Lung cancer is the second-most common cancer worldwide and the leading cause of cancer morbidity and mortality in men (1). Non-small cell lung cancer (NSCLC) accounts for approximately 85% of all lung cancers, and adenocarcinoma is the most common histologic type of NSCLC (2). Lung adenocarcinoma can be further subdivided into preinvasive lesions [atypical adenomatous hyperplasia (AAH) and adenocarcinoma in situ (AIS)], minimally invasive adenocarcinoma (MIA), and invasive adenocarcinoma (IAC), according to the degree of invasion (3). The different subtypes of lung adenocarcinoma strongly impact the choice of surgical methods and the prognosis of the patient. Previous studies have reported that in stage I lung adenocarcinoma, patients with the AIS and MIA subtypes have a 5-year disease-free survival (DFS) rate close to 100%, while this rate is only 38%–86% for those with IAC (4, 5).

Lobectomy remains the standard surgical treatment for lung adenocarcinoma. However, non-IAC patients also have the option of limited surgical resection (6). Several prospective, multicenter clinical studies have investigated the effect of lobectomy vs. sublobectomy on the survival rates of AIS and MIA patients. They have found that sublobectomy is the optimal choice, producing similar survival rates along with better quality of life (7, 8). If we could precisely determine the degree of invasiveness of lung adenocarcinoma intraoperatively, the less damaging sublobectomy would be feasible in more cases, based on guaranteed curative resection. However, Yeh et al. have reported that interobserver agreement in terms of discriminating between AIS, MIA, and invasive adenocarcinomas using frozen sections (FSs) is not satisfactory (κ = 0.378, fair agreement) (9). As such, the identification of AIS and MIA by examination of intraoperative FSs is not recommended by the prevailing guidelines.

Non-mucinous invasive lung adenocarcinoma is mainly divided into lepidic, acinar, papillary, micropapillary, and solid predominant patterns; this pattern is increasingly considered to be a powerful predictor of prognosis in lung adenocarcinoma (10). IAC, with a predominant lepidic pattern, has been found to reflect indolent behavior in terms of the biological behavior of the tumor, which is associated with a favorable prognosis (11). In contrast, studies have shown that prognosis is significantly worse in the case of sublobectomy than in the case of lobectomy among IAC patients with micropapillary composition ≥5% (11–13). Furthermore, the presence of micropapillary or solid components has been found to be an independent predictor for unfavorable recurrence-free survival (RFS) rates in cT1N0M0 lung adenocarcinoma patients undergoing lobectomy (14). Preoperatively and intraoperatively, accurate histologic subtyping is helpful to enable thoracic surgeons to decide on the extent of surgical resection. The accuracy of FS for prediction of the predominant histologic subtype is not satisfactory (κ = 0.565, moderate agreement), and the sensitivity of FS analysis, in terms of the presence or absence of micropapillary and solid patterns, is poor (micropapillary: 37%, κ = 0.321, fair agreement; solid: 69%, κ = 0.67, substantial agreement) (9).

To date, studies on intraoperative FS have demonstrated low consistency in the determination of pathological subtype (9, 15–17). Therefore, it is desirable to identify preoperative biomarkers that can distinguish the histologic patterns of early-stage lung adenocarcinoma.

Circulating tumor cell (CTC) level has proven utility in the early diagnosis of cancer, in monitoring recurrence and metastasis, in determining the prognosis of surgical and systemic interventions, and in the selection of postoperative adjuvant therapy (18). Folate receptor-positive CTCs (FR⁺CTCs) are a type of circulating rare cell with high expression of FRs on the cell surface; levels of these cells can be reliably quantified using commercially available kits (19). Lung cancer patients of all stages have been found to exhibit significantly higher FR⁺CTC levels than patients with benign lung disease and healthy subjects. Studies have also shown that FR⁺CTC level can be used to diagnose lung cancer with high sensitivity (79.6%) and specificity (88.2%) (19–21). Furthermore, FR⁺CTC level is significantly higher in patients with MIA and IAC than in AIS patients (22). However, the correlation between FR⁺CTCs and the histologic patterns of cTis-T1N0M0 lung adenocarcinoma remains to be explored, and the question of whether FR⁺CTC level can help thoracic surgeons to determine the extent of surgical resection required for early-stage lung adenocarcinoma is worthy of investigation.

