兽研简要总结
猫认知功能障碍综合征是一种与年龄相关的疾病，可引起行为改变。该病难以诊断，因为临床症状可能较轻微，且也可与其他医学状况同时出现。脑磁共振成像可能有助于识别与脑老化相关的客观变化，但猫的正常参考信息有限。本研究检查了丘脑间粘合，这是一个已在犬和人类认知功能障碍中研究过的小型大脑结构。研究回顾了112只年龄在4个月至18岁之间、未报告认知功能障碍体征的猫的脑部扫描。该结构的厚度随年龄增长而减小，在年龄最大的组别中尤其更小。在考虑了脑部尺寸、性别和绝育状态、品种组和MRI扫描仪后，以及在排除幼猫或短头品种猫后，这种关联仍然存在。这些发现为接受临床成像的猫的年龄相关脑部测量提供了额外信息。然而，它们并未确定该测量能否区分无认知功能障碍的老化与猫认知功能障碍综合征。在确定其诊断价值之前，需要使用标准化行为评估和成像方案，对伴有和不伴有确诊认知功能障碍的猫进行比较的进一步研究。
关键词：丘脑间粘合，脑萎缩，认知功能障碍综合征，猫，老化
1. 引言
兽医护理的进步促进了家猫预期寿命的延长，导致与年龄相关疾病的患病率升高，包括猫认知功能障碍综合征（FCDS）。该病的患病率随年龄增长而增加，其临床症状常被主人解释为正常老化的一部分。FCDS的特征是与年龄相关的认知功能下降，表现为不能归因于其他医学状况的行为改变[1]，如疼痛、内分泌病（例如甲状腺功能亢进）和影响中枢神经系统的疾病[2]。先前的一篇综述报告称，主人在36%的7-11岁猫、28%的11-14岁猫和50%的15岁以上猫中观察到至少一种行为改变；在16-19岁猫中报告的患病率达到88%[3]。
尽管具有临床相关性，但关于FCDS的诊断、其危险因素及其与脑萎缩的关联，现有数据有限[4]。猫的正常老化过程中已记录到脑萎缩，这与犬和人类的发现一致。然而，FCDS与脑萎缩之间的关系仍未得到充分研究，尽管怀疑其发挥重要作用[5]。
在犬中，认知功能障碍综合征已被更广泛地研究，类似的诊断挑战促使人们评估脑萎缩的客观成像标志物。在这些标志物中，背腹侧丘脑间粘合厚度（IAT）已成为一个可能有用的参数。丘脑间粘合是连接第三脑室两侧左右丘脑的圆形中线灰质桥。在犬中，丘脑间粘合大小与健康个体的年龄呈负相关[6,7]，并且据报道在患有认知功能障碍的犬中较低[7,8,9]，支持其作为脑萎缩指标的使用。
然而，在猫中，丘脑间粘合大小的年龄相关变化仍不清楚。已发表的研究得出了相互矛盾的结果。一项使用基于体素的形态学测量的研究未发现健康猫中丘脑间粘合大小显著减小[5]，尽管该发现应谨慎解读。该基于体素的形态学分析并非设计为对丘脑间粘合的专门评估，作者指出，猫脑小型局部区域的细微变化可能无法通过该方法检测到。相反，另一项研究报告老年猫（≥12岁）的IAT显著较低；然而，该发现基于仅分布在三个年龄类别中的小样本[10]。此外，一项将年龄作为连续变量评估的近期研究也报告，随着年龄增加，丘脑间粘合面积和IAT与脑高度之比显著降低[11]。然而，该研究也纳入了小样本的猫。
因此，本研究主要旨在评估年龄（建模为连续变量）与无明显认知功能障碍猫的IAT之间的关联，其次描述猫各生命阶段的IAT。建立此类基线值对于未来在健康猫与表现出符合FCDS临床症状的年龄匹配猫之间进行比较至关重要。作者假设IAT会随年龄增加而显著降低。
2. 材料与方法
2.1. 研究设计与病例选择
本回顾性横断面研究在Oniris兽医医院（法国南特）小动物诊断影像科开展。2007年至2023年间因临床指征接受脑或颅骨MRI检查的客户所有的任意年龄猫均符合纳入条件。若MRI显示明显的脑实质病变（如占位性病变、脑炎或脑积水），或主人报告提示认知功能障碍的行为改变，则排除该猫。对每只猫记录年龄、体重、性别、品种、MRI检查指征及最终诊断。
按照最新版AAHA/AAFP猫生命阶段指南[12]，将猫分为四个年龄组：幼猫（<1岁）、青年成猫（1至<7岁）、成熟成猫（7至<11岁）和老年猫（≥11岁）。
2.2. MRI操作
2007年至2015年8月间进行的MRI检查使用1.0特斯拉扫描仪（Magnetom Harmony，Siemens Medical Solutions，Erlangen，Germany），而2015年9月至2023年间进行的检查使用配备标准头部线圈的1.5特斯拉扫描仪（Magnetom Essenza，Siemens Medical Solutions，Erlangen，Germany）。所有检查均采集横断面T1加权（T1W）和T2加权（T2W）序列。
在1.0 T系统上，图像矩阵为384 × 228像素，DICOM像素间距为0.469 × 0.469 mm/像素，对应平面内空间分辨率约为0.47 mm/像素。采集矩阵为384 × 228。层厚为3.0 mm，相邻层面间距为3.0 mm，对应无层间隔的连续层面。采样百分比为100%，相位视野百分比为60%。T1W图像采用重复时间（TR）575 ms和回波时间（TE）12 ms获取。T2W图像采用TR 4300 ms和TE 95 ms采集。
在1.5 T系统上，重建图像矩阵为448 × 448像素，DICOM像素间距为0.379 × 0.379 mm/像素，对应平面内空间分辨率约为0.38 mm/像素。采集矩阵为448 × 359。层厚为3.0 mm，相邻层面间距为3.3 mm，对应0.3 mm的层间隔。采样百分比为80%，相位视野百分比为100%。T1W图像采用TR 475 ms和TE 12 ms获取。T2W图像采用TR 4850 ms和TE 94 ms采集。
MRI检查均在猫全身麻醉下进行，所有图像采集时猫取背卧位。
2.3. 丘脑间粘合厚度、脑宽度和脑高度测量
IAT测量使用Horos软件（版本4.0.1；Horos Project）由两名独立观察者完成：一名诊断影像学教授和一名已完成专项培训的兽医专业毕业年级学生。观察者对年龄组及所有其他临床信息设盲。IAT在脑最大背腹径层面的横断面T1W和T2W MR图像上测量（图1）。每只猫的脑高度和脑宽度由该毕业年级兽医学生在用于IAT测量的同一横断面T1W图像上测量，分别取脑实质背侧缘与腹侧缘之间的背腹距离以及右侧缘与左侧缘之间的横向距离，以获得整体脑尺寸用于后续形态计量分析，并评估脑大小与IAT之间的潜在关联。
[图：横断面MR图像T1W序列上丘脑间粘合厚度（白线之间）的测量（A]
2.4. 统计分析
统计分析使用R（版本4.6.0；R Foundation for Statistical Computing，Vienna，Austria）在RStudio（版本2026.09.0+174；Posit Software，PBC，Boston，MA，USA）中完成。
2.4.1. 测量可重复性评估
通过评估两名操作者独立从同一T1W和T2W MR图像获得的IAT测量的观察者间一致性，评估测量方案的可重复性。
一致性采用Bland–Altman分析、Lin一致性相关系数以及基于两名观察者测量均值的双向混合效应绝对一致性相关系数[ICC(3,k)]进行评估。Bland–Altman图通过绘制两名观察者测量差值对其均值的散点图评估一致性，并估计平均偏倚和95%一致性界限。计算Lin一致性相关系数（ρc）及其准确度分量（Cb）和精密度分量（Pearson相关系数，ρ）以量化一致性。计算ICC值及其95%置信区间（CIs）以总结观察者间一致性。
The standard error of measurement (SEM) and minimal detectable change at the 95% confidence level (MDC95) were estimated separately for each scanner and sequence using paired measurements from the two observers. Assuming independent random measurement errors of equal variance, SEM was calculated as the standard deviation of the differences between observers divided by
