Clinicopathological and Prognostic Disparities in Part‐Solid Versus Pure Solid Pulmonary Nodules as Early‐Stage Lung Adenocarcinoma

ABSTRACT

Purpose

The optimal surgical strategy for early‐stage lung adenocarcinoma appearing as part‐solid nodules (PSN) versus pure solid nodules (SPN) remains debated. This study directly compares their clinicopathological features, surgical outcomes, and recurrence patterns to inform personalized management based on CT characteristics.

Methods

This retrospective study analyzed 710 surgical patients with clinical stage IA lung adenocarcinoma. Nodules were categorized as PSN or SPN by the presence of a GGO component on preoperative CT. Interaction‐effect analysis assessed CT‐attenuation dynamics in solid components. Propensity‐score matching balanced intergroup confounders, and Cox regression identified recurrence risk factors for each type.

Results

Mean CT‐solid value (−13.2 Hu vs. 22.2 Hu, p < 0.001) and the trend of change in CT‐solid value during follow‐up (Interaction effect Index: −0.011078, p < 0.001) were significantly different between PSN and SPN. Upon histological examination, PSN had significantly lower incidences of visceral pleural invasion (11.7% vs. 30.4%, p < 0.001) and lymph node metastasis (11.2% vs. 20.6%, p = 0.008) than SPN. After propensity‐score matching, patients with PSN had significantly better recurrence‐free survival (86.8% vs. 70.7%, p < 0.001) than those with SPN. In Cox regression analysis, sublobar resection and lymph node sampling‐only were associated with increased risk of loco‐regional recurrence in SPN, but not in PSN.

Conclusion

PSN and SPN show distinct clinicopathological features and prognostic profiles, suggesting potentially different biological behaviors. CT features need to be incorporated with clinical tumor stage in deciding surgical strategies on resection extent and lymph node examination for early‐stage lung adenocarcinoma.

For clinical stage IA lung adenocarcinoma, part‐solid nodules (PSN) show lower CT‐solid value (−13.2 HU), less lymph node metastasis (11.2%), and better 5‐year recurrence‐free survival (87.6%) than pure solid nodules (SPN; 22.2 HU, 20.6%, 79.8%). Sublobar resection with limited nodal sampling appears safe for PSN, whereas SPN may require lobectomy and systematic nodal dissection to reduce local recurrence. CT morphology should be integrated with clinical T stage to guide individualized surgical strategies.

1. Introduction

The application of Computed Tomography (CT) screening has led to early detection of lung cancers, especially those appearing as small, ground glass opacity (GGO)‐containing nodules [1]. Since the 8th TNM classification for Non‐small Cell Lung Cancer (NSCLC), clinical T category definition has been shifted to the size of the solid component on the CT scan [2]. Surgical resection is the mainstay in the treatment of early‐stage Lung Adenocarcinoma (LUAD), with the goal of satisfying radical removal of the disease while preserving as much healthy lung parenchyma as possible. Current guidelines on resection extent for early‐stage lung cancer are primarily based on clinical T stage. Because of the results from the recently released phase III trials, sublobar resection has replaced anatomical lobectomy as the guideline‐recommended procedure for peripherally located T1a‐bN0M0 NSCLC [3]. The Japanese Clinical Oncology Group (JCOG) trials focused mostly on GGO‐containing lesions (namely part‐solid nodules, PSN) [4, 5, 6]. With only half of the enrolled patients having pure solid lesions (solid pulmonary nodules, SPN), the JCOG0802 trial found better overall survival (OS) with segmentectomy than with lobectomy in patients with cT1a‐b LUAD [7]. In contrast, the Cancer and Leukemia Group B (CALGB) 140 503 trial, which did not specify nodule features on CT scan, found similar OS between sublobar resections and lobectomy [8]. While the European Society of Thoracic Surgeons (ESTS) recommends systematic nodal dissection in all lung cancer cases [9], the CALGB 140503 trial found that the extent of lymph node dissection had no impact on recurrence‐free survival (RFS) [10]. It remains unclear whether this conclusion holds true for the distinct entities of PSN and SPN. A recent study on GGO‐containing nodules has shown that selective lymph node dissection or omitting dissection can reduce surgical trauma without undermining oncological outcome [11]. These discrepancies limit the real‐world application of results from clinical trials.

