Differences in tumor angiogenesis and related factors between lung adenocarcinomas manifesting as pure ground glass opacity and solid nodules

Abstract

Introduction

The prognosis of ground glass opacity featured lung adenocarcinomas (GGO-LUAD) is significantly better than that of solid nodule featured lung adenocarcinomas (SN-LUAD), but the specific reasons behind their indolent tumor behavior are still unclear. The purpose of this study is to investigate their differences in intratumoral microvessels, related angiogenic factors and important stromal cells.

Methods

Thirty patients (15 paired patients only with GGO or SN) diagnosed with pathological stage 0-I lung adenocarcinoma who underwent surgical treatment were included into this study. Immunohistochemistry was performed to stain the blood vessel markers (CD31, CD34 and CD105), LYVE-1, the cancer-associated fibroblasts (CAFs) markers (α-SMA and S100A4), TGF-β and HIF-1α from 30 patients tissue sections. At the same time, Ki67 Labeling Index (LI) extracted from pathological report of all patients was also analyzed.

Results

GGO-LUAD is more abundant than SN-LUAD in lymphatic vessel density (LVD), but similar in total microvessel density (CD31 + MVD). However, GGO-LUAD is significantly lower than SN-LUAD in CD34 + MVD and CD105 + MVD. In terms of TGF-β, HIF-1α expression and Ki67 LI level, GGO-LUAD was also significantly weaker than SN-LUAD. Moreover, the distribution of CAFs in GGO-LUAD is less than that in SN-LUAD. Regardless of the pathological type (adenocarcinoma in situ (AIS) or minimally invasive adenocarcinoma (MIA) or invasive lung adenocarcinoma (IAC)), there is no difference in any of the above indicators in GGO-LUAD.

Conclusions

Our finding displays that GGO-LUAD was significantly lower than SN-LUAD in CD34 + MVD and CD105 + MVD reflecting tumor angiogenesis, and the distribution of CAFs and factors related to tumor angiogenesis were also significantly lower in GGO-LUAD, which may indicate that the weak ability of angiogenesis might be the reason for the good prognosis of GGO-LUAD.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12672-025-01898-5.

Introduction

Lung cancer currently remains the leading cause of cancer-related mortality in the world [1]. In order to decrease mortalities, numerous broad-based lung cancer screening programs have been implemented globally [2–5]. These programs have helped in identifying many cases of lung cancer at an early stage, mainly early-stage lung adenocarcinomas manifesting predominantly as ground-glass opacities. Several studies [6–9] demonstrate that ground glass opacity featured lung adenocarcinomas (GGO-LUAD) have a better prognosis when compared to those solid nodules featured lung adenocarcinomas (SN-LUAD). However, the reasons explaining why GGO-LUAD have a better prognosis as compared to SN-LUAD still remain unknown.

Angiogenesis is necessary for cell survival, growth and metastasis of the tumor [10–12]. Tumor microvessel density is the most commonly used indicator of tumor angiogenesis. Many studies have shown that high tumor microvessel density is closely associated with poor prognosis in a variety of solid tumors, including lung cancer [13, 14], breast cancer [15], prostate cancer [16] and kidney cancer [17], etc. Therefore, we hypothesize that the microvessel distribution of GGO-LUAD may be significantly different from that of SN-LUAD. To date, there is no study on the microvessel distribution of early-stage lung adenocarcinoma, especially GGO-LUAD.

Cancer-associated fibroblasts (CAFs) are important components of solid tumor stroma. In recent years, numerous investigations have focused on the role of CAFs in tumor microenvironment. Accumulated evidence [18–22] displays that CAFs play an important role in tumor progression and metastasis, including tumor cell growth, tumor angiogenesis and extracellular matrix remodeling. However, the distribution of CAFs in early-stage lung adenocarcinoma, especially GGO-LUAD, are also poorly understood.