[内容请查看原文]
[内容请查看原文]
2.4.2. IAT、年龄与形态计量变量之间的关联
对每只猫，后续分析中使用的IAT值分别计算为两名观察者对T1W和T2W图像测量的均值。
在单独的线性回归模型中评估了T1W图像中IAT与脑宽度、脑高度和体重之间的关联，每个模型均对年龄和扫描仪进行了校正。纳入年龄是为了将这些形态学变量的关联与IAT的年龄相关变化区分开来，而纳入扫描仪是为了考虑两种MRI系统之间的差异。
年龄作为连续变量建模（MRI时的精确年龄，以年为单位）。采用简单线性回归分别评估T1W和T2W图像中年龄与IAT的关联。分别对T1W和T2W测量值拟合多变量线性回归模型，纳入年龄、脑高度、脑宽度、性别/绝育状态、品种组（短头或非短头）和扫描仪。采用校正广义方差膨胀因子评估共线性。
为探究IAT随年龄下降可能存在的加速现象，在线性模型中添加二次项（年龄2），并使用赤池信息准则（AIC）和嵌套模型的F检验比较两个模型。以R2总结模型拟合度，并通过残差图和Q-Q图目视评估线性、正态性和残差同方差性假设。
通过排除幼猫（<1岁）后重复年龄分析进行敏感性分析。额外的敏感性分析分别在排除短头猫（波斯猫和异国短毛猫品种）后评估年龄与IAT之间的线性关联，分别针对T1W和T2W测量值进行。
2.4.3. 年龄类别间的比较
为便于比较生命阶段，还比较了不同年龄类别间的平均IAT：幼猫、青年、成熟和老年。采用线性模型评估年龄类别与IAT之间的关联，其中年龄类别作为分类解释变量纳入。通过目视检查Q-Q图评估模型残差的正态性。进行单因素方差分析（ANOVA）以评估各年龄组间平均IAT是否存在差异。统计显著性设定为p < 0.05。当检测到年龄的显著效应时，使用Tukey事后检验进行年龄组间的两两比较。
3. 结果
3.1. 研究群体
共纳入112只猫，年龄4个月至18岁（9只幼猫、56只青年、24只成熟和23只老年）。各组的年龄、体重、性别、品种分布以及脑宽度和脑高度总结于表1。
Table 1. Age (years), body weight (kg), sex, breed characteristics, brain width (mm) and brain height (mm) of sampled cats in the four age categories. BW: body weight; F: female; FS: spayed female; M: male; MC: castrated male; SD: standard deviation.
| Kitten [<1 y] ( n = 9) | Young [1 to <7 y] ( n = 56) | Mature [7 to <11 y] ( n = 24) | Senior [≥11 y] ( n = 23) | |
|---|---|---|---|---|
| Median Age [min–max] | 0.63 [0.31–0.96] | 3.31 [1.18–6.88] | 8.04 [7.00–10.78] | 13.58 [11.02–17.96] |
| Mean BW [min–max] | 2.7 [2.0–3.4] | 4.4 [2.4–8.1] | 5.2 [3.5–8.6] | 4.2 [2.6–6.5] |
| Sex | F (2), M (3), FS (1), MC (3) | F (5), M (5), FS (18), MC (28) | F (1), M (0), FS (8), MC (15) | F (0), M (0), FS (13), MC (10) |
| Breed | European Shorthair (7), Maine Coon (1), Siamese (1) | European Shorthair (41), Birman (5), Maine Coon (3), Siamese (2), Chartreux (1), Exotic Shorthair (1), Norwegian Forest cat (1), Persian (1), Sphynx (1) | European Shorthair (21), Chartreux (1), Siamese (1), Persian (1) | European Shorthair (20), Siamese (1), Persian (2) |
| Mean brain width +/− SD | 38.00 +/− 1.22 | 38.59 +/− 2.67 | 39.12 +/− 1.83 | 38.70 +/− 1.94 |
| Mean brain height +/− SD | 30.33 +/− 3.04 | 29.54 +/− 1.67 | 30.12 +/− 2.01 | 29.30 +/− 2.20 |
使用1.0 T扫描仪检查的猫年龄中位数为5.56岁（n = 29；第25–75百分位数，1.51–8.06岁；范围，0.31–14.50岁），使用1.5 T扫描仪检查的猫年龄中位数为5.46岁（n = 83；第25–75百分位数，2.79–10.20岁；范围，0.35–17.96岁）。
纳入的猫在MRI上均无明显脑实质病变，包括占位性病变、脑炎和脑积水。记录的诊断包括特发性癫痫、耳部疾病（外耳、中耳或内耳）、鼻腔疾病、鼻咽息肉、颅颈交界畸形、颈段脊髓疾病、中毒或未发现异常。
3.2. 测量可重复性
对于T1W图像，Bland–Altman分析显示平均偏倚为−0.08 mm，95%一致性界限范围为−0.68至0.53 mm。Lin一致性相关系数为ρc = 0.87（0.82–0.91），精密度ρ = 0.88，准确度Cb = 0.99。ICC为0.93（95% CI：0.90–0.95），表明观察者间一致性极好。总体而言，观察者之间观察到可接受的一致性。
对于T2W图像，Bland–Altman分析显示平均偏倚为−0.07 mm，95%一致性界限范围为−0.62至0.47 mm。Lin一致性相关系数为ρc = 0.89（0.85–0.92），精密度ρ = 0.90，准确度Cb = 0.99。ICC为0.94（95% CI：0.92–0.96），表明观察者间一致性极好。总体而言，观察者之间观察到可接受的一致性。
扫描仪特异性观察者间SEM估计值为0.179至0.238 mm，相应的MDC95估计值为0.497至0.658 mm。在1.0 T图像上（n = 29），T1W测量的SEM和MDC95分别为0.188和0.520 mm，T2W测量分别为0.238和0.658 mm。在1.5 T图像上（n = 83），T1W测量的相应值为0.228和0.633 mm，T2W测量为0.179和0.497 mm。
3.3. IAT、年龄与形态学变量之间的关联
在分别校正年龄和扫描仪的线性回归模型中，T1W IAT与脑宽度（B = 0.032 mm/mm，95% CI −0.011至0.074，p = 0.142）、脑高度（B = 0.045 mm/mm，95% CI −0.002至0.092，p = 0.059）或体重（B = 0.060 mm/kg，95% CI −0.011至0.131，p = 0.097）均无显著关联。
单变量线性回归显示，年龄与IAT在T1W图像上呈显著负相关（B = −0.071 mm/年，95% CI −0.091至−0.050，p < 0.001；R2 = 0.30），在T2W图像上亦呈显著负相关（B = −0.067 mm/年，95% CI −0.088至−0.046，p < 0.001；R2 = 0.27）（图2）。
[图：年龄与丘脑间黏合厚度（IAT）在（A）T1W和（B）T2W图像上的关联。每个点代表]
在校正脑高度、脑宽度、性别/绝育状态、品种组和扫描仪后，年龄在T1W图像上仍与IAT呈负相关（B = −0.074 mm/年，95% CI −0.094至−0.053，p < 0.001），在T2W图像上亦呈负相关（B = −0.067 mm/年，95% CI −0.087至−0.046，p < 0.001）。两个模型均纳入112只猫。
