PO-1655 Tuning deep learning models for automatic segmentation of head and neck cancers in PET/CT images

Determining progression-free survival (PFS) for head and neck squamous cell carcinoma (HNSCC) patients is a challenging but pertinent task that could help stratify patients for improved overall outcomes. PET/CT images provide a rich source of anatomical and metabolic data for potential clinical biomarkers that would inform treatment decisions and could help improve PFS. In this study, we participate in the 2021 HECKTOR Challenge to predict PFS in a large dataset of HNSCC PET/CT images using deep learning approaches. We develop a series of deep learning models based on the DenseNet architecture using a negative log-likelihood loss function that utilizes PET/CT images and clinical data as separate input channels to predict PFS in days. Internal model validation based on 10-fold cross-validation using the training data (N=224) yielded C-index values up to 0.622 (without) and 0.842 (with) censoring status considered in C-index computation, respectively. We then implemented model ensembling approaches based on the training data cross-validation folds to predict the PFS of the test set patients (N=101). External validation on the test set for the best ensembling method yielded a C-index value of 0.694. Our results are a promising example of how deep learning approaches can effectively utilize imaging and clinical data for medical outcome prediction in HNSCC, but further work in optimizing these processes is needed.

Download Full-text

PO-0967 Prediction of treatment outcome for head and neck cancers using radiomics of PET/CT images

Radiotherapy and Oncology ◽

10.1016/s0167-8140(19)31387-8 ◽

2019 ◽

Vol 133 ◽

pp. S526

Author(s):

A. Rosvoll Groendahl ◽

A.D. Midtfjord ◽

G.S. Elvatun Rakh Langberg ◽

O. Tomic ◽

U.G. Indahl ◽

...

Keyword(s):

Treatment Outcome ◽

Head And Neck ◽

Ct Images ◽

Head And Neck Cancers ◽

Pet Ct

Download Full-text

Fully Automated Delineation of Gross Tumor Volume for Head and Neck Cancer on PET-CT Using Deep Learning: A Dual-Center Study

Contrast Media & Molecular Imaging ◽

10.1155/2018/8923028 ◽

2018 ◽

Vol 2018 ◽

pp. 1-12 ◽

Cited By ~ 14

Author(s):

Bin Huang ◽

Zhewei Chen ◽

Po-Man Wu ◽

Yufeng Ye ◽

Shi-Ting Feng ◽

...

Keyword(s):

Head And Neck Cancer ◽

Deep Learning ◽

Head And Neck ◽

Neck Cancer ◽

Gold Standard ◽

Gross Tumor Volume ◽

Tumor Volume ◽

Ct Images ◽

Dice Similarity Coefficient ◽

Pet Ct

Purpose. In this study, we proposed an automated deep learning (DL) method for head and neck cancer (HNC) gross tumor volume (GTV) contouring on positron emission tomography-computed tomography (PET-CT) images. Materials and Methods. PET-CT images were collected from 22 newly diagnosed HNC patients, of whom 17 (Database 1) and 5 (Database 2) were from two centers, respectively. An oncologist and a radiologist decided the gold standard of GTV manually by consensus. We developed a deep convolutional neural network (DCNN) and trained the network based on the two-dimensional PET-CT images and the gold standard of GTV in the training dataset. We did two experiments: Experiment 1, with Database 1 only, and Experiment 2, with both Databases 1 and 2. In both Experiment 1 and Experiment 2, we evaluated the proposed method using a leave-one-out cross-validation strategy. We compared the median results in Experiment 2 (GTVa) with the performance of other methods in the literature and with the gold standard (GTVm). Results. A tumor segmentation task for a patient on coregistered PET-CT images took less than one minute. The dice similarity coefficient (DSC) of the proposed method in Experiment 1 and Experiment 2 was 0.481∼0.872 and 0.482∼0.868, respectively. The DSC of GTVa was better than that in previous studies. A high correlation was found between GTVa and GTVm (R = 0.99, P<0.001). The median volume difference (%) between GTVm and GTVa was 10.9%. The median values of DSC, sensitivity, and precision of GTVa were 0.785, 0.764, and 0.789, respectively. Conclusion. A fully automatic GTV contouring method for HNC based on DCNN and PET-CT from dual centers has been successfully proposed with high accuracy and efficiency. Our proposed method is of help to the clinicians in HNC management.

