Prognostic factors and survival prediction of resected non-small cell lung cancer with ipsilateral pulmonary metastases: a study based on the Surveillance, Epidemiology, and End Results (SEER) database

Ipsilateral pulmonary metastasis (IPM), also known as separate or additional tumor nodule(s), is a form of intrapulmonary metastasis in non-small cell lung cancer (NSCLC), including tumors with additional nodule(s) in the same lobe (PM1) and tumors with additional nodule(s) in the ipsilateral different lobe (PM2) [1]. The incidence of IPM in NSCLC ranges from 3–10%, with PM1 occurring in 2–7% of cases and PM2 occurring in 1–3% [2‐6]. Since the latest classification change for PM in the 7th edition American Joint Committee on Cancer (AJCC) tumor-node-metastasis (TNM) staging system, PM1 and PM2 were reclassified as the descriptor of T3 and T4 from T4 and M1, respectively [1].

Although the current AJCC staging system still relies on the location relationship between IPM and the primary tumor [7], previous studies have demonstrated that prognostic factors of IPM are not solely related to its location [3‐6, 8]. In recent years, there has been a rise in both surgical and systemic adjuvant therapy for patients with IPM [9‐11]. Research has shown that NSCLC patients with IPM tend to have a better prognosis compared to those with other descriptors of the T3 or T4 stage [12, 13]. Some recent studies have recommended further downstaging of NSCLC patients with IPM [14, 15]. Therefore, studying the survival outcome and prognostic factors of these patients after surgery remains critical.

Nomogram is a convenient and accurate method to predict survival outcomes, which has become very popular in malignant tumor survival prediction in recent years [16]. While many predictive models for extrapulmonary metastasis in NSCLC have been established, and their accuracy and discrimination had been validated by internal and external data [17‐19], there is a dearth of research on intrapulmonary metastasis with predictive models, and the sample size of prognostic studies for these patients is relatively small. To fill this gap, we plan to construct a nomogram that can accurately predict the survival of these patients using data from the Surveillance, Epidemiology, and End Results (SEER) database.

IPM is a heterogeneous category in NSCLC and survival outcomes can be affected by various factors [5, 6, 8]. However, many previous studies have had limited sample sizes and often reported improved survival outcomes due to surgical treatment resulting in downstaging of IPM. Recent research suggests that further downstaging of IPM may be beneficial, emphasizing the importance of continued surveillance of these patients' prognosis [15]. Fortunately, the SEER database contains a wealth of information regarding the treatment and survival outcomes of IPM patients. Leveraging the large sample size provided by the SEER database, we constructed a nomogram that more accurately and conveniently predicts survival outcomes based on patient characteristics.

Our data suggest that patients over 65 years old have a worse prognosis, which is consistent with other studies on lung cancer [20, 21]. However, the prognostic value of age in relation to IPM has been reported differently [5, 6, 8‐11]. Older patients may have worse outcomes due to an increased prevalence of comorbidities among lung cancer patients [22‐24]. In terms of sex, male patients are widely accepted as having a worse prognosis in other lung cancer studies [25]. Our study found that patients with IPM had poorer survival compared to female patients (HR 1.845, 95% CI 1.558–2.184), which is consistent with previous research [9, 26].

Tumor size is a well-known prognostic factor for NSCLC, and larger tumors have a greater risk of invasion and distant metastasis. In our study, we subdivided primary tumor size into four categories to facilitate statistical analysis by nomogram. We found that primary tumor size was an independent prognostic factor for NSCLC with IPM, with patients having a primary tumor size of ≤ 3 cm having better overall survival (P < 0.001). Nakagawa et al. [5] analyzed 48 pathologically diagnosed IPM patients and found that those with tumor sizes > 30 mm had a worse prognosis compared to those with tumor sizes ≤ 30 mm (HR 2.578, 95% CI 1.006–6.608). For PM2, Wang et al. [15] reported that patients with tumor sizes ≤ 3 cm had better overall survival than T4 patients (HR 0.629, 95% CI 0.455–0.869), and their survival outcome was comparable to that of T3 patients. Patients with tumor sizes > 3 cm had a similar survival outcome to T4 patients. However, Ucvet et al. [6] reported that tumor size was not a prognostic factor for survival outcome, which may be due to the fact that 86% of their enrolled patients had tumor sizes of 1 cm or smaller.

The status of lymph node metastases, as reflected by the N status, is a critical prognostic factor for NSCLC. Lee et al. [27] found that patients with N0 had better survival outcomes compared to N1 and N2 patients with IPM. Nakagawa et al. [5] also reported that mediastinal lymph node metastases were a worse prognostic factor for surgically treated patients with IPM. In our study, N status was identified as the most important prognostic factor with the longest line and highest risk score in the nomogram, particularly for N3 in comparison to N0. However, some studies on IPM have reported no negative effects of lymph node involvement. This discrepancy in our findings could be due to the relatively small sample size of these studies [6, 8].