脑宽度与IAT在T1W（B = 0.055 mm/mm，95% CI 0.011至0.099，p = 0.014）和T2W图像（B = 0.072 mm/mm，95% CI 0.027至0.117，p = 0.002）上呈正相关。脑高度与IAT在T2W图像上呈正相关（B = 0.067 mm/mm，95% CI 0.021至0.113，p = 0.005），但其系数在T1W图像上无统计学显著性（p = 0.086）。
短头猫在T1W（B = −0.818 mm，95% CI −1.288至−0.349，p < 0.001）和T2W图像（B = −0.567 mm，95% CI −1.046至−0.089，p = 0.021）上的校正IAT较低。在T1W图像上，1.5 T扫描仪与IAT高于1.0 T扫描仪相关（B = 0.222 mm，95% CI 0.015至0.429，p = 0.035），但在T2W图像上无显著差异（p = 0.922）。在任一序列上，各性别/绝育状态与去势雄性之间的个体比较均无统计学显著性。校正GVIF范围为1.03至1.14，表明无实质性共线性。
在T1W图像上，二次模型比线性模型拟合数据更好（ΔAIC = 4.9；F检验，p = 0.010），二次项为负（B = −0.0057，95% CI −0.0100至−0.0014），提示年龄与IAT之间存在非线性关联。
在T2W图像上，二次项未改善模型拟合（ΔAIC = 0.2；p = 0.14），保留线性模型。
排除幼猫后，年龄与IAT的单变量线性关联在T1W（B = −0.076 mm/年，95% CI −0.098至−0.054，p < 0.001；R2 = 0.31）和T2W图像（B = −0.075 mm/年，95% CI −0.097至−0.052，p < 0.001；R2 = 0.30）上仍显著。T1W二次项仍显著（B = −0.0056，95% CI −0.0104至−0.0007，p = 0.025），但T2W二次项不显著（p = 0.46）。
排除5只短头猫后，年龄与IAT的单变量线性关联在T1W图像（B = −0.061 mm/年，95% CI −0.082至−0.040，p < 0.001；R2 = 0.25）和T2W图像（B = −0.062 mm/年，95% CI −0.084至−0.040，p < 0.001；R2 = 0.23）上仍均显著。
3.4. 年龄类别之间的比较
结果汇总于表2和图3。对于T1W分析，幼猫（<1岁）、青年（1至<7岁）、成熟（7至<11岁）和老年（≥11岁）组的平均IAT值分别为5.50 mm（4.62–5.99）、5.48（4.29–7.01）、5.28（3.99–6.33）和4.68（3.18–6.17）。单因素方差分析显示，T1W图像上各年龄组间IAT存在显著差异（p < 0.001）。老年组IAT显著低于青年组（平均差 −0.80 mm，95% CI −1.14至−0.47，p < 0.001）、成熟组（−0.61 mm，95% CI −1.00至−0.21，p < 0.001）和幼猫组（−0.82 mm，95% CI −1.35至−0.29，p < 0.001）。成熟组与青年组之间未发现显著差异（−0.20 mm，95% CI −0.53至0.13，p = 0.407）。幼猫组与青年组（p = 1.000）或成熟组（p = 0.707）无显著差异。
Table 2. Mean (min-max) and median (25th–75th percentiles) values for interthalamic adhesion thickness in the four different age categories on the T1W and T2W sequences.
| Kittens ( n = 9) | Young ( n = 56) | Mature ( n = 24) | Senior ( n = 23) | |
|---|---|---|---|---|
| IAT T1W (mm) | Mean 5.50 (4.62–5.99) Median 5.63 (5.25–5.84) | Mean 5.48 (4.29–7.01) Median 5.47 (5.18–5.74) | Mean 5.28 (3.99–6.33) Median 5.41 (4.95–5.59) | Mean 4.68 (3.18–6.17) Median 4.71 (4.37–5.05) |
| IAT T2W (mm) | Mean 5.11 (4.16–5.87) Median 5.11 (5.00–5.17) | Mean 5.24 (4.04–6.42) Median 5.23 (4.98–5.51) | Mean 4.89 (3.45–6.08) Median 4.91 (4.55–5.26) | Mean 4.43 (3.55–5.15) Median 4.42 (4.08–4.74) |
[图：按年龄类别分组的丘脑间黏合厚度（IAT）在（A）T1W和（B）T2W图像上。箱体显示四分位]
在T2W分析中，幼猫（<1岁）、青年（1至<7岁）、成熟（7至<11岁）和老年（≥11岁）组的平均IAT值分别为5.11（4.16–5.87）、5.24（4.04–6.42）、4.89（3.45–6.08）和4.43（3.55–5.15）。单因素方差分析发现，T2W图像上各年龄组间IAT存在显著差异（p < 0.001）。老年组的IAT显著低于青年组（平均差−0.81 mm，95% CI −1.14至−0.47，p < 0.001）、成熟组（−0.45 mm，95% CI −0.85至−0.06，p = 0.017）和幼猫组（−0.67 mm，95% CI −1.20至−0.14，p = 0.007）。成熟猫的IAT也低于青年猫（−0.35 mm，95% CI −0.68至−0.02，p = 0.030）。幼猫组与青年组（p = 0.889）或成熟组（p = 0.697）无显著差异。
4. 讨论
本研究在一个更大的回顾性临床MRI队列中，进一步描述了无明显认知功能障碍的猫在既定生命阶段中IAT的年龄相关变化，样本量较先前报道更大。
我们的结果发现，年龄（作为连续变量建模）与T1W和T2W图像上的IAT均呈负相关。在调整脑高度、脑宽度、性别/绝育状态、品种组和扫描仪后，年龄与IAT的负相关在两个序列上均持续存在。在生命阶段间的次要比较中，老年猫在两个序列上的IAT均低于幼猫、青年成年猫和成熟成年猫。成熟成年猫在T2W图像上的IAT也低于青年成年猫，而该比较在T1W图像上无统计学显著性。幼猫与青年或成熟成年猫之间未发现统计学显著差异。
这些发现与先前两项研究的结果一致：一项报告12岁及以上猫的平均IAT显著低于1至6岁猫[10]，另一项报告丘脑间黏合面积及IAT与脑高度比值随年龄增长显著下降[11]。然而，这两项研究的样本量均较小。相比之下，我们的结果与Hamamoto等人的结果不同，后者未发现IAT随年龄显著下降[5]。这一差异可能归因于方法学差异，包括使用基于体素的形态学分析和较小的样本量（65只1至16岁猫）。基于体素的形态学分析对于检测丘脑间黏合（一个小的中线结构）的细微变化可能敏感性有限，因为空间标准化、组织分割和平滑处理可能降低解剖特异性并掩盖小的局灶性差异。
在本研究中，年龄在主要分析中作为连续变量建模，而改编自AAHA/AAFP猫生命阶段指南[12]的分类保留用于次要比较和临床解读。鉴于先前报道提示，归因于FCDS的年龄相关脑改变和/或行为改变在15岁以上的猫中更为明显[3,10,13]，将老年类别细分为15岁以下和15岁及以上猫将是有意义的。然而，仅4只猫年龄在15岁及以上，被认为不足以支持这些亚组之间的可靠比较。因此，这些猫被保留在单一老年类别中。
加入负二次年龄项改善了T1W图像的模型拟合，提示在较大年龄时IAT下降更陡。排除幼猫后该发现仍然存在。然而，二次项在T2W图像上无统计学显著性，且年龄范围上限的猫数量较少限制了晚年估计曲线的精确度。因此，这种曲率证据应被视为探索性的。
