Abstract
AI-based methods have shown great promise in a variety of biomedical research fields, including neurooncologic imaging. For example, machine learning methods have offered informative ...predictions of overall survival (OS) and progression-free survival (PFS), differentiation between pseudoprogression (PsP) and progressive disease (PD), and estimation of mutational status from imaging data. Despite their promise, AI, and especially the emerging deep learning (DL) methods, are challenged by several factors, including imaging heterogeneity across scanners and lack of sufficiently large and diverse training datasets, which limits their reproducibility and general acceptance. These challenges prompted the development of the ReSPOND (Radiomics Signatures for PrecisiON Diagnostics) consortium on glioblastoma, a growing effort to bring together a community of researchers sharing imaging, demographic, clinical and (currently) limited molecular data in order to address the following aims: 1) pool and harmonize data across diverse hospitals and patient populations worldwide; 2) derive robust and generalizable AI models for prediction of (initially) OS, PFS, PsP vs. PD, and recurrence; 3) test these predictive models across multiple sites. In its first phase, ReSPOND aims to pool together approximately 3,000MRI scans (from 10institutions plus TCIA), along with demographics, KPS, and (for a subset) MGMT/IDH1 status. We present initial results testing the generalization of a previously trained model of OS on 505Penn datasets to 2independent cohorts from Case Western Reserve University and University Hospitals (N=44), and Penn (N=67). The results indicate good generalization, with correlation coefficients between OS/predicted-OS between 0.25 to 0.5, depending on variable availability, which is comparable to cross-validated accuracy previously obtained from the training set itself (N=505). Additional preliminary studies evaluating prediction of future recurrence from baseline pre-operative scans in de novo patients (Penn model applied to CWR) indicated potential for guiding targeted dose escalation and supra-total resection (excellent predictions in 6/12 patients, modest in 1/12, and poor in 5/12).
Abstract
Epidermal Growth Factor Receptor (EGFR) missense-mutations in glioblastoma (GBM) typically occur within the extracellular domain (ECD) and, to-date, have not shown a clinical impact. These ...ECD mutations, while not as common as wild-type amplification or EGFRvIII, do represent a distinct patient population in need of investigation. This study investigates the oncogenic effect of the most common EGFR ECD missense-mutations (R108K, A289D/T/V, and G598V) in a retrospective cohort of 260 de novo GBM patients from the University of Pennsylvania. Extrapolation of each individual missense-mutation revealed a significant reduction in patient survival for the EGFRA289D/T/V mutants (p=0.028). This effect was confirmed with a follow-up cohort of 111 patients. The Cancer Imaging Phenomics Toolkit (CaPTk – www.cbica.upenn.edu/captk) was used to extract 2104 quantitative imaging phenomic (QIP) features, from multiparametric magnetic resonance imaging, across the various tumor sub-regions comprising the enhancing and non-enhancing tumor core, as well as the peritumoral edematous/invaded tissue. Multivariate machine learning techniques integrated these features, revealing radiographic signatures suggestive of increased invasion and proliferation for patients bearing the A289 mutation. The underlying mechanism of EGFRA289V in tumor growth was examined using engineered U87 glioma cells and patient-derived HK281 glioma spheres simulating expression of wild-type-EGFR or EGFRA289V. Corroborating our findings in patients, mice bearing intracranial tumors with EGFRA289V mutations revealed significantly worse survival accompanied by highly invasive tumors. Cells expressing EGFRA289V yielded a highly active EGFR/ERK signaling pathway, resulting in increased expression and functionality of the matrix metalloproteinase MMP1, which can be attenuated by the use of ERK pathway inhibitors. Collectively, the findings of this study highlight a highly invasive and proliferative phenotype associated with the EGFRA289V missense-mutation. Moreover, given the tumor-specific and extracellular nature of the mutation, we postulate that it may be amenable to targeted therapy. Key data presented is from a manuscript with preliminary acceptance to Cancer Cell. *equal contribution
To construct a multi-institutional radiomic model that supports upfront prediction of progression-free survival (PFS) and recurrence pattern (RP) in patients diagnosed with glioblastoma multiforme ...(GBM) at the time of initial diagnosis.
