Abstract
Objective:
To differentiate pre-invasive lesion from invasive pulmonary adenocarcinoma (IPA) appearing as ground-glass nodules (GGNs) using CT features.
Methods:
149 GGNs were enrolled in this study, with 74 pure GGNs (p-GGNs) and 75 mixed GGNs (m-GGNs). Firstly, univariate analysis was used to analyse the difference of CT features between pre-invasive lesion and IPA. Then, multivariate analysis was conducted to identify variables that could independently differentiate pre-invasive lesion from IPA. Receiver operating characteristic curve analysis was performed to evaluate the differentiating value of identified variables.
Results:
In the p-GGNs, multivariate analysis showed that the amount of blood vessels was an independent risk factor. Using the amount of blood vessels “≥1” as the diagnostic criterion, we could diagnose IPA with a sensitivity of 100%. Using the amount of blood vessels “=0” as the diagnostic criterion, we could diagnose pre-invasive lesions with a specificity of 100%. In the m-GGNs, multivariate analysis showed that the volume of solid portion (VSolid) and pleural indentation were two independent risk factors. One further model was constructed using these two variables: model = 2.508 × (VSolid + 1.407) × (pleural indentation − 1.016). Using the new model, improved diagnostic ability was achieved compared with using VSolid or pleural indentation alone.
Conclusion:
The amount of blood vessels through the p-GGNs would be an important criterion during clinical management, while VSolid and pleural indentation seemed important for m-GGNs. Moreover, the new model could further improve the differentiating value for m-GGNs.
Advances in knowledge:
CT features are useful in differentiating pre-invasive lesion from IPA appearing as GGNs.
INTRODUCTION
According to the new pathological classification constituted in 2011, lung adenocarcinoma was divided into the pre-invasive lesion group and the invasive pulmonary adenocarcinoma (IPA) group.1,2 Significant difference existed between these two groups regarding the surgical method and the scope of lymph node dissection.2–5 Sublobar resection could be acceptable for pre-invasive lesion, while the standard surgical treatment for IPA should be lobectomy.2–5 Skip metastases involving mediastinal lymph nodes, without hilar lymph nodes appeared mostly in the IPA group, thus the scope of lymph node dissection for the IPA group should be larger than for the pre-invasive lesion group.2–5 Therefore, accurate differentiation between pre-invasive and IPA lesions before surgery was crucial, particularly for surgery planning, prognosis assessment and doctor–patient communication.
CT is the most commonly used imaging modality for detection, differentiation and follow-up of lung ground-glass nodules (GGNs).6–9 A previous study by Lee et al10 had shown that CT features could help differentiate pre-invasive lesion from IPA; however, they did not adopt pulmonary GGN-related vessels on CT images as an independent evaluation factor, while several studies showed that vasculature remodelling or neovascularization was one of the initiating events occurring in the early stages of tumour development.11,12 Meanwhile, a previous study by Gao et al13 demonstrated that different relationships between GGNs and blood vessels could be found within different kinds of GGNs owing to variations in developmental biology and behaviour. Therefore, the relationships between pulmonary GGNs and blood vessels might also be important to differentiate pre-invasive lesions from IPA.
Therefore, in this study, we tried to differentiate pre-invasive lesion from IPA using multimodal CT features, especially combined utilization of the relationships between pulmonary GGNs and blood vessels.
METHODS AND MATERIALS
Patient selection
From January 2010 to May 2014, a search of the radiology information systems of our hospital (the First Affiliated Hospital of Nanjing Medical University) by one radiologist for patients with pulmonary GGNs identified on chest CT scans was performed. Our study population was selected in the following steps: first, we selected all CT scans for which the reports included the words “GGN”, “part-solid nodule”, “nonsolid nodule”, “ground-glass opacity”, or “ground-glass nodule”. A total of 9431 CT scans in 7952 patients were found. Of the 9431 CT scans, only 5493 scans in 4890 patients with thin-section images (section thickness ≤1.5 mm) were included. Second, two radiologists reviewed all of the CT scans and excluded diffuse GGNs, very small GGNs <5 mm, large lesions >3 cm in size and transient GGNs, leaving 176 patients with 197 GGNs. Among the 197 GGNs, an additional 48 GGNs that were not pathologically confirmed were excluded. Finally, 135 patients with 149 pathologically proved GGNs constituted our study population (Figure 1). The 135 patients consisted of 42 male and 93 female patients, with a mean age of 59.28 ± 10.14 years.
Figure 1.
Flowchart of study population. Numbers in parentheses are the number of patients. GGN, ground-glass nodule; GGO, ground-glass opacity.