Pulmonary resection with lymph node removal is the standard treatment for NSCLC [28]. Nwogu et al. [29] suggested that resecting more lymph nodes was associated with better patient survival. Samayoa et al. [30] also reported that removing fewer than ten lymph nodes was associated with a 12% increased risk of death in early-stage NSCLC patients. The previous studies concerning IPM needed more relevant information on lymph node removal. Our data showed that removing more regional lymph nodes was associated with better survival in multivariate analysis, highlighting the importance of thoracic lymphadenectomy for IPM.

Tumor histology, specifically the degree of differentiation of tumor tissue, is another important factor affecting prognosis. Sun et al. [31] reported that histologic grade was significantly associated with survival in NSCLC, with undifferentiated carcinoma carrying an 80% increased risk of death compared to well-differentiated carcinoma. Chung et al. [32] found that poorly differentiated tumors had higher lymph node metastasis rates and worse survival compared to well- or moderately differentiated tumors. For NSCLC with IPM, Salazar et al. [11] found that only patients with tumor grade III had statistically significant better survival outcomes than patients with tumor grade I in PM1, while Li et al. [9] found that tumor grade was not a prognostic factor for 3-year OS in PM2 patients. In our study, patients with a histologic grade of well-differentiated were associated with better overall survival.

In terms of treatment for surgically treated patients, chemotherapy is indicated for those with stage II and IIIa non-small cell lung cancer. Salazar et al. [11] investigated the National Cancer Database for NSCLC patients with PM1, including 528 patients, and found that chemotherapy increased 3-year OS. Similarly, Park et al. [10] enrolled 142 patients with PM1 and demonstrated that chemotherapy improved their 5-year OS and disease-free survival, except for patients with tumors smaller than 4 cm. However, Baum documented patients with PM1 from Heidelberg and Berlin databases and suggested that adjuvant therapy did not improve long-term survival if PM1 was the only pT3N0 descriptor [13]. For PM2, Li et al. [9] reported that adjuvant chemotherapy was associated with improved 3-year OS compared to surgery alone. In our study, we found that patients with IPM had significantly improved survival compared to treatment with only surgery (HR 0.805, 95% CI 0.652–0.994, P = 0.044).

In our study, we meticulously incorporated several independent prognostic factors, including age, sex, primary tumor size, N status, number of regional lymph nodes removed, tumor grade, and chemotherapy, into the construction of our nomogram. To assess its clinical utility, we rigorously evaluated the nomogram's performance using various metrics such as the C-index, calibration curves, ROC curves, and AUC. These comprehensive evaluations consistently demonstrated that our nomogram exhibits exceptional accuracy, discrimination, and practicality when it comes to predicting overall survival. It's worth noting that recent work by Wang et al. [15] introduced a modified classification for ipsilateral pulmonary metastases, which considers primary tumor size and IPM location. While this approach provides valuable insights, it still relies on the AJCC TNM staging system and may not fully capture the nuances of high-risk patient stratification. Notably, our study takes a pioneering step by introducing the first nomogram tailored specifically for predicting the OS of non-small cell lung cancer patients with IPM. What sets our nomogram apart is its ability to effectively distinguish high-risk patients from low-risk ones based on the total nomogram points. This distinction is underscored by the significant differences in OS observed between these two patient groups in our Kaplan–Meier analysis and log-rank test. With a total nomogram point threshold of 96, clinicians can readily identify high-risk patients among those diagnosed with IPM, facilitating more precise treatment decisions and improved patient care.

Our study has several limitations worth noting. Firstly, due to the retrospective nature of our research, there is a potential for selection bias, and certain crucial variables, such as the size and number of separate tumor nodules, were lacking in the SEER database. It is important to highlight that these variables have not consistently demonstrated significant prognostic associations in the literature [6, 15, 33‐35]. Secondly, our study solely relied on data from the SEER database, which may not fully represent the broader population of NSCLC patients with IPM. Therefore, to ensure the generalizability of our nomogram, it is imperative to validate our findings in independent cohorts. Lastly, while our nomogram exhibited strong calibration and discrimination, external validation is indispensable to confirm its performance and clinical utility. Additionally, it's crucial to acknowledge that the SEER database lacks comprehensive data, such as clinical and pathological tumor stages, surgical margins, complete clearance timing of tumors and metastases, specific details regarding the mode, method, and dosage of chemoradiotherapy. Due to insufficient pathological data, we were unable to study the postoperative number of positive lymph nodes and pN. Furthermore, data pertaining to genetic testing and immunotherapy, which are relevant to cancer patient prognosis [36‐38], remain incomplete. Despite these limitations, our study offers a valuable tool for predicting survival outcomes in NSCLC patients with IPM, relying on readily available clinical factors.

In conclusion, age, sex, primary tumor size, N status, number of regional lymph nodes removed, tumor grade, and chemotherapy of patients were the independent prognostic factors for NSCLC patients with IPM. We constructed and internally validated a nomogram to predict 1-year, 3-year, and 5-year OS for NSCLC patients with IPM according to independent prognostic factors. After internal validation, this novel nomogram demonstrated good calibration, discrimination, clinical utility, and practical decision-making effects for the prognosis of NSCLC patients with IPM.