Overall survival and progression-free survival of patients with non-small cell lung cancer and either limited or extensive synchronous metastatic spread in Germany—a population-based cancer registry cohort study

Introduction

Background

Worldwide lung cancer is among the most frequently diagnosed cancers and among the leading causes of death (1). The majority of lung cancer cases is diagnosed with non-small cell lung cancer (NSCLC) (2). Despite a variety of new therapeutic measures, such as targeted treatments, antibody or immunotherapies or highly specific local measures (surgery, stereotactic or norm fractionated radiation), the overall survival (OS) of patients with lung cancer remains poor, with a 5-year absolute survival estimate of around 19% for men and 25% for women (3). In Germany, around half of the 57,000 incident cases of lung cancer diagnosed each year present with distant metastasis (3). However, these are not exclusively patients with a generalized or polymetastatic disease, but rather a heterogeneous patient population. Patients with a limited number of metastases affecting a limited number of organ systems—commonly referred to as oligometastatic disease—are of particular clinical interest. Studies have shown that these patients may benefit from local ablative therapy to distant metastases and primary tumour. Reports indicate an improved survival prognosis that is distinct from patients with polymetastatic disease and that is comparable to that of patients with stage III disease (4-6).

According to the consensus paper by Dingemans et al. (7) and the American Society for Radiation Oncology (ASTRO)/ European Society for Radiotherapy and Oncology (ESTRO) guidelines (8), oligometastasis in NSCLC is defined as the presence of a maximum of five metastases in no more than three organs. As German cancer registries do not record the exact number of metastases, this study took a pragmatic approach based on the TNM-8 classification and the number of affected organ systems documented. The underlying cohort has already been characterised in detail by Bertram et al. (9), who divided it into four groups: three subgroups with limited metastasis [limited intra-thoracic metastasis (M1a), extra-thoracic metastasis (M1b), and multi-organ metastasis (M1c) with a maximum of three affected organ systems], and one group with extensive metastasis [extensive multi-organ metastasis (EMS) with more than three affected organ systems].

Rationale and knowledge gap

Accurate prognostic classification of metastatic NSCLC remains a major challenge in clinical practice. While the 8th edition of the tumor-node-metastasis (TNM) classification (10) introduced important refinements by distinguishing between intra-thoracic (M1a), single-organ extra-thoracic (M1b) and multi-organ metastatic disease (M1c), the prognostic implications of further stratifying M1c patients based on the number of organ systems involved remain underexplored.

Evidence—such as from the German CRISP registry published by Metzenmacher et al. [2023] (11)—suggests that both the extent and distribution of metastases, in particular the distinction between limited (LMOI) and extensive multi-organ involvement (EMOI), may influence survival outcomes [both, OS and progression-free survival (PFS)]. However, robust population-based data to support this distinction are still lacking.

Objective

This study forms part of a series analysing the clinical characteristics, treatment strategies and survival outcomes of patients with stage IV NSCLC and limited or extensive metastasis. This analysis aims to investigate OS and PFS within the subgroups defined by Bertram et al. (limited versus extensive metastasis) (9).

The focus is not on evaluating the effects of individual treatments, but rather on the prognostic significance of subgroup formation. The objective is to determine whether this classification reflects clinically significant differences in survival. Additionally, the influence of the organ systems involved on OS and PFS will be analysed within each subgroup to identify possible prognostic differences between different metastasis locations. The influence of different treatment strategies will be analysed in a later study. We present this article in accordance with the STROBE reporting checklist (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-344/rc).

Methods

Study design and data source

The study is registered with the German Register for Clinical Studies under the reference DRKS00028698. Details of the study design and the data source have been published elsewhere (9). In summary, a retrospective cohort study was conducted based on data from the three German population-based cancer registries. About 27% of the German population live in the catchment area of the registries of Bavaria, Hesse and Schleswig-Holstein. The registries operate on the basis of a nation-wide legal foundation (12) and federal state-specific laws (13) and share the same dataset definition (13). After receipt of the data from the cancer registries, we harmonized and pooled the data. A detailed description of our approach is described by Waldmann et al. [2025] (14).

We included patients with diagnosis of lung cancer (ICD-10 C34) between 2016 and 2020, aged 18 years and older at diagnosis, resident in the respective federal state, and information on the location of the distant metastasis at primary diagnosis (synchronous metastases). Another inclusion criterion was the documentation of tumour, lymph node and distant metastases according to the 8th edition of the TNM classification (10). We excluded DCO cases (death certificate only) and cases with histology other than NSCLC. A detailed overview of the inclusion process can be found in the flowchart in the accompanying paper by Bertram et al. (9).