The purpose of this study was to evaluate surgical strategies for patients with clinical stage IA LUAD appearing as PSN or SPN, focusing on resection extent, lymph node dissection, and their impact on patient outcomes. The rationale is that CT features are often the key preoperative determinant on surgical strategy in addition to clinical tumor stage. By isolating CT features as the decision‐making parameter, we aimed to investigate differences in imaging and histological characteristics between PSN and SPN and their association with tumor recurrence after different types of surgery.

2. Methods

2.1. Study Population

From a prospectively kept database, 3218 consecutive patients who underwent curative surgery for clinical stage IA LUAD at Shanghai Chest Hospital between January 2013 and October 2016 were retrieved for this study. The inclusion criteria were pathologically confirmed invasive LUAD appearing as peripheral PSN or SPN on preoperative CT scan, and surgery with a curative intent. The exclusion criteria were pure GGO lesions, patients having CT scans at other hospitals, an interval of over 1 month between CT examination and surgery, suspected synchronous multiple primary lung nodules, non‐adenocarcinoma histology, those with positive surgical margins (R1/R2 resection), and those patients who had received neoadjuvant therapy before surgery. According to the radiological appearance on CT scan, patients were divided into a PSN group and a SPN group.

2.2. Radiological and Pathological Evaluations

CT scans were performed from the apex to the base of the lung at suspended maximum inspiration using a 64‐detector CT row scanner (Brilliance 64; Philips, Eindhoven, The Netherlands). Soft tissue and lung algorithms with a thickness of 1 mm were used to reconstruct the images. Consolidation was defined as a homogeneous increase in pulmonary parenchymal attenuation that obscured the margins of airway walls and vessels, while GGO was defined as hazy increased opacity of lung, with preservation of bronchial and vascular margins [12]. All measurements were conducted with a lung window setting (lung window center, −520 HU/lung window width, 1450 HU). The longest diameters of the total lesion and the consolidation part of PSN were measured on the planes displaying the largest lesion and the largest portion of solid component, respectively. CT value of solid component (CT‐solid value) was measured with region‐of‐interest cursors, which drew the edge of the solid component on the plane displaying the largest portion of the solid component [2, 13]. The ROI encompassed the entire solid component on that slice, and the mean Hounsfield unit was recorded. Reproducibility of CT‐solid measurements was not formally evaluated; to minimize bias, all measurements were performed by consensus of two experienced thoracic radiologists.

Histological types of the tumors were classified according to the International Association for the Study of Lung Cancer (IASLC) /American Thoracic Society (ATS) /European Respiratory Society (ERS) classification of LUADs [14]. Tumor grading followed the IASLC system: Grade 1 (well‐differentiated, lepidic‐predominant, < 20% high‐grade patterns), Grade 2 (moderately‐differentiated, acinar/papillary‐predominant, < 20% high‐grade patterns), and Grade 3 (poorly‐differentiated, ≥ 20% high‐grade patterns) [15].

2.3. Surgery and Postoperative Follow‐Up

All patients received lobectomy, segmentectomy, or wedge resection according to the surgeons' decision, with an intention‐to‐cure. For patients undergoing sublobar resection (segmentectomy or wedge resection), the intraoperative margin distance was routinely assessed. In all such cases, the tumor distance from the surgical margin was > 1.5 cm or greater than the maximum tumor diameter. Extent of lymph node dissection was retrieved from the prospectively recorded data. Systematic nodal dissection (SND) was defined as complete dissection of at least three mediastinal nodal stations, one of which should be those nodes in the subcarinal position, along with the excision of hilar and intrapulmonary lymph nodes. Adjuvant therapies were jointly determined by the surgeons and the physicians after surgery according to guideline. Patients were followed every 6 months for the first 2 years after surgery and then annually afterwards. Physical examination, chest CT scan, ultrasonography of abdominal and neck regions, and serum tumor marker test were checked routinely. Magnetic resonance imaging of brain was checked yearly in patients with SPN. Bone scans and PET‐CT were reserved only if recurrence was suspected. Loco‐regional recurrence was defined as recurrent disease limited to the ipsilateral hemithorax and regional lymph nodes, including hilar, mediastinal, and cervical lymph nodes regardless of the side of the primary tumor. All the other recurrences, including contralateral or bilateral pulmonary metastases, were defined as distant recurrences [16].

