Neuroimaging and fluid biomarkers are used to differentiate frontotemporal dementia (FTD) from Alzheimer’s disease (AD). We implemented a machine learning algorithm that provides individual ...probabilistic scores based on magnetic resonance imaging (MRI) and cerebrospinal fluid (CSF) data. We investigated whether combining MRI and CSF levels could improve the diagnosis confidence. 215 AD patients, 103 FTD patients, and 173 healthy controls (CTR) were studied. With MRI data, we obtained an accuracy of 82 % for AD vs. FTD. A total of 74 % of FTD and 73 % of AD participants have a high probability of accurate diagnosis. Adding CSF-NfL and 14–3–3 levels improved the accuracy and the number of patients in the confidence group for differentiating FTD from AD. We obtain individual diagnostic probabilities with high precision to address the problem of confidence in the diagnosis. We suggest when MRI, CSF, or the combination are necessary to improve the FTD and AD diagnosis. This algorithm holds promise towards clinical applications as support to clinical findings or in settings with limited access to expert diagnoses.
•Machine Learning applied to structural MRI features differentiates FTD from AD.•Probabilistic ML allows the distribution of cases along a spectrum between groups.•Combining MRI and CSF improves accuracy and confidence in diagnosing FTD from AD.
Background
MRI atrophy predicts cognitive status in AD. However, this relationship has not been investigated in early-onset AD (EOAD, < 65 years) patients with a biomarker-based diagnosis.
Methods
...Forty eight EOAD (MMSE ≥ 15; A + T + N +) and forty two age-matched healthy controls (HC; A − T − N −) from a prospective cohort underwent full neuropsychological assessment, 3T-MRI scan and lumbar puncture at baseline. Participants repeated the cognitive assessment annually. We used linear mixed models to investigate whether baseline cortical thickness (CTh) or subcortical volume predicts two-year cognitive outcomes in the EOAD group.
Results
In EOAD, hemispheric CTh and ventricular volume at baseline were associated with global cognition, language and attentional/executive functioning 2 years later (
p
< 0.0028). Regional CTh was related to most cognitive outcomes (
p
< 0.0028), except verbal/visual memory subtests. Amygdalar volume was associated with letter fluency test (
p
< 0.0028). Hippocampal volume did not show significant associations.
Conclusion
Baseline hemispheric/regional CTh, ventricular and amygdalar volume, but not the hippocampus, predict two-year cognitive outcomes in EOAD.
Background
One of the clinical problems for biomarkers' clinical use is the ability to differentiate between Alzheimer’s disease (AD), frontotemporal dementia (FTD), and healthy subjects (CTR). This ...clearly challenges diagnosis and prognosis. We implemented a ML algorithm that provides individual probabilistic diagnoses for these dementias based on magnetic resonance imaging (MRI) and we correlated the results with biochemical markers.
Method
We studied 3T‐T1w MRI of 432 subjects. A subset of these participants had cerebrospinal fluid (CSF) and plasma biomarkers (Table 1). We obtained regional subcortical gray matter volumes and cortical thickness measures using Freesurfer. We implemented a calibrated classifier with a Support Vector Machine with only the MRI data. We tested paired‐wise classification and classification across the 3 groups. We obtained individual probabilities associated with group correspondence. We studied the correlation between these probabilities and CSF and plasma biomarkers. For this, we subdivided the groups into true‐group (subjects with classification according to clinical diagnosis) and false‐group (subjects which did not coincide with clinical diagnosis). Finally, we implemented a permutation test to find the importance of each region in the classification.
Result
We obtained accuracies of 90.7 ± 6.7% in the CTR vs AD classification, 88.6 ± 4.5% for CTR vs FTD, 79.3 ± 8.8% for AD vs FTD, and 79.9 ± 5.1% when discriminating the 3 groups. We obtained a significant positive correlation for plasma p‐tau181 for the false‐AD in the comparison AD vs CTR (Figure 1). The correlation of the false‐CTR was significantly positive for the CSF and plasma NfL. Finally, in the AD vs FTD, the true‐FTD had a significant negative correlation with CSF NfL. The other biochemical biomarkers did not provide additional information. The most important regions for classification are shown in Figure 2.
Conclusion
The ML algorithm gave high accuracies. Within wrongly classified AD patients, probabilities correlated positively with plasma p‐tau181, suggesting hidden pathological processes associated with subjects clinically classified as CTR. Finally, the group probability within well‐classified FTD patients in comparison with AD negatively correlated with CSF NfL. We suggest that this approach can be used as a tool to try to develop personalized diagnoses.