Definition of limited and extensive synchronous metastases

Our analysis focused on patients with synchronous metastases, defined as all metastases diagnosed either at the time of primary tumour diagnosis or within 92 days thereafter. The 92-day time window allows for staging delays that may occur in routine clinical practice. This ensures that any metastases that are present at diagnosis, but which are not detected until later due to diagnostic constraints, are still classified as synchronous. This provides a more accurate representation of the extent of the initial disease.

To approximate oligometastatic disease, we followed the definition proposed in the consensus report by Dingemans et al. (7) and the ASTRO/ESTRO guidelines (8), which define oligometastatic as the presence of one to five metastases in up to three different organ systems. As the exact number of metastases is not consistently available in cancer registry data, we used the 8th edition of the TNM classification—which for the first time distinguishes between single and multiple metastases—in combination with information on the number of organ systems involved.

We categorised patients into the following subgroups of limited and extensive metastatic spread (EMS):

• IT-LMS (intra-thoracic limited metastatic spread): patients with metastases confined to the thoracic cavity, including tumour nodule in the contralateral lung lobe, pleural metastasis or malignant pleural or pericardial effusion. This category corresponds to TNM stage M1a.
• ET-LMS (extra-thoracic limited metastatic spread): patients with a single metastasis outside the thoracic region. This category corresponds to TNM stage M1b.
• LMOI: patients with multiple distant metastases involving up to three different organ systems. This category is based on TNM stage M1c and represents a limited metastatic burden.
• EMOI: patients with metastases in more than three organ systems or those classified as having generalised metastatic disease [coded “GEN” (12)]. This category is based on TNM stage M1c and represents an extensive metastatic burden.

Subgroups 1 to 3 represent an approximation of oligometastatic disease and are collectively referred to as patients with limited metastatic spread (LMS), while subgroup 4 reflects a pattern of extensive, polymetastatic disease (EMS).

Follow-up information (progression, death)

Information on progression, recurrence, the diagnosis of distant metastasis and death during follow-up is subject to mandatory reporting by physicians to the cancer registries (12). Further, the cancer registries regularly compare their databases with the residents’ registration offices (automated, probabilistic data linkage process) and have access to all death certificates in the respective federate state, which means that information on vital status (vital status alive/deceased; date of comparison/linkage or date of death) is available in the registry in a standardized form and systematically assessed. The last systematic vital status assessment relevant for data provision took place in May 2020 in Hesse, in June 2021 in Bavaria, and in Schleswig-Holstein in February 2022.

Statistical analyses

The analyses were stratified according to the four prespecified subgroups (see Section “Definition of limited and extensive synchronous metastases”). Patient characteristics were summarised using medians and interquartile ranges (IQRs, 25th–75th percentiles) for continuous variables and absolute and relative frequencies for categorical variables. Descriptive variables included demographic and clinical parameters such as age, sex, Eastern Cooperative Oncology Group (ECOG) performance status, histological subtype, tumour location, tumour stage, nodal and metastatic status, and metastatic sites. Proportion of missing information for each variable is also given.

OS and PFS were estimated using the Kaplan-Meier method. OS was defined as the time from primary tumour diagnosis (corresponding to the date of metastasis) to death, censoring, or the date of the last systematic vital status update. Only patients with at least one day of follow-up were included in the OS analysis.

To reduce immortal time bias in PFS estimation, a landmark method was applied. For PFS, a minimum of 93 days of follow-up was required to assess progression. Patients who progressed within 92 days of diagnosis were excluded from the PFS analysis. PFS was calculated from a landmark of 92 days after diagnosis, until the earliest occurrence of progression, death, censoring or the last vital status update. Patients without an event were censored at their last follow-up.

When the number of patients still at risk in a subgroup felt to 10 or fewer, the Kaplan-Meier curve was truncated from that point. Median and interquartile survival times, the number of patients still at risk at specific time points (12, 24, 36 and 48 months) after the specified index date, and 1- to 4-year survival rates were reported.

Finally, multivariable Cox proportional hazards models were used to analyse the association between survival and the extent and location of metastases. First, a model was fitted that considered the extent of metastatic spread as the only explanatory variable. Subsequently, separate models were calculated for each subgroup (IT-LMS, ET-LMS, LMOI and EMOI), taking the respective metastasis sites into account. Demographic and clinical covariates (age, sex, histology, tumour stage, tumour location and ECOG performance status) were included in the models as confounders, but these are not reported in the results section. The selection of covariates was based on the clinical expertise of the study team and supported by findings from previous literature, including the study by Metzenmacher et al. (11) or Wiesweg et al. (15). Missing values in categorical variables were handled by introducing a separate ‘unknown’ category. Hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated for all models. We assessed the proportional hazards assumption visually by inspecting the Schoenfeld residuals for each independent variable. Multicollinearity, outliers and influential data points were also examined, and a sensitivity analysis was performed excluding identified outliers.

