A Rare Subtype of Non-small Cell Lung Cancer: Report of 159 Resected Stage I-IIIA Pulmonary Lymphoepithelioma-like Carcinoma Cases

Background The current study analyzed resected stage I-IIIA pulmonary lymphoepithelioma-like carcinoma (LELC) cases to dene the clinical characteristics, prognosis and long-term outcomes of LELC, with the purpose of guiding clinical management for this rare tumor. Methods Resected stage I-IIIA LELC, adenocarcinoma (ADC), squamous cell carcinoma (SCC) and adenosquamous carcinoma (ASC) cases from our center were enrolled. Propensity score matching (PSM) was applied to minimize the selection bias. Overall survival (OS) and disease-free survival (DFS) were compared between groups. Multivariate analyses were performed to identify the prognostic factors, and a nomogram was developed. Results A total of 159 LELCs, 2,757 ADCs, 1,331 SCCs and 155 ASCs were included. LELC, dominated among younger patients and nonsmokers, always presented without typical imaging manifestations of lung cancer. LELC was a poorly differentiated disease that lacked driver gene mutations and was positive for immunohistochemistry indicators of squamous cell lineage. Survival analyses revealed that OS was signicantly better for LELC than for other common non-small cell lung cancer (NSCLC) both before PSM (all P < 0.001) and after PSM (all P < 0.05). Further analyses revealed that early pathological node stage and preoperative albumin level ≥ 35 were identied as independent prognostic factors favoring OS and DFS. Conclusions LELC, dominated among younger and nonsmoking populations, showed a lower extent of malignancy regarding CT characteristics. It lacked driver gene mutations and was positive for immunohistochemistry indicators of squamous cell lineage. The survival outcome of PSC was better than other common NSCLCs.


Introduction
Primary pulmonary lymphoepithelioma-like carcinoma (LELC), a rare subtype of non-small cell lung cancer (NSCLC), accounts for less than 1% of all lung neoplasms (1) and was rst described in 1987 by Begin (2). According to the World Health Organization (WHO) Classi cation in 2015, it was removed from the subgroup of large cell lung cancer and reclassi ed as a unique subgroup of NSCLC (3). Owing to the inherent rarity and the lack of prospective clinical trials, the natural course, prognosis and management strategy of LELC requires in-depth investigation.
LELC is an Epstein-Barr virus (EBV)-associated and undifferentiated nasopharyngeal-like carcinoma (2,4,5). Previous literature demonstrated that most LELC cases were documented in Southeast Asia including Guangdong Province, Taiwan, Hong Kong and Singapore (6)(7)(8)(9)(10)(11)(12). LELC is more prevalent among younger and nonsmoking populations without sexual predilection (6,9,11,12). In addition, several clinical series suggested that LELC has a favorable survival outcome when compared with other lung cancers (6,8,11,12). Although many efforts have been devoted to LELC research in the past few decades, the general demographics and prognosis remain enigmatic, and larger datasets are warranted to tailor the clinical practice guidelines for this rare disease.
In the current study, we retrospectively reviewed 159 resected stage I-IIIA LELC cases to sketch an outline of the clinicopathological characteristics of the disease. We also compared the overall survival (OS) of LELC with other common lung cancers both before and after propensity score matching (PSM) with the purpose of helping clinicians estimate individual survival and select a proper treatment strategy.

Patient Selection
Consecutive resected patients diagnosed with LELC between 1990 and 2016 from the Sun Yat-sen University Cancer Center (SYSUCC) were retrospectively enrolled. In addition, resected patients diagnosed with adenocarcinoma (ADC), squamous cell carcinoma (SCC) and adenosquamous carcinoma (ASC) from 2001 to 2016 were also included in this study.
All included cases t the following criteria: [1] pathologically diagnosed as stage I-IIIA disease and [2] surgical resection was performed. The exclusion criteria were as follows: [1] previous or concurrent other primary cancers; [2] age < 18 years old; [3] underwent neoadjuvant therapy and [4] clinicopathological information was unavailable.
This study was approved by the Institutional Review Board of Sun Yat-sen University Cancer Center. All the included patients signed the informed consent. All methods involved in our article were performed in accordance with the Declaration of Helsinki. The authenticity of this article has been validated by uploading the key raw data onto the Research Data Deposit public platform (www.researchdata.org.cn), with the approval RDD number as RDDA2020001729. The dataset generated for this study are available on request to the corresponding authors.