Histological evaluation
All resected specimens were formalin fixed and stained with haematoxylin–eosin in accordance with the routine regulations of our hospital. One board-certified pathologist (with 10 years' experience performing pathological diagnosis of lung cancer) reviewed specimens and recorded the pathological subtype of each tumour according to the International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society classification of lung adenocarcinomas.1,2
According to the new adenocarcinoma classification and pathological findings, GGNs were divided into two groups: the pre-invasive lesion group and the IPA group.1,2 Pre-invasive lesions included atypical adenomatous hyperplasia (AAH) and adenocarcinoma in situ (AIS), which were defined as lesions showing no stromal, vascular or pleural invasion.1,2 IPA, which was defined as adenocarcinoma containing an invasive component, included minimally invasive adenocarcinoma and invasive adenocarcinoma with several subtypes such as lepidic predominant, acinar predominant or papillary predominant adenocarcinoma.1,2 In our study, all histological subtypes were included in our study. No histological subtypes were excluded from the study.
CT imaging
Chest CT scans were performed using SOMATOM® Definition-64 or SOMATOM Emotion-16 (Siemens Healthcare, Erlangen, Germany). CT scans were obtained in all patients in the supine position at full inspiration. CT parameters were as follows: tube voltage, 120 kVp; tube current, 100–300 mA; beam pitch, 0.875–1.500; detector collimation, 1.0–2.5 mm; rotation time, 0.5–1.0 s and a reconstruction kernel with a high-frequency algorithm. Reconstruction thicknesses and intervals were 1.5 and 1.5 mm. All images were reviewed with the standard lung window (window width, 1200 HU; window level, −600 HU) and mediastinal window (window width, 400 HU; window level, 50 HU).
Imaging assessment
GGNs are defined as hazily increased attenuation in the lung with preservation of intact bronchial and vascular structures.14–16 In our study, GGNs were divided into the mixed GGNs (m-GGNs) group and pure GGNs (p-GGNs) group according to whether an internal solid component existed or not.10,14
The CT features that were analysed for each GGN included: (1) volume of GGNs (VGGN), (2) volume of solid portion (VSolid), (3) solid proportion (VSolid/VGGN), (4) margin (spiculation, non-spiculation), (5) border (lobulation, non-lobulation), (6) the presence of bubble lucency, (7) the presence of pleural retraction, (8) the presence of air bronchogram, (9) amount of blood vessels through the GGNs. All CT findings were evaluated based on high-resolution CT (HRCT) images.
The volume of GGNs was measured in Siemens' volume measurement software using interactive model. In the interactive model, we set the thickness as 1.5 mm; the upper limit of assessment as 1000 HU; and the lower limit of assessment as −1000 HU. The lesions in the same section were delineated by two authors with consensus by using an operator-defined region of interest (ROI) on each HRCT axial image. The volume was obtained by multiplying the lesions areas by the thicknesses and interval. The volume measurement of the solid portion was similar to the volume measurement of GGNs.
Blood vessel analysis was performed in terms of the amount of blood vessels through the GGNs. The amount of blood vessels was divided into zero, 1, 2, 3 and ≥4. “0” meant no vessel passed through GGNs (Figure 2). “1” meant one blood vessel passed through GGNs and so on (Figures 3–5). “≥4” meant that four blood vessels or more passed through GGNs (Figure 6). The amount of blood vessels was analysed based on axial images, multiplanar reformation images, maximum-intensity projection images and volume-rendering technique images.
Figure 2.
(a) CT image of pure ground-glass nodules (GGNs) in the right upper lobe of a 33-year-old male. (b, c) Maximum-intensity projection image and volume-rendering technique image show that no vessel passed through the GGNs. (d) Pathological results after wedge resection confirmed the diagnosis of adenocarcinoma in situ.
Figure 3.
(a) CT image of pure ground-glass nodules (GGNs) in the right middle lobe of a 58-year-old male. (b, c) Maximum-intensity projection and volume-rendering technique images show that one blood vessel passed through the GGNs (arrows). (d) Pathological results after lobectomy confirmed the diagnosis of minimally invasive adenocarcinoma.
Figure 5.
(a) CT image of mixed ground-glass nodules (GGNs) with pleural indentation and lobulation in the right upper lobe of a 65-year-old female. (b, c) Maximum-intensity projection and volume-rendering technique images show that three blood vessels passed through the GGNs (arrows). (d) Pathological results after lobectomy confirmed the diagnosis of invasive pulmonary adenocarcinoma.
Figure 6.
(a) CT image of mixed ground-glass nodules (GGNs) with lobulation in the right upper lobe of a 73-year-old male. (b, c) Four blood vessels were found through the GGNs based on maximum-intensity projection and volume-rendering technique images (arrows). (d) Pathological results after lobectomy confirmed the diagnosis of invasive pulmonary adenocarcinoma.
