Lung cancer is still a heavy health and economic burden in China, with over 80% being non-small cell lung cancer (NSCLC) (1, 2), which mainly comprises adenocarcinoma (ADC) and squamous carcinoma (SCC). In recent decades, although overall survival (OS) for locally advanced NSCLC and those with remote lesions has greatly improved due to the success of immunotherapy-based regimens and targeted therapies (3–5), radical resection is still the optimal choice for early-stage cases, specifically stage I. Fortunately, with the success of clinical trials, including ADAURA (6), Keynote091 (7), and IMpower 010 (8), the survival for stage IB cases has been further guaranteed. Nonetheless, the 5-year OS for stage IA patients ranges from 77-92%, declining up to 68% for stage IB patients (9). Although many innovative prognostic indicators, such as molecular residual disease (MRD), have been reported and are very helpful in subsequent treatment strategy decisions in these cases (10, 11), searching for easily accessible and highly cost-effective prognostic indicators is still needed at present in practice.

Carcinoembryonic antigen (CEA) is a classic tumor marker in NSCLC (12) that plays an important role not only in diagnosis (13) but also in prognosis (14). However, the usefulness of CEA in practice was also blocked due to its relatively low sensitivity, particularly in stage I cases, as previous studies suggested a positive rate ranging from 16.22% (145/894 in stage IA (15)) to 33.62% [274/815 in stage I (16)]. Some studies have tried to minimize its cutoff points to improve its prognostic efficacy in stage I patients (17, 18), but the results were not extensively validated. Thus, other reliable prognostic indicators are still needed for stage I cases, the majority of which present a normal CEA level. Interestingly, it was found that some inflammatory cells and cytokines played a key role in early tumorigenesis of lung cancer (19–21), which suggested that they may also have a role in prognosis. Previously, some studies found that indicators such as the neutrophil-lymphocyte ratio (NLR) (22, 23) and lymphocyte to monocyte ratio (LMR) (24) could have prognostic value in stage I NSCLC; however, these studies also included some patients with abnormal CEA. Neutrophils are the main fractions in peripheral blood and have broad functions in cancer development, such as promoting metastasis (25, 26) and progression (27), and enhancing cancer cell adhesion (28). The prognostic usefulness of the absolute neutrophil count (ANC) has been addressed in many cancers (29–33). In lung cancer, neutrophils were also found to be a promoter of disease progression (34), and they could also stimulate T-cell responses in early-stage disease (35). In line with the aforementioned studies (29–33), the prognostic value of ANC in lung cancer was also established in chemo-naïve stage IIIB or IV cases (36) or previously treated and subsequently received anlotinib patients (37). However, the value of ANC in CEA normal stage I cases is still largely unknown.

In this study, we aimed to detect the prognostic value of the ANC in CEA normal stage I NSCLC (ADC+SCC) patients.

In this study, preoperative ANC was found to be the only significant prognostic indicator in predicting survival in CEA normal stage I NSCLC patients, in contrast to ALC, AMC, and PLC. Patients with a relatively low ANC before surgery had significantly better outcomes than those with a high ANC; however, it was not validated as an independent prognostic factor for both DFS and OS. To the best of our knowledge, this is the first report concerning the prognostic value of a specific peripheral blood fraction in CEA normal stage I NSCLC.

Previously, the prognostic role of ANC in cancer has been under extensive study, and the expansion of these cells in peripheral blood commonly predicts poor survival. Examples could be found in advanced gastric cancer, localized prostate cancer, metastatic colorectal cancer, and some head and neck malignancies (29–33). In NSCLC, Teramukai et al. studied 388 chemo-naïve stage IIIB or IV patients and found that in contrast to ALC and AMC, ANC was the only significant indicator for survival, and pretreatment high ANC (with a cutoff at 4500/mm³) was significantly correlated with poor OS and progression-free survival (PFS) (36). In addition, Zer et al., in a study with 88 advanced cases, received PD-1 inhibitor treatment and found a declined ANC during such therapy, indicating good disease control and therapy response (40); similarly, Murakami et al., reached a similar conclusion in a study with 213 patients receiving nivolumab treatment, (41). In addition, Chen et al. in a study with 71 advanced patients who received anlotinib treatment and found that compared to ALC, ANC was the only significant indicator correlated with both OS and PFS (37). Except for reports in advanced scenarios without surgery, there is also a report conducted in resected stage I-IIIA patients who found that ANC was positively associated with increased tumor burden and that surgical removal of the lesion resulted in a decrease in these cells in peripheral blood. The increase in these cells was independently correlated with poor OS (42). Nonetheless, few studies have explored the prognostic usefulness of ANC in stage I NSCLC, let alone in CEA normal background. Interestingly, two reports have indicated that a low preoperative NLR [cutoff points: 2.50 (22), 2.84 (23)] was significantly correlated with good DFS and OS in stage I NSCLC, including some cases with elevated CEA [16.73% (43/257) (22), 20.56% (37/180) (23)]. Although not conclusive, a low NLR in these studies may partially include some cases with a low ANC (accompanied by a normal ALC or high ALC), which may give some support to our results. In addition, we found that low ANC was more common in some features like females (92.50% (148/160)) and in those without a smoking history (88.39% (198/224)). As in previous studies in stage I NSCLC, the female sex and never-smokers were significant protective factors for good survival (43, 44). Moreover, we also found some significant correlations of ANC with other blood fractions, particularly AMC. Although these fractions did not show any prognostic value for either DFS or OS in our study, a study demonstrated the prognostic usefulness of AMC in stage I NSCLC (44). These results may contribute to the explanation of the positive role of ANC in survival in our study.

In recent years, cancer dissemination was found to be an early event (45), and these detached cells, also known as circulating tumor cells (CTCs), were found to act as precursors of metastasis and play a key role in recurrence and treatment failure in many cancers (46–48), including lung cancer (49, 50). It is also notable that these cells are found to be a powerful generator of CEA in lung cancer (51, 52). Taking into consideration the intrinsic degradation of CEA in the liver in patients, it was plausible that stage I cases would have a low frequency of CTCs when CEA is maintained in the normal range. Interestingly, neutrophils were found to play an important role in regulating lung cancer cells; for example, they could manipulate tumor angiogenesis and enhance the hypoxic microenvironment and Snail expression, which could then promote cell growth and disease progression (34, 53). In addition, they can also generate a unique structure, namely, neutrophil extracellular traps (NETs) (54), which can support metastasis (55). Furthermore, although not reported in lung cancer, neutrophils could interact with CTCs and escort these cells to enable cell cycle progression (56) and contribute to their survival by inhibiting peripheral leukocyte activation (57) or the formation of metastatic lesions (58, 59). Based on these facts, we speculate that the expansion of neutrophils, in particular a specific subset of these cells in peripheral blood, could increase the opportunity for their interaction with CTCs, although they presented with a low frequency in stage I cases, and promote the development of these cells as well as the formation of metastatic sites. All these biological processes would then result in poor survival in the patients; however, it was also notable that ANC level was not validated as an independent risk factor for both DFS and OS. In fact, the neutrophils are heterogeneous clusters with different functions in cancer development. For example, these cells in circulation can be divided into high- or low-density neutrophils, in which the low-density populations include mature and immature neutrophils (60). Although not reported in stage I cases, low-density neutrophils are likely to play a more important role in promoting resistance to immunotherapy in NSCLC (61). In recent years, increasing evidence has indicated that these cells are plastic during cancer development (62, 63). The dynamic change of different clusters of these cells can be an explanation for its failure as an independent risk factor in prognosis in present study. Additionally, it was long time established that smoking induced chronic inflammation can resulted in delayed neutrophil clearance (64); whereas quitting smoking after diagnosis can not only resulted in decreased white blood cells (65), but also improved survival in NSCLC patients irrespective of stage (66). It was notable that nearly half of the patients quitted smoking (33/77, data not shown) after surgery in our study, which may be the underlying reason for aforementioned change of ANC level in our study.

Our study also has some limitations except its retrospective nature and relatively small sample size. First, due to the relative short duration of follow up, the events for DFS and OS are rare, which could largely impair the statistical power and the reliability of the conclusions; second, although adjuvant chemotherapy was still under debate in stage IB cases (67), some patients did accept adjuvant treatment according to aforementioned clinical trials (6–8); however, the influence of such treatment cannot be further evaluated in our study due to the absent of these information. Building on these facts, our results should be further validated in future.

Overall, we found that preoperative ANC was the only significant prognostic indicator in stage I NSCLC compared to other peripheral blood cell fractions. Although ANC was found to be a useful prognostic indicator in CEA normal stage I NSCLC; however, it was not validated as an independent prognostic factor and additional studies for its role in prognosis for these patients are still needed in future.