In the present study, we mainly analyzed the distribution of tumor vasculature (tumor microvessel and lymphatic vessel density) of GGO-LUAD and SN-LUAD by immunohistochemistry (IHC), and also evaluated the important stroma cells-CAFs in the tumor microenvironment, hoping to understand the reasons behind their indolent behavior.

Materials and methods

Patient selection

Thirty patients (15 pairs of patients with either pure GGO or SN) who were diagnosed with stage 0-I lung adenocarcinoma between January 2021 and December 2021 were included in this study. All patients underwent surgical resection, and none received neoadjuvant therapy prior to surgery. The patient selection process and study design are detailed in Fig. 1. Ethical approval for the study was obtained from the Institutional Review Board of Tongji Medical College, Huazhong University of Science and Technology, and written informed consent was provided by all participants. The study was conducted in compliance with the Declaration of Helsinki (revised 2013).

Fig. 1.

Flowchart of patient selection and study design

Radiological and histological evaluation

Lesions were assessed through high-resolution computed tomography (HRCT) scans. All chest CT images were obtained during full inspiration and reviewed retrospectively, with a focus on detecting GGO. Tumor size was measured as the maximum axial diameter of the nodule in the lung window. Consolidation was defined as a uniform increase in pulmonary parenchymal density that obscures airway and vessel walls, while GGO was identified by a hazy increase in lung opacity that preserves bronchial and vascular structures. Two experienced thoracic surgeons (Y.C. and X.F.) independently and blindly examined the preoperative CT scans.

Two independent pathologists (S.Q. and J.X.), unaware of the patients’ clinical outcomes, examined hematoxylin and eosin (H&E)-stained tissue slides. Tumor histological subtypes were determined using the classification system from the International Association for the Study of Lung Cancer (IASLC), American Thoracic Society (ATS), and European Respiratory Society (ERS) for lung adenocarcinomas, categorizing them as AIS, MIA, or IAC. The staging of the disease was determined based on the 8th edition of the American Joint Committee on Cancer (AJCC) TNM Staging Manual [24].

Immunohistochemistry

Formalin-fixed, paraffin-embedded (FFPE) tissue samples were used for consecutive sectioning. IHC was conducted as previously outlined [25]. Briefly, 4 μm sections were deparaffinized, rehydrated, and underwent microwave antigen retrieval using EDTA buffer at pH 9.0, except for CD34 and HIF-1α staining, which used citrate buffer at pH 6.0. Endogenous peroxidase activity was blocked using 3% hydrogen peroxide for 15 min, followed by a 30-min block with 5% bovine serum albumin to prevent non-specific binding. The slides were incubated with the primary antibodies at 4 °C overnight. Afterward, the sections were treated with the secondary antibodies at room temperature for 30 min. The immunoreactivity was visualized using the DAB method (GK600505, GeneTech, Shanghai, China). Slides lacking primary antibodies served as negative controls.

The following antibodies were employed for the IHC analysis: CD31 (Ab76533, 1:400 dilution; Abcam, MA, USA), CD34 (Ab110643, 1:1000 dilution; Abcam, MA, USA), CD105 (Ab169545, 1:2000 dilution; Abcam, MA, USA), LYVE-1 (Ab219556, 1:20000 dilution; Abcam, MA, USA), HIF-1α (Ab8366, 1:250 dilution; Abcam, MA, USA), α-SMA (1A4, pre-prepared manufacturer dilution; Dako, Denmark), FSP1 (polyclonal, pre-prepared manufacturer dilution; Dako, Denmark), and TGF-β (21898-1-AP, 1:250 dilution; Proteintech, China). The detailed procedure was performed according to the manufacturer’s instructions.

Immunohistochemical analysis and scoring of molecular markers

The H-score method was applied to quantify the expression of α-SMA in stromal cells and TGF-β and HIF-1α in tumor cells. The H-score was calculated by multiplying the percentage of stained cells by an intensity score, as previously outlined [26]. Specifically, the staining intensity was graded as follows: 1 + for weak, 2 + for moderate, and 3 + for strong staining. The percentage of positive cells was recorded based on the stained area within the tumor, ranging from 0 to 100. The H-score was determined as the product of intensity and percentage, with a possible range of 0 to 300. Three independent evaluators (R.Q., along with pathologists S.Q. and J.X.) assessed the slides without knowledge of patient outcomes.

