CLDN18.2 expression identifies a gastric-type phenotype and a potential therapeutic target in pulmonary invasive mucinous adenocarcinoma

Abstract

Aims

This study aimed to investigate the clinicopathological and immunophenotypic significance of Claudin18.2 (CLDN18.2) in pulmonary invasive mucinous adenocarcinoma (IMA), with particular focus on its association with gastric-type differentiation and potential implications for diagnosis and targeted therapy.

Methods

Immunohistochemistry for CLDN18.2, MUC5AC, MUC6, and PD-L1 (22C3-TPS) was performed in 117 surgically resected lung IMAs. Associations with clinicopathological parameters and overall survival were analysed. Correlation analyses among CLDN18.2, MUC5AC, and MUC6 expression were conducted using Spearman’s rank coefficients to assess the relationship between CLDN18.2 and gastric-type mucin phenotype.

Results

CLDN18.2 expression was detected in 71 of 117 tumors (60.7%), including 42 with high (35.9%) and 29 with low (24.8%) expression. MUC5AC and MUC6 positivity was observed in 75.2% and 34.2% of cases, respectively, while all tumors were PD-L1 negative. Spread through air spaces (STAS) was less frequent in CLDN18.2-positive (40.8% vs. 60.9%, P = 0.034) and MUC5AC-positive tumors (42% vs. 69%, P = 0.012). High CLDN18.2 expression was associated with earlier pathological stage (81% vs. 68%) and fewer nodal metastases (4.8% vs. 18%). CLDN18.2 expression correlated positively with MUC5AC (ρ = 0.469, P < 0.0001) and MUC6 (ρ = 0.43, P < 0.0001). Neither CLDN18.2 nor MUC5AC had prognostic impact, whereas MUC6 positivity tended to indicate shorter overall survival (P = 0.088).

Conclusions

CLDN18.2 is frequently expressed in lung IMA and is closely associated with gastric-type mucin differentiation and reduced aerogenous spread. Although not prognostic, its consistent expression in PD-L1-negative tumors suggests a distinct biological subset and supports the potential use of CLDN18.2 as a biomarker for molecular stratification and therapeutic exploration in this rare subtype of lung adenocarcinoma.

Introduction

Invasive mucinous adenocarcinoma (IMA) of the lung is a relatively uncommon subtype of lung adenocarcinoma, accounting for approximately 2–10% of cases. It is characterised by tumor cells containing abundant intracytoplasmic mucin and often presents a “pneumonia-like” or multifocal distribution on imaging [1–3]. This variant frequently exhibits spread through air spaces (STAS), a histological pattern associated with local recurrence and adverse prognosis [4, 5]. Molecularly, IMA is defined by a high frequency of KRAS mutations, whereas canonical driver alterations such as EGFR are relatively rare, limiting the applicability of EGFR tyrosine kinase inhibitors [1, 6]. In addition, most IMAs show low or absent PD-L1 expression, suggesting limited benefit from immune checkpoint blockade [1, 7–9]. For patients with advanced or recurrent disease, systemic treatment—primarily platinum-based chemotherapy—provides modest benefit, and overall clinical outcomes remain suboptimal, highlighting the need for alternative therapeutic strategies [3, 10].

Claudin18.2 (CLDN18.2) is a tight-junction protein physiologically confined to the apical membrane of differentiated gastric glandular epithelium. Loss of polarity and cellular dedifferentiation in malignancy can expose this epitope, rendering it targetable [11]. In gastrointestinal cancers, phase III clinical trials (SPOTLIGHT and GLOW) have demonstrated that adding zolbetuximab, a monoclonal antibody against CLDN18.2, to platinum–fluoropyrimidine chemotherapy significantly improves progression-free and overall survival in CLDN18.2-positive, HER2-negative advanced gastric or gastro-oesophageal junction adenocarcinoma [12–14]. The standard threshold for positivity in these trials was defined as ≥ 75% of tumor cells showing moderate-to-strong membranous staining. Importantly, treatment did not compromise health-related quality of life. In the lung, CLDN18.2 expression is enriched in IMA. A recent cohort reported a positivity rate of 76.2%, with nearly half of the cases meeting the ≥ 75% staining threshold, while multi-tumor mucinous series have confirmed its frequent expression and potential translational relevance in pulmonary mucinous tumors [15, 16].

