Evaluation of high-grade growth patterns in invasive adenocarcinomas manifesting as ground-glass lesions via clinical and qualitative and quantitative computed tomography features

Introduction

Lung cancer is the leading cause of cancer-related death worldwide (1). Adenocarcinoma accounts for nearly half of all lung cancers, and invasive adenocarcinoma (IAC), characterized by its high heterogeneity, is the most common histological subtype (2). A growing number of studies are indicating that the histological heterogeneity of IAC also reflects the biological diversity of individual tumors and influences patient prognosis (3,4). Moreover, it has been established that regardless of the dominant histological pattern of IAC, the presence of high-grade growth patterns (HGPs) significantly increases tumor aggressiveness, leading to a relatively poor prognosis (5-7). Therefore, evaluating the presence of HGPs in IACs is critical for accurate risk stratification and treatment planning.

In order to further clarify the correlation between HGPs and tumor aggressiveness, the International Association for the Study of Lung Cancer (IASLC) proposed a new pathological grading system in 2020 (8). By incorporating the presence of HGPs and applying a 20% threshold, the IASLC system recategorized IACs as well differentiated, moderately differentiated, and poorly differentiated. Several studies have confirmed the efficacy of this grading system in predicting the prognosis of the patients with IACs (9-11). Poorly differentiated IACs account for 34–55% of resected adenocarcinomas and are associated with significantly worse prognosis as compared to their well differentiated and moderately differentiated counterparts, with recurrence rates of 16.7%, 0.0%, and 6.0%, respectively (12). Hence, the identification of poorly differentiated IACs may hold critical clinical significance that warrants further investigation.

Thin-section computed tomography (TSCT) is recommended as the primary screening modality for lung cancer due to its rapid scanning and high-resolution capabilities (13). Neoplastic lesions manifesting with ground-glass opacity (GGO) are commonly detected during screening and are generally associated with a favorable prognosis (14-16). Therefore, compared to that of solid lung adenocarcinomas, the management of adenocarcinomas with GGO is relatively conservative (17). However, recent research suggests that the presence of HGPs in patients with GGO-IACs is an adverse prognostic factor, which may be associated with a relatively poorer prognosis (7). Consequently, determining the presence of HGPs in lesions and the corresponding pathological grade may be crucial for identifying and managing potentially high-risk GGO-IACs.

Previous studies (18,19) reported that qualitative and quantitative features of ground-glass nodules on computed tomography (CT) can reliably predict their invasiveness, while only a few studies have focused on identifying CT-based indicators of IACs containing HGPs (20-22). However, this research has exclusively examined GGO-IACs and solid IACs (S-IACs), but no study dedicated specifically to GGO-IACs or the evaluation of their pathological grade has been conducted. Moreover, although the prediction of HGPs in GGO-IACs has primarily been achieved through radiomics and machine learning (23,24), the clinical practicality of these models remains limited. Therefore, this study aimed to further examine and verify the clinical and CT indicators for predicting the presence of HGPs and the pathological grade related to HGPs in GGO-IACs. We present this article in accordance with the STROBE and TRIPOD reporting checklists (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-aw-2490/rc).

Methods

This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments and was approved by the Institutional Review Board of The First Affiliated Hospital of Chongqing Medical University (No. 2025-118-01). The requirement for informed consent was waived due to the retrospective nature of the analysis.

Patients

A search of the electronic medical record system was conducted to identify patients who underwent resection of a pulmonary lesion in the thoracic surgery department from September 2017 to December 2024, and 27,984 patients were identified. First, a fellowship-trained thoracic radiologist (S.T.Y.) reviewed the electronic medical records. A total of 13,863 patients with noninvasive or benign lesions were excluded, comprising 4,033 with precursor lesions, 3,098 with fibrous tissue hyperplasia, 2,858 with inflammatory lesions, 1,736 with infectious lesions, and 2,138 with other benign lesions (including lymphoid tissue, granuloma, hamartoma, organizing pneumonia, etc.). Subsequently, we excluded 3,661 patients with squamous cell carcinoma, 729 with small-cell carcinoma, 592 with metastatic carcinoma, 644 with other malignant tumors (including adenosquamous carcinoma, large-cell carcinoma, and sarcomatoid carcinoma), 2,404 with minimally invasive adenocarcinoma (MIA), and 152 patients with IAC who lacked complete pathological data. The previously mentioned radiologist (S.T.Y.) and an additional senior thoracic radiologist (Z.G.C.) jointly reviewed the CT images of these patients, excluding 1,137 patients without thin-section CT images (≤1 mm) within 1 month before surgery and 2,875 patients with solid lesions. After preliminary selection, the remaining patients with pure ground-glass lesions (n=305) and part-solid lesions (n=1,622) were considered candidates for this study [based on the absence or presence of solid components obscuring the underlying lung architecture (25)]. From these 1927 patients, those with lesions having large or multiple cystic spaces affecting evaluation (n=29) and those with CT images containing artifacts or noise that compromised accurate evaluation (n=17) were excluded. The final cohort included 1,881 patients (mean age 60.5±9.8 years; 688 men and 1,291 women) comprising 1,979 GGO-IACs. The flowchart of patient selection is shown in Figure 1.