衰老猫的组织病理学研究已报道神经元丢失、突触变性、β-淀粉样蛋白沉积和tau病理[14,15,16,17]。较低的IAT可能反映该结构内的神经元丢失，也可能与丘脑体积减小有关，如犬[18]和人类[19]研究所示。综上所述，这些观察提示IAT可能是评估猫年龄相关脑形态学变化的相关参数。
观察到的IAT随年龄增长而下降支持进一步研究IAT作为猫年龄相关脑变化的定量影像学标志物，如先前在犬中所述[7]。事实上，在健康犬中，已有报道IAT随年龄下降[6,7,8]。尽管本研究聚焦于无明显认知功能障碍的猫，但探讨FCDS猫的脑萎缩是否更为明显将是有意义的，正如在猫中已被怀疑[14]并在犬中得到证实[7]。鉴于诊断该疾病的挑战，特别是细微行为变化可能被主人忽视或误解，诸如IAT等客观影像学测量可能有助于更客观的评估并促进早期发现。
在解读这些发现时，应考虑若干局限性。首先，本回顾性研究在单一转诊医院进行；因此，研究人群可能无法代表一般猫群体。猫因一系列临床指征接受MRI检查，最终诊断异质，尽管排除了有明显脑实质病变或主人报告行为改变的猫，但仍可能引入选择偏差。短头猫在两个序列上的调整IAT均较低，但仅纳入五只，限制了该发现的解读。排除这些猫后，年龄与IAT之间的负相关仍然显著。
尽管总体样本量较大，但猫在各年龄组的分布不均匀。在老年组中，仅四只猫年龄在15岁或以上，这限制了年龄范围高端估计的精确性。
此外，根据回顾性医疗记录判断猫无认知功能障碍，未进行标准化行为评估或特定神经学检查。因此，可能存在细微的认知或行为改变，但未被主人识别。
另外，MRI检查在较长时期内使用不同扫描仪进行（2015年8月前为1.0特斯拉，之后为1.5特斯拉），导致图像分辨率存在差异。这种技术异质性可能影响了解剖结构的显示和IAT测量的精确性，从而增加了测量变异性。扫描仪作为协变量纳入多变量模型，且与T1W图像上的IAT相关，但与T2W图像无关。调整扫描仪和其他协变量后，年龄与IAT之间的负相关仍然存在。
最后，在3 mm厚横断面图像上使用单一背腹侧测量也引入了与部分容积效应和切片定位或方向相关的潜在局限性，这可能影响丘脑间粘合的勾画及其最大厚度的测量。扫描仪特异性SEM和MDC95估计量化了同一图像上观察者间测量变异性的随机成分，但未考虑两位观察者共有的系统误差或不同MRI检查之间的变异性。因此，这些估计不应被解释为检测个体猫纵向变化的验证阈值。本研究结果支持年龄与IAT之间的组水平关联，而非个体水平的诊断用途。正中矢状面面积测量、三维体积测量以及结合脑部尺寸的方法值得作为替代或补充的形态计量学方法进行评估。在假定其优于线性IAT测量之前，应直接评估其相对可重复性和对年龄相关变化的敏感性。
5. 结论
总体而言，在这个回顾性临床MRI队列中，猫的IAT随年龄增长而下降。这些发现支持进一步评估IAT作为猫年龄相关脑变化的简单形态计量学标志物。在推荐临床诊断使用之前，需要进行涉及确诊FCDS的猫和年龄匹配的对照猫、标准化行为评估和统一MRI协议的前瞻性研究。
缩写
本手稿中使用以下缩写：
| FCDS | Feline cognitive dysfunction syndrome |
| IAT | Interthalamic adhesion thickness |
| MRI | Magnetic resonance imaging |
| T1W | T1-weighted |
| T2W | T2-weighted |
| TR | Repetition time |
| TE | Echo time |
| FOV | Field of view |
作者贡献
概念化，O.T.、R.K.-C.和M.F.；方法学，O.T.、R.K.-C.和M.F.；验证，R.K.-C.和M.F.；形式分析，R.K.-C.；调查，O.T.；数据管理，O.T.和R.K.-C.；写作—初稿准备，R.K.-C.；写作—审查和编辑，R.K.-C.、O.T.和M.F.；可视化，R.K.-C.；监督，M.F.；项目管理，M.F.。所有作者已阅读并同意手稿的出版版本。
机构审查委员会声明
本回顾性观察性研究不需要伦理审查和批准，因为它仅涉及审查现有的匿名临床记录和预先存在的MRI检查，这些检查是在Oniris兽医医院作为常规诊断护理的一部分从客户拥有的猫获得的。本研究未进行额外程序、干预或影像检查。
知情同意声明
从Oniris获得了对Oniris匿名临床数据进行回顾性使用的知情同意。
数据可用性声明
由于与回顾性临床记录相关的保密限制，本研究呈现的数据可根据要求从通讯作者处获得。
利益冲突
作者声明无利益冲突。
资助声明
本研究未获得外部资助。
脚注
[内容请查看原文]
参考文献
[内容请查看原文]
关联数据
数据可用性声明
由于与回顾性临床记录相关的保密限制，本研究呈现的数据可根据要求从通讯作者处获得。
查看英文原文▾
Simple Summary
Feline cognitive dysfunction syndrome is an age-related condition that can cause changes in behavior. It is difficult to diagnose because the clinical signs may be subtle and can also occur with other medical conditions. Brain magnetic resonance imaging may help identify objective changes associated with brain aging, but normal reference information in cats is limited. This study examined the interthalamic adhesion, which is a small cerebral structure already studied for cognitive dysfunction in dogs and humans. Brain scans from 112 cats aged 4 months to 18 years without reported signs of cognitive dysfunction were reviewed. The thickness of this structure decreased with age and was particularly smaller in the oldest age groups. This association persisted after accounting for brain dimensions, sex and neuter status, breed group, and MRI scanner, and after excluding kittens or cats of brachycephalic breeds. These findings provide additional information on age-associated brain measurements in cats undergoing clinical imaging. However, they do not establish whether this measurement can distinguish aging without cognitive dysfunction from feline cognitive dysfunction syndrome. Further studies comparing cats with and without confirmed cognitive dysfunction, using standardized behavioral assessments and imaging protocols, are needed before its diagnostic value can be determined.