Against this background, we aimed to characterise the prevalence and staining patterns of CLDN18.2 (high vs. low) in a resected cohort of 117 IMAs; to examine its concordance with gastric-type mucin markers (MUC5AC and MUC6); to assess associations with key clinicopathological features, including tumor stage, lymph-node status, STAS, and PD-L1 expression; and to explore correlations with overall survival. Defining the clinicopathological and diagnostic significance of CLDN18.2 in pulmonary IMA is clinically relevant, as PD-L1 negativity is common and therapeutic options are limited, providing a rationale for biomarker-based stratification and potential implementation of CLDN18.2-directed therapy in this subtype.

Materials and methods

Case selection and clinicopathological data

A consecutive series of 117 patients who underwent curative-intent resection for invasive mucinous adenocarcinoma (IMA) of the lung at the Affiliated Hospital of Nantong University between 2018 and 2020 were retrospectively reviewed. Diagnosis of IMA was confirmed by two experienced pulmonary pathologists according to the 2021 WHO Classification of Thoracic Tumors [17].

Inclusion criteria were: (i) availability of adequate formalin-fixed, paraffin-embedded (FFPE) tumor tissue; (ii) no history of neoadjuvant radiotherapy, chemotherapy, targeted therapy, or immunotherapy; and (iii) complete clinicopathological and follow-up information retrievable from the institutional electronic record system. Exclusion criteria included: (i) insufficient tumor tissue or severe artefacts precluding evaluation; and (ii) concomitant primary malignancy of indeterminate origin.

Clinicopathological parameters recorded comprised age, sex, maximal tumor diameter, AJCC 8th edition pathological stage, visceral pleural invasion, lymph-node metastasis, perineural and vascular invasion, spread through air spaces (STAS), and survival outcome. STAS was defined as the presence of tumor cells within air spaces in the lung parenchyma beyond the edge of the main tumor. Morphologically, STAS was identified on hematoxylin–eosin-stained sections as micropapillary clusters, solid nests, or single tumor cells floating within alveolar spaces without direct connection to the main tumor [18, 19].

The study protocol was reviewed and approved by the Institutional Review Board of the Affiliated Hospital of Nantong University (Approval No. 2022-L078) and conducted in accordance with the Declaration of Helsinki (as revised in 2013). Owing to the retrospective design and anonymised data, the requirement for written informed consent was waived.

Histology and immunohistochemistry

Tissue samples were fixed in 10% neutral-buffered formalin, routinely processed, embedded in paraffin, and sectioned at 4 μm for haematoxylin–eosin (H&E) staining. Immunohistochemistry (IHC) was carried out using an EnVision detection system (Dako, Agilent Technologies) with 3,3′-diaminobenzidine (DAB) as the chromogen. Primary antibodies were Claudin18.2 (clone 43–14 A; Roche/Ventana), MUC5AC and MUC6 (ready-to-use; Fuzhou Maixin), and PD-L1 (clone 22C3; Agilent Technologies). All staining procedures were performed on Leica automated stainers and a Ventana BenchMark Ultra platform under standardised laboratory conditions.

Scoring criteria

CLDN18.2: Staining was considered positive when ≥ 10% of tumor cells showed ≥ 1 + membranous staining, or any proportion exhibited ≥ 2 + membranous staining [20]. High expression was defined as ≥ 75% of tumor cells displaying ≥ 2 + membranous staining [15], while low expression referred to positive cases that did not reach the threshold for high expression. Specifically, membranous staining intensity was graded as 1+ (weak), 2+ (moderate), and 3+ (strong).

MUC5AC and MUC6: Immunoreactivity was assessed semi-quantitatively, adapting commonly used scoring approaches for gastric/gastrointestinal mucin markers. Staining intensity was graded as 1+ (weak), 2+ (moderate), and 3+ (strong) according to the cytoplasmic staining of tumor cells, and the proportion of stained tumor cells was recorded. Positive expression was defined as ≥ 10% of tumor cells showing 2 + or 3 + cytoplasmic staining, with or without apical–lateral membrane accentuation. Cases with no staining, only 1 + staining, or staining in < 10% of tumor cells were regarded as negative [21–23].

