E-Locus - Institutional Repository of the University of Crete

Home Collections Type of Work Doctoral theses

Doctoral theses

Search command : Author="Δαλακούρα" And Author="Κατερίνα"

Current Record: 6 of 2440

[Add to Basket]

Identifier

000463762

Title

Multi-modal data integration using machine and deep learning techniques for predicting high-level clinical outcomes of non-small cell lung cancer patients

Alternative Title

Πολυτροπική ενσωμάτωση δεδομένων με χρήση τεχνικών μηχανικής και βαθιάς μάθησης για την πρόβλεψη κλινικών αποτελεσμάτων ασθενών με μη μικροκυτταρικό καρκίνο του πνεύμονα

Author

Τριβιζάκης, Ελευθέριος

Thesis advisor

Σουγλάκος, Ιωάννης

Reviewer

Καραντάνας, Απόστολος
Μαριάς, Κωνσταντίνος

Abstract

Contemporary research on multi-omic analysis is expected to play an important role in the correlation of genotypes with imaging phenotypes by enhancing the impact of medical image analysis and multi-modal data integration in oncology towards precision medicine. The association of medical data with molecular or genetic features of the examined disease can lead to more accurate diagnosis, enhancement of personalized treatment in patients with lung cancer, and will be a key contribution in identifying the underlining molecular mechanisms that cause the disease’s pathogenesis. Furthermore, using machine learning-based multimodal integration to predict the response of specific treatments can spare patients from unnecessary procedures and improve their quality of life. Additionally, this can lead to an optimized cost-benefit ratio, especially when considering new and expensive health care treatments. The examined population of this study consists of patients with primary non-small cell lung cancer with available imaging examinations, transcriptomic, and clinical/pathological data such as smoking history, tumor staging, and mutational burden of specific oncogenes. The latter is particullarly important since the genetic expression of various targeted oncogenes has been likend to therapy response for lung cancer patients. In this thesis, advanced treatment decision systems based on deep learning and artificial intelligence have been developed to predict the high-level clinical outcomes of patients with nonsmall cell lung cancer, including molecural subtypes, survival prediction, and therapy response.

Language

English

Subject

Computed tomography

Image analysis

Multi-omics

Radiomics

Transcriptomics

Ανάλυση εικόνας

Αξονική τομογραφία

Μεταγράγωμα

πολυ-ομική