Lesion-vessel spatial relationship: a significant indicator for differentiation and evaluation of pulmonary ground-glass nodules

Introduction

With the widespread use of computed tomography (CT), the detection rate of pulmonary ground-glass nodules (GGNs) has gradually increased (1-3). GGNs are diverse in nature and can be divided into neoplastic and non-neoplastic types. The clinical management strategies for neoplastic and non-neoplastic GGNs are significantly different. Neoplastic nodules typically necessitate regular monitoring or surgical resection, whereas non-neoplastic nodules generally do not require special intervention, with a substantial proportion potentially resolving during follow-up. (4). Therefore, accurately determining the nature of GGNs is crucial for their appropriate management. However, the similar imaging features of neoplastic and non-neoplastic GGNs complicate the differential diagnosis, underscoring the importance of precise evaluation (5,6).

Previous studies on differentiating neoplastic and non-neoplastic GGNs have revealed their distinct differences in morphological features, including size, density, shape, boundary, margin, pleural indentation, vacuole sign, and vascular changes (6-10). Since the lung architecture is not obscured by ground-glass opacity, the internal vessels of GGNs can be observed. In addition to the morphological features, vessel characteristics associated with GGNs have been found to be valuable for differential diagnosis. Moreover, it has been reported that abnormal vessel changes within GGNs, such as rigidity, distortion, and dilation, indicate a high likelihood of neoplastic lesions and are closely related to invasive adenocarcinoma (IAC) (11-14).

Additionally, studies have demonstrated that the vascular convergence sign is useful for distinguishing between benign and malignant pulmonary nodules (11,15). However, this sign is less frequently observed in GGNs and is primarily associated with solid nodules (16,17). Notably, not all vessels within neoplastic GGNs display such abnormalities, and many internal vessels remain normal. Furthermore, another study found that only the presence of internal vessels in GGNs did not correlate with the nature of the nodules (18). Therefore, the role of the lesion-vessel relationship in differentiating GGNs is still being investigated, and further studies are necessary.

The relationships between pulmonary vessels and GGNs can be clearly observed and assessed in CT images obtained with modern multidetector scanners (4). These relationships can be classified as either adherent or surrounded. Pulmonary vessels are encased by interstitium, and the vessel walls, along with the surrounding interstitium, may block the progression of pulmonary lesions (19). It is hypothesized that variations in vessel diameters result in differences in the blocking effect exerted by the vessel walls and surrounding interstitium. Consequently, the spatial relationship between lesions and vessels of varying diameters may potentially provide valuable information for differential diagnosis, especially for those GGNs lacking typical radiological features, thereby offering new insights for further management.

The purpose of this study was to investigate and characterize the relationship between lesions and vessels of different diameters for differentiating and evaluating neoplastic and non-neoplastic GGNs. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-427/rc).

Methods

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

Patients

A retrospective data collection was conducted for patients with GGNs who underwent CT examinations in The First Affiliated Hospital of Chongqing Medical University from January 2019 to September 2023 using the electronic medical record system and picture archiving and communications system (PACS). Neoplastic nodules were surgically resected and confirmed by pathological examination, while the non-neoplastic nodules were confirmed as benign by pathological examination, complete absorption, or a significant reduction in size or density during follow-up. The non-neoplastic GGNs confirmed by pathological examination were collected from January 2019 to January 2023, while those confirmed by follow-up were gathered from January 2019 to September 2023. Neoplastic GGNs were collected between January 2021 and June 2022. The inclusion criteria were as follows: (I) nodules appeared as GGNs on lung window images and (II) an interval between chest CT examination and surgery less than 2 weeks. Regarding the enrolled GGNs, the spatial relationship between lesions and pulmonary vessels was evaluated. Cases that did not exhibit a definite relationship with vessels (no vessels observed within the lesion or a certain distance between any part of the lesion and the adjacent vessels) were excluded. For the GGNs that had a relationship with vessels (lesions that appeared to be in contact with adjacent vessels or encircling vessels), the exclusion criteria were as follows: (I) lack of thin-section images with a thickness of ≤1.25 mm; (II) image artifacts that affected image analysis; and (III) incomplete clinical data for surgically resected patients. The neoplastic GGNs were matched with non-neoplastic GGNs based on the size range of the non-neoplastic ones (4–22 mm), resulting in 70 larger (>22 mm) neoplastic GGNs being excluded. This ensured that there was no significant size difference between the neoplastic and non-neoplastic GGNs because nodule size is related to the presence of vessels within and around it (18), and larger nodules have greater coverage and are more likely to adhere to or surround vessels. Finally, a total of 1,038 patients comprising 1,062 GGNs were enrolled in this study, including 528 patients comprising 531 neoplastic GGNs and 510 patients comprising 531 non-neoplastic GGNs. Figure 1 shows the flow of patient selection. Neoplastic GGNs included atypical adenomatous hyperplasia (AAH), adenocarcinoma in situ (AIS), minimally invasive adenocarcinoma (MIA), and IAC. Non-neoplastic GGNs included lesions confirmed by follow-up, nonspecific inflammation, focal fibrosis, tuberculosis, and focal alveolar hemorrhage.

Figure 1 Flowchart of patient selection. AAH, atypical adenomatous hyperplasia; AIS, adenocarcinoma in situ; CT, computed tomography; GGN, ground-glass nodule; IAC, invasive adenocarcinoma; MIA, minimally invasive adenocarcinoma.