PD-L1 (clone 22C3): Expression was evaluated using the tumor proportion score (TPS), calculated as (membrane-positive tumor cells / all evaluable tumor cells) × 100%. A TPS ≥ 1% was regarded as positive [24]. Appropriate positive controls were included for PD-L1 immunohistochemistry.

All stained slides were reviewed by experienced pathologists according to predefined criteria, and discrepant cases were re-evaluated by joint review to reach a consensus.

Follow-up

Follow-up information was obtained from the institutional electronic medical record system and confirmed by telephone interviews. The median follow-up duration was 32 months (range, 3–70 months). All 117 patients were successfully followed, with no loss to follow-up. Overall survival was defined as the interval from the date of surgery to death from any cause or the last follow-up.

Statistical analysis

Statistical analyses were performed using R software (version 4.4.1; R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were expressed as mean ± standard deviation (SD) or median (interquartile range, IQR), and categorical variables as frequencies and percentages. Group comparisons were made using the χ² test or Fisher’s exact test for categorical variables and the t-test or Mann–Whitney U test for continuous variables, as appropriate. Correlations between CLDN18.2 and MUC5AC/MUC6 expression were evaluated using Spearman’s rank correlation (ρ, two-sided P values). Overall survival was analysed by the Kaplan–Meier method and compared using the log-rank test. All tests were two-sided, and P < 0.05 was considered statistically significant.

Results

Expression of CLDN18.2, MUC5AC, MUC6 and PD-L1 in the lung mucinous adenocarcinoma

A consecutive series of 117 patients who underwent curative-intent resection for invasive mucinous adenocarcinoma (IMA) of the lung at the Affiliated Hospital of Nantong University between 2018 and 2020 was retrospectively reviewed. The diagnosis of IMA was confirmed by two experienced pulmonary pathologists according to the 2021 WHO Classification of Thoracic Tumors. Expression of CLDN18.2, MUC5AC, and MUC6 was evaluated by immunohistochemistry. CLDN18.2 staining was considered positive when ≥ 10% of tumor cells exhibited membranous staining of ≥ 1 + intensity, or when any proportion of tumor cells demonstrated ≥ 2 + membranous staining; high expression was defined as ≥ 75% of tumor cells showing membranous staining of ≥ 2 + intensity. MUC5AC and MUC6 expression was defined as positive when ≥ 10% of tumor cells displayed 2 + or 3 + cytoplasmic staining, with or without apical–lateral membrane accentuation. Among the 117 cases of pulmonary mucinous adenocarcinoma, CLDN18.2 positivity was observed in 71 cases (60.7%), with high expression detected in 42 cases (35.9%), MUC5AC positivity in 88 cases (75.2%), and MUC6 positivity in 40 cases (34.2%). PD-L1 immunohistochemical staining was also assessed in this cohort. All 117 cases were negative for PD-L1 expression, with TPS < 1% in all tumors. Accordingly, no PD-L1-positive case was identified in this cohort.

Correlation between the expression of CLDN18.2, MUC5AC, and MUC6 and clinicopathological parameters in lung mucinous adenocarcinoma

The expression of CLDN18.2, MUC5AC, and MUC6 showed no significant associations with most clinicopathological features. Regarding sex distribution, no significant differences were observed between CLDN18.2 or MUC5AC expression groups, whereas MUC6 positivity was significantly more frequent in female patients compared with the negative group (70% vs. 50.6%, P = 0.045). No significant differences were detected in age, tumor size, stage, visceral pleural invasion, lymph node metastasis, perineural invasion, vascular invasion, or survival status (P > 0.05). Borderline associations were observed between MUC6 expression and tumor size as well as stage; however, these findings did not reach definitive statistical significance. Notably, the incidence of spread through air spaces (STAS) was significantly lower in the CLDN18.2-positive group (40.8% vs. 60.9%, P = 0.034) and in the MUC5AC-positive group (42% vs. 69%, P = 0.012), suggesting that CLDN18.2 and MUC5AC positivity may be associated with reduced tumor aerogenous spread (Table 1).