Figure 1 Flowchart of patient inclusion. CT, computed tomography; IAC, invasive adenocarcinoma; MIA, minimally invasive adenocarcinoma; TSCT, thin-section computed tomography.

Imaging acquisition

The image analysis was based on plain CT scans, which were performed with one of the following CT scanners: SOMATOM Perspective (Siemens Healthineers, Erlangen, Germany), SOMATOM Definition Flash (Siemens Healthineers), SOMATOM Force (Siemens Healthineers), Discovery CT750 HD (GE HealthCare, Chicago, IL, USA), and Aquilion ONE pureViSION (Canon Medical Systems, Otawara, Japan). All patients were placed in the supine position with both upper limbs raised and were scanned from the thoracic inlet to the lung base at the end of inspiration during a single breath-hold. The plain CT images were obtained under the following parameters: tube voltage, 110–120 kVp; tube current, 50–140 mAs; beam pitch, 1.0–1.1; scanning slice thickness, 5 mm; rotation time, 0.5–0.6 s; collimation, 0.6–0.625 mm; reconstruction slice thickness and spacing, 0.625 or 1 mm; matrix, 512×512; and iterative reconstruction with standard-sharpness (GE HealthCare CT scanners) or medium-sharpness (Siemens Healthineers scanners) algorithms.

Histopathological evaluation

Hematoxylin and eosin (HE)-stained sections were evaluated, and a semiquantitative estimation method with 5% increments for all patterns was employed first at 40× magnification to observe the lesion across the entire section and then at 200× magnification to observe tumor growth patterns and their respective proportions (26). HGPs include solid, micropapillary, cribriform, and fused gland growth patterns (8,27,28). Based on the presence or absence of HGPs, all lesions were divided into two groups: group A (containing HGPs) and group B (no HGPs). According to the pathological grading system proposed by the IASLC, lesions were reclassified as well differentiated (lepidic predominant with <20% HGPs), moderately differentiated (acinar or papillary predominant with <20% HGPs), and poorly differentiated (any histological patterns with ≥20% HGPs) (8).

Clinical data and image analysis

Information on the electronic medical record system, including the patients’ clinical data (age, gender, smoking history, lung cancer history, other malignant tumor history, and family malignant tumor history), was recorded. All the CT images were reviewed on a picture archiving and communication system workstation by two chest radiologists (S.T.Y and Z.G.C), with Z.G.C being the one blinded to the pathology results. Any discrepancy between the two radiologists was resolved by discussion until a consensus was reached.

The following CT features of the lesions were analyzed: (I) size (the mean of the longest diameter and the perpendicular diameter on axial CT images); (II) distribution (upper lobe, middle lobe, or lower lobe); (III) shape (round, oval, or irregular); (IV) CT pattern (pure ground-glass lesion or part-solid lesion); (V) boundary (well defined or ill defined); (VI) distribution of solid component (marginal or nonmarginal); (VII) consolidation-to-tumor ratio (CTR); (VIII) vacuole sign; (IX) air bronchogram sign; (X) lobulation; (XI) spiculation; (XII) pleural indentation; and (XIII) pleural tag. The distribution of the solid component in lesions included marginal distribution (the solid component closely opposed to the normal lung tissue and partially surrounded by GGO component) and nonmarginal distribution (the solid component totally surrounded by a GGO component) (Figure 2). CTR was calculated as the ratio of the solid component area to the total lesion area on the axial section with the maximum lesion area.

Figure 2 TSCT images of lesions and the distribution of solid components. (A) Pure ground-glass lesion without any solid component. (B) Part-solid lesion with a nonmarginally distributed solid component (the solid component totally surrounded by the ground-glass opacity component). (C) Part-solid lesion with a marginally distributed solid component (the arrow points to the solid component that is partly surrounded by ground-glass opacity component). TSCT, thin-section computed tomography.

