Deep Learning Use for Differentiation of Low-grade vs High-Grade Glioma in Intraoperative Squash Smears

2021 ◽  
Author(s):  
Bhushan Diwakar Thombre ◽  
B N Nandeesh ◽  
Vikas Vazhayil ◽  
A R Prabhuraj
Keyword(s):  
Deep Learning ◽  
Diagnostic Tool ◽  
Learning Algorithm ◽  
High Grade Glioma ◽  
Rapid Screening ◽  
Training Dataset ◽  
Low Grade ◽  
High Grade ◽  
Current Standard ◽  
Feature Visualization

Abstract ObjectiveAutomated diagnosis using Artificial Intelligence (AI) techniques would be a useful addition to the intraoperative squash smear diagnosis. A robust diagnostic tool would enhance capabilities in centres where there is limited expertise for the diagnosis of intracranial lesions. The study aims to explore possibilities of deep learning technique-based models to classify squash smear images of glioma into high- and low-grade tumors.Methods500 Scanned images of squash smear were obtained intraoperatively and dataset was built. Image dataset was then pre-processed and fed into a CNN (Convolutional Neural Network) model for training and validation. The dataset consisted of 10,000 images of high (6000) and low (4000) grade gliomas, divided into three sets of training, validation and testing. ResultsCNN model based on deep learning algorithm was built and trained on training dataset to get accuracy of 96.2%. On a testing dataset which contains images previously unseen by trained model, it could achieve accuracies of 91% for diagnosing high grade glioma and 77% for low grade glioma. A positive predictive value of 86.6% and F1-score of 0.887 was achieved. Feature visualization technique was applied at the end to visualize regions of interest.ConclusionDeep Learning techniques can be applied as diagnostic tool if proper standardized images are obtained for reporting of squash smears of gliomas. The diagnostic accuracies of such tools can reach up to current standard diagnostic accuracies by conventional ways of reporting. Feature visualization techniques applied which can be used for rapid screening of slides or section of slide to assist in rapid diagnosis.

scholarly journals Study on Radar Echo-Filling in an Occlusion Area by a Deep Learning Algorithm

2021 ◽  
Vol 13 (9) ◽  
pp. 1779
Author(s):  
Xiaoyan Yin ◽  
Zhiqun Hu ◽  
Jiafeng Zheng ◽  
Boyong Li ◽  
Yuanyuan Zuo
Keyword(s):  
Deep Learning ◽  
Loss Function ◽  
Learning Algorithm ◽  
Weather Radar ◽  
Loss Functions ◽  
Training Dataset ◽  
Echo Intensity ◽  
Common Mean ◽  
Deep Learning Algorithm ◽  
Radar Beam

Radar beam blockage is an important error source that affects the quality of weather radar data. An echo-filling network (EFnet) is proposed based on a deep learning algorithm to correct the echo intensity under the occlusion area in the Nanjing S-band new-generation weather radar (CINRAD/SA). The training dataset is constructed by the labels, which are the echo intensity at the 0.5° elevation in the unblocked area, and by the input features, which are the intensity in the cube including multiple elevations and gates corresponding to the location of bottom labels. Two loss functions are applied to compile the network: one is the common mean square error (MSE), and the other is a self-defined loss function that increases the weight of strong echoes. Considering that the radar beam broadens with distance and height, the 0.5° elevation scan is divided into six range bands every 25 km to train different models. The models are evaluated by three indicators: explained variance (EVar), mean absolute error (MAE), and correlation coefficient (CC). Two cases are demonstrated to compare the effect of the echo-filling model by different loss functions. The results suggest that EFnet can effectively correct the echo reflectivity and improve the data quality in the occlusion area, and there are better results for strong echoes when the self-defined loss function is used.

scholarly journals Deep Learning Classification of Canine Behavior Using a Single Collar-Mounted Accelerometer: Real-World Validation

Animals ◽  
2021 ◽  
Vol 11 (6) ◽  
pp. 1549
Author(s):  
Robert D. Chambers ◽  
Nathanael C. Yoder ◽  
Aletha B. Carson ◽  
Christian Junge ◽  
David E. Allen ◽  
...  
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Real World ◽  
Learning Algorithm ◽  
Drinking Behavior ◽  
True Positive Rate ◽  
Training Dataset ◽  
Activity Levels ◽  
Accelerometer Data ◽  
Activity Monitors

Collar-mounted canine activity monitors can use accelerometer data to estimate dog activity levels, step counts, and distance traveled. With recent advances in machine learning and embedded computing, much more nuanced and accurate behavior classification has become possible, giving these affordable consumer devices the potential to improve the efficiency and effectiveness of pet healthcare. Here, we describe a novel deep learning algorithm that classifies dog behavior at sub-second resolution using commercial pet activity monitors. We built machine learning training databases from more than 5000 videos of more than 2500 dogs and ran the algorithms in production on more than 11 million days of device data. We then surveyed project participants representing 10,550 dogs, which provided 163,110 event responses to validate real-world detection of eating and drinking behavior. The resultant algorithm displayed a sensitivity and specificity for detecting drinking behavior (0.949 and 0.999, respectively) and eating behavior (0.988, 0.983). We also demonstrated detection of licking (0.772, 0.990), petting (0.305, 0.991), rubbing (0.729, 0.996), scratching (0.870, 0.997), and sniffing (0.610, 0.968). We show that the devices’ position on the collar had no measurable impact on performance. In production, users reported a true positive rate of 95.3% for eating (among 1514 users), and of 94.9% for drinking (among 1491 users). The study demonstrates the accurate detection of important health-related canine behaviors using a collar-mounted accelerometer. We trained and validated our algorithms on a large and realistic training dataset, and we assessed and confirmed accuracy in production via user validation.

scholarly journals HGG-12. A CASE OF PEDIATRIC SPINAL HIGH-GRADE GLIOMA WITH NTRK1 GENE FUSION

2020 ◽  
Vol 22 (Supplement_3) ◽  
pp. iii346-iii346
Author(s):  
Tamaki Morisako ◽  
Daisuke Umebayashi ◽  
Kazuaki Kamata ◽  
Hiroyuki Yamamoto ◽  
Takumi Yamanaka ◽  
...  
Keyword(s):  
Spinal Cord ◽  
Mr Imaging ◽  
Tumor Progression ◽  
Gene Fusion ◽  
High Grade Glioma ◽  
Gene Panel ◽  
Partial Resection ◽  
Low Grade ◽  
High Grade ◽  
Surgical Specimen

Abstract INTRODUCTION Tumors arising from the spinal cord are uncommon, especially high-grade tumors in pediatric patients. We report a case of high-grade glioma in the spinal cord harboring NTRK1 gene fusion, who received effective entrectinib therapy. CASE REPORT: A 5-year-old boy presented right hemiparesis and MR imaging revealed an intramedullary enhancing mass at the vertebral body level between C3 and Th1. He underwent microsurgical partial resection and the histological diagnosis was low-grade astrocytoma. After the first-line chemotherapy with vincristine and carboplatin, his right hemiparesis deteriorated and recurrent MR imaging showed growth of the tumor. He underwent microsurgical partial resection again and the histological examination was high-grade glioma with endothelial proliferation and necrosis. The chemoradiotherapy with temozolomide and focal irradiation of 50.4 Gy were given, and his neurological symptom slightly improved. One month later, he presented respiratory disturbance and required assisted ventilation with tracheostomy. MR imaging showed tumor progression invading upward to medulla oblongata. NTRK1 gene fusion was detected in the previous surgical specimen by a gene panel testing, and he received entrectinib, a potent inhibitor of tropomyosin receptor kinase (TRK). Since then, no tumor progression has been demonstrated for several months by MRI and he has been stable neurologically. CONCLUSION High-grade spinal cord tumors are rare and effective treatment strategies have not been addressed. Although the frequency of the gene fusion is very low in pediatric gliomas, identification of the driver gene aberration like in this case by a gene panel can provide potential targeted therapies for selected patients.

scholarly journals OS3.6 Prediagnostic symptoms and signs of adult glioma: the patients’ view

2019 ◽  
Vol 21 (Supplement_3) ◽  
pp. iii9-iii10
Author(s):  
M C M Peeters ◽  
L Dirven ◽  
J A F Koekkoek ◽  
E G Gortmaker ◽  
L Fritz ◽  
...  
Keyword(s):  
Health Care ◽  
Health Care Professionals ◽  
High Grade Glioma ◽  
Median Number ◽  
Low Grade ◽  
High Grade ◽  
Health Care Usage ◽  
Patients Experience ◽  
Symptoms And Signs ◽  
Chronic Complaints