CT examinations

All patient CT data analysis were derived from nonenhanced scans performed with the Discovery CT 750 HD (GE HealthCare, Chicago, IL, USA), SOMATOM Perspective (Siemens Healthineers, Erlangen, Germany), SOMATOM Force (Siemens Healthineers), and SOMATOM Definition Flash (Siemens Healthineers) CT scanner. Patients were placed in the supine position, with their upper limbs elevated above their heads, and instructed to hold their breath after deep inspiration. The scan range was from the thoracic entrance to the costophrenic angle. The scanning parameters were as follows: tube voltage, 110–130 kvp; tube current, 50–140 mAs (with automatic current modulation technology being used); slice thickness, 5 mm; rotation time, 0.5 s; pitch, 1–1.1; collimation, 0.6–0.625 mm; reconstruction slice thickness and interval, 0.625–1.25 mm; and matrix, 512×512.

Clinical data and image analysis

Clinical data, including the patient’s age, sex, smoking history, clinical symptoms (cough, expectoration, hemoptysis, chest pain, back pain, and fever), history of individual malignant tumor, history of family malignant tumor, and other concomitant lung diseases were recorded from the electronic medical record system. The CT data were analyzed with the Carestream Vue PACS (Philips Healthcare, Amsterdam, the Netherlands) under lung window settings [width, 1,200–1,600 Hounsfield unit (HU); level, −500 to −700 HU]. Two senior radiologists evaluated the GGNs on axial and multiplanar reconstruction (MPR) images without knowledge of the pathological results. Any discrepancies between the radiologists were resolved by consensus.

The following CT features of GGNs were assessed: (I) nodule size (average of the longest and perpendicular diameters on axial CT images); (II) location (upper lobe, middle lobe, or lower lobe); (III) shape [regular (round or oval) or irregular]; (IV) boundary (well-defined or ill-defined); (V) lobulation sign (yes or no); (VI) spiculation sign (yes or no); (VII) pleural indentation sign (yes or no); and (VIII) vacuole sign (yes or no). Lobulation was defined as an abrupt bulging of the contour of the lesion. Spiculation was defined as the presence of linear strands extending from the nodule margin into the lung parenchyma without reaching the pleural surface (20). Pleural indentation was defined as linear areas of high attenuation and originating from the nodules extending peripherally to contact the pleura (17). Vacuole sign was defined as the small spot area of prominent low attenuation similar to air attenuation within the lesion (21).

Based on axial images, the relationship between GGNs and vessels was initially evaluated. For GGNs suspected to be related to vessels, MPR was performed to obtain coronal and sagittal images, as well as images parallel and perpendicular to the long axis of the related vessels. This further clarified the relationship between GGNs and vessels, categorizing it into two conditions: adherence and surrounding. Adherence was defined as the edge of a GGN being in contact with the wall of a nearby vessel, while surrounding was defined as the GGN encircling the vessel. On the images parallel and perpendicular to the long axis of the vessels, adherence was considered to the vessel edge closely contacting the contour of the GGN, with the vessel wall tangent to the edge of the GGN; meanwhile, surrounding was considered to be the vessel passing through the GGN, with the vessel being completely surrounded by lesions. If the vessel was partially surrounded by the GGN and the extent of the encircled portion was less than half of the vessel, the relationship was classified as adherence; if the encircled portion was equal to or greater than half, it was classified as surrounding (Figure 2). The lesion-vessel spatial relationship was classified into three types: a GGN only adhering to vessels (type I), a GGN only surrounding the vessels (type II), and a GGN both adhering to and surrounding the vessels (type III). The vessel diameter was measured in the vertical plane of the long axis of the vessels. Anatomically, pulmonary vessels with a diameter of 1 mm are classified as distal small arteries or veins, located near the pre-capillary level, and these vessels have thin smooth muscle and reduced elastic fibers (22). In contrast, vessels that are 2 mm or larger are categorized as medium-sized pulmonary elastic arteries, or medium-to-high order branch pulmonary veins. These arteries are characterized by multilayer elastic fibers and thicker smooth muscle, while the veins contain more smooth muscle and collagen fibers (22,23). Consequently, 1 mm and 2 mm were designated as cutoff points for grouping vessel diameters, resulting in the following categories: (I) <1 mm; (II) ≥1 but <2 mm; and (III) ≥2 mm.

Figure 2 Classification and evaluation of lesion-vessel relationship. (A) GGN adhering to a vessel (red circle). (B) An axial image showing a 14.5-mm GGN in the right upper lobe, with a vessel (arrow) adjacent to it. (C) An MPR image showing a GGN in close contact with the long axis of this vessel (arrow). (D) A section perpendicular to the long axis of this vessel closely attached to the edge of the GGN (arrow). (E) GGN surrounding a vessel (red circle). (F) An axial image showing a 17.5-mm GGN in the right upper lobe, with a vessel (arrow) passing through it. (G) An MPR image of the long axis of the vessel showing the GGN surrounding this vessel (arrow) and its branches. (H) A section perpendicular to the long axis of this vessel completely encircled by the GGN (arrow). (I) GGN adhering to a vessel (red circle). (J) An axial image showing a 9-mm GGN in the left lower lobe, with a vessel (arrow) adjacent to this lesion. (K) An MPR image showing the GGN in close contact with the long axis of this vessel (arrow). (L) A section perpendicular to the long axis of this vessel (arrow) is partially embedded in the lesion, with the overlapped portion being less than half of the vessel (arrow). (M) GGN surrounding a vessel (red circle). (N) An axial image showing a 12.5-mm GGN in the right upper lobe, with a vessel (arrow) encircled by this lesion. (O) An MPR image showing the GGN surrounding the long axis of this vessel (arrow). (P) A section perpendicular to the long axis of this vessel (arrow) partially embedded in the lesion, with the overlapped portion being more than half of the vessel (arrow). GGN, ground-glass nodule; MPR, multi-planar reconstruction.

