Virtual Library
Start Your Search
Debora Gil
Moderator of
-
+
SH02 - Highlight of the Previous Day (ID 99)
- Event: WCLC 2019
- Type: Highlight of the Previous Day Session
- Track:
- Presentations: 6
- Now Available
- Moderators:Grace Yang, Debora Gil
- Coordinates: 9/10/2019, 11:30 - 13:00, Melbourne (1991)
-
+
SH02.01 - Locally Advanced (Now Available) (ID 3667)
11:30 - 13:00 | Presenting Author(s): Rolf Stahel
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
SH02.02 - Pathology (Now Available) (ID 3663)
11:30 - 13:00 | Presenting Author(s): Francisco Perez Ochoa
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
SH02.03 - Thymoma and Other Thoracic Malignancies (Now Available) (ID 3664)
11:30 - 13:00 | Presenting Author(s): Virginie Westeel
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
SH02.04 - Screening (Now Available) (ID 3665)
11:30 - 13:00 | Presenting Author(s): Laureano Molins
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
SH02.05 - SCLC/NET (Now Available) (ID 3666)
11:30 - 13:00 | Presenting Author(s): Antonio Calles Blanco
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
SH02.06 - Diagnosis/Staging (Now Available) (ID 3856)
11:30 - 13:00 | Presenting Author(s): Sabita Jiwnani
- Abstract
- Presentation
Abstract not provided
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
Author of
-
+
ES08 - Critical Concerns in Screening (ID 11)
- Event: WCLC 2019
- Type: Educational Session
- Track: Screening and Early Detection
- Presentations: 1
- Now Available
- Moderators:Ugo Pastorino, John Kirkpatrick Field
- Coordinates: 9/08/2019, 13:30 - 15:00, Copenhagen (1980)
-
+
ES08.05 - Advances in Artificial Intelligence - How Lung Cancer CT Screening Will Progress? (Now Available) (ID 3195)
13:30 - 15:00 | Presenting Author(s): Debora Gil
- Abstract
- Presentation
Abstract
Predictive models for personalized medicine (also known as radiomics) is a recent discipline that uses sophisticated image analysis and artificial intelligence (AI) methods to obtain quantitative image-based features that correlate to final diagnosis and treatment outcome [1].
The application of radiomics in lung cancer screening can represent a critical shift in this field. Some recent studies, like [2-3], show that radiomic features (including tumor shape descriptors and texture analysis) extracted from CT scans have significantly better predictive value than volumetry alone (AUC= 0.9 vs 0.74). Texture analysis reflects tumour heterogeneity and has recently introduced in PET images. In fact, PET texture analysis has demonstrated its value in establishing survival [4], predicting distant metastasis [5], detecting mutations and establishing radiotherapy doses [6]. However, and despite the promising results, there are some limitations like the low reliability of heterogeneity parameters in tumours with small volume, the low repeatability and reproducibility of textural features in the clinical setting and the limitation of the analytic methods.
A multi-radiomic model that could integrate morphological features from the CT together with biological characteristics from the PET and clinical risk factors (age, smoking history, contact with asbestos or family cancer background), would become a highly accurate diagnostic and prognostic method and, thus, make lung cancer screening programs cost-effective. However, in order that radiomics become the cornerstone for clinical decision-making, new machine learning and statistical strategies adapted to the specific requirements of clinical applications should be formulated.
A main pitfall in current state of the art AI methods is the use of generic machine learning and statistical tools borrowed from other fields of application which fall short under clinical conditions [7]. Predictive radiomic models for personalized medicine should address several specific challenges different from the ones common to other application areas of artificial intelligence. First, models should collect and integrate diverse multimodal data sources in a quantitative manner that delivers unambiguous clinical predictions. Second, models should also be easily interpreted from a clinical point of view to allow the analysis of the clinical factors that have an impact on the clinical decision. Third, predictions should be robust concerning data uncertainties due to the impact of collection conditions (like acquisition parameters or variability in manual annotations) and the presence of rare and/or outlying cases, which become highly influential for minority classes lead to overfitting.
This work reviews state-of-the-art AI methods for radiomics, the specific challenges that they must face in medical imaging applications and the latest advances for reliable personalized early diagnosis of lung cancer.
References
[1] P Lambin, et al, Radiomics: the bridge between medical imaging and personalized medicine, Nature Reviews,12, 749-53, 2017.
[2] Hawkins et al. Prediction of pathological nodal involvement by CT-based Radiomic features of the primary tumor in patients with clinically node-negative peripheral lung adenocarcinomas, Med. Phys. 45 (6), 2018.
[3] Peikert T et al. Novel high-resolution computed tomography-based radiomic classifier for screen-identified pulmonary nodules in the National Lung Screening Trial, PLOS ONE 13(10), 2018.
[4] Ohri N, Duan F, Snyder BS, Wei B, Machtay M, Alavi A, et al. Pretreatment 18F-FDG PET textural features in locally Advanced non-small cell lung cancer: secondary analysis of ACRIN 6668/RTOG 0235. J Nucl Med.57:842–8, 2016.
[5] Wu J,Aguilera, et al. Early-stage non-small cell lung cancer: quantitative imaging characteristics of (18)F fluorodeoxyglucose PET/CT allow prediction of distant metastasis. Radiology, 281:270–8, 2016.
[6] Yip SS, et al. Associations between somatic mutations and metabolic imaging phenotypes in non-small cell lung cancer. J Nucl Med. 58:569–76, 2017.
[7] JP. Cohen et al, Distribution matching losses can hallucinate features in medical image translation, MICCAI 2018.
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