Achieving consistency between intraoperative FS and postoperative final pathology in patients with early-stage lung adenocarcinomas remains a challenging problem. Marchevsky et al. report that intraoperative FS has lower diagnostic accuracy for AIS and MIA than for IAC (51% vs. 97%) (25). He et al. have also found that intraoperative FS pathology offers poor differentiation between AAH and AIS and between AIS and MIA. The possible reasons include the deformation of tissues and cells caused by freezing during the operation and the limitations of FS sampling leading to an incomplete representation of the entire lesion (26). In our study, consistency between intraoperative FS and postoperative pathology was only 62.3%, lower than observed in a previous study (64%) (9). The FR⁺CTC test was approved by the NMPA to assist in diagnosing pulmonary lesions and has showed high sensitivity and specificity. In our study, FR⁺CTC level could not be used to distinguish between AIS, MIA, and IAC, which is in line with the results of the study by Ding et al. However, in their study, FR⁺CTCs could be used to distinguish between AIS and MIA (22); this inconsistency of findings may be related to the fact that the participants included in our study were all in the early stage and the sample size was small. Neither FR⁺CTC level nor intraoperative FS can accurately distinguish the pathological subtypes of lung adenocarcinoma. However, larger samples in prospective clinical studies are needed to provide validation in the future.

Some studies have shown that pathological subtype, degree of differentiation, and clinical stage are closely correlated with prognosis in lung cancer (27, 28). However, few studies have explored the ability of FR⁺CTC level to distinguish growth pattern subtypes, as this one has done, and we propose that this measure can be used a reference in determining the extent of surgical resection required in patients with early invasive adenocarcinoma. Adenocarcinoma patients with tumors exhibiting different growth patterns would ideally receive different forms of treatment, and they have different prognoses. In a study by Tsuta et al., 904 cases of surgically resected adenocarcinomas were investigated; it was found that lepidic, acinar, papillary, micropapillary, and solid predominant adenocarcinomas were associated with different prognoses, and their 5-year OS rates were 93%, 67%, 74%, 62%, and 58%, respectively (29). A recent study consisting of 697 patients with pN0M0/papillar/papillar–acinar predominant lung adenocarcinomas with diameter ≤3 cm who underwent curative resection found that the presence of micropapillary and solid patterns as minor components had a negative impact on prognosis. For MP/S⁺ Lep⁻, MP/S⁺ Lep⁺, MP/S⁻ Lep⁻, and MP/S⁻ Lep⁺ types, the 5-year RFS rates were 81.9%, 94.0%, 94.4%, and 98.7%, respectively (P < 0.001), and the 5-year OS rates were 87.7%, 96.6%, 94.4%, and 98.4%, respectively (P < 0.001). Moreover, a multivariate analysis suggested that the MP/S⁺ Lep⁻ subtype was an independent poor prognostic factor for both RFS and OS (30). The sensitivity and specificity of intraoperative FS for diagnosis of solid primary adenocarcinoma were 79% and 94%, respectively. For the diagnosis of micropapillary primary lung adenocarcinoma, specificity reached 99%, but sensitivity was only 21%. When only the presence or absence of micropapillary subtypes was considered, the diagnostic sensitivity of FS was still only 37% (9).

In our study, we found that there was no difference in level of FR⁺CTCs between the acinar, lepidic, papillary, micropapillary, and solid subtypes in the non-mucinous IAC group. Levels were higher in the acinar subtype compared to the lepidic subtype; in the solid subtype compared to the acinar and papillary subtypes; and in the papillary subtype compared to the lepidic subtype. Furthermore, when only the presence or absence of micropapillary or solid subtypes was considered, there was a notable difference in FR⁺CTC level between the groups (P = 0.017 and P = 0.022, respectively). The presence of micropapillary or solid patterns was found to significantly increase the risk of nodal upstaging in the multivariable analysis (P < 0.001 and P = 0.001, respectively), and these patterns were independently associated with poor RFS (P = 0.041 and P < 0.001) among patients with cT1N0M0 lung adenocarcinoma (14). Micropapillary or solid components have been proven to be poor prognostic factors. Therefore, these patients should be cautious when considering sublobectomy. According to the guidelines, the standard surgery for IAC is still lobectomy; however, in our study, the proportions of patients who underwent lobectomy in high-risk groups with micropapillary and solid components were only 64.1% and 68.3%, respectively ( Supplementary Table 3 ). Therefore, it is necessary to identify a biomarker that can aid preoperative pathological judgment and improve the accuracy of selection of surgical method. According to the results of our study, FR⁺CTCs may be a promising auxiliary biomarker for determination of whether solid and micropapillary components are present. However, the diagnostic efficacy is not very good, and the sensitivity and specificity of FR⁺CTC level as a test for the presence of micropapillary components and solid components are relatively low. The AUCs are well below 0.7 (0.574 and 0.590). This means that FR⁺CTC level can only function as a reference in distinguishing whether micropapillary or solid components are present, and does not have a strong ability to make this distinction. Prospective clinical studies with larger samples are needed to provide validation of this finding. The proportion of tumors of the complex gland subtype is small among patients with early-stage adenocarcinoma; thus, there is insufficient evidence to draw a conclusion on its role in prognosis after sublobar resection at present.