Data Collection
Clinical, pathological, CT and immunohistochemistry (IHC) data were retrieved from patients' medical records. Clinical variables included age, sex, main complaint, smoking status, tumor history, preoperative albumin level, preoperative complications, surgical approach, surgical type and adjuvant therapy. In terms of age, LELC cases were assigned to 2 groups (≤ 60 years old and > 60 years old) based on the optimal cutoff value determined by X-tile software (13). The preoperative albumin level was dichotomized according to the lower limit of normal. CT features included tumor diameter, location, morphology, speculation, lobulation, pleural indentation, obstructive pneumonia, cavity, clinical tumor (cT) stage, clinical node (cN) stage and clinical tumor-node-metastasis (cTNM) stage. All the relevant data were collected from imaging reports. TNM staging was performed according to the 8th edition of the American Joint Committee on Cancer (AJCC) TNM staging system (14). Pathological characteristics included tumor site, tumor diameter, grade, pleural invasion, lymphovascular invasion, perineural invasion, examined lymph nodes (ELNs), positive lymph nodes (PLNs), pathological tumor (pT) stage, pathological node (pN) stage and pathological tumor-node-metastasis (pTNM) stage. ELNs and PLNs were also dichotomized according to the cutoff value determined by X-tile software. IHC features included CK, CK5/6, CK7, thyroid transcription factor (TTF)-1, P63, Epstein-Bar virus-encoded RNA (EBER), epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK). EGFR testing was performed by the Ampli cation Refractory Mutation System, and ALK testing was performed by in situ hybridization.

Follow-up
In general, postoperative follow-up was carried out every 3 months for the rst 2 years, every 6 months for the next 3-5 years, and annually thereafter (15)(16)(17)(18). At each follow-up visit, a physical examination and chest and abdominal CT scans were performed (15). If the patient had speci c symptoms, the examination was performed as soon as possible for a more careful assessment (15)(16)(17)(18). Follow-up information was updated in October 2020 to determine patients' vital status.

Statistical Analysis
All statistical analyses were performed using R version 3.5.2 (The R Foundation for Statistical Computing, Vienna, Austria; http://www.r-project.org), IBM SPSS Statistics (version 25.0, IBM Corp, Armonk, NY, USA), X-tile software (13) and GraphPad Prism 8 software. OS was de ned as the interval from the date of surgery to the date of death from any cause or the last follow-up. DFS was de ned as the time from the date of surgery to the date of tumor recurrence or death from any cause. All survival outcomes were estimated by the Kaplan-Meier method with a log-rank test. Univariate and multivariate Cox analyses were used to identify the prognostic factors, and a nomogram was formulated. The concordance index (C-index) was performed to verify the predicted effect of the nomogram (19). A one to one propensity score matching (PSM) method based on age, sex, smoking status, surgical type, ELNs, pT stage, pN stage, pTNM stage and adjuvant therapy was employed to reduce bias (20). Pearson's χ2 test or Fisher's exact test was used to compare categorical variables between groups. X-tile software was used to determine the cutoff value (13). A two-sided P < 0.05 was considered statistically signi cant.