Abstract BACKGROUND Little is known about the symptoms and signs glioma patients experience in the year before diagnosis, either or not resulting in health care usage. The objective of this study was to determine the incidence of several symptoms and signs glioma patients experienced in the year prior to diagnosis, as well as visits to a general practitioner (GP) related to these issues. MATERIAL AND METHODS This was a cross-sectional study, including adults diagnosed with a glioma <12 months ago. Patients were asked to complete a 30-item study-specific questionnaire, if possible with input of a proxy, focusing on symptoms and signs they experienced in the 12 months before diagnosis. For each indicated symptom/sign, patients were asked whether they consulted the GP for this issue. In addition, the presence of comorbidity and other chronic complaints were assessed, as well as consulted health care professionals (HCPs) in the year prior to diagnosis. The statistical analyses were corrected for multiple testing. RESULTS Between July 2016 and March 2019, 58 patients completed the questionnaires, 54 (93%) with input of a proxy. Forty-one (72%) patients were men, with a median age of 60 years (range 43–78), and the median time since diagnosis was 4 months (range 1–12). Forty (69%) patients had a comorbidity or chronic complaint, and the median number of consulted HCPs was 2 (range 0–8). The median number of symptoms/signs experienced in the year before diagnosis was 8 (range 2–19) in low-grade and 5 (range 0–24) in high-grade glioma (p=0.258). The five most frequently mentioned problems were fatigue (34/58, 59%), mental tiredness (28/58, 48%), sleeping disorder (23/58, 40%), headache (22/58, 38%) and stress (20/58, 34%), with no differences between low- and high grade glioma. Twenty-five (43%) patients had visited the GP with at least one issue. We found that patients who did consult their GP reported significantly more often muscle weakness (11 vs 3, p=0.002), paralysis in for example a hand or leg (10 vs 3, p=0.006), or a change in consciousness (9 vs 3, p=0.013) compared to those that did not consult the GP. However, they did not differ with respect to the number of symptoms (median 7 vs 5), comorbidities and chronic complaints (median 1 vs 1), or overall health care usage (median 3 vs 2). CONCLUSION Glioma patients experience a range of problems in the year prior to diagnosis, but patients who consult the GP report significantly more often neurological problems.

scholarly journals Research on Feature Extracted Method for Flutter Test Based on EMD and CNN

2021 ◽  
Vol 2021 ◽  
pp. 1-10
Author(s):  
Hua Zheng ◽  
Zhenglong Wu ◽  
Shiqiang Duan ◽  
Jiangtao Zhou
Keyword(s):  
Feature Extraction ◽  
Deep Learning ◽  
Wind Tunnel ◽  
Signal Analysis ◽  
Learning Algorithm ◽  
Structural Information ◽  
Structural Response ◽  
Training Dataset ◽  
Measured Signal ◽  
Test Dataset

Due to the inevitable deviations between the results of theoretical calculations and physical experiments, flutter tests and flutter signal analysis often play significant roles in designing the aeroelasticity of a new aircraft. The measured structural response from aeroelastic models in both wind tunnel tests and real fight flutter tests contain an abundance of structural information, but traditional methods tend to have limited ability to extract features of concern. Inspired by deep learning concepts, a novel feature extraction method for flutter signal analysis was established in this study by combining the convolutional neural network (CNN) with empirical mode decomposition (EMD). It is widely hypothesized that when flutter occurs, the measured structural signals are harmonic or divergent in the time domain, and that the flutter modal (1) is singular and (2) its energy increases significantly in the frequency domain. A measured-signal feature extraction and flutter criterion framework was constructed accordingly. The measured signals from a wind tunnel test were manually labeled “flutter” and “no-flutter” as the foundational dataset for the deep learning algorithm. After the normalized preprocessing, the intrinsic mode functions (IMFs) of the flutter test signals are obtained by the EMD method. The IMFs are then reshaped to make them the suitable size to be input to the CNN. The CNN parameters are optimized though the training dataset, and the trained model is validated through the test dataset (i.e., cross-validation). The accuracy rate of the proposed method reached 100% on the test dataset. The training model appears to effectively distinguish whether or not the structural response signal contains flutter. The combination of EMD and CNN provides effective feature extraction of time series signals in flutter test data. This research explores the connection between structural response signals and flutter from the perspective of artificial intelligence. The method allows for real-time, online prediction with low computational complexity.

scholarly journals Low and high grade glioma segmentation in multispectral brain MRI data

2018 ◽  
Vol 10 (1) ◽  
pp. 110-132 ◽  
Author(s):  
László Szilágyi ◽  
David Iclănzan ◽  
Zoltán Kapás ◽  
Zsófia Szabó ◽  
Ágnes Győrfi ◽  
...  
Keyword(s):  
Early Stage ◽  
Brain Mri ◽  
High Grade Glioma ◽  
Detection Algorithm ◽  
Gabor Wavelet ◽  
Low Grade ◽  
Data Sets ◽  
Morphological Operations ◽  
High Grade ◽  
Mass Screening System