Table 1.

Correlation between the expression of CLDN18.2, MUC5AC, and MUC6 and clinicopathological parameters in lung mucinous adenocarcinoma

Variables	Total (n = 117)	CLDN18.2 Negative (n = 46)	CLDN18.2 Positive (n = 71)	P	MUC5AC Negative (n = 29)	MUC5AC Positive (n = 88)	P	MUC6 Negative (n = 77)	MUC6 Positive (n = 40)	P
Sex, n (%)				0.201			0.793			0.045*
Female	67 (57.3)	23 (50)	44 (62)		16 (55.2)	51 (58)		39 (50.6)	28 (70)
Male	50 (42.7)	23 (50)	27 (38)		13 (44.8)	37 (42)		38 (49.4)	12 (30)
Age, n (%)				0.377			0.175			0.77
< 60	36 (30.8)	12 (26.1)	24 (33.8)		6 (20.7)	30 (34.1)		23 (29.9)	13 (32.5)
≥ 60	81 (69.2)	34 (73.9)	47 (66.2)		23 (79.3)	58 (65.9)		54 (70.1)	27 (67.5)
Tumor size, n (%)				0.137			0.969			0.055
< 3 cm	77 (65.8)	34 (73.9)	43 (60.6)		19 (65.5)	58 (65.9)		46 (59.7)	31 (77.5)
≥ 3 cm	40 (34.2)	12 (26.1)	28 (39.4)		10 (34.5)	30 (34.1)		31 (40.3)	9 (22.5)
Stage, n (%)				0.101			0.26			0.083
I	90 (76.9)	39 (84.8)	51 (71.8)		22 (75.9)	68 (77.3)		55 (71.4)	35 (87.5)
II	21 (17.9)	7 (15.2)	14 (19.7)		7 (24.1)	14 (15.9)		18 (23.4)	3 (7.5)
III	6 ( 5.1)	0 (0)	6 (8.5)		0 (0)	6 (6.8)		4 (5.2)	2 (5)
Visceral pleural invasion (VPI), n (%)				0.377			0.301			0.24
No	89 (76.1)	33 (71.7)	56 (78.9)		20 (69)	69 (78.4)		56 (72.7)	33 (82.5)
Yes	28 (23.9)	13 (28.3)	15 (21.1)		9 (31)	19 (21.6)		21 (27.3)	7 (17.5)
Lymph node metastasis, n (%)				0.873			0.54			0.161
No	101 (86.3)	40 (87)	61 (85.9)		24 (82.8)	77 (87.5)		64 (83.1)	37 (92.5)
Yes	16 (13.7)	6 (13)	10 (14.1)		5 (17.2)	11 (12.5)		13 (16.9)	3 (7.5)
Spread through air spaces, n (%)				0.034*			0.012*			0.174
No	60 (51.3)	18 (39.1)	42 (59.2)		9 (31)	51 (58)		36 (46.8)	24 (60)
Yes	57 (48.7)	28 (60.9)	29 (40.8)		20 (69)	37 (42)		41 (53.2)	16 (40)
Perineural invasion (PNI), n (%)				0.393			1			1
No	116 (99.1)	45 (97.8)	71 (100)		29 (100)	87 (98.9)		76 (98.7)	40 (100)
Yes	1 ( 0.9)	1 (2.2)	0 (0)		0 (0)	1 (1.1)		1 (1.3)	0 (0)
Vascular invasion, n (%)				0.628			0.77			0.533
No	99 (84.6)	38 (82.6)	61 (85.9)		24 (82.8)	75 (85.2)		64 (83.1)	35 (87.5)
Yes	18 (15.4)	8 (17.4)	10 (14.1)		5 (17.2)	13 (14.8)		13 (16.9)	5 (12.5)
Survival status, n (%)				0.865			0.234			0.316
Dead	17 (14.5)	7 (15.2)	10 (14.1)		2 (6.9)	15 (17)		13 (16.9)	4 (10)
Alive	100 (85.5)	39 (84.8)	61 (85.9)		27 (93.1)	73 (83)		64 (83.1)	36 (90)

*Indicates a statistically significant difference at p < 0.05

Clinicopathological features of patients with different CLDN18.2 expression levels