We retrospectively identified data for patients with newly diagnosed GBM from two institutions (institution 1, n = 65; institution 2, n = 15) who underwent gross total resection followed by standard adjuvant chemoradiation therapy, with pathologically confirmed recurrence, sufficient follow-up magnetic resonance imaging (MRI) scans to reliably determine PFS, and available presurgical multiparametric MRI (MP-MRI). The advanced software suite Cancer Imaging Phenomics Toolkit (CaPTk) was leveraged to analyze standard clinical brain MP-MRI scans. A rich set of imaging features was extracted from the MP-MRI scans acquired before the initial resection and was integrated into two distinct imaging signatures for predicting mean shorter or longer PFS and near or distant RP. The predictive signatures for PFS and RP were evaluated on the basis of different classification schemes: single-institutional analysis, multi-institutional analysis with random partitioning of the data into discovery and replication cohorts, and multi-institutional assessment with data from institution 1 as the discovery cohort and data from institution 2 as the replication cohort.
These predictors achieved cross-validated classification performance (ie, area under the receiver operating characteristic curve) of 0.88 (single-institution analysis) and 0.82 to 0.83 (multi-institution analysis) for prediction of PFS and 0.88 (single-institution analysis) and 0.56 to 0.71 (multi-institution analysis) for prediction of RP.
Imaging signatures of presurgical MP-MRI scans reveal relatively high predictability of time and location of GBM recurrence, subject to the patients receiving standard first-line chemoradiation therapy. Through its graphical user interface, CaPTk offers easy accessibility to advanced computational algorithms for deriving imaging signatures predictive of clinical outcome and could similarly be used for a variety of radiomic and radiogenomic analyses.
It has become increasingly common to incorporate adjuvant chemotherapy with radiotherapy in the treatment of resected anaplastic astrocytoma based on results from recent phase II/III randomized ...trials. However, whether or not combined chemoradiotherapy is associated with improved survival outcome in patients who undergo "biopsy only" is less clear.
The US National Cancer Database was used to identify patients with histologically confirmed, biopsy-only anaplastic astrocytoma who received either radiotherapy alone or combined chemoradiotherapy from 2006 through 2014.
In total, 1896 patients with biopsy-only anaplastic astrocytoma were included, among whom 363 (19.1%) received radiotherapy alone and 1533 (80.9%) received combined chemoradiotherapy. The median age at diagnosis was 60 years. Combined chemoradiotherapy was associated with a significant survival benefit when compared with radiotherapy alone on univariable analysis (median, 13.2
5.6 months; hazard ratio HR, 0.57; 95% confidence interval CI, 0.50-0.65;
< 0.001) and on multivariable analysis (HR, 0.62; 95% CI, 0.55-0.71;
< 0.001). A significant survival benefit for combined chemoradiotherapy persisted in a propensity score-matched analysis (HR, 0.67; 95% CI, 0.56-0.78;
<0.001).
Our results suggest that combined chemoradiotherapy may be associated with significantly improved survival over radiotherapy alone in patients with anaplastic astrocytoma who undergo biopsy only.
Histopathologic evaluation provides clinicians with vital information for accurate quantification of disease. Hematoxylin and Eosin slide staining is the gold standard in histopathology, revealing ...tissue morphology, structure, and cellular composition. However, variation in materials, equipment, and staining protocols, make histology slide staining rife with color nonconformity. Although, pathologists compensate for stain color variations, these disparities introduce inaccuracies in automated whole slide image (WSI) analysis, degrading neural network model generalization and accentuating data domain shift. Digital pathology state-of-the-art stain color normalization methods employ a single WSI as color reference, but selecting a single WSI representative of a WSI-cohort is infeasible, inadvertently introducing a normalization bias. Goal: We demonstrate a composite or aggregate of H&E density histograms and stain vectors obtained from a randomly selected population of WSIs (WSI-Cohort subset) are representative of a WSI-cohort, therefore suitable as a normalization reference. Methods: In a High-Performance Computing (HPC) environment, this study utilized 1864 Ivy GAP WSIs as a WSI-cohort. First, 200 WSI-cohorts subsets varying in size from 1 to 200 WSI-pairs were built using randomly selected WSIs. Then, the WSI-pairs mean Wasserstein Distances and WSI-cohort subsets standard deviations were calculated. Consequently, the optimal WSI-cohort subset size was approximated using the Pareto Principle 80/20 rule. Lastly, the complete WSI-cohort was color normalized using the structure-preserving color normalization algorithm and the optimal WSI-cohort subset histogram and stain vectors aggregates. Results: We show a WSI-cohort subset aggregates are representative of a WSI-cohort through WSI-cohort CIELAB color space intensity channel swiftly convergence, as a result of the law of large numbers theorem and shown as a power law distribution. In addition, we show stain color normalization at the optimal WSI-cohort subset size (Pareto principle 80/20 rule) and corresponding CIELAB color space intensity channel convergence through stain normalization permutations. The latter three-fold: a) Quantitatively, using 500 WSI-cohorts, b) Quantitatively, using 8100 WSI-regions of interest and c) Qualitative, using 30 Cellular Tumor patch stain normalization permutations. Conclusions: WSI-cohort subset aggregate based stain color normalization is a solution for reproducible, accurate, and efficient WSI-cohort normalization, increasing computational pathology robustness and results integrity
Abstract
PURPOSE
Differentiation of true tumor progression from pseudoprogression (PsP) is a major unmet need in patients with glioblastoma (GBM). 18FFluciclovine is a synthetic amino acid PET ...radiotracer that is FDA approved in the setting of biochemical recurrence in prostate cancer. The aim of this study was to assess the value of 18FFluciclovine PET in differentiation of true tumor progression and PsP in post-treatment of glioblastoma.
