兽研简要摘要
猫可发生持续时间较长的鼻及鼻窦炎症，称为慢性鼻窦炎，该病通常难以治疗，有时还会引发是否需要手术的问题。兽医通常使用计算机断层扫描（CT）检查鼻及鼻窦，但CT表现的严重程度并不总是与猫的病情严重程度相符。在这项纳入10只猫的小型初步研究中，我们探讨了CT表现与临床症状之间的相互关系。两者并不一致，且总体CT评分未能识别出存在鼻窦引流局部阻塞的猫；这种阻塞而非总体CT评分，似乎是考虑手术时更具信息价值的CT特征。我们还描述了四种结合CT与临床发现的初步分型模式。这些观察结果为初步性质，旨在为更大规模的研究提供方向，而非直接指导治疗决策。
1. 引言
猫慢性鼻窦炎（CRS）是一种常见但临床处理困难的疾病，其特征为持续性或反复性鼻分泌物、打喷嚏以及不同程度的上呼吸道阻塞[1]。尽管在猫科临床中发病率较高，CRS仍是一种令人沮丧的疾病，因为许多患者对治疗反应不完全且频繁复发[1]。
传统上，猫CRS与既往病毒性上呼吸道感染相关，尤其是猫疱疹病毒1型（FHV-1），据认为该病毒可启动上皮损伤和黏膜功能障碍[1]。然而，越来越多的证据表明，仅持续性病毒感染并不能完全解释该病的慢性化和变异性[1,2]。相反，CRS似乎是初始黏膜损伤、继发性细菌定植、黏液纤毛清除功能受损以及鼻腔和鼻旁窦进行性结构重塑之间复杂相互作用的结果[1,3,4]。
近期研究强调了结构损伤在疾病进展中的作用。早年发生CRS的猫往往表现出更严重的鼻甲溶解和鼻窦异常，提示疾病表型可能因发病年龄和病程而异[5]。这些发现支持以下概念：猫CRS并非单一疾病实体，而是涉及炎症和结构改变不同贡献的异质性谱系[1,2]。
计算机断层扫描（CT）已成为评估猫鼻窦疾病的重要诊断工具，能够详细评估鼻甲结构、鼻窦受累情况及骨性重塑[3,6]。然而，在临床实践中，临床医生经常遇到CT表现与临床严重程度之间的不一致[3,6]。一些结构破坏明显的猫仅表现出轻微临床症状，而另一些CT异常轻微的猫却表现出严重的临床疾病[3]。这种不匹配提示，CT表现可能主要反映累积性结构损伤，而临床症状与当前炎症活动的关系更为密切[1,3,6]。
单一疾病标签可能涵盖机制上不同亚型的概念，在人类慢性鼻窦炎中已被日益认识，传统上伴有或不伴有鼻息肉的CRS二分法正被基于表型和内型的分类系统所补充[7,8]。人类CRS中提出的分类框架越来越多地纳入疾病的解剖分布（局部 vs. 弥漫）和炎症内型优势（2型 vs. 非2型），后者由组织嗜酸性粒细胞计数、细胞因子谱和免疫球蛋白E等分子生物标志物支持[7,8,9]。这些内型驱动的方法与不同的治疗反应相关，并已开始指导靶向生物制剂的使用[9,10]。值得注意的是，影像学严重程度与症状严重程度之间相关性较差在人类CRS中也有充分记录，进一步证实结构性疾病负担和炎症活动代表不同的病理生理领域；然而，这种分离尚未被采纳为人类CRS的主要分类轴，在人类CRS中基于可测量生物标志物的内型分类是可行的[11]。CT与临床分离的模式是否有助于表征猫CRS尚未被系统探讨。由于在人类CRS中实现内型驱动分类所需的分子生物标志物基础设施在兽医实践中尚未建立，基于临床可及工具——计算机断层扫描和临床严重程度评分——的探索性方法是研究结构损伤与炎症活动之间区别的合理第一步[7,8,9]。
考虑猫慢性鼻窦炎（CRS）中CT与临床表现分离的模式可能对治疗选择具有意义。目前，猫CRS的管理通常采用多模式方案，包括抗菌治疗、抗炎或免疫调节治疗以及支持性护理[1,12,13]。尽管采用了这些方法，治疗反应往往不完全，部分患者出现难治性[1,12]。重要的是，这些治疗策略针对的是一个复杂且多因素疾病过程的不同组成部分，没有任何单一模式能够持续治愈[1,2,4,12]。手术干预已在部分病例中有过描述，特别是涉及鼻窦阻塞或腔内滞留物质的病例，但结局不一且适应证仍不明确[14,15]。这种差异性可能反映了 underlying 疾病机制的差异，并凸显了更好地识别以机械性阻塞和鼻窦引流受损为主要作用的患者的必要性。
区分结构性损伤与炎症活动可能对治疗具有意义：疾病以炎症为主而无局灶性阻塞的猫可采用药物治疗，而局灶性阻塞和鼻窦引流受损则可能需要考虑手术以恢复通气和引流，无论是否合并炎症活动。如果手术考虑仅取决于CT严重程度，那么CT评分最高的猫将一律成为候选者；临床经验提示并非如此，这意味着可能需要对影像进行功能性而非纯定量解读。当前决策中的一个主要空白是缺乏一种将影像学发现与临床表现相联系的方法，以帮助识别哪些猫可考虑手术 versus 继续药物治疗[1,3]。
因此，本探索性初步研究的目标为：
(1) 探索CRS猫中基于CT的结构严重程度与临床严重程度之间的关系模式；
(2) 基于整合的CT和临床评估描述暂定的临床影像学模式；
(3) 探索CT上局灶性阻塞的证据而非整体CT严重程度是否可提供与手术考虑相关的额外信息。
我们假设CT表现与临床严重程度之间的差异部分反映了结构性损伤与炎症活动之间的区别，并且局灶性阻塞或鼻窦引流受损而非整体CT严重程度可能是与手术考虑更相关的CT特征[1,3]。鉴于样本量小、回顾性设计和新开发的评分系统，本研究旨在产生假设而非建立分类或手术决策框架。
2. 材料与方法
2.1. 研究设计与病例选择
本回顾性观察性病例系列研究在一家兽医转诊医院进行。检索了2024年2月至2026年2月期间因慢性上呼吸道症状（包括鼻分泌物、打喷嚏、喘鸣和鼻阻塞）就诊并接受鼻腔和鼻旁窦CT检查的客户拥有猫的医疗记录。在研究期间识别出22只具有慢性上呼吸道症状且可获得CT影像的猫，若CT及相关检查确定诊断为肿瘤或喉气管狭窄而非CRS，或影像质量不足以进行半定量评分，则排除这些猫。其余10只最终诊断为CRS的猫符合本病例系列的以下纳入标准：(1) 上呼吸道症状持续或反复发作至少4周；(2) 在转诊机构进行鼻腔和鼻旁窦CT检查；(3) 在CT检查时或2周内记录有临床严重程度评估。
筛查的22只猫中有10只符合纳入标准并纳入最终分析。从医疗记录中提取品种、年龄、性别、体重、临床病史和诊断检测结果（聚合酶链反应[PCR]、细胞学、细菌/真菌培养、组织病理学）；病毒学（PCR）结果与细胞学发现分开记录。还记录了随访数据，包括治疗反应和后续影像学（如可获得）。在所有10例中，细胞学标本仅通过鼻灌洗作为诊断程序获取；鼻灌洗并非用于或作为治疗干预，所有10只猫均可获得细胞学结果。微生物学检测（细菌/真菌培养和PCR）仅在一部分猫中进行。组织病理学需要额外的侵入性操作，因此仅在一部分病例中获得；在可获得的情况下，其被用作模式判定的支持性证据而非独立的鉴别标准。
2.2. CT采集与评分
CT examinations were performed under general anesthesia with cats positioned in sternal recumbency, using a 16-row multidetector CT scanner (Somatom Emotion, Siemens Medical Systems, Forchheim, Germany) in helical mode before and after intravenous administration of iohexol (Omnipaque 300, GE Healthcare, Shanghai, China; 2.0 mL/kg). Acquisition parameters were 130 kVp, 150 mAs, and pitch 0.9, with 1-mm soft-tissue reconstruction. A novel semiquantitative CT severity scoring system (0–12) was developed for this study, comprising four subscores each ranging from 0 to 3. This system differs from the Lund-Mackay staging system used in human CRS [16], which assigns per-sinus scores of 