Supplementary analyses and results are available on the Open Science Framework (OSF), including the univariate and full multivariable models with all covariate effects: https://osf.io/5kcvx/. Data management and analysis was performed in R (version 4.1.3) (16). The main packages used were tidyverse (version 2.0.0) (17), survival (version 3.7-0; cox model assumptions cox proportional hazard model) (18), survminer (version 0.4.9; median survival time, x-year estimates) (19), ggsurvfit (version 1.0.0; Kaplan-Meier plots) (20) and finalfit (version 1.0.8; cox model, hazard ratio table) (21). The code used in this study can be found on the OSF platform at the following link: https://osf.io/5kcvx/.

Ethics

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. In Germany, there is an obligation to register cancer patients. At the time of diagnosis, patients are informed by their doctors about the registration of information on their disease, treatment and the course of the disease in the cancer registry of the respective federal state. The cancer registry laws of the federal states permit the transfer of de facto anonymous data to third parties for the purpose of health services research (22-25). At the time of data retrieval from the cancer registries, no ethics vote was required in Germany for analyses based on de facto anonymous data without patient contact. Nevertheless, we reported the research project to the ethics committee of the University of Lübeck. The ethics committee approved the study protocol (file number: Az 22-083).

Results

Description of the patients

A total of 8,033 patients were included in the analysis. Of these, 7,277 had LMS, including 1,767 patients with IT-LMS, 1,314 with ET-LMS and 4,196 with LMOI. The remaining 756 patients were classified as having an EMS. The median age varied by subgroup, ranging from 66 to 70 years in the LMS groups and 64 years in the EMS group. Starting from the date of diagnosis, patients with LMS had a median follow-up of 8 months (IQR, 3 to 16 months), compared to 4 months (IQR, 2 to 9 months) in the EMS group (see the characteristics table of the cohort on OSF: https://osf.io/5kcvx).

A detailed description of the cohorts is provided elsewhere (9).

OS and PFS of patients with synchronous metastatic NSCLC according to the extent of metastatic spread

Figure 1 shows the Kaplan-Meier curves for OS (Figure 1A) and PFS (Figure 1B) in patients with NSCLC, stratified by the extent of synchronous metastatic spread.

Figure 1 Kaplan-Meier curves for overall (A) and progression-free survival (B), stratified by extent of synchronous metastatic spread in NSCLC. OS measured from diagnosis (n=8,033); PFS from day 93 using landmark method (n=5,336). EMOI, extensive multi-organ involvement; ET-LMS, extra-thoracic limited metastatic spread; IT-LMS, intra-thoracic limited metastatic spread; LMOI, limited multi-organ involvement; NSCLC, non-small cell lung cancer; OS, overall survival; PFS, progression-free survival.

The median OS from diagnosis was 15 months for patients with IT-LMS, 12 months for ET-LMS and 8 months for LMOI. By contrast, patients with EMOI had a median OS of just 4 months. The differences in survival between the subgroups were statistically significant (P<0.001 for each).

At 2 years, OS in the LMS subgroups ranged from 21.5% to 33.4%, compared to 11.3% in the EMS group. At 4 years, OS in the LMS subgroups ranged from 8.5% to 13.0%. Survival could not be reliably estimated for the EMS group, due to the small number of patients at risk (n<10; see Table 1).

Patients who died or experienced disease progression within 92 days after diagnosis were excluded from the PFS analysis, resulting in a study population of 5,336 patients (4,988 with LMS and 348 with EMS). PFS was analysed using a landmark method starting 92 days after diagnosis. Median PFS was 8.5 months for IT-LMS, 7.1 months for ET-LMS and 5.8 months for LMOI. EMOI patients had a median PFS of 4.0 months (P<0.001). At 2 years, PFS in the LMS subgroups ranged from 14.5% to 19.8%, compared to 8.9% in the EMS group. Four-year PFS could not be estimated for any subgroup due to the small number of patients at risk (see Table 1).

Cox proportional hazard regression model for OS and PFS

Multivariable Cox regression analyses of OS and PFS, adjusted for age, sex, histology, tumour stage, tumour location, and ECOG performance status, showed significantly poorer survival in all subgroups compared to IT-LMS. The highest risk was observed in the EMOI group (HR =1.87, 95% CI: 1.54–2.28 for OS and HR =1.91, 95% CI: 1.67–2.18 for PFS), followed by the LMOI group (HR =1.56, 95% CI: 1.45–1.68 for OS and HR =1.35, 95% CI: 1.25–1.46 for PFS). A smaller but still significant effect was seen for the ET-LMS group regarding OS (HR =1.16, 95% CI: 1.06–1.27) (see Table 2).

