izpis_h1_title_alt

Napovedovanje prihodnjega napredovanja bolezni multiple skleroze z morfometrično analizo magnetnoresonančne slike glave
ID Zagožen, Veronika (Author), ID Špiclin, Žiga (Mentor) More about this mentor... This link opens in a new window, ID Dular, Lara (Comentor)

.pdfPDF - Presentation file, Download (2,87 MB)
MD5: CD589284C8814939EE43F8B599BC6065

Abstract
Opis in motivacija problema: Multipla skleroza (ang. Multiple sclerosis, MS) je ena izmed najpogostejših avtoimunskih bolezni osrednjega živčevja pri mladih odraslih osebah in nastaja zaradi avtoimunskega odziva proti mielinski ovojnici živčevja. Patologija bolezni MS se odraža kot klasično vnetje v okolici kapilar in mielinske ovojnice, ki se na magnetnoresonančnih (MR) slikah kaže v obarvanih področjih, imenovanih lezije. Bolezen MS običajno poteka eni izmed treh oblik: (i) recidivno-reminentna faza, (ii) sekundarno-progresivna faza ter (iii) primarno-progresivna faza. Za namen diagnostike in spremljanja poteka bolezni MS se uporablja predvsem trodimenzionalne (3D) MR slike možganov. Priporočljiv standardiziran protokol za zajem 3D MR slik možganov so T1 utežena (ang. T1-weighted, T1w) in FLAIR modaliteta zajema slik. Za namene napovedovanja prihodnjega poteka bolezni MS se uporablja meritve oziroma biomarkerje pridobljene iz MR slik, kot so meritve volumnov celotne možganovine, atrofija posameznih regij možganov, spremembe v volumnu sive možganovine, spremembe v volumnu in številu lezij po času glede na razvoj bolezni MS, itd. Klinična ocena napredovanja je podana z razširjeno lestvico prizadetosti (ang. expanded disability status scale, EDSS). Napovedovanje prihodnjega napredovanja prizadetosti zaradi bolezni MS oz. poteka EDSS je aktualna znanstvena tema, pri čemer zadnje raziskave v znanstvenih člankih poročajo sposobnost razvrščanja bolnikov z napredovanjem v bolezni MS z do 80% točnostjo, kar kaže na potencialno uporabnost metodologije za implementacijo le te kot prognostičnega orodja v procesu zdravljenja bolezni MS. Podatki: Podatkovna baza uporabljena v magistrski nalogi je nastala v okviru študije Artificial Intelligence in predicting Progression in Multiple Sclerosis study (AI ProMiS) in vsebuje 3D magnetno resonančne (MR) slike bolnikov z multiplo sklerozo (MS), demografske in klinične podatke o bolnikih. Vsebuje 1284 T1w in FLAIR MR slik pridobljenih za 486 bolnikov, od tega 71,3% ženskega in 28,7% moškega spola, s povprečno starostjo 39,7 ± 10,3 let. Končna množica podatkov sestoji iz meritev volumna in števila lezij, informacij o pacientu, kot so starost, spol, vrednosti EDSS, ter meritve volumnov možganskih struktur ter pripadajoče normalizirane volumne glede na volumen znotrajlobanjskega prostora možganskih struktur in meritve asimetrije med pripadajočimi strukturami možganov v levi in desni hemisferi. Razmerje bolnikov s prihodnjim napredovanjem v bolezni MS oz. brez napredovanja je v dani podatkovni množici v razmerju 1:4. Metode: Zasnova učenja in vrednotenja napovednih modelov vključuje tri glavne korake: (i) izbira relevantnih značilnic, (ii) preslikava prostora značilnic v prostor nižje dimenzije ter (iii) metodo razvrščanja. Uporabljene so bile metode za izbiro značilnic na podlagi korelacijskega filtra (ang. Correlation-based Feature Selection, CFS), metoda z rekurzivno eliminacijo značilnic (ang. Recursive Feature Elimination, RFE), metoda LASSO (ang. least absolute shrinkage and selection operator, LASSO) in genetski algoritem (ang. Genetic Algorithm, GA). Za namene zmanjšanja in preslikave dimenzij prostora vhodnih značilnic se je uporabila metoda analize glavnih komponent (ang. Principal Component Analysis, PCA). Uporabljeni modeli za razvrščanje vzorcev v razrede so bili metoda K-najbližjih sosedov (ang. k-nearest neighbors, KNN), naključni gozdovi (ang. Random Forest, RF) in metoda podpornih vektorjev (ang. Support Vector Machines, SVM). Kombinacije metod izbire značilnic, z ali brez PCA, in modelov razvrščanja smo učili in vrednotili s štirikoračno križno validacijo in kot mere sposobnosti razvrščanja izračunali površino pod delovno karakteristiko sprejemnika (ang. area under the receiver operating curve, AUC), točnost, občutljivost in specifičnost. Rezultati: Z vidika primerjave med izbranimi metodami za izbiro optimalne podskupine značilnic se je kot najuspešnejša pokazala metoda s korelacijskim filtrom, ki daje v kombinaciji z vsemi tremi razvrščevalnimi modeli najboljše rezultate. Najboljše rezultate smo pridobili z uporabo metode s korelacijskim filtrom v povezavi s PCA metodo ter SVM razvrščevalnikom (ang. support vector classifier, SVC), ki je podal vrednost metrik uspešnosti razvrščanja vzorcev v razrede: AUC 0,77, točnosti 0,69, občutljivosti 0,72 in specifičnosti 0,68. Iz rezultatov je viden tudi pozitiven vpliv implementacije metode za zmanjšanje dimenzije oz. preslikavo vhodnih podatkov PCA, ki v splošnem izboljšajo delovanje modelov. Zaključek: Rezultati eksperimentov pridobljeni tekom magistrske naloge so navkljub različni in v našem primeru precej heterogeni množici MRI slik, pridobljeni na petih različnih skenerjih iz štirih različnih inštitucij, razmeroma primerljivi z najboljšimi rezultati pretekle študije [45] in potrjujejo hipotezo o zmožnosti napovedi prihodnjega napredovanja bolezni MS na osnovi meritev možganskih struktur pridobljenih iz T1w in FLAIR MR slik.