METHODS
15 patients with GBM with new contrast-enhancing lesions or lesions showing increased enhancement (>25% increase) on standard MRI after completion of radiation underwent 60-minutes dynamic 18FFluciclovine PET imaging. Patients subsequently (within 1 week) underwent resection of the enhancing lesion and the tumor percentage vs treatment-related changes were quantified on histopathology. Patients were considered true tumor progression if tumor represented ≥ 50% of the resected specimen and considered PsP if treatment-related changes represented ≥70% of the resected specimen. Summed 30- to 40-minute post-injection PET images were used to measure SUVpeak, SUVmax, and 50% threshold SUVmean.
RESULTS
10 patients with true tumor progression and 5 patients with PsP were included. Patients who demonstrated true tumor progression had significantly higher SUVpeak compared to patients with PsP (5.3±1.4 vs 3.1± 0.9, p=0.002, AUC=0.92, p<0.0001). SUVpeak cut-off of 3.5 provided 100% sensitivity, 80% specificity and 93% accuracy for differentiation of true tumor progression from PsP. There was a moderate to strong correlation between SUVpeak and tumor percentage on histopathology (Rho= 0.68, p=0.004). Alternative SUV measures had similar performance.
DISCUSSION
Our preliminary results indicated that 18FFluciclovine PET imaging can accurately differentiate true tumor progression from PsP. Further studies are required to confirm these promising early results and determine the optimal criteria for interpreting 18FFluciclovine PET to distinguish PsP from true tumor progression.
Elevated levels of choline are generally emphasized as marker of increased cellularity and cell membrane turnover in gliomas. In this study, we investigated the incidence rate of lack of ...choline/creatine and choline/water elevation in a population of grade I–III gliomas. A cohort of 41 patients with histopathologically confirmed gliomas underwent multi-voxel proton magnetic resonance spectroscopy on a 3 T magnetic resonance system prior to treatment. Peak areas for choline and myoinositol were measured from all voxels that exhibited hyperintensity on fluid-attenuated inversion recovery images and were normalized to creatine and unsuppressed water from each voxel. The average metabolite/creatine and metabolite/water ratios from these voxels were then computed. Similarly, average metabolite ratios were computed from normal brain parenchyma. Gliomas were considered for lack of choline elevation when choline/creatine and choline/water ratios from neoplastic regions were less than those from normal brain parenchyma regions. Six of 41 (14.6%) grade I–III gliomas showed lack of elevation for choline/creatine and choline/water ratios compared to normal brain parenchyma. Four of these six gliomas also demonstrated elevated levels of myoinositol/creatine ratio. All other gliomas (n = 35) had elevated choline levels from neoplastic regions relative to normal parenchyma. The sensitivity of choline/creatine or choline/water in determining a grade I–III glioma was 85.4%. These findings suggest that a lack of choline/creatine or choline/water elevation may be seen in some gliomas and low choline levels should not prevent us from considering the possibility of a grade I–III glioma.
Abstract
Differentiation of true tumor progression from treatment-related effects remains a major unmet need in caring for patients with glioblastoma. Here, we report how the intraoperative ...combination of MRI with18F-fluciclovine PET guided surgical sampling in 2 patients with recurrent glioblastoma.18F-Fluciclovine PET is FDA approved for use in prostate cancer and carries an orphan drug designation in glioma. To investigate its utility in recurrent glioblastoma, we fused PET and MRI images using 2 different surgical navigation systems and performed targeted stereotactic biopsies from the areas of high (“hot”) and low (“cold”) radiotracer uptake. Concordant histopathologic and imaging findings suggest that a combined18F-fluciclovine PET-MRI–guided approach can guide neurosurgical resection of viable recurrent glioblastoma in the background of treatment-related effects, which can otherwise look similar on MRI.