2.4. Statistical Analysis

Statistical analyses were performed using the SPSS Statistics 23.0 software (SPSS Inc., Chicago, IL, USA) and R software (version 3.6.3, http://www.r‐project.org). Continuous variables were compared using the Student t‐test or Wilcoxon rank‐sum test, depending on whether the variables were normally distributed. Categorical variables were analyzed with the Chi‐square test or the Fisher exact test when appropriate. The interaction effect analysis on the CT‐solid value used the Generalized Linear Models (GLMs) with Gamma distribution including nodule type and solid component size. The CT‐solid values in the interaction effect analysis were adjusted to continuous positive values by using the following formula:

$$\begin{array}{c}\text{Adjusted}\mathrm{CT}-\text{solid value}=\mathrm{CT}-\text{solid value}\hfill \\ +\mathrm{abs}\left(\min \left(\mathrm{CT}-\text{solid value}\right)\right)+1\hfill \end{array}$$

RFS was defined as the time between the day of surgery and the day of first recurrence, patient death, or last follow‐up. RFS was calculated with Kaplan–Meier curves and compared by log‐rank test. To analyze the imaging and histological features of PSN and SPN, their solid component size was incorporated into the 1st propensity‐score matching (PSM). In the analysis for patient prognosis, a 2nd PSM was carried out including patient sex, age, smoking history, American Society of Anesthesiologists (ASA) score, forced expiratory volume in 1 s, resection extent and lymph node sampling or dissection, along with solid component size. Thus, the first PSM aimed to enable a size‐independent comparison of tumor biology, while the second PSM aimed to enable a clinical and surgery‐balanced comparison of long‐term outcomes. A ratio of 1:1 match with a caliper size of 0.04 was applied for the 1st PSM, using the nearest‐neighbor matching algorithm without replacement. For the 2nd PSM, the caliper size was set as 0.01 in order to reduce interference among factors and further improve the reliability of the analysis. Risk factors for recurrence were analyzed by Cox proportional hazards regression model. Potential risk factors with p < 0.1 in univariate analysis, along with factors deemed clinically plausible by clinicians (even if p exceeded 0.1), were included in the multivariate risk analysis. All tests were two‐tailed with a statistical significance of 0.05.

3. Results

In total, the study included 710 patients with clinical stage IA LUAD. There were 356 patients in the PSN group and 354 patients in the SPN group. Overall, clinical and pathological characteristics differed markedly between the PSN and SPN groups. There were significantly more female patients in the PSN group than in the SPN group. The solid component size was significantly smaller in the PSN group than in the SPN group. Notably, the CT‐solid value in PSN was significantly lower than in SPN. More sublobar resections and less SND were performed in the PSN group (Table 1). Upon pathological examination, PSN had significantly lower histological grading, less lymph node metastasis, LVI, and VPI than SPN.

TABLE 1.

Clinical characteristics of patients with PSN and SPN.

		Before 1st PSM			After 1st PSM
		PSN	SPN	p	PSN	SPN	p
		(n = 356)	(n = 354)		(n = 214)	(n = 214)
Sex	Female	246 (69.1%)	176 (49.7%)	< 0.001	144 (67.3%)	106 (49.5%)	< 0.001
	Male	110 (30.9%)	178 (50.3%)		70 (32.7%)	108 (50.5%)
Age, year		60.9 ± 9.1	61.7 ± 9.7	0.272	61.8 ± 8.8	61.5 ± 10.2	0.953
Smoking history	Smoker	51 (14.3%)	52 (14.7%)	0.891	31 (14.5%)	26 (12.1%)	0.477
	Non‐smoker	305 (85.7%)	302 (85.3%)		183 (85.5%)	188 (87.9%)
FEV1 pred	≥ 80%	294 (82.6%)	275 (77.7%)	0.102	176 (82.2%)	167 (78.0%)	0.276
	< 80%	62 (17.4%)	79 (22.3%)		38 (17.8%)	47 (22.0%)
ASA score	Grade 1	176 (49.4%)	188 (53.1%)	0.286	101 (47.2%)	116 (54.2%)	0.177
	Grade 2	164 (46.1%)	157 (44.4%)		104 (48.6%)	94 (43.9%)
	Grade 3	16 (4.5%)	9 (2.5%)		9 (4.2%)	4 (1.9%)
Resection extent	Sublobar resection	108 (30.3%)	78 (22.0%)	0.012	53 (24.8%)	54 (25.2%)	0.911
	Lobectomy	248 (69.7%)	276 (78.0%)		161 (75.2%)	160 (74.8%)
Extent of lymph node dissection	Sampling‐only	153 (43.0%)	106 (29.9%)	< 0.001	81 (37.9%)	77 (36.0%)	0.689
	SND	203 (57.0%)	248 (70.1%)		133 (62.1%)	137 (64.0%)