Intratumoral microvessel density (MVD, CD31, CD34 and CD105) was assessed according to the criteria described previously, with some modifications [12, 27]. Briefly, the three most vascularized areas (hotspots) within each sections were selected for quantification of blood vessels at a magnification of × 100 (a lower power microscope) with a Carl Zeiss microscope (Baden-Württemberg, Germany). Subsequently, three views were selected from hotspots at a magnification of × 400 (a higher power microscope). Therefore, a total of nine hotspots was evaluated from the sections for each patient. Any brown-staining endothelial cell or endothelial cell cluster that clearly separated from adjacent microvessels, tumor cells, and connective elements, but not single endothelial cells was considered as a single, countable microvessel regardless of whether a vessel lumen was seen. The average from nine high-power fields was calculated as the final MVD value for each patient. For intratumoral lymphatic vessel density (LVD) analysis (LYVE-1), the final LVD value was also counted by the above method.

To assess positive cells (fibroblast-specific protein 1, FSP1, also called S100A4 + cells), tissue sections were examined under a low-power microscope (× 100), and three regions with the highest concentration of marker-positive cells were identified. From each of these regions, three random fields were chosen for closer examination under a higher magnification (× 400). Marker-positive cells within the intratumoral areas were counted across nine high-power fields. The average number of positive cells from these fields was used to determine the overall cell density.

The expression of Ki67 protein in the patient's tumor is routinely detected by the pathologist after surgery. The expression level of Ki67 protein can be represented by the Ki67 labeling index (LI), which refers to the percentage of Ki67-positive nuclei under the microscope in IHC tests. Ki67 LI of all patients included in this study was obtained from postoperative pathology reports.

Statistical analysis

Bar graph data are displayed as the mean with standard deviation (SD). For comparing continuous variables, we utilized Student’s unpaired t-test, the Mann–Whitney U test, or One-Way ANOVA. Categorical variables were analyzed using the chi-square (χ2) test or Fisher’s exact test. All P-values were two-tailed, with a threshold of P < 0.05 indicating statistical significance (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001). The statistical analyses were conducted using SPSS version 23.0 (SPSS, Chicago, IL, USA) and GraphPad Prism 7.0 (GraphPad Software, Inc.).

Results

Patient and tumor characteristics

Between January and December 2021, FFPE samples from 30 early-stage lung adenocarcinoma patients (15 with pure GGO-LUAD and 15 with SN-LUAD) were gathered for IHC analysis. The majority of the patients in both groups were female, and no significant differences were found in terms of gender, age, or family history. In the pure GGO-LUAD group, final pathology revealed 3 cases of AIS, 7 of MIA, and 5 of IAC, whereas all patients in the SN-LUAD group had IAC. Overall, the pure GGO-LUAD group was predominantly composed of pre-invasive or minimally invasive adenocarcinomas, indicating that GGO-LUAD tends to be a less aggressive form of the tumor. Detailed patient and tumor characteristics are listed in Table 1.

Table 1.

Clinical characteristics of paired Patients

Variables	pGGO (n = 15)	SN (n = 15)	P value
Age (Year ± SD)	58.3 ± 11.8	60.1 ± 7.7	0.363
Gender
Male	4	7	0.449
Female	11	8
Smoking status
No	14	6	0.005
Current or former	1	9
Family history
Yes	3	2	> 0.999
No	12	13
Tumor size (mean ± SD, mm)	13.6 ± 1.1	19.3 ± 1.4	0.003
Histology
AIS	3	0	< 0.001
MIA	7	0
IAC	5	15
pTNM stage
0	3	0	< 0.001
IA1	9	1
IA2	3	7
IA3	0	7
EGFR status
WT	6	5	> 0.999
Mutation	9	10

pGGO pure ground glass opacity, SN solid nodule, AIS adenocarcinoma in situ, MIA minimally invasive adenocarcinoma, IAC invasive adenocarcinoma, WT wild type

