Nature Communications 8: Article number: 14356 (2017); Published: 16 February 2017; Updated: 18 April 2017 The original version of this Article contained an error in the spelling of the author Carlos S. Moreno, which was incorrectly given as Carlos Moreno. This has now been corrected in both the PDFand HTML versions of the Article.
Abstract Diffuse glioma is characterized by a poor prognosis and a universal resistance to therapy, though the evolutionary processes behind this resistance remain unclear. The Glioma Longitudinal Analysis (GLASS) Consortium has previously demonstrated that therapy-induced selective pressures shape the genetic evolution of glioma in a stochastic manner. However, single-cell studies have revealed that malignant glioma cells are highly plastic and transition their cell state in response to diverse challenges, including changes in the microenvironment and the administration of standard-of-care therapy. To interrogate the factors driving therapy resistance in diffuse glioma, we collected and analyzed RNA- and/or DNA-sequencing data from temporally separated tumor pairs of over 300 adult patients with IDH-wild-type or IDH-mutant glioma. In a subset of these tumor pairs, we additionally performed multiplex immunofluorescence to capture the spatial relationship between tumor cells and their microenvironment. Recurrent tumors exhibited diverse changes that were attributable to changes in histological features, somatic alterations, and microenvironment interactions. IDH-wild-type tumors overall were more invasive at recurrence and exhibited increased expression of neuronal signaling programs that reflected a possible role for neuronal interactions in promoting glioma progression. In contrast, recurrent IDH-mutant tumors exhibited a significant increase in proliferative expression programs that correlated with discrete genetic changes. Hypermutation and acquired CDKN2A homozygous deletions associated with an increase in proliferating stem-like malignant cells at recurrence in both glioma subtypes, reflecting active tumor expansion. A transition to the mesenchymal phenotype was associated with the presence of a specific myeloid cell state defined by unique ligand-receptor interactions with malignant cells, providing opportunities to target this transition through therapy. Collectively, our results uncover recurrence-associated changes in genetics and the microenvironment that can be targeted to shape disease progression following initial diagnosis. Citation Format: Frederick S. Varn, Kevin C. Johnson, Jan Martinek, Jason T. Huse, MacLean P. Nasrallah, Pieter Wesseling, Lee A. Cooper, Tathiane M. Malta, Taylor E. Wade, Thais S. Sabedot, Daniel J. Brat, Peter V. Gould, Adelheid Wöehrer, Kenneth Aldape, Azzam Ismail, Floris P. Barthel, Hoon Kim, Emre Kocakavuk, Nazia Ahmed, Kieron White, Santhosh Sivajothi, Indrani Datta, Jill S. Barnholtz-Sloan, Spyridon Bakas, Fulvio D'Angelo, Hui K. Gan, Luciano Garofano, Mustafa Khasraw, Simona Migliozzi, D. Ryan Ormond, Sun Ha Paek, Erwin G. Van Meir, Annemiek M. Walenkamp, Colin Watts, Michael Weller, Tobias Weiss, Karolina Palucka, Lucy F. Stead, Laila M. Poisson, Houtan Noushmehr, Antonio Iavarone, Roel G. Verhaak, The GLASS Consortium. Longitudinal analysis of diffuse glioma reveals cell state dynamics at recurrence associated with changes in genetics and the microenvironment [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2168.
Ensembled machine learning paradigms enable base learners to provide more accurate predictions than a standard approach using a single learner. Though the ensemble learning decreases variance or bias, improving predictions, limited literatures have been reported with an active learning strategy narrowing uncertainty in prediction. We present an ensemble based active learning approach for breast cancer detection, averaging predictions from the start of the art machine learning models on histopathology images. We demonstrate that the ensemble based active learning approach outperforms other approaches on breast cancer detection.
Glioblastoma is the most common primary adult brain tumor, with a grim prognosis - median survival of 12-18 months following treatment, and 4 months otherwise. Glioblastoma is widely infiltrative in the cerebral hemispheres and well-defined by heterogeneous molecular and micro-environmental histopathologic profiles, which pose a major obstacle in treatment. Correctly diagnosing these tumors and assessing their heterogeneity is crucial for choosing the precise treatment and potentially enhancing patient survival rates. In the gold-standard histopathology-based approach to tumor diagnosis, detecting various morpho-pathological features of distinct histology throughout digitized tissue sections is crucial. Such "features" include the presence of cellular tumor, geographic necrosis, pseudopalisading necrosis, areas abundant in microvascular proliferation, infiltration into the cortex, wide extension in subcortical white matter, leptomeningeal infiltration, regions dense with macrophages, and the presence of perivascular or scattered lymphocytes. With these features in mind and building upon the main aim of the BraTS Cluster of Challenges https://www.synapse.org/brats2024, the goal of the BraTS-Path challenge is to provide a systematically prepared comprehensive dataset and a benchmarking environment to develop and fairly compare deep-learning models capable of identifying tumor sub-regions of distinct histologic profile. These models aim to further our understanding of the disease and assist in the diagnosis and grading of conditions in a consistent manner.
The diagnosis of diffuse gliomas requires the careful inspection of large amounts of visual data. Identifying tissue regions that inform diagnosis is a cumbersome task for human reviewers and is a process prone to inter-reader variability. In this paper we present an automatic method for identifying critical diagnostic regions within whole-slide microscopy images of gliomas. We frame the problem of critical region identification as a texture-based content retrieval task in the sense that each image is represented by a set of texture features. Both linear and nonlinear dimensionality reduction techniques are utilized to explore the intrinsic dimensionality of the feature space where images are classified by classification and regression trees with performances improved by a newly extended multi-class gentle boosting (MCGB) mechanism. The proposed method is demonstrated on 1200 sample regions using a five-fold cross validation, achieving a 96.25% classification accuracy.
