Induced Sputum KL-6 and HRCT Scoring in Idiopathic Pulmonary Fibrosis

Background and Study Objective

Zhang and colleagues investigated whether KL-6 measured in induced sputum can help with the diagnosis and assessment of idiopathic pulmonary fibrosis (IPF). Specifically, the study asked two linked questions:

1. Can induced sputum KL-6 differentiate IPF patients from healthy subjects?

2. Does induced sputum KL-6 reflect disease severity, particularly when considered alongside a structured high-resolution CT (HRCT) fibrosis score?

Study Design, Setting, and Participants

This was a prospective, observational, single-center study conducted from October 2021 to April 2023. The authors enrolled:

• 20 patients with newly diagnosed IPF, not receiving antifibrotic therapy

• 20 age-matched healthy subjects as the control group

Induced Sputum Collection and KL-6 Quantification

Induced sputum was obtained using a standardized protocol:

• Salbutamol pre-treatment

• Induction with 5% hypertonic saline

Samples were processed using PBS dilution (the protocol did not include DTT). KL-6 concentrations were then measured by ELISA.

Physiologic Testing

Participants underwent pulmonary function evaluation including:

• Spirometry

• Diffusing capacity (DLCO)

The authors assessed relationships between induced sputum KL-6 and multiple physiologic indices, including DLCO-related parameters.

HRCT Fibrosis Scoring Method

HRCT scans were quantified with a structured scoring system:

• Lungs were divided into six regions

• Each region was scored from 0 to 5

• Total score ranged from 0 to 30

Two radiologists scored the scans, and the average score was used for analysis.

Statistical Analysis Framework

The study used:

• Spearman correlation to examine associations among KL-6, lung function measures, and HRCT fibrosis score

• Receiver operating characteristic (ROC) analysis to evaluate discrimination for:

  • KL-6 alone

  • HRCT score alone

  • Combined approaches (KL-6 + HRCT score)

• Bootstrap internal validation with 1000 resamples

Results

Between-Group Comparison: Induced Sputum KL-6

Induced sputum KL-6 concentrations were higher in IPF than in healthy controls:

• Median ~776 U/mL in IPF vs 322 U/mL in healthy subjects (P < 0.001)

This group separation supported the authors’ central premise that induced sputum KL-6 differs between IPF patients and healthy individuals.

Association With Physiologic Impairment

Higher induced sputum KL-6 was associated with worse pulmonary function. Reported correlations included:

• DLCO/VA: r = –0.872

• DLCO% predicted: r = –0.783

These findings indicate that, within the IPF group, higher KL-6 aligned with more impaired gas transfer metrics.

Association With HRCT Fibrosis Burden

Induced sputum KL-6 showed a strong positive correlation with imaging fibrosis severity:

• KL-6 vs HRCT score: r = 0.908

This relationship was a key result in the paper, linking a sputum-based biomarker to a structured imaging-based fibrosis score.

Diagnostic Discrimination (ROC Analyses)

KL-6 Alone

• AUC 0.844

• Cutoff ~623.78 U/mL

• Sensitivity 90%, specificity 67.5%

HRCT Score Alone

• AUC 0.899

• Cutoff 7.75

• Sensitivity 80%, specificity 85%

Combined Model: KL-6 + HRCT Score

• AUC 0.936

• Sensitivity 80%, specificity 97.5%
  The authors reported similar performance when the model was adjusted for age/smoking.

The combined approach produced the highest AUC and the highest specificity in this IPF-versus-healthy comparison.

Interpretation and Practical Implications Within This Study

Within the scope of this dataset, the results support two conclusions:

• Induced sputum KL-6 is elevated in IPF compared with healthy subjects.

• Induced sputum KL-6 aligns closely with both physiologic impairment and HRCT fibrosis scoring, and combining KL-6 with HRCT scoring improved discriminatory performance relative to either measure alone.

Study Constraints (Relevant to Interpretation)

The paper evaluates IPF against healthy controls and includes a small sample (20 per group) from a single center, with performance assessed using internal bootstrap validation.