Abstract Several hundreds of thousand humans are diagnosed with brain cancer every year, and the majority dies within the next two years. The chances of survival could be easiest improved by early diagnosis. This is why there is a strong need for reliable algorithms that can detect the presence of gliomas in their early stage. While an automatic tumor detection algorithm can support a mass screening system, the precise segmentation of the tumor can assist medical staff at therapy planning and patient monitoring. This paper presents a random forest based procedure trained to segment gliomas in multispectral volumetric MRI records. Beside the four observed features, the proposed solution uses 100 further features extracted via morphological operations and Gabor wavelet filtering. A neighborhood-based post-processing was designed to regularize and improve the output of the classifier. The proposed algorithm was trained and tested separately with the 54 low-grade and 220 high-grade tumor volumes of the MICCAI BRATS 2016 training database. For both data sets, the achieved accuracy is characterized by an overall mean Dice score > 83%, sensitivity > 85%, and specificity > 98%. The proposed method is likely to detect all gliomas larger than 10 mL.

PATH-41. POLYMORPHOUS LOW-GRADE NEUROEPITHELIAL TUMOR OF THE YOUNG WITH FGFR3-TACC3 FUSION MIMICKING HIGH-GRADE GLIOMA: CASE REPORT AND SERIES OF HIGH-GRADE CORRELATES

2021 ◽  
Vol 23 (Supplement_6) ◽  
pp. vi124-vi124
Author(s):  
Danielle Golub ◽  
Peter C Pan ◽  
Benjamin Liechty ◽  
Cheyanne Slocum ◽  
Tejus Bale ◽  
...  
Keyword(s):  
Progression Free Survival ◽  
High Grade Glioma ◽  
Proliferation Index ◽  
Molecular Profile ◽  
Low Grade ◽  
High Grade ◽  
Molecular Features ◽  
Tert Promoter ◽  
Tert Promoter Mutations ◽  
Promoter Mutations

Abstract BACKGROUND Polymorphous low-grade neuroepithelial tumor of the young (PLNTY) is a recently-described entity that can occasionally histologically and molecularly mimic high-grade glioma. The histologic and molecular features that predict aggressive behavior in FGFR3-TACC3 altered tumors are unclear. CASES We present a rare case of an indolent neuroepithelial neoplasm in a 59-year-old female with imaging initially suggestive of high-grade glioma and analyze common molecular features between this case and a series of high-grade gliomas. After total resection, pathology of the case patient revealed predominantly low-grade cytomorphology, abundant microcalcifications, unusual neovascularization, and a low proliferation index. The lesion was diffusely CD34 immunoreactive and harbored both an FGFR3-TACC3 fusion and a TERT promoter mutation. Based on the overall histologic and molecular profile, a diagnosis of PLNTY was favored. The patient was thereafter observed without adjuvant therapy with no evidence of progression at 15-month follow-up. In contrast, a series of eight adult patients with glioblastomas harboring FGFR3-TACC3 fusions and correspondingly aggressive clinical courses are also presented. Common molecular findings included IDH-wildtype status, absence of 1p19q codeletion, and CDKN2A loss. TERT promoter mutations and lack of MGMT promoter methylation were also frequently observed. These patients demonstrated a median 15-month overall survival and a 6-month progression-free survival. CONCLUSIONS PLNTY is a rare low-grade entity that can display characteristics of high-grade glioma, particularly in adults. The potential for a unique entity mimicking PLNTY which may act as a precursor lesion for a more malignant phenotype should be considered in cases with FGFR3-TACC3 fusions and other high-grade features.

KS02.4.A Olaparib in Recurrent IDH-mutant High-Grade Glioma (OLAGLI)

2021 ◽  
Vol 23 (Supplement_2) ◽  
pp. ii4-ii4
Author(s):  
F Ducray ◽  
M Sanson ◽  
O Chinot ◽  
M Fontanilles ◽  
R Rivoirard ◽  
...  
Keyword(s):  
Median Time ◽  
High Grade Glioma ◽  
Median Number ◽  
Parp Inhibition ◽  
Low Grade ◽  
Type I ◽  
High Grade ◽  
Rano Criteria ◽  
Anaplastic Gliomas ◽  
New Treatments