Keywords: interthalamic adhesion, cerebral atrophy, cognitive dysfunction syndrome, feline, aging
1. Introduction
Advances in veterinary care have contributed to increased life expectancy in domestic cats, resulting in a higher prevalence of age-related diseases, including feline cognitive dysfunction syndrome (FCDS). The prevalence of this condition increases with age, and its clinical signs are frequently interpreted by owners as part of normal aging. FCDS is characterized by an age-associated decline in cognitive function that manifests as behavioral changes not attributable to other medical conditions [1], such as pain, endocrinopathies (e.g., hyperthyroidism), and diseases affecting the central nervous system [2]. A previous review reported that owners observed at least one behavioral change in 36% of cats aged 7–11 years, 28% of cats aged 11–14 years, and 50% of cats older than 15 years; the reported prevalence reached 88% in cats aged 16–19 years [3].
Despite its clinical relevance, limited data are available regarding the diagnosis of FCDS, its risk factors, and its association with brain atrophy [4]. Brain atrophy has been documented during normal aging in cats, consistent with findings in dogs and humans. However, the relationship between FCDS and brain atrophy remains insufficiently investigated, although it is suspected to play a significant role [5].
In dogs, cognitive dysfunction syndrome has been investigated more extensively, and similar diagnostic challenges have prompted the evaluation of objective imaging markers of brain atrophy. Among these markers, the dorsoventral interthalamic adhesion thickness (IAT) has emerged as a potentially useful parameter. The interthalamic adhesion is a rounded midline bridge of gray matter connecting the right and left thalami across the third ventricle. In dogs, interthalamic adhesion size has been negatively correlated with age in healthy individuals [6,7], and has been reported to be lower in dogs with cognitive dysfunction [7,8,9], supporting its use as an indicator of cerebral atrophy.
In cats, however, age-related changes in interthalamic adhesion size remain unclear. Published studies have yielded conflicting results. A study using voxel-based morphometry did not identify a significant decrease in interthalamic adhesion size in healthy cats [5], although this finding should be interpreted cautiously. This voxel-based morphometry analysis was not designed as a dedicated assessment of the interthalamic adhesion, and the authors noted that subtle changes in small localized regions of the feline brain may not be detectable with this approach. Conversely, another study reported significantly lower IAT in older cats (≥12 years); however, this finding was based on a small sample distributed across only three age categories [10]. Additionally, a recent study that evaluated age as a continuous variable also reported a significant decrease in interthalamic adhesion area and in the ratio of IAT to brain height with increasing age [11]. However, that study also included a small sample of cats.
Therefore, this study primarily aimed to evaluate the association between age, modeled as a continuous variable, and IAT in cats without apparent cognitive dysfunction, and secondarily to describe IAT across feline life stages. Establishing such baseline values is essential for future comparisons between healthy cats and age-matched cats exhibiting clinical signs consistent with FCDS. The authors hypothesized that IAT would significantly decrease with increasing age.
2. Materials and Methods
2.1. Study Design and Case Selection
This retrospective cross-sectional study was conducted at the Small Animal Diagnostic Imaging Department of the Veterinary Hospital, Oniris (Nantes, France). Client-owned cats of any age that underwent clinically indicated brain or skull MRI between 2007 and 2023 were eligible for inclusion. Cats were excluded if MRI revealed an apparent cerebral parenchymal lesion, such as a mass lesion, encephalitis or hydrocephalus, or if the owner reported behavioral changes suggestive of cognitive dysfunction. For each cat, age, body weight, sex, breed, MRI indications, and final diagnosis were recorded.
The cats were classified into four age categories, as described in the most recent AAHA/AAFP Feline Life Stage Guidelines [12]: kitten (<1 year), young adult (1 to <7 years), mature adult (7 to <11 years) and senior (≥11 years).
2.2. MRI Procedure
MRI examinations performed between 2007 and August 2015 were acquired with a 1.0 Tesla scanner (Magnetom Harmony, Siemens Medical Solutions, Erlangen, Germany), whereas examinations performed between September 2015 and 2023 were acquired with a 1.5 Tesla scanner (Magnetom Essenza, Siemens Medical Solutions, Erlangen, Germany) equipped with a standard head coil. In all examinations, transverse T1-weighted (T1W) and T2-weighted (T2W) sequences were acquired.
On the 1.0 T system, the image matrix was 384 × 228 pixels, with a DICOM pixel spacing of 0.469 × 0.469 mm/pixel, corresponding to an in-plane spatial resolution of approximately 0.47 mm/pixel. The acquisition matrix was 384 × 228. Slice thickness was 3.0 mm, with a 3.0 mm spacing between successive slices, corresponding to contiguous slices without an interslice gap. The percent sampling was 100%, and the percent phase field of view was 60%. T1W images were obtained using a repetition time (TR) of 575 ms and an echo time (TE) of 12 ms. T2W images were acquired using a TR of 4300 ms and a TE of 95 ms.
On the 1.5 T system, the reconstructed image matrix was 448 × 448 pixels, with a DICOM pixel spacing of 0.379 × 0.379 mm/pixel, corresponding to an in-plane spatial resolution of approximately 0.38 mm/pixel. The acquisition matrix was 448 × 359. Slice thickness was 3.0 mm, with a 3.3 mm spacing between successive slices, corresponding to a 0.3 mm interslice gap. The percent sampling was 80%, and the percent phase field of view was 100%. T1W images were obtained with a TR of 475 ms and a TE of 12 ms. T2W images were acquired with a TR of 4850 ms and a TE of 94 ms.
MRI examinations were performed with cats under general anesthesia and positioned in dorsal recumbency for all image acquisitions.
2.3. Interthalamic Adhesion Thickness, Brain Width and Brain Height Measurements
IAT measurements were performed using Horos software (version 4.0.1; Horos Project) by two independent observers: a professor of diagnostic imaging and a final-year veterinary student who had completed specific training. The observers were blinded to age category and all other clinical information. IAT was measured on transverse T1W and T2W MR images of the brain at the level of its maximal dorsoventral dimension (Figure 1). Brain height and width were measured for each cat by the final-year veterinary student on the same transverse T1W image used for IAT measurement, as the dorsoventral distance between the dorsal and ventral margins of the brain parenchyma and the transverse distance between its right and left lateral margins, respectively, to obtain global brain dimensions for subsequent morphometric analyses and to evaluate potential associations between brain size and IAT.
[图：Measurement of the interthalamic adhesion thickness (between white lines) on transverse MR images, on a T1W sequence ( A]
2.4. Statistical Analysis
Statistical analyses were performed using R (version 4.6.0; R Foundation for Statistical Computing, Vienna, Austria) within RStudio (version 2026.09.0+174; Posit Software, PBC, Boston, MA, USA).