Statistical analysis

Data processing and statistical analysis were conducted with SPSS version 26.0 (IBM Corp., Armonk, NY, USA) and R version 3.2.5 software (The R Foundation for Statistical Computing, Vienna, Austria). Continuous variables are expressed as the mean ± standard deviation, with statistical differences analyzed via the Mann-Whitney U test or Kruskal-Wallis test. Categorical variables are expressed as counts and percentages and were analyzed with the Pearson chi-squared test or Fisher exact test, with the Benjamini-Hochberg procedure being applied to adjust for multiple comparisons. Univariate analysis and multivariate logistic regression analyses were employed to identify independent indicators for predicting the presence of HGPs and the pathological grade in GGO-IACs. Before multivariate logistic regression was conducted, the multicollinearity of variables was examined via the variance inflation factor. The cutoff values of the independent continuous variables identified in the multivariate logistic regression were further calculated via receiver operating characteristic (ROC) curve analysis. To evaluate the robustness and generalization of the models, fivefold cross-validation was performed for each model. The intraclass correlation coefficient (ICC) and Cohen kappa coefficient were used to assess the interobserver agreement of continuous and categorical variables, respectively. Interobserver agreement according to ICC was classified as poor (<0.500), moderate (0.500–0.740), good (0.750–0.890), or excellent (≥0.900). Interobserver agreement according to kappa coefficients was categorized as poor (<0.000), slight (0.000–0.200), fair (0.210–0.400), moderate (0.410–0.600), substantial (0.610–0.800), or almost perfect (0.810–1.000) (29).

The clinical and CT models were constructed based on the results of the multivariable logistic regression analysis, and area under the curve (AUC) was used to evaluate the predictive effectiveness of these models. A P value <0.05 was considered statistically significant.

Results

Clinical characteristics of patients and the pathological and CT features of lesions

Among the 1,979 lesions in 1,881 patients, 490 contained HGPs (group A), while 1,489 had no HGPs (group B). Table 1 summarizes the patients’ clinical characteristics and the pathological and CT features of lesions. Compared to group B, group A had a greater proportion of males (40.2% vs. 32.9%) and smokers (29.1% vs. 21.5%) (P<0.001).

Lesions in group A, as compared to those in group B, more frequently exhibited part-solid manifestation (85.5% vs. 81.5%), a marginally distributed solid component (45.3% vs. 10.5%), ill-defined boundary (28.5% vs. 11.6%), irregular shape (14.2% vs. 7.7%), spiculation (34.4% vs. 11.2%), vacuole sign (38.3% vs. 29.4%), pleural tag (41.2% vs. 29.0%), and pleural indentation (37.7% vs. 16.3%, P<0.05). Additionally, the lesions in group A, as compared to those in group B, had a significantly higher mean total diameter (18.7±7.6 vs. 16.6±6.6 mm) and CTR (13.9%±13.7% vs. 26.9%±22.1%, P<0.05, Figure 3).

Figure 3 A 47-year-old male with moderately differentiated IAC without HGPs (A-C) and a 66-year-old male smoker with moderately differentiated IAC with HGPs (D-F). (A) An axial TSCT unenhanced image showed a 17-mm, well-defined and lobulated part-solid nodule located in the right upper lobe. The internal solid component was nonmarginally distributed. (B) A photomicrograph (HE, ×5.4) indicated that this nodule consisted of a predominant lepidic growth pattern (black asterisks). Within it, a region of increased cellular density was noted (white asterisk), which corresponded to the solid component of the nodule seen on the TSCT image. (C) At higher magnification (HE, ×100), the densely cellular region was by the acinar growth pattern (arrow). (D) An axial TSCT unenhanced image showed a 32-mm, well-defined and lobulated part-solid tumor located in the right upper lobe, with spiculation, pleural indentation, and air bronchogram sign. The internal solid component was marginally-distributed. (E) A photomicrograph (HE, ×3.4) revealed a region of increased cellular density within the tumor (asterisks), which corresponded to the solid component of the tumor observed on the TSCT image. (F) At higher magnification (HE, ×100), the densely cellular region was characterized by a predominant acinar growth pattern (white arrow) and a less predominant micropapillary growth pattern (approximately 8%; black arrow). HE, hematoxylin and eosin; HGP, high-grade growth pattern; IAC, invasive adenocarcinoma; TSCT, thin-section computed tomography.

Prior to multivariate logistic regression analysis, multicollinearity assessment was performed among the candidate variables (Table S1). In the multivariate logistic regression analysis, part-solid manifestation, marginally distributed solid component, ill-defined boundary, spiculation, pleural indentation, pleural tag, and higher CTR were identified as independent predictors of HGPs (P<0.001). For the CTR, ROC analysis yielded an optimal cutoff value of 21.5% for predicting HGPs (AUC =0.662; sensitivity =54.3%; specificity =76.4%). The AUC of this model was 0.792 (95% CI: 0.767–0.816; sensitivity =73.7%; specificity =73.7%; P<0.001, Figure 4). Fivefold cross-validation internal validation was performed to further confirm the robustness and generalization of the model, yielding an AUC of 0.788 (95% CI: 0.765–0.809; fold-wise standard deviation =0.028).

