Diagnostic performance for pulmonary adenocarcinoma on CT: comparison of radiologists with and without three-dimensional convolutional neural network

To compare diagnostic performance for pulmonary invasive adenocarcinoma among radiologists with and without three-dimensional convolutional neural network (3D-CNN). Enrolled were 285 patients with adenocarcinoma in situ (AIS, n = 75), minimally invasive adenocarcinoma (MIA, n = 58), and invasive adenocarcinoma (IVA, n = 152). A 3D-CNN model was constructed with seven convolution-pooling and two max-pooling layers and fully connected layers, in which batch normalization, residual connection, and global average pooling were used. Only the flipping process was performed for augmentation. The output layer comprised two nodes for two conditions (AIS/MIA and IVA) according to prognosis. Diagnostic performance of the 3D-CNN model in 285 patients was calculated using nested 10-fold cross-validation. In 90 of 285 patients, results from each radiologist (R1, R2, and R3; with 9, 14, and 26 years of experience, respectively) with and without the 3D-CNN model were statistically compared. Without the 3D-CNN model, accuracy, sensitivity, and specificity of the radiologists were as follows: R1, 70.0%, 52.1%, and 90.5%; R2, 72.2%, 75%, and 69%; and R3, 74.4%, 89.6%, and 57.1%, respectively. With the 3D-CNN model, accuracy, sensitivity, and specificity of the radiologists were as follows: R1, 72.2%, 77.1%, and 66.7%; R2, 74.4%, 85.4%, and 61.9%; and R3, 74.4%, 93.8%, and 52.4%, respectively. Diagnostic performance of each radiologist with and without the 3D-CNN model had no significant difference (p > 0.88), but the accuracy of R1 and R2 was significantly higher with than without the 3D-CNN model (p < 0.01). The 3D-CNN model can support a less-experienced radiologist to improve diagnostic accuracy for pulmonary invasive adenocarcinoma without deteriorating any diagnostic performances. • The 3D-CNN model is a non-invasive method for predicting pulmonary invasive adenocarcinoma in CT images with high sensitivity. • Diagnostic accuracy by a less-experienced radiologist was better with the 3D-CNN model than without the model.

[1]  W. Willett,et al.  Misinterpretation and misuse of the kappa statistic. , 1987, American journal of epidemiology.

[2]  E. DeLong,et al.  Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. , 1988, Biometrics.

[3]  N. Dubrawsky.  Cancer statistics , 1989, CA: a cancer journal for clinicians.

[4]  E. Berg,et al.  World Health Organization Classification of Tumours , 2002 .

[5]  A. Borczuk,et al.  Invasive Size is an Independent Predictor of Survival in Pulmonary Adenocarcinoma , 2009, The American journal of surgical pathology.

[6]  H. Asamura,et al.  Histological Scoring for Small Lung Adenocarcinomas 2 cm or Less in Diameter: A Reliable Prognostic Indicator , 2010, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer.

[7]  C. Gatsonis,et al.  Reduced Lung-Cancer Mortality with Low-Dose Computed Tomographic Screening , 2012 .

[8]  Y. Takeshima,et al.  The prognostic role of pathologic invasive component size, excluding lepidic growth, in stage I lung adenocarcinoma. , 2013, The Journal of thoracic and cardiovascular surgery.

[9]  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.

[10]  C. Wilshire,et al.  Radiologic Evaluation of Small Lepidic Adenocarcinomas to Guide Decision Making in Surgical Resection. , 2015, The Annals of thoracic surgery.

[11]  Brady J. McKee,et al.  Performance of ACR Lung-RADS in a clinical CT lung screening program. , 2015, Journal of the American College of Radiology : JACR.

[12]  Michael S. Bernstein,et al.  ImageNet Large Scale Visual Recognition Challenge , 2014, International Journal of Computer Vision.

[13]  J. Crowley,et al.  The IASLC Lung Cancer Staging Project: Proposals for Revision of the TNM Stage Groupings in the Forthcoming (Eighth) Edition of the TNM Classification for Lung Cancer , 2016, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer.

[14]  A. Nicholson,et al.  The IASLC Lung Cancer Staging Project: Proposals for Coding T Categories for Subsolid Nodules and Assessment of Tumor Size in Part‐Solid Tumors in the Forthcoming Eighth Edition of the TNM Classification of Lung Cancer , 2016, Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer.

[15]  Subhashini Venugopalan,et al.  Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs. , 2016, JAMA.

[16]  Y. Kluger,et al.  Deep Survival: A Deep Cox Proportional Hazards Network , 2016, ArXiv.

[17]  A. Borczuk.  Prognostic considerations of the new World Health Organization classification of lung adenocarcinoma , 2016, European Respiratory Review.

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

[19]  Sebastian Thrun,et al.  Dermatologist-level classification of skin cancer with deep neural networks , 2017, Nature.

[20]  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 .

[21]  Seiichi Tagawa,et al.  Classification of C2C12 cells at differentiation by convolutional neural network of deep learning using phase contrast images , 2017, Human Cell.

[22]  B. van Ginneken,et al.  Observer variability for Lung-RADS categorisation of lung cancer screening CTs: impact on patient management , 2018, European Radiology.

[23]  Daniel S. Kermany,et al.  Identifying Medical Diagnoses and Treatable Diseases by Image-Based Deep Learning , 2018, Cell.

[24]  Lanfen Lin,et al.  A deep 3D residual CNN for false-positive reduction in pulmonary nodule detection. , 2018, Medical physics.

[25]  Uri Shaham,et al.  DeepSurv: personalized treatment recommender system using a Cox proportional hazards deep neural network , 2016, BMC Medical Research Methodology.

[26]  Weidong Cai,et al.  Knowledge-based Collaborative Deep Learning for Benign-Malignant Lung Nodule Classification on Chest CT , 2019, IEEE Transactions on Medical Imaging.

[27]  G. Corrado,et al.  End-to-end lung cancer screening with three-dimensional deep learning on low-dose chest computed tomography , 2019, Nature Medicine.

[28]  Lu Wang,et al.  Automated pulmonary nodule detection in CT images using 3D deep squeeze-and-excitation networks , 2019, International Journal of Computer Assisted Radiology and Surgery.

[29]  A. Jemal,et al.  Cancer statistics, 2019 , 2019, CA: a cancer journal for clinicians.

[30]  B. Zheng,et al.  A deep residual learning network for predicting lung adenocarcinoma manifesting as ground-glass nodule on CT images , 2019, European Radiology.

[31]  M. Noguchi,et al.  Application of deep learning (3-dimensional convolutional neural network) for the prediction of pathological invasiveness in lung adenocarcinoma , 2019, Medicine.

[32]  Y. She,et al.  Development and Validation of a Deep Learning Model for Non–Small Cell Lung Cancer Survival , 2020, JAMA network open.

[33]  Eui Jin Hwang,et al.  Deep learning algorithm for surveillance of pneumothorax after lung biopsy: a multicenter diagnostic cohort study , 2020, European Radiology.