Patient characteristics

There were 84 (58.3%) females and 60 (41.7%) males among the 144 patients, with the mean age of 59.18 ± 10.29 years (range, 32–78 years), and the mean tumor size of 11.95 ± 5.46 mm (range, 4.20-27.75 mm). We divided all patients into T1a (< 1 cm), cT1b (≥ 1 cm; < 2 cm), and cT1c (≥ 2 cm; ≤ 3 cm) according to the tumor size, the number of patients in each group was 52, 65, and 27, respectively. The percentage of SN in each group were 71.2%, 89.2%, and 88.9%, while the LN metastasis rates were 3.8%, 21.5%, and 40.7%, according to postoperative pathology, respectively. Table 1 summarizes the clinical data and pathological characteristics of 144 patients in each group. No significant differences were found in gender, family history of lung cancer, history of malignancy, smoking index, surgery, and location among the three groups (p > 0.05).

Table 1 Patient characteristicsComparison of the growth characteristics of pSN and SN

Table 2 summarizes the CT features and follow-up changes of SPN. We divided them into pSN and SN for comparison, according to the imaging staging of pulmonary nodules. Among 25 pSN, the mean size was 12.03 ± 6.35 mm with a mean follow-up time of 170.4 ± 91.1 days, while the mean size was 14.28 ± 5.91 mm with a mean follow-up time of 180.5 ± 92.7 days among 119 SN.

The changes in diameter, volume, and mass during the follow-up period of pSN and SN were statistically significant. The mean diameter, volume, and mass increases in the pSN and SN were 0.95 vs. 2.15 mm (p < 0.001), 195.31 vs. 696.25mm3 (p < 0.001), and 0.20 vs. 0.77 g (p < 0.001), respectively. The relative diameter, volume, and mass change during follow-up period were greater in the SN compared to pSN, 19.29% vs. 10.18% (p < 0.001), 45.85% vs. 17.12% (p < 0.001), and 51.78% vs. 22.71% (p < 0.001), respectively. 80 of all follow-up nodules exhibited growth, and we calculated volume doubling time of the whole nodule (VDTt) and mass doubling time of the whole nodule (MDTt) for these growing nodules. The VDTt for pSN and SN were 541 vs. 337 days (P = 0.005), while the MDTt was 458 vs. 298 days (p = 0.018), respectively (Fig. 2).

Comparison of the growth characteristics of the solid component of pSN and SN

Without considering the ground-glass component, we compared the growth difference between the solid component of pSN and the SN, SN showed faster growth compared to the solid component of pSN (Table 2). During the follow-up period, the mean diameter, volume and mass increase of the solid component of pSN and SN were 0.92 vs. 2.15 mm (p < 0.001), 58.23 vs. 696.25 mm3 (p = 0.002), and 0.45 vs. 0.77 g (p = 0.002), respectively. The relative diameter, volume and mass change of the solid component of pSN and SN were 19.29% vs. 15.72% (p = 0.260), 24.74% vs. 45.85% (p = 0.007), and 29.82% vs. 51.78% (p = 0.008), respectively, the relative solid component volume and mass change were statistically differences. Volume doubling time of the solid component (VDTc) and mass doubling time of the solid component (MDTc) were calculated for nodules exhibited growth, the VDTc for the solid component of pSN and SN was 496 vs. 337 days (P = 0.004), while the MDTc was 453 vs. 298 days (p = 0.003), respectively (Fig. 2).

Table 2 Growth features of pSN and SNFig. 2

(A-B) Box plot shows VDT and MDT for the whole nodule of PSN and SN. The median VDTt and MDTt of SN was significantly shorter than pSN. (C-D) Box plot shows the solid component of PSN and SN. The median VDT and MDT of SN was significantly shorter than the solid component of pSN. ∗ represent p < 0.05. ∗∗ represent p < 0.01. pSN, part-solid nodules. SN, solid nodules. VDTt, volume doubling time of the whole nodule. MDTt, mass doubling time of the whole nodule. VDTc, volume doubling time of the solid component. MDTc, mass doubling time of the solid component