Language:Slovenian
Keywords:Multipla skleroza, avtoimunska bolezen, osrednji živčni sistem, magnetna resonanca (MRI), prihodnje napredovanje bolezni, izbor značilnic, analiza glavnih komponent (PCA), metode podpornih vektorjev (SVM), k-najbližjih sosedov (KNN), naključni gozdovi (RF), EDSS.
Work type:Master's thesis/paper
Organization:FE - Faculty of Electrical Engineering
Year:2023
PID:20.500.12556/RUL-148792 This link opens in a new window
COBISS.SI-ID:165739267 This link opens in a new window
Publication date in RUL:31.08.2023
Views:1321
Downloads:183
Metadata:XML RDF-CHPDL DC-XML DC-RDF
:
Copy citation
Share:Bookmark and Share

Secondary language

Language:English
Title:Prognosis of future progression of multiple sclerosis disease based on morphometric analysis of brain magnetic resonance scan
Abstract:
Problem motivation: Multiple sclerosis (MS) is one of the most common autoimmune diseases of the central nervous system in young adults, causing an autoimmune response against myelin nerve sheath. The pathology of MS is located entirely in the central nervous system and presents as classic inflammation around capillaries and myelin sheaths, visible as colored areas or lesions on magnetic resonance imaging (MRI) scans. The disease typically progresses in one of three phases: (i) relapsing-remitting, (ii) secondary-progressive, and (iii) primary-progressive. For the purpose of diagnosing and monitoring the progression of MS, three-dimensional (3D) MRI brain scans are commonly used. The recommended standardized protocol for acquiring 3D MRI brain images includes T1-weighted (T1w) and FLAIR modalities. For predicting the future course of MS, particularly the risk of increased disability according to the expanded disability status scale (EDSS), measurements or biomarkers obtained from MRI scans are used. These biomarkers include measurements of total brain volume, atrophy of specific brain regions, changes in ventricular volume, changes in gray matter volume, changes in lesion volume and count over time with the progression of MS, etc. The topic of predicting the future progression of MS, namely the course of EDSS, is an active research field, where the recent research articles report the ability of predictive models to classify patients with disease progression with up to 80% accuracy, indicating the potential usefulness of these models as a diagnostic tool in the treatment process of MS. Data: The database used in this master's thesis was created as part of the Artificial Intelligence in predicting Progression in Multiple Sclerosis study (AI ProMiS) and contains 3D MRI scans of patients with MS, along with demographic and clinical patient data [16]. The dataset consists of 1284 T1w and FLAIR MRI scans obtained for 486 patients, with 71.3% being female and 28.7% male, with an average age of 39.7 ± 10.3 years. The final dataset consists of measurements of brain volumes and lesion counts, patient information such as age, gender, EDSS scores, as well as volumes of healthy brain structure, corresponding volumes normalized relative to the intracranial volume, and asymmetry of corresponding left-right brain regions across the hemispheres. The ratio of patients with future disease progression in MS and those without is 1:4. Methods: The design of training and evaluation of predictive models included three main steps: (i) selection of relevant features, (ii) mapping of feature space into a lower-dimensional space, and (iii) classification method. The methods used for feature selection were the Correlation-based Feature Selection (CFS) method, Recursive Feature Elimination (RFE) method, Least Absolute Shrinkage and Selection Operator (LASSO) method, and Genetic Algorithms (GA). For the purpose of reducing the dimensionality of input features, the Principal Component Analysis (PCA) method was used. The employed models for classifying features into classes of patient with and without future disease progression were the k-nearest neighbors (KNN) method, Random Forest (RF), and Support Vector Machines (SVM) method. Combinations of feature selection, with or without the use of PCA, and classification methods were trained and test using four-fold cross-validation and then computed the overall performance metrics like the area under the receiver operating curve (AUC), accuracy, sensitivity and specificity. Results: From the perspective of comparing the selected methods for choosing the optimal subset of features, the correlation-based feature selection method proved optimal, providing the best results in combination with all three classification models. The best overall results were obtained using the method with the correlation filter in combination with the PCA and the SVC classifier, which yielded an AUC metric value of 0.77, accuracy of 0.69, sensitivity of 0.72, and specificity of 0.68 . Furthermore, the results indicate the positive impact of implementing the PCA method for reducing the dimensionality and/or mapping of input data, which generally improved the performance of the models. Conclusion: Despite the use of different validation datasets, with our being composed of MRIs from five different scanners and four different institutions and thus the most heterogeneous, the obtained results of this master's thesis demonstrate rather good reproducibility of the best findings of a previous study [45] and thereby reaffirm the hypothesis regarding the ability to predict the progression of MS based on data from measurements of healthy and pathological brain regions obtained from T1w and FLAIR MRI scans.

Keywords:Multiple sclerosis, Autoimmune disease, Central nervous system, Magnetic resonance imaging (MRI), Disease progression, Feature selection, Principal Component Analysis (PCA), Support Vector Machines (SVM), k-nearest neighbors (KNN), Random Forest (RF), EDSS.

Similar documents

Similar works from RUL:
Similar works from other Slovenian collections:

Back