Statistical analysis

Continuous variables are expressed as the mean ± standard deviation, while categorical variables are expressed as numbers and percentages. For comparing variables between neoplastic and non-neoplastic GGNs, the Mann-Whitney test was used for patients’ age and GGN size, whereas the Pearson Chi-squared test or Fisher’s exact test was employed for patients’ gender, GGN distribution, frequencies of different CT features, and lesion-vessel relationships. One-way analysis of variance was performed to compare invasiveness scores in neoplastic GGNs. Variables with statistically significant differences in the univariate analysis were further included in multivariate logistic regression analysis to identify independent factors. The final logistic regression model was created via the enter method. The robustness and generalization of the model were evaluated via the fivefold cross-validation. Receiver operating characteristic (ROC) analysis was performed to evaluate the discrimination efficiency of the logistic regression model, and the DeLong test was used to compare the performance of the ROC curves. Interreader agreement for lesion-vessel relationships, lesion-vessel spatial patterns, and classification of vessel diameter was evaluated via Cohen’s kappa coefficient, which was categorized as slight (κ≤0.20), fair (κ=0.21−0.40), moderate (κ=0.41−0.60), substantial (κ=0.61−0.80), or almost perfect (κ≥0.81). All statistical analyses were performed with SPSS version 26.0 (IBM Corp., Armonk, NY, USA), R 3.2.5 software (The R Foundation for Statistical Computing), and MedCalc version 20.0 (MedCalc Software, Ostend, Belgium), with a P value of <0.05 being considered statistically significant.

Results

Patients’ clinical characteristics and CT features of GGNs

Among the 531 non-neoplastic GGNs in 510 patients, 320 (60.3%) were pathologically confirmed after surgical resection, while 211 (39.7%) showed complete absorption or a significant reduction in size or density during follow-up. The 531 neoplastic GGNs in 528 patients comprised 15 AAHs, 168 AISs, 198 MIAs, and 150 IACs.

The patients’ clinical characteristics and CT features of GGNs are listed in Table 1. Compared to patients with non-neoplastic GGNs, the proportion of females and nonsmokers was higher among patients with neoplastic GGNs (females: 71.8% vs. 55.9%; nonsmokers: 75.6% vs. 82.2%; both P values <0.05). Additionally, neoplastic GGNs were more frequently located in the upper lobes and exhibited CT features including well-defined boundary, lobulation sign, pleural indentation sign, and vacuole sign (all P values <0.001). Multivariate analysis of clinical and CT indicators revealed that the independent risk factors for predicting neoplastic GGNs were female sex [odds ratio (OR) 1.675; 95% confidence interval (CI): 1.134−2.420; P=0.009], well-defined boundary (OR 3.164; 95% CI: 2.313−4.328; P<0.001), lobulation sign (OR 1.785; 95% CI: 1.239−2.571; P=0.002), and vacuole sign (OR 3.837; 95% CI: 2.033−7.244; P<0.001). The ROC analysis of this regression model 1 (prediction model developed using clinical and CT features.) yielded an AUC of 0.728 (sensitivity, 62.9%; specificity, 72.2%; 95% CI: 0.693−0.762; P<0.001).

Spatial relationship between GGNs and pulmonary vessels

Agreement was substantial for the classification of lesion-vessel relationships (κ=0.873; 95% CI: 0.847–0.898), lesion-vessel spatial patterns (κ=0.862; 95% CI: 0.835–0.889), and the diameter of vessels (κ=0.859; 95% CI: 0.833–0.884). Regarding the spatial relationship between GGNs and vessels, type I, II, and III were identified in 346 (65.2%), 127 (23.9%), and 58 (10.9%) non-neoplastic GGNs, respectively, while in neoplastic GGNs, these types were found in 79 (14.9%), 408 (76.8%), and 44 (8.3%), respectively. Type I was significantly more common in non-neoplastic GGNs (P<0.001), while type II was predominantly found in neoplastic GGNs (P<0.001).

Among the 404 non-neoplastic GGNs adhering to vessels (type I and type III), there were 110 (25.0%), 271 (67.1%), and 98 (24.3%) adhering to vessels with diameters of <1, ≥1 but <2, and ≥2 mm, respectively; meanwhile, among the 123 neoplastic GGNs, there were 33 (26.8%), 91 (74.0%), and 3 (2.4%), respectively. Compared to neoplastic GGNs, non-neoplastic ones were significantly more likely to adhere to vessels with a diameter of ≥2 mm (P<0.001) (Figure 3).

Figure 3 A 47-year-old female with a non-neoplastic GGN located in the right upper lobe. An axial (A) image revealed a vessel (arrow) adjacent to the GGN. Oblique axial (B) and oblique sagittal (C) images parallel to and perpendicular to the long axis of this vessel (arrow) confirmed that the lesion adhered to the vessel with a diameter of 2.5 mm. After 5 months, it was completely absorbed (D). GGN, ground-glass nodule.