Figure 4.
(a) CT image of mixed ground-glass nodules (GGNs) with pleural indentation and lobulation in the right upper lobe of a 71-year-old male. (b, c) Two blood vessels were found through the GGNs based on the maximum-intensity projection and volume-rendering technique images (arrows). (d) Pathological results after lobectomy confirmed the diagnosis of invasive pulmonary adenocarcinoma.
Imaging analysis was conducted independently by two experienced thoracic radiologists (with 10 and 9 years' experience performing chest image interpretations, respectively) who were blinded to the pathological results. Discrepancies in interpretation between observers were resolved by consensus.
Statistical analysis
Statistical analysis was performed separately for m-GGN and p-GGN groups owing to the different malignancy potentialities of these two groups.10,14 For each group, independent sample t-test was used to statistically analyse the difference between pre-invasive lesions and IPA, regarding VGGN, VSolid, solid proportion and the amount of blood vessels. χ2 test was used to demonstrate the difference regarding CT morphological features including margin, border, bubble lucency, pleural retraction and air bronchogram. To determine the magnitude of colinearity of the independent variables, Pearson correlation was used. Variables with a p-value of <0.05 at univariate analysis were used as the input variables for multiple logistic regression analysis. Multiple logistic regression analysis was conducted to identify variables that could independently differentiate pre-invasive lesion from IPA. In multiple logistic regression analysis, a backwards stepwise selection mode was used. The removal of variables was based on likelihood ratio statistics with a probability of 0.10. Receiver operating characteristic (ROC) curve analysis was calculated to evaluate the diagnostic value of identified variables in differentiating pre-invasive lesion from IPA.
RESULTS
Among all the 135 patients, a total of 149 GGNs were detected, with 74 p-GGNs and 75 m-GGNs. The p-GGN group included AAH (n = 26), AIS (n = 30) and IPA (n = 18). The m-GGN group included AAH (n = 4), AIS (n = 14) and IPA (n = 57).
THE PURE GROUND-GLASS NODULES GROUP
Univariate analysis of pure ground-glass nodules
Among CT findings, statistical differences between pre-invasive lesion and IPA were found in the VGGN, the border (lobulation), pleural indentation and the amount of blood vessels through the GGNs (p < 0.05). Statistical differences between the two groups were not found in other CT morphological features (bubble-like lucency, spiculation, air bronchogram) (p ≥ 0.05) (Tables 1 and 2).
Table 1.
CT findings of pre-invasive lesion and invasive pulmonary adenocarcinoma (IPA) in pure-ground-glass nodules (GGNs)
| Feature | Pre-invasive lesion (n = 56) | IPA (n = 18) | p-value |
|---|---|---|---|
| VGGN (cm3) | 1.31 ± 1.43 | 2.81 ± 2.17 | 0.002a |
| Lobulated (±) | 22/34 | 14/4 | 0.004a |
| Spiculated (±) | 7/49 | 2/16 | 1.0 |
| Bubble-like lucency (±) | 15/41 | 7/11 | 0.328 |
| Air bronchogram (±) | 14/42 | 7/11 | 0.256 |
| Pleural indentation (±) | 13/43 | 9/9 | 0.031a |
VGGN, volume of GGNs.
Quantitative variables are expressed as the means ± standard deviation. Categorical variables are expressed as numbers.
Significant difference (p < 0.05).
Table 2.
The amount of blood vessels of pre-invasive lesion and invasive pulmonary adenocarcinoma (IPA) in pure ground-glass nodules
| Amount of blood vessels | Pre-invasive lesion (n = 56) | IPA (n = 18) | p-value |
|---|---|---|---|
| 0 | 17 | 0 | 0.001a |
| 1 | 23 | 8 | |
| 2 | 15 | 5 | |
| 3 | 1 | 3 | |
| ≥4 | 0 | 2 |
Significant difference (p < 0.05).
Logistic regression analysis of pure ground-glass nodules
Pearson correlation analysis results demonstrated that there was a significant positive correlation between the vessel number and lesion size (p < 0.05), but the correlation coefficient was weak (r = +0.539). The existing weak correlation between the vessel number and lesion size would not have any influence on further multivariate logistic regression. Therefore, VGGN, lobulation, pleural indentation and the amount of blood vessels were used as input variables for the multivariate logistic regression analysis. Multivariate logistic regression analysis revealed that only the amount of blood vessels was a significant independent risk factor in differentiating pre-invasive lesion from IPA (p < 0.05). The adjusted odds ratio for the amount of blood vessel was 3.13.
