Abstract
Purpose
This study aimed to identify independent predictors of infield and outfield progression in stage IV non–small cell lung cancer (NSCLC) patients receiving thoracic radiation therapy (RT) combined with immune checkpoint inhibitors (ICI) and evaluate whether local tumor response mediates systemic disease control.
Materials and Methods
This retrospective study included 98 patients with stage IV NSCLC receiving thoracic RT combined with ICI between January 2018 and December 2023. Infield and outfield progression were analyzed using a competing risk analysis and survival outcomes were estimated using the Kaplan-Meier method.
Results
With a median follow-up time of 11.2 months (range, 1.4 to 84.4), the median overall survival was 24.9 months (95% confidence interval [CI], 18.4 to 31.3) and the median progression-free survival was 4.5 months (95% CI, 3.5 to 6.7). Infield objective response rate was 77.6%. On multivariable analysis, equivalent doses in 2 Gy fractions ≥40 Gy (subdistribution hazard ratio [SHR], 0.32; p = 0.002), concurrent chemotherapy with ICI (SHR, 0.30; p = 0.005), and squamous histology (SHR, 2.33; p = 0.025) were independent predictors of infield progression. Infield response (SHR, 0.28; p = 0.001), ICI-first sequencing (SHR, 0.27; p = 0.002), and Eastern Cooperative Oncology Group performance status 2 (SHR, 5.32; p < 0.001) independently predicted outfield progression.
Conclusion
Higher RT dose and infield tumor response independently predicted infield and outfield disease control, consistent with a potential dose-dependent abscopal association in stage IV NSCLC treated with thoracic RT and ICI.
Keywords: Non-small-cell lung carcinoma, Thoracic radiotherapy, Immune checkpoint inhibitors, Abscopal effect, Competing risk analysis
Introduction
Lung cancer remains the leading cause of cancer-related mortality worldwide, with non–small cell lung cancer (NSCLC) accounting for approximately 85% of all lung cancer diagnoses [1]. Nearly half of patients present with stage IV disease at diagnosis, a setting historically associated with a median overall survival (OS) of less than one year with platinum-based chemotherapy [2]. The introduction of immune checkpoint inhibitors (ICI) targeting the programmed cell death 1 (PD-1) and programmed cell death ligand 1 (PD-L1) axis has fundamentally transformed the treatment landscape of stage IV NSCLC. Pembrolizumab monotherapy demonstrated superior OS over chemotherapy in patients with PD-L1 expression ≥50%, and ICI-chemotherapy combinations have extended survival across broader, PD-L1–unselected populations [3,4]. Despite these advances, durable responses are achieved in only a subset of patients and the majority ultimately experience disease progression, underscoring the persistent need for strategies that augment or restore ICI efficacy [3,4]. Thoracic radiation therapy (RT) is frequently employed in stage IV NSCLC for disease control at symptomatic or high-burden intrathoracic sites, and its immunomodulatory properties have attracted growing interest as a potential complement to ICI in this setting.
Beyond its direct cytotoxic effects, RT promotes immunogenic tumor cell death, enhances tumor antigen release, and modulates the tumor immune microenvironment in ways that may synergize with PD-1/PD-L1 blockade [5]. Local RT-induced immune activation has been associated with tumor responses at distant, non-irradiated sites—known as the abscopal effect—providing a conceptual framework for RT as a systemic immune primer when combined with ICI [6]. This rationale found its strongest clinical validation in stage III NSCLC, where the PACIFIC trial demonstrated that durvalumab administered after definitive concurrent chemoradiation therapy significantly improved both progression-free survival (PFS) and OS, establishing the RT-ICI combination as a standard of care in that setting [7]. However, stage IV NSCLC represents a fundamentally different and underrepresented population. Thoracic RT is commonly delivered with palliative or salvage intent alongside ICI across heterogeneous sequencing strategies, yet this population was excluded from the landