Convolutional neural networks can accurately distinguish four histologic growth patterns of lung adenocarcinoma in digital slides

During the diagnostic workup of lung adenocarcinomas (LAC), pathologists evaluate distinct histological tumor growth patterns. The percentage of each pattern on multiple slides bears prognostic significance. To assist with the quantification of growth patterns, we constructed a pipeline equipped with a convolutional neural network (CNN) and soft-voting as the decision function to recognize solid, micropapillary, acinar, and cribriform growth patterns, and non-tumor areas. Slides of primary LAC were obtained from Cedars-Sinai Medical Center (CSMC), the Military Institute of Medicine in Warsaw and the TCGA portal. Several CNN models trained with 19,924 image tiles extracted from 78 slides (MIMW and CSMC) were evaluated on 128 test slides from the three sites by F1-score and accuracy using manual tumor annotations by pathologist. The best CNN yielded F1-scores of 0.91 (solid), 0.76 (micropapillary), 0.74 (acinar), 0.6 (cribriform), and 0.96 (non-tumor) respectively. The overall accuracy of distinguishing the five tissue classes was 89.24%. Slide-based accuracy in the CSMC set (88.5%) was significantly better (p < 2.3E-4) than the accuracy in the MIMW (84.2%) and TCGA (84%) sets due to superior slide quality. Our model can work side-by-side with a pathologist to accurately quantify the percentages of growth patterns in tumors with mixed LAC patterns.

[1]  Hilla Peretz,et al.  Ju n 20 03 Schrödinger ’ s Cat : The rules of engagement , 2003 .

[2]  Wilko Weichert,et al.  Prognostic Impact and Clinicopathological Correlations of the Cribriform Pattern in Pulmonary Adenocarcinoma , 2015, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer.

[3]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[4]  Andrew H. Beck,et al.  Diagnostic Assessment of Deep Learning Algorithms for Detection of Lymph Node Metastases in Women With Breast Cancer , 2017, JAMA.

[5]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[6]  Daisuke Komura,et al.  Machine Learning Methods for Histopathological Image Analysis , 2017, Computational and structural biotechnology journal.

[7]  Pawel Badura,et al.  Information Technologies in Medicine - 5th International Conference, ITIB, 2016 Kamień Śląski, Poland, June 20-22, 2016 Proceedings, Volume 1 , 2016, ITIB.

[8]  Matti Pietikäinen,et al.  Identification of tumor epithelium and stroma in tissue microarrays using texture analysis , 2012, Diagnostic Pathology.

[9]  Aristotelis Tsirigos,et al.  Classification and Mutation Prediction from Non-Small Cell Lung Cancer Histopathology Images using Deep Learning , 2017, bioRxiv.

[10]  Mahadev Satyanarayanan,et al.  OpenSlide: A vendor-neutral software foundation for digital pathology , 2013, Journal of pathology informatics.

[11]  Akihiko Yoshizawa,et al.  Impact of proposed IASLC/ATS/ERS classification of lung adenocarcinoma: prognostic subgroups and implications for further revision of staging based on analysis of 514 stage I cases , 2011, Modern Pathology.

[12]  P. Schil.  Faculty Opinions recommendation of The 2015 world health organization classification of lung tumors: impact of genetic, clinical and radiologic advances since the 2004 classification. , 2017 .

[13]  A. Gazdar,et al.  Comprehensive analysis of lung cancer pathology images to discover tumor shape features that predict survival outcome , 2018, bioRxiv.

[14]  Iver Petersen,et al.  Reproducibility of histopathological subtypes and invasion in pulmonary adenocarcinoma. An international interobserver study , 2012, Modern Pathology.

[15]  Ehsan Kazemi,et al.  Deep Convolutional Neural Networks Enable Discrimination of Heterogeneous Digital Pathology Images , 2017, bioRxiv.

[16]  J. Austin,et al.  The 2015 World Health Organization Classification of Lung Tumors: Impact of Genetic, Clinical and Radiologic Advances Since the 2004 Classification. , 2015, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer.

[17]  A. Yoshizawa,et al.  Validation of the IASLC/ATS/ERS Lung Adenocarcinoma Classification for Prognosis and Association with EGFR and KRAS Gene Mutations: Analysis of 440 Japanese Patients , 2013, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer.

[18]  Joel H. Saltz,et al.  Patch-Based Convolutional Neural Network for Whole Slide Tissue Image Classification , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[19]  E. van Marck,et al.  Different growth patterns of non-small cell lung cancer represent distinct biologic subtypes. , 2008, The Annals of thoracic surgery.

[20]  A. Gazdar,et al.  Comprehensive analysis of lung cancer pathology images to discover tumor shape and boundary features that predict survival outcome , 2018, Scientific Reports.

[21]  Zoe Wainer,et al.  Does Lung Adenocarcinoma Subtype Predict Patient Survival?: A Clinicopathologic Study Based on the New International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society International Multidisciplinary Lung Adenocarcinoma Classification , 2011, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer.

[22]  Akihiko Yoshizawa,et al.  A Grading System of Lung Adenocarcinomas Based on Histologic Pattern is Predictive of Disease Recurrence in Stage I Tumors , 2010, The American journal of surgical pathology.

[23]  David M. W. Powers,et al.  Evaluation: from precision, recall and F-measure to ROC, informedness, markedness and correlation , 2011, ArXiv.

[24]  W. Travis,et al.  Pathology of lung cancer. , 2011, Clinics in chest medicine.

[25]  Gwénaël Le Teuff,et al.  Subtype Classification of Lung Adenocarcinoma Predicts Benefit From Adjuvant Chemotherapy in Patients Undergoing Complete Resection. , 2015, Journal of clinical oncology : official journal of the American Society of Clinical Oncology.

[26]  Yang Zhang,et al.  The prognostic and predictive value of solid subtype in invasive lung adenocarcinoma , 2014, Scientific Reports.

[27]  Junzhou Huang,et al.  Comprehensive Computational Pathological Image Analysis Predicts Lung Cancer Prognosis , 2017, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer.

[28]  Dmitrii Bychkov,et al.  Deep learning based tissue analysis predicts outcome in colorectal cancer , 2018, Scientific Reports.

[29]  D. Lynch,et al.  The National Lung Screening Trial: overview and study design. , 2011, Radiology.

[30]  Bram van Ginneken,et al.  Epithelium segmentation using deep learning in H&E-stained prostate specimens with immunohistochemistry as reference standard , 2018, Scientific Reports.

[31]  W. Travis.  Pathology of lung cancer. , 2002, Clinics in chest medicine.

[32]  R. Figlin,et al.  A novel machine learning approach reveals latent vascular phenotypes predictive of renal cancer outcome , 2017, Scientific Reports.

[33]  Dumitru Erhan,et al.  Going deeper with convolutions , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[34]  N. Razavian,et al.  Classification and mutation prediction from non–small cell lung cancer histopathology images using deep learning , 2018, Nature Medicine.

[35]  S. Horie,et al.  Excellent prognosis of lepidic-predominant lung adenocarcinoma: low incidence of lymphatic vessel invasion as a key factor. , 2014, Anticancer research.

[36]  Erik Reinhard,et al.  Color Transfer between Images , 2001, IEEE Computer Graphics and Applications.

[37]  B. van Ginneken,et al.  Deep learning as a tool for increased accuracy and efficiency of histopathological diagnosis , 2016, Scientific Reports.

[38]  Andrew Janowczyk,et al.  Deep learning for digital pathology image analysis: A comprehensive tutorial with selected use cases , 2016, Journal of pathology informatics.

[39]  David L. Olson,et al.  Advanced Data Mining Techniques , 2008 .

[40]  I. Wistuba,et al.  Histologic patterns and molecular characteristics of lung adenocarcinoma associated with clinical outcome , 2012, Cancer.

[41]  Olaf Hellwich,et al.  Deep convolutional neural networks for automatic classification of gastric carcinoma using whole slide images in digital histopathology , 2017, Comput. Medical Imaging Graph..

[42]  D. Brat,et al.  Predicting cancer outcomes from histology and genomics using convolutional networks , 2017, Proceedings of the National Academy of Sciences.

[43]  Catarina Eloy,et al.  Classification of breast cancer histology images using Convolutional Neural Networks , 2017, PloS one.

[44]  Zhipeng Jia,et al.  Large scale tissue histopathology image classification, segmentation, and visualization via deep convolutional activation features , 2017, BMC Bioinformatics.

[45]  Iver Petersen,et al.  Training increases concordance in classifying pulmonary adenocarcinomas according to the novel IASLC/ATS/ERS classification , 2012, Virchows Archiv.

[46]  Prasad S Adusumilli,et al.  The cribriform pattern identifies a subset of acinar predominant tumors with poor prognosis in patients with stage I lung adenocarcinoma: a conceptual proposal to classify cribriform predominant tumors as a distinct histologic subtype , 2014, Modern Pathology.

[47]  Thomas J. Fuchs,et al.  Terabyte-scale Deep Multiple Instance Learning for Classification and Localization in Pathology , 2018, ArXiv.

[48]  H. Asamura,et al.  The utility of the proposed IASLC/ATS/ERS lung adenocarcinoma subtypes for disease prognosis and correlation of driver gene alterations. , 2013, Lung cancer.