Scholarlink

Introduction <p>This study aimed to compare different algorithms based on the Edinburgh Cognitive and Behavioural ALS Screen (ECAS) to classify patients with amyotrophic lateral sclerosis (ALS) according to their neuropsychological phenotype to identify possible discrepancies among these systems.</p> Methods <p>ECAS-Cognitive and -Carer Interview (ECAS-C/-CI) scores of <i>N</i> = 901 patients with ALS without a formal diagnosis of dementia were retrospectively retrieved. Patients were classified, pursuant to Strong et al.’s criteria, as cognitively and behaviourally normal (ALScbn), cognitively and/or behaviourally impaired (ALSci/bi/cbi), or Possible ALS-FTD, according the following ECAS-based algorithms: (1) Abrahams’, solely addressing ECAS-C total and ALS-Specific subtotals; (2) Poletti et al.’s, addressing single task-level ECAS-C scores; (3) “Subscale”, addressing ECAS-C subscales (i.e., Language, Executive, Fluency, Memory and Visuospatial). All algorithms relied on single-item-level ECAS-CI scores for behavioural classifications.</p> Results <p>Whilst agreement rates among these classifications were moderate to high (84–86%; Cohen’s <i>k</i> = 0.78–0.81), and some discrepancies emerged: (1) “ALScbn-to-ALSci” and “ALSci-to-ALScbn” re-classifications occurred across the three comparisons, ranging from ~ 11% to ~ 24%; (2) the most classificatory disagreements (~ 43%) occurred for the ALScbi category when comparing single task-level (Poletti) to total-level (Abrahams) algorithms, with patients being re-classified as either ALSbi or Possible ALS-FTD; (3) ~ 24% of Abraham’s Possible ALS-FTD cases were re-classified as either ALScbi or ALSbi by the Subscale approach.</p> Conclusions <p>Different ECAS-based algorithms for deriving Strong’s phenotypes might yield slight discrepancies that could under- or overestimate a given classification.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

ECAS-Based Neuropsychological Phenotyping in Amyotrophic Lateral Sclerosis: A Retrospective Study Comparing Different Algorithms

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月20日

查看全文
  • Barbara Poletti,
  • Edoardo N. Aiello,
  • Monica Consonni,
  • Barbara Iazzolino,
  • Silvia Torre,
  • Veronica Faltracco,
  • Alessandra Telesca,
  • Francesca Palumbo,
  • Beatrice Curti,
  • Giulia De Luca,
  • Arianna Moreschi,
  • Francesca Frisco,
  • Eleonora Dalla Bella,
  • Enrica Bersano,
  • Nilo Riva,
  • Federico Verde,
  • Stefano Messina,
  • Alberto Doretti,
  • Alessio Maranzano,
  • Claudia Morelli,
  • Stefano F. Cappa,
  • Andrea Calvo,
  • Michael J. Strong,
  • Vincenzo Silani,
  • Giuseppe Lauria,
  • Adriano Chiò,
  • Nicola Ticozzi

摘要

Introduction

This study aimed to compare different algorithms based on the Edinburgh Cognitive and Behavioural ALS Screen (ECAS) to classify patients with amyotrophic lateral sclerosis (ALS) according to their neuropsychological phenotype to identify possible discrepancies among these systems.

Methods

ECAS-Cognitive and -Carer Interview (ECAS-C/-CI) scores of N = 901 patients with ALS without a formal diagnosis of dementia were retrospectively retrieved. Patients were classified, pursuant to Strong et al.’s criteria, as cognitively and behaviourally normal (ALScbn), cognitively and/or behaviourally impaired (ALSci/bi/cbi), or Possible ALS-FTD, according the following ECAS-based algorithms: (1) Abrahams’, solely addressing ECAS-C total and ALS-Specific subtotals; (2) Poletti et al.’s, addressing single task-level ECAS-C scores; (3) “Subscale”, addressing ECAS-C subscales (i.e., Language, Executive, Fluency, Memory and Visuospatial). All algorithms relied on single-item-level ECAS-CI scores for behavioural classifications.

Results

Whilst agreement rates among these classifications were moderate to high (84–86%; Cohen’s k = 0.78–0.81), and some discrepancies emerged: (1) “ALScbn-to-ALSci” and “ALSci-to-ALScbn” re-classifications occurred across the three comparisons, ranging from ~ 11% to ~ 24%; (2) the most classificatory disagreements (~ 43%) occurred for the ALScbi category when comparing single task-level (Poletti) to total-level (Abrahams) algorithms, with patients being re-classified as either ALSbi or Possible ALS-FTD; (3) ~ 24% of Abraham’s Possible ALS-FTD cases were re-classified as either ALScbi or ALSbi by the Subscale approach.

Conclusions

Different ECAS-based algorithms for deriving Strong’s phenotypes might yield slight discrepancies that could under- or overestimate a given classification.