Resting state functional MRI (RS-fMRI) and arterial spin labeling (ASL) MRI are widely used in research field in order to investigate functional connectivity (FC) and brain perfusion (namely, cerebral blood flow, CBF), respectively. In this study, two major neurodegenerative pathologies, Parkinson’s disease (PD) and remitting relapsing multiple sclerosis (RRMS), are considered by single modalities and by merging of the two MRI techniques, in order to assess if altered CBF and neurovascular uncoupling (NVU) may bias FC results. Previous studies (Pillai and Mikulis, 2015; Para et al., 2017) have proved that the physiological cascade leading to the hemodynamic manifestation of neural activation could be modified in presence of pathologies that cause perfusion impairments, thus leading to the lack or reduction of BOLD signal, which, in turn, can be misinterpreted as decreased or absent neural activation. This study was conducted in the MRI Laboratory of Don C. Gnocchi Foundation, IRCCS Santa Maria Nascente, in Milan. Data analysis was performed by means of FMRIB’s Software Library (FSL). Two different studies were considered: twenty-five PD patients and twenty age-matched healthy control (HC) subjects (HC-PD); thirty-four RRMS patients and thirty-four age-matched HC subjects (HC-RRMS). Concerning ASL analysis, a sub-set of forty-nine subjects belonging to the HC-RRMS study underwent acquisition in hypercapnia condition for cerebrovascular reactivity (CVR) evaluation. MRI data were acquired by a 1.5 T Siemens Magnetom Avanto scanner (Siemens AG Medical Solutions, Erlangen, Germany) equipped with a 32-channel headcoil. A preliminary, though major, part of the work addressed the assessment of acquisition/processing pipelines robustness in the addressed clinical studies. In RS-fMRI, image distortion correction, denoising, and robustness of independent component analysis (ICA) were tested. In ASL repeatability was assessed by a scan/rescan procedure involving subject repositioning. The presence of pathology-related alterations was evaluated both in FC and in CBF/CVR. Firstly, single modality analyses were performed. Next, merging and data fusion methods were applied, including a specific ICA-based algorithm (Linked ICA). The main results showed a reduction of FC in sensory motor network for PD patients, whereas perfusion alteration was absent. This, in turn, was confirmed by the fused data analyses, thus concluding that the FC alteration in the PD group was due to neural activity damage with no bias due to perfusion impairment. Concerning the study involving RRMS, no alterations were found, principally because of the early stage of the disease and for the insufficient statistical power. In conclusion, the proposed protocol merging FC and CBF was shown to be sufficiently robust for clinical application and effective in the evaluation of PD, with potential extension to other neurodegenerative diseases.
La RM funzionale ‘resting state’ e la RM ‘arterial spin labeling’ (ASL) sono tecniche ampiamente utilizzate in campo di ricerca per investigare rispettivamente la connettività funzionale (CF) e la perfusione (comunemente flusso sanguigno cerebrale, FSC). In questo studio, vengono considerate due principali patologie neurodegenerative: la malattia di Parkinson e la sclerosi multipla recidivante-remissiva, tramite analisi a singola modalità e fusione delle due tecniche di RM, per la valutazione dell’influenza delle alterazioni di FSC sulla CF dovuta a disaccoppiamento neuro vascolare. Studi precedenti (Pillai and Mikulis, 2015; Para et al., 2017) hanno dimostrato che la cascata di eventi fisiologici che conduce alla manifestazione emodinamica dell’attivazione neurale potrebbe risultare modificata in presenza di patologie che causano danni di perfusione, e dunque condurre all’assenza o riduzione del segnale di BOLD, il quale, a sua volta, può essere male interpretato come diminuzione o assenza di attivazione. Lo studio è stato condotto presso il laboratorio di RM alla Fondazione Don Gnocchi, IRCCS Santa Maria Nascente, a Milano. L’analisi dei dati è stata effettuata tramite FMRIB’s Software Library (FSL). Sono stati considerati due differenti studi: venticinque pazienti PD e venti soggetti sani; trentaquattro pazienti RRMS e trentaquattro soggetti sani. Per lo studio ASL un sottogruppo di quarantanove soggetti (HC-RRMS) è stato sottoposto ad acquisizione in condizioni di ipercapnia per la valutazione della reattività cerebrovascolare (RCV). I dati di risonanza sono stati acquisiti tramite scanner Siemens Magnetom Avanto (Siemens AG Medical Solutions, Erlangen, Germania). Uno studio preliminare, sebbene importante, è stato rivolto alla valutazione della robustezza delle pipeline di acquisizione/elaborazione nello specifico studio clinico. In