The clinicopathologic characteristics of lung mucinous adenocarcinomas with high CLDN18.2 expression were compared with those showing low expression. No significant differences were observed in sex, age, tumor size, stage, visceral pleural invasion, STAS, or survival status (P > 0.05). However, a higher proportion of stage I cases was observed in the high-expression group compared with the low-expression group (81% vs. 68%), while the frequency of lymph node metastasis was lower (4.8% vs. 18%). Although these differences did not reach statistical significance, a trend toward earlier tumor stage and fewer nodal metastases was observed in the high-expression group, which may in part reflect the limited sample size of the present cohort (Table 2). Immunohistochemical analysis revealed considerable inter-tumoral variation in CLDN18.2 expression, with staining intensity and extent ranging from negative to low to high across different cases (Fig. 1). Moreover, MUC5AC and MUC6 expression levels progressively increased with higher CLDN18.2 expression, whereas PD-L1 expression remained generally low and showed no significant variation among different CLDN18.2 groups (Fig. 2).

Table 2.

Correlation between CLDN18.2 expression levels and clinicopathological parameters in lung mucinous adenocarcinoma

Variables	Total (n = 71)	CLDN18.2 Low (n = 50)	CLDN18.2 High (n = 21)	P
Sex, n (%)				0.281
Female	44 (62.0)	33 (66)	11 (52.4)
Male	27 (38.0)	17 (34)	10 (47.6)
Age, n (%)				0.62
< 60	24 (33.8)	16 (32)	8 (38.1)
≥ 60	47 (66.2)	34 (68)	13 (61.9)
Tumor size, n (%)				0.495
< 3 cm	43 (60.6)	29 (58)	14 (66.7)
≥ 3 cm	28 (39.4)	21 (42)	7 (33.3)
Stage, n (%)				0.375
I	51 (71.8)	34 (68)	17 (81)
II	14 (19.7)	12 (24)	2 (9.5)
III	6 ( 8.5)	4 (8)	2 (9.5)
Visceral pleural invasion (VPI), n (%)				0.527
No	56 (78.9)	38 (76)	18 (85.7)
Yes	15 (21.1)	12 (24)	3 (14.3)
Lymph node metastasis, n (%)				0.262
No	61 (85.9)	41 (82)	20 (95.2)
Yes	10 (14.1)	9 (18)	1 (4.8)
Spread through air spaces, n (%)				0.404
No	42 (59.2)	28 (56)	14 (66.7)
Yes	29 (40.8)	22 (44)	7 (33.3)
Perineural invasion (PNI), n (%)
No	71 (100.0)	50 (100)	21 (100)	1
Yes
Vascular invasion, n (%)				0.262
No	61 (85.9)	41 (82)	20 (95.2)
Yes	10 (14.1)	9 (18)	1 (4.8)
Survival status, n (%)				0.712
Dead	10 (14.1)	8 (16)	2 (9.5)
Alive	61 (85.9)	42 (84)	19 (90.5)

Fig. 1.

Representative CLDN18.2 expression patterns in pulmonary invasive mucinous adenocarcinoma (IMA). A-C CLDN18.2-negative group: no membranous CLDN18.2 staining in tumor cells (A, H&E, ×100; B, IHC, ×100; C, IHC, ×400). D-F CLDN18.2 low-expression group: weak-to-moderate membranous staining present in 1%–74% of tumor cells (D, H&E, ×100; E, IHC, ×100; F, IHC, ×400). G-I CLDN18.2 high-expression group: diffuse moderate-to-strong membranous staining in ≥ 75% of tumor cells (G, H&E, ×100; H, IHC, ×100; I, IHC, ×400). In the whole cohort (n = 117), 46 cases were classified as CLDN18.2-negative and 71 as CLDN18.2-positive, including 50 low-expression and 21 high-expression cases. Representative images are shown

Fig. 2.