To evaluate how individual metastasis locations influence OS and PFS, separate multivariable Cox models were calculated for each of the four subgroups (see Table 3).

Notably, in the LMOI group, there was a clear association between certain metastasis locations and survival. Metastases in the adrenal glands (HR =1.16; 95% CI: 1.05–1.27), the brain (HR =1.27; 95% CI: 1.16–1.40), the liver (HR =1.26; 95% CI: 1.15–1.39), the bones (HR =1.24; 95% CI: 1.14–1.36), and other locations (HR =1.23; 95% CI: 1.11–1.36) were each significantly associated with poorer OS.

A comparable pattern was also observed for PFS within the LMOI group. Significantly unfavourable HRs were found for metastases in the adrenal glands (HR =1.15; 95% CI: 1.03–1.29), the brain (HR =1.19; 95% CI: (.06–1.34), the liver (HR =1.19; 95% CI: 1.06–1.34), the skeletal system (HR =1.20; 95% CI: 1.08–1.33), and other organ manifestations (HR =1.19; 95% CI: 1.06–1.35).

In the ET-LMS group, the presence of metastases in other organs was also significantly associated with poorer PFS (HR =1.64; 95% CI: 1.05–2.55).

Discussion

Key findings and comparison to similar researches

Lung cancer is a disease that is associated with a high mortality. Patients with a limited number of metastases in a limited number of organ systems, known as oligometastatic disease, represent a distinct subgroup that may benefit from targeted treatment approaches. In recent years, small studies have been conducted to investigate the clinical benefit of local ablative therapy in oligometastatic disease, highlighting the potential to improve survival in this patient population. However, the limited reference literature combined with the lack of a consensus definition, has led to inconsistent definitions of oligometastatic disease patients, particularly with regard to the number of metastases, the organs involved and the exclusion criteria used (26,27).

In our study, we analysed OS and PFS in approximately 8,000 patients diagnosed with stage IV NSCLC between 2016 and 2020. To approximate oligometastatic and polymetastatic disease, we stratified patients based on the M category of the 8th edition TNM classification and the number of organ systems involved. We defined three subgroups representing LMS: intra-thoracic LMS (IT-LMS; M1a), single extra-thoracic LMS (ET-LMS; M1b), and limited multi-organ involvement (LMOI; M1c with metastases in up to three organ systems). Patients with metastases in more than three organ systems or with generalised disease (EMOI) were considered as having extensive, polymetastatic disease. As the cancer registry data used did not consistently capture the exact number of metastases per organ, this classification served as a pragmatic approximation based on the available information.

We believe that these are the key findings of our study:

• We approximated oligometastatic disease with a pragmatic TNM-based classification and termed it LMS. This approach allowed reproducible stratification of patients according to the extent and location of metastases and provided a clear prognostic distinction from extensive, polymetastatic disease in terms of OS and PFS, despite the lack of precise metastasis counts in the registry data.
• Within the M1c category, further stratification based on the number of organ systems involved (≤3 vs. >3) showed significant survival differences—with a median OS difference of 4 months. This suggests that the M1c group is not prognostically homogeneous and may benefit from additional subclassification.
• Multivariable Cox regression analyses confirmed that the extent and location of metastases were both independently associated with survival outcomes. Patients with LMOI or EMOI experienced significantly poorer OS and PFS than those with IT-LMS. Within the LMOI subgroup, specific metastatic sites, including the liver, brain, bones, adrenal glands and others, were found to be significantly associated with a poorer prognosis after adjusting for age, sex, histology, tumour stage, tumour location and ECOG performance status.

Survival differences between subgroups with limited and extensive metastases (key findings 1 and 2)

Based on our large unselected, population-based cohort of stage IV NSCLC patients, we were able to demonstrate that differentiating the extent of metastatic spread is clinically relevant for both OS and PFS. The observed survival differences between the TNM categories M1a, M1b and M1c confirm the prognostic relevance of this classification, which was introduced with the introduction of the 8th edition of the TNM.

The median OS was longest for patients with M1a disease (IT-LMS), at 15 months, followed by M1b (ET-LMS) at 12 months, M1c ≤3 organs (LMOI) at 8 months, and M1c >3 organs (EMOI) at 4 months. This clear distinction highlights the prognostic value of subclassifying M1c, a factor that has only been considered in a few studies to date.

