An official website of the United States government
Abstract Non-small-cell lung cancer (NSCLC) represents approximately 80–85% of lung cancer diagnoses and is the leading cause of cancer-related death worldwide. Recent studies indicate that image-based radiomics features from positron emission tomography/computed tomography (PET/CT) images have predictive power for NSCLC outcomes. To this end, easily calculated functional features such as the maximum and the mean of standard uptake value (SUV) and total lesion glycolysis (TLG) are most commonly used for NSCLC prognostication, but their prognostic value remains controversial. Meanwhile, convolutional neural networks (CNN) are rapidly emerging as a new method for cancer image analysis, with significantly enhanced predictive power compared to hand-crafted radiomics features. Here we show that CNNs trained to perform the tumor segmentation task, with no other information than physician contours, identify a rich set of survival-related image features with remarkable prognostic value. In a retrospective study on pre-treatment PET-CT images of 96 NSCLC patients before stereotactic-body radiotherapy (SBRT), we found that the CNN segmentation algorithm (U-Net) trained for tumor segmentation in PET and CT images, contained features having strong correlation with 2- and 5-year overall and disease-specific survivals. The U-Net algorithm has not seen any other clinical information (e.g. survival, age, smoking history, etc.) than the images and the corresponding tumor contours provided by physicians. In addition, we observed the same trend by validating the U-Net features against an extramural data set provided by Stanford Cancer Institute. Furthermore, through visualization of the U-Net, we also found convincing evidence that the regions of metastasis and recurrence appear to match with the regions where the U-Net features identified patterns that predicted higher likelihoods of death. We anticipate our findings will be a starting point for more sophisticated non-intrusive patient specific cancer prognosis determination. For example, the deep learned PET/CT features can not only predict survival but also visualize high-risk regions within or adjacent to the primary tumor and hence potentially impact therapeutic outcomes by optimal selection of therapeutic strategy or first-line therapy adjustment.
more »
« less
This content will become publicly available on June 11, 2026
Improving lung cancer diagnosis and survival prediction with deep learning and CT imaging
Lung cancer is a major cause of cancer-related deaths, and early diagnosis and treatment are crucial for improving patients’ survival outcomes. In this paper, we propose to employ convolutional neural networks to model the non-linear relationship between the risk of lung cancer and the lungs’ morphology revealed in the CT images. We apply a mini-batched loss that extends the Cox proportional hazards model to handle the non-convexity induced by neural networks, which also enables the training of large data sets. Additionally, we propose to combine mini-batched loss and binary cross-entropy to predict both lung cancer occurrence and the risk of mortality. Simulation results demonstrate the effectiveness of both the mini-batched loss with and without the censoring mechanism, as well as its combination with binary cross-entropy. We evaluate our approach on the National Lung Screening Trial data set with several 3D convolutional neural network architectures, achieving high AUC and C-index scores for lung cancer classification and survival prediction. These results, obtained from simulations and real data experiments, highlight the potential of our approach to improving the diagnosis and treatment of lung cancer.
more »
« less
- Award ID(s):
- 2210388
- PAR ID:
- 10623858
- Editor(s):
- Yanwu, Xu
- Publisher / Repository:
- plos.org
- Date Published:
- Journal Name:
- PLOS One
- Volume:
- 20
- Issue:
- 6
- ISSN:
- 1932-6203
- Page Range / eLocation ID:
- e0323174
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Lung cancer is the second most common cancer in the world. The aim of this study is to identify biomarkers for lung cancer that can aid in its diagnosis and treatment. The gene expression profiles from GEO database were analyzed by GEO2R to identify Differentially Expressed Genes (DEGs) and further analyzed using Cytoscape. The data was divided into two categories: non-treatment and treatment groups. A total of 407 DEGs (254 upregulated and 153 downregulated) and 259 DEGs (124 upregulated and 135 downregulated) were isolated for non-treatment and treatment studies respectively. The significant Gene Ontologies and pathways enriched with DEGS were identified using Cytoscape apps, BiNGO and ReactomeFIPlugIn, respectively. Hub genes based on network parameters - Degree, Closeness and Betweenness - were isolated using CytoHubba. In conclusion, DEGs identified in this study may play an important role in early diagnosis or as biomarkers of lung cancer.more » « less
-
null (Ed.)Positron emission tomography and computed tomography (PET-CT) dual-modality imaging provides critical diagnostic information in modern cancer diagnosis and therapy. Automated accurate tumor delineation is essentially important in computer-assisted tumor reading and interpretation based on PET-CT. In this paper, we propose a novel approach for the segmentation of lung tumors that combines the powerful fully convolutional networks (FCN) based semantic segmentation framework (3D-UNet) and the graph cut based co-segmentation model. First, two separate deep UNets are trained on PET and CT, separately, to learn high level discriminative features to generate tumor/non-tumor masks and probability maps for PET and CT images. Then, the two probability maps on PET and CT are further simultaneously employed in a graph cut based co-segmentation model to produce the final tumor segmentation results. Comparative experiments on 32 PET-CT scans of lung cancer patients demonstrate the effectiveness of our method.more » « less
-
null (Ed.)Abstract We propose a data and knowledge driven approach for SPECT by combining a classical iterative algorithm of SPECT with a convolutional neural network.The classical iterative algorithm, such as ART and ML-EM, is employed to provide the model knowledge of SPECT.A modified U-net is then connected to exploit further features of reconstructed images and data sinograms of SPECT.We provide mathematical formulations for the architecture of the proposed networks.The networks are trained by supervised learning using the technique of mini-batch optimization.We apply the trained networks to the problems of simulated lung perfusion imaging and simulated myocardial perfusion imaging, and numerical results demonstrate their effectiveness of reconstructing source images from noisy data measurements.more » « less
-
Over the last decade, both early diagnosis and targeted therapy have improved the survival rates of many cancer patients. Most recently, immunotherapy has revolutionized the treatment options for cancers such as melanoma. Unfortunately, a significant portion of cancers (including lung and breast cancers) do not respond to immunotherapy, and many of them develop resistance to chemotherapy. Molecular characterization of non-responsive cancers suggest that an embryonic program known as epithelial-mesenchymal transition (EMT), which is mostly latent in adults, can be activated under selective pressures, rendering these cancers resistant to chemo- and immunotherapies. EMT can also drive tumor metastases, which in turn also suppress the cancer-fighting activity of cytotoxic T cells that traffic into the tumor, causing immunotherapy to fail. In this review, we compare and contrast immunotherapy treatment options of non-small cell lung cancer (NSCLC) and triple negative breast cancer (TNBC). We discuss why, despite breakthrough progress in immunotherapy, attaining predictable outcomes in the clinic is mostly an unsolved problem for these tumors. Although these two cancer types appear different based upon their tissues of origin and molecular classification, gene expression indicate that they possess many similarities. Patient tumors exhibit activation of EMT, and resulting stem cell properties in both these cancer types associate with metastasis and resistance to existing cancer therapies. In addition, the EMT transition in both these cancers plays a crucial role in immunosuppression, which exacerbates treatment resistance. To improve cancer-related survival we need to understand and circumvent, the mechanisms through which these tumors become therapy resistant. In this review, we discuss new information and complementary perspectives to inform combination treatment strategies to expand and improve the anti-tumor responses of currently available clinical immune checkpoint inhibitors.more » « less
