Introduction

Esophageal cancer, a notably aggressive and lethal malignancy, ranks as the seventh most common cancer globally. It is broadly categorized into squamous cell carcinoma and adenocarcinoma1,2. Despite advancements in treatment modalities such as surgery, chemotherapy, radiotherapy, and immunotherapy, esophageal cancer remains a challenging disease with poor prognosis3,4. Recent advancements in systemic treatments, particularly in chemotherapy and immunotherapy, have significantly improved patient outcomes. For instance, immune checkpoint inhibitors (ICIs) like pembrolizumab and nivolumab, especially when combined with other therapies, have led to better survival rates and reduced recurrence in esophageal cancer patients5,6. Additionally, studies in non-small cell lung cancer (NSCLC) with high PD-L1 expression suggest that chemoimmunotherapy may offer better response rates and progression-free survival compared to ICIs alone, though overall survival rates are similar7. Radiotherapy plays a critical role across various stages of the disease and has established itself as a cornerstone in the multifaceted approach to esophageal cancer treatment, offering a blend of therapeutic benefits across various disease stages8,9,10. While radiotherapy enhances local control and potentially improves survival, it also introduces the risk of potential long-term side effects and second primary cancers11,12. As the survival rates for esophageal cancer patients improve due to advancements in treatment modalities, several studies conducted across populations have shown an elevated risk for the development of Second Primary Malignancies (SPM)13,14. The observed Standardized Incidence Ratios (SIRs) for these cases range between 1.15 and 3.53, indicating a higher likelihood compared to the general populace14,15,16. Particularly, the risk of developing Second Primary Lung Cancer (SPLC) have emerged as a significant concern16,17. Research has identified increased incidences of Secondary Primary Malignancies (SPMs) in association with radiotherapy across various cancer types18,19. A study indicated that postoperative radiotherapy could increase the risk of SPLC, adversely impacting survival rates in lung cancer survivors20. Given the proximity of the radiation field to the lungs, understanding radiotherapy's long-term safety profile is critical. However, studies on SPLC after radiotherapy for esophageal cancer are relatively rare.

The Surveillance, Epidemiology, and End Results (SEER) program provides a pivotal resource for studying these long-term outcomes21. Its comprehensive data facilitates an in-depth examination of SPLC incidence post-esophageal cancer treatment, highlighting the nuanced relationship between treatment modalities and secondary cancer risks22.

This study utilizes the SEER database to quantify SPLC incidence among esophageal cancer survivors and specifically examines the impact of radiotherapy. It also identifies factors that influence the development of SPLC and assesses patient survival outcomes. Through this analysis, we aim to refine treatment and surveillance strategies, contributing to a deeper understanding of cancer survivorship and the need for balancing effective treatment against long-term risks.

Method

Data source, study population, and group definitions

Our research utilized the Surveillance, epidemiology, and end results (SEER) program’s database, an extensive repository managed by the National Cancer Institute, representing approximately 34.6% of the U.S. population. We extracted data from the SEER 17 Registries Research Data, covering diagnoses from 2000 to 2020, to conduct an epidemiological analysis of esophageal cancer and subsequent SPLC incidence.

Individuals diagnosed with esophageal cancer were identified through ICD-O-3 topographical codes ‘C15.0’, ‘C15.3’, ‘C15.4’, ‘C15.5’, ‘C15.8’, and ‘C15.9’, with inclusion strictly limited to cases with histological confirmation. The study focused on patients classified as having a single primary esophageal cancer or esophageal cancer as the first of multiple primaries, ensuring clarity in the sequence of cancer diagnoses. Through a rigorous selection process, outlined in a flowchart (Fig. 1) for transparency, we refined our cohort to 56,493 individuals, stratified by radiotherapy status into non-radiotherapy (NRT) and radiotherapy (RT) groups.

Figure 1
figure 1

Flowchart illustrating the selection process for the study population. The initial cohort consisted of individuals diagnosed with esophageal cancer from 2000 to 2020 from the SEER database. Patients were then excluded based on specific criteria, resulting in the final study population. The study cohort was further divided based on the history of radiotherapy for esophageal cancer. Finally, patients were categorized into groups based on the development of Second Primary Lung Cancer (SPLC) during follow-up.

SPLC cases were defined as lung cancers (ICD-O-3 codes C34.0-C34.9) diagnosed at least two years following esophageal cancer treatment, distinguishing them from metastatic or synchronous tumors. The criteria facilitated the division of our cohort into specific groups for comparative analysis:

  • NRT and RT groups: Based on the absence or presence of radiotherapy treatment for esophageal cancer.