A recent monocentric cohort study by Wiesweg et al. [2025] (15) with 218 patients observed significantly longer survival times using the TNM8 classification: M1a (37 months), M1b (28 months) and M1c (21 months) for ≤2 organs. The discrepancy between these results and ours is likely due to the selection of fitter patients (ECOG 0–1 in 84.9% vs. <10% in our cohort) and the high proportion of patients with local treatment (70%) in Wiesweg et al.’s (15) study. By contrast, only 43.2% of patients in our cohort received systemic therapy without additional local treatment (see the OSF link: https://osf.io/5kcvx/). The definition of M1c was also stricter (a maximum of two organ systems affected). A further breakdown of the M1c category has been considered in smaller studies, such as the analysis by Metzenmacher et al. [2023] (11), according to the number of affected organ systems. However, this breakdown remains rarely used in clinical or population-based research. We used an approach comparable to Metzenmacher et al. (11) and subdivided the M1c group into LMOI (limited multiorgan involvement) and EMOI (extensive multiorgan involvement) with the aim to replicate and confirm this stratification in a large, real-world cohort of over 8,000 patients. Metzenmacher et al. (11) also investigated the prognostic significance of the TNM-M categories in a Germany-wide, real-world cohort (the CRISP registry) comprising 2,082 patients with NSCLC. In addition to TNM8-based stratification, the M1c category was differentiated further according to the number of organs affected (≤3 vs. >3) and the presence of liver metastases. Our results are similar to the results reported in their study in many areas: For M1a, the median OS was 16.4 months (versus 15 months in our study); for M1c with ≤3 affected organs, it was 8.2 months (versus 8 months in our study); and for M1c with >3 affected organs, it was 6.6 months (versus 4 months in our study). The median PFS was also in a similar range: around 7.7 months for M1a (versus 8.5 months in our study) and 5.0 months for M1c with ≤3 organs affected (versus 5.8 months).

However, it is striking that, in the CRISP cohort, patients with M1b disease showed a higher median OS (18.8 months without liver metastases) than M1a patients—a finding that could not be reproduced in our analysis (M1b: 12 months; M1a: 15 months). This discrepancy may be due to methodological differences: while we calculated OS time from the time of diagnosis (PFS per landmark approach from day 93), follow-up in Metzenmacher et al.’s study began with the start of first-line systemic therapy. Differences in the definition of synchronous metastasis [not specified by Metzenmacher et al. (11)] and the potential selection bias resulting from the exclusion of ECOG >1 patients from the CRISP registry may also have contributed to the observed differences in survival outcomes. Although the general trend of decreasing survival with increasing metastatic burden is consistent with previous literature, the absolute survival estimates in our real-world cohort are generally lower than those reported in clinical trials (4,28-30) as well as in the analysis by Wiesweg et al. (15) and Metzenmacher et al. (11). For example, Guberina et al. (4) reported a 5-year OS of 28.3% in patients with synchronous oligometastatic NSCLC (cT3/cT4 or cN2/cN3 stage), while Metzenmacher et al. (11) reported a 4-year OS of 20.1% in the M1a cohort and 25.8% in the M1b cohort. In comparison, the 4-year OS in our cohort ranged from 8.5% to 13% depending on the LMS subgroup. These differences probably reflect the broader inclusion criteria and the more heterogeneous, unselected patient population in our study based on population-based cancer registry data. Our cohort included a higher proportion of patients with M1c disease (LMOI and EMOI: 57%), an older age at diagnosis, and a larger proportion of patients with poorer overall health (ECOG >0) compared to Guberina et al. (4). In addition, patients in the Guberina et al. (4) study received an intensive multimodal treatment regimen including local ablative therapy, surgery, and chemoimmunotherapy—approaches presumably reserved for fitter patients.

The treatment details of our cohort have not yet been systematically evaluated. However, initial evaluations suggest that 43.2% of patients received systemic therapy alone, including those with extensive metastasis (EMOI). In addition to patient-related factors such as age and general health, differences in treatment could help to explain the comparatively lower survival rates.

Prognostic relevance of the extent of metastatic spread and metastasis location (key finding 3)

Five multivariable Cox models were calculated to investigate the prognostic relevance of the extent of the metastatic spread and the affected organ systems: one model for metastasis extent and one model for metastasis location within each subgroup. All models were adjusted for age, sex, histology, tumour stage, tumour location and ECOG performance status.

In the metastasis extent model, the ET-LMS (M1b), LMOI (M1c with ≤3 affected organs) and EMOI (M1c with >3 affected organs) subgroups exhibited significantly poorer OS than the IT-LMS (M1a) reference group. For PFS, only the results for the LMOI and EMOI subgroups were significant.

Similar analyses were conducted by Metzenmacher et al. (11), who took into account comparable covariates (excluding tumour location but including BMI, the Charlson Comorbidity Index and smoking status). In their study, the difference in OS/ PFS between M1a and M1b was not significant, which may be due to the smaller M1b cohort (n=249 vs. 629 in M1a), amongst other things. In our analysis, the HR for M1b versus M1a in OS was significant (HR =1.16, 95% CI: 1.06–1.27), albeit of limited clinical relevance.