Distribution of intratumoral lymphatic vessels and microvessels in GGO-LUAD and SN-LUAD

We first performed a staining analysis of intratumoral lymphatic vessel marker, LYVE-1. The results showed that GGO-LUAD is more abundant than SN-LUAD in intratumoral LVD (Figure S1B). In different pathological subtype GGO-LUAD, there was no significant difference in intratumoral LVD (Figure S1C). Representative images of immunohistochemistry staining are shown in Figure S1A.

Next, we use vascular endothelial markers that label intratumoral microvessels in different states to analyze the intratumoral MVD. We found that there is no significant difference between GGO-LUAD and SN-LUAD in CD31 + MVD (CD31, a pan endothelial marker) (Fig. 2B). Interestingly, GGO-LUAD is significantly lower than SN-LUAD in CD34 + MVD (CD34, a differentiated endothelial marker) (Fig. 3B) and CD105 + MVD (CD105, a proliferation-related and activated endothelial marker) (Fig. 4B). However, in GGO-LUAD, there was no difference in intratumoral MVD between different pathological subtypes (Figs. 2C, 3C and 4C). Representative images of immunohistochemistry staining are shown in Figs. 2A, 3A, 4A and Figure S2.

Fig. 2.

Distribution of Intratumoral CD31 + microvessels in tumor specimens. A Representative immunohistochemistry images for distribution of CD31 + MVD in different radiological patterns and pathological subtypes. Black scale bar: 100 μm. Red scale bar: 25 μm. B CD31 + MVD were counted among GGO-LUAD and SN-LUAD. C CD31 + MVD were counted were counted among GGO-LUAD with different pathological subtypes

Fig. 3.

Distribution of Intratumoral CD34 + microvessels in tumor specimens. A Representative immunohistochemistry images for distribution of CD34 + MVD in different radiological patterns and pathological subtypes. Black scale bar: 100 μm. Red scale bar: 25 μm. B CD34 + MVD were counted among GGO-LUAD and SN-LUAD. (C) CD34 + MVD were counted were counted among GGO-LUAD with different pathological subtypes

Fig. 4.

Distribution of Intratumoral CD105 + microvessels in tumor specimens. A Representative immunohistochemistry images for distribution of CD105 + MVD in different radiological patterns and pathological subtypes. Black scale bar: 100 μm. Red scale bar: 25 μm. B CD105 + MVD were counted among GGO-LUAD and SN-LUAD. (C) CD105 + MVD were counted were counted among GGO-LUAD with different pathological subtypes

SN-LUAD is associated with severe hypoxia and higher tumor cell proliferation index

Hypoxia is a common characteristic of many cancers. The master regulator of tumor hypoxia, hypoxia-inducible factor-1α (HIF-1α), controls many aspects of tumorigenesis, including proliferation, metabolism, metastasis, differentiation, and response to radiation therapy, especially tumor angiogenesis. Therefore, we further analyzed the tumor hypoxia in two groups by HIF-1α staining via immunohistochemistry. The results showed that SN-LUAD was significantly higher than GGO-LUAD in the expression of HIF-1α (Fig. 5B), which was consistent with the fact that SN-LUAD had more abundant tumor neovascularization. However, in GGO-LUAD, there was no difference in the expression of HIF-1α between different pathological subtypes (Fig. 5C). Representative images of immunohistochemistry staining are shown in Fig. 5A.

Fig. 5.

