Recent studies of Alzheimer’s disease (AD) spread suggest that pathology may be transmitted from one brain area to another either via local diffusion or long-way transport via white matter pathways. However, this hypothesis requires more confirmations, and it’s even more unclear whether such models are applicable in non-amnestic AD (naAD), a group of AD phenotypes characterized by relative spared episodic memory at onset and domain-specific cognitive impairments. Few studies to date have in fact addressed the longitudinal spread of disease in naAD, and all of them considering no more than two variants. At first we compared 240 T1-weighted anatomical MRIs from 129 AD patients with elderly controls’ scans to assess atrophy in each of 120 regions-of-interest (ROIs); then we computed disease progression models separately for each phenotype: typical amnestic AD (aAD), logopenic variant primary progressive aphasia (lvPPA), posterior cortical atrophy (PCA), corticobasal syndrome (CBS), and frontal-variant AD (fvAD). All patients had autopsy or cerebrospinal fluid (CSF) evidence of AD pathology. Results from the amnestic cohort were used to determine appropriate parameters for the phase assignment algorithm, based on association with Braak pathology staging. For each AD variant, 4 phases of regional atrophy were defined based on decreasing frequency of atrophy across participants. We observed a unique distribution of accumulating atrophy for each phenotype. Phase 1 ROIs in our model represent the anatomical origin for each phenotype, including: medial temporal lobe (MTL) for the aAD group (spared in the other phenotypes), left lateral temporal lobe for lvPPA, occipito-parietal cortex for PCA, temporo-parietal cortex for CBS, and fronto-temporal cortex for fvAD. We subsequently assigned a phase to each patient MRI scan based on the similarity of regional atrophy patterns with atrophy predicted for the corresponding phenotype at each phase. ROI phases were strongly correlated with available pathological factors, while MRI phase was significantly correlated with demographic and clinical measures. Then we decided to investigate grey matter (GM) change over time in MRIs within a cohort of patients partly overlapping with the sample used for the cross-sectional study, with the exception of CBS patients (insufficient longitudinal data): 17 aAD, 25 lvPPA, 20 PCA, and 12 fvAD patients, compared to 37 matched controls. We analyzed GM volume and its longitudinal change in phase 1 ROIs from the cross-sectional study for naAD variants, and in MTL for aAD. We also investigated longitudinal atrophy outside these areas through an accessory whole-brain analysis, and we compared phenotypes between each other. We observed unique regional patterns of initial atrophy and longitudinal neocortical disease spread with different rates in lvPPA, PCA, and fvAD, which correlated with cognitive impairments. Atrophy spread over time included both proximal and distant regions from the hypothesized focus of disease onset, thus suggesting that multiple mechanisms of disease progression may have been involved; for what concerns the second mechanism, in particular, a measurement of structural connectivity predicted the severity of longitudinal atrophy, thus corroborating the hypothesis of long-distance fiber pathways. In MTL regions, naAD patients had less severe atrophy than aAD patients at baseline, but longitudinal rates did not differ between groups; MTL sparing in naAD may be due to later onset of MTL degeneration than in aAD, considering that older age was associated with atrophy in this area, independent of group. The current study corroborated probable areas of early disease for naAD and showed that each phenotype has a different pattern of atrophy progression across the cortex, providing also important data about pathology transmission.

