Machine Learning-Assisted Classification of Paraffin-Embedded Brain Tumors with Raman Spectroscopy

Raman spectroscopy (RS) has demonstrated its utility in neurooncological diagnostics, spanning from intraoperative tumor detection to the analysis of tissue samples peri- and postoperatively. In this study, we employed Raman spectroscopy (RS) to monitor alterations in the molecular vibrational characteristics of a broad range of formalin-fixed, paraffin-embedded (FFPE) intracranial neoplasms (including primary brain tumors and meningiomas, as well as brain metastases) and considered specific challenges when employing RS on FFPE tissue during the routine neuropathological workflow. We spectroscopically measured 82 intracranial neoplasms on CaF2 slides (in total, 679 individual measurements) and set up a machine learning framework to classify spectral characteristics by splitting our data into training cohorts and external validation cohorts. The effectiveness of our machine learning algorithms was assessed by using common performance metrics such as AUROC and AUPR values. With our trained random forest algorithms, we distinguished among various types of gliomas and identified the primary origin in cases of brain metastases. Moreover, we spectroscopically diagnosed tumor types by using biopsy fragments of pure necrotic tissue, a task unattainable through conventional light microscopy. In order to address misclassifications and enhance the assessment of our models, we sought out significant Raman bands suitable for tumor identification. Through the validation phase, we affirmed a considerable complexity within the spectroscopic data, potentially arising not only from the biological tissue subjected to a rigorous chemical procedure but also from residual components of the fixation and paraffin-embedding process. The present study demonstrates not only the potential applications but also the constraints of RS as a diagnostic tool in neuropathology, considering the challenges associated with conducting vibrational spectroscopic analysis on formalin-fixed, paraffin-embedded (FFPE) tissue.

Computational Assessment of Spectral Heterogeneity within Fresh Glioblastoma Tissue Using Raman Spectroscopy and Machine Learning Algorithms

Understanding and classifying inherent tumor heterogeneity is a multimodal approach, which can be undertaken at the genetic, biochemical, or morphological level, among others. Optical spectral methods such as Raman spectroscopy aim at rapid and non-destructive tissue analysis, where each spectrum generated reflects the individual molecular composition of an examined spot within a (heterogenous) tissue sample. Using a combination of supervised and unsupervised machine learning methods as well as a solid database of Raman spectra of native glioblastoma samples, we succeed not only in distinguishing explicit tumor areas-vital tumor tissue and necrotic tumor tissue can correctly be predicted with an accuracy of 76%-but also in determining and classifying different spectral entities within the histomorphologically distinct class of vital tumor tissue. Measurements of non-pathological, autoptic brain tissue hereby serve as a healthy control since their respective spectroscopic properties form an individual and reproducible cluster within the spectral heterogeneity of a vital tumor sample. The demonstrated decipherment of a spectral glioblastoma heterogeneity will be valuable, especially in the field of spectroscopically guided surgery to delineate tumor margins and to assist resection control.

From Research to Diagnostic Application of Raman Spectroscopy in Neurosciences: Past and Perspectives

In recent years, Raman spectroscopy has been more and more frequently applied to address research questions in neuroscience. As a non-destructive technique based on inelastic scattering of photons, it can be used for a wide spectrum of applications including neurooncological tumor diagnostics or analysis of misfolded protein aggregates involved in neurodegenerative diseases. Progress in the technical development of this method allows for an increasingly detailed analysis of biological samples and may therefore open new fields of applications. The goal of our review is to provide an introduction into Raman scattering, its practical usage and also commonly associated pitfalls. Furthermore, intraoperative assessment of tumor recurrence using Raman based histology images as well as the search for non-invasive ways of diagnosis in neurodegenerative diseases are discussed. Some of the applications mentioned here may serve as a basis and possibly set the course for a future use of the technique in clinical practice. Covering a broad range of content, this overview can serve not only as a quick and accessible reference tool but also provide more in-depth information on a specific subtopic of interest.

Adult-born granule cells mature through two functionally distinct states

Adult-born granule cells (ABGCs) are involved in certain forms of hippocampus-dependent learning and memory. It has been proposed that young but functionally integrated ABGCs (4-weeks-old) specifically contribute to pattern separation functions of the dentate gyrus due to their heightened excitability, whereas old ABGCs (>8 weeks old) lose these capabilities. Measuring multiple cellular and integrative characteristics of 3- 10-week-old individual ABGCs, we show that ABGCs consist of two functionally distinguishable populations showing highly distinct input integration properties (one group being highly sensitive to narrow input intensity ranges while the other group linearly reports input strength) that are largely independent of the cellular age and maturation stage, suggesting that ‘classmate’ cells (born during the same period) can contribute to the network with fundamentally different functions. Thus, ABGCs provide two temporally overlapping but functionally distinct neuronal cell populations, adding a novel level of complexity to our understanding of how life-long neurogenesis contributes to adult brain function.