Abbreviations: ASA, American Society of Anesthesiologists; FEV1 pred, forced expiratory volume in 1 s; PSM, propensity score match; SND, systematic nodal dissection.

After the 1st PSM, 214 pairs of PSNs and SPNs were retained for further analysis. Solid component size was similar between the two groups. The PSN group still had more female patients than the SPN group (Table 1). The CT imaging patterns between PSN and SPN still had significant differences after PSM. While PSNs had significantly larger total lesion size than SPNs, their CT‐solid was significantly lower than the latter; meanwhile, PSNs had a much larger distribution range than SPNs (PSN vs. SPN: −13.2 ± 69.6 vs. 22.2 ± 29.0, p < 0.001; Figure 1A). In addition, PSNs showed marked increase in CT‐solid along with the growth of solid component, while change in CT‐solid was less obvious in SPNs (Figure 1B).

FIGURE 1.

Comparison of CT‐solid value between PSNs and SPNs. (A) PSNs had a much larger distribution range in CT‐solid compared with SPNs. (B) Scatter plot of solid component size and their CT‐solid value in PSNs and SPNs; the fitting curves showed more marked change in CT‐solid value of PSN than that of SPNs.

Upon pathological examination, there was still significantly less Grade 3 histology in PSNs than in SPNs (9.8% vs. 34.1%, p < 0.001) after PSM. Incidence of VPI (11.7% vs. 30.4%, p < 0.001) and lymph node metastasis (11.2% vs. 20.6%, p = 0.011) was also significantly lower in the PSN group than in the SPN group (Table 2).

TABLE 2.

Imaging and pathological characteristics for PSN and SPNs.

		Before PSM			After PSM
		PSN	SPN	p	PSN	SPN	p
		(n = 356)	(n = 354)		(n = 214)	(n = 214)
Total lesion size, mm		21.8 ± 9.2	18.0 ± 5.9	< 0.001	24.9 ± 8.9	15.7 ± 5.7	< 0.001
Solid component size, mm		12.1 ± 6.4	18.0 ± 5.9	< 0.001	15.6 ± 5.7	15.7 ± 5.7	0.857
CT‐solid value, Hu		−30.7 ± 72.6	27.1 ± 28.1	< 0.001	−13.2 ± 69.6	22.2 ± 29.0	< 0.001
Histological grading	Grade 1	53 (14.9%)	3 (0.8%)	< 0.001	27 (12.6%)	3 (1.4%)	< 0.001
	Grade 2	275 (77.2%)	214 (60.5%)		166 (77.6%)	138 (64.5%)
	Grade 3	28 (7.9%)	137 (38.7%)		21 (9.8%)	73 (34.1%)
LVI		9 (2.5%)	21 (5.9%)	0.032	8 (3.7%)	6 (2.8%)	0.587
VPI		30 (8.4%)	113 (31.9%)	< 0.001	25 (11.7%)	65 (30.4%)	< 0.001
Pathological N1/2		26 (7.3%)	87 (24.6%)	< 0.001	24 (11.2%)	44 (20.6%)	0.008

Abbreviations: LVI, lymphovascular invasion; PSM, propensity score match; VPI, visceral pleural invasion.

The 2nd PSM was conducted for the prognosis analysis. Detailed information is shown in Table S1. With a median follow‐up of 55 months in the PSN group and 51 months in the SPN group, the 5‐year RFS of the PSN patients was significantly better than that of the SPN patients (87.6% vs. 79.8%, p < 0.001) (Figure 2A). In further stratified analysis by clinical T category, the 5‐year RFS of the PSN patients was significantly better than that of the SPN patients in each T category (Figure 2B–D).

FIGURE 2.