Recenti studi sulla progressione della malattia d’Alzheimer(AD)suggeriscono che la patologia possa essere trasmessa da un’area all’altra del cervello tramite diffusione locale o lungo le fibre assonali.Tuttavia,quest’ipotesi necessita di maggiori conferme ed è ancora meno chiaro se questi modelli possano essere applicabili alle varianti non amnesiche di AD(naAD),un gruppo di fenotipi AD caratterizzati da relativo risparmio all’esordio della memoria episodica e deficit cognitivi dominio-specifici.Pochi studi ad oggi hanno infatti analizzato la progressione longitudinale della malattia in naAD, tutti al massimo con due sole di queste varianti. Inizialmente abbiamo confrontato 240 RMN pesate in T1 da 129 pazienti AD con RMN di controlli sani per stabilire l’atrofia in 120 regioni di interesse(ROI); poi abbiamo calcolato modelli di progressione di malattia separatamente per ogni fenotipo: AD amnesici(aAD),variante logopenica di afasia primaria progressiva(lvPPA),atrofia corticale posteriore(PCA)e variante frontale di AD(fvAD).Tutti i pazienti avevano evidenza di patologia AD tramite dati autoptici o da liquor cefalorachidiano(CSF).I risultati dalla coorte di aAD sono stati utilizzati per determinare i parametri per l’algoritmo di assegnazione delle fasi,basato sull’associazione con lo staging di patologia di Braak. Per ogni variante AD,4 fasi di atrofia regionale sono state definite sulla base della frequenza decrescente di atrofia tra i soggetti.Abbiamo osservato pattern unici di atrofia cumulativa per ogni fenotipo.Le ROI di fase 1 nel nostro modello rappresentano l’origine anatomica di ogni fenotipo,inclusi: il lobo temporale mesiale(MTL)per il gruppo aAD(risparmiato negli altri gruppi),il lobo temporale sinistro nella lvPPA,la corteccia parieto-occipitale nella PCA,quella temporo-parietale per la CBS e le aree fronto-temporali per la fvAD.Successivamente abbiamo assegnato una fase ad ogni RMN dei pazienti in base alla somiglianza dei pattern di atrofia regionale con l’atrofia predetta per il fenotipo corrispondente ad ogni fase.Le fasi ROI erano correlate con parametri patologici disponibili,mentre le fasi RMN erano correlate con misure demografiche e cliniche. Dopodiché abbiamo deciso di studiare le modifiche nel tempo della sostanza grigia(GM)in una coorte di pazienti in parte sovrapposta a quella usata per lo studio trasversale,ad esclusione di CBS: 17 aAD,25 lvPPA,20 PCA e 12 fvAD,con 37 controlli abbinati.Abbiamo analizzato il volume della GM e le sue modifiche longitudinali nelle ROI di fase 1 dallo studio trasversale per naAD e l’MTL per aAD.Abbiamo anche studiato l’atrofia longitudinale al di fuori di queste aree tramite un’analisi accessoria all’intero cervello e abbiamo comparato i fenotipi tra loro. Abbiamo osservato pattern regionali unici di atrofia iniziale e diffusione longitudinale nella neocorteccia con tassi differenti in lvPPA,PCA e fvAD,che correlavano con i deficit cognitivi.La progressione di atrofia nel tempo ha coinvolto aree sia prossimali sia distanti dal sito d’esordio della malattia,suggerendo quindi più meccanismi di diffusione della patologia coinvolti; per quanto riguarda il secondo,in particolare,una misura di connettività strutturale prediceva la severità di atrofia longitudinale,corroborando quindi l’ipotesi delle vie lunghe.Nell’MTL,i pazienti naAD mostravano al basale meno atrofia dei pazienti aAD,ma i tassi longitudinali non erano diversi tra gruppi; il relativo risparmio dell’MTL in naAD potrebbe quindi essere dovuto a un esordio più tardivo della degenerazione nell’MTL rispetto all’aAD,considerando che l’età più avanzata era associata con atrofia in quest’area,indipendentemente dal gruppo. Il presente studio ha corroborato probabili aree di malattia precoce in naAD e mostrato che ogni fenotipo ha un diverso pattern di progressione di atrofia lungo la corteccia,fornendo anche dati importanti sulla trasmissione della patologia.

(2020). An MRI-based analysis of the longitudinal progression of atrophy in amnestic and non-amnestic phenotypes of Alzheimer’s disease. (Tesi di dottorato, Università degli Studi di Milano-Bicocca, 2020).

