Classification and Mutation Prediction from Non-Small Cell Lung Cancer Histopathology Images using Deep Learning

Coudray, Nicolas; Moreira, André L.; Sakellaropoulos, Theodore; Fenyö, David; Razavian, Narges; Tsirigos, Aristotelis

doi:10.1101/197574

Cited by 67 publications

(81 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…65 SETBP1 forms a heterodimer with SET protein and inhibited the tumor suppressive functions of PP2A, SETBP1 mutation is detected in AML and non-small cell lung cancer, which generally blocks its ubiquitination and elicits abnormal high expression. [66][67][68] RHOA is a GTPases that influences multiple biological processes, upregulation of RHOA is associated with tumorigenesis, while instead of intestinal subtype, RHOA mutation is specifically identified in diffuse subtype of GC. 69,70 MUC6 encodes a secretory mucin and protects gastric mucosa, whose reduction inclines chronic mucosal injury and promotes carcinogenesis.…”

Section: Discussionmentioning

confidence: 99%

Reduced m6A modification predicts malignant phenotypes and augmented Wnt/PI3K‐Akt signaling in gastric cancer

et al. 2019

View full text Add to dashboard Cite

Background As the most abundant epigenetic modification on mRNAs and long non‐coding RNAs, N6‐methyladenosine (m6A) modification extensively exists in mammalian cells. Controlled by writers (methyltransferases), readers (signal transducers), and erasers (demethylases), m6A influences mRNA structure, maturation, and stability, thus negatively regulating protein expression in a post‐translational manner. Nevertheless, current understanding of m6A's roles in tumorigenesis, especially in gastric cancer (GC) remains to be unveiled. In this study, we assessed m6A's clinicopathological relevance to GC and explored the underlying mechanisms. Methods By referring to a proteomics‐based GC cohort we previously generated and the TCGA‐GC cohort, we merged expressions of canonical m6A writers (METTL3/METTL14), readers (YTHDF1/YTHDF2/YTHDF3), and erasers (ALKBH5/FTO), respectively, as W, R, and E signatures to represent m6A modification. We stratified patients according to these signatures to decipher m6A's associations with crucial mutations, prognosis, and clinical indexes. m6A's biological functions in GC were predicted by gene set enrichment analysis (GSEA) and validated by in vitro experiments. Results We discovered that W and R were potential tumor suppressive signatures, while E was a potential oncogenic signature in GC. According to W/R/E stratifications, patients with low m6A‐indications were accompanied with higher mutations of specific genes ( CDH1 , AR , GLI3 , SETBP1 , RHOA , MUC6 , and TP53 ) and also demonstrated adverse clinical outcomes. GSEA suggested that reduced m6A was correlated with oncogenic signaling and phenotypes. Through in vitro experiments, we proved that m6A suppression (represented by METTL14 knockdown) promoted GC cell proliferation and invasiveness through activating Wnt and PI3K‐Akt signaling, while m6A elevation (represented by FTO knockdown) reversed these phenotypical and molecular changes. m6A may also be involved in interferon signaling and immune responses of GC. Conclusions Our work demonstrated that low‐m6A signatures predicted adverse clinicopathological features of GC, while the reduction of RNA m6A methylation activated oncogenic Wnt/PI3K‐Akt signaling and promoted malignant phenotypes of GC cells.

show abstract

Section: Discussionmentioning

confidence: 99%

Reduced m6A modification predicts malignant phenotypes and augmented Wnt/PI3K‐Akt signaling in gastric cancer

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Deep learning, as the core technology of the rising artificial intelligence (AI) in recent years, has been reported with significantly diagnostic accuracy in medical imaging for automatic detection of lung diseases [13,14,15]. It surpassed human-level performance on the ImageNet image classification task with one million images for training in 2015 [16], showed dermatologist-level performance on classifying skin lesions in 2017 [17] and obtained very impressive results for lung cancer screening in 2019 [13].…”

Section: Introductionmentioning

confidence: 99%

Deep Learning-based Detection for COVID-19 from Chest CT using Weak Label

Zheng

Deng

et al. 2020

Preprint

519

443

View full text Add to dashboard Cite

Accurate and rapid diagnosis of COVID-19 suspected cases plays a crucial role in timely quarantine and medical treatment. Developing a deep learning-based model for automatic COVID-19 detection on chest CT is helpful to counter the outbreak of SARS-CoV-2. A weakly-supervised deep learning-based software system was developed using 3D CT volumes to detect COVID-19. For each patient, the lung region was segmented using a pre-trained UNet; then the segmented 3D lung region was fed into a 3D deep neural network to predict the probability of COVID-19 infectious. 499 CT volumes collected from Dec. 13, 2019, to Jan. 23, 2020, were used for training and 131 CT volumes collected from Jan 24, 2020, to Feb 6, 2020, were used for testing. The deep learning algorithm obtained 0.959 ROC AUC and 0.976 PR AUC. There was an operating point with 0.907 sensitivity and 0.911 specificity in the ROC curve. When using a probability threshold of 0.5 to classify COVID-positive and COVID-negative, the algorithm obtained an accuracy of 0.901, a positive predictive value of 0.840 and a very high negative predictive value of 0.982. The algorithm took only 1.93 seconds to process a single patient's CT volume using a dedicated GPU. Our weaklysupervised deep learning model can accurately predict the COVID-19 infectious probability in chest CT volumes without the need for annotating the lesions for training. The easily-trained and highperformance deep learning algorithm provides a fast way to identify COVID-19 patients, which is beneficial to control the outbreak of SARS-CoV-2. The developed deep learning software is available at https://github.com/sydney0zq/covid-19-detection.

show abstract

“…Overall, the dataset consisted of 141 WSIs from 56 patients, 28 patients with FCD IIb, and 28 patients with genetically confirmed TSC. H&E stainings were included due to the proven potential of CNNs to extract information not visible to the human observer in H&E slides, 5,22 thus eliminating the need for more complex and expensive immunostainings.…”

Section: Dataset and Region Of Interestmentioning

confidence: 99%

“…Based on these tasks, more abstract functions like disease grading, prognosis prediction, and imaging biomarkers for genetic subtype identification have been established. 4,5 Successful examples include utilization in different types of cancer detection/classification/grading, 6,7 classification of liver cirrhosis, 8 heart failure detection, 9 and classification of Alzheimer plaques. 10 The most commonly used deep learning architectures are convolutional neural networks (CNNs; Figure 1C).…”

Section: Introductionmentioning

confidence: 99%

Same same but different: A Web‐based deep learning application revealed classifying features for the histopathologic distinction of cortical malformations

et al. 2020

View full text Add to dashboard Cite

Objective The microscopic review of hematoxylin‐eosin–stained images of focal cortical dysplasia type IIb and cortical tuber of tuberous sclerosis complex remains challenging. Both entities are distinct subtypes of human malformations of cortical development that share histopathological features consisting of neuronal dyslamination with dysmorphic neurons and balloon cells. We trained a convolutional neural network (CNN) to classify both entities and visualize the results. Additionally, we propose a new Web‐based deep learning application as proof of concept of how deep learning could enter the pathologic routine. Methods A digital processing pipeline was developed for a series of 56 cases of focal cortical dysplasia type IIb and cortical tuber of tuberous sclerosis complex to obtain 4000 regions of interest and 200 000 subsamples with different zoom and rotation angles to train a neural network. Guided gradient‐weighted class activation maps (Guided Grad‐CAMs) were generated to visualize morphological features used by the CNN to distinguish both entities. Results Our best‐performing network achieved 91% accuracy and 0.88 area under the receiver operating characteristic curve at the tile level for an unseen test set. Novel histopathologic patterns were found through the visualized Guided Grad‐CAMs. These patterns were assembled into a classification score to augment decision‐making in routine histopathology workup. This score was successfully validated by 11 expert neuropathologists and 12 nonexperts, boosting nonexperts to expert level performance. Significance Our newly developed Web application combines the visualization of whole slide images with the possibility of deep learning–aided classification between focal cortical dysplasia IIb and tuberous sclerosis complex. This approach will help to introduce deep learning applications and visualization for the histopathologic diagnosis of rare and difficult‐to‐classify brain lesions.

show abstract

Classification and Mutation Prediction from Non-Small Cell Lung Cancer Histopathology Images using Deep Learning

Abstract: 25Visual analysis of histopathology slides of lung cell tissues is one of the main methods used by

Cited by 67 publications

References 24 publications

Reduced m6A modification predicts malignant phenotypes and augmented Wnt/PI3K‐Akt signaling in gastric cancer

Reduced m6A modification predicts malignant phenotypes and augmented Wnt/PI3K‐Akt signaling in gastric cancer

Deep Learning-based Detection for COVID-19 from Chest CT using Weak Label

Same same but different: A Web‐based deep learning application revealed classifying features for the histopathologic distinction of cortical malformations

Contact Info

Product

Resources

About