For the ET-LMS cohort (M1b), our model showed no significant effect of individual metastasis locations on OS or PFS. Metzenmacher et al. (11) reported a significantly adverse effect of liver metastases (compared to other sites), but not for bone, brain, or lymph node metastases. One possible reason for this discrepancy is the different modelling approach: While Metzenmacher et al. (11) used a categorical reference group (‘other sites’), we examined the binary effect of liver metastases (present/absent) in our models.

For the M1c cohort, Metzenmacher et al. (11) did not consider individual metastases, but rather analysed the number of affected organs (≤3 vs. >3). The involvement of more than three organs was associated with a significantly worse prognosis for OS and PFS, which is consistent with our results for the EMOI group.

In our analysis, significant effects of individual metastatic locations were only observed in the LMOI cohort (M1c with ≤3 affected organs), rather than in the EMOI group (>3 organs). These differences may be due to interactions between locations, which are more difficult to discern with greater organ involvement.

Campos-Balea et al. [2020] (31) conducted a large-scale SEER-based analysis of 46,030 patients with metastatic NSCLC to examine the prognostic impact of individual metastasis locations (brain, liver, lung and bone; each categorised as present or absent). This analysis also included cases with multiple organ involvement. A significantly poorer prognosis was also observed for liver, brain and bone metastases in this study, a finding supported by our results for the EMOI group, which also showed poorer survival.

Our findings regarding the poor prognosis of liver, brain and bone metastases in the LMOI cohort are consistent with a systematic review by Wang et al. [2025] (32), which confirmed the negative prognostic impact of these metastases in metastatic NSCLC based on numerous studies. This underscores the general relevance of metastasis location for risk stratification.

Strengths and limitations

To our knowledge, this is the largest population-based analysis of NSCLC patients with limited or extensive metastatic disease using real-world population-based cancer registry data from multiple regions in Germany. A major strength is the standardised assessment of vital status across registries and the inclusion of an unselected cohort without the restrictive inclusion and exclusion criteria typically found in randomised controlled trials, which enhances the generalisability of our findings. However, short follow-up periods were observed in some patients, i.e. in those with diagnosis in 2019/2020 and the last systematic assessment of vital status in 2020/2021. This may lead to an underestimation of OS and PFS, especially in cases where patients have not yet experienced disease progression or death, and may introduce bias into survival estimates.

In addition, important clinical variables such as smoking history, pre-existing conditions, comorbidities or socioeconomic status are not documented in cancer registry data. Although ECOG performance status is sometimes recorded, it is missing in more than 50% of cases. As these variables are relevant to prognosis, their absence limits the ability to fully adjust for potential confounders, particularly in subgroup analyses. To assess the robustness of our results, we performed a sensitivity analysis in which we excluded ECOG from the multivariable models. The results remained largely consistent, indicating that the observed associations between metastatic patterns and survival are not solely dependent on performance status. However, we acknowledge that residual confounding factors cannot be completely ruled out, and we recommend that this aspect be further explored in future analyses with more complete clinical data.

A key methodological weakness is the lack of documentation of the exact number of metastases in each organ. This made it impossible to apply the consensus definitions for oligometastatic NSCLC as defined by Dingemans et al. (7). Our classification is therefore based on the M category of the TNM classification (8th edition), combined with the number of organ systems affected. ET-LMS (M1b) and EMOI (affecting more than three organs) can be clearly assigned to oligometastatic or polymetastatic disease, respectively. However, in the ET-LMS and LMOI groups, it remains unclear whether individual patients have more than five metastases and should therefore be considered polymetastatic by definition. Nevertheless, the clear differences in survival between the subgroups highlight the prognostic relevance of the chosen classification. In future, standardised recording of the number and location of metastases is needed to adequately reflect the growing importance of oligometastatic diseases.

Including only patients who had received at least one treatment in our analysis may have introduced selection bias, resulting in overly optimistic survival estimates. The survival curves for patients without treatment information can be found at: https://osf.io/5kcvx/. These curves show lower OS and PFS in this subgroup. However, the OS and PFS estimates presented here are broadly consistent with those reported in the literature on real-world survival.

One of the central aims of this study was to investigate whether the chosen classification, based on the extent and location of metastases, is associated with survival, regardless of treatment. Treatment was therefore deliberately excluded from the multivariable models. A separate analysis will be conducted to investigate the influence of different treatment modalities on survival prospects.

However, the lack of adjustment for treatment represents a methodological limitation, as differences in survival could be at least partially explained by different treatment strategies. To provide an initial overview of the treatment received by our cohort, we are making information on the entire cohort available via the following OSF link: https://osf.io/5kcvx/.