Diagnostic value of the amount of blood vessels in differentiating pre-invasive lesion from IPA
ROC analysis showed that the area under the curve (AUC) for the amount of blood vessels was 0.738 [95% confidence interval (CI), 0.612–0.864]. The sensitivity and specificity in differentiating pre-invasive lesions from IPA were 100% and 30.4%, respectively. Using the amount of blood vessels “≥1” as the diagnostic criterion, we could diagnose IPA with a sensitivity of 100%. Using the amount of blood vessels “=0” as the diagnostic criterion, we could diagnose pre-invasive lesions with a specificity of 100%.
THE MIXED GROUND-GLASS NODULES GROUP
Univariate analysis of mixed ground-glass nodules
Among CT findings, statistical differences between pre-invasive lesion and IPA were found in VGGN, VSolid, pleural indentation and the amount of blood vessels (p < 0.05). Statistical differences were not found in the other CT morphological features (bubble-like lucency, spiculation, air bronchogram, loblation) and solid proportion (p ≥ 0.05) (Tables 3 and 4).
Table 3.
CT findings of pre-invasive lesion and invasive pulmonary adenocarcinoma (IPA) in mixed ground-glass nodules (GGNs)
| Feature | Pre-invasive lesion (n = 18) | IPA (n = 57) | p-value |
|---|---|---|---|
| VGGN (cm3) | 2.19 ± 1.33 | 5.12 ± 8.03 | 0.008a |
| VSolid (cm3) | 0.38 ± 0.26 | 1.28 ± 2.32 | 0.002a |
| Solid proportion (%) | 22.97 ± 18.31 | 23.32 ± 12.08 | 0.339 |
| Lobulated (±) | 14/4 | 47/10 | 0.923 |
| Spiculated (±) | 6/12 | 24/33 | 0.508 |
| Bubble-like lucency (±) | 8/10 | 17/40 | 0.251 |
| Air bronchogram (±) | 9/9 | 31/26 | 0.745 |
| Pleural indentation (±) | 7/11 | 43/14 | 0.004a |
VGGN, volume of GGNs; VSolid, volume of solid portion.
Quantitative variables are expressed as the means ± standard deviation. Categorical variables are expressed as numbers.
Significant difference (p < 0.05).
Table 4.
The amount of blood vessels of pre-invasive lesion and invasive pulmonary adenocarcinoma (IPA) in mixed ground-glass nodules
| Amount of blood vessels | Pre-invasive lesion (n = 18) | IPA (n = 57) | p-value |
|---|---|---|---|
| 0 | 1 | 0 | 0.034a |
| 1 | 5 | 13 | |
| 2 | 9 | 18 | |
| 3 | 3 | 17 | |
| ≥4 | 0 | 9 |
Significant difference (p < 0.05).
Logistic regression analysis of mixed ground-glass nodules
Pearson correlation analysis results demonstrated that there was a significant positive correlation between the amount of blood vessels and VGGN (p < 0.05), but the correlation coefficient was weak (r = + 0.291). Pearson correlation analysis results demonstrated that the amount of blood vessels had no association with VSolid (p > 0.05). Therefore, VGGN, VSolid, pleural indentation and the amount of blood vessels were used as input variables for multivariate logistic regression analysis. After multivariate stepwise logistic regression, VSolid and pleural indentation were found as independent risk factors in differentiating pre-invasive lesions from IPA (p < 0.05). Then, a model was constructed using these two variables and the formula of this model calculated as follows: model = 2.508 × (VSolid + 1.407) × (pleural indentation − 1.016) (VSolid expressed as cm3) (Table 5).
Table 5.
Multivariate logistic regression analysis of independent risk factors in differentiating pre-invasive lesion from invasive pulmonary adenocarcinoma in the mixed ground-glass nodules
| Variable | Coefficient | Odds ratio | 95% confidence interval | p-value |
|---|---|---|---|---|
| VSolid | 2.508 | 12.285 | 1.641–91.649 | 0.014a |
| Pleural retraction | 1.407 | 4.084 | 1.214–13.734 | 0.023a |
| Constant | −1.016 | 0.362 | 0.000–0.000 | 0.122 |
VSolid, volume of solid portion.
Significant difference (p < 0.05).
Diagnostic value of VSolid or pleural indentation alone at univariate analysis in differentiating pre-invasive lesion from IPA
ROC analysis showed that AUC for VSolid was 0.746 (95% CI: 0.631–0.862). Using the optimal cut-off value of VSolid = 0.66 cm3 as the diagnostic criterion to differentiate pre-invasive lesion from IPA, a sensitivity of 56% and a specificity of 89% could be acquired, respectively (Figure 7).
Figure 7.
