Glial Tumors Classification Essay

INTRODUCTION

Intraventricular tumors represent a subgroup of intracranial lesions with typical and unique features, which may be considered apart from the classical subdivisions of tumors into intra- and extra-axial lesions(1). In spite of the fact that such lesions are easy to visualize, the differential diagnosis may be difficult without the knowledge on the types of tissues that give rise to such tumors(2).

The ventricles are surrounded by a layer of ependymal cells and a subependymal plate formed by glial cells. Such layers give origin to ependymomas, subependymomas and subependymal giant cell astrocytomas. Such a lining and the septum pellucidum that is located between the corpus callosum and the fornix, separating the lateral ventricles, also give origin to central neurocytoma, a unique glial neuronal tumor of the ventricular systems(2).

The choroid plexus is the most vascularized portion of the ventricular system and produces the cerebrospinal fluid. Primary neoplasias of this tissue are highly vascularized and are commonly associated with hydrocephalus due increased production of cerebrospinal fluid. Such lesions have a benign presentation (choroid plexus papillomas) and, less frequently, a malignant presentation (choroid plexus carcinomas). Tumors such as meningiomas and metastases may also occur in this location.

Masses are more frequently found in the posterior portion of the lateral ventricles(2), but their location may vary according to the type of tumor. Choroid plexus papillomas occur mainly in children, with predilection for the lateral ventricles in this age range, while in adults it usually is more frequently found in the fourth ventricle. Ependymomas are most frequent in the posterior fossa in children, and in adults they are generally supratentorial.

Many times, inflammatory/infectious lesions are observed within the ventricular system, and among them neurocysticercosis is very common in Brazil. Other less frequent conditions, such as histoplasmosis, may also be observed.

In the present essay, the authors have gathered images obtained along the last 15 years at the Radiology Service of Hospital de Clínicas - Universidade Estadual de Campinas. The study was approved by the Committee for Ethics in Research of the Institution.

RADIOLOGICAL FINDINGS

Ependymoma

Between 50% and 75% of ependymomas occur in the posterior fossa(1). Among those occurring in the intraventricular region, 58% originate in the fourth ventricle, while the other 42% are located in the lateral and third ventricles(2). Such tumors develop as a relatively soft mass, filling the ventricular lumen and molding into the ventricular cavity(1).

At non-contrast-enhanced computed tomography (CT), intraventricular ependymomas are usually isodense, with areas of calcifications (40-80% of cases). Occasionally, intratumoral hemorrhages may produce a blood-fluid level. At contrast-enhanced CT, there is heterogeneous contrast enhancement.(2-4)

At magnetic resonance imaging (MRI) (Figure 1), ependymoma may appear as either a solid tumor or as a mixed solid cystic tumor. The solid portion of the lesion presents hypo or isosignal at T1-weighted sequences, and hypersignal at T2-weighted sequences, while the cystic portion presents a signal similar to the signal of the cerebrospinal fluid at T1- and T2-weighted sequences. The main feature of such tumors is signal heterogeneity. Areas of spontaneous hypersignal corresponding to hemorrhage may be observed at T1-weighted sequences. At susceptibility-weighted imaging, foci of hyposignal are commonly observed, corresponding to calcifications or hematic products. After paramagnetic contrast agent injection, a generally heterogeneous enhancement of this tumor is observed(1,3). MRI is considered the modality of choice to evaluate such lesions(2).

Figure 1 Male, 53-year-old patient. Sagittal MRI T1-weighted sequence (A) shows a plastic lesion located in the fourth ventricle, extending down to the Luschka’s, Magendie’s and magnum foramens, with areas of hypersignal (hemorrhage) and contrast uptake (B,C,D). At T2-weighted sequence (E,F), the lesion is hyperintense, with foci of hyposignal (hemorrhage/small calcifications). Anatomopathological analysis revealed the presence of an ependymoma. 

Low-grade astrocytoma

The most common location of such tumors is in the temporal and frontal lobes. Intraventricular low-grade astrocytomas are located in the frontal horn, third ventricle, atrium and fourth ventricle. They may form focal masses with regular margins(1).