Stratified analysis of RFS in PSN and SPN by different T categories after PSM.

Cox regression analysis showed that the presence of GGO (HR = 0.59, 95% CI: 0.38–0.90, p = 0.014) was an independent protective factor for recurrence. And T1c (HR = 6.01, 95% CI: 3.34–10.8, p < 0.001) was identified as an independent risk factor for increased recurrence (Table S2).

In PSNs, a stratified analysis by different clinical T categories was conducted (Figure S1). PSN with a similar size of solid component had similar RFS (For cT1a PSN, p = 0.816; For cT1b PSN, p = 0.762; For cT1c PSN, p = 0.888), regardless of the extent of GGO.

Patterns of recurrence were also different between the PSN and SPN groups (Table S3). PSN had much lower rates of loco‐regional recurrence and distant metastasis than SPN. In the PSN group, there was no difference in loco‐regional recurrence or distant metastasis between patients who received lobectomy or sublobar resections, or between patients who had SND or lymph node sampling only. On the contrary, the loco‐regional recurrence rate in SPNs was significantly higher after sublobar resections than after lobectomy (29.5% vs. 10.3%, p < 0.001), and was also significantly higher after nodal sampling‐only than after SND (25.5% vs. 9.8%, p < 0.001). But no significant difference was observed in distant metastasis rate among different surgical procedures (Table 3).

TABLE 3.

Comparison in recurrence status among different surgical treatment.

	Resection extent		p	Lymph node dissection		p
	Sublobar	Lobectomy		Sampling‐only	SND
Loco‐regional recurrence*
PSN	5 (4.6%)	10 (4.0%)	0.796	5 (3.3%)	10 (4.9%)	0.441
SPN	23 (29.5%)	28 (10.3%)	< 0.001	27 (25.5%)	24 (9.8%)	< 0.001
Distant recurrence**
PSN	7 (6.5%)	19 (7.7%)	0.694	7 (4.6%)	19 (9.4%)	0.086
SPN	11 (14.1%)	49 (17.9%)	0.426	17 (16.0%)	43 (17.6%)	0.730

^*Loco‐regional recurrence: Patients with loco‐regional recurrence with/without distant recurrence.

^**Distant recurrence: Patients with distant recurrence with/without loco‐regional recurrence.

Stratified cox regression analyses for loco‐regional or distant recurrence in PSN and SPN were then carried out (Tables [Link], [Link]). For PSN, solid component size was the only independent predictor for both loco‐regional and distant recurrence (Figure 3). In SPN, lobectomy and SND were found to be protective factors for loco‐regional recurrence, but not for distant metastasis. The lesion size was the only risk factor for the distant recurrence in SPN (Figure 3).

FIGURE 3.

Multivariate Cox regression analysis for different recurrence patterns. (A) Loco‐regional recurrence in PSN. (B) Distant metastasis in PSN. (C) Loco‐regional recurrence in SPN. (D) Distant metastasis in SPN.

4. Discussion

Our study revealed significant differences between PSN and SPN across clinical, radiological, and pathological dimensions. Despite the solid components of PSN and SPN both appearing as solid images on CT scan, the CT‐solid of PSN was significantly lower than SPN. Moreover, while the CT‐solid in PSN increased remarkably with the increase in size of solid component, the CT‐solid of SPN showed less obvious change during the course. After PSM, PSN showed significantly lower rates of high‐grade histology, VPI, and lymph node metastasis compared with SPN. Patients with PSN had significantly better 5‐year RFS, consistently across all clinical T1 categories, than patients with SPN after PSM. Sublobar resection and limited lymph node sampling only were associated with increased risk of loco‐regional recurrence in SPN, but not in PSN. These findings underscore the need to differentiate PSN and SPN in clinical evaluation and surgical planning.