Download Full-text

A comparison of fully automatic segmentation of tumors and involved nodes in PET/CT of head and neck cancers

Physics in Medicine and Biology ◽

10.1088/1361-6560/abe553 ◽

2021 ◽

Author(s):

Aurora Rosvoll Groendahl ◽

Ingerid Skjei Knudtsen ◽

Bao Ngoc Huynh ◽

Martine Mulstad ◽

Yngve Mardal Mardal Moe ◽

...

Keyword(s):

Head And Neck ◽

Automatic Segmentation ◽

Head And Neck Cancers ◽

Pet Ct ◽

Fully Automatic

Download Full-text

A deep learning method for automatic segmentation of the bony orbit in MRI and CT images

Scientific Reports ◽

10.1038/s41598-021-93227-3 ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Jared Hamwood ◽

Beat Schmutz ◽

Michael J. Collins ◽

Mark C. Allenby ◽

David Alonso-Caneiro

Keyword(s):

Deep Learning ◽

Human Performance ◽

Automatic Segmentation ◽

Ct Images ◽

Human Observer ◽

In Series ◽

Magnetic Resonance Imaging Mri ◽

High Degree ◽

Orbital Region ◽

Ct And Mri

AbstractThis paper proposes a fully automatic method to segment the inner boundary of the bony orbit in two different image modalities: magnetic resonance imaging (MRI) and computed tomography (CT). The method, based on a deep learning architecture, uses two fully convolutional neural networks in series followed by a graph-search method to generate a boundary for the orbit. When compared to human performance for segmentation of both CT and MRI data, the proposed method achieves high Dice coefficients on both orbit and background, with scores of 0.813 and 0.975 in CT images and 0.930 and 0.995 in MRI images, showing a high degree of agreement with a manual segmentation by a human expert. Given the volumetric characteristics of these imaging modalities and the complexity and time-consuming nature of the segmentation of the orbital region in the human skull, it is often impractical to manually segment these images. Thus, the proposed method provides a valid clinical and research tool that performs similarly to the human observer.

Download Full-text

Vulnerability in Deep Transfer Learning Models to Adversarial Fast Gradient Sign Attack for COVID-19 Prediction from Chest Radiography Images

Applied Sciences ◽

10.3390/app11094233 ◽

2021 ◽

Vol 11 (9) ◽

pp. 4233

Author(s):

Biprodip Pal ◽

Debashis Gupta ◽

Md. Rashed-Al-Mahfuz ◽

Salem A. Alyami ◽

Mohammad Ali Moni

Keyword(s):

Deep Learning ◽

Transfer Learning ◽

Chest Radiography ◽

High Sensitivity ◽

Ct Images ◽

Classification Performance ◽

Learning Models ◽

X Ray ◽

Image Capturing ◽

Fast Gradient

The COVID-19 pandemic requires the rapid isolation of infected patients. Thus, high-sensitivity radiology images could be a key technique to diagnose patients besides the polymerase chain reaction approach. Deep learning algorithms are proposed in several studies to detect COVID-19 symptoms due to the success in chest radiography image classification, cost efficiency, lack of expert radiologists, and the need for faster processing in the pandemic area. Most of the promising algorithms proposed in different studies are based on pre-trained deep learning models. Such open-source models and lack of variation in the radiology image-capturing environment make the diagnosis system vulnerable to adversarial attacks such as fast gradient sign method (FGSM) attack. This study therefore explored the potential vulnerability of pre-trained convolutional neural network algorithms to the FGSM attack in terms of two frequently used models, VGG16 and Inception-v3. Firstly, we developed two transfer learning models for X-ray and CT image-based COVID-19 classification and analyzed the performance extensively in terms of accuracy, precision, recall, and AUC. Secondly, our study illustrates that misclassification can occur with a very minor perturbation magnitude, such as 0.009 and 0.003 for the FGSM attack in these models for X-ray and CT images, respectively, without any effect on the visual perceptibility of the perturbation. In addition, we demonstrated that successful FGSM attack can decrease the classification performance to 16.67% and 55.56% for X-ray images, as well as 36% and 40% in the case of CT images for VGG16 and Inception-v3, respectively, without any human-recognizable perturbation effects in the adversarial images. Finally, we analyzed that correct class probability of any test image which is supposed to be 1, can drop for both considered models and with increased perturbation; it can drop to 0.24 and 0.17 for the VGG16 model in cases of X-ray and CT images, respectively. Thus, despite the need for data sharing and automated diagnosis, practical deployment of such program requires more robustness.

Download Full-text