Representative immunohistochemical staining of MUC5AC, MUC6, and PD-L1 according to CLDN18.2 expression levels in pulmonary invasive mucinous adenocarcinoma (IMA). A-C CLDN18.2-negative group: weak MUC5AC staining in a few glandular cells, with both MUC6 and PD-L1 negative (IHC, ×200). D-F CLDN18.2 low-expression group: focal weak-to-moderate MUC5AC staining, with MUC6 and PD-L1 remaining negative (IHC, ×200). G-I CLDN18.2 high-expression group: diffuse moderate-to-strong MUC5AC staining in most tumor cells and partial positivity for MUC6, whereas PD-L1 remained negative (IHC, ×200). The corresponding group sizes were as follows: CLDN18.2-negative, n = 46; CLDN18.2 low-expression, n = 50; and CLDN18.2 high-expression, n = 21. Representative images are shown

Correlation between CLDN18.2 and gastric-type mucin markers

Correlation analysis revealed that CLDN18.2 expression was significantly and positively correlated with MUC5AC (ρ = 0.469, P < 0.0001) and MUC6 (ρ = 0.43, P < 0.0001). Specifically, MUC5AC positivity was observed in 90.1% of CLDN18.2-positive cases compared with only 9.9% in CLDN18.2-negative cases, while MUC6 positivity was significantly higher in CLDN18.2-positive than in negative cases (52.1% vs. 6.5%). These results indicate that CLDN18.2 positivity is closely associated with gastric-type mucin phenotypes defined by MUC5AC and MUC6 expression (Table 3).

Table 3.

Correlation analysis between CLDN18.2 expression and MUC5AC/MUC6 in lung mucinous adenocarcinoma

Variables	Total (n = 117)	CLDN18.2 Negative (n = 46)	CLDN18.2 Positive (n = 71)	Spearman’s coefficient of rank correlation (rho)	P
MUC5AC, n (%)				0.469	< 0.0001
Negative	29 (24.8)	22 (47.8)	7 (9.9)
Positive	88 (75.2)	24 (52.2)	64 (90.1)
MUC6, n (%)				0.43	< 0.0001
Negative	77 (65.8)	43 (93.5)	34 (47.9)
Positive	40 (34.2)	3 (6.5)	37 (52.1)

Survival analysis

Kaplan-Meier survival analysis demonstrated that neither CLDN18.2 positivity nor high expression was significantly associated with overall survival (OS) (P = 0.42 and P = 0.62, respectively). Patients with MUC5AC-positive tumors showed a trend toward poorer OS compared with MUC5AC-negative patients, although the difference was not statistically significant (P = 0.19). In contrast, MUC6 positivity was associated with a more pronounced decline in OS, with borderline significance (P = 0.088), suggesting that MUC6 expression may serve as a potential marker of unfavorable prognosis (Fig. 3).

Fig. 3.

Association of CLDN18.2, MUC5AC, and MUC6 expression with overall survival in pulmonary invasive mucinous adenocarcinoma (IMA) (Kaplan-Meier analysis). A Overall survival according to CLDN18.2 expression status: CLDN18.2-positive (n = 71) versus CLDN18.2-negative tumors (n = 46); no significant difference was observed (P = 0.42). B Overall survival according to MUC5AC expression status: MUC5AC-positive (n = 88) versus MUC5AC-negative tumors (n = 29); a trend toward poorer survival was observed in the positive group, but the difference was not statistically significant (P = 0.19). C Overall survival according to MUC6 expression status: MUC6-positive (n = 40) versus MUC6-negative tumors (n = 77); reduced overall survival was observed in the positive group, with borderline statistical significance (P = 0.088). D Overall survival according to CLDN18.2 expression level among CLDN18.2-positive cases: high expression (n = 21) versus low expression (n = 50); no significant difference was observed (P = 0.62). Survival curves were generated using the Kaplan–Meier method and compared using the log-rank test

In summary, CLDN18.2 expression in lung mucinous adenocarcinoma is strongly associated with gastric-type mucin markers MUC5AC and MUC6. CLDN18.2 positivity correlates with lower STAS incidence and trends toward earlier stage and fewer nodal metastases. All cases were PD-L1 negative, supporting the immune-cold nature of this subtype. Although CLDN18.2 itself does not significantly affect patient survival, MUC6 expression may serve as a potential indicator of adverse prognosis in this distinctive histological variant.