Receiver operating characteristic curve result of using new diagnostic model, VSolid and pleural indentation to differentiate pre-invasive lesions from invasive pulmonary adenocarcinoma. Compared with the diagnostic value of VSolid or pleural indentation alone, improved value was achieved using our new diagnostic model (area under the curve, 0.794 vs 0.746 and 0.683).
Using pleural retraction as the diagnostic criterion to differentiate pre-invasive lesion from IPA, an AUC of 0.683 (95% CI: 0.535–0.830), a sensitivity of 75% and a specificity of 61% could be acquired, respectively (Figure 7).
Diagnostic value of the new diagnostic model (combination of pleural indentation and VSolid) in differentiating pre-invasive lesion from IPA
ROC analysis showed that AUC for the new diagnostic model (combination of pleural indentation and VSolid) was 0.794 (95% CI: 0.688–0.900) (Figure 7). Using this new diagnostic model, the sensitivity and specificity in differentiating pre-invasive lesions from IPA were 47.4% and 100%, respectively. Statistical difference between the new diagnostic model and pleural indentation alone was found (p < 0.05). The statistical difference between the new diagnostic model and VSolid alone was not found (p ≥ 0.05).
DISCUSSION
Pre-invasive lesions and IPA commonly demonstrate different biological behaviour, such as the mediastinal lymph node involvement. In our study cohort, the rate of mediastinal lymph node involvement was 12.8% (9/70) in the IPA group, which is significantly higher than in the pre-invasive lesion group, 1.5% (1/65); both are similar to the results from previous studies.2–5 Different biological behaviour means different treatment plan, therefore, to find a simple and objective method to accurately differentiate pre-invasive lesions and IPA before surgery is very important for the treatment plan decision. So, we tried to use the imaging features on pre-surgery CT images to predict whether GGNs are IPA or not. As a result, we firstly found that in the p-GGNs group, the amount of blood vessels was a significant independent risk factor in differentiating pulmonary lesions of these two groups. Using the amount of blood vessels “≥1” as the diagnostic criterion, we could diagnose IPA with a sensitivity of 100%. Using the amount of blood vessels “=0” as the diagnostic criterion, we could diagnose pre-invasive lesions with a specificity of 100%. Secondly, in the m-GGNs group, VSolid and pleural indentation were independent risk factors in differentiating two kinds of lesions. Using our newly established diagnostic model that fully considered VSolid and pleural indentation, we could get a high diagnostic specificity even to 100%.
For the p-GGNs, the amount of blood vessel was a significant independent risk factor in differentiating pre-invasive lesion from IPA. Using the amount of blood vessels “≥1” as diagnostic criterion, we could diagnose IPA with a sensitivity of 100%. Meanwhile, using the amount of blood vessels “=0” as diagnostic criterion, we could diagnose pre-invasive lesions with a specificity of 100%. Our research results suggested that the amount of blood vessel could assist in differentiating the grade of tumour, which is similar to the viewpoint of a previous study.13 Previously, Gao et al13 subdivided the relationships between GGNs and vessels into four types, including: Type I, vessels passing by GGNs; Type II, intact vessels passing through GGNs; Type III, distorted, dilated or tortuous vessels seen within GGNs; Type IV, more complicated vasculature other than those described above. Strong correlation between invasive adenocarcinoma and Type III and IV relationships was found in their study. However, their image assessment focused on the location and morphological change of the vessels, which was an absolutely subjective and qualitative process. Thus, considering the repeatability of image assessment, we transformed the subjective and qualitative image analysis to objective and quantitative measurement of the amount of related vessels; furthermore, we also found a strong correlation between invasive GGNs and increased amount of related vessels. On why increased amounts of the vessels could be an indicator of invasive GGNs, in our opinion, the potential mechanism might include the following two aspects: firstly, the foci of fibrosis would increase when the invasive grade of tumour increased. More fibrosis found in the tumours would further lead to vascular convergence around the tumour, which were recognized on CT as vascular convergence.17–19 Secondly, invasive tumour mostly demonstrated higher metabolism than pre-invasive tumour. The essential pre-condition of high metabolism should be enough blood supply, which could only be achieved by vessel proliferation.
Actually in our opinion, our finding that increased numbers of tumour-related vessels could be an indicator of IPA fit with the concept of neovascularization. Angiogenesis is one of the hallmarks of cancer.20,21 Tumours above a critical size of 2–3 mm3 are strongly dependent on the active recruitment of new blood vessels.22,23 A previous study showed that the extent of enhancement at dynamic enhanced CT reflected the underlying nodular angiogenesis, and significant positive correlations were found between the extent of peak enhancement and microvessel density and histological grading.24 Similar results have also been reported between standardized uptake value and microvessel density and histological grading.25 Therefore, CT and positron emission tomography (PET)/CT may play an important role in evaluating the neovascularization of pulmonary nodules. However, a comparative study on the value of routine CT scan with dynamic CT or PET-CT in evaluating the neovascularization of pulmonary nodules is still lacking.