Among the GGNs that surrounded vessels (type II and type III), those surrounding vessels with diameters of <1, ≥1 but <2, and ≥2 mm were observed in 111 (60.0%), 81 (43.8%), and 8 (4.3%) of the 185 non-neoplastic GGNs, respectively; meanwhile, among 452 neoplastic GGNs, the corresponding numbers were 167 (36.9%), 317 (70.1%), and 106 (23.5%), respectively. Compared to non-neoplastic GGNs, neoplastic ones were significantly more likely to surround vessels with a diameter of ≥1 but <2 mm (P<0.001) and those with a diameter of ≥2 mm (P<0.001). In contrast, lesions surrounding vessels with a diameter of <1 mm were more commonly observed in non-neoplastic GGNs than in neoplastic ones (P<0.001).

Among the 58 non-neoplastic GGNs and 44 neoplastic GGNs that surrounded and adhered to vessels (type III) (Figure S1), those adhering to vessels with a diameter of ≥2 mm and vessels with a diameter of <2 mm were more common in non-neoplastic GGNs than in neoplastic ones (15.6% vs. 2.3%; P=0.040); meanwhile, those adhering to vessels with a diameter of <2 mm and those surrounding vessels with a diameter of ≥1 mm were significantly more common in neoplastic GGNs than in non-neoplastic ones (79.5% vs. 46.6%; P=0.001).

Multivariable logistic regression for predicting GGNs

Multivariable logistic regression analysis was performed on lesion-vessel relationships with statistical significance in the univariate analysis (Table 2). This logistic regression model showed that the independent indicators for neoplastic GGNs were a type II lesion-vessel relationship (OR 3.747; 95% CI: 2.321−6.049; P<0.001), surrounding vessels with a diameter ≥1 but <2 mm (OR 2.671; 95% CI: 1.841−3.875; P<0.001), and surrounding vessels with a diameter ≥2 mm (OR 5.918; 95% CI: 2.683−13.053; P<0.001); meanwhile, independent predictor of non-neoplastic GGNs was adherence to vessels with a diameter of ≥2 mm (OR 11.868; 95% CI: 3.643–38.668; P<0.001). The ROC analysis of this model 2 (prediction model developed using vascular features only) yielded an AUC of 0.825 (sensitivity, 83.6%; specificity, 71.0%; 95% CI: 0.800−0.849; P<0.001). Model 2 demonstrated significantly superior predictive efficiency compared to the model 1, which relied solely on traditional CT features of lesions (P<0.001).

Table 3 shows the multivariable logistic regression analysis of CT features and the lesion-vessel relationship for identifying independent radiological indicators neoplastic and non-neoplastic GGNs. The independent indicators for neoplastic GGNs were well-defined boundary (OR 4.550; 95% CI: 3.063−6.759; P<0.001), vacuole sign (OR 2.907; 95% CI: 1.361−6.210; P=0.006), type II lesion-vessel relationship (OR 5.011; 95% CI: 2.753−9.124; P <0.001), surrounding vessels with a diameter ≥1 but <2 mm (OR 2.497; 95% CI: 1.519−4.106; P<0.001), and surrounding vessels with a diameter ≥2 mm (OR 8.375; 95% CI: 2.574−27.244; P<0.001); meanwhile, the independent predictor of non-neoplastic GGNs was adherence to vessels with a diameter of ≥2 mm (OR 13.082; 95% CI: 3.826−44.737; P<0.001). The AUC for this model was 0.878 (sensitivity, 80.6%; specificity, 79.7%; 95% CI: 0.855−0.901; P<0.001) (Figure 4). Internal validation via fivefold cross-validation was used for further confirming the robustness and generalization of this model and yielded an AUC of 0.868 (95% CI: 0.832–0.903; standard deviation across folds, 0.047). The predictive efficiency of model 3 (prediction model combined vascular indicators and clinical and CT features) was also significantly higher than that of model 1 (P<0.001).

Figure 4 ROC curves showing the performance of model 1, model 2, and model 3 in distinguishing neoplastic and non-neoplastic GGNs. AUCs for model 1, model 2, and model 3 were 0.728 (95% CI: 0.693–0.762; P<0.001), 0.825 (95% CI: 0.800–0.849; P<0.001), and 0.878 (95% CI: 0.855–0.901; P<0.001), respectively. Model 1: clinical and CT features; model 2: vascular indicators only; model 3: combination of vascular indicators and clinical and CT features. AUC, area under the curve; CI, confidence interval; GGN, ground-glass nodule; ROC, receiver operating characteristic.

The surrounded vessels in different subtypes of neoplastic GGNs

Among the 531 neoplastic GGNs that surrounded vessels (type II and type III), which included 137 AAHs/AISs, 167 MIAs, and 139 IACs, the proportions of lesions surrounding vessels with diameters of ≥1 but <2 mm significantly increased with greater invasiveness (from 49.6% to 72.6%; P=0.014), as did the proportion of lesions surrounding vessels with diameters of ≥2 mm (from 5.8% to 36.0%; P<0.001) (Figure 5). Conversely, the proportion of GGNs surrounding vessels with a diameter of <1 mm was significantly decreased with greater invasiveness (from 56.2% to 17.9%; P<0.001). The invasiveness score based on the diameter of surrounding vessel was significantly higher in AAH/AIS than in IAC (P<0.001) (Table 4, Figure S2).