Figure 4 ROC curves for the identification of HGP status and GGO-IAC differentiation. (A) Prediction of GGO-IACs with HGPs. (B) Prediction of poorly differentiated GGO-IACs. (C) Prediction of moderately differentiated GGO-IACs. AUC, area under the curve; GGO-IAC, ground-class opacity invasive adenocarcinoma; HGP, high-grade growth pattern; ROC, receiver operating characteristic.

Clinical characteristics of patients and the pathological and CT features of the well/moderately and poorly differentiated lesions

Among the 1,979 GGO-IACs, 1,861 (94.0%) were well/moderately differentiated and 118 (6.0%) were poorly differentiated. Table 2 lists the clinical characteristics of the patients and the pathological and CT features of lesions. Compared to the group of well/moderately differentiated GGO-IACs, the group with poorly differentiated GGO-IACs had a higher proportion of males (45.7% vs. 34.0%, P<0.05).

Compared to the well/moderately GGO-IACs, poorly differentiated GGO-IACs had a higher CTR (29.2%±23.5% vs. 16.4%±16.4%; P<0.001) and more frequently exhibited a marginally distributed solid component (46.6% vs. 17.4%), ill-defined boundary (24.5% vs. 15.2%), spiculation (38.9% vs. 15.5%), vacuole sign (40.6% vs. 31.0%), and pleural indentation (33.0% vs. 20.9%) but less frequently showed lobulation (17.9% vs. 27.9%, P<0.05, Figure 5).

Figure 5 A 73-year-old female with well-differentiated IAC without HGPs (A-C) and another 73-year-old female with a poorly differentiated IAC with HGPs (D-F). (A) An axial TSCT unenhanced image showed a 23-mm ill-defined and lobulated part-solid nodule located in the right upper lobe. The internal solid component was nonmarginally distributed. (B) A photomicrograph (HE, ×5.5) indicated that this nodule consisted of a predominant lepidic growth pattern (black asterisks). Within it, a region of increased cellular density was observed (white asterisk), which corresponded to the solid component of the nodule seen on the TSCT image. (C) At higher magnification (HE, ×100), the densely cellular region was characterized by a predominant lepidic growth pattern (white arrow) and a less predominant acinar pattern (approximately 30%; black arrow). (D) An axial TSCT unenhanced image showed a 22-mm well-defined and lobulated part-solid nodule located in the left upper lobe, with spiculation, pleural indentation, and air bronchogram sign. The internal solid component was marginally distributed. (E) A photomicrograph (HE, ×4.4) revealed that this nodule consisted of a lepidic growth pattern (black asterisk) and extensive fibrosis (gray asterisk). Within the fibrotic area, a region of increased cellular density was noted (white asterisk), and both of them corresponded to the solid component of the nodule observed on the TSCT image. (F) At higher magnification (HE, ×100), the densely cellular region was characterized by the micropapillary growth pattern (arrow). IAC, invasive adenocarcinoma; HE, hematoxylin and eosin; HGP, high-grade growth pattern; TSCT, thin-section computed tomography.

Prior to multivariate logistic regression analysis, multicollinearity assessment was performed among the candidate variables (Table S2). In multivariate logistic regression analysis, absence of lobulation, spiculation, marginally distributed solid component, ill-defined boundary, and higher CTR were identified as independent predictors of poorly differentiated GGO-IACs (P<0.05). For the CTR, ROC analysis yielded an optimal cutoff value of 28.5% for predicting HGPs (AUC =0.650; sensitivity =51.7%; specificity =79.8%). The AUC of this model was 0.739 (95% CI: 0.689–0.790; sensitivity =61.0%; specificity =75.3%; P<0.001, Figure 4). Fivefold cross-validation internal validation was performed to further confirm the robustness and generalization of the model, yielding an AUC of 0.722 (95% CI: 0.691–0.762; fold-wise standard deviation =0.047).

Clinical characteristics of patients and the pathological and CT features of well- and moderately differentiated lesions

Of the 1979 GGO-IACs, 1629 (82.3%) were moderately differentiated and 232 (11.7%) were well differentiated. Table S3 presents the patients’ clinical characteristics and the pathological and CT features of lesions. Compared to well-differentiated GGO-IACs, the moderately differentiated ones had a higher CTR (16.7%±16.6% vs. 13.8%±14.7%; P<0.05) and more frequently demonstrated a well-defined boundary (85.7% vs. 77.5%) (P<0.05). In multivariate logistic regression analysis, a well-defined boundary was the only independent predictor of moderately differentiated GGO-IACs. The AUC of this model was 0.576 (95% CI: 0.537–0.614, P<0.001, Figure 4).