0–2 based on opacification extent, in three key respects: (1) it incorporates turbinate destruction as a distinct subscore, reflecting the central role of conchal pathology in feline sinonasal disease; (2) it grades intraluminal material by attenuation characteristics rather than by presence or absence of opacification alone; and (3) it includes bony remodeling as a separate domain to capture the osteoproliferative changes characteristic of chronic feline CRS. Content validity was established through iterative review and consensus among the investigators—including a board-certified veterinary radiologist and a small animal internal medicine specialist—with reference to previously published feline CT scoring approaches [3,6]. The four subscores are defined as follows:
[内容请查看原文]
鼻甲破坏（0–3）：0 = 结构正常；1 = 轻度、斑片状丢失；2 = 中度、部分破坏；3 = 重度、广泛或完全破坏。
[内容请查看原文]
鼻窦受累（0–3）：0 = 无鼻窦异常；1 = 单侧或一个鼻窦内少量液体/液平；2 = 双侧或中度液体积聚；3 = 双侧、弥漫性或完全鼻窦浑浊。
[内容请查看原文]
腔内物质（0–3）：0 = 无；1 = 仅为低衰减液体；2 = 中等软组织密度物质；3 = 高衰减、矿化或浓缩物质。
[内容请查看原文]
骨质重塑（0–3）：0 = 无；1 = 轻度骨膜或骨内膜反应；2 = 中度骨增生或骨变薄；3 = 重度重塑、骨质增生或超出鼻甲的骨质溶解。
根据CT证据定性评估局灶性阻塞的存在或缺失，作为二元变量：鼻窦内滞留高衰减浓缩物质、鼻咽部解剖性阻塞或其他表明鼻窦引流受损且适合手术干预的特征。
CT评分和阻塞评估由Jinhwa Chang，DVM，PhD（兽医放射科医师）进行，其了解临床表现，因为综合评估是研究设计的一部分；因此评分未设盲；然而，阻塞作为二元变量独立于手术候选分类进行评估。对于有连续CT检查的病例，使用索引扫描（最近或最完整的评估）进行主要分析；随访扫描使用相同的半定量系统评分并与索引扫描比较，并审查CT采集参数在各时间点之间的一致性。纵向变化以叙述方式描述。
2.3. 临床严重程度评分
A novel semiquantitative clinical severity scoring system (0–12) was developed for this study, based on the documented clinical assessment at or near the time of CT. Content validity was established through consensus among the investigators (Gippeum Lee, Jinhwa Chang, and Jeongmin Lee), with subscore domains selected to capture the principal clinical manifestations of feline CRS, and the system underwent face validation by two clinicians—Jeongmin Lee, DVM, PhD (small animal internal medicine specialist), and Jinhwa Chang, DVM, PhD (veterinary radiologist)—prior to retrospective application. The scoring system has not been formally validated, and inter- and intraobserver agreement were not assessed. The four subscores are defined as follows:
[内容请查看原文]
鼻分泌物严重程度（0–3）：0 = 无；1 = 间歇性浆液性或轻度分泌物；2 = 持续性黏液性或黏液脓性分泌物；3 = 大量脓性分泌物、结痂、血性分泌物或喷射状模式。
[内容请查看原文]
打喷嚏频率（0–3）：0 = 无；1 = 偶尔；2 = 每日、反复；3 = 频繁或严重打喷嚏，包括影响睡眠的打喷嚏或反复打喷嚏发作。
[内容请查看原文]
气流阻塞/喘鸣（0–3）：0 = 无；1 = 轻度、间歇性呼吸音；2 = 持续性喘鸣或临床记录的单侧鼻阻塞证据；3 = 明显鼻阻塞、归因于阻塞的睡眠障碍或明显的呼吸用力。该子评分反映临床记录的阻塞（例如，可听见的喘鸣、观察到的呼吸用力），而非CT定义的结构性阻塞，后者单独评分（第2.2节）。
[内容请查看原文]
功能/生活质量影响（0–3）：0 = 无明显功能影响；1 = 轻度影响；2 = 中度影响，需要持续管理；3 = 对日常活动、睡眠、食欲或其他重要生活质量方面有明确影响，基于同期病历证据，如睡眠障碍、进食或活动受损、需要反复就诊或持续管理、停止治疗时临床迅速恶化或主人明确报告的恶化。
临床评分由兽医、理学硕士Gippeum Lee（猫科医学专家）根据病历进行回顾性评定；评估者对CT结果未设盲，因为在临床实践中临床与影像数据是结合解读的。
2.4. 探索性临床影像学模式分类
根据以下内容的综合解读，将病例归入暂定的临床影像学模式（暂定表型）：（1）CT总严重程度评分及子评分特征，（2）临床总严重程度评分，（3）CT-临床分离的方向和幅度（CT评分减去临床评分），（4）CT上局灶性阻塞的定性评估，以及（5）鼻腔灌洗细胞学（所有猫均可获得）以及（在可获得的情况下）微生物学数据（PCR结果、细菌/真菌培养）。分组由研究者在审查每例病例的所有可用数据后达成共识完成，未设盲。由于同一数据被用于定义和描述这些模式，该分组为描述性的，未经验证。
四种暂定模式的主要判别轴如下。CT上局灶性阻塞或窦道引流受损的证据是区分阻塞型与其他三型的最明确标准。在非阻塞性病例中，鼻甲结构的保留与破坏是下一个判别指标：以鼻甲破坏为主的病例归入结构/重塑型，而鼻甲结构保留的病例则考虑归入活动/渗出型或免疫/黏膜型。在这两种鼻甲保留型之间，CT和临床严重程度水平以及细胞学模式（中性粒细胞性与淋巴浆细胞性炎症）作为进一步的判别标准：具有明显腔内渗出物、中性粒细胞性细胞学和相对较高临床严重程度的病例归入活动/渗出型，而CT和临床表现轻微且细胞学呈淋巴浆细胞性模式的病例归入免疫/黏膜型。
2.5. 概念性手术候选类别
手术候选类别是预先基于概念定义的，依据是CT上局灶性阻塞或窦道引流受损的证据，而非CT总严重程度评分。病例被归类为：内科治疗（未确定手术靶点；有内科管理指征）、保守治疗（存在结构损伤但无局灶性阻塞；适合继续保守管理）、考虑手术干预（边缘性阻塞特征；若内科管理失败需重新考虑手术选项）或手术候选（存在明确的阻塞靶点，适合手术引流或减压）。该逻辑基于人慢性鼻窦炎手术管理临床实践指南中阐述的原则，即内科治疗难以控制的窦道引流阻塞是手术干预的主要理由[17]。因此，局灶性阻塞既是阻塞型模式的定义标准，也是较高手术候选类别的定义标准，二者之间的重叠是定义性的而非研究发现。利用这些数据探索的问题是，本身并不定义阻塞的CT严重程度评分总和是否与阻塞状态相一致。本系列中无猫接受手术（例如额窦环钻术）；这些类别是对CT发现的概念性解读，尚未针对手术结局进行验证。