At CT, astrocytomas are visualized as a hypodense, solidcystic mass(5). Such lesions are hypovascular tumors and usually present less intense enhancement than other intraventricular masses(6). At MRI (Figure 2), astrocytomas are homogeneous, presenting iso- or hyposignal at T1-weighted sequences and hypersignal at T2-weighted sequences. Cysts may be observed within the lesion. Contrast-enhancement is variable(1).

Figure 2 Male, 3-year-old patient. Axial (A) and sagittal (B) MRI T1-weighted sequences show the presence of a heterogeneous, solid-cystic lesion located in the fourth ventricle, with isosignal and some foci of hypersignal at T1-weighted sequence (hemorrhage). Axial (C) and sagittal (D) sections at contrast-enhanced T1-weighted sequences show intense contrast uptake in the solid component of the lesion. At T2-weighted sequence (E), the solid component shows predominant hypersignal, with some foci of hyposignal (hemorrhage). Biopsy revealed pilocytic astrocytoma. 

Central neurocytoma

This tumor is frequently located in the ventricular system, filling the lateral ventricle from the foramen of Monro. At CT, the tumor presents in the form of either a polylobulated or round-shaped lesion that may be spontaneously hyperdense, isodense or mixed(3,7). They are seen attached to the septum pellucidum, and may extend to both lateral ventricles(1,8). Subtle or moderate enhancement is observed after contrast agent injection. Calcifications and cysts may be present(8).

At MRI (Figure 3), the typical finding is that of an intraventricular mass, frequently located in the foramen of Monro, in contact with the septum pellucidum. The tumor is frequently isointense in relation to the parenchyma at T1weighted sequence, and subtly hypointense at T2-weighted sequence. Cystic zones may be observed(8). The lesion appearance is spongy(3). After contrast agent injection, subtle or moderate enhancement of the lesion is observed(7).

Figure 3 Male, 33-year-old patient. Sagittal MRI T1-weighted sequence (A) presents a heterogeneous, solid lesion with isosignal/areas of hyposignal (cystic areas) located in the right lateral ventricle, adjacent to the septum pellucidum, extending toward the foramen of Monro and left lateral ventricle, determining hydrocephalus. At contrast-enhanced T1-weighted sequence (B,C) a subtle enhancement is observed. On the axial and coronal sections of T2-weighted sequences (D,E), the lesion presents isosignal/hypersignal. Biopsy revealed the presence of a central neurocytoma. 

Ganglioglioma

Gangliomas are benign tumors affecting principally children and young adults, most frequently located in the temporal lobes. Exclusively intraventricular location is rarely observed. MRI findings are nonspecific. Most commonly, the mass presents iso- or hyposignal on T1-weighted images, and iso- or hypersignal on T2-weighted images. Peripheral gross ("bizarre") calcifications may be observed. The pattern of enhancement of the mass by the paramagnetic contrast agent ranges from none to intense(1) (Figure 4).

Figure 4 Female, 10-year-old patient. Axial MRI T1-weighted sequences (A) demonstrate the presence of a solid-cystic lesion located in the right lateral ventricle, with solid component presenting isosignal and cystic component with hyposignal. Also, an isointense lesion is observed in the optic chiasm (arrow on B) at axial T1-weighted image. After contrast injection (D), intense, but heterogeneous uptake is observed. Axial FLAIR sequence (C) shows a lesion with hyperintense signal, and at T2- weighted sequence (E) the lesion presents with heterogeneous, mixed signal, with internal hydrated foci. There is a lesion with similar characteristics in the optic chiasm (B,D – arrows). Anatomopathological analysis was conclusive for ganglioglioma in both lesions. 

Choroid plexus papilloma

Amongst intracranial tumors, choroid plexus mass lesions are rare, representing about 0.4-0.6% of cases in patients of all ages. Such tumors occur predominantly in the first decade of life (38%), especially in the first two years. Papillomas are the most common choroid plexus mass lesions in children, principally before the fifth year of life and predominantly located in the lateral ventricles. The second most frequent location is the fourth ventricle, most commonly in adult individuals, and rarely in the third ventricle(1,9).

At CT, papillomas are isodense or slightly hyperdense to the gray matter. Calcifications are found in 25% of cases. The contours are lobulated, with slightly irregular margins(3). The contrast enhancement is intense and slightly heterogeneous. At MRI, a large lobulated mass isointense to the white matter is observed at T1-weighted sequences. Calcifications and intralesional flow voids may be observed. Intense enhancement is observed after intravenous contrast injection(1,9) (Figure 5).