Objective
Although the presymptomatic stages of frontotemporal dementia (FTD) provide a unique chance to delay or even prevent neurodegeneration by early intervention, they remain poorly defined. ...Leveraging a large multicenter cohort of genetic FTD mutation carriers, we provide a biomarker‐based stratification and biomarker cascade of the likely most treatment‐relevant stage within the presymptomatic phase: the conversion stage.
Methods
We longitudinally assessed serum levels of neurofilament light (NfL) and phosphorylated neurofilament heavy (pNfH) in the Genetic FTD Initiative (GENFI) cohort (n = 444), using single‐molecule array technique. Subjects comprised 91 symptomatic and 179 presymptomatic subjects with mutations in the FTD genes C9orf72, GRN, or MAPT, and 174 mutation‐negative within‐family controls.
Results
In a biomarker cascade, NfL increase preceded the hypothetical clinical onset by 15 years and concurred with brain atrophy onset, whereas pNfH increase started close to clinical onset. The conversion stage was marked by increased NfL, but still normal pNfH levels, while both were increased at the symptomatic stage. Intra‐individual change rates were increased for NfL at the conversion stage and for pNfH at the symptomatic stage, highlighting their respective potential as stage‐dependent dynamic biomarkers within the biomarker cascade. Increased NfL levels and NfL change rates allowed identification of presymptomatic subjects converting to symptomatic disease and capture of proximity‐to‐onset. We estimate stage‐dependent sample sizes for trials aiming to decrease neurofilament levels or change rates.
Interpretation
Blood NfL and pNfH provide dynamic stage‐dependent stratification and, potentially, treatment response biomarkers in presymptomatic FTD, allowing demarcation of the conversion stage. The proposed biomarker cascade might pave the way towards a biomarker‐based precision medicine approach to genetic FTD. ANN NEUROL 2022;91:33–47
Early-onset Alzheimer’s disease (EOAD) and frontotemporal dementia (FTD) have a high proportion of genetically determined cases. Next-generation sequencing technologies have triggered the discovery ...of new mutations and genetic variants in dementia-causal genes. We performed whole-exome sequencing and selective analysis of known genes causative of EOAD and FTD in a well-characterized Spanish cohort of 103 patients (60 EOAD, 43 FTD) to find genetic variants associated to patients’ phenotype. In EOAD patients, a new likely pathogenic variant in PSEN1 gene (p.G378R) was found. In FTD patients, 2 likely pathogenic variants were found, one in MAPT gene (p.P397S) and one in VCP gene (p.R159H). In our series, 2% of early-onset dementia without criteria for clinical genetic testing according to current guidelines presented a likely pathogenic mutation. We have also detected 13 additional variants of uncertain significance in causal genes, as well as rare variants in risk genes for dementia (ABCA7, SORL1, SQSTM1, and TREM2). Next-generation technologies in neurodegenerative diseases constitute a powerful tool that significantly contributes to patients’ diagnosis.
•Whole-exome sequencing analysis of 103 patients with early-onset dementia.•A novel likely pathogenic variant is described in PSEN1 gene (p.G378R) in AD.•Novel likely pathogenic variants are found in MAPT (p.P397S) and VCP (p.R159H) in FTD.•We have found likely pathogenic variants in 2% of patients without early-onset family history of disease.•Clinical criteria for genetic testing may disappear as WES becomes more available.
Background
Blood‐based biomarkers have recently emerged as minimally‐invasive, accessible and relatively inexpensive diagnostic and prognostic tools for people with cognitive impairment. Before being ...routinely implemented in clinical practice, the diagnostic performance of distinct commercially available assays should be studied in real‐world cohorts. We aimed to study and compare the diagnostic accuracy of different plasma biomarkers measured using two different assay platforms in a memory clinic cohort.
Method
Participants were selected from a prospective memory clinic cohort; all had Alzheimer’s disease (AD) CSF biomarkers performed. Plasma p‐tau181, GFAP and NfL were measured using Simoa (Quanterix), while plasma p‐tau181, Aβ1‐40 and Aβ1‐42 were quantified using Lumipulse G (Fujirebio). Clinical diagnoses were made according to published criteria, blinded to plasma biomarkers. Aβ status (‐/+) was defined by CSF Aβ concentration using local cutoffs.
Result
One hundred and ten participants were included (mean age standard deviation 66 7.8 years, 56% women). Diagnostic categories included 10 cognitively unimpaired controls, 24 with suspected non‐neurodegenerative cause of cognitive impairment (SND), 53 AD, 9 Lewy body disease (LBD, 4 Aβ+) and 14 frontotemporal dementia (FTD, 1 Aβ+).