Scoring of HIF-1α expression in tumor specimens. A Representative immunohistochemistry images for HIF-1α expression in different radiological patterns and pathological subtypes. Black scale bar: 100 μm. Red scale bar: 25 μm. B H-score of HIF-1α expression were analyzed among GGO-LUAD and SN-LUAD. C H-score of HIF-1α expression were analyzed among GGO-LUAD with different pathological subtypes

At the same time, we also found that the Ki67 LI of SN-LUAD was significantly higher than that of GGO-UAD (Figure S3A). However, as shown in Figure S3B, no matter what pathological subtype, as long as the radiologic manifestation is GGO, there is no significant difference in Ki67 LI. The above results show that GGO-LUAD has extremely significant indolent biological behavior.

The distribution of CAFs in GGO-LUAD is less than that in SN-LUAD

Accumulating evidence from human cancer studies [28–30] suggests that CAFs are closely related to MVD and play a key role in tumor angiogenesis and progression. Therefore, we further investigated the infiltration of CAFs in GGO-LUAD and SN-LUAD. CAFs were identified by S100A4 and α‐SMA staining via immunohistochemistry. We observed no significant difference in distribution of α‐SMA + activated CAFs between GGO-LUAD and SN-LUAD, but GGO-LUAD is still less than SN-LUAD in the distribution of α‐SMA + activated CAFs (Fig. 7B). However, we found that the distribution of S100A4 + CAFs in SN-LUAD was significantly higher than that in GGO-LUAD (Fig. 6B). Also in GGO-LUAD, there was no difference in the distribution of S100A4 + CAFs and α‐SMA + activated CAFs between different pathological subtypes (Figs. 6C and 7C). Representative images of immunohistochemistry staining are shown in Figs. 6A and 7A.

Fig. 7.

Distribution of α‐SMA + cancer-associated fibroblasts (CAFs) in tumor specimens. A Representative immunohistochemistry images for distribution of α‐SMA + activated CAFs in different radiological patterns and pathological subtypes. Black scale bar: 100 μm. Red scale bar: 25 μm. B α‐SMA + activated CAFs were analyzed among GGO-LUAD and SN-LUAD. C α‐SMA + activated CAFs were analyzed among GGO-LUAD with different pathological subtypes

Fig. 6.

Distribution of S100A4 + cancer-associated fibroblasts (CAFs) in tumor specimens. A Representative immunohistochemistry images for distribution of S100A4 + CAFs in different radiological patterns and pathological subtypes. Black scale bar: 100 μm. Red scale bar: 25 μm. B S100A4 + CAFs were analyzed among GGO-LUAD and SN-LUAD. (C) S100A4 + CAFs were analyzed among GGO-LUAD with different pathological subtypes

TGF-β exhibits lower expression levels in GGO-LUAD compared to SN-LUAD

TGF-β plays an important role in the occurrence, development, and metastasis of tumors, including coordinating the development of tumor stroma, promoting angiogenesis, immune evasion, and ECM remodeling. More importantly, TGF-β can affect tumor angiogenesis not only by regulating VEGF expression [31], but also by promoting the migration and conversion of resident fibroblasts into CAFs as well as promoting endothelial cells to become CAFs through endothelial-to-mesenchymal transition (EndMT) [32–34]. Therefore, we were interested to see how the expression of TGF-β might differ in GGO-LUAD and SN-LUAD tumor tissues. Interestingly, we found that GGO-LUAD had significantly lower expression of TGF-β than SN-LUAD (Fig. 8B). However, in GGO-LUAD, there was no difference in TGF-β expression between different pathological subtypes (Fig. 8C). Representative images of immunohistochemistry staining are shown in Fig. 8A.

Fig. 8.

Scoring of TGF-β expression in tumor specimens. A Representative immunohistochemistry images for TGF-β expression in different radiological patterns and pathological subtypes. Black scale bar: 100 μm. Red scale bar: 25 μm. B H-score of TGF-β expression were analyzed among GGO-LUAD and SN-LUAD. C H-score of TGF-β expression were analyzed among GGO-LUAD with different pathological subtypes. (Due to the limited tumor tissue in one patient in the GGO-LUAD group, tissue slices could not be made again. Therefore, in this experiment, only tumor slices from 14 patients were used in the GGO-LUAD group)

Discussion

Growing evidence [6–9] has demonstrated that GGO is an independent factor for good prognosis of lung adenocarcinoma. However, it is still unclear why the prognosis of GGO-LUAD is good. Tumor angiogenesis plays a crucial role in the occurrence, development, and metastasis of tumors, which is closely related to the prognosis of tumors. To the best of our knowledge, this study is the first to explore the difference between GGO-LUAD and SN-LUAD from the perspective of tumor microvascular environment, which may explain the reason behind the good prognosis of GGO-LUAD.