For the m-GGNs, pleural indentation and the VSolid were two independent risk factors for IPA. Central fibrosis and resultant tissue contraction could cause fibrotic strands around the tumour, which were recognized on CT as pleural indentations.26 As we mentioned before, the foci of fibrosis would increase when the invasive grade of tumour increased. This mechanism explained that pleural indentation significantly increased as histological grades advanced, which is similar with those of Takashima et al.26 Meanwhile, we found that larger VSolid might be an indicator for IPA for m-GGNs, which was similar to the previous studies of Lee et al10 and Tsutani et al.27 Tsutani et al27 proposed that the risk of IPA increased when the diameter of the solid region exceeded 8 mm. However, both of them used the two-dimensional measurement of the size of the solid region, whereas we used the volume of the solid region as the risk variable for analysis. Previous studies have showed that three-dimensional volume measurements could evaluate the size and growth of tumours better than the conventional two-dimensional diameter measurement; meanwhile, better interobserver reproducibility could be achieved during three-dimensional volume measurements.28–30 Furthermore in our study, we established a new diagnostic model that combined pleural indentation and VSolid. Using the new model calculated as 2.508 × (VSolid + 1.407) × (pleural indentation − 1.016), the diagnostic value could be further improved compared with using pleural indentation and VSolid alone (AUC, 0.794 for new model vs 0.746 for VSolid and 0.683 for pleural indentation), especially the diagnostic specificity (specificity, 100% for new model vs 89% for VSolid and 61% for pleural indentation). The high diagnostic value, especially the high specificity meant that the GGNs would indeed be the invasive lesions once it met our newly proposed diagnostic model, which was crucial for surgery planning, prognosis assessment.
Our study had several limitations. Firstly, potential selection bias might exist owing to the retrospective design of our study. Secondly, further subdivision of pulmonary arteries or veins involved was not conducted during the assessment of the amount of blood vessels. Previous studies indicated that lung cancer should be suspected if pulmonary veins are involved with solid pulmonary nodules, and open biopsy needs to be performed even when the cytological and histological findings are negative.31–33 Therefore, further subdivision of types of vessels when analysing the diagnostic value of the amount of blood vessels would be significant. Thirdly, the number of cases in this study was limited, especially IPA lesions in the p-GGNs group and pre-invasive lesions in the m-GGNs group, which may compromise the diagnostic power. Further studies with larger sample numbers would be desired to verify the research results of our study.
In conclusion, the amount of blood vessels through the GGNs would be the most important criterion during clinical management of p-GGNs. When the amount of vascular convergence increased, the possibility of IPA also increased. Meanwhile, VSolid and pleural indentation would be the important criterion during clinical management of m-GGNs. Moreover, the combination diagnosis of VSolid and pleural indentation can improve diagnostic accuracy and specificity.
FUNDING
This article was funded by Tong-fu Yu.
Contributor Information
J Liang, Email: liangjingxyz@163.com.
X-Q Xu, Email: njmu-xuxiaoquan@163.com.
H Xu, Email: njmu-xuhai@163.com.
M Yuan, Email: njmu-yuanmei@163.com.
W Zhang, Email: njmu-zhangwei@163.com.
Z-F Shi, Email: njmu-shizhaofei@163.com.
T-F Yu, Email: njmu_yutongfu@163.com.