Figure 5 Patients with neoplastic GGNs surrounding vessels of varying diameters. (A-C) A 67-year-old female with an 11.5-mm GGN located in right upper lobe. Axial (A), oblique coronal (B), and oblique axial (C) images parallel and perpendicular to the long axis of the related vessel indicated that the GGN surrounded a vessel (arrows) with a diameter of 1.2 mm. After operation, it was confirmed as AIS. (D-F) A 69-year-old female with a 14-mm GGN located in right upper lobe. Axial (D), oblique coronal (E), and oblique sagittal (F) images parallel and perpendicular to the long axis of the related vessel indicated that the GGN surrounded a vessel (arrows) with a diameter of 1.7 mm. After operation, it was confirmed as MIA. (G-I) A 54-year-old female with a 13-mm GGN located in left upper lobe. Axial (G), oblique axial (H), and oblique sagittal (I) images parallel and perpendicular to the long axis of the related vessel indicated that the GGN surrounded a vessel (arrows) with a diameter of 3.2 mm and its branches. After operation, it was confirmed as IAC. AIS, adenocarcinoma in situ; GGN, ground-glass nodule; IAC, invasive adenocarcinoma; MIA, minimally invasive adenocarcinoma.

Patients’ clinical characteristics and CT features of non-neoplastic GGNs

The patients’ clinical characteristics and CT features of non-neoplastic GGNs are summarized in Table 5. Compared to the pathologically confirmed GGNs, those confirmed by follow-up were smaller in size, more frequently located in the lower lobes, and had an ill-defined boundary but exhibited less spiculation and lobulation signs (all P values <0.05).

There was no significant difference in different types of lesion-vessel relationships between pathologically confirmed non-neoplastic GGNs and those confirmed by follow-up (P=0.210). Among the 250 pathologically confirmed non-neoplastic GGNs (group 1) and 157 non-neoplastic GGNs confirmed by follow-up (group 2) that adhered to vessels, cases adhering to vessels with a diameter of ≥1 mm but <2 mm were more frequently found in group 1 (73.2% vs. 56.1%; P<0.001); meanwhile, those adhering to vessels with a diameter of <1 mm were more common in group 2 (34.4% vs. 22.4%; P=0.008). There was no significant difference in lesions adhering to vessels with a diameter of ≥2 mm between these two groups (26.0% vs. 21.0%; P=0.253). In addition, there was no significant difference in the surrounding vessels with different diameters between the 114 pathologically confirmed non-neoplastic GGNs and the 71 non-neoplastic GGNs confirmed by follow-up.

Discussion

In this study, we comprehensively evaluated the role of the spatial relationship between GGNs and vessels of varying diameters in differentiating neoplastic from non-neoplastic GGNs. The findings indicate that neoplastic GGNs primarily surround vessels, while non-neoplastic GGNs are more likely to adhere to them. Furthermore, non-neoplastic GGNs frequently exhibited lesions adhering to vessels with a diameter of ≥2 mm, whereas neoplastic ones were more common among lesions surrounding vessels with a diameter of ≥1 mm. These vascular indicators were identified as independent factors for differentiating neoplastic from non-neoplastic GGNs; they exhibited a higher diagnostic efficiency and can be combined with traditional morphological features to enhance their differential diagnostic efficiency. Additionally, the likelihood of surrounding larger vessels increased with greater neoplastic invasiveness. Our findings confirm lesion-vessel spatial relationship as a crucial indicator that should be considered in the differentiation and evaluation of GGNs.

Regarding larger vessels with a diameter of ≥2 mm, non-neoplastic GGNs tend to adhere to them, while neoplastic GGNs are more likely to surround them. This difference is likely attributed to the structural characteristics of the vessels. As vessel diameter increases, the vessel walls and surrounding interstitium become thicker, resulting in a more pronounced blocking effect on lesions (19,22,24). Consequently, non-neoplastic GGNs are confined to one side of the vessel and exhibit a tendency to adhere to it. Additionally, non-neoplastic GGNs are often associated with slight and localized inflammation (19), rendering them more susceptible to obstruction by larger vessels. In contrast, neoplastic GGNs display an invasive growth pattern that allows them to grow without being obstructed by pulmonary vessels and the surrounding interstitium (25,26). This leads to the gradual involvement of lung tissue around the vessels, ultimately resulting in neoplastic GGNs surrounding the vessels.

Neoplastic GGNs surround vessels with a diameter of <1 mm less frequently than do non-neoplastic GGNs, while they more commonly surround vessels with a diameter of ≥1 but <2 mm. This difference may be attributed to the higher density of neoplastic GGNs (27), where the internal solid components may obscure small vessels, making them difficult to identify on CT images. Furthermore, previous studies have indicated that vessels exhibiting abnormal changes are more prevalent in neoplastic GGNs. The high metabolic activity of neoplasms and their tendency to infiltrate the bronchovascular bundle or the interlobular septa may contribute to vessel thickening (17,28,29). This suggests that vessels with an initial diameter of <1 mm may thicken to a range of 1 mm to 2 mm in neoplastic GGNs. Additionally, the blocking effect of small vessels, especially those with a diameter <1 mm, is relatively weak, allowing non-neoplastic GGNs to easily surround them. Consequently, the differences between neoplastic and non-neoplastic GGNs in relation to the surrounded vessels with a diameter of <1 mm may lack specificity for differentiation.