Relationship between different pathological types of nodules and LN Metastasis

Among the nodules we followed up, 3 of adenocarcinoma in situ (AIS) and 19 of minimally invasive adenocarcinoma (MIA) were all N0, while in invasive adenocarcinoma (IA), there were 95 N0 and 27 N1/2. The percentage composition of different pathological types in LN metastasis was statistically different (p = 0.039), Pearson chi-square test analysis showed a correlation between pathological type and LN metastasis with a column coefficient of 0.20, but according to the results of the regression analysis, pathology type was not an independent risk factor for LN metastasis (Table 3). We investigating the VDTt and MDTt of different pathological types of nodules, and found that both AIS and MIA were N0, with the VDTt of 790 days and the MDTt of 695 days for AIS, 541 and 458 days for MIA, respectively. The growth of IA showed significant differences in LN metastasis, with the VDTt of 386 vs. 248 days for the N0 and N1/2, respectively (p = 0.019), and the MDTt of 348 vs. 228 days for the N0 and N1/2, respectively (p = 0.034) (Fig. 3).

Fig. 3

Graph shows the growth characteristics of the different pathological types of nodules in N0 and N1/2, where both AIS and MIA were N0. And both the mean VDTt and MDTt for IA were shorter in the N1/2 group than in N0, with the difference being statistically different (p < 0.05). AIS, adenocarcinoma in situ. MIA, minimally invasive adenocarcinoma. IA, invasive adenocarcinoma. VDTt, volume doubling time of the whole nodule. MDTt, mass doubling time of the whole nodule

Univariate and multivariate analysis for LN Metastasis

We divided all pulmonary nodules into two observation groups, N0 and N1/N2, according to the LN metastasis. Table 3 summarizes the patient characteristics and the results of the univariate analysis. Univariate analysis showed that gender (p = 0.04), smoking index (p = 0.037), nodule types (p = 0.008), pathological type (p = 0.039), initial diameter (p = 0.01), consolidation (p = 0.047), C/T ratio (p < 0.001), initial volume (p < 0.001), initial mass (p < 0.001), diameter increase (p < 0.001), consolidation increase (p < 0.001), volume increase (p < 0.001), mass increase (p < 0.001), relative diameter change (p < 0.001), relative volume change (p < 0.001), relative mass change (p < 0.001), VDTt (p = 0.001), and MDTt (p = 0.001) were all related to LN metastasis. On multivariate analysis, initial diameter (OR, 1.352; 95% CI, 1.141–1.584, p < 0.001), consolidation increase (OR, 0.543; 95% CI, 0.246–1.152, p = 0.019), volume increase (OR, 1.008; 95% CI, 1.001–1.030, p = 0.020), mass increase (OR, 1.019; 95% CI, 1.000-1.077, p = 0.021), VDTt (OR, 0.921; 95% CI, 0.878–0.965, p = 0.002), and MDTt (OR, 1.080; 95% CI, 1.017–1.134, p = 0.004) were independent factors for LNs metastasis (Table 4).

When drawing the ROC curve to evaluate the ability of VDTt and MDTt to predict the LN metastasis, the area under the curve (AUC) for VDTt was 0.860 (95% CI, 0.778–0.943; p < 0.001), if VDTt was used to predict the LN metastasis and assuming a threshold of 307 days for defining high risk for LN metastasis, with a sensitivity and specificity of 81.5% and 82.9%, respectively. The AUC for MDTt was 0.848 (95% CI, 0.759–0.936; p < 0.001), if MDTt was used to predict the LN metastasis and assuming a threshold of 254 days for defining high risk for LN metastasis, with a sensitivity and specificity of 77.8% and 84.6%, respectively (Fig. 4).

Table 3 Univariate analysis for metastasisTable 4 Multivariate analysis for LN metastasisFig. 4

Receiver-operating characteristic curves. Reference line: (AUC 0.5); VDTt: volume doubling time of the whole nodule (AUC 0.860); MDTt: mass doubling time of the whole nodule (AUC 0.848). AUC: area under the curve