  • EC_SPLC (esophageal cancer with SPLC): Comprises patients from both the NRT and RT groups who developed SPLC, designed to explore the cumulative incidence of SPLC across different treatment backgrounds.

  • OPLC (only primary lung cancer): This group consists of individuals diagnosed with only primary lung cancer, without a history of esophageal cancer or any other primary malignancies. It is utilized as a control group to compare the survival outcomes and incidences of lung cancer in a population not previously affected by esophageal cancer.

  • EC_SPLC_RT (Esophageal Cancer with SPLC who received Radiotherapy): Patients who had undergone radiotherapy for esophageal cancer and subsequently developed SPLC fall into this category, allowing for an in-depth analysis of the impact of radiotherapy on SPLC risk.

  • EC_SPLC_NRT (Esophageal Cancer with SPLC who did not receive Radiotherapy): This group includes esophageal cancer survivors who developed SPLC but did not receive radiotherapy, serving as a counterpoint to EC_SPLC_RT in evaluating the role of radiotherapy in the development of secondary lung cancer.

The endpoint for follow-up was determined by the earliest occurrence of SPLC diagnosis, all-cause mortality, or completion of a 24-year follow-up, ensuring a comprehensive assessment of long-term outcomes.

Competing risks model and SIRs

To evaluate the differential risk of developing SPLC among esophageal cancer survivors, we employed a competing risks statistical framework. This model was crucial for accurately accounting for the fact that patients could experience one of several mutually exclusive outcomes, each of which precludes the occurrence of the others. The cumulative incidence function (CIF) was used to estimate the probability of developing SPLC over time, taking these competing risks into account. The Standardized Incidence Ratios (SIRs) were calculated separately for patients with and without radiotherapy (RT and NRT groups), enabling us to assess the specific impact of radiotherapy on SPLC risk. Statistical significance was determined through confidence intervals and p-values generated by the software.

Propensity score matching (PSM) and survival outcome analysis

We utilized Propensity Score Matching (PSM) to adjust for confounding factors across patient groups, ensuring a balanced comparison for our survival outcome analysis. This was achieved by matching patients on a 1:1 ratio using nearest-neighbor matching, based on propensity scores calculated from factors such as age, sex, cancer stage, and received treatments. Following PSM, we assessed survival outcomes to understand the impact of various factors, including treatment differences. Survival analyses were performed using Kaplan–Meier estimates and Cox proportional hazards models, providing a broad overview of how these factors influence overall and cancer-specific survival across the groups.

Statistical analysis

In this investigation, we utilized R software (version 4.3.3) for an in-depth statistical analysis aimed at elucidating the impact of treatment on survivors of esophageal cancer and quantifying the incidence of SPLC. The analysis of SIRs was conducted with SEER*Stat software (version 8.4.3), comparing the observed incidence of SPLC in our patient cohorts to expected rates grounded in general population statistics, with adjustments made for age, sex, and calendar period. To address the complexity of competing events in the assessment of SPLC cumulative incidence, the Fine-Gray competing risk regression model was meticulously applied. Propensity Score Matching (PSM) was implemented to ensure comparability between study groups, utilizing a 1:1 nearest-neighbor matching algorithm. This was achieved with the aid of specific R packages: ‘cmprsk’ for competing risks analysis and ‘MatchIt’ for PSM, enhancing the precision of our analyses. Categorical variables underwent evaluation via the Chi-square test, while continuous variables were examined using the Mann–Whitney test for distributions not adhering to normality and the t-test for those that were normally distributed.

Results

The baseline characteristics

Our study commenced with an initial dataset of 57,215 individuals diagnosed with esophageal cancer, confirmed via pathological examination, from the SEER database for the period 2000–2020. Subsequent exclusions due to lack of survival data, unidentified ethnicity, age below 20, or incomplete treatment details refined the cohort to 56,493 subjects. This group was then divided into those who had not received radiotherapy (NRT group, n = 25,490) and those who had undergone radiotherapy (RT group, n = 31,003), based on their treatment history. The median age for the entire cohort was determined to be 66 years, with an interquartile range of 58 to 75 years, indicating a predominance of older adults. The majority of the cohort was male (78.83%) and White (84.01%), reflecting the demographic trends in esophageal cancer incidence within the selected population. Comprehensive details on the demographic and clinical characteristics of the cohort, as well as the treatment specifics, have been methodically documented in (Table 1). Additionally, for those esophageal cancer patients who developed secondary primary lung cancer (EC_SPLC), specific details are further delineated in the supplementary Table S1.