 

The translation of the preceding English text in Chinese:

 

背景与研究目标

Zhang 及其同事探讨了在诱导痰中测定的 KL-6 是否有助于特发性肺纤维化（IPF）的诊断与评估。具体而言，本研究提出了两个相互关联的问题：

• 诱导痰 KL-6 能否区分 IPF 患者与健康受试者？

• 诱导痰 KL-6 能否反映疾病严重程度，尤其是在与结构化的高分辨率 CT（HRCT）纤维化评分结合考虑时？

研究设计、研究场所与参与者

本研究为前瞻性、观察性、单中心研究，开展时间为 2021 年 10 月至 2023 年 4 月。作者纳入：

• 20 例新诊断 IPF 患者（未接受抗纤维化治疗）

• 20 例年龄匹配的健康受试者作为对照组

诱导痰采集与 KL-6 定量

诱导痰采用标准化流程获取：

• 沙丁胺醇预处理

• 使用 5% 高渗盐水诱导

样本采用 PBS 稀释处理（流程未包含 DTT）。随后使用 ELISA 测定胆量 KL-6 浓度。

生理学检测

参与者接受肺功能评估，包括：

• 肺活量测定（spirometry）

• 弥散功能（DLCO）

作者评估了诱导痰 KL-6 与多项生理学指标之间的关系，其中包括与 DLCO 相关的参数。

HRCT 纤维化评分方法

HRCT 扫描采用结构化评分系统进行量化：

• 将肺部分为六个区域

• 每个区域评分范围为 0 到 5 分

• 总评分范围为 0 到 30 分

两名放射科医师对影像进行评分，并采用平均值用于分析。

统计分析框架

本研究采用：

• Spearman 相关分析，用于检验 KL-6、肺功能指标与 HRCT 纤维化评分之间的关联

• 受试者工作特征（ROC）曲线分析，用于评估以下指标的区分能力：

  • 单独 KL-6

  • 单独 HRCT 评分

  • 联合方法（KL-6 + HRCT 评分）

• 采用 1000 次重抽样的自助法（bootstrap）进行内部验证

结果

组间比较：诱导痰 KL-6

IPF 组的诱导痰 KL-6 浓度高于健康对照组：

• IPF 组中位数约 776 U/mL，健康受试者为 322 U/mL（P < 0.001）

这种组间差异支持了作者的核心观点：IPF 患者与健康个体之间的诱导痰 KL-6 存在差异。

与生理功能受损的关联

更高的诱导痰 KL-6 与更差的肺功能相关。报告的相关性包括：

• DLCO/VA：r = –0.872

• DLCO% 预计值：r = –0.783

这些发现提示，在 IPF 组内，更高的 KL-6 与更明显的气体交换受损指标相一致。

与 HRCT 纤维化负荷的关联

诱导痰 KL-6 与影像纤维化严重程度呈强正相关：

• KL-6 与 HRCT 评分：r = 0.908

这一关系是论文中的关键结果之一，将基于痰液的生物标志物与结构化的影像纤维化评分联系起来。

诊断区分能力（ROC 分析）

仅 KL-6

• AUC 0.844

• 截断值约 623.78 U/mL

• 敏感性 90%，特异性 67.5%

仅 HRCT 评分

• AUC 0.899

• 截断值 7.75

• 敏感性 80%，特异性 85%

联合模型：KL-6 + HRCT 评分

• AUC 0.936

• 敏感性 80%，特异性 97.5%

作者报告：在对年龄/吸烟进行调整后，该模型表现相近。

在本研究的 IPF 与健康对照比较中，联合方法获得了最高的 AUC 和最高的特异性。

本研究范围内的解读与实际意义

在该数据集范围内，结果支持两点结论：

• 与健康受试者相比，IPF 患者的诱导痰 KL-6 升高。

• 诱导痰 KL-6 与生理功能受损及 HRCT 纤维化评分均密切一致，且将 KL-6 与 HRCT 评分联合可较单独使用任一指标获得更好的区分表现。

研究限制（与结果解读相关）

本文将 IPF 与健康对照进行比较，样本量较小（每组 20 例），且为单中心研究；模型表现通过内部 bootstrap 验证进行评估。

---

Reference:

Bingxin Zhang, Dejun Zhao, Danping Hu

Induced sputum KL-6 combined with HRCT scoring for diagnosing and monitoring idiopathic pulmonary fibrosis.

Biomol Biomed [Internet]. 2025 Sep. 12 [cited 2025 Dec. 29];26(3):452–461.

Available from: https://www.bjbms.org/ojs/index.php/bjbms/article/view/12667

---

Additional information:

We invite submissions for our upcoming thematic issues, including:

• Immune Prediction and Prognostic Biomarkers in Immuno-Oncology
• Artificial Intelligence and Machine Learning in disease diagnosis and treatment target identification

More news: Blog