2.6. 统计分析
鉴于样本量小（n=10），所有统计分析均视为探索性和假设生成性的。计算Spearman秩相关系数（rs）以评估CT与临床严重程度评分之间的单调关联。使用10,000次重抽样估计rs的偏差校正和加速（BCa）bootstrap 95%置信区间；选择BCa方法而非更简单的百分位数方法，是因为它同时校正bootstrap分布中的偏倚和偏态[18,19]。尽管如此，在n=10时，置信区间应谨慎解读，因为在极小样本中bootstrap置信区间的覆盖度——即使采用BCa方法——也会降低[18,20]。描述性统计以中位数和范围（或在指明处为四分位距[IQR]）报告。相关系数、置信限和p值报告至小数点后三位，评分的中位数和四分位距报告至小数点后一位。p值<0.05被视为统计学显著性阈值，但需注意样本量小大幅限制了统计效能，并增加了I类和II类错误的风险。鉴于所有分析的探索性性质，未进行多重比较校正。分析使用Python、SciPy和NumPy进行，结果使用Python 3.13.16、SciPy 1.18.1和NumPy 2.5.3确认。
3. 结果
3.1. 研究群体
纳入10只经诊断为CRS的客户所有猫。品种包括9只韩国短毛猫和1只英国短毛猫。样本包括4只绝育雌性、4只去势雄性和2只性别未记录的猫。CT检查时年龄范围为1至8岁（中位数3.5岁）；体重范围为3.0至5.77 kg（中位数4.85 kg）。所有猫均有持续超过4周的上呼吸道症状，表现为持续性或复发性，并在单一转诊机构接受了CT检查。
3.2. CT与临床严重程度评分分布
CT严重程度评分范围为2至10（中位数8.0；IQR 5.3–10.0），临床严重程度评分范围为4至11（中位数6.5；IQR 6.0–9.3）。最高CT评分（10/12）见于4只猫（病例2、4、6和10），最低评分（2/12）见于病例3。最高临床评分（11/12）记录于病例1，最低评分（4/12）见于病例3。各病例评分及CT子评分见表1。
3.3. CT–临床相关性
CT与临床严重程度之间的Spearman秩相关系数为rs = 0.250（p = 0.486），bootstrap 95%置信区间为−0.558至0.885（10,000次重采样）。该区间极宽，几乎跨越了相关系数的整个可能范围，无法解释关联的方向或幅度，反映了n = 10时相关性估计的不精确性；无统计学显著相关性不应被解释为无关系的证据。各病例在CT–临床谱上分布广泛，无一致的单调趋势（图1）。
CT–临床分离的方向和幅度在各病例间存在差异（CT—临床评分，表1）。差值范围为−5至+5。负值表示临床严重程度超过CT所见，在病例1中最为明显（−5），该病例临床评分最高（11/12），但CT所见仅为中度（6/12）。正值表示CT所见超过临床严重程度，在病例6（+5）和病例4（+4）中最为明显，两者CT评分均为10/12，但临床评分相对较轻（分别为5/12和6/12）。
3.4. 初步临床影像学模式
综合解读CT所见、临床严重程度、CT—临床分离模式及细胞学数据，并结合可获得的微生物学数据，将10只猫归为四种初步临床影像学模式（表2、图2和图3）。各模式简要总结如下；病例级评分见表1。
[图：四种初步临床影像学模式的鼻腔和鼻窦代表性横断面CT图像。]
活动性/渗出性（n = 2；病例1和9）。CT评分分别为6和5，临床评分分别为11和6。两只猫均保留鼻甲结构（鼻甲子评分1），双侧鼻窦积液/渗出物（鼻窦子评分3）而无骨质重塑，灌洗液细胞学为中性粒细胞性（病例9为轻度）。另外，病例9检测到FHV-1（病毒学结果，非细胞学发现）。两只猫均被认为适合内科管理。
结构性/重塑性（n = 4；病例2、4、5和8）。这些猫有鼻甲破坏（子评分2–3）和骨质重塑（子评分2–3），CT评分为5–10（中位数9.0），临床评分为6–7（中位数6.5），CT–临床差异分别为+3、+4、+1和−1。均无局灶性阻塞，均归类为保守管理候选。病例2和4有系列CT资料（图4）：病例2在12个月内显示进行性鼻甲破坏和新发额窦矿化，而病例4在6年内显示稳定的重塑外观，组织病理学显示慢性淋巴细胞浆细胞性鼻炎伴骨质增生。
阻塞性（n = 3；病例6、7和10）。CT评分分别为10、8和10，临床评分分别为5、10和10。阻塞情况在各猫间不同（图5）：左侧额窦内高衰减（约490 Hounsfield单位）滞留物伴内骨膜重塑（病例6）；软腭背侧偏斜导致鼻咽前部完全闭塞（病例7）；双侧鼻窦滞留物伴额窦骨质肥厚和筛骨溶解（病例10）。病例6被归类为“考虑手术干预”，病例7和10为手术候选（概念性分类；未进行手术）。
免疫/黏膜性（n = 1；病例3）。该猫CT评分最低（2/12），临床评分低（4/12），鼻甲保留，鼻窦受累轻微，细胞学模式为淋巴细胞浆细胞性，罕见中性粒细胞，无嗜酸性粒细胞。作为单只猫，仅描述以记录一种额外的探索性模式。
3.5. 阻塞状态、候选类别与CT严重程度
根据设计，CT上的局灶性阻塞定义了阻塞型模式以及“考虑外科干预”和“外科候选”类别；因此，所有三只存在阻塞的猫均被归入这些类别，而七只无阻塞的猫均未被归入（表1）。这种重叠是定义性的，并非作为研究发现呈现。所关注的问题是与其CT严重程度总评分的关系，该评分本身并不定义阻塞。有阻塞的猫CT评分数值上高于无阻塞的猫（中位数10.0；范围8–10 vs. 中位数6.0；范围2–10），但范围重叠，且由于组样本量小（3 vs. 7），未进行正式比较。
两只无阻塞的猫（病例2和病例4，均为结构/重塑型；临床评分分别为7/12和6/12）CT评分最高，为10/12，而病例7尽管存在鼻咽部阻塞，CT评分较低（8/12）。因此，在这个小样本系列中，高CT总评分并不与局灶性阻塞一致。
4. 讨论
这项针对10只CRS猫的探索性初步研究考察了基于CT的结构严重程度与临床严重程度之间的关系，描述了四种暂定的临床影像学模式，并探讨了CT上局灶性阻塞的证据是否在总体CT严重程度之外提供了与外科考虑相关的信息。主要观察结果为：（1）CT和临床严重程度评分分布广泛，相关性估计过于不精确，无法支持方向性结论；（2）综合审查识别出四种暂定模式，每种模式均基于极少数猫，且均未经验证；（3）最高CT评分并不与局灶性阻塞一致，提示引流阻塞的定性CT特征可能携带总严重程度评分所不包含的信息。这些发现是描述性的，具有假设生成性质。
4.1. CT与临床的分离
弱正相关（rs = 0.250）伴随极宽的bootstrap 95% CI（−0.558至0.885），在n = 10时无法对CT与临床关系作出任何推断，也不应被解读为无关系的证据。然而，病例的离散程度（CT与临床差异从−5到+5）提供了描述性证据，表明CT严重程度和临床严重程度在个体猫中可能不一致。这与人类CRS中报道的影像学严重程度与症状严重程度之间相关性差的情况相平行[11]，Lund–Mackay分期系统本身即是在认识到CT表现与症状之间关系不可预测的背景下开发的[16]。在猫中，Schoenborn等人发现，对于62只鼻窦疾病猫，CT在确定疾病范围方面比X线摄影更敏感，但在检测有相关临床体征的猫的异常方面并不更敏感[3]，提示临床疾病可能在没有明显CT改变的情况下存在。
极端病例说明了这种分离：病例1临床评分最高（11/12），但CT评分为6/12，鼻甲保留且以渗出物为主的病变；而病例4 CT评分为10/12，但随访6年后临床评分为6/12。在这两只猫中，单一评分均未能反映疾病状态；这并不能确定其潜在机制。
单独影像学在区分炎症亚型方面的能力也有限：在一项77只猫的X线研究中，鼻甲破坏与肿瘤相关，但在慢性鼻炎中表现多变且非特异性[21]，而结合鼻腔和内侧咽后淋巴结CT特征可提高鼻炎与肿瘤的鉴别能力[22]。基于CT的严重程度评分本身是否可靠，将在下文讨论。
4.2. CT技术因素和判读陷阱