Discussion

In this postoperative cohort of 117 patients with pulmonary invasive mucinous adenocarcinoma (IMA), three principal observations were made. First, CLDN18.2 expression was frequent (overall positivity 60.7%; high expression 35.9%) and exhibited a strong association with the gastric-type mucin phenotype, showing moderate correlations with MUC5AC and MUC6 expression. Second, tumors positive for CLDN18.2 or MUC5AC showed a lower frequency of spread through air spaces (STAS). Third, all cases were negative for PD-L1 (22C3-TPS), indicating an immune-cold profile. Collectively, these findings suggest that CLDN18.2 serves primarily as a lineage and potentially translational biomarker rather than a prognostic marker in pulmonary IMA.

Molecular co-expression and gastric-type differentiation

CLDN18.2 expression closely paralleled that of MUC5AC and MUC6, delineating a gastric-type differentiation spectrum within pulmonary IMA. The progressive increase in expression intensity and the moderate positive correlations (ρ = 0.469 and 0.43, p < 0.0001) support the presence of a gastric-type differentiation programme, consistent with previous studies [22, 23, 25]. However, reported inter-study variability in CLDN18.2 expression highlights the need for harmonised immunohistochemical protocols and standardised interpretation thresholds to ensure inter-laboratory comparability.

Relation to existing classification frameworks

Current classifications of lung adenocarcinoma predominantly focus on genetic alterations such as KRAS, EGFR, ALK, or NTRK [26]. Our findings complement these molecular frameworks by identifying a phenotypic subset defined by protein-level co-expression of CLDN18.2, MUC5AC, and MUC6. This subset is typically KRAS-mutant, TTF-1–negative, and PD-L1–low, resembling the “gastric-type” pattern seen in gastric adenocarcinoma. The analogy underscores a cross-organ continuum wherein mucin-associated pathways govern both morphologic and biological behavior [17].

Clinicopathologic behavior and STAS correlation

STAS is recognized as an adverse invasive pattern associated with postoperative recurrence, particularly following sublobar resection [18, 19, 27–30]. The reduced STAS frequency in CLDN18.2- and MUC5AC-positive tumors accords with the notion that gastric-type differentiation and preserved junctional integrity may impede aerogenous spread. Mechanistically, CLDN18.2-mediated tight junctions and the accompanying mucinous matrix may confer structural resistance to detachment. Functional studies using cell-based or organoid models are required to verify this hypothesis.

Immune landscape and therapeutic implications

Uniform PD-L1 negativity in this cohort reinforces the immune-cold phenotype of IMA, in line with prior reports showing PD-L1 TPS ≥ 1% in only 6.1% of IMA compared with 59.7% of non-mucinous adenocarcinomas [31]. Multi-omic data further support a KRAS-driven biology with low immune infiltration, underscoring the need for non-PD-L1-dependent therapeutic targets [32, 33]. CLDN18.2 may represent a biomarker of potential translational interest in this regard. At present, no lung-specific enrollment criterion has been established for CLDN18.2-targeted therapy in pulmonary IMA; therefore, the gastric/gastroesophageal benchmark of moderate-to-strong membranous staining in ≥ 75% of tumor cells may serve as a practical reference, although its clinical applicability in lung IMA remains to be validated. In the FAST trial for gastric/gastroesophageal adenocarcinoma, zolbetuximab an anti-CLDN18.2 monoclonal antibody—combined with chemotherapy improved both progression-free and overall survival in CLDN18.2-positive tumors [34]. Applying a similar threshold (≥ 75% tumor cells with ≥ 2 + membranous staining), 35.9% of our IMA cases fell within a range that may be of interest for future biomarker stratification studies [35]. Co-expression with MUC5AC and MUC6 may further delineate a biologically homogeneous PD-L1–negative subset.

Translational and clinical relevance

From a diagnostic perspective, an immunohistochemical panel incorporating CLDN18.2, MUC5AC, and MUC6 is readily implementable in routine surgical pathology. Such a panel facilitates lineage attribution in small biopsies and aids identification of clinically meaningful subgroups. The CLDN18.2-positive, gastric-type IMA subset typically exhibits low PD-L1 expression, reduced STAS, and an immune-cold microenvironment—features that may inform both diagnostic reporting and future biomarker stratification. Future investigations integrating spatial transcriptomics and multiplex immunohistochemistry are warranted to delineate the architectural and immune context of this phenotype and to validate the reproducibility of CLDN18.2 scoring algorithms.