Abstract BACKGROUND There is a need to develop new treatments in IDH-mutant high-grade gliomas recurring after radiotherapy and chemotherapy. Based on preclinical studies showing that IDH-mutant tumors could be vulnerable to PARP inhibition we launched a phase II study to test the efficacy of olaparib (Lynparza) monotherapy in this population. METHODS Adults with recurrent high-grade IDH-mutant gliomas after radiotherapy and at least one line of alkylating chemotherapy (PCV or TMZ), KPS &gt; 60, normal organ function were enrolled. The primary endpoint was 6 months PFS according to RANO criteria. Patients were treated with olaparib 300 mg twice daily. We used a single-stage Fleming design with p0 = 30%, p1 = 50%, a type I unilateral error rate of 5% and a power of 80%. RESULTS 35 patients with recurrent IDH-mutant gliomas (IDH1R132H-mutant n = 32, other IDH mutation n = 3, 1p/19 codeleted n = 16, 1p/19q non-codeleted n = 14) were enrolled (malignantly transformed low-grade gliomas n = 21, anaplastic gliomas n = 8, glioblastomas n = 6). Median time since diagnosis was 7.4 years (1–22 years), median time since radiotherapy was 2.8 years (0.6–18 years), median number of previous chemotherapy lines was 2 (1–5). With a median follow-up of 11 months, 30 patients had stopped treatment due to tumor progression and 2 patients were still on treatment 16 to 18 months after treatment start. At 6 months, 11/35 patients were progression-free (31 %). According to RANO criteria, based on local investigator analysis, 2 patients (5%) had a partial response and 14 patients a stable disease (37%) with a median duration of response of 9 months (4–18+). Median PFS and OS were 2.3 and 15.9 months and were similar in 1p/19q codeleted and non-codeleted patients. A grade 3 olaparib-related adverse event was observed in 5 patients (14%, lymphopenia n = 3, fatigue n = 2, diarrhea n = 1) and a grade 2 in 15 patients (43%), most frequently consisting in fatigue (23%), gastrointestinal disorders (20%) and lymphopenia (20%). No patient definitively stopped olaparib due to side effects. CONCLUSIONS In this heavily pre-treated population of recurrent IDH-mutant gliomas, olaparib monotherapy was well tolerated and resulted in some activity supporting its evaluation in association with alkylating chemotherapy in recurrent IDH-mutant gliomas in future studies.

scholarly journals SURG-32. THE USE OF 5-AMINOLEVULINIC ACID IN LOW-GRADE GLIOMA RESECTION: A SYSTEMATIC REVIEW

2019 ◽  
Vol 21 (Supplement_6) ◽  
pp. vi246-vi246
Author(s):  
Ahmad Almekkawi ◽  
Tarek El Ahmadieh ◽  
Karl Abi-Aad ◽  
Salah Aoun ◽  
Najib EL Tecle ◽  
...  
Keyword(s):  
Systematic Review ◽  
Quality Index ◽  
Aminolevulinic Acid ◽  
High Grade Glioma ◽  
Extent Of Resection ◽  
Low Grade Glioma ◽  
Low Grade ◽  
High Grade ◽  
Visualization Technology ◽  
Glioma Resection

Abstract BACKGROUND 5-aminolevulinic acid is a reliable tool for optimizing high-grade glioma resection. However, its efficacy in low-grade glioma resection remains unclear. OBJECTIVE To study the role of 5-aminolevulinic acid in low-grade glioma resection and assess positive fluorescence rates and effect on the extent of resection. METHODS A systematic review of PubMed, Google Scholar, and Cochrane was performed from the date of inception to February 1, 2019. Studies that correlated 5-aminolevulinic acid fluorescence with low-grade glioma in the setting of operative resection were selected. Studies with biopsy only were excluded. Positive fluorescence rates were calculated. Quality index of the selected papers using the Downs and Black criteria checklist was provided. RESULTS Twelve articles met the selection criteria with 244 histologically-confirmed low-grade glioma patients who underwent microsurgical resection. All patients received 20 mg/kg body weight of 5-aminolevulinic acid. Only 60 patients (n=60/244; 24.5%) demonstrated visual intra-operative 5-aminolevulinic acid fluorescence. The extent of resection was reported in 4 studies, however, the data combined low- and high-grade tumors. Only 2 studies reported on tumor location. Only 3 studies reported on clinical outcomes. The Zeiss OPMI Pentero microscope was most commonly used across all studies. The average quality index was 14.58 (range: 10–17) which correlated with an overall good quality. CONCLUSION There is an overall low correlation between 5-aminolevulinic acid fluorescence and low-grade glioma. Advances in visualization technology and using standardized fluorescence quantification methods may further improve the visualization and reliability of 5-aminolevulinic acid fluorescence in low-grade glioma resection.

Sign in / Sign up

Export Citation Format

Share Document