For type III GGNs, those adhering to vessels with a diameter ≥2 mm and surrounding vessels with a diameter <2 mm were more commonly associated with non-neoplastic GGNs. In contrast, GGNs surrounding vessels with a diameter ≥1 mm and adhering to vessels with a diameter <2 mm were more frequently observed in neoplastic GGNs. These differences may serve as valuable references for differentiating GGNs that both surround and adhere to vessels. Specifically, GGNs that adhere to vessels with a diameter of ≥2 mm should primarily be classified as non-neoplastic. Conversely, in the absence of this condition, GGNs that surround vessels of ≥1 mm should be considered neoplastic. Therefore, the diameter of vessels adhering to lesions should be considered in diagnosing type III GGNs.

Among neoplastic GGNs that surround vessels, the probability of surrounding vessels with diameters of ≥1 mm but <2 mm and ≥2 mm significantly increases as the invasiveness of the lesions rises. This pattern is likely attributable to the stronger infiltration associated with higher invasiveness, leading to a greater incidence of larger diameter vessels being surrounded. Therefore, the diameter of the surrounded vessels may provide valuable information for assessing the invasiveness of neoplastic GGNs. GGNs that surround vessels with a diameter of ≥1 mm, particularly those with a diameter ≥2 mm, should be considered to have a high likelihood of being IAC. This finding may serve as an additional indicator for evaluating the invasiveness of neoplastic GGNs.

In non-neoplastic GGNs, those confirmed by follow-up tended to be smaller, more frequently located in the lower lobe, and have an ill-defined boundary and were less likely to exhibit spiculation sign or lobulation sign as compared to the pathologically confirmed ones. However, there was no significant difference in the types of lesion-vessel relationships between the two groups. Among the GGNs that adhered to vessels, those confirmed by follow-up were more likely to adhere to vessels with a diameter of <1 mm than were the pathologically confirmed ones. In contrast, lesions adhering to vessels with a diameter of ≥1 but <2 mm were more commonly found in the pathologically confirmed GGNs. This may be attributed to the relatively smaller size of follow-up-confirmed GGNs, indicating very mild infiltration, which limits their opportunity to contact larger vessels and increases the likelihood of being obstructed by small vessels. Therefore, non-neoplastic GGNs that adhere to vessels <1 mm may suggest mild inflammation and could interpreted as a sign of absorption during follow-up.

Previous studies have reported that patients with neoplastic GGNs are primarily females and nonsmokers (27,29). For CT features, neoplastic GGNs are predominantly located in the upper lobe and are more frequently associated with higher CT attenuation, larger size, lobulation, well-defined boundary, and vacuole sign; meanwhile, non-neoplastic GGNs tend to exhibit ill-defined boundary, coarse margin, irregular shape, and pleural thickening (30,31). These findings are consistent with our study’s results. In addition, several studies have indicated that GGNs exhibiting abnormal vascular changes, such as rigidity, distortion, and dilation, can aid in differential diagnosis (13,14,32,33). Gao et al. (11) investigated the relationship between GGNs and vascular morphological changes. Their findings indicated that vascular distortion, dilation, or convergence in GGNs were strong indicators of malignancy; however, these signs were observed in only 40.78% of malignant GGNs. In our study, we further classified and assessed the spatial relationships between lesions and vessels and found that 76% of neoplastic GGNs exhibited a type II relationship, which was more prevalent. Additionally, the integration of spatial relationships and vessel diameter proved to be significantly beneficial for differential diagnosis, enhancing the evaluation of vascular changes, and improving the diagnostic efficiency of traditional morphological features, particularly for GGNs lacking abnormal vascular changes. Furthermore, several studies have shown that computer-aided detection can serve as a valuable supplement to visual reading for the detection and diagnosis of pulmonary nodules (34-36). Consequently, incorporating the information of the lesion-vessel spatial relationship into artificial intelligence software and structured reporting system for lung nodule assessment may provide an additional critical reference in the differentiation of GGNs.

The study involved several limitations that should be mentioned. First, we employed a retrospective design, and thus further evaluation and validation of the diagnostic value of the findings are warranted. Second, the use of single center might have introduced selection bias. Although the cross-validation supports the model’s internal validity, the findings should be confirmed in multicenter cohorts. Third, although this study assessed the relationships between GGNs and vessels, it did not conduct a comprehensive investigation of vascular morphological changes or vessel types. Fourth, we did not account for the location of lesions within the lung field, but the diameter of vessels typically varies across the inner, middle, and outer zones. Finally, we employed a range of classifications rather than relying on absolute values to evaluate vessel diameters. This approach helped to minimize the impact of measurement variability and enhanced clinical applicability; however, manual measurement errors and observer variability might have hindered the accurate classification of vessels, particularly since most of them were thin. Integrating artificial intelligence software for more precise vessel diameter measurements may eliminate these influences in future studies.