An MRI-based analysis of the longitudinal progression of atrophy in amnestic and non-amnestic phenotypes of Alzheimer’s disease

DA RE, FULVIO
2020

Abstract

Recent studies of Alzheimer’s disease (AD) spread suggest that pathology may be transmitted from one brain area to another either via local diffusion or long-way transport via white matter pathways. However, this hypothesis requires more confirmations, and it’s even more unclear whether such models are applicable in non-amnestic AD (naAD), a group of AD phenotypes characterized by relative spared episodic memory at onset and domain-specific cognitive impairments. Few studies to date have in fact addressed the longitudinal spread of disease in naAD, and all of them considering no more than two variants. At first we compared 240 T1-weighted anatomical MRIs from 129 AD patients with elderly controls’ scans to assess atrophy in each of 120 regions-of-interest (ROIs); then we computed disease progression models separately for each phenotype: typical amnestic AD (aAD), logopenic variant primary progressive aphasia (lvPPA), posterior cortical atrophy (PCA), corticobasal syndrome (CBS), and frontal-variant AD (fvAD). All patients had autopsy or cerebrospinal fluid (CSF) evidence of AD pathology. Results from the amnestic cohort were used to determine appropriate parameters for the phase assignment algorithm, based on association with Braak pathology staging. For each AD variant, 4 phases of regional atrophy were defined based on decreasing frequency of atrophy across participants. We observed a unique distribution of accumulating atrophy for each phenotype. Phase 1 ROIs in our model represent the anatomical origin for each phenotype, including: medial temporal lobe (MTL) for the aAD group (spared in the other phenotypes), left lateral temporal lobe for lvPPA, occipito-parietal cortex for PCA, temporo-parietal cortex for CBS, and fronto-temporal cortex for fvAD. We subsequently assigned a phase to each patient MRI scan based on the similarity of regional atrophy patterns with atrophy predicted for the corresponding phenotype at each phase. ROI phases were strongly correlated with available pathological factors, while MRI phase was significantly correlated with demographic and clinical measures. Then we decided to investigate grey matter (GM) change over time in MRIs within a cohort of patients partly overlapping with the sample used for the cross-sectional study, with the exception of CBS patients (insufficient longitudinal data): 17 aAD, 25 lvPPA, 20 PCA, and 12 fvAD patients, compared to 37 matched controls. We analyzed GM volume and its longitudinal change in phase 1 ROIs from the cross-sectional study for naAD variants, and in MTL for aAD. We also investigated longitudinal atrophy outside these areas through an accessory whole-brain analysis, and we compared phenotypes between each other. We observed unique regional patterns of initial atrophy and longitudinal neocortical disease spread with different rates in lvPPA, PCA, and fvAD, which correlated with cognitive impairments. Atrophy spread over time included both proximal and distant regions from the hypothesized focus of disease onset, thus suggesting that multiple mechanisms of disease progression may have been involved; for what concerns the second mechanism, in particular, a measurement of structural connectivity predicted the severity of longitudinal atrophy, thus corroborating the hypothesis of long-distance fiber pathways. In MTL regions, naAD patients had less severe atrophy than aAD patients at baseline, but longitudinal rates did not differ between groups; MTL sparing in naAD may be due to later onset of MTL degeneration than in aAD, considering that older age was associated with atrophy in this area, independent of group. The current study corroborated probable areas of early disease for naAD and showed that each phenotype has a different pattern of atrophy progression across the cortex, providing also important data about pathology transmission.
FERRARESE, CARLO
GROSSMAN, MURRAY
Malattia d'Alzheimer; RMN; longitudinale; atrofia; AD non amnesico
Alzheimer's disease; MRI; longitudinal; atrophy; AD non amnesico
MED/26 - NEUROLOGIA
English
30-gen-2020
NEUROSCIENZE
32
2018/2019
open
(2020). An MRI-based analysis of the longitudinal progression of atrophy in amnestic and non-amnestic phenotypes of Alzheimer’s disease. (Tesi di dottorato, Università degli Studi di Milano-Bicocca, 2020).
File in questo prodotto:
File Dimensione Formato  
phd_unimib_073768.pdf

accesso aperto

Descrizione: tesi di dottorato
Dimensione 3.17 MB
Formato Adobe PDF
3.17 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10281/261941
Citazioni
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
Social impact