Plasma p‐tau181Quanterix and Aβ1‐42/Aβ1‐40 had the highest diagnostic accuracy (Figure 1) to discriminate between SND and AD (AUC CI 0.94 0.89‐0.99 and 0.94 0.85‐1), followed by GFAP (0.93 0.87‐0.99), p‐tau181Fujirebio (0.90 0.82‐0.98) and Aβ1‐42 (0.71 0.58‐0.85). Plasma NfL performed the best to differentiate FTD from SND and AD (AUC 0.95 0.88‐1 and 0.85 0.71‐0.99, respectively).
For Aβ status discrimination (Figure 2), p‐tau181Quanterix had an AUC CI of 0.91 0.85‐0.96, followed by p‐tau181Fujirebio (0.86 0.79‐0.93), Aβ1‐42/Aβ1‐40 (0.85 0.76‐0.93) and GFAP (0.84 0.77‐0.92) with no statistically significant differences in AUCs. Balanced (Youden index) cut‐offs were calculated to study diagnostic performance, resulting in sensitivities of 79‐83%, specificities of 74‐83% and accuracies of 76‐83%. No combination of plasma biomarkers resulted in a significantly increased discriminative accuracy for Aβ status. All plasma biomarkers were moderately correlated with p‐tau181Quanterix (ρ = 0.40‐0.75, Figure 3).
Conclusion
In our cohort, p‐tau181Quanterix, p‐tau181Fujirebio, Aβ1‐42/Aβ1‐40 and GFAP had a high diagnostic performance to discriminate CSF‐defined Aβ status. Plasma NfL identified individuals with FTD. Further studies comparing different plasma biomarkers are needed before implementation in clinical practice.
Background
Little is known about the influence of age at onset (AAO) on plasma biomarkers and their use as prognostic biomarkers in Alzheimer’s disease (AD).
Method
We selected patients with AD ...diagnosis with available neuropsychological testing (NPS) at time of diagnosis and two years later, and plasma biomarkers at baseline.
NPS battery included Free and Cued Selective Reminding Test (FCSRT), Landscape test (visual memory), Boston Naming Test, Semantic Fluency, BDAE auditory comprehension, Constructional and Ideomotor Praxis, Visual Object and Space Perception (VOSP) Incomplete Letters and Number Location subtests, Trail Making Test (TMT) A and B, Phonemic Fluency, and Digit Span Forward and Backward. NPS scores were compared by AAO: early‐onset AD (EOAD; <65 years) vs. late onset AD (LOAD; >65y).
We analyzed plasma biomarkers phosphorilated‐tau181 (p‐tau181), total tau (t‐tau), neurofilament light chain (NfL), glial fibrillary acidic protein (GFAP) and ubiquitin C‐terminal hydrolase L1 (UCHL‐1) using the Quanterix Simoa p‐tau181 Advantage V2 and Neurology 4‐Plex A assays. Group comparisons and linear regressions adjusted by years of education (YOE) were performed in Stata/IC 16.1.
Result
Forty‐two participants were included, 23 LOAD and 19 EOAD. Plasma p‐tau181 and GFAP levels were higher in LOAD (Table 1). We did not find differences between LOAD and EOAD in NPS tests at baseline or +2 years (Table 2).
Plasma ptau‐181 was associated with progression in MMSE globally, VOSP‐Incomplete letters globally and in EOAD. Plasma NfL were associated to Boston Naming test globally and in EOAD, Semantic fluency test globally, VOSP‐incomplete letters in EOAD, and Free and Total Learning of FCSRT in LOAD. Plasma GFAP was associated to MMSE globally and in EOAD, Free learning of FCSRT in LOAD and VOSP‐Incomplete letters and number location globally. Plasma UCHL‐1 was associated to Semantic fluency test in LOAD (Table 3). Praxis and attention and executive function tests loss were not associated to plasma biomarkers.
Conclusion
Plasma p‐tau18, NfL, GFAP and UCHL‐1 were associated to pogression in memory, language and visual tests. They were predominantly associated to memory loss in LOAD and language and visual function loss in EOAD. Results need to be interpreted cautiously due to small sample size.
Background
The application of blood‐based biomarkers for the identification of Alzheimer’s disease (AD) and the development of novel digital technologies as cognitive screening tests are critical to ...moving toward a reliable, more accessible early diagnosis. Our aim was to evaluate the diagnostic performance of a machine learning‐based cognitive assessment known as Altoida’s digital neuro‐signature (DNS) in patients with non‐degenerative mild cognitive impairment (ndMCI) and MCI due to AD (prodromal AD) and its association with CSF and plasma biomarkers.