2.4.1. Assessment of Measurement Reproducibility
Reproducibility of the measurement protocol was assessed by evaluating interobserver agreement for IAT measurements independently obtained by two operators from the same T1W and T2W MR images.
Agreement was assessed using Bland–Altman analysis, the Lin concordance correlation coefficient, and a two-way mixed-effects correlation coefficient (ICC) for absolute agreement based on the mean of the two observers’ measurements [ICC(3,k)]. Bland–Altman plots were used to evaluate agreement by plotting the difference between the two observers’ measurements against their mean and to estimate mean bias and 95% limits of agreement. The Lin concordance correlation coefficient (ρc), together with its accuracy component (Cb) and precision component (Pearson’s correlation coefficient, ρ), was calculated to quantify agreement. ICC values and their 95% confidence intervals (CIs) were calculated to summarize interobserver agreement.
The standard error of measurement (SEM) and minimal detectable change at the 95% confidence level (MDC95) were estimated separately for each scanner and sequence using paired measurements from the two observers. Assuming independent random measurement errors of equal variance, SEM was calculated as the standard deviation of the differences between observers divided by
[内容请查看原文]
. MDC95 was calculated as 1.96 ×[内容请查看原文]
× SEM.2.4.2. Associations Between IAT, Age and Morphometric Variables
For each cat, IAT values used in subsequent analyses were calculated as the mean of the two observers’ measurements for T1W and T2W images separately.
Associations between T1W IAT and brain width, brain height and body weight were assessed in separate linear regression models, each adjusted for age and scanner. Age was included to distinguish associations with these morphometric variables from the age-related variation in IAT, while scanner was included to account for differences between the two MRI systems.
Age was modeled as a continuous variable (exact age in years at the time of MRI). The association between age and IAT was assessed with simple linear regression, separately for T1W and T2W images. Multivariable linear regression models, including age, brain height, brain width, sex/neuter status, breed group (brachycephalic or non-brachycephalic) and scanner, were fitted separately for T1W and T2W measurements. Collinearity was assessed using adjusted generalized variance inflation factors.
To explore a possible acceleration of IAT decline with age, a quadratic term (age2) was added to the linear model, and the two models were compared using Akaike’s information criterion (AIC) and an F-test for nested models. Model fit was summarized with R2, and the assumptions of linearity, normality and homoscedasticity of residuals were assessed visually with residual plots and Q–Q plots.
Sensitivity analyses were performed by repeating the age analyses after exclusion of kittens (<1 year). Additional sensitivity analyses assessed the linear association between age and IAT after exclusion of brachycephalic cats (Persian and Exotic Shorthair breeds), separately for T1W and T2W measurements.
2.4.3. Comparisons Among Age Categories
To allow comparison between life stages, mean IAT was also compared among age categories: kitten, young, mature and senior. The association between age category and IAT was assessed with a linear model in which age category was included as a categorical explanatory variable. Normality of model residuals was assessed by visual inspection of Q–Q plots. A one-way analysis of variance (ANOVA) was performed to assess whether mean IAT differed among age groups. Statistical significance was set at p < 0.05. When a significant effect of age was detected, pairwise comparisons among age groups were performed using Tukey’s post hoc test.
3. Results
3.1. Study Population
A total of 112 cats aged 4 months to 18 years were included (9 kittens, 56 young, 24 mature and 23 senior). Age, body weight, sex, breed distribution and brain width and height for each group are summarized in Table 1.
Table 1. Age (years), body weight (kg), sex, breed characteristics, brain width (mm) and brain height (mm) of sampled cats in the four age categories. BW: body weight; F: female; FS: spayed female; M: male; MC: castrated male; SD: standard deviation.
| Kitten [<1 y] ( n = 9) | Young [1 to <7 y] ( n = 56) | Mature [7 to <11 y] ( n = 24) | Senior [≥11 y] ( n = 23) | |
|---|---|---|---|---|
| Median Age [min–max] | 0.63 [0.31–0.96] | 3.31 [1.18–6.88] | 8.04 [7.00–10.78] | 13.58 [11.02–17.96] |
| Mean BW [min–max] | 2.7 [2.0–3.4] | 4.4 [2.4–8.1] | 5.2 [3.5–8.6] | 4.2 [2.6–6.5] |
| Sex | F (2), M (3), FS (1), MC (3) | F (5), M (5), FS (18), MC (28) | F (1), M (0), FS (8), MC (15) | F (0), M (0), FS (13), MC (10) |
| Breed | European Shorthair (7), Maine Coon (1), Siamese (1) | European Shorthair (41), Birman (5), Maine Coon (3), Siamese (2), Chartreux (1), Exotic Shorthair (1), Norwegian Forest cat (1), Persian (1), Sphynx (1) | European Shorthair (21), Chartreux (1), Siamese (1), Persian (1) | European Shorthair (20), Siamese (1), Persian (2) |
| Mean brain width +/− SD | 38.00 +/− 1.22 | 38.59 +/− 2.67 | 39.12 +/− 1.83 | 38.70 +/− 1.94 |
| Mean brain height +/− SD | 30.33 +/− 3.04 | 29.54 +/− 1.67 | 30.12 +/− 2.01 | 29.30 +/− 2.20 |
The median age was 5.56 years in cats examined using the 1.0 T scanner (n = 29; 25th–75th percentiles, 1.51–8.06 years; range, 0.31–14.50 years) and 5.46 years in cats examined using the 1.5 T scanner (n = 83; 25th–75th percentiles, 2.79–10.20 years; range, 0.35–17.96 years).
No included cats had apparent cerebral parenchymal lesions on MRI, including mass lesions, encephalitis and hydrocephalus. Recorded diagnoses included idiopathic epilepsy, ear disease (external, middle or internal ear), nasal disease, nasopharyngeal polyps, craniocervical junction malformations, cervical spinal cord disease, intoxication or no abnormalities identified.
3.2. Measurement Reproducibility
For T1W images, Bland–Altman analysis showed a mean bias of −0.08 mm, with 95% limits of agreement ranging from −0.68 to 0.53 mm. The Lin concordance correlation coefficient was ρc = 0.87 (0.82–0.91), precision ρ = 0.88, and accuracy Cb = 0.99. The ICC was 0.93 (95% CI: 0.90–0.95), indicating excellent interobserver agreement. Overall, acceptable agreement was observed between observers.
For T2W images, Bland–Altman analysis showed a mean bias of −0.07 mm, with 95% limits of agreement ranging from −0.62 to 0.47 mm. The Lin concordance correlation coefficient was ρc = 0.89 (0.85–0.92), precision ρ = 0.90, and accuracy Cb = 0.99. The ICC was 0.94 (95% CI: 0.92–0.96), indicating excellent interobserver agreement. Overall, acceptable agreement was observed between observers.
Scanner-specific estimates of the interobserver SEM ranged from 0.179 to 0.238 mm, and the corresponding MDC95 estimates ranged from 0.497 to 0.658 mm. On the 1.0 T images (n = 29), SEM and MDC95 were 0.188 and 0.520 mm for T1W measurements and 0.238 and 0.658 mm for T2W measurements, respectively. On the 1.5 T images (n = 83), the corresponding values were 0.228 and 0.633 mm for T1W measurements and 0.179 and 0.497 mm for T2W measurements.