Comparison of patients’ clinical characteristics and the pathological and CT features of lesions between different pathological grades in group A

Among the 490 lesions in group A, 45 (9.1%) were well differentiated, 327 (66.7%) were moderately differentiated, and 118 (24.0%) were poorly differentiated. Table S4 shows the clinical characteristics of patients and the pathological and CT features of lesions. Both univariate and multivariate analyses revealed no significant differences across the pathological grades.

Interobserver agreement

Table S5 summarizes the interobserver agreement for the CT features. For continuous features, agreement was almost perfect for size (ICC =0.964) and CTR (ICC =0.966). For categorical features, agreement was also almost perfect (κ=0.819–0.964).

Discussion

This study investigated the utility of clinical and CT features for the qualitative and quantitative assessment of HGPs in GGO-IACs. Compared to clinical characteristics, the CT features of lesions were more closely associated with the presence of HGPs and the pathological grade associated with HGPs. It was found that GGO-IACs with HGPs were typically associated with part-solid lesions exhibiting a marginally distributed solid component, ill-defined boundary, spiculation, pleural indentation, pleural tag, and higher CTR. Moreover, lesions with a marginally distributed solid component, ill-defined boundary, spiculation, and higher CTR but without lobulation had a higher possibility of being poorly differentiated.

These findings indicate that the abundance of HGPs in GGO-IACs could be preliminarily evaluated through their CT features.

Among the various CT features, CTR serves as a well-known critical factor associated with the prognosis of patients with IACs (30,31). Previous studies have confirmed that the CTR of IACs is associated with lymph node metastasis (LNM) (31-33). This study further revealed that GGO-IACs with a higher CTR were more likely to contain HGPs. This finding indicates that the LNM in GGO-IACs with a higher CTR may be associated with a higher frequency of HGPs within these lesions. Li et al. (34) found that lesions with scattered or eccentric solid components are more frequently associated with the presence of HGPs and with recurrence. Notably, we found a similar finding that a marginally distributed solid component on CT was an independent predictor of lesions containing HGPs, and it demonstrated a closer association with HGPs than did the CTR. We further conducted an analysis of the lesions with LNM and found that larger lesion size, a higher CTR (>31.5%), and pleural indentation were the risk factors for LNM (Table S6). This finding further confirms the relationship between the distribution of solid components and the presence of HGPs. Therefore, the proportion and distribution of solid components should be simultaneously evaluated for the prediction of HGPs.

Previous studies on this subject have evaluated the presence of HGPs based on clinical and CT features (20-22). In our work, CT features were also associated with the pathological grade in GGO-IACs. An Ill-defined boundary, marginally distributed solid components, spiculation, and a higher CTR served as predictors of poorly differentiated GGO-IACs (amount of HGPs ≥20%). This indicates that CT features are also associated with the abundance of HGPs in GGO-IACs. It is worth noting that lesions without lobulation were more common in poorly differentiated GGO-IACs despite it being an important predictor of lung cancer, and this may be because more poorly differentiated lesions were irregular, with their lobulation being incorporated into the irregular shape during assessment.

Three previous studies examined the association between conventional CT features and HGPs in both GGO-IACs and S-IACs (20-22). It was found that solid lesion, spiculation, and larger lesion diameter were predictors of HGPs, with solid lesion showing the strongest correlation. This suggests that the solid manifestation of lesions may represent higher invasiveness. In this study, when only GGO-IACs were included, a relatively greater number of CT indicators correlating with the presence of HGPs were identified. Furthermore, it was found that the CT features were only significantly different between poorly differentiated and moderately/well-differentiated GGO-IACs. However, for GGO-IACs with different pathological grades in group A, no differences were observed in either clinical characteristics or CT features. This finding indicates that for GGO-IACs containing HGPs, despite differences in their pathological grades, their CT features reflecting tumor invasiveness are consistent, implying that their invasiveness is comparable (35,36). This further suggests that once HGPs are present, their quantity may not be reflected in gross CT morphology. This may also indicate that the impact of HGPs on the invasiveness of GGO-IACs is weaker than that on S-IACs. A previous study also confirmed that GGO-IACs of different sizes and CTRs are not significantly different in terms of prognosis (37). Thus, it can be speculated that HGPs exert a less pronounced prognostic effect on GGO-IACs than on S-IACs. While GGO-IACs exhibit marked heterogeneity in histopathological characteristics compared to S-IACs, they may display a similar degree of invasiveness and be associated with generally better prognosis. However, further investigation is needed to confirm this.