Table 1 Characteristics of esophageal cancer survivors included in the study cohort.

Cumulative incidence of SPLC in EC patients

Utilizing a competing risks model, we assessed the differential cumulative incidences of second primary lung cancer (SPLC) in esophageal cancer survivors who were treated with and without radiotherapy (RT and NRT, respectively). The cumulative incidence curves indicated a distinct separation between the NRT and RT groups, which was statistically significant (p < 0.01) (Fig. 2). Specifically, the RT group had a cumulative incidence of 0.32% at 50 months, escalating to 1.04% by 250 months. In comparison, the NRT group presented with a 0.19% cumulative incidence at 50 months, which rose to 0.57% at the 250-month mark (Table 2). This comparative analysis underscores the increased long-term risk of developing SPLC associated with RT. Further analysis using a multivariable competing risks regression model elucidated the influence of several covariates on the incidence of SPLC. Radiotherapy (RT) significantly increased the risk of SPLC, with a hazard ratio (HR) of 1.41 (95% CI 1.06, 1.88; p = 0.018), indicating a 41% higher risk of SPLC in patients receiving RT compared to those who did not (Table 3).

Figure 2
figure 2

Cumulative incidence of second primary lung cancer (SPLC) among esophageal cancer survivors stratified by radiotherapy status.

Table 2 Cumulative incidence of SPLC among esophageal cancer survivors.
Table 3 Multivariable competing risks regression analysis showing hazard ratios (HR) for the development of SPLC in esophageal cancer survivors.

Subgroup analyses demonstrated that while radiotherapy was generally associated with an increased incidence of EC_SPLC, there were notable exceptions in certain subgroups. These findings indicate that while RT remains a significant factor in increasing the risk of SPLC across the majority of subgroups, individual subgroup characteristics might influence the extent of this risk (Fig. 3).

Figure 3
figure 3

Forest plot of subgroup analysis evaluating the influence of radiotherapy on the risk of developing second primary lung cancer (SPLC) in esophageal cancer survivors. The plot presents hazard ratios (HRs) for SPLC across various subgroups, including demographic characteristics and clinical factors such as age, sex, histology, and treatment received. Each point represents the HR for a subgroup, with horizontal lines depicting 95% confidence intervals. This plot highlights the differential risk and significance of radiotherapy across patient subgroups.

SIRs for EC_SPLC.

Further analysis revealed the SIRs for developing second primary lung cancer (SPLC) among esophageal cancer survivors, distinctly highlighting the elevated risk associated with radiotherapy. Our findings, based on SIR calculations comparing observed versus expected SPLC cases, underscored a markedly increased risk in the RT group (SIR = 2.68, 95% CI 2.35–3.04) relative to the NRT cohort (SIR = 1.53, 95% CI 1.27–1.83). This pattern persisted across various demographics, with the most pronounced disparities observed by sex—females in the RT group faced a disproportionately higher risk (SIR = 4.37, 95% CI 3.46–5.46) compared to males. Age-related analysis indicated younger EC survivors (< 60 years) undergoing RT were the most susceptible to SPLC, while latency findings highlighted an increased SPLC risk persisting beyond a decade post-treatment. In the segmented analysis by year of diagnosis, a pronounced increase in SIR was documented among EC patients treated with RT in the most recent cohort (2015–2019), with the ratio escalating to 3.05 (95% CI 2.01–4.44) (Table 4).

Table 4 Standardized incidence ratios (SIRs) for second primary lung cancer (SPLC) in esophageal cancer survivors, stratified by radiotherapy status.

PSM and survival outcome.

