Identifying individuals with early-stage Alzheimer’s disease (AD) at greater risk of steeper clinical decline would enable better-informed medical, support, and life planning decisions. Despite accumulating evidence on the clinical prognostic value of tau positron emission tomography (PET) in typical late-onset amnestic AD, its utility in predicting clinical decline in individuals with atypical forms of AD remains unclear. Across heterogeneous clinical phenotypes, patients with atypical AD consistently exhibit abnormal tau accumulation in the posterior nodes of the default mode network of the cerebral cortex. This evidence suggests that tau burden in this functional network could be a common imaging biomarker for prognostication across the syndromic spectrum of AD. Here, we examined the relationship between baseline tau PET signal and the rate of subsequent clinical decline in a sample of 48 A+/T+/N+ patients with mild cognitive impairment or mild dementia due to AD with atypical clinical phenotypes: Posterior Cortical Atrophy (n = 16), logopenic variant Primary Progressive Aphasia (n = 15), and amnestic syndrome with multi-domain impairment and young age of onset < 65 years (n = 17). All patients underwent magnetic resonance imaging (MRI), tau PET, and amyloid PET scans at baseline. Each patient’s longitudinal clinical decline was assessed by calculating the annualized change in the Clinical Dementia Rating Sum-of-Boxes (CDR-SB) scores from baseline to follow-up (mean time interval = 14.55 ± 3.97 months). Atypical early AD patients showed an increase in CDR-SB by 1.18 ± 1.25 points per year: t(47) = 6.56, p < .001, Cohen’s d = 0.95. Across clinical phenotypes, baseline tau in the default mode network was the strongest predictor of clinical decline (R2 = .30), outperforming a simpler model with baseline clinical impairment and demographic variables (R2 = .10), tau in other functional networks (R2 = .11-.26), and the magnitude of cortical atrophy (R2 = .20) and amyloid burden (R2 = .09) in the default mode network. Overall, these findings point to the contribution of default mode network tau to predicting the magnitude of clinical decline in atypical early AD patients one year later. This simple measure could aid the development of a personalized prognostic, monitoring, and treatment plan, which would help clinicians not only predict the natural evolution of the disease but also estimate the effect of disease-modifying therapies on slowing subsequent clinical decline given the patient’s tau burden while still early in the disease course.

中文翻译：

---

默认模式网络 tau 预测非典型早期阿尔茨海默病的未来临床下降


识别患有早期阿尔茨海默病 （AD） 的个体，其临床衰退风险更大，将有助于做出更明智的医疗、支持和人生规划决策。尽管积累了关于 tau 正电子发射断层扫描 （PET） 在典型晚发性遗忘性 AD 中的临床预后价值的证据，但其在预测非典型 AD 个体临床下降方面的效用仍不清楚。在异质性临床表型中，非典型 AD 患者在大脑皮层默认模式网络的后节点中始终表现出异常的 tau 积累。这一证据表明，该功能网络中的 tau 负荷可能是 AD 综合征谱中预后的常见影像生物标志物。在这里，我们检查了基线 tau PET 信号与随后临床下降率之间的关系，样本包括 48 例 A+/T+/N+ 轻度认知障碍或因 AD 引起的轻度痴呆患者，临床表型不典型：后皮质萎缩 （n = 16），对数减少型变异原发性进行性失语症 （n = 15） 和遗忘综合征伴多域损伤和发病年龄小 < 65 岁 （n = 17）。所有患者在基线时都接受了磁共振成像 （MRI） 、 tau PET 和淀粉样蛋白 PET 扫描。通过计算临床痴呆评定量和 （CDR-SB） 评分从基线到随访的年化变化来评估每位患者的纵向临床下降 （平均时间间隔 = 14.55 ± 3.97 个月）。非典型早期 AD 患者的 CDR-SB 每年增加 1.18 ± 1.25 点： t（47） = 6.56，p < .001，Cohen 的 d = 0.95。在临床表型中，默认模式网络中的基线 tau 是临床下降的最强预测因子 （R2 = .30），优于具有基线临床损伤和人口统计变量 （R2 = .10）、其他功能网络中的 tau （R2 = .11-.26） 以及皮质萎缩的幅度 （R2 = .20） 和淀粉样蛋白负荷 （R2 = .09） 在默认模式网络中。总体而言，这些发现表明默认模式网络 tau 对预测一年后非典型早期 AD 患者临床下降幅度的贡献。这种简单的措施可以帮助制定个性化的预后、监测和治疗计划，这不仅可以帮助临床医生预测疾病的自然演变，还可以评估疾病缓解疗法对减缓后续临床衰退的影响，因为患者的 tau 负担仍处于病程早期。