This study involved several limitations that should be acknowledged. First, we employed a single-center, retrospective design. The model developed for poorly differentiated GGO-IACs showed moderate discriminatory performance. Moreover, poorly differentiated GGO-IACs accounted for only 6% of the total sample, suggesting its direct clinical application may currently be limited. These findings need to be further validated in other centers. Second, the number of lesions containing HGPs and those lacking HGPs was imbalanced in the sample, which might have introduced bias into the results. Third, while CT features demonstrated potential in predicting the presence of HGPs and poorly differentiated lesions in GGO-IACs, manual feature assessment used is inherently subjective and poses challenges to reproducibility in a multicenter setting. In future work, we aim to examine the integration of deep learning architectures to enable the automated extraction of such semantic features; via advanced deep learning models, feature labeling can be standardized to ensure consistent diagnostic performance across multicenter settings. Finally, the follow-up period for GGO-IACs in the study was short, at only 2 to 3 years, making it difficult to evaluate the long-term prognosis. We will conduct prognostic analysis in subsequent research to supplement this study.

Conclusions

There was a close correlation between CT features and HGPs in GGO-IACs. Among the suspected GGO-IACs, the part-solid ones with an ill-defined boundary, marginally distributed solid components, spiculation, and CTR >28.5% had a higher possibility of containing HGPs and being poorly differentiated. Such lesions may exhibit greater aggressiveness and thus require more aggressive clinical management.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE and TRIPOD reporting checklists. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-aw-2490/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-aw-2490/dss

