Deep Convolutional Neural Network-Based Positron Emission Tomography Analysis Predicts Esophageal Cancer Outcome

In esophageal cancer, few prediction tools can be confidently used in current clinical practice. We developed a deep convolutional neural network (CNN) with 798 positron emission tomography (PET) scans of esophageal squamous cell carcinoma and 309 PET scans of stage I lung cancer. In the first stage, we pretrained a 3D-CNN with all PET scans for a task to classify the scans into esophageal cancer or lung cancer. Overall, 548 of 798 PET scans of esophageal cancer patients were included in the second stage with an aim to classify patients who expired within or survived more than one year after diagnosis. The area under the receiver operating characteristic curve (AUC) was used to evaluate model performance. In the pretrain model, the deep CNN attained an AUC of 0.738 in identifying patients who expired within one year after diagnosis. In the survival analysis, patients who were predicted to be expired but were alive at one year after diagnosis had a 5-year survival rate of 32.6%, which was significantly worse than the 5-year survival rate of the patients who were predicted to survive and were alive at one year after diagnosis (50.5%, p < 0.001). These results suggest that the prediction model could identify tumors with more aggressive behavior. In the multivariable analysis, the prediction result remained an independent prognostic factor (hazard ratio: 2.830; 95% confidence interval: 2.252–3.555, p < 0.001). We conclude that a 3D-CNN can be trained with PET image datasets to predict esophageal cancer outcome with acceptable accuracy.

[1]  Peter Smeets,et al.  Predictive and prognostic value of metabolic tumour volume and total lesion glycolysis in solid tumours , 2012, European Journal of Nuclear Medicine and Molecular Imaging.

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

[3]  Ronald M. Summers,et al.  Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning , 2016, IEEE Transactions on Medical Imaging.

[4]  C. Compton,et al.  Survival prediction tools for esophageal and gastroesophageal junction cancer: A systematic review , 2018, The Journal of thoracic and cardiovascular surgery.

[5]  Shuhei Nomura,et al.  Automatic anatomical classification of esophagogastroduodenoscopy images using deep convolutional neural networks , 2018, Scientific Reports.

[6]  V. Goh,et al.  Predicting Response to Neoadjuvant Chemotherapy with PET Imaging Using Convolutional Neural Networks , 2015, PloS one.

[7]  Richard K. G. Do,et al.  Convolutional neural networks: an overview and application in radiology , 2018, Insights into Imaging.

[8]  P. Lambin,et al.  Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach , 2014, Nature Communications.

[9]  L Cozzi,et al.  PET Radiomics in NSCLC: state of the art and a proposal for harmonization of methodology , 2017, Scientific Reports.

[10]  R. Cerfolio,et al.  Maximum standardized uptake values on positron emission tomography of esophageal cancer predicts stage, tumor biology, and survival. , 2006, The Annals of thoracic surgery.

[11]  Michael V. McConnell,et al.  Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning , 2017, Nature Biomedical Engineering.

[12]  H. Zaidi,et al.  Towards enhanced PET quantification in clinical oncology. , 2018, The British journal of radiology.

[13]  Philippe Lambin,et al.  Pre-treatment CT radiomics to predict 3-year overall survival following chemoradiotherapy of esophageal cancer , 2018, Acta oncologica.

[14]  Walter Noordzij,et al.  Prediction of Response to Neoadjuvant Chemotherapy and Radiation Therapy with Baseline and Restaging 18F-FDG PET Imaging Biomarkers in Patients with Esophageal Cancer. , 2018, Radiology.

[15]  M. Schwaiger,et al.  Metabolic imaging predicts response, survival, and recurrence in adenocarcinomas of the esophagogastric junction. , 2006, Journal of clinical oncology : official journal of the American Society of Clinical Oncology.

[16]  Patrick Granton,et al.  Radiomics: extracting more information from medical images using advanced feature analysis. , 2012, European journal of cancer.

[17]  Amy Maxmen Deep learning sharpens views of cells and genes , 2018, Nature.

[18]  Wei-min Li,et al.  Development and clinical application of radiomics in lung cancer , 2017, Radiation Oncology.

[19]  H. Haenssle,et al.  Man against machine: diagnostic performance of a deep learning convolutional neural network for dermoscopic melanoma recognition in comparison to 58 dermatologists , 2018, Annals of oncology : official journal of the European Society for Medical Oncology.

[20]  O. Hoekstra,et al.  FDG-PET Parameters as Prognostic Factor in Esophageal Cancer Patients: A Review , 2012, Indian Journal of Surgical Oncology.

[21]  R. Subramaniam,et al.  PET/Computed Tomography Scanning and Precision Medicine: Esophageal Cancer. , 2017, PET clinics.

[22]  R. Larue,et al.  External validation of a prognostic model incorporating quantitative PET image features in oesophageal cancer. , 2019, Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology.

[23]  Impact of sarcopenia on outcomes of locally advanced esophageal cancer patients treated with neoadjuvant chemoradiation followed by surgery. , 2017, Journal of gastrointestinal oncology.

[24]  O. Hoekstra,et al.  FDG-PET Parameters as Prognostic Factor in Esophageal Cancer Patients: A Review , 2011, Annals of surgical oncology.

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

[26]  Zhonghua Chen,et al.  Comparison of machine learning methods for classifying mediastinal lymph node metastasis of non-small cell lung cancer from 18F-FDG PET/CT images , 2017, EJNMMI Research.

[27]  Kenji Suzuki,et al.  Overview of deep learning in medical imaging , 2017, Radiological Physics and Technology.

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

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

[30]  R. Gillies,et al.  Deep learning for lung cancer prognostication: A retrospective multi-cohort radiomics study , 2018, PLoS medicine.

[31]  Adelin Albert,et al.  FDG PET/CT radiomics for predicting the outcome of locally advanced rectal cancer , 2017, European Journal of Nuclear Medicine and Molecular Imaging.

[32]  R. Hills,et al.  Evaluation of prognostic models developed using standardised image features from different PET automated segmentation methods , 2018, EJNMMI Research.

[33]  Vicky Goh,et al.  Quantifying tumour heterogeneity in 18F-FDG PET/CT imaging by texture analysis , 2012, European Journal of Nuclear Medicine and Molecular Imaging.

[34]  M. Orringer,et al.  Decreased core muscle size is associated with worse patient survival following esophagectomy for cancer. , 2013, Diseases of the esophagus : official journal of the International Society for Diseases of the Esophagus.