REFERENCES
- 1.Travis WD, Brambilla E, Noguchi M, Nicholson AG, Geisinger KR, Yatabe Y, et al. International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society international multidisciplinary classification of lung adenocarcinoma. J Thorac Oncol 2011; 6: 244–85. doi: 10.1097/JTO.0b013e318206a221 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Van Schil PE, Asamura H, Rusch VW, Mitsudomi T, Tsuboi M, Brambilla E, et al. Surgical implications of the new IASLC/ATS/ERS adenocarcinoma classification. Eur Respir J 2012; 39: 478–86. doi: 10.1183/09031936.00027511 [DOI] [PubMed] [Google Scholar]
- 3.Graham AN, Chan KJ, Pastorino U, Goldstraw P. Systematic nodal dissection in the intrathoracic staging of patients with non-small cell lung cancer. J Thorac Cardiovasc Surg 1999; 117: 246–51. doi: 10.1016/S0022-5223(99)70419-8 [DOI] [PubMed] [Google Scholar]
- 4.Ohde Y. The proportion of consolidation to ground-glass opacity on high resolution CT is a good predictor for distinguishing the population of non-invasive peripheral adenocarcinoma. Lung Cancer 2003; 42: 303–10. doi: 10.1016/j.lungcan.2003.07.001 [DOI] [PubMed] [Google Scholar]
- 5.Fukui T, Katayama T, Ito S, Abe T, Hatooka S, Mitsudomi T. Clinicopathological features of small-sized non-small cell lung cancer with mediastinal lymph node metastasis. Lung Cancer 2009; 66: 309–13. doi: 10.1016/j.lungcan.2009.03.004 [DOI] [PubMed] [Google Scholar]
- 6.Aoki T, Tomoda Y, Watanabe H, Nakata H, Kasai T, Hashimoto H, et al. Peripheral lung adenocarcinoma: correlation of thin-section CT findings with histologic prognostic factors and survival. Radiology 2001; 220: 803–9. doi: 10.1148/radiol.2203001701 [DOI] [PubMed] [Google Scholar]
- 7.Takahashi S, Tanaka N, Okimoto T, Tanaka T, Ueda K, Matsumoto T, et al. Long term follow-up for small pure ground-glass nodules: implications of determining an optimum follow-up period and high-resolution CT findings to predict the growth of nodules. Jpn J Radiol 2012; 30: 206–17. doi: 10.1007/s11604-011-0033-8 [DOI] [PubMed] [Google Scholar]
- 8.Yamagami T, Yoshimatsu R, Miura H, Yamada K, Takahata A, Matsumoto T, et al. Diagnostic performance of percutaneous lung biopsy using automated biopsy needles under CT-fluoroscopic guidance for ground-glass opacity lesions. Br J Radiol 2013; 86: 20120447. doi: 10.1259/bjr.20120447 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Stigt JA, Uil SM, Oostdijk AH, Boers JE, van den Berg JW, Groen HJ. A diagnostic program for patients suspected of having lung cancer. Clin Lung Cancer 2012; 13: 475–81. doi: 10.1016/j.cllc.2012.03.001 [DOI] [PubMed] [Google Scholar]
- 10.Lee SM, Park CM, Goo JM, Lee HJ, Wi JY, Kang CH. Invasive pulmonary adenocarcinomas versus preinvasive lesions appearing as ground-glass nodules: differentiation by using CT features. Radiology 2013; 268: 265–73. doi: 10.1148/radiol.13120949 [DOI] [PubMed] [Google Scholar]
- 11.Uramoto H, Yamada S, Tanaka F. Angiogenesis of lung cancer utilizes existing blood vessels rather than developing new vessels using signals from carcinogenesis. Anticancer Res 2013; 33: 1913–16. [PubMed] [Google Scholar]
- 12.Jia WR, Chai WM, Tang L, Wang Y, Fei XC, Han BS, et al. Three-dimensional contrast enhanced ultrasound score and dynamic contrast-enhanced magnetic resonance imaging score in evaluating breast tumor angiogenesis: correlation with biological factors. Eur J Radiol 2014; 83: 1098–105. doi: 10.1016/j.ejrad.2014.03.027 [DOI] [PubMed] [Google Scholar]
- 13.Gao F, Li M, Ge X, Zheng X, Ren Q, Chen Y, et al. Multi-detector spiral CT study of the relationships between pulmonary ground-glass nodules and blood vessels. Eur Radiol 2013; 23: 3271–7. doi: 10.1007/s00330-013-2954-3 [DOI] [PubMed] [Google Scholar]
- 14.Henschke CI, Yankelevitz DF, Mirtcheva R, McGuinness G, McCauley D, Miettinen OS; ELCAP Group. CT screening for lung cancer: frequency and significance of part-solid and nonsolid nodules. AJR Am J Roentgenol 2002; 178: 1053–7. doi: 10.2214/ajr.178.5.1781053 [DOI] [PubMed] [Google Scholar]
- 15.Kim HY, Shim YM, Lee KS, Han J, Yi CA, Kim YK. Persistent pulmonary nodular ground-glass opacity at thin-section CT: histopathologic comparisons. Radiology 2007; 245: 267–75. doi: 10.1148/radiol.2451061682 [DOI] [PubMed] [Google Scholar]