基于CT的评分也受到技术伪影的影响。在实验犬和猫研究中，周围液体降低CT上鼻甲的可见度，与骨完整性无关[23]，因此分泌物过多的猫鼻甲显示不清并不一定提示骨质溶解。这可能影响了渗出性病例的鼻甲子评分，是我们评分的一个局限性。未来研究应使用标准化CT方案，包括Uosyte等人推荐的中频和高频重建核[23]。
4.3. 暂定模式的解读
这四种模式是每组1至4只猫的描述性分组，系事后从其所描述的同一数据集中得出，使用了未经验证的评分和非盲法共识评估。因此，最好将其视为探索性观察而非既定表型，其中基于单只猫的免疫/黏膜模式尤为暂定。它们大致类似于人类CRS中基于内型的生物标志物所捕捉的结构损伤与炎症活动之间的区分[7,8,9]，但此处无生物标志物数据，且细胞学和微生物学结果与CT和临床数据一起参与了模式归类。
活动性/渗出性模式（n = 2）表现为鼻甲保留与渗出物为主的CT表现并存，临床严重程度超过CT严重程度，且细胞学为中性粒细胞性；其中性粒细胞性特征与人类非2型CRS内型相似[7,9]，但此比较仅为推测。结构/重塑模式（n = 4）表现为鼻甲破坏与骨重塑并存，CT评分等于或高于临床评分，且无局灶性阻塞，符合累积性结构改变超过当前炎症活动的表现。阻塞性模式（n = 3）以局灶性阻塞为定义特征，其机制异质（窦内滞留物、软腭偏斜致鼻咽阻塞、双侧窦内滞留物伴骨肥厚）；因此将其视为功能性 designation 而非单一解剖实体更为恰当。免疫/黏膜模式（n = 1）CT改变轻微，细胞学为淋巴浆细胞性，类似猫的淋巴浆细胞性鼻炎[24]，但其为独立实体还是非结构性谱系的轻型，仅凭一只猫无法确定。
4.4. 纵向观察
仅两只结构/重塑模式的猫有连续CT资料，任何关于时间演变的推断均为推测。病例2在12个月内显示进行性鼻甲破坏和新发额窦矿化，而病例4在6年内显示重塑外观稳定，细胞学炎症减轻。这些对比性病程提示单次CT仅代表可变轨迹中的一个快照，这与更广泛的气道重塑文献一致，即结构改变可能持续超过炎症刺激[25,26]，但我们的数据不允许就模式转换得出结论。病例2中那样的进行性矿化是否会导致滞留物和阻塞尚不清楚，需要前瞻性连续影像学检查。
4.5. 局灶性阻塞、CT严重程度与手术考量
手术候选类别系预先根据CT显示局灶性阻塞或引流受损而非根据CT总评分定义，局灶性阻塞也定义了阻塞性模式。因此，阻塞与候选资格之间的关联已内置于分类中，并未独立证明，且因无猫接受手术而无法在此验证。应区分两个不同问题：CT总评分衡量整体结构负担，而局灶性阻塞是引流的定性特征。由于阻塞的定义独立于总评分，两者可以比较，在本系列中总评分未能识别出有阻塞的猫：两只评分为最高10/12的猫无局灶性阻塞（病例2和4），而病例7评分为8/12，却有鼻咽阻塞。按CT表现对猫进行分类与此观察并不矛盾：未能追踪阻塞或临床严重程度的是总评分，而非CT作为模态，而各模式则综合了特定CT特征（鼻甲、窦、滞留物、骨和阻塞）与临床及细胞学表现进行解读。因此我们的观察仅限于描述性发现：在这一小系列中，较高的CT总评分与阻塞并不一致；这并不表明CT对手术结局的预测能力差。
这一推理与人类CRS手术指南一致，该指南建议不要仅根据任意CT阈值来规划鼻窦手术范围，并将药物治疗难治的引流阻塞作为主要手术指征[17]，也与功能性内镜鼻窦手术的逻辑一致，其靶点是窦口鼻道阻塞而非发炎的黏膜[27]；荟萃分析数据表明术前CT严重程度不能可靠预测术后症状改善[11]。在猫中，近期一例病例报告描述了 obstructive、治疗难治性CRS猫在额窦环钻和反复冲洗后获得持续缓解[28]；该猫不属于本系列。从人类指南和单例病例报告外推到猫科实践属于推测。
病例6（CT 10/12，临床5/12）说明阻塞性特征的手术权重可能不同：其高衰减额窦滞留物被归类为“考虑手术干预”，因为它并未明确导致引流受损，且该猫经内科管理后改善。在缺乏验证标准（如衰减阈值）的情况下，此类区分依赖临床判断。
4.6. 局限性
本研究存在重要局限性。样本量很小（n=10），且各模式仅包含一至四只猫，因此无法得出确定性结论，相关性估计也不精确。CT严重度评分和临床严重度评分均为本研究开发，尚未经过正式验证，也未评估观察者间或观察者内一致性；CT由一名放射科医师评分，临床评分由一名临床医师完成，二者均未对其他发现设盲，且主要的定量分析是将这两个新评分相互关联。模式分组是基于其所描述同一数据的事后构建，手术候选类别为概念性分类，设计上部分存在循环论证（第4.5节），系从人类指南外推而来[17]，且因无猫接受手术而未经手术结局验证。
其他局限性包括回顾性单中心设计、诊断数据异质（所有10只猫均有鼻腔灌洗细胞学检查，而微生物学检测和组织病理学仅部分猫有）、两只猫性别未记录、CT采集未标准化，以及渗出性病例中可能因液体导致鼻甲可见性被低估[23]。由于组间样本量限制，未进行组间统计学比较。
4.7. 未来方向
未来研究应：(1) 验证CT评分和临床评分，并在更大规模、理想为多中心前瞻性队列中检验其观察者间和观察者内一致性（对于约0.4的Spearman相关系数，95%CI宽度约±0.3需要大约40–50例[29]）；(2) 采用盲法评估和标准化CT方案，包括双核重建[23]；(3) 按模式记录结局，包括手术结局，以检验局灶性阻塞是否能识别从手术中获益的猫；(4) 加入细胞学、内镜和生物标志物数据以细化模式。
5. 结论
在这项纳入10只慢性鼻窦炎猫的初步研究中，CT严重度评分与临床严重度评分在个体猫中存在差异，无一致的单调关联，尽管相关性估计过于不精确而无法作出任何推断（rs=0.250，95%CI −0.558至0.885）。CT总评分最高的猫不一定有局灶性阻塞的CT证据，这提示阻塞可能携带总评分未包含的信息；由于阻塞也定义了阻塞性模式和概念性手术候选类别，这种重叠是设计使然，且无手术结局数据可用。
四种临床放射学模式各基于一至四只猫，属暂定性和假设生成性，这些发现并未建立手术决策框架。它们支持将CT局灶性阻塞证据而非仅整体CT严重度作为候选变量，在具有验证评分和手术结局的更大规模前瞻性研究中加以评估。
查看英文原文▾
Simple Summary
Cats can develop long-lasting inflammation of the nose and sinuses, called chronic rhinosinusitis, which is often difficult to treat and sometimes raises the question of surgery. Veterinarians commonly use computed tomography (CT) to examine the nose and sinuses, but how severe the CT findings appear does not always match how unwell the cat is. In this small pilot study of 10 cats, we explored how CT findings and clinical signs related to one another. The two did not consistently match, and the overall CT score did not identify the cats in which a localized obstruction of sinus drainage was present; this obstruction, rather than the overall CT score, appeared to be the more informative CT feature for considering surgery. We also describe four provisional patterns that combine CT and clinical findings. These observations are preliminary and are intended to guide larger studies rather than to inform treatment decisions directly.