Figure 5 Male, 10-year-old patient. Coronal MRI T1-weighted sequence (A) demonstrates a solid, microlobulated, cauliflower-like lesion located in the left lateral ventricle, with isosignal at T1-weighet sequence, with foci of hyposignal (calcifications or vessels). At contrast-enhanced T1-weighted sequence (B,C,D) intense contrast uptake is observed. Axial T2-weighted sequence (E) shows isosignal with foci of hyposignal. Hydrocephalus is present. Anatomopathological analysis revealed choroid plexus papilloma. 

Primitive neuroectodermal tumors

Primitive neuroectodermal tumor (PNET) is a generic name for the classification that includes medulloblastomas and histologically indistinguishable neoplasms located in the central nervous system, at other sites than the cerebellum. Medulloblastoma is a type of PNET that most frequently affects the central nervous system, particularly in the first decade of life. It is located in the posterior fossa, typically filling the fourth ventricle, with about 67%-93% being located in the cerebellar vermis(1).

At CT, medulloblastomas are seen as a spontaneously hyperdense lesion, and evidence of vasogenic edema may be found(5). In children, MRI demonstrates a usually intraventricular mass located on the median or paramedian line, with relatively homogeneous signal intensity. Usually, isosignal is observed on T1-weighted sequences, and iso- or hyposignal on T2-weighted sequences, besides typical diffusion restriction and intense enhancement following contrast agent injection. In adult individuals, the spectrum of signal intensity is similar to that of children, frequently presenting isosignal on T2-weighted sequences(1).

Meningioma

It is the most common benign neoplasm of the central nervous system(1), representing 33% of all (asymptomatic) intracranial incidentalomas. In adult individuals, intraventricular meningiomas are amongst the most common tumors found in the lateral ventricles.

CT reveals a sharply delineated, lobulated mass with periventricular edema(4,5). Focal or diffuse ventricular dilatation may be present, depending on the degree of obstruction to the drainage of cerebrospinal fluid (CSF). Calcifications are commonly found (in 50% of cases)(2). At MRI (Figure 7), the lesion may be seen as an iso- to hypointense mass on T1-weighted sequences and, in general isointense on T2weighted sequences(2), with intense contrast-enhancement(5).

Figure 6 Male, 10-year-old patient. Sagittal MRI (A) T1-weighted sequence shows heterogeneous lesion with hyposignal located in the fourth ventricle, suprasellar cistern, pineal gland, septum pellucidum, lateral ventricle and hypophyseal stalk. At contrast-enhanced T1-weighted sequence (B,C) there is contrast uptake by the lesion and also an intraparenchymal mass or liquoric lesions in the right fronto-temporo-insular region (intraventricular metastasis from primary central nervous system lesion). Histopathological analysis revealed PNET dissemination. 

Figure 7 Female, 67-year-old patient. Pre-contrast axial CT (A) and post-contrast axial CT (B,C) demonstrate spontaneously hyperdense, lobulated lesion with intense contrast uptake. Axial MRI T1-weighted sequence (D) shows expansile lesion with isosignal located in the posterior and inferior horns of the left lateral ventricle infiltrating the adjacent parenchyma, characterized by extensive nodularity with intense contrast enhancement (F,G). At T2-weighted sequence (E) the lesion presents hyposignal. The patient underwent surgery for resection of intraventricular meningioma in the left lateral ventricle (images not available), which revealed the presence of an atypical meningioma with frequent figures of mitosis positive for the Ki-67 marker in 20% of the nuclei, which indicates a high degree of cell proliferation. The images refer to the lesion recurrence. 

Epidermoid tumor

Intracranial epidermoid tumors are congenital inclusion cysts corresponding to 0.2-1.8% of primary intracranial neoplasias(10) (it is the most common congenital tumor of the nervous central system(1)). Its most common site is the cerebellopontine angle cistern(10) and the incidence peak is at the fourth decade of life(1). Such tumors develop within the ventricular spaces, surrounding adjacent vessels and nerves(10). The lesion margins are generally irregular and associated obstructive hydrocephalus may be observed. Calcifications are present on the lesion borders in 25% of cases(1).