Funding: This work was supported by the Project of Chongqing Natural Science Foundation (No. CSTB2024NSCQ-MSX0655) and the Senior Medical Talents Program of Chongqing for Young and Middle-aged from Chongqing Health Commission (to Z.G.C.).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-aw-2490/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of The First Affiliated Hospital of Chongqing Medical University (No. 2025-118-01), and individual consent for this retrospective analysis was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, Jemal A. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Travis WD, Brambilla E, Noguchi M, Nicholson AG, Geisinger K, Yatabe Y, Powell CA, Beer D, Riely G, Garg K, Austin JH, Rusch VW, Hirsch FR, Jett J, Yang PC, Gould MAmerican Thoracic Society. International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society: international multidisciplinary classification of lung adenocarcinoma: executive summary. Proc Am Thorac Soc 2011;8:381-5. [Crossref] [PubMed]
3. Russell PA, Wainer Z, Wright GM, Daniels M, Conron M, Williams RA. Does lung adenocarcinoma subtype predict patient survival?: A clinicopathologic study based on the new International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society international multidisciplinary lung adenocarcinoma classification. J Thorac Oncol 2011;6:1496-504. [Crossref] [PubMed]
4. Russell PA, Barnett SA, Walkiewicz M, Wainer Z, Conron M, Wright GM, Gooi J, Knight S, Wynne R, Liew D, John T. Correlation of mutation status and survival with predominant histologic subtype according to the new IASLC/ATS/ERS lung adenocarcinoma classification in stage III (N2) patients. J Thorac Oncol 2013;8:461-8. [Crossref] [PubMed]
5. Mikubo M, Tamagawa S, Kondo Y, Hayashi S, Sonoda D, Naito M, Shiomi K, Ichinoe M, Satoh Y. Micropapillary and solid components as high-grade patterns in IASLC grading system of lung adenocarcinoma: Clinical implications and management. Lung Cancer 2024;187:107445. [Crossref] [PubMed]
6. Takahashi Y, Eguchi T, Kameda K, Lu S, Vaghjiani RG, Tan KS, Travis WD, Jones DR, Adusumilli PS. Histologic subtyping in pathologic stage I-IIA lung adenocarcinoma provides risk-based stratification for surveillance. Oncotarget 2018;9:35742-51. [Crossref] [PubMed]
7. Li R, Li Z, Yang Z, Qiu B, Tan F, Xue Q, Gao S, He J. The presence of micropapillary and/or solid subtypes is an independent prognostic factor for patients undergoing curative resection for stage I lung adenocarcinoma with ground-glass opacity. Transl Lung Cancer Res 2024;13:256-68. [Crossref] [PubMed]
8. Moreira AL, Ocampo PSS, Xia Y, Zhong H, Russell PA, Minami Y, et al. A Grading System for Invasive Pulmonary Adenocarcinoma: A Proposal From the International Association for the Study of Lung Cancer Pathology Committee. J Thorac Oncol 2020;15:1599-610. [Crossref] [PubMed]
9. Choi Y, Kim J, Park H, Kim HK, Kim J, Jeong JY, Ahn JH, Lee HY. Rethinking a Non-Predominant Pattern in Invasive Lung Adenocarcinoma: Prognostic Dissection Focusing on a High-Grade Pattern. Cancers (Basel) 2021;13:2785. [Crossref] [PubMed]
10. Fujikawa R, Muraoka Y, Kashima J, Yoshida Y, Ito K, Watanabe H, Kusumoto M, Watanabe SI, Yatabe Y. Clinicopathologic and Genotypic Features of Lung Adenocarcinoma Characterized by the International Association for the Study of Lung Cancer Grading System. J Thorac Oncol 2022;17:700-7. [Crossref] [PubMed]
11. Feng J, Shao X, Gao J, Ge X, Sun Y, Shi Y, Wang Y, Niu R. Application and progress of non-invasive imaging in predicting lung invasive non-mucinous adenocarcinoma under the new IASLC grading guidelines. Insights Imaging 2025;16:4. [Crossref] [PubMed]
12. Cho IS, Shim HS, Lee HJ, Suh YJ. Clinical implication of the 2020 International Association for the Study of Lung Cancer histologic grading in surgically resected pathologic stage 1 lung adenocarcinomas: Prognostic value and association with computed tomography characteristics. Lung Cancer 2023;184:107345. [Crossref] [PubMed]
13. Henschke CI, Yip R, Shaham D, Markowitz S, Cervera Deval J, Zulueta JJ, Seijo LM, Aylesworth C, Klingler K, Andaz S, Chin C, Smith JP, Taioli E, Altorki N, Flores RM, Yankelevitz DFInternational Early Lung Cancer Action Program Investigators. A 20-year Follow-up of the International Early Lung Cancer Action Program (I-ELCAP). Radiology 2023;309:e231988. [Crossref] [PubMed]
14. Miyoshi T, Aokage K, Katsumata S, Tane K, Ishii G, Tsuboi M. Ground-Glass Opacity Is a Strong Prognosticator for Pathologic Stage IA Lung Adenocarcinoma. Ann Thorac Surg 2019;108:249-55. [Crossref] [PubMed]
15. Choe J, Lee SM, Park S, Choi S, Do KH, Seo JB. The prognostic value of lymphovascular invasion for stage I lung adenocarcinoma based on the presence of ground-glass opacity. Eur Radiol 2025;35:2256-64. [Crossref] [PubMed]
16. Nakamura S, Fukui T, Kawaguchi K, Fukumoto K, Hirakawa A, Yokoi K. Does ground glass opacity-dominant feature have a prognostic significance even in clinical T2aN0M0 lung adenocarcinoma? Lung Cancer 2015;89:38-42. [Crossref] [PubMed]