1. Introduction
Feline chronic rhinosinusitis (CRS) is a common yet clinically challenging condition characterized by persistent or recurrent nasal discharge, sneezing, and varying degrees of upper airway obstruction [1]. Despite its high prevalence in feline practice, CRS remains a frustrating disease to manage, as many patients show incomplete response to therapy and frequent relapse [1].
Traditionally, feline CRS has been associated with antecedent viral upper respiratory infection, particularly feline herpesvirus type 1 (FHV-1), which is believed to initiate epithelial injury and mucosal dysfunction [1]. However, increasing evidence suggests that persistent viral infection alone does not fully explain the chronicity and variability of the disease [1,2]. Instead, CRS appears to result from a complex interaction between initial mucosal injury, secondary bacterial colonization, impaired mucociliary clearance, and progressive structural remodeling of the nasal cavity and paranasal sinuses [1,3,4].
Recent studies have emphasized the role of structural damage in disease progression. Cats that develop CRS at an early age tend to exhibit more severe turbinate lysis and sinus abnormalities, suggesting that disease phenotype may vary depending on age at onset and disease duration [5]. These findings support the concept that feline CRS is not a single disease entity, but rather a heterogeneous spectrum involving varying contributions of inflammation and structural change [1,2].
Computed tomography (CT) has become an essential diagnostic tool for evaluating feline sinonasal disease, enabling detailed assessment of turbinate architecture, sinus involvement, and osseous remodeling [3,6]. However, in clinical practice, clinicians frequently encounter a discrepancy between CT findings and clinical severity [3,6]. Some cats with marked structural destruction exhibit only mild clinical signs, whereas others with minimal CT abnormalities present with severe clinical disease [3]. This mismatch suggests that CT findings may primarily reflect cumulative structural damage, while clinical signs are more closely associated with current inflammatory activity [1,3,6].
The concept that a single disease label may encompass mechanistically distinct subtypes has been increasingly recognized in human chronic rhinosinusitis, where the traditional dichotomy of CRS with versus without nasal polyps is being supplemented by phenotype- and endotype-based classification systems [7,8]. Proposed classification frameworks in human CRS increasingly incorporate anatomical distribution of disease (local vs. diffuse) and inflammatory endotype dominance (type 2 vs. non-type 2), with the latter supported by molecular biomarkers such as tissue eosinophil counts, cytokine profiles, and immunoglobulin E [7,8,9]. These endotype-driven approaches have been linked to differential treatment responses and are beginning to inform the use of targeted biologic therapies [9,10]. Notably, the poor correlation between imaging severity and symptom severity has also been well documented in human CRS, reinforcing that structural disease burden and inflammatory activity represent distinct pathophysiologic domains; however, this dissociation has not been adopted as a primary classification axis in human CRS, where endotype classification based on measurable biomarkers is feasible [11]. Whether patterns of CT–clinical dissociation could help characterize feline CRS has not been systematically explored. Because the molecular biomarker infrastructure that enables endotype-driven classification in human CRS is not yet established in veterinary practice, an exploratory approach based on clinically accessible tools—computed tomography and clinical severity scoring—is a reasonable first step to examine the distinction between structural damage and inflammatory activity [7,8,9].
Considering CT–clinical dissociation patterns in feline CRS may have implications for treatment selection. Currently, management of feline CRS is typically multimodal, including antimicrobial therapy, anti-inflammatory or immunomodulatory treatment, and supportive care [1,12,13]. Despite these approaches, treatment response is often incomplete, and a subset of patients becomes refractory [1,12]. Importantly, these therapeutic strategies address different components of a complex and multifactorial disease process, and no single modality is consistently curative [1,2,4,12]. Surgical intervention has been described in selected cases, particularly those involving sinus obstruction or retained intraluminal material, but outcomes are variable and indications remain poorly defined [14,15]. This variability likely reflects differences in underlying disease mechanisms, and highlights the need to better identify patients in whom mechanical obstruction and impaired sinus drainage play a primary role.
Distinguishing structural damage from inflammatory activity may have treatment implications: cats whose disease is primarily inflammatory without focal obstruction may be managed medically, whereas focal obstruction and impaired sinus drainage may warrant consideration of surgery to restore ventilation and drainage, regardless of concurrent inflammatory activity. If surgical consideration depended on CT severity alone, cats with the highest CT scores would uniformly be candidates; clinical experience suggests otherwise, implying that a functional rather than purely quantitative interpretation of imaging may be needed. A major gap in current decision-making is the lack of an approach that links imaging findings with clinical presentation and helps identify cats for which surgery might be considered versus continued medical management [1,3].
Therefore, the objectives of this exploratory pilot study were:
(1) to explore the pattern of relationship between CT-based structural severity and clinical severity in cats with CRS;
(2) to describe provisional clinicoradiologic patterns based on integrated CT and clinical assessment;
(3) to explore whether CT evidence of focal obstruction, rather than overall CT severity, may provide additional information relevant to surgical consideration.
We hypothesized that the discrepancy between CT findings and clinical severity reflects, in part, the distinction between structural damage and inflammatory activity, and that focal obstruction or impaired sinus drainage, rather than overall CT severity, may be the more relevant CT feature for surgical consideration [1,3]. Given the small sample size, retrospective design, and newly developed scores, this study is intended to generate hypotheses rather than to establish a classification or a surgical decision-making framework.
2. Materials and Methods
2.1. Study Design and Case Selection
This retrospective observational case series was conducted at a single veterinary referral hospital. Medical records were searched for client-owned cats evaluated between February 2024 and February 2026 for chronic upper respiratory signs (including nasal discharge, sneezing, stertor, and nasal obstruction) that underwent CT examination of the nasal cavity and paranasal sinuses. Of 22 cats identified with chronic upper respiratory signs and available CT imaging during the study period, cats were excluded if CT and associated work-up established a diagnosis of neoplasia or laryngotracheal stenosis rather than CRS, or if imaging was of insufficient quality for semiquantitative scoring. The remaining 10 cats with a final diagnosis of CRS met the following inclusion criteria for this case series: (1) persistent or recurrent upper respiratory signs for a minimum of 4 weeks’ duration; (2) CT examination of the nasal cavity and paranasal sinuses performed at the referral institution; and (3) clinical severity assessment documented at or within 2 weeks of the CT examination.
Ten of the 22 screened cats met the inclusion criteria and were included in the final analysis. Breed, age, sex, body weight, clinical history, and results of diagnostic testing (polymerase chain reaction [PCR], cytology, bacterial/fungal culture, histopathology) were extracted from medical records; virological (PCR) results were recorded separately from cytologic findings. Follow-up data, including treatment response and subsequent imaging where available, were also recorded. In all 10 cases, cytologic specimens were obtained via nasal lavage performed as a diagnostic procedure only; nasal lavage was not intended or used as a therapeutic intervention, and cytologic results were available for all 10 cats. Microbiologic testing (bacterial/fungal culture and PCR) was available only in a subset of cats. Histopathology required an additional invasive procedure and was therefore available in only a subset of cases; where available, it was used as supportive evidence for pattern assignment rather than as an independent discriminating criterion.