The most common form of metastasis for lung cancer is lymph node metastasis. In the early stages of tumor spread, malignant cells spread from the primary site to regional lymph nodes. Hence, the lymphatic system plays an essential role in cancer biology. The lymphatic vessel density (LVD) is the most widely used indicator to measure tumor lymphangiogenesis. In the present study, we first stained LYVE-1 (lymphatic vessel marker) with immunohistochemistry technology, focusing on analyzing the differences in the lymphatic vessel density in GGO-LUAD and SN-LUAD. We found that GGO-LUAD has more abundant lymphatic vessels than SN-LUAD. However, the role of LVD as a prognostic predictor in NSCLC remains controversial. Vamos et al. [35] found that high LVD correlated with poor overall survival in NSCLC. Another study [36] demonstrated that lung cancer patients with high expression of LYVE-1 had better prognosis than those with low expression. In recent years, there is more evidence that lymphatic vessels not only play a role in tumor metastasis, but also boost of antitumor immunity by activating more cytotoxic T lymphocytes [37–39]. Therefore, the high density of lymphatic vessels may be one of the reasons for the good prognosis of GGO-LUAD.

One of the ten most important biological features of cancer is tumor angiogenesis [40]. In malignant tumors, cancer cells acquire invasive behavior and trigger a stromal response involving powerful angiogenesis [41]. Therefore, we hypothesized that the good prognosis of GGO-LUAD may be closely related to tumor angiogenesis. Previous studies [14, 17, 42] have shown that tumor microvessels in different status play different roles in tumors, even the opposite. CD34 is only expressed in differentiated endothelial cells [43] and CD105 is a glycoprotein expressed on activated endothelial cells in tissues participating in angiogenesis [44]. The microvessels marked by the above two molecules can better reflect the ability of tumor angiogenesis, which is closely related to tumor growth and metastasis [45]. Hence, we stained the marker of microvessels in different status, and found that although GGO-LUAD is similar to SN-LUAD in the distribution of CD31 + MVD, GGO-LUAD was significantly lower than SN-LUAD in CD34 + MVD and CD105 + MVD. It is well known that hypoxia is the most critical factor leading to tumor angiogenesis [46]. Next, we also found that GGO-LUAD was significantly lower than SN-LUAD in HIF-1α expression. At the same time, in terms of the proliferation ability (Ki-67 expression) of tumor cells, GGO-LUAD is also significantly weaker than SN-LUAD. This is consistent with the research results of Bos et al. which suggest that HIF-1α overexpression is closely related to tumor proliferation [47]. Therefore, the weak angiogenesis and proliferation ability of GGO-LUAD may be the reason for its remarkable biological inertness.

As an important component of tumor stroma, cancer-associated fibroblasts also play an important role in tumor angiogenesis and progression [28–30]. In order to explore the distribution of CAFs in GGO-LUAD, we use the two most commonly used molecules (S100A4 and α‐SMA) in many studies to label CAFs [48]. We ultimately found that GGO-LUAD had a less distribution of CAFs in both molecular markers, especially in the S100A4 + CAFs, which was significantly lower than SN-LUAD. This fully demonstrates that CAFs may play an important role in the progress of GGO-LUAD. Various proximal and distal healthy cells can be reprogrammed into CAFs, including bone marrow stromal cells, adipocytes, resident fibroblasts, pericyte, endothelial cells and mesothelial cells [49]. Therefore, further research is needed to determine which cell derived CAFs play a role in GGO-LUAD.