Table 5 delineates the equilibrium of baseline covariates before and after propensity score matching (PSM) for two specific patient groups: esophageal cancer (EC) patients who underwent radiotherapy and subsequently developed second primary lung cancer (hereinafter referred to as the EC_SPLC_RT group) and individuals diagnosed with only primary lung cancer (OPLC), without any previous history of other primary malignancies. This meticulous comparison highlights mean or proportion values, standardized mean differences (SMDs), and p-values for each group both prior to and following PSM. The marked reduction in SMDs post-PSM underscores the efficacy of the matching process in harmonizing covariate distributions between the two cohorts, thereby enabling a more accurate evaluation of comparative outcomes while minimizing the impact of confounding variables. Subsequent to PSM, survival outcomes between the EC_SPLC_RT group and the OPLC group were systematically analyzed. Kaplan–Meier estimates facilitated the calculation of survival probabilities and median survival times, with the survival curves providing a visual comparison of survival rates between the groups. The log-rank test, indicating p-values of 0.359 for overall survival (OS) and 0.011 for cancer-specific survival (CSS), revealed significant differences in CSS between the groups. Additionally, Cox proportional hazards regression analyses were conducted to estimate hazard ratios (HR) with 95% confidence intervals (CI), yielding a HR of 1.111 (95% CI 0.881–1.402) for OS, and a distinctly different HR of 0.695 (95% CI 0.525–0.920) for CSS (Fig. 4). This discrepancy underscores the varied impact of radiotherapy followed by the development of SPLC on patient survival compared to those with OPLC.

Table 5 Baseline characteristics of EC_SPLC_RT compared to OPLC patients before and after PSM.
Figure 4
figure 4

Kaplan–Meier survival curves for EC_SPLC_RT versus OPLC patients. (A) cancer-specific survival (CSS) comparison. (B) overall survival (OS) comparison.

Supplementary materials (Table S2, Table S3, Figure S1 and Figure S2) detail survival comparisons for EC patients with second primary lung cancer, with or without prior radiotherapy (EC_SPLC and EC_SPLC_NRT groups), against those with only primary lung cancer (OPLC). Notably, the EC_SPLC group showed significant differences in CSS from OPLC, with no impact on OS. No significant survival differences were found between the EC_SPLC_NRT group and OPLC. Further details are available in the supplementary tables and figures.

Discussion

Leveraging the extensive SEER database, our study embarked on a pioneering examination of SPLC incidence in esophageal cancer survivors, with a particular focus on the impact of radiotherapy. By offering a detailed analysis of the SPLC incidence and an exploration of the effects of radiotherapy, our research stands out for its depth in examining the temporal risk and the complex interplay of various factors affecting SPLC risk. Our findings not only corroborate but also expand upon previous research, emphasizing the importance of precise, long-term monitoring and management of esophageal cancer survivors to enhance their prognosis.

In our univariate competing risks model analysis, we observed a significant increase in the risk of developing SPLC among patients treated with radiotherapy for esophageal cancer, demonstrating a cumulative incidence of SPLC of 1.04% at 250 months post-radiotherapy, starkly higher than the 0.57% observed in patients who did not undergo radiotherapy. A previous study, focusing on resectable lung cancer, similarly highlights the significant role radiotherapy plays in elevating the risk of secondary malignancies. For instance, their research found that radiotherapy was substantially related to a higher risk of major second solid malignancies (RR = 1.21; 95% CI 1.08 to 1.35)18, echoing our observations of a notable increase in SPLC risk post-radiotherapy in esophageal cancer patients. Further, in the multivariate competing risks model, radiotherapy (RT) significantly elevated the risk of SPLC, with a hazard ratio (HR) of 1.41 (95% CI 1.06, 1.88; p = 0.018), aligning with the study’s emphasis on radiotherapy’s broad impact across different cancer types, including esophageal cancer18.

The subgroup analysis illuminates the intricate dynamics among various demographic, clinical, and treatment-related variables impacting the risk of SPLC among survivors of esophageal cancer. Middle-aged individuals (50–69 years) may have a unique response to radiotherapy compared to older or younger patients, potentially affecting their risk of developing SPLC. This age group’s longer post-treatment survival also extends the opportunity for SPLC to emerge. In gender subgroup analysis, notable differences were observed in radiotherapy exposure levels and dosages between male and female patients. These variances might result in differential exposure of surrounding tissues, particularly the lungs, to radiation23, thereby potentially affecting SPLC risk. Furthermore, radiotherapy's impact may be further modulated by gender-specific comorbidities and lifestyle factors, such as smoking history24 and occupational exposures, contributing to the complexity of assessing SPLC risk. The variation in SPLC incidence across different ethnic groups underscores the multifaceted interplay of genetic diversity, lifestyle factors, socioeconomic status, access to healthcare, biological responses to radiation, and environmental exposures. In histology subgroup analysis, squamous cell carcinomas are more commonly located in the upper parts of the esophagus, closer to the lungs25, potentially exposing lung tissue to higher radiation doses during treatment. This exposure could contribute to the higher observed rates of SPLC. According to NCCN Clinical Practice Guidelines, patients undergoing surgery might receive diverse combinations of treatments, including chemotherapy and targeted therapy26, complicating the assessment of radiotherapy's sole impact on SPLC risk. The combination of chemotherapy and radiotherapy may have a synergistic effect, thereby increasing the risk of Second Primary Lung Cancer (SPLC). This synergistic effect is likely due to the concurrent damage to DNA caused by both treatment methods27, leading to an increased risk of mutations and carcinogenesis, particularly in the surrounding healthy tissues, such as the lungs. Consequently, among patients in the chemotherapy group, the presence or absence of radiotherapy shows a notably significant difference in the incidence of SPLC. In groups not receiving chemotherapy, the dominance of patients not treated with radiotherapy implies a smaller sample of those who were, possibly reducing the statistical power to detect significant differences in SPLC incidence between the two.