2.2. CT Acquisition and Scoring
CT examinations were performed under general anesthesia with cats positioned in sternal recumbency, using a 16-row multidetector CT scanner (Somatom Emotion, Siemens Medical Systems, Forchheim, Germany) in helical mode before and after intravenous administration of iohexol (Omnipaque 300, GE Healthcare, Shanghai, China; 2.0 mL/kg). Acquisition parameters were 130 kVp, 150 mAs, and pitch 0.9, with 1-mm soft-tissue reconstruction. A novel semiquantitative CT severity scoring system (0–12) was developed for this study, comprising four subscores each ranging from 0 to 3. This system differs from the Lund-Mackay staging system used in human CRS [16], which assigns per-sinus scores of 0–2 based on opacification extent, in three key respects: (1) it incorporates turbinate destruction as a distinct subscore, reflecting the central role of conchal pathology in feline sinonasal disease; (2) it grades intraluminal material by attenuation characteristics rather than by presence or absence of opacification alone; and (3) it includes bony remodeling as a separate domain to capture the osteoproliferative changes characteristic of chronic feline CRS. Content validity was established through iterative review and consensus among the investigators—including a board-certified veterinary radiologist and a small animal internal medicine specialist—with reference to previously published feline CT scoring approaches [3,6]. The four subscores are defined as follows:
[内容请查看原文]
Turbinate destruction (0–3): 0 = normal architecture; 1 = mild, patchy loss; 2 = moderate, partial destruction; 3 = severe, extensive or complete destruction.
[内容请查看原文]
Sinus involvement (0–3): 0 = no sinus abnormality; 1 = unilateral or minimal fluid/fluid-level in one sinus; 2 = bilateral or moderate fluid accumulation; 3 = bilateral, diffuse, or complete sinus opacification.
[内容请查看原文]
Intraluminal material (0–3): 0 = none; 1 = low-attenuation fluid only; 2 = moderate soft tissue density material; 3 = high-attenuation, mineralized, or inspissated material.
[内容请查看原文]
Bony remodeling (0–3): 0 = none; 1 = mild periosteal or endosteal reaction; 2 = moderate osseous proliferation or bone thinning; 3 = severe remodeling, hyperostosis, or osteolysis beyond turbinates.
The presence or absence of focal obstruction was assessed qualitatively as a binary variable based on CT evidence of retained high-attenuation inspissated material within a sinus, anatomical obstruction of the nasopharynx, or other features indicating impaired sinus drainage amenable to surgical intervention.
CT scoring and obstruction assessment were performed by Jinhwa Chang, DVM, PhD (veterinary radiologist), with knowledge of clinical presentation, as integrated assessment was part of the study design; scoring was therefore not blinded; however, obstruction was assessed as a binary variable independently of surgical candidacy classification. In cases with serial CT examinations, the index scan (most recent or most complete evaluation) was used for the primary analysis; follow-up scans were scored using the identical semiquantitative system and compared to the index scan, and CT acquisition parameters were reviewed for consistency across time points. Longitudinal changes were described narratively.
2.3. Clinical Severity Scoring
A novel semiquantitative clinical severity scoring system (0–12) was developed for this study, based on the documented clinical assessment at or near the time of CT. Content validity was established through consensus among the investigators (Gippeum Lee, Jinhwa Chang, and Jeongmin Lee), with subscore domains selected to capture the principal clinical manifestations of feline CRS, and the system underwent face validation by two clinicians—Jeongmin Lee, DVM, PhD (small animal internal medicine specialist), and Jinhwa Chang, DVM, PhD (veterinary radiologist)—prior to retrospective application. The scoring system has not been formally validated, and inter- and intraobserver agreement were not assessed. The four subscores are defined as follows:
[内容请查看原文]
Nasal discharge severity (0–3): 0 = none; 1 = intermittent serous or mild discharge; 2 = persistent mucoid or mucopurulent discharge; 3 = profuse purulent discharge, crusting, blood-tinged discharge, or spraying pattern.
[内容请查看原文]
Sneezing frequency (0–3): 0 = none; 1 = occasional; 2 = daily, recurrent; 3 = frequent or severe sneezing, including sleep-disturbing sneezing or repeated sneezing bursts.
[内容请查看原文]
Airflow obstruction/stertor (0–3): 0 = none; 1 = mild, intermittent respiratory noise; 2 = persistent stertor or clinically documented evidence of unilateral nasal obstruction; 3 = marked nasal blockage, sleep disturbance attributable to obstruction, or overt respiratory effort. This subscore reflects clinically documented obstruction (e.g., audible stertor, observed respiratory effort) rather than CT-defined structural obstruction, which was scored separately (Section 2.2).
[内容请查看原文]
Functional/quality-of-life impact (0–3): 0 = no discernible functional impact; 1 = mild impact; 2 = moderate impact requiring ongoing management; 3 = clear impact on daily activity, sleep, appetite, or other important aspects of quality of life, based on contemporaneous medical-record evidence such as sleep disturbance, impaired eating or activity, need for repeated visits or ongoing management, rapid clinical worsening upon treatment discontinuation, or explicit owner-reported deterioration.
Clinical scores were assigned retrospectively from medical records by Gippeum Lee, DVM, MS (feline medicine specialist); the evaluator was not blinded to CT findings, as clinical and imaging data were interpreted together in clinical practice.
2.4. Exploratory Clinicoradiologic Pattern Classification
Cases were grouped into provisional clinicoradiologic patterns (provisional phenotypes) based on integrated interpretation of (1) total CT severity score and subscore profile, (2) total clinical severity score, (3) the direction and magnitude of CT–clinical dissociation (CT score minus clinical score), (4) qualitative assessment of focal obstruction on CT, and (5) nasal lavage cytology (available for all cats) and, where available, microbiologic data (PCR results, bacterial/fungal culture). The grouping was performed by consensus among the investigators after reviewing all available data for each case, without blinding. Because the same data were used to define and describe the patterns, the grouping is descriptive and was not validated.
The principal discriminating axes underlying the four provisional patterns were as follows. CT evidence of focal obstruction or impaired sinus drainage was the clearest criterion distinguishing the obstructive pattern from the other three. Among non-obstructive cases, preservation versus destruction of turbinate architecture was the next discriminant: cases with predominant turbinate destruction were assigned to the structural/remodeling pattern, whereas cases with preserved turbinate architecture were considered for either the active/exudative or immune/mucosal pattern. Between these two turbinate-preserved patterns, the level of CT and clinical severity together with the cytologic pattern (neutrophilic versus lymphoplasmacytic inflammation) served as the further discriminating criterion: cases with prominent intraluminal exudate, neutrophilic cytology, and comparatively high clinical severity were assigned to the active/exudative pattern, whereas cases with mild CT and clinical findings and a lymphoplasmacytic cytologic pattern were assigned to the immune/mucosal pattern.
2.5. Conceptual Surgical Candidacy Categories
Surgical candidacy categories were defined a priori and conceptually, based on CT evidence of focal obstruction or impaired sinus drainage rather than on the total CT severity score. Cases were categorized as: Medical (no surgical target identified; medical management indicated), Conservative (structural damage present but no focal obstruction; continued conservative management appropriate), Consider surgical intervention (borderline obstructive features; surgical option to be reconsidered if medical management fails), or Surgical candidate (clear obstructive target amenable to surgical drainage or decompression). This logic was grounded in the principle, articulated in clinical practice guidelines for the surgical management of human chronic rhinosinusitis, that sinus drainage obstruction refractory to medical therapy is a principal justification for surgical intervention [17]. Focal obstruction was therefore a defining criterion of both the obstructive pattern and the upper candidacy categories, and the overlap between them is definitional rather than a finding. The question explored with these data was whether the summed CT severity score, which does not itself define obstruction, tracked with obstruction status. No cat in the present series underwent surgery (e.g., frontal sinus trephination); the categories are a conceptual interpretation of CT findings and have not been validated against surgical outcomes.