Implications and actions needed

In a previous analysis, we were able to describe that it is feasible and clinically meaningful to distinguish patients according to the number of organs involved and the location of distant metastases, as these subgroups differ in terms of baseline characteristics. In the current analysis, we have found that these subgroups also differ in terms of OS and PFS. These findings—based on real-world, population-based data—may help clinicians to identify groups of patients with better OS and PFS who may benefit most from aggressive, local ablative treatment strategies.

The association of different treatment strategies with OS and PFS needs to be further elucidated.

Conclusions

We successfully stratified a heterogeneous population of patients with limited metastatic disease (approximating oligometastatic disease) into clinically meaningful subgroups using a pragmatic classification based on TNM staging and the number of organ systems affected. These subgroups exhibited distinct survival outcomes, with a greater metastatic burden being associated with poorer OS and PFS. Our findings suggest that patients with limited metastatic disease may benefit from more intensive treatment approaches that target both the primary lung tumour and metastatic sites, rather than solely palliative strategies. Further studies are needed to determine whether survival differences are due to the effects of treatment and to refine treatment selection for these patient subgroups.

Acknowledgments

We would like to thank the cancer registries in Bavaria, Hesse and Schleswig-Holstein for providing the data for this research project.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-344/rc

Peer Review File: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-344/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-344/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of the University of Lübeck, Germany (No. Az 22-083).