Strengths and limitations

The strengths of this study include the use of a uniformly resected cohort, a trial-aligned CLDN18.2 scoring system, and systematic assessment of STAS and clinical outcomes. However, several limitations should be acknowledged. First, this was a retrospective single-center study, and the genotyping data were incomplete, particularly with regard to detailed KRAS allele information, which limited further molecular subclassification. Second, although PD-L1 was evaluated in all cases and all tumors were classified as negative according to the predefined 1% cutoff, TPS values below 1% were not further stratified. In addition, the lack of broader immune profiling beyond PD-L1 also limited a more comprehensive characterization of the tumor immune microenvironment. Furthermore, because only surgically resected cases were included, without advanced-stage disease, treatment response data, or functional validation, the therapeutic relevance of CLDN18.2 expression in pulmonary invasive mucinous adenocarcinoma should be interpreted with caution. Finally, immunohistochemical interpretation is inherently semi-quantitative, and some degree of interobserver variability is unavoidable, particularly in cases with heterogeneous or borderline staining; the lack of formal interobserver agreement analysis therefore represents another limitation of the present study. Future multicenter prospective studies integrating genomic, spatial, and clinical treatment data are warranted to determine whether high CLDN18.2 expression defines a biologically distinct and potentially clinically relevant subset of pulmonary IMA.

Conclusions

This study demonstrates that CLDN18.2 is frequently expressed in pulmonary invasive mucinous adenocarcinoma (IMA) and strongly associated with gastric-type mucin differentiation markers (MUC5AC, MUC6). CLDN18.2-positive tumors showed lower STAS frequency. All cases were PD-L1 negative, highlighting an immune-cold microenvironment. Although not prognostic, CLDN18.2 identifies a distinct phenotypic subset that may have potential translational relevance as a biomarker in pulmonary IMA.

These findings provide a basis for further studies to clarify the biological significance and possible therapeutic relevance of CLDN18.2 in clinically and molecularly well-characterized cohorts.

Abbreviations

CLDN18.2

Claudin18.2

IMA

Invasive Mucinous Adenocarcinoma

PD-L1

Programmed Death-Ligand 1

STAS

Spread Through Air Spaces

AJCC

American Joint Committee on Cancer

FFPE

Formalin-Fixed, Paraffin-Embedded

H&E

Haematoxylin and Eosin

IHC

Immunohistochemistry

DAB

3,3′-diaminobenzidine

TPS

Tumor Proportion Score

OS

Overall Survival

KRAS

Kirsten rat sarcoma viral oncogene homolog

EGFR

Epidermal Growth Factor Receptor

ALK

Anaplastic Lymphoma Kinase

NTRK

Neurotrophic Receptor Tyrosine Kinase

TTF-1

Thyroid Transcription Factor-1

HER2

Human Epidermal Growth Factor Receptor 2

Authors’ contributions

Yifei Liu conceived and designed the study. Tingting Bian and Xiaoyu Ren drafted the manuscript. Daishan Jiang, Jianguo Zhang, Sheng Xiao and Yuan Li participated in the experimental design and conception. All authors reviewed and approved the final manuscript.

Funding

This study was supported by the Jiangsu Provincial Research Hospital (YJXYY202204-YSC01, YJXYY202204-YSB01), Nantong Basic Research Plan Project (MS2023067), and National Natural Science Foundation of China (No. 82273422).

Data availability

Data will be available upon request to the corresponding author.

Declarations

Ethics approval and consent to participate

The study protocol was reviewed and approved by the Institutional Review Board of the Affiliated Hospital of Nantong University (Approval No. 2022-L078) and conducted in accordance with the Declaration of Helsinki (as revised in 2013). Owing to the retrospective design and anonymised data, the requirement for written informed consent was waived.

Consent for publication

This manuscript is original, has not been published elsewhere, and is not under consideration by any other publication. All authors have read and approved the final version of the manuscript and agree with its submission to Diagnostic Pathology.

Competing interests

The authors declare no competing interests.