tumor VMI-fECV and IDI-fECV were positively related to the effects on OS and PFS (both p< 0.05). The tumor IDI-fECV-Maximum was the only congruent independent predictor in patients with HCC after TACE in the multivariate analysis on OS (p = 0.000) and PFS (p = 0.028). Patients with higher IDI-fECV-Maximum values had better survival rates above the optimal cutoff values, which were 35.42% for OS and 29.37% for PFS. The quantified fECV determined by the equilibrium-phase contrast-enhanced DECT can potentially predict the survival outcomes of patients with HCC following TACE treatment.

C-Arm Computed Tomographic Image Fusion for Repetitive Transarterial Chemoembolization of Hepatocellular Carcinoma.

The aim of this study was to evaluate the potential implications of fusion imaging with C-arm computed tomography (CACT) scans for repetitive conventional transarterial chemoembolization (cTACE) for hepatocellular carcinoma. Fifty-six cTACE sessions were performed using fusion CACT images from September 2020 to June 2021 in a tertiary referral center, and the data were retrospectively analyzed. Fusion of unenhanced and enhanced CACT images was considered when previously accumulated iodized oil hampered the identification of local tumor progression or intrahepatic distant metastasis (indication A), when a tumor was supplied by multiple arteries with different origins from the aorta and missing tumor enhancement was suspected (indication B), or when iodized oil distribution on immediate post-cTACE CACT images needed to be precisely compared with the pre-cTACE images (indication C). Fusion image quality, initial tumor response, time to local progression (TTLP) of index tumors, and time to progression (TTP) were evaluated. The fusion quality was satisfactory with a mean misregistration distance of 1.4 mm. For the 40 patients with indication A, the initial tumor responses at 3 months were nonviable, equivocal, and viable in 27 (67.5%), 4 (10.0%), and 9 (22.5%) index tumors, respectively. The median TTLP and TTP were 14.8 months and 4.5 months, respectively. For 10 patients with indication B, the median TTLP and TTP were 8.3 months and 2.6 months, respectively. Among the 6 patients with indication C, 2 patients were additionally treated at the same cTACE session after confirming incomplete iodized oil uptake on fusion imaging. Fusion CACT images are useful in patients with hepatocellular carcinoma undergoing repetitive cTACE.