The tumor microenvironment is not only composed of tumor cells, but also a variety of stromal components, including a variety of immune cells, fibroblasts, vascular cells (endothelial cells and pericytes) and extracellular matrix [50]. As a pleiotropic cytokine, the importance of TGF-β for vasculogenesis has been known for some time [51]. Meanwhile, TGF-β can affect tumor angiogenesis not only by regulating VEGF expression [31], but also by promoting the migration and conversion of resident fibroblasts into CAFs as well as promoting endothelial cells to become CAFs through EndMT [32–34]. In our work, we found that GGO-LUAD had significantly lower expression of TGF-β than SN-LUAD, which may be one of the factors leading to the weaker angiogenesis ability of GGO-LUAD.

Several shortcomings of our study should be mentioned. Firstly, the sample size in this study remains relatively small, even though it is the first to investigate the differences between GGO-LUAD and SN-LUAD in relation to the tumor microvascular environment. Therefore, the findings should be interpreted with caution, and further research with a larger sample size is needed to validate these results. Secondly, in immunohistochemistry analysis, although the slides were independently evaluated by multiple people in a blinded fashion, there is still inevitable selection bias. Lastly, in this study, we did not explore the specific molecular mechanisms by which TGF-β and CAFs affect tumor angiogenesis in GGO-LUAD, more research is urgently needed in the future.

Conclusions

In summary, our findings suggest that GGO-LUAD was significantly lower than SN-LUAD in CD34 + MVD and CD105 + MVD reflecting tumor angiogenesis, and the distribution of CAFs and factors related to tumor angiogenesis (TGF-β and HIF-1α) were also significantly lower, which may indicate that the weak ability of angiogenesis might be the reason for the good prognosis of GGO-LUAD. Considering the small sample size in our study, future analyses with a larger cohort are still needed to confirm these findings.

Abbreviations

AIS

Adenocarcinoma in situ

CAFs

Cancer-associated fibroblasts

ECM

Extracellular matrix

EndMT

Endothelial-to-mesenchymal transition

FFPE

Formalin-fixed, paraffin-embedded

GGO

Ground glass opacity

GGO-LUAD

Ground glass opacity featured lung adenocarcinomas

HRCT

High-resolution computed tomography

IAC

Invasive lung adenocarcinoma

LVD

Lymphatic vessel density

MVD

Microvessel density

MIA

Minimally invasive adenocarcinoma

SN

Solid nodule

SN-LUAD

Solid nodule featured lung adenocarcinomas

Author contributions

Rirong Qu: Conceptualization, Data curation, Formal analysis, Software, Investigation, Methodology, Supervision, Visualization, Roles/Writing—original draft, Writing—review & editing. Yang Zhang: Data curation, Formal analysis, Software, Writing—review & editing. Shenghui Qin: Data curation, Formal analysis, Project administration, Visualization. Jing Xiong: Data curation, Project administration, Visualization. Xiangning Fu: Data curation, Formal analysis, Investigation, Funding acquisition, Writing—review & editing. Lequn Li: Data curation, Formal analysis, Investigation, Validation, Visualization, Writing—review & editing. Dehao Tu: Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Roles/Writing—original draft, Writing—review & editing. Yixin Cai: Conceptualization, Data curation, Formal analysis, Investigation, Funding acquisition, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Roles/Writing—original draft, Writing—review & editing.

Funding

The study is funded by the China Postdoctoral Science Foundation under Grant Number 2024M761036, the Tongji Hospital Clinical Research Flagship Program (No.2019CR107) and Chen Xiaoping Foundation Youth Research Fund in Hubei Province (CXPJJH11900018-01). The funding sources had no influence on the analysis and interpretation of data or on the contents of the manuscript.

Data availability

Data is provided within the manuscript or supplementary information files.The data sets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethical approval and consent to participate

This study was approved by the institutional review board of Tongji Medical College of Huazhong University of Science and Technology and written informed consent was provided by all participants.

Competing interests

The authors declare no competing interests.