At CT, the typical appearance is that of an extra-axial hypodense mass without venous contrast uptake. At MRI such tumors may presents with iso- or subtle hypersignal to the CSF on T1- and T2-weighted sequences (Figure 8). The main differential diagnosis is made with arachnoid cyst, generally by means of FLAIR and diffusion-weighted sequences. The arachnoid cyst follows the CSF signal intensity at all sequences, while epidermoid tumors are not hypointense at FLAIR sequences, showing areas of hypersignal to CSF. Contrarily to arachnoid cysts, epidermoid tumors typically present diffusion restriction(10).

Figure 8 Female, 60-year-old patient. Sagittal MRI T1-weighted sequence (A) demonstrates cystic lesion with lobulated margins, irregular contours and low signal intensity. At contrast-enhanced T1-weighted sequence (B), the lesion presented enhancement, and at T2-weighted sequence (C) was hyperintense. Heterogeneous signal is observed at FLAIR sequence (D), with hypersignal at diffusion-weighted image and hyposignal at ADC mapping (diffusion restriction) (E,F). The lesion is located in the fourth ventricle. Anatomopathological analysis revealed epidermoid tumor. 

CONCLUSION

Imaging findings of intraventricular tumors are quite variable, probably because of the diversity of tissues in the central nervous system which give origin to such lesions. Thus, the study of the skull, particularly by MRI, playa a relevant role in the attempt to define the differential diagnoses based on their location and characteristics of signal at the different sequences, as well as in the detection of hemorrhagic elements and calcifications. It is the responsibility of the radiologist to know the main imaging findings of each lesion in the attempt to narrow the differential diagnosis.

1. Introduction

Brain tumors can be cancerous or non-cancerous. In 2016, the World Health Organization (WHO) reclassified tumor types of the central nervous system into a more accurate system of brain tumor classification by integrating molecular information with the traditional histology markers [1,2]. Some of the most common brain tumors are gliomas, and they form from supportive cells in the brain called glial cells. There are different types of glial tumors such as: astrocytoma, oligodendroglioma and glioblastoma. The astrocytoma is the most common type of glioma and is formed by star-shaped cells known as astrocytes [3]. The overall classification of astrocytomas by the World health Organization is into four grades according to how abnormal the tumor cells look under the microscope and their rate of growth (I–IV) [4]. Grade I cancerous growths can be usually cured by surgical resection, as they have little proliferative capability. Grade II cancerous growths have a patient survival average of 5–15 years, as they have a relatively small proliferative potential. Grade III cancerous growths have greater malignancy, and they exhibit nuclear atypia and brisk mitotic capacity. Grade IV gliomas, which are known also as Glioblastoma Multiform (GBM), are considered as the most aggressive cancer subtype with the presence of microvascular proliferation and pseudopalisading necrosis. More importantly, grading of brain tumors is crucial for determining the survival rate; e.g., Grade I has the highest overall survival, and Grade IV has the poorest overall survival. The grade of the glioma tumor during initial diagnosis and prognosis is essential to determine appropriate treatment options [5,6].

Typically, for the initial characterization of the tumor, Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) scans are used to produce detailed images of the brain. However, for the further classification of tumors into sub-grades, a biopsy of the tumor is necessary for detailed examination by a pathologist [7]. Tissue biopsy is an invasive procedure and is time consuming, since it requires the tissue sample to be sent to a laboratory where a pathologist conducts extensive examination and classification [8]. The technology for imaging tissue has increased in resolution to identify smaller lesions causing a greater dependency on imaging for disease diagnosis. This has resulted in the use of multiple imaging technologies for cross-referencing a suspected clinical case and achieving greater accuracy in diagnosis. The consequence of more numerous imaging studies per patient is the requirement for the integration of additional technologies to aid in the diagnosis. Technologies such as computer-assisted diagnosis methods are being developed to provide supportive diagnostic tools for analyzing medical images and identifying disease, as well as grading of brain tumors [9].