3.3. Associations Between IAT, Age and Morphometric Variables
In separate linear regression models adjusted for age and scanner, T1W IAT was not significantly associated with brain width (B = 0.032 mm/mm, 95% CI −0.011 to 0.074, p = 0.142), brain height (B = 0.045 mm/mm, 95% CI −0.002 to 0.092, p = 0.059), or body weight (B = 0.060 mm/kg, 95% CI −0.011 to 0.131, p = 0.097).
Univariable linear regression identified a significant negative association between age and IAT on T1W images (B = −0.071 mm/year, 95% CI −0.091 to −0.050, p < 0.001; R2 = 0.30) and on T2W images (B = −0.067 mm/year, 95% CI −0.088 to −0.046, p < 0.001; R2 = 0.27) (Figure 2).
[图：Association between age and interthalamic adhesion thickness (IAT) on ( A ) T1W and ( B ) T2W images. Each point represe]
After adjustment for brain height, brain width, sex/neuter status, breed group and scanner, age remained negatively associated with IAT on T1W images (B = −0.074 mm/year, 95% CI −0.094 to −0.053, p < 0.001) and T2W images (B = −0.067 mm/year, 95% CI −0.087 to −0.046, p < 0.001). Both models included 112 cats.
Brain width was positively associated with IAT on T1W (B = 0.055 mm/mm, 95% CI 0.011 to 0.099, p = 0.014) and T2W images (B = 0.072 mm/mm, 95% CI 0.027 to 0.117, p = 0.002). Brain height was positively associated with IAT on T2W images (B = 0.067 mm/mm, 95% CI 0.021 to 0.113, p = 0.005), but its coefficient was not statistically significant on T1W images (p = 0.086).
Brachycephalic cats had lower adjusted IAT on T1W (B = −0.818 mm, 95% CI −1.288 to −0.349, p < 0.001) and T2W images (B = −0.567 mm, 95% CI −1.046 to −0.089, p = 0.021). The 1.5 T scanner was associated with higher IAT than the 1.0 T scanner on T1W images (B = 0.222 mm, 95% CI 0.015 to 0.429, p = 0.035), but not significantly on T2W images (p = 0.922). None of the individual sex/neuter-status comparisons with castrated males was statistically significant on either sequence. Adjusted GVIFs ranged from 1.03 to 1.14, indicating no substantial collinearity.
On T1W images, a quadratic model fitted the data better than a linear model (ΔAIC = 4.9; F-test, p = 0.010), with a negative quadratic term (B = −0.0057, 95% CI −0.0100 to −0.0014), suggesting a nonlinear association between age and IAT.
On T2W images, the quadratic term did not improve model fit (ΔAIC = 0.2; p = 0.14), and a linear model was retained.
After the exclusion of kittens, the univariable linear association between age and IAT remained significant on T1W (B = −0.076 mm/year, 95% CI −0.098 to −0.054, p < 0.001; R2 = 0.31) and on T2W images (B = −0.075 mm/year, 95% CI −0.097 to −0.052, p < 0.001; R2 = 0.30). The T1W quadratic term remained significant (B = −0.0056, 95% CI −0.0104 to −0.0007, p = 0.025), but not the T2W quadratic term (p = 0.46).
After the exclusion of the five brachycephalic cats, the univariable linear association between age and IAT remained significant on both T1W images (B = −0.061 mm/year, 95% CI −0.082 to −0.040, p < 0.001; R2 = 0.25) and T2W images (B = −0.062 mm/year, 95% CI −0.084 to −0.040, p < 0.001; R2 = 0.23).
3.4. Comparisons Among Age Categories
Results are summarized in Table 2 and Figure 3. For the T1W analysis, mean IAT values were 5.50 mm (4.62–5.99), 5.48 (4.29–7.01), 5.28 (3.99–6.33), and 4.68 (3.18–6.17) in the kitten (<1 year), young (1 to <7 years), mature (7 to <11 years), and senior (≥11 years) groups, respectively. A one-way ANOVA identified a significant difference in IAT among age groups on T1W images (p < 0.001). IAT was significantly lower in the senior group than in the young (mean difference −0.80 mm, 95% CI −1.14 to −0.47, p < 0.001), mature (−0.61 mm, 95% CI −1.00 to −0.21, p < 0.001) and kitten groups (−0.82 mm, 95% CI −1.35 to −0.29, p < 0.001). No significant differences were identified between the mature and young groups (−0.20 mm, 95% CI −0.53 to 0.13, p = 0.407). The kitten group did not differ significantly from the young (p = 1.000) or from the mature groups (p = 0.707).
Table 2. Mean (min-max) and median (25th–75th percentiles) values for interthalamic adhesion thickness in the four different age categories on the T1W and T2W sequences.
| Kittens ( n = 9) | Young ( n = 56) | Mature ( n = 24) | Senior ( n = 23) | |
|---|---|---|---|---|
| IAT T1W (mm) | Mean 5.50 (4.62–5.99) Median 5.63 (5.25–5.84) | Mean 5.48 (4.29–7.01) Median 5.47 (5.18–5.74) | Mean 5.28 (3.99–6.33) Median 5.41 (4.95–5.59) | Mean 4.68 (3.18–6.17) Median 4.71 (4.37–5.05) |
| IAT T2W (mm) | Mean 5.11 (4.16–5.87) Median 5.11 (5.00–5.17) | Mean 5.24 (4.04–6.42) Median 5.23 (4.98–5.51) | Mean 4.89 (3.45–6.08) Median 4.91 (4.55–5.26) | Mean 4.43 (3.55–5.15) Median 4.42 (4.08–4.74) |
[图：Interthalamic adhesion thickness (IAT) by age category on ( A ) T1W and ( B ) T2W images. Boxes show the interquartile r]
For the T2W analysis, mean IAT values were 5.11 (4.16–5.87), 5.24 (4.04–6.42), 4.89 (3.45–6.08), and 4.43 (3.55–5.15) in the kitten (<1 year), young (1 to <7 years), mature (7 to <11 years), and senior (≥11 years) groups, respectively. A one-way ANOVA identified a significant difference in IAT among age groups on T2W images (p < 0.001). IAT was significantly lower in the senior group than in the young (mean difference −0.81 mm, 95% CI −1.14 to −0.47, p < 0.001), mature (−0.45 mm, 95% CI −0.85 to −0.06, p = 0.017) and kitten groups (−0.67 mm, 95% CI −1.20 to −0.14, p = 0.007). Mature cats also had a lower IAT than young cats (−0.35 mm, 95% CI −0.68 to −0.02, p = 0.030). The kitten group did not differ significantly from the young (p = 0.889) or from the mature group (p = 0.697).
4. Discussion
This study further characterizes age-related variation in IAT across defined feline life stages in a larger retrospective clinical MRI cohort of cats without apparent cognitive dysfunction than previously reported.
Our results identified a negative association between age, modeled as a continuous variable, and IAT on both T1W and T2W images. The negative association between age and IAT persisted on both sequences after adjustment for brain height, brain width, sex/neuter status, breed group, and scanner. In the secondary comparisons among life stages, senior cats had lower IAT than kittens, young adults and mature adults on both sequences. Mature adults also had lower IAT than young adults on T2W images, whereas this comparison was not statistically significant on T1W images. No statistically significant differences were identified between kittens and either young or mature adults.