2.6. Statistical Analysis
Given the small sample size (n = 10), all statistical analyses were considered exploratory and hypothesis-generating. Spearman’s rank correlation coefficient (rs) was calculated to assess the monotonic association between CT and clinical severity scores. A bias-corrected and accelerated (BCa) bootstrap 95% confidence interval for rs was estimated using 10,000 resamples; the BCa method was selected over the simpler percentile method because it corrects for both bias and skewness in the bootstrap distribution [18,19]. Nonetheless, at n = 10, the confidence interval should be interpreted with caution, as coverage of bootstrap confidence intervals—even with the BCa method—is reduced in very small samples [18,20]. Descriptive statistics are reported as median and range (or interquartile range [IQR] where indicated). Correlation coefficients, confidence limits, and p values are reported to three decimal places, and medians and interquartile ranges of scores to one decimal place. A p value < 0.05 was considered the threshold for statistical significance, with the caveat that the small sample size substantially limits statistical power and increases the risk of both type I and type II error. No adjustment for multiple comparisons was applied, given the exploratory nature of all analyses. Analyses were performed using Python, SciPy, and NumPy, and the results were confirmed using Python 3.13.16, SciPy 1.18.1, and NumPy 2.5.3.
3. Results
3.1. Study Population
Ten client-owned cats diagnosed with CRS were included. Breeds comprised nine Korean Shorthair and one British Shorthair. The sample included four spayed females, four neutered males, and two cats of unrecorded sex. Age at the time of CT ranged from 1 to 8 years (median 3.5 years); body weight ranged from 3.0 to 5.77 kg (median 4.85 kg). All cats had persistent or recurrent upper respiratory signs exceeding 4 weeks and underwent CT at a single referral institution.
3.2. CT and Clinical Severity Score Distribution
CT severity scores ranged from 2 to 10 (median 8.0; IQR 5.3–10.0) and clinical severity scores ranged from 4 to 11 (median 6.5; IQR 6.0–9.3). The highest CT scores (10/12) were observed in four cats (Cases 2, 4, 6, and 10), and the lowest (2/12) in Case 3. The highest clinical score (11/12) was recorded in Case 1 and the lowest (4/12) in Case 3. Individual case scores and CT subscores are presented in Table 1.
3.3. CT–Clinical Correlation
Spearman’s rank correlation between CT and clinical severity was rs = 0.250 (p = 0.486), with a bootstrap 95% confidence interval of −0.558 to 0.885 (10,000 resamples). This extremely wide interval, spanning nearly the entire possible range of correlation coefficients, precludes interpretation of either the direction or the magnitude of the association and reflects the imprecision of correlation estimates at n = 10; the absence of a statistically significant correlation should not be interpreted as evidence of no relationship. Cases were widely dispersed across the CT–clinical spectrum without a consistent monotonic trend (Figure 1).
The direction and magnitude of CT–clinical dissociation varied across cases (CT—Clinical score, Table 1). The difference ranged from −5 to +5. Negative values, indicating clinical severity exceeding CT findings, were most pronounced in Case 1 (−5), the case with the highest clinical score (11/12) but only moderate CT findings (6/12). Positive values, indicating CT findings exceeding clinical severity, were most pronounced in Case 6 (+5) and Case 4 (+4), both with CT scores of 10/12 but relatively mild clinical scores (5/12 and 6/12, respectively).
3.4. Provisional Clinicoradiologic Patterns
Integrated interpretation of CT findings, clinical severity, the pattern of CT—clinical dissociation, and cytologic data, together with microbiologic data where available, grouped the 10 cats into four provisional clinicoradiologic patterns (Table 2, Figure 2 and Figure 3). Each pattern is summarized briefly below; case-level scores are given in Table 1.
[图：Representative transverse CT images of the nasal cavity and sinuses for the four provisional clinicoradiologic patterns.]
Active/exudative (n = 2; Cases 1 and 9). CT scores were 6 and 5 and clinical scores 11 and 6. Both cats had preserved turbinate architecture (turbinate subscore 1), bilateral sinus fluid/exudate (sinus subscore 3) without bony remodeling, and neutrophilic lavage cytology (mild in Case 9). Separately, FHV-1 was detected in Case 9 (virological result, not a cytologic finding). Both cats were considered candidates for medical management.
Structural/remodeling (n = 4; Cases 2, 4, 5, and 8). These cats had turbinate destruction (subscores 2–3) and bony remodeling (subscores 2–3), with CT scores of 5–10 (median 9.0), clinical scores of 6–7 (median 6.5), and CT–clinical differences of +3, +4, +1, and −1. None had focal obstruction, and all were categorized as conservative-management candidates. Serial CT was available for Cases 2 and 4 (Figure 4): Case 2 showed progressive turbinate destruction and new frontal sinus mineralization over 12 months, whereas Case 4 showed a stable remodeled appearance over 6 years, with histopathology showing chronic lymphoplasmacytic rhinitis with osseous proliferation.
Obstructive (n = 3; Cases 6, 7, and 10). CT scores were 10, 8, and 10 and clinical scores 5, 10, and 10, respectively. The obstruction differed among cats (Figure 5): high-attenuation (~490 Hounsfield units) retained material in the left frontal sinus with endosteal remodeling (Case 6); complete obliteration of the rostral nasopharynx by dorsal soft palate deviation (Case 7); and bilateral retained sinus material with frontal sinus hyperostosis and ethmoidal bone lysis (Case 10). Case 6 was categorized as “consider surgical intervention” and Cases 7 and 10 as surgical candidates (conceptual categories; no surgery was performed).
Immune/mucosal (n = 1; Case 3). This cat had the lowest CT score (2/12) and a low clinical score (4/12), with preserved turbinates, minimal sinus involvement, and a lymphoplasmacytic cytologic pattern with rare neutrophils and no eosinophils. Being a single cat, it is described only to document an additional exploratory pattern.
3.5. Obstruction Status, Candidacy Categories, and CT Severity
By design, focal obstruction on CT defined the obstructive pattern and the “consider surgical intervention” and “surgical candidate” categories; all three cats with obstruction were therefore assigned to these categories, and none of the seven cats without obstruction were (Table 1). This overlap is definitional and is not presented as a finding. The question of interest is the relation to the summed CT severity score, which does not itself define obstruction. Cats with obstruction had numerically higher CT scores (median 10.0; range 8–10) than cats without obstruction (median 6.0; range 2–10), but the ranges overlapped, and no formal comparison was made because of the group sizes (3 vs. 7).
Two cats without obstruction (Cases 2 and 4, both structural/remodeling; clinical scores 7/12 and 6/12) had the maximum CT score of 10/12, whereas Case 7 had a lower CT score (8/12) with nasopharyngeal obstruction. In this small series, therefore, a high summed CT score did not coincide with focal obstruction.
4. Discussion
This exploratory pilot study of 10 cats with CRS examined the relationship between CT-based structural severity and clinical severity, described four provisional clinicoradiologic patterns, and explored whether CT evidence of focal obstruction adds information relevant to surgical consideration beyond overall CT severity. The principal observations were: (1) CT and clinical severity scores were widely dispersed, and the correlation estimate was too imprecise to support directional conclusions; (2) integrated review identified four provisional patterns, each based on very few cats and none validated; and (3) the highest CT scores did not coincide with focal obstruction, suggesting that qualitative CT features of drainage obstruction may carry information that a summed severity score does not. These findings are descriptive and hypothesis-generating.
4.1. CT–Clinical Dissociation
The weak positive correlation (rs = 0.250) with an extremely wide bootstrap 95% CI (−0.558 to 0.885) does not allow any inference about the CT–clinical relationship at n = 10 and should not be read as evidence of no relationship. Nevertheless, the dispersion of cases, with CT–clinical differences from −5 to +5, provides descriptive evidence that CT severity and clinical severity can diverge in individual cats. This parallels the poor correlation between imaging severity and symptom severity reported in human CRS [11], where the Lund–Mackay staging system was itself developed in recognition of the unpredictable relationship between CT appearance and symptoms [16]. In cats, Schoenborn et al. found that CT was more sensitive than radiography for defining disease extent in 62 cats with sinonasal disease but not for detecting abnormalities in cats with referable clinical signs [3], suggesting that clinical disease may exist without marked CT changes.
The extreme cases illustrate this divergence: Case 1 had the highest clinical score (11/12) but a CT score of 6/12 with preserved turbinates and exudate-predominant disease, whereas Case 4 had a CT score of 10/12 but a clinical score of 6/12 after 6 years of follow-up. Neither score alone captured disease status in these two cats; this does not establish the underlying mechanisms.
