Introduction

Cancer is the second leading cause of death globally and is responsible for an estimated 9.6 million deaths in 2018. Lung cancer is the leading cause of cancer death in the United States with an estimated 160,000 deaths in the past year.[1] Early detection of cancer hence plays a key role in diagnosis which, in turn, improves the long-term survival rates.

There are several barriers to the early detecting of cancer, such as a global shortage of radiologists. In addition to the shortage, detecting malignant tumors in X-rays can be difficult and challenging even for experienced radiologists. This time-consuming process typically leads to fatigue-based diagnostic errors and discrepancies[2].

Our project focuses on detecting the presence of malignant tumors in chest X-rays. In order to aid radiologists around the world, we propose to exploit supervised and unsupervised Machine Learning algorithms for lung cancer detection. We aim to showcase ‘explainable’ models [3] that could perform close to human accuracy levels for cancer-detection. We envision our models being used to assist radiologists and scaling cancer detection to overcome the lack of diagnostic bandwidth in this domain. We can also potentially export our models to personal devices, which would allow for easier, cheaper and more accessible cancer detection.

Supervised Learning

Grad CAM

Unsupervised Learning

Dataset

Results

References

1. Rajpurkar, P., Irvin, J., Zhu, K., Yang, B., Mehta, H., Duan, T. et.al., M. P. (2017). Chexnet: Radiologist-level pneumonia detection on chest x-rays with deep learning.
2. Gulshan V, Peng L, Coram M, et al. Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs. JAMA. 2016
3. R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh and D. Batra, "Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization," 2017 IEEE International Conference on Computer Vision (ICCV), Venice, 2017, pp. 618-626.
4. Kingma P, Welling M., An Introduction to Variational Autoencoders, arXiv:1906.02691
5. Ardila, D., Kiraly, A.P., Bharadwaj, S. et al. End-to-end lung cancer screening with three-dimensional deep learning on low-dose chest computed tomography. Nat Med 25, 954–961 (2019)
6. Yongsik Sim, Myung Jin Chung et al. Deep Convolutional Neural Network–based Software Improves Radiologist Detection of Malignant Lung Nodules on Chest Radiographs, Radiology, 2019