Discussion
In the present study, the patient characteristics, survival and prognosis of resected stage I-IIIA LELC were retrospectively investigated. Our data demonstrated that LELC was more prevalent in younger patients and nonsmokers, with no obvious gender predisposition. LELC often presents as a peripheral irregular lung mass without typical imaging manifestations of lung cancer. Moreover, LELC is a poorly differentiated disease that lacks typical driver gene mutations and is positive for IHC indicators of squamous cell lineage. In further analyses, LELC had a better survival outcome than other common lung cancers both before and after PSM. Finally, multivariate analyses revealed that both early pN stage and preoperative albumin level ≥ 35 were prognostic factors favoring OS and DFS.
In previous study, several clinical series suggested that LELC is often identi ed in younger nonsmokers (4,12,21), and there was no sexual predilection (4,21,22), which was akin to our ndings. The abovementioned result suggested that unlike SCC, smoking might not be the main etiology of LELC (7,12). Most tumors in our cohort were peripheral and had irregular morphology, echoing previous reports (23,24), but con icting with Qin et al.'s study (7). In line with our study, Chen et al. reviewed 42 LELC and 134 SCC cases and demonstrated that LELC lacks typical imaging manifestations of lung cancer such as cavity, calci cation and vascular convergence (8). In our study, CT characteristics indicated that LELC had a lower extent of malignancies. We proposed that if the CT imaging presented as a peripheral irregular mass without typical manifestations of lung cancer, malignancies such as LELC should be suspected.
In our study, almost all the cases were diagnosed as poorly differentiated disease, which was in accordance with previous ndings that LELC is characterized by poorly differentiated tumor cells with prominent nucleoli and large vesicular nuclei (24,25). IHC data showed that our results were similar to those of Jiang et al, where the authors investigated 43 resected LELC patients and concluded that the tumor is typically positive for CK, CK5/6 and P63, which suggests squamous cell lineage, but is negative for TTF-1 and CK7 (26). Similar scenarios were also seen in Qin et al's study (7) and Liang et al's study (4). Owing to the similar morphology and IHC indicators, LELC is often misdiagnosed as SCC (27). Previous reports demonstrated that the presence of EBV in the nuclei of LELC tumor cells is critical for diagnosis. This can be con rmed by EBER in situ hybridization testing (8,28). In our research, EBER was positive in 99.3% of all the tested patients. From our perspective, if the patient originated from an area with a prevalence of EBV infection and presented with a peripheral lung mass, EBER testing was preferred in the pretreatment examination.
In our study, molecular testing revealed that LELC lacked target agent-sensitive mutations (EGFR and ALK). In the study by Hong et al, the authors explored the genetic landscape of LELC and demonstrated a low percentage of typical driver mutations, such as EGFR, BRAF and KRAS (29). The same scenarios were also observed in Wang et al's study (30) (33), coronary heart disease (34) and respiratory diseases (35), which may reduce life expectancy.
The multivariate analysis revealed that pN stage and preoperative albumin level were correlated with OS and DFS of resected stage I-IIIA LELC in our study. It is evidenced that nodal stage is an important in uencing factor for LELC patient survival (8,12,14). For albumin level, Liang et al investigated the outcomes of 52 resected LELCs and demonstrated that the serum albumin level was an independent prognostic factor (4), which was similar to our ndings. Surprisingly, pT stage, pleural invasion, lymphovascular invasion and tumor grade, four important prognosis predictors in other NSCLCs, were not correlated with OS and DFS in our study, suggesting that the natural course and biology of LELC might be different from those of other common NSCLCs.
To the best of our knowledge, this study represents the rst comprehensive and concurrent analysis of resected stage I-IIIA LELC. In addition, the virtues of this study were that it included the largest cohort size and had a long-term follow-up. Additionally, the evaluation of a wide range of clinicopathological variables allowed us to better understand the demographic trends and prognosis of the disease.
However, our study also had some limitations. First, in the era of precision therapy, molecular indicators such as PD-1, PD-L1, KRAS and BRAF were not involved in our study. Second, despite the signi cant advantages provided by a larger case number than has ever been reported before, the cohort size was still limited. Finally, the retrospective nature may have contributed to selection bias. Further efforts on prospective data collection and incorporation of the abovementioned factors are warranted.

Conclusions
In conclusion, LELC is a rare distinct subtype of NSCLC, that prevails in young nonsmokers. It often presents as a peripheral lung mass without typical imaging manifestations of lung cancer. Pathological and IHC ndings con rmed that LELC, a poorly differentiated diseases, lacked typical driver gene mutations and was positive for squamous cell lineage IHC indicators. Further analyses revealed that LELC had a better survival outcome than other common lung cancers.