17. Zhong F, Xu J, Wu L, Zhao S. Comparative long-term prognosis of early surgery and surgery after surveillance for patients with ground-glass nodule adenocarcinomas. Sci Rep 2024;14:18785. [Crossref] [PubMed]
18. Xie B, Wang R, Fu K, Wang Q, Liu Z, Peng W. The value of predicting the invasiveness and degree of infiltration of pulmonary ground-glass nodules based on computed tomography features and enhanced quantitative analysis. Quant Imaging Med Surg 2024;14:6767-79. [Crossref] [PubMed]
19. Meng Q, Zheng C, Guo L, Gao P, Liu W, Ge H, Liu T, Peng H, Lu J, Chen X. Construction and validation of a risk score system for diagnosing invasive adenocarcinoma presenting as pulmonary pure ground-glass nodules: a multi-center cohort study in China. Quant Imaging Med Surg 2024;14:4864-77. [Crossref] [PubMed]
20. Xu J, Liu L, Ji Y, Yan T, Shi Z, Pan H, Wang S, Yu K, Qin C, Zhang T. Enhanced CT-Based Intratumoral and Peritumoral Radiomics Nomograms Predict High-Grade Patterns of Invasive Lung Adenocarcinoma. Acad Radiol 2025;32:482-92. [Crossref] [PubMed]
21. Dong H, Yin LK, Qiu YG, Wang XB, Yang JJ, Lou CC, Ye XD. Prediction of high-grade patterns of stage IA lung invasive adenocarcinoma based on high-resolution CT features: a bicentric study. Eur Radiol 2023;33:3931-40. [Crossref] [PubMed]
22. Wang Z, Zhang N, Liu J, Liu J. Predicting micropapillary or solid pattern of lung adenocarcinoma with CT-based radiomics, conventional radiographic and clinical features. Respir Res 2023;24:282. [Crossref] [PubMed]
23. Wang X, Zhang L, Yang X, Tang L, Zhao J, Chen G, Li X, Yan S, Li S, Yang Y, Kang Y, Li Q, Wu N. Deep learning combined with radiomics may optimize the prediction in differentiating high-grade lung adenocarcinomas in ground glass opacity lesions on CT scans. Eur J Radiol 2020;129:109150. [Crossref] [PubMed]
24. Choi Y, Aum J, Lee SH, Kim HK, Kim J, Shin S, Jeong JY, Ock CY, Lee HY. Deep Learning Analysis of CT Images Reveals High-Grade Pathological Features to Predict Survival in Lung Adenocarcinoma. Cancers (Basel) 2021;13:4077. [Crossref] [PubMed]
25. Liang ZR, Lv FJ, Fu BJ, Lin RY, Li WJ, Chu ZG. Reticulation Sign on Thin-Section CT: Utility for Predicting Invasiveness of Pure Ground-Glass Nodules. AJR Am J Roentgenol 2023;221:69-78. [Crossref] [PubMed]
26. Travis WD, Brambilla E, Nicholson AG, Yatabe Y, Austin JHM, Beasley MB, Chirieac LR, Dacic S, Duhig E, Flieder DB, Geisinger K, Hirsch FR, Ishikawa Y, Kerr KM, Noguchi M, Pelosi G, Powell CA, Tsao MS, Wistuba I, Panel WHO. The 2015 World Health Organization Classification of Lung Tumors: Impact of Genetic, Clinical and Radiologic Advances Since the 2004 Classification. J Thorac Oncol 2015;10:1243-60. [Crossref] [PubMed]
27. Warth A, Muley T, Kossakowski C, Stenzinger A, Schirmacher P, Dienemann H, Weichert W. Prognostic impact and clinicopathological correlations of the cribriform pattern in pulmonary adenocarcinoma. J Thorac Oncol 2015;10:638-44. [Crossref] [PubMed]
28. Mäkinen JM, Laitakari K, Johnson S, Mäkitaro R, Bloigu R, Pääkkö P, Lappi-Blanco E, Kaarteenaho R. Histological features of malignancy correlate with growth patterns and patient outcome in lung adenocarcinoma. Histopathology 2017;71:425-36. [Crossref] [PubMed]
29. Benchoufi M, Matzner-Lober E, Molinari N, Jannot AS, Soyer P. Interobserver agreement issues in radiology. Diagn Interv Imaging 2020;101:639-41. [Crossref] [PubMed]
30. Sun F, Huang Y, Yang X, Zhan C, Xi J, Lin Z, Shi Y, Jiang W, Wang Q. Solid component ratio influences prognosis of GGO-featured IA stage invasive lung adenocarcinoma. Cancer Imaging 2020;20:87. [Crossref] [PubMed]
31. Li W, Zhou F, Wan Z, Li M, Zhang Y, Bao X, Zhang L, Shi J. Clinicopathologic features and lymph node metastatic characteristics in patients with adenocarcinoma manifesting as part-solid nodule exceeding 3 cm in diameter. Lung Cancer 2019;136:37-44. [Crossref] [PubMed]
32. Xu S, He Z, Li X, He J, Ni H, Ren D, Ren F, Li T, Chen G, Chen L, Chen J. Lymph Node Metastases in Surgically Resected Solitary Ground-Glass Opacities: A Two-Center Retrospective Cohort Study and Pooled Literature Analysis. Ann Surg Oncol 2023;30:3760-8. [Crossref] [PubMed]
33. Yip R, Li K, Liu L, Xu D, Tam K, Yankelevitz DF, Taioli E, Becker B, Henschke CI. Controversies on lung cancers manifesting as part-solid nodules. Eur Radiol 2018;28:747-59. [Crossref] [PubMed]
34. Li M, Xi J, Zhang H, Jin X, Fan Z, Zhan C, Feng M, Tan L, Wang Q. Ground glass nodules with scattered or eccentric island-shaped consolidations may have poor outcomes. Cancer Innov 2023;2:148-58. [Crossref] [PubMed]
35. Sun Q, Li P, Zhang J, Yip R, Yankelevitz DF, Henschke CI. CT Predictors of Angiolymphatic Invasion in Non-Small Cell Lung Cancer 30 mm or Smaller. Radiology 2025;314:e240599. [Crossref] [PubMed]
36. McWilliams A, Tammemagi MC, Mayo JR, Roberts H, Liu G, Soghrati K, et al. Probability of cancer in pulmonary nodules detected on first screening CT. N Engl J Med 2013;369:910-9. [Crossref] [PubMed]
37. Hattori A, Matsunaga T, Takamochi K, Oh S, Suzuki K. Neither Maximum Tumor Size nor Solid Component Size Is Prognostic in Part-Solid Lung Cancer: Impact of Tumor Size Should Be Applied Exclusively to Solid Lung Cancer. Ann Thorac Surg 2016;102:407-15. [Crossref] [PubMed]