RS-fMRI sono state testate la correzione delle distorsioni, la rimozione del rumore e la robustezza dell’analisi delle componenti principali. Riguardo l’analisi ASL, è stata testata l’acquisizione ripetuta (2 scan successive) con riposizionamento. La presenza di alterazioni legate alle patologie è stata valuta sia riguardo alla CF che riguardo al FSC/RCV. Inizialmente, è stata effettuata un’analisi a singola modalità. Successivamente, sono stati applicati metodi di unione e fusione dei dati, incluso l’utilizzo di uno specifico algoritmo ICA (Linked ICA). I risultati principali hanno mostrato una riduzione di CF nella rete sensoriale-motoria per i pazienti PD, mentre non sono state riscontrate alterazioni nella perfusione. Questo, a sua volta, è stato confermato dalle analisi di fusione dei dati, perciò si è potuto dedurre che l’alterazione di CF nei pazienti PD fosse dovuta ad un danno relativo all’attivazione neurale stessa che non risulta essere stata influenzata da modificazioni di perfusione. Riguardo lo studio che comprende pazienti RRMS, non sono state rilevate alterazioni, principalmente a causa dello stadio precoce della patologia e dell’insufficiente potere statistico. In conclusione, il protocollo di fusione di analisi di CF e FSC è risultato sufficientemente robusto per l’applicazione clinica ed efficace nella valutazione della malattia di Parkinson e potenzialmente estendibile ad altre patologie neurodegenerative.
Study of brain perfusion via arterial spin labeling as possible confounding factor of functional connectivity in neurodegenerative diseases
PIRASTRU, ALICE
2016/2017
Abstract
Resting state functional MRI (RS-fMRI) and arterial spin labeling (ASL) MRI are widely used in research field in order to investigate functional connectivity (FC) and brain perfusion (namely, cerebral blood flow, CBF), respectively. In this study, two major neurodegenerative pathologies, Parkinson’s disease (PD) and remitting relapsing multiple sclerosis (RRMS), are considered by single modalities and by merging of the two MRI techniques, in order to assess if altered CBF and neurovascular uncoupling (NVU) may bias FC results. Previous studies (Pillai and Mikulis, 2015; Para et al., 2017) have proved that the physiological cascade leading to the hemodynamic manifestation of neural activation could be modified in presence of pathologies that cause perfusion impairments, thus leading to the lack or reduction of BOLD signal, which, in turn, can be misinterpreted as decreased or absent neural activation. This study was conducted in the MRI Laboratory of Don C. Gnocchi Foundation, IRCCS Santa Maria Nascente, in Milan. Data analysis was performed by means of FMRIB’s Software Library (FSL). Two different studies were considered: twenty-five PD patients and twenty age-matched healthy control (HC) subjects (HC-PD); thirty-four RRMS patients and thirty-four age-matched HC subjects (HC-RRMS). Concerning ASL analysis, a sub-set of forty-nine subjects belonging to the HC-RRMS study underwent acquisition in hypercapnia condition for cerebrovascular reactivity (CVR) evaluation. MRI data were acquired by a 1.5 T Siemens Magnetom Avanto scanner (Siemens AG Medical Solutions, Erlangen, Germany) equipped with a 32-channel headcoil. A preliminary, though major, part of the work addressed the assessment of acquisition/processing pipelines robustness in the addressed clinical studies. In RS-fMRI, image distortion correction, denoising, and robustness of independent component analysis (ICA) were tested. In ASL repeatability was assessed by a scan/rescan procedure involving subject repositioning. The presence of pathology-related alterations was evaluated both in FC and in CBF/CVR. Firstly, single modality analyses were performed. Next, merging and data fusion methods were applied, including a specific ICA-based algorithm (Linked ICA). The main results showed a reduction of FC in sensory motor network for PD patients, whereas perfusion alteration was absent. This, in turn, was confirmed by the fused data analyses, thus concluding that the FC alteration in the PD group was due to neural activity damage with no bias due to perfusion impairment. Concerning the study involving RRMS, no alterations were found, principally because of the early stage of the disease and for the insufficient statistical power. In conclusion, the proposed protocol merging FC and CBF was shown to be sufficiently robust for clinical application and effective in the evaluation of PD, with potential extension to other neurodegenerative diseases.| File | Dimensione | Formato | |
|---|---|---|---|
|
2017_07_Pirastru.pdf
solo utenti autorizzati dal 05/07/2020
Descrizione: Testo della tesi
Dimensione
4.27 MB
Formato
Adobe PDF
|
4.27 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/135436