Previous studies have demonstrated that the CT value of pulmonary nodules could serve as a preoperative predictor of invasiveness in early‐stage lung cancer. Tamura et al. found that the development of invasiveness in clinical stage IA non‐small cell lung cancer was accompanied by an increase in the average CT value of the lesion (non‐invasive group at −643.6 ± 9.4 HU, invasive group at −365.9 ± 11.4 HU, p < 0.0001) [17]. But they used the average CT value of the entire lesion in their study, which failed to consider the more important prognostic impact of the solid component. It is already known that GGO component typically represents more indolent lepidic histology, while the solid component may indicate more aggressive tumor [18, 19, 20]. But it is unclear whether the solid component in PSN is equivalent to SPN of the same size. We first did a PSM analysis to eliminate the impact from the size of solid components. The CT‐solid in PSN was still significantly lower than that in SPN (−13.2 HU vs. 22.2 HU, p < 0.001). We further analyzed the CT‐solid value escalation kinetics during tumor growth in PSN versus SPN. The solid components in PSNs showed progressive consolidation with markedly increasing CT‐solid value along with enlarging solid component size (p < 0.001), while there was no such obvious consolidation‐size correlation in SPN. This difference in CT‐solid value change between the two types of nodules may be explained by different patterns of carcinogenesis and development of LUAD, including the stepwise progression pathway and the non‐stepwise (De novo) pathway, which have been suggested in the tumorigenesis process of early‐stage LUAD [21]. PSN demonstrated progressive increase in CT‐solid during solid component growth, corresponding to the stepwise progression pathway wherein invasive adenocarcinoma evolves from pre‐invasive GGO component. In contrast, SPN exhibited size‐independent stable attenuation, characteristic of the non‐stepwise (De novo) pathway where malignancies manifest as inherently solid lesions without precursor involvement.

Solid component within PSN is known to correlate with regions showing infiltrative proliferation of tumor cells [22, 23]. It is well‐established that the solid component dictates tumor aggressiveness, leading to its adoption as the determinant for clinical T definition since the 8th staging [2]. After matching for other potential confounding factors, we compared the RFS between PSN and SPN with a similar size of solid component. Our results showed that PSN still had significantly superior RFS compared with SPN across the entire T1 stage. SPN is usually considered to be associated with worse histological characteristics compared with PSN [24, 25]. Our results further demonstrated that when solid component size was matched, PSN exhibited significantly lower invasiveness than SPN (lower IASLC grading, less visceral pleural invasion, and less lymph node metastasis). And such significant histopathological differences were associated with distinct clinical prognostic outcomes. Fu et al. found that the patients with PSN had better prognosis among stage I NSCLC [26]. Miyoshi et al. also reported that the 5‐year overall survival of patients with GGO‐containing lesions was 97% compared to 84% in patients without GGO components in stage IA LUAD [27]. However, resection extent and extent of lymph node dissection were not incorporated as stratified variables in these studies' prognostic analyses. And only overall survival was analyzed without stratification into loco‐regional versus distant recurrence, thereby failing to elucidate distinct recurrence patterns between nodule types. The dynamic changes in CT attenuation values relative to solid component size were not explored, leaving the radiological evolutionary differences between PSN and SPN unclear. Better prognosis in PSN than in SPN was observed in other studies, though PSM was not applied in these studies [28, 29, 30]. In contrast, a key strength of our study is the use of systematic PSM. This methodological rigor reinforces that the superior RFS observed in PSN is attributable to its distinct biological behavior rather than to imbalanced clinical factors, underscoring the prognostic significance of CT morphology in early‐stage lung adenocarcinoma.

In our study, PSN patients experienced significantly less overall recurrence and loco‐regional recurrence compared to SPN patients. SPN patients had higher loco‐regional recurrence after sublobar resection than after lobectomy and after lymph node sampling than after systematic dissection. In the JCOG0802 trial, patients with PSN and SPN were grouped together. Loco‐regional recurrence rates were twice as high after segmentectomy than after lobectomy in the overall cohort, mostly in patients with lesions exhibiting a consolidation‐to‐tumor ratio (CTR) of 1.0 (i.e., SPN), which is also consistent with the findings in our study [7, 31]. In Cox regression analysis for recurrence risk factors, only solid component size was predictive of loco‐regional and distant recurrence rates in the PSN group, while the extent of surgical resection and the thoroughness of lymph node dissection were independent risk factors for loco‐regional recurrence in the SPN group. Limited resection (sublobar vs. lobectomy) and less comprehensive lymph node dissection (sampling vs. systematic dissection) were associated with a significantly higher rate of loco‐regional recurrence in SPN. Patients with SPN who underwent lobectomy had a significantly lower loco‐regional recurrence rate than those who had sublobar resections, a result that is also consistent with the findings of JCOG0802 loco‐regional recurrence risk factors analysis [32]. In the additional analysis of lymph node dissection from the JCOG0802 study, Maniwa et al. reported that only 1.5% of PSN patients exhibited lymph node metastasis, with a particularly low likelihood of metastasis to non‐adjacent lobar or mediastinal lymph nodes [33]. A recent multicenter prospective study demonstrated that systematic lymph node dissection may be omitted in PSN patients [11]. Compared with patients undergoing lymph node dissection, the no‐dissection group showed comparable survival outcomes, along with shorter surgical time, fewer postoperative complications, and shorter hospital stays [11]. This finding suggests that systematic lymph node dissection may offer limited prognostic benefit for PSN, as the risk of nodal involvement is minimal. In contrast, SPN patients demonstrated a significantly higher rate of lymph node metastasis, underscoring the necessity of thorough nodal evaluation in this subgroup. This distinction suggests that imaging‐specific lymph node dissection strategies may be considered: limited sampling for PSN and systematic dissection for SPN, tailored to their inherent biological behaviors.