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Cancer (IARC) TIA for R on. Global Cancer Observatory. Available online: https://gco.iarc.fr/ (accessed 29 April 2024).
2. Açikgöz Ö, Bilici A, Özyükseler DT, et al. Survival outcomes of patients with oligometastatic non-small cell lung cancer who were treated with radical therapy: a multicenter analysis. Turk J Med Sci 2023;53:949-61. [Crossref] [PubMed]
3. Robert Koch-Institut; Gesellschaft der epidemiologischen Krebsregister in Deutschland e.V. Krebs in Deutschland für 2019/2020. 14th edn. Berlin. Available online: https://doi.org/10.25646/11357 (2023).
4. Guberina M, Pöttgen C, Guberina N, et al. Long-Term Survival in Patients with Oligometastatic Non-Small Cell Lung Cancer by a Multimodality Treatment-Comparison with Stage III Disease. Cancers (Basel) 2024;16:1174. [Crossref] [PubMed]
5. Petrelli F, Ghidini A, Ghidini M, et al. Better survival of patients with oligo- compared with polymetastatic cancers: a systematic review and meta-analysis of 173 studies. F1000Res 2021;10:423. [Crossref] [PubMed]
6. Parikh RB, Cronin AM, Kozono DE, et al. Definitive primary therapy in patients presenting with oligometastatic non-small cell lung cancer. Int J Radiat Oncol Biol Phys 2014;89:880-7. [Crossref] [PubMed]
7. Dingemans AC, Hendriks LEL, Berghmans T, et al. Definition of Synchronous Oligometastatic Non-Small Cell Lung Cancer-A Consensus Report. J Thorac Oncol 2019;14:2109-19. [Crossref] [PubMed]
8. Iyengar P, All S, Berry MF, et al. Treatment of Oligometastatic Non-Small Cell Lung Cancer: An ASTRO/ESTRO Clinical Practice Guideline. Pract Radiat Oncol 2023;13:393-412. [Crossref] [PubMed]
9. Bertram S, Schliemann A, Katalinic A, et al. Characteristics of patients with non-small cell lung cancer and either limited or extensive synchronous metastatic spread in Germany—a population-based cancer registry cohort study. Transl Lung Cancer Res 2025; [Crossref]
10. Wittekind C. TNM Klassifikation maligner Tumoren. 8. Auflage. Weinheim: Wiley-VCH Verlag. Available online: https://www.wiley-vch.de/de/fachgebiete/medizin-und-gesundheit/tnm-klassifikation-maligner-tumoren-978-3-527-34280-8 (2017).
11. Metzenmacher M, Griesinger F, Hummel HD, et al. Prognostic factors in nonsmall cell lung cancer: insights from the German CRISP registry. Eur Respir J 2023;61:2201336. [Crossref] [PubMed]
12. Gesetz zur Weiterentwicklung der Krebsfrüherkennung und zur Qualitätssicherung durch klinische Krebsregister (Krebsfrüherkennungs- und -registergesetz – KFRG). Bundesgesetzblatt Teil 2013;I:617.
13. Bundesministerium für Gesundheit. Bekanntmachung: Aktualisierter einheitlicher onkologischer Basisdatensatz der Arbeitsgemeinschaft Deutscher Tumorzentren e.V. (ADT) und der Gesellschaft der epidemiologischen Krebsregister in Deutschland e.V. (GEKID) Vom 27. März 2014. BAnz AT 2014;28.04.2014: 1–28.
14. Waldmann A, Labohm L, Baltus H, et al. Merging of comprehensive data and completeness of information on diagnosis, treatment and progression from several population-based cancer registries in Germany – initial experiences using the example of lung cancer. GMS Med Inform Biom Epidemiol 2025;21:Doc02. DOI: 10.3205/mibe000274.
15. Wiesweg M, Küter C, Schnorbach J, et al. Oligometastatic non-small cell lung cancer: Impact of local and contemporary systemic treatment approaches on clinical outcome. Int J Cancer 2025;156:776-87. [Crossref] [PubMed]
16. R Core Team. R: A language and environment for statistical computing, R Foundation for Statistical Computing. Available online: https://www.R-project.org/ (2021).
17. tidyverse 2.0.0. Available online: https://www.tidyverse.org/blog/2023/03/tidyverse-2-0-0/ (accessed 19 February 2025).
18. Therneau TM, until 2009) TL (original S->R port and R maintainer, Elizabeth A, et al. survival: Survival Analysis. Available online: https://cran.r-project.org/web/packages/survival/index.html (2024, accessed 30 July 2024).
19. Kassambara A, Kosinski M, Biecek P, et al. survminer: Drawing Survival Curves using ‘ggplot2. Available online: https://cran.r-project.org/web/packages/survminer/index.html (2021, accessed 30 July 2024).
20. Sjoberg DD, Baillie M, Fruechtenicht C, et al. ggsurvfit: Flexible Time-to-Event Figures. Available online: https://cran.r-project.org/web/packages/ggsurvfit/index.html (2024, accessed 30 July 2024).
21. Harrison E, Drake T, Pius R. finalfit: Quickly Create Elegant Regression Results Tables and Plots when Modelling. Available online: https://cran.r-project.org/web/packages/finalfit/index.html (2024, accessed 10 October 2024).
22. Gesetz über das Krebsregister des Landes Schleswig-Holstein (Krebsregistergesetz - KRG SH) Vom 4. November 2015.
23. Bayerisches Krebsregistergesetz (BayKRegG) - Vom 7. März 2017 (GVBl. S. 26), Available online: https://www.gesetze-bayern.de/Content/Document/BayKRegG/true (accessed 2 August 2024).
24. Hessisches Krebsregistergesetz Vom 15. Oktober 2014 (GVBl. S. 241). Available online: https://www.rv.hessenrecht.hessen.de/bshe/document/jlr-KRGHE2014V4P1 (accessed 2 August 2024).
25. Gesetz- und Verordnungsblatt für Schleswig-Holstein, Ausgabe Nr. 16, Kiel, 15. Dezember 2022. Available online: https://www.schleswig-holstein.de/DE/landesregierung/ministerien-behoerden/IV/Service/GVOBl/GVOBl/2022/gvobl_16_2022.pdf?__blob=publicationFile&v=5
26. Wang XS, Bai YF, Verma V, et al. Randomized Trial of First-Line Tyrosine Kinase Inhibitor With or Without Radiotherapy for Synchronous Oligometastatic EGFR-Mutated Non-Small Cell Lung Cancer. J Natl Cancer Inst 2023;115:742-8. [Crossref] [PubMed]
27. Palma DA, Olson R, Harrow S, et al. Stereotactic Ablative Radiotherapy for the Comprehensive Treatment of Oligometastatic Cancers: Long-Term Results of the SABR-COMET Phase II Randomized Trial. J Clin Oncol 2020;38:2830-8. [Crossref] [PubMed]
28. Griffioen GH, Toguri D, Dahele M, et al. Radical treatment of synchronous oligometastatic non-small cell lung carcinoma (NSCLC): patient outcomes and prognostic factors. Lung Cancer 2013;82:95-102. [Crossref] [PubMed]
29. Tönnies M, Pfannschmidt J, Bauer TT, et al. Metastasectomy for synchronous solitary non-small cell lung cancer metastases. Ann Thorac Surg 2014;98:249-56. [Crossref] [PubMed]
30. Friedes C, Mai N, Hazell S, et al. Consolidative Radiotherapy in Oligometastatic Lung Cancer: Patient Selection With a Prediction Nomogram. Clin Lung Cancer 2020;21:e622-32. [Crossref] [PubMed]
31. Campos-Balea B, de Castro Carpeño J, Massutí B, et al. Prognostic factors for survival in patients with metastatic lung adenocarcinoma: An analysis of the SEER database. Thorac Cancer 2020;11:3357-64. [Crossref] [PubMed]
32. Wang S, Tang W, Jin F, et al. Comprehensive Analysis of Lung Cancer Metastasis: Sites, Rates, Survival, and Risk Factors-A Systematic Review and Meta-Analysis. Clin Respir J 2025;19:e70107. [Crossref] [PubMed]