FR⁺CTC level was correlated with lymph node metastasis in IAC, with the level being higher in patients with lymph node metastasis than in patients without metastasis (P = 0.035), which was consistent with a previous study (27). The sensitivity and specificity of FR⁺CTC level for diagnosis of lymph node metastasis were 77.14% and 47.88%, respectively, and the AUC was 0.606. Thus, at a cutoff value of 9.72 FU/3 ml, this test cannot reliably distinguish whether lymph node metastasis has occurred. Evidence has shown that invasion may occur very early in the tumor process, and CTCs are released into circulation in the early phase of cancer. Lymph node infiltration is related to poor prognosis (28). Many researchers have emphasized the importance of detection of CTCs (31–33). Our results also showed that FR⁺CTC level was correlated with degree of differentiation in IAC. Until now, few studies have focused on the prognostic value of biomarkers in lung tumors with different degrees of differentiation, so this finding is of limited significance because a more accurate method of discrimination has already been developed.

VPI is a high-risk factor that affects prognosis in lung adenocarcinoma. Nevertheless, identification of VPI is still reliant on elastic fiber staining, which is time-consuming and difficult to perform in intraoperative FS. The accuracy, sensitivity, and specificity of intraoperative FS for diagnosis of VPI have been found to be 75%, 47.4%, and 97.3%, respectively (34). Furthermore, our results indicated that FR⁺CTC level was correlated with VPI (P = 0.003). However, when 10.25 FU/3 ml was used as the cutoff value, the sensitivity and specificity of FR⁺CTC level for the diagnosis of VPI were 61.54% and 58.81%, respectively. The associated AUC was 0.623, which means that FR⁺CTC is almost capable of distinguishing whether VPI has occurred with sufficient accuracy.

Two studies have shown that the sensitivity of STAS detection in intraoperative FS is only 44%-54%. At the same time, specificity is as high as 80%-91%, and multifactorial logistic regression analysis has found that artifacts are the only relevant factor in the misdiagnosis of STAS by FS (35, 36). However, in the present study, there was no significant difference in FR⁺CTC level between cases with and without STAS (P = 0.224). The possible reasons are a relatively small sample size and the inclusion of patients from a population at a specific stage. Ways to improve the sensitivity of STAS detection preoperatively or intraoperatively are worthy of future research.

In our study, we found that the proportions of patients with advanced subtypes (micropapillary or solid or complex glands), lymph node metastasis, poor differentiation, and VPI who underwent lobectomy were 64.8%, 82.9%, 63.5%, and 75.0%, respectively ( Supplementary Table 3 ); the proportion did not reach 100% in any of these subgroups. In the case of stage I IAC patients who underwent sublobectomy because high-risk pathological factors (advanced subtypes, VIP, STAS, lymph node metastasis, and so on) could not be identified intraoperatively, postoperative adjuvant therapy or supplementary lobectomy was suggested if the patients agreed (16, 37). FR⁺CTCs can provide helpful information in distinguishing aggressive growth patterns of IAC, lymph node metastasis, and VPI, which could help to identify the optimal surgical strategy and avoid infliction of secondary damage on patients. In the meantime, FR⁺CTC level can be used as a biomarker and possible prognostic factor in early-stage lung adenocarcinoma.

This study had some limitations. The study was preliminary and retrospective in nature, and patient outcomes were not collected and analyzed. In subsequent studies, we will collect long-term prognostic data and analyze the value of FR⁺CTC level in predicting outcomes in early-stage patients. Furthermore, prospective randomized studies are required to truly discern the value of FR⁺CTC level in the preoperative differentiation of histological subtypes and as a tool for use in the selection of surgical method.

To conclude, FR⁺CTC level may be used as an additional biomarker in identifying patients with micropapillary components, solid components, advanced subtypes, poorly differentiated tumors, VPI, and lymph node metastasis. Lobectomy should be selected as the preferred surgical option for these high-risk patients.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the Shanghai Chest Hospital. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

Conception and design: CZ and WL. Administrative support: AW and WL. Provision of study materials or patients: CZ and WL. Collection and assembly of data: CZ and RZ. Data analysis and interpretation: CZ and RYZ. Manuscript writing and final approval of the manuscript: all authors.