- 16.Ko JP. Lung nodule detection and characterization with multi-slice CT. J Thorac Imaging 2005; 20: 196–209. doi: 10.1097/01.rti.0000171625.92574.8d [DOI] [PubMed] [Google Scholar]
- 17.Noguchi M, Morikawa A, Kawasaki M, Matsuno Y, Yamada T, Hirohashi S, et al. Small adenocarcinoma of the lung. Histologic characteristics and prognosis. Cancer 1995; 75: 2844–52. doi: [DOI] [PubMed] [Google Scholar]
- 18.Zwirewich CV, Vedal S, Miller RR, Müller NL. Solitary pulmonary nodule: high-resolution CT and radiologic-pathologic correlation. Radiology 1991; 179: 469–76. doi: 10.1148/radiology.179.2.2014294 [DOI] [PubMed] [Google Scholar]
- 19.Zhang L, Yankelevitz DF, Carter D, Henschke CI, Yip R, Reeves AP. Internal growth of nonsolid lung nodules: radiologic-pathologic correlation. Radiology 2012; 263: 279–86. doi: 10.1148/radiol.11101372 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Folkman J. Angiogenesis in cancer, vascular, rheumatoid and other disease. Nat Med 1995; 1: 27–31. doi: 10.1038/nm0195-27 [DOI] [PubMed] [Google Scholar]
- 21.Carmeliet P, Jain RK. Molecular mechanisms and clinical applications of angiogenesis. Nature 2011; 473: 298–307. doi: 10.1038/nature10144 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 2011; 144: 646–74. doi: 10.1016/j.cell.2011.02.013 [DOI] [PubMed] [Google Scholar]
- 23.Bergers G, Benjamin LE. Tumorigenesis and the angiogenic switch. Nat Rev Cancer 2003; 3: 401–10. doi: 10.1038/nrc1093 [DOI] [PubMed] [Google Scholar]
- 24.Yi CA, Lee KS, Kim EA, Han J, Kim H, Kwon OJ, et al. Solitary pulmonary nodules: dynamic enhanced multi-detector row CT study and comparison with vascular endothelial growth factor and microvessel density. Radiology 2004; 233: 191–9. doi: 10.1148/radiol.2331031535 [DOI] [PubMed] [Google Scholar]
- 25.Cappabianca S, Porto A, Petrillo M, Greco B, Reginelli A, Ronza F, et al. Preliminary study on the correlation between grading and histology of solitary pulmonary nodules and contrast enhancement and [18F]fluorodeoxyglucose standardised uptake value after evaluation by dynamic multiphase CT and PET/CT. J Clin Pathol 2011; 64: 114–19. doi: 10.1136/jcp.2010.076562 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Takashima S, Maruyama Y, Hasegawa M, Yamanda T, Honda T, Kadoya M, et al. CT findings and progression of small peripheral lung neoplasms having a replacement growth pattern. AJR Am J Roentgenol 2003; 180: 817–26. doi: 10.2214/ajr.180.3.1800817 [DOI] [PubMed] [Google Scholar]
- 27.Tsutani Y, Miyata Y, Nakayama H, Okumura S, Adachi S, Yoshimura M, et al. Prediction of pathologic node-negative clinical stage IA lung adenocarcinoma for optimal candidates undergoing sublobar resection. J Thorac Cardiovasc Surg 2012; 144: 1365–71. doi: 10.1016/j.jtcvs.2012.07.012 [DOI] [PubMed] [Google Scholar]
- 28.Koike W, Iwano S, Matsuo K, Kitano M, Kawakami K, Naganawa S. Doubling time calculations for lung cancer by three-dimensional computer-aided volumetry: effects of inter-observer differences and nodule characteristics. J Med Imaging Radiat Oncol 2014; 58: 82–8. doi: 10.1111/1754-9485.12128 [DOI] [PubMed] [Google Scholar]
- 29.Revel MP, Bissery A, Bienvenu M, Aycard L, Lefort C, Frija G. Are two-dimensional CT measurements of small noncalcified pulmonary nodules reliable? Radiology 2004; 231: 453–8. doi: 10.1148/radiol.2312030167 [DOI] [PubMed] [Google Scholar]
- 30.Moss M, Wellman DA, Cotsonis GA. An appraisal of multivariable logistic models in the pulmonary and critical care literature. Chest 2003; 123: 923–8. doi: 10.1378/chest.123.3.923 [DOI] [PubMed] [Google Scholar]
- 31.Mori K, Saitou Y, Tominaga K, Yokoi K, Miyazawa N, Okuyama A, et al. Small nodular lesions in the lung periphery: new approach to diagnosis with CT. Radiology 1990; 177: 843–9. doi: 10.1148/radiology.177.3.2243999 [DOI] [PubMed] [Google Scholar]
- 32.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. doi: 10.1007/s003300050929 [DOI] [PubMed] [Google Scholar]
- 33.Kakinuma R, Ohmatsu H, Kaneko M, Eguchi K, Naruke T, Nagai K, et al. Detection failures in spiral CT screening for lung cancer: analysis of CT findings. Radiology 1999; 212: 61–6. doi: 10.1148/radiology.212.1.r99jn1461 [DOI] [PubMed] [Google Scholar]