Conclusions

There are differences in the spatial relationships between neoplastic and non-neoplastic GGNs and pulmonary vessels of varying diameters. Neoplastic GGNs tend to primarily surround vessels, whereas non-neoplastic GGNs are more likely to adhere to them. Specifically, GGNs adhering to vessels, especially those that adhere to vessels with a diameter of ≥2 mm, are likely to be non-neoplastic, while those surrounding vessels, particularly those surrounding vessels with a diameter of ≥1 mm, are more likely to be neoplastic. Furthermore, neoplastic GGNs that surround larger vessels are associated with higher invasiveness. These lesion-vessel relationships should be integrated with other CT features to provide comprehensive indicators for a more accurate differential diagnosis of GGNs.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-427/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-427/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-2025-427/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 Ethics Committee 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. Park CM, Goo JM, Lee HJ, Lee CH, Chun EJ, Im JG. Nodular ground-glass opacity at thin-section CT: histologic correlation and evaluation of change at follow-up. Radiographics 2007;27:391-408. [Crossref] [PubMed]
2. Ten Haaf K, Jeon J, Tammemägi MC, Han SS, Kong CY, Plevritis SK, Feuer EJ, de Koning HJ, Steyerberg EW, Meza R. Risk prediction models for selection of lung cancer screening candidates: A retrospective validation study. PLoS Med 2017;14:e1002277. [Crossref] [PubMed]
3. Primack SL, Müller NL, Mayo JR, Remy-Jardin M, Remy J. Pulmonary parenchymal abnormalities of vascular origin: high-resolution CT findings. Radiographics 1994;14:739-46. [Crossref] [PubMed]
4. Mazzone PJ, Lam L. Evaluating the Patient With a Pulmonary Nodule: A Review. JAMA 2022;327:264-73. [Crossref] [PubMed]
5. Naidich DP, Bankier AA, MacMahon H, Schaefer-Prokop CM, Pistolesi M, Goo JM, Macchiarini P, Crapo JD, Herold CJ, Austin JH, Travis WD. Recommendations for the management of subsolid pulmonary nodules detected at CT: a statement from the Fleischner Society. Radiology 2013;266:304-17. [Crossref] [PubMed]
6. Li Q. Recent progress in computer-aided diagnosis of lung nodules on thin-section CT. Comput Med Imaging Graph 2007;31:248-57. [Crossref] [PubMed]
7. Chu ZG, Li WJ, Fu BJ, Lv FJ. CT Characteristics for Predicting Invasiveness in Pulmonary Pure Ground-Glass Nodules. AJR Am J Roentgenol 2020;215:351-8. [Crossref] [PubMed]
8. Chang YC, Chen PT, Hsieh MS, Huang YS, Ko WC, Lin MW, Hsu HH, Chen JS, Chang YC. Discrimination of invasive lung adenocarcinoma from Lung-RADS category 2 nonsolid nodules through visual assessment: a retrospective study. Eur Radiol 2024;34:3453-61. [Crossref] [PubMed]
9. Zhang Y, Shen Y, Qiang JW, Ye JD, Zhang J, Zhao RY. HRCT features distinguishing pre-invasive from invasive pulmonary adenocarcinomas appearing as ground-glass nodules. Eur Radiol 2016;26:2921-8. [Crossref] [PubMed]
10. Niu R, Shao X, Shao X, Wang J, Jiang Z, Wang Y. Lung Adenocarcinoma Manifesting as Ground-Glass Opacity Nodules 3 cm or Smaller: Evaluation With Combined High-Resolution CT and PET/CT Modality. AJR Am J Roentgenol 2019;213:W236-45. [Crossref] [PubMed]
11. Gao F, Li M, Ge X, Zheng X, Ren Q, Chen Y, Lv F, Hua Y. Multi-detector spiral CT study of the relationships between pulmonary ground-glass nodules and blood vessels. Eur Radiol 2013;23:3271-7. [Crossref] [PubMed]
12. Zhang C, Li J, Sun M, Li S, Li J, Li Q, Zhu Z. Peripheral vessel and air bronchograms for detecting the pathologic patterns of subsolid nodules. Clin Imaging 2019;56:63-8. [Crossref] [PubMed]
13. Meng Q, Ren P, Gao P, Dou X, Chen X, Guo L, Song Y. Effectiveness and Feasibility of Complementary Lung-RADS Version 1.1 in Risk Stratification for pGGN in LDCT Lung Cancer Screening in a Chinese Population. Cancer Manag Res 2020;12:189-98. [Crossref] [PubMed]
14. Liu J, Yang X, Li Y, Xu H, He C, Qing H, Ren J, Zhou P. Development and validation of qualitative and quantitative models to predict invasiveness of lung adenocarcinomas manifesting as pure ground-glass nodules based on low-dose computed tomography during lung cancer screening. Quant Imaging Med Surg 2022;12:2917-31. [Crossref] [PubMed]
15. Gao F, Sun Y, Zhang G, Zheng X, Li M, Hua Y. CT characterization of different pathological types of subcentimeter pulmonary ground-glass nodular lesions. Br J Radiol 2019;92:20180204. [Crossref] [PubMed]