Method
Altoida’s MCI‐DNS is a 10‐minute cognitive test battery evaluating activities of daily living via motoric and augmented reality tasks. The test consists of placing and finding virtual objects in a real environment and its final score is obtained by weighting multi‐modal digital data features, such as hands’ micro‐movements, speed, reaction times, or navigation trajectories. We included 81 participants, classified according to their clinical status and CSF AD biomarker profile as: cognitively unimpaired controls, CTR (n = 10; age = 68.5±5.9; MMSE = 29.4±1.1), ndMCI (n = 25; age = 67.6±7.2; MMSE = 26.9±1.9) and prodromal AD (n = 46; age = 70.8±4.9; MMSE = 24.3±3.3). We further investigated a subsample of participants classified according to their plasma pTau181 levels as measured by SiMoA cutoff = 1.37 pg/mL (Sarto et al., 2022): pTau181 negative (n = 27; age = 68.5±5.9; MMSE = 26.9±2.7) or pTau181 positive (n = 30; age = 70.3±5.5; MMSE = 24.1±3.7).
Result
Significant differences were found in MCI‐DNS scores between CTR group and ndMCI (F = 23.5; p<0.001) and prodromal AD (F = 114.4; p<0.001) groups. Also, ndMCI showed higher MCI‐DNS scores than the prodromal AD group (F = 4.53; p<0.05, Fig. 1). ROC curves showed an excellent diagnostic accuracy of the MCI‐DNS in the discrimination between CTR vs. ndMCI (AUC = 0.879) and CTR vs. prodromal AD (AUC = 0.975) (Fig. 1). Further analyses showed differences in MCI‐DNS scores between CSF Aβ42 negative and CSF Aβ42 positive (F = 18.9; p<0.001; Fig. 2), as well as between plasma pTau181 negative and plasma pTau181 positive (F = 6.16; p<0.01; Fig. 2). Finally, MCI DNS scores significantly correlated with CSF Aβ42, CSF Aβ42/pTau ratio, CSF neurofilament‐light chain (NfL) and plasma pTau181 concentrations (Fig. 3).
Conclusion
Altoida’s MCI‐DNS test allows excellent discrimination between CTR and patients with MCI. MCI‐DNS scores significantly correlate with CSF AD core biomarkers, biomarkers of neurodegeneration and blood‐based biomarkers (i.e., plasma pTau181).
Background
The diagnosis of early‐onset neurodegenerative dementias (<65 years) can represent a challenge due to their lower frequency respect to late‐onset dementias and atypical forms of ...presentation. Cognitive impairment has emerged as a frequent complaint after COVID‐19 infection.
Method
We retrospectively reviewed (2016‐2021) the demographic and clinical data of the new referrals at our early onset dementia clinic (Hospital Clínic Barcelona). We used Fisher’s Exact test and ANOVA in Stata/IC 16.1 to analyze differences between patients visited in 2021, 2020 and the period 2016‐2019.
Result
We evaluated 296 patients in 2021, 104 in 2020 and 98 patients/year in 2016‐2019. In 2021, patients had an age at onset (AAO) of 50.1 years, lower than 2020 (53.4) and the period 2016‐2019 (53.0) (p<0.05). The percentage of women in 2021 (69.6%) was higher than 2020 (57.7) and 2016‐2019 (56.0) (p<0.05). Diagnostic delay was lower in 2021 (2.1 years) than 2020 (3.3) and 2016‐2019 (3.0) (p<0.05). No differences were found in Minimental (MMSE) scores (Table 1).
In the period 2016‐2021, the number of neurodegenerative diseases (ND) remained steady, the number of subjective cognitive decline (SCD) decreased and the number of non neurodegenerative causes (NNC) experienced a large increase (Table 1), representing 77.7% of visits in 2021.
We did not find differences in the type of ND diagnosis in each period (Figure 1A). ND subgroups did not show differences in AAO, sex or MMSE. In 2021, NND presented lower AAO, higher percentage of women, lower diagnostic delay (Figure 1B) and higher MMSE scores than previous years. No differences were found in the SCD group.
Cognitive impairment after Covid‐19 accounted for 16.7% of NND in 2020 (n = 8, AAO 50.6 (11.8), 62.5% female, MMSE 26.8(2.3)) and 66.6% of NND in 2021 (n = 153, AAO 49,0 (10.0), 80.1% female, MMSE 27.8 (2.6)).
Conclusion
In 2021 we visited approximately three times more patients than in 2016‐2020, among which we observed an increase in NND, mainly patients with cognitive impairment after Covid‐19. On contrast, we found a similar number of ND diagnosis and reduction in SCD.