These findings are consistent with those of two previous studies: one reported significantly lower mean IAT in cats aged 12 years or older compared with cats aged 1 to 6 years [10], and the other reported a significant decrease in interthalamic adhesion area and in the ratio of IAT to brain height with increasing age [11]. However, both studies included smaller samples. In contrast, our results differ from those of Hamamoto et al., who did not identify a significant decrease in IAT with age [5]. This discrepancy may be attributable to methodological differences, including the use of voxel-based morphometry and a smaller sample size (65 cats aged 1 to 16 years). Voxel-based morphometry may have limited sensitivity for detecting subtle changes in the interthalamic adhesion, a small midline structure, because spatial normalization, tissue segmentation, and smoothing may reduce anatomic specificity and obscure small focal differences.
In the present study, age was modeled as a continuous variable in the primary analyses, while categories adapted from the AAHA/AAFP Feline Life Stage Guidelines [12] were retained for secondary comparisons and clinical interpretation. Subdividing the senior category into cats younger than 15 years and those aged 15 years or older would have been of interest, given previous reports suggesting that age-related cerebral changes and/or behavioral changes attributed to FCDS become more pronounced in cats older than 15 years [3,10,13]. However, only four cats were aged 15 years or older, which was considered insufficient to support reliable comparisons between these subgroups. These cats were therefore retained within a single senior category.
The addition of a negative quadratic age term improved model fit on T1W images, suggesting a steeper reduction in IAT at older ages. This finding persisted after the exclusion of kittens. However, the quadratic term was not statistically significant on T2W images, and the small number of cats at the upper end of the age range limits the precision of the estimated curve in late life. Evidence for this curvature should therefore be considered exploratory.
Histopathological studies in aging cats have reported neuronal loss, synaptic degeneration, β-amyloid deposition and tau pathology [14,15,16,17]. Lower IAT may reflect neuronal loss within this structure and may also be associated with a reduced thalamic volume, as suggested by studies conducted in dogs [18] and humans [19]. Taken together, these observations suggest that IAT may be a relevant parameter for assessing age-related cerebral morphometric changes in cats.
The observed decrease in IAT with advancing age supports further investigation of IAT as a quantitative imaging marker of age-related cerebral changes in cats, as previously described in dogs [7]. Indeed, in healthy dogs, IAT has been reported to decrease with age [6,7,8]. Although the present study focused on cats without apparent cognitive dysfunction, it would be of interest to investigate whether brain atrophy is more pronounced in cats with FCDS, as has been suspected in cats [14] and demonstrated in dogs [7]. Given the challenges associated with diagnosing this condition, particularly because subtle behavioral changes may go unnoticed or be misinterpreted by owners, objective imaging measurements such as IAT may potentially contribute to a more objective assessment and facilitate earlier detection.
Several limitations should be considered when interpreting these findings. First, this retrospective study was conducted at a single referral hospital; therefore, the study population may not be representative of the general feline population. Cats underwent MRI for a range of clinical indications and had heterogeneous final diagnoses, which may have introduced selection bias despite the exclusion of cats with overt cerebral parenchymal lesions or owner-reported behavioral changes. Brachycephalic cats had lower adjusted IAT on both sequences, but only five were included, limiting the interpretation of this finding. The negative association between age and IAT remained significant after their exclusion.
Although the overall sample size was large, the distribution of cats across age groups was uneven. In the senior group, only four cats were aged 15 years or older, which limits the precision of estimates at the upper end of the age range.
Furthermore, cats were considered free of cognitive dysfunction based on retrospective medical records, without standardized behavioral assessment or a specific neurological examination. Subtle cognitive or behavioral changes may therefore have been present but unrecognized by owners.
In addition, MRI examinations were performed over an extended period with different scanners (1.0 Tesla until August 2015 and 1.5 Tesla thereafter), resulting in variation in image resolution. This technical heterogeneity may have affected the visualization of anatomic structures and the precision of IAT measurements, thereby contributing to measurement variability. Scanner was included as a covariate in the multivariable models and was associated with IAT on T1W but not T2W images. The negative association between age and IAT persisted after adjustment for scanner and the other covariates.
Finally, the use of a single dorsoventral measurement on 3 mm thick transverse images also introduces potential limitations related to partial-volume effects and slice positioning or orientation, which may affect the delineation of the interthalamic adhesion and measurement of its maximal thickness. The scanner-specific SEM and MDC95 estimates quantify the random component of interobserver measurement variability on the same images, but do not account for systematic errors shared by both observers or variability between separate MRI examinations. These estimates should therefore not be interpreted as validated thresholds for detecting longitudinal change in an individual cat. The present findings support a group-level association between age and IAT rather than individual-level diagnostic use. Midsagittal area measurements, three-dimensional volumetric measurements and approaches incorporating brain dimensions warrant evaluation as alternative or complementary morphometric methods. Their relative reproducibility and sensitivity to age-related changes should be assessed directly before superiority over a linear IAT measurement can be assumed.
5. Conclusions
Overall, IAT decreased with age in cats in this retrospective clinical MRI cohort. These findings support further evaluation of IAT as a simple morphometric marker of age-related cerebral changes in cats. Prospective studies involving cats with confirmed FCDS and age-matched control cats, standardized behavioral assessments and harmonized MRI protocols are required before clinical diagnostic use can be recommended.
Abbreviations
The following abbreviations are used in this manuscript:
| FCDS | Feline cognitive dysfunction syndrome |
| IAT | Interthalamic adhesion thickness |
| MRI | Magnetic resonance imaging |
| T1W | T1-weighted |
| T2W | T2-weighted |
| TR | Repetition time |
| TE | Echo time |
| FOV | Field of view |
Author Contributions
Conceptualization, O.T., R.K.-C. and M.F.; methodology, O.T., R.K.-C. and M.F.; validation, R.K.-C. and M.F.; formal analysis, R.K.-C.; investigation, O.T.; data curation, O.T. and R.K.-C.; writing—original draft preparation, R.K.-C.; writing—review and editing, R.K.-C., O.T. and M.F.; visualization, R.K.-C.; supervision, M.F.; project administration, M.F. All authors have read and agreed to the published version of the manuscript.
Institutional Review Board Statement
Ethical review and approval were not required for this retrospective observational study because it involved only the review of existing, anonymized clinical records and pre-existing MRI examinations obtained from client-owned cats as part of routine diagnostic care at the Veterinary Hospital, Oniris. No additional procedures, interventions, or imaging examinations were performed for the purpose of this study.
Informed Consent Statement
Informed consent for the retrospective use of anonymized clinical data from Oniris was obtained from Oniris.
Data Availability Statement
The data presented in this study are available on request from the corresponding author due to confidentiality restrictions associated with retrospective clinical records.
Conflicts of Interest
The authors declare no conflicts of interest.
Funding Statement
This research received no external funding.
Footnotes
[内容请查看原文]
References
[内容请查看原文]
Associated Data
Data Availability Statement
The data presented in this study are available on request from the corresponding author due to confidentiality restrictions associated with retrospective clinical records.