16. Takashima S, Maruyama Y, Hasegawa M, Yamanda T, Honda T, Kadoya M, Sone S. CT findings and progression of small peripheral lung neoplasms having a replacement growth pattern. AJR Am J Roentgenol 2003;180:817-26. [Crossref] [PubMed]
17. Zwirewich CV, Vedal S, Miller RR, Müller NL. Solitary pulmonary nodule: high-resolution CT and radiologic-pathologic correlation. Radiology 1991;179:469-76. [Crossref] [PubMed]
18. Fu BJ, Lv FJ, Li WJ, Lin RY, Zheng YN, Chu ZG. Significance of intra-nodular vessel sign in differentiating benign and malignant pulmonary ground-glass nodules. Insights Imaging 2021;12:65. [Crossref] [PubMed]
19. Webb WR. Thin-section CT of the secondary pulmonary lobule: anatomy and the image--the 2004 Fleischner lecture. Radiology 2006;239:322-38. [Crossref] [PubMed]
20. Xu DM, van Klaveren RJ, de Bock GH, Leusveld A, Zhao Y, Wang Y, Vliegenthart R, de Koning HJ, Scholten ET, Verschakelen J, Prokop M, Oudkerk M. Limited value of shape, margin and CT density in the discrimination between benign and malignant screen detected solid pulmonary nodules of the NELSON trial. Eur J Radiol 2008;68:347-52. [Crossref] [PubMed]
21. Lee HJ, Kim YT, Kang CH, Zhao B, Tan Y, Schwartz LH, Persigehl T, Jeon YK, Chung DH. Epidermal growth factor receptor mutation in lung adenocarcinomas: relationship with CT characteristics and histologic subtypes. Radiology 2013;268:254-64. [Crossref] [PubMed]
22. Townsley MI. Structure and composition of pulmonary arteries, capillaries, and veins. Compr Physiol 2012;2:675-709. [Crossref] [PubMed]
23. Horsfield K, Gordon WI. Morphometry of pulmonary veins in man. Lung 1981;159:211-8. [Crossref] [PubMed]
24. Le L, Narula N, Zhou F, Smereka P, Ordner J, Theise N, Moore WH, Girvin F, Azour L, Moreira AL, Naidich DP, Ko JP. Diseases Involving the Lung Peribronchovascular Region: A CT Imaging Pathologic Classification. Chest 2024;166:802-20. [Crossref] [PubMed]
25. Yang Z, Sone S, Takashima S, Li F, Honda T, Yamanda T. Small peripheral carcinomas of the lung: thin-section CT and pathologic correlation. Eur Radiol 1999;9:1819-25. [Crossref] [PubMed]
26. Chu ZG, Yang ZG, Shao H, Zhu ZY, Deng W, Tang SS, Chen J, Li Y. Small peripheral lung adenocarcinoma: CT and histopathologic characteristics and prognostic implications. Cancer Imaging 2011;11:237-46. [Crossref] [PubMed]
27. He XQ, Li X, Wu Y, Wu S, Luo TY, Lv FJ, Li Q. Differential Diagnosis of Nonabsorbable Inflammatory and Malignant Subsolid Nodules with a Solid Component ≤5 mm. J Inflamm Res 2022;15:1785-96. [Crossref] [PubMed]
28. Noguchi M, Morikawa A, Kawasaki M, Matsuno Y, Yamada T, Hirohashi S, Kondo H, Shimosato Y. Small adenocarcinoma of the lung. Histologic characteristics and prognosis. Cancer 1995;75:2844-52. [Crossref] [PubMed]
29. Kobayashi Y, Ambrogio C, Mitsudomi T. Ground-glass nodules of the lung in never-smokers and smokers: clinical and genetic insights. Transl Lung Cancer Res 2018;7:487-97. [Crossref] [PubMed]
30. Hu H, Wang Q, Tang H, Xiong L, Lin Q. Multi-slice computed tomography characteristics of solitary pulmonary ground-glass nodules: Differences between malignant and benign. Thorac Cancer 2016;7:80-7. [Crossref] [PubMed]
31. Fu BJ, Zhang XC, Lv FJ, Chu ZG. Potential Role of Intrapulmonary Concomitant Lesions in Differentiating Non-Neoplastic and Neoplastic Ground Glass Nodules. J Inflamm Res 2023;16:6155-66. [Crossref] [PubMed]
32. Xue W, Kong L, Zhang X, Xin Z, Zhao Q, He J, Wu W, Duan G. Tumor blood vessel in 3D reconstruction CT imaging as an risk indicator for growth of pulmonary nodule with ground-glass opacity. J Cardiothorac Surg 2023;18:333. [Crossref] [PubMed]
33. Hsu WC, Huang PC, Pan KT, Chuang WY, Wu CY, Wong HF, Yang CT, Wan YL. Predictors of Invasive Adenocarcinomas among Pure Ground-Glass Nodules Less Than 2 cm in Diameter. Cancers (Basel) 2021;13:3945. [Crossref] [PubMed]
34. Silva M, Milanese G, Ledda RE, Pastorino U, Sverzellati N. Screen-detected solid nodules: from detection of nodule to structured reporting. Transl Lung Cancer Res 2021;10:2335-46. [Crossref] [PubMed]
35. Chen X, Feng B, Chen Y, Liu K, Li K, Duan X, Hao Y, Cui E, Liu Z, Zhang C, Long W, Liu X. A CT-based radiomics nomogram for prediction of lung adenocarcinomas and granulomatous lesions in patient with solitary sub-centimeter solid nodules. Cancer Imaging 2020;20:45. [Crossref] [PubMed]
36. Silva M, Schaefer-Prokop CM, Jacobs C, Capretti G, Ciompi F, van Ginneken B, Pastorino U, Sverzellati N. Detection of Subsolid Nodules in Lung Cancer Screening: Complementary Sensitivity of Visual Reading and Computer-Aided Diagnosis. Invest Radiol 2018;53:441-9. [Crossref] [PubMed]