DOI:http://dx.doi.org/10.7554/eLife.03104.001

Remembering what happened on different occasions involves a process in the brain called pattern separation, which allows us to separate and distinguish our memories. One part of the brain where pattern separation occurs is called the dentate gyrus, which sits in the hippocampus—the brain region that is in charge of certain forms of learning and memory.

Neurons called granule cells are thought to play a central role in hippocampal pattern separation. These cells, unlike the majority of nerve cells, can form at any time, and those that form in the mature brain are called adult born granule cells (ABGCs). Although it usually takes 10 weeks for these cells to fully mature, they are capable of communicating with each other about 3–4 weeks after being generated. Previously, it had been reported that while young, 4-week-old ABGCs are required for pattern separation, slightly older (8 week old) ABGCs are not.

What intrinsic properties make ABGCs capable of contributing to pattern separation? Is this property defined by the fate (i.e. a predetermined program) of the cell, or by the cell's experiences and activities?

To investigate these questions, Brunner et al. labeled ABGCs with a fluorescent tag when these neurons were born in adult male rats. Then, when the tagged cells were aged between 3 and 10 weeks old, the electrical properties of the labeled cells were measured from thin brain slices.

Brunner et al. found that ABGCs respond to input signals with two different levels of sensitivity. The youngest cells (3–5 weeks old) are exceptionally sensitive to a narrow range of input signal strengths, which is useful for pattern separation. The oldest investigated cells (10 weeks old), on the other hand, respond incrementally to a wide range of different input signal strengths. Under these experimental conditions, the cells changed how they respond to input signals some time between 5 and 9 weeks after being born. However, they either behaved like the youngest or like the oldest cells: no intermediate behavior was seen.

Unexpectedly, the switch is not directly related to the age of the cells: cells born at the same time don't necessarily change behavior at the same time, and cells born at different times may behave similarly. Thus, Brunner et al. suggest that it is the experience of the cells, and not their fate, that determines how they help the dentate gyrus function during the investigated period.

DOI:http://dx.doi.org/10.7554/eLife.03104.002

Imaging neurite development of adult-born granule cells

Neural stem/progenitor cells (NSPCs) generate new neurons throughout life in the mammalian hippocampus. Newborn granule cells mature over several weeks to functionally integrate into the pre-existing neural circuitry. Even though an increasing number of genes that regulate neuronal polarization and neurite extension have been identified, the cellular mechanisms underlying the extension of neurites arising from newborn granule cells remain largely unknown. This is mainly because of the current lack of longitudinal observations of neurite growth within the endogenous niche. Here we used a novel slice culture system of the adult mouse hippocampal formation combined with in vivo retroviral labeling of newborn neurons and longitudinal confocal imaging to analyze the mode and velocity of neurite growth extending from immature granule cells. Using this approach we show that dendritic processes show a linear growth pattern with a speed of 2.19±0.2 μm per hour, revealing a much faster growth dynamic than expected by snapshot-based in vivo time series. Thus, we here identified the growth pattern of neurites extending from newborn neurons within their niche and describe a novel technology that will be useful to monitor neuritic growth in physiological and disease states that are associated with altered dendritic morphology, such as rodent models of epilepsy.

Zif268/egr1 gene controls the selection, maturation and functional integration of adult hippocampal newborn neurons by learning

New neurons are continuously added to the dentate gyrus of the adult mammalian brain. During the critical period of a few weeks after birth when newborn neurons progressively mature, a restricted fraction is competitively selected to survive in an experience-dependent manner, a condition for their contribution to memory processes. The mechanisms that control critical stages of experience-dependent functional incorporation of adult newborn neurons remain largely unknown. Here, we identify a unique transcriptional regulator of the functional integration of newborn neurons, the inducible immediate early gene zif268/egr1. We show that newborn neurons in zif268-KO mice undergo accelerated death during the critical period of 2-3 wk around their birth and exhibit deficient neurochemical and morphological maturation, including reduced GluR1 expression, increased NKCC1/KCC2b chloride cotransporter ratio, altered dendritic development, and marked spine growth defect. Investigating responsiveness of newborn neurons to activity-dependent expression of zif268 in learning, we demonstrate that in the absence of zif268, training in a spatial learning task during this critical period fails to recruit newborn neurons and promote their survival, leading to impaired long-term memory. This study reveals a previously unknown mechanism for the control of the selection, functional maturation, and experience-dependent recruitment of dentate gyrus newborn neurons that depends on the inducible immediate early gene zif268, processes that are critical for their contribution to hippocampal-dependent long-term memory.