Our study has several limitations that should be considered when interpreting the results. First, its retrospective design inherently introduces selection bias. Although PSM balanced measured confounders, it cannot correct for unmeasured factors—therefore our findings are hypothesis‐generating and require prospective validation. Second, the CT evaluation was based on two‐dimensional measurements, which might not fully capture the three‐dimensional heterogeneity of the solid components. Third, interobserver reproducibility for CT‐solid measurements was not formally assessed, which might affect the reliability of our CT‐based quantitative analysis. Forth, the absence of molecular profiling precludes a deeper understanding of the biological mechanisms underlying the observed clinicopathological differences between PSN and SPN. Fifth, wedge resection and segmentectomy were not analyzed separately due to limited subgroup sample sizes after PSM, which would preclude meaningful statistical comparison. Future studies incorporating volumetric or radiomic CT analyses and molecular characterization would help clarify the distinct pathogenesis of PSN and SPN and further refine individualized management strategies.

In conclusion, PSN and SPN may likely represent LUADs of distinct origins, which accounts for the significant differences observed in their radiological features, pathological characteristics, and clinical outcomes. Definitive conclusions regarding distinct biological entities will require molecular and genomic validation. Our findings further confirm that the solid components of PSN and SPN are fundamentally distinct in their biological behavior, rather than merely representing “lesions of similar size with different CT phenotypes”. Therefore, these findings suggest that differentiated treatment strategies may be considered for the two imaging subtypes, although confirmation in prospective, protocol‐driven studies is required before clinical implementation.

Author Contributions

Songyuan Guo: data curation, investigation, writing – review and editing. Yiyang Wang: project administration, resources, writing – review and editing. Wentao Fang: conceptualization, supervision, project administration, resources, writing – review and editing. Zhitao Gu: project administration, resources, writing – review and editing. Runhua Yu: data curation, investigation, formal analysis, methodology, writing – original draft, writing – review and editing. Hanyue Li: conceptualization, supervision, writing – original draft, writing – review and editing, data curation, investigation, formal analysis, methodology.

Funding

This study was financially supported by Shanghai Shenkang Hospital Development Center SHDC2020CR1023B.

Ethics Statement

The study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice Guidelines and was approved by the Institutional Review Board of Shanghai Chest Hospital (No. IS25276) and was registered in the Chinese Clinical Trial Registry (ChiCTR2600116838). Due to the retrospective nature of the research, the requirement for written informed consent was waived by the ethics committee. All patient data were anonymized and handled with strict confidentiality to protect participant privacy throughout the study.

Consent

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Comparison of RFS among patients with PSN in the same T category but with different GGO size.

Table S1: Clinicopathologic features after 2nd Propensity score match.

Table S2: Univariate/Multivariate Cox regression analysis for Recurrence.

Table S3: Recurrence pattern in all patients.

Table S4: Univariate/Multivariate Cox regression analysis for Loco‐regional recurrence in PSN.

Table S5: Univariate/Multivariate Cox regression analysis for Distant recurrence in PSN.

Table S6: Univariate/Multivariate Cox regression analysis for Loco‐regional recurrence in SPN.

Table S7: Univariate/Multivariate Cox regression analysis for Distant recurrence in SPN.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.