Our study shows that the radiotherapy group demonstrating an SIR of 2.68 (95% CI 2.35–3.04), distinctly higher than the non-radiotherapy cohort, which had an SIR of 1.53 (95% CI 1.27–1.83). This increase aligns with past research that particularly highlighted a significant rise in SPLC risk in patients treated with radiotherapy, where a follow-up of 5 to 9 years showed an SIR of 3.46 (95% CI 2.41–4.82)14. These cumulative insights not only validate the accuracy of our findings but also emphasize the critical need for long-term monitoring of SPLC risk in esophageal cancer survivors, providing a comprehensive quantitative framework for assessing this risk in clinical practice.

We employed PSM to rigorously adjust for potential confounders, ensuring a robust comparison between esophageal cancer survivors and those diagnosed with only primary lung cancer (OPLC group). Post-PSM analysis showed that the cohorts of esophageal cancer survivors (including EC_SPLC, EC_SPLC_RT, and EC_SPLC_NRT) and OPLC groups were well-matched, allowing for an accurate and unbiased evaluation of survival outcomes. This meticulous matching process underscores the analytical rigor of our study, ensuring that observed differences in survival outcomes are attributable to the impact of radiotherapy and not confounded by baseline discrepancies.

The survival analysis revealed significant differences in cancer-specific survival (CSS) between groups, with the EC_SPLC_RT group exhibiting better CSS compared to the OPLC group, despite similar overall survival (OS) rates. This suggests that radiotherapy in esophageal cancer survivors may increase SPLC risk without affecting overall mortality, highlighting the efficacy of routine post-treatment CT scans in early tumor detection and treatment. Although no significant CSS and OS differences were found between the EC_SPLC_NRT group and the OPLC group, possibly due to smaller sample sizes, the EC_SPLC_NRT group’s CSS trended higher. Our analysis aligns with previous research indicating better prognosis for lung squamous cell carcinoma patients with second primary malignancies (SPMs)28, attributed to early detection and comprehensive treatment. Thus, our findings, together with those from previous studies, emphasize the importance of vigilant post-treatment surveillance for esophageal cancer survivors to enhance early detection and treatment of SPLC, potentially improving patient outcomes.

Our study reveals important insights but also faces limitations such as its retrospective nature, potential for unmeasured confounding factors, and limited detail on treatment specifics like radiation dose. These challenges highlight the difficulties in establishing causality and drawing precise dose-response relationships. Moreover, the findings' generalizability may be affected by the demographic scope of the SEER database and evolving cancer treatment modalities. Therefore, future research is crucial for a deeper understanding of SPLC risk, emphasizing the importance of considering comprehensive treatment details and lifestyle factors to accurately assess and mitigate this risk. Future studies should delve into prospective validations, mechanisms behind SPLC post-radiotherapy, and radiotherapy techniques to mitigate SPLC risks in esophageal cancer survivors. Exploring intervention strategies and the impact of treatments on quality of life is crucial, aiming for personalized care that balances therapeutic benefits against long-term risks. This multifaceted approach will enhance our understanding and improve outcomes for survivors.

Conclusion

This cohort study, utilizing the SEER database, not only highlights the increased risk of SPLC following radiotherapy in esophageal cancer survivors but also provides survival analysis outcomes. Specifically, our findings reveal that the prognosis of SPLC in esophageal cancer survivors with a history of radiotherapy does not fare worse than those with OPLC, underscoring the necessity for comprehensive long-term monitoring and management.