Imaging alone also has limited ability to separate inflammatory subtypes: in a radiographic study of 77 cats, turbinate destruction was associated with neoplasia but was variable and non-specific in chronic rhinitis [21], and combining nasal and medial retropharyngeal lymph node CT features improved discrimination between rhinitis and neoplasia [22]. Whether CT-based severity scores are themselves reliable measurements is addressed next.
4.2. CT Technical Considerations and Interpretive Pitfalls
CT-based scoring is also subject to technical artifacts. In experimental canine and feline studies, surrounding fluid reduced turbinate visibility on CT irrespective of bone integrity [23], so poorly visible turbinates in cats with copious secretions do not necessarily indicate osteolysis. This may have affected the turbinate subscores in exudative cases and is a limitation of our scoring. Future studies should use standardized CT protocols, including medium- and high-frequency reconstruction kernels as recommended by Uosyte et al. [23].
4.3. Interpretation of the Provisional Patterns
The four patterns are descriptive groupings of one to four cats each that were derived post hoc from the same dataset they describe, using unvalidated scores and non-blinded consensus assessment. They are therefore best regarded as exploratory observations rather than established phenotypes, and the immune/mucosal pattern, based on a single cat, is particularly provisional. They are loosely analogous to the distinction between structural damage and inflammatory activity that endotype-based biomarkers capture in human CRS [7,8,9], but no biomarker data were available here, and cytologic and microbiologic findings contributed to pattern assignment alongside CT and clinical data.
The active/exudative pattern (n = 2) combined preserved turbinates and exudate-predominant CT findings with clinical severity exceeding CT severity and neutrophilic cytology; the neutrophilic profile resembles non-type 2 CRS endotypes described in humans [7,9], although this comparison is speculative. The structural/remodeling pattern (n = 4) combined turbinate destruction and bone remodeling with CT scores equal to or higher than clinical scores and no focal obstruction, consistent with cumulative structural change outpacing current inflammatory activity. The obstructive pattern (n = 3) was defined by focal obstruction, and its mechanisms were heterogeneous (retained sinus material, nasopharyngeal obstruction by soft palate deviation, and bilateral retained sinus material with hyperostosis); it is therefore better regarded as a functional designation than as a single anatomical entity. The immune/mucosal pattern (n = 1) had minimal CT changes and lymphoplasmacytic cytology, resembling lymphoplasmacytic rhinitis described in cats [24], but whether it is a distinct entity or a mild form of the non-structural spectrum cannot be determined from one cat.
4.4. Longitudinal Observations
Serial CT was available for only two cats with the structural/remodeling pattern, and any inference about temporal evolution is speculative. Case 2 showed progressive turbinate destruction and new frontal sinus mineralization over 12 months, whereas Case 4 showed a stable remodeled appearance over 6 years with declining cytologic inflammation. These contrasting courses suggest that a single CT represents a snapshot along a variable trajectory, in line with the broader airway-remodeling literature in which structural change may persist beyond the inflammatory stimulus [25,26], but our data do not permit conclusions about pattern conversion. Whether progressive mineralization such as that in Case 2 can lead to retained material and obstruction is unknown and would require prospective serial imaging.
4.5. Focal Obstruction, CT Severity, and Surgical Consideration
Surgical candidacy categories were defined a priori by CT evidence of focal obstruction or impaired drainage rather than by the summed CT score, and focal obstruction also defined the obstructive pattern. The association between obstruction and candidacy is therefore built into the classification, was not independently demonstrated, and cannot be validated here because no cat underwent surgery. Two different questions should be separated: the summed CT score measures overall structural burden, whereas focal obstruction is a qualitative feature of drainage. Because obstruction was defined independently of the summed score, the two can be compared, and in this series the summed score did not identify the cats with obstruction: two cats with the maximum score of 10/12 had no focal obstruction (Cases 2 and 4), whereas Case 7, with a score of 8/12, had nasopharyngeal obstruction. Classifying cats by CT findings is not inconsistent with this observation: it is the summed score, not CT as a modality, that did not track obstruction or clinical severity, whereas the patterns drew on specific CT features (turbinate, sinus, material, bone, and obstruction) interpreted together with clinical and cytologic findings. Our observation is therefore limited to the descriptive finding that a higher total CT score did not coincide with obstruction in this small series; it does not show that CT is a poor predictor of surgical outcome.
This reasoning mirrors the human CRS surgical guideline, which advises against planning the extent of sinus surgery solely on arbitrary CT thresholds and treats drainage obstruction refractory to medical therapy as a principal surgical indication [17], and the logic of functional endoscopic sinus surgery, whose target is ostiomeatal obstruction rather than inflamed mucosa [27]; meta-analytic data indicate that preoperative CT severity does not reliably predict symptom improvement after surgery [11]. In cats, a recent case report described sustained resolution after frontal sinus trephination and repeated irrigation in a cat with obstructive, treatment-refractory CRS [28]; that cat was not part of the present series. Extrapolation from human guidelines and a single case report to feline practice is speculative.
Case 6 (CT 10/12, clinical 5/12) illustrates that obstructive features may differ in surgical weight: its high-attenuation frontal sinus material was categorized as “consider surgical intervention” because it was not clearly causing drainage impairment, and the cat improved with medical management. In the absence of validated criteria (e.g., attenuation thresholds), such distinctions rely on clinical judgment.
4.6. Limitations
This study has important limitations. The sample is very small (n = 10) and the patterns contain one to four cats each, so no definitive conclusions can be drawn and the correlation estimate is imprecise. The CT and clinical severity scores were developed for this study, have not been formally validated, and were not assessed for inter- or intraobserver agreement; CT was scored by one radiologist and clinical scores by one clinician, neither of whom was blinded to other findings, and the principal quantitative analysis relates these two new scores to each other. The pattern grouping is a post hoc construct derived from the same data it describes, and the surgical candidacy categories are conceptual, partly circular by design (Section 4.5), extrapolated from human guidelines [17], and unvalidated against surgical outcomes because no cat underwent surgery.
Additional limitations include the retrospective single-center design, heterogeneous diagnostic data (nasal lavage cytology was available for all 10 cats, whereas microbiologic testing and histopathology were available only in a subset), two cats with unrecorded sex, non-standardized CT acquisition, and possible fluid-related underestimation of turbinate visibility in exudative cases [23]. Between-group statistical comparisons were not performed because of the group sizes.
4.7. Future Directions
Future studies should (1) validate the CT and clinical scores and test their inter- and intraobserver agreement in a larger, ideally multicenter and prospective cohort (a 95% CI width of about ±0.3 for a Spearman correlation of approximately 0.4 requires roughly 40–50 cases [29]); (2) use blinded assessment and standardized CT protocols, including dual-kernel reconstruction [23]; (3) record outcomes by pattern, including surgical outcomes, to test whether focal obstruction identifies cats that benefit from surgery; and (4) add cytologic, endoscopic, and biomarker data to refine the patterns.
5. Conclusions
In this pilot study of 10 cats with chronic rhinosinusitis, CT severity and clinical severity scores diverged in individual cats without a consistent monotonic association, although the correlation estimate was too imprecise for any inference (rs = 0.250, 95% CI −0.558 to 0.885). Cats with the highest summed CT scores did not necessarily show CT evidence of focal obstruction, which suggests that obstruction may carry information that a summed score does not; because obstruction also defined the obstructive pattern and the conceptual surgical candidacy categories, this overlap is by design, and no surgical outcomes were available.
The four clinicoradiologic patterns, each based on one to four cats, are provisional and hypothesis-generating, and these findings do not establish a surgical decision-making framework. They support evaluating CT evidence of focal obstruction, rather than overall CT severity alone, as a candidate variable in larger prospective studies with validated scores and surgical outcomes.