An increasing number of medical imaging techniques which align with computer-based classification and segmentation algorithms are also being examined and validated by researchers. These analytic methodologies are being applied to different types of medical images for different clinical purposes including cancer tumors’ staging [10]. Such innovations also address the challenge to grading tumors by human interpretation of images from multiple readers since there is the possibility of inter-reader variability when determining the tumors’ grade based on the visual features of lesions. Therefore, an automated image analytic process is sought for the classification of brain tumors that would have the capacity to quantitatively assist in more objective diagnosis [11]. The task of brain tumor image analysis is very challenging with the use of traditional machine learning algorithms for glioma cancers since there are no well-defined characteristics of the tumor, and it requires accurate differentiation of the lesion from normal tissues surrounding the cancer. The efforts to utilize image analysis for cancer diagnosis are in contrast to more recent work on genetic analysis of tissue samples [12]. Such genomic analysis requires oncogenes to be identified using functional assays from tumors and is based on the premise that cancers are a genomic disease. The Cancer Genome Atlas (TCGA) Research Network has analyzed numerous human tumors to identify molecular alterations at the DNA, RNA, protein and epigenetic levels [13]. Similar to image-guided diagnosis and tumor classification, this molecular methodology is intended to classify the tumors and guide appropriate therapies. Additionally, there are many genomic applications available in all clinical disciplines today. One such example is in work by Verhaak et al., where they classify glioblastoma subtypes with specific alterations in genetic markers such as neurofibromatosis type 1 (NF1) and platelet-derived growth factor receptor/isocitrate dehydrogenase 1 (PDGFRA/IDH1) [14]. The intent is to better understand how the genetic alterations can be linked to alternative cells of tumor origin. These cumulative data are anticipated to serve as a framework for investigation of targeted therapies to block these molecular alterations. Additional efforts by TCGA to explore GBM suggest that Methylguanyl Methyltransferase (MGMT) can shift the genetic mutation spectrum for GBM and possibly provide options for alkylating treatment. While it is beyond the scope of this report to cover all the possible methodologies being researched to better identify brain tumors, it is worth mentioning that hyperspectral imaging has also been explored as a mechanism for tumor identification. This effort was a collaborative effort that ran till 2016 to generate a large image data stack at different spectral wavelengths. The project is described as having the capacity to “discriminate between healthy and malignant tissues in real-time during surgical procedures” and thus could be classified more as an image-guided therapy modality for cancers [15].

Literature Review

In surveying the image analysis published literature, a series of research studies describe the use of imaging methodologies in conjunction with with computer-aided diagnosis for the classification and segmentation of brain tumors. For example, in work by Pereira et al. the authors implemented the Convolutional Neural Network (ConvNet) as an artificial neural network for the purpose of segmenting the brain tumor based on pixel-level labeling [16]. They obtained a Dice similarity coefficient metric of 78%, 65% and 75% for the complete, core and enhancing brain tumor regions, respectively. This work is contrasted with the work proposed in the current study since their classification of brain tumor did not consider sub-grades of tumor classification using image-level labeling. Similarly, in Ertosun’s study, the authors presented an automatic grading tool of glioma using deep learning [5]. However, their tool used digital pathology images, which are acquired from an invasive imaging technique. The tool proposed in the current report uses a non-invasive imaging technique, which is based on fluid-attenuated inversion recovery (FLAIR)-weighted MR images.

In studies by Tate et al. the authors used Linear Discriminant Analysis (LDA) to classify brain tumors using 1H short-echo spectra [17]. They obtained near 90% correct classification for two out of three datasets they tested with their algorithm. In additional studies by Majós et al. the authors used single voxel proton MR spectroscopy at different values of Echo Time (TE) to classify the signal into four classes including meningioma, low-grade astrocytoma, anaplastic astrocytoma and glioblastoma-metastases [18]. They obtained an accurate classification rate of 81% on the dataset of short TE values. The classification methodology addressed in the study by Ranjith et al. attempted to classify the samples into two classes: benign and malignant. In this study, the database utilized consisted of MR spectroscopy data [19]. After implementing several machine learning methods in this study, a sensitivity of 86.1% was achieved using the random forest method. The classification solutions in both these earlier works were based on the use of single voxel proton MR spectroscopy unlike the solution proposed in our work, where we used FLAIR-weighted MR images [18,19] .

Additional studies published in the literature demonstrate the use of traditional machine learning methods to provide segmentation for the brain tumor, but not classification or grading. For example, the Bayesian Model used by Corso et al. was for the purpose of detecting and segmenting brain tumor from adjacent edema in multichannel MR 3D scans [20]. This Bayesian model was demonstrated to be computationally efficient with a segmentation accuracy of up to 88% compared to earlier studies analyzing MR images [21,22,23,24]. This model also had the second shortest computation time as compared to the previous studies [21,22,23,24]. Additional studies used expectation-maximization for a fully-automated tumor segmentation [25]. Expectation-maximization estimates the Probability Density Functions (PDFs) of the brain tissue classes and the intensity heterogeneity based on using T1- and T2-weighted MR images. The results of their approach were compared with a manual and semi-automatic approaches, where it gave a comparable, but less accurate performance as compared to the performance in 3D object segmentation of brain tumors [26,27].

In image analysis studies by Zacharaki the authors applied the Support Vector Machine (SVM) algorithm to classify glioma tumors [28]. They achieved a varying accuracy between 85% and 90% on the dataset they used. Both T1- and FLAIR-weighted MR images were used to assess the designed SVM classifier. The classification accuracy, sensitivity and specificity, were respectively 85%, 87% and 79% for discrimination of metastases from gliomas and 88%, 85% and 96% for discrimination of high-grade from low-grade gliomas. The study by Lawrence et al. inspired us to choose ConvNet as the deep Artificial Neural Network (ANN) for our analysis [29]. ConvNet is the most studied and validated methodology for image analysis tasks. ConvNet is well known for its special property called spatial invariance, such that the network is able to learn invariant features that make the convolution process invariant to translation, rotation and shifting. As a result, the differences between the same classed tumors due to the translation, rotation and shifting are dealt with using the spatial invariance property of ConvNet. Additionally, this makes the classification of brain tumors into different classes more feasible. In addition, the classification in this work is performed on image-level analysis. This function enables ConvNet to predict a single label representing the class of the patient’s MR image. These characteristics contrast with the segmentation option of analysis where the classification is performed at the image’s pixel level. The image-level classification that is used in this work does not require advanced computational processing capacity to train the deep network and tune its weights and parameters.

The aims of this study were as follows:
  • Propose a potential noninvasive replacement technique for the traditional invasive methods of grading brain tumors.

  • Address the classification of brain tumor using MR images in conjunction with deep learning with artificial neural networks.

  • Demonstrate a baseline application for using ConvNet in brain tumor grading and prove its efficiency for brain tumor staging.

2. Materials and Methods

2.1. Classification of Brain MR Images

In this study, an algorithm is designed using ConvNet for the aim of classifying varied brain images into categories of healthy brains, brains with low-grade tumor and brains with high-grade tumor. The presented work uses ConvNet as a classifier. This classifier is able to distinguish between the different categories based on the features that ConvNet learned automatically during the training process.

In theory, distinctions between healthy brains, brains with low-grade tumor and brains with high-grade tumor could be sorted by ConvNet because these differences are already apparent to the human eye. The main features for such distinction in classification are the existence of a necrotic core and enhancing rim around the tumor. Therefore, the target in this study is to utilize a neural network to automatically classify the MR images of the brain into three sub-classes:
  • MR brain images of healthy subjects.

  • MR brain images of glioma patients having low-grade glioma tumor.

  • MR brain images of glioma patients having high-grade glioma tumor.

Figure 1, which shows the FLAIR-weighted MR images, gives an example of how the brain of a healthy subject differs from the patient having a glioma cancer. One of the effects when glioma tumor spreads in the brain is that the distribution of fluids present in the brain is changed due to the formation of swelling, or edema around the necrotic or cystic core of the tumor, as portrayed in the figure below.

2.2. ConvNet Designed Architecture

ConvNet is considered as one of the feed-forward ANNs. It is used in this study for classifying brain tumors. The ConvNet is derived from the biological arrangement of the visual cortex of mammals where the main feature extracted as the input data to the network is the full connectedness style of neurons in the network architecture. ConvNet has a set of unique characteristics such as the non-fully-connected network, which is different from traditional ANN. This means that the data neurons are linked with a smaller part of the prior layer. Another important feature of ConvNet is the depth of architecture where each layer looks like a hexahedron of neurons of specific dimensions regarding height, width and depth [30].

Figure 2 illustrates the design of our ConvNet, including the different layers in the ConvNet architecture. The input image passes through different sets of layers. These layers consist of the convolution layer, max-pooling layer and rectified linear unit (ReLU) layer. Moving to the network’s rear part, the architecture includes a fully-connected layer and a softmax loss layer, which ensures that the output of the network represents the tissue class to which the input image belongs.

We implemented a modified version of AlexNet where the size of the input image is 160 × 160 kernels [31] as shown in Figure 3. More specifically, the layers’ configurations relative to the types and specifications of our network architecture are as follows, and Appendix A can be referred to for these layers’ definitions.
  • Convolutional layers: The filter parameters of the convolutional layer are initialized by giving random numbers from a Gaussian distribution with a spatial resolution of a 20 × 20 kernel array. The inputs to the convolutional layer are gray-scale images taken so that the filter depth at the input layer is 1. The number of filters is also set to 96, so that it spans the entire area of the input image.

  • Pooling layers: This is a different type of pooling layer, wherein a maximum number of pooling layers shows the best performance. The size of the sliding window is set to 3 × 3 kernels, and the value of the stride is set to 2. The stride of 2 means the image will be re-sampled with a value of 2. The max-pooling layer, however, decreases the spatial resolution of the image equally with the stride value. The size of the sliding window and the stride value remains the same throughout different max-pooling layers in the architecture. Then, the rectified linear unit as the non-linearity layer is used. This layer converts the entire pixel of negative values to zero, which leads to making the computations simpler and also avoids further complications due to moving forward along the other layers in the ConvNet. The convolutional, max-pooling and the non-linearity layers are replicated along the architecture.

  • Fully-connected layers: They consist of two layers where each one of them decreases the spatial size of the input to 1 × 1. The filter depth is equal to 3, corresponding to our three classes: healthy subjects, patients having a low-grade tumor and patients having a high-grade tumor.

  • Softmax loss layer: This layer is responsible for estimating the performance of the network and updating the network weights through the back-propagation process, which is based on the derivation of the loss function. This simplifies classification, because outputs are either close to 0 or 1.

3. Implementation and Experimentation

This section describes the dataset used, how the data input is provided to the network and the preprocessing methods that are applied to the input images.

3.1. Dataset

The dataset used in this study is from the Cancer Imaging Archive (TCIA) [32,33]. The data were originally annotated and labeled by experts from Thomas Jefferson University and Henry Ford Hospitals. This dataset is composed of MR scans for 130 subjects belonging to three classes including low-grade, high-grade and healthy subjects. This public dataset is one of the most trusted online datasets; however, it has a few limitations. Firstly, the dataset is composed of MR scans from 130 subjects. However, the ground truths were only provided for 126 subjects of the total number of subjects available in the dataset. Secondly, the ground truths provided were for the entire MR scan, which implies that the information at the slice level is missing. Finally, the segmentation information was not provided, thus, the labels at the pixel level were not given. Neurologists have done a two-phase examination. The first phase consists of verifying all the annotations given by Thomas Jefferson University and Henry Ford Hospitals’ experts by acting as second observers. Furthermore, they labeled the 4 unlabeled MR scans. It took approximately 2–3 min to examine each scan. This time includes the time required for loading, analyzing and providing feedback. The second phase of reviewing these scans was focused on the slice-level details. In this report, we used 2D ConvNet for the training and testing of the Computer-Aided Diagnostics (CAD) tool. Not all the slices of a full brain MRI scan of a patient with glioma tumor would show the lesion. Thus, neurologists selected only those few slices from each scan of the low-grade and high-grade tumor class that contained the lesion. The neurologists went through each scan, selecting an average of 31 slices from each MRI scan. The whole process was performed very carefully making sure that the selected slices from the low-grade and high-grade glioma class did not include any healthy slices. In this phase, It took approximately 5–7 min to analyze each scan.

In our dataset, the low-grade class consisted of Astrocytoma II and Oligodendroglioma II. The high-grade dataset class was a combination of GBM, Astrocytoma III and Oligodendroglioma III. The third and final class of the dataset was comprised of healthy subjects. These image archives from 130 subjects resulted in 4069 2D image samples in total. On average, 31 2D slices were selected out of each brain MR scan. The entire data break down of low-grade and high-grade gliomas is shown in the Table 1

One thought on “Glial Tumors Classification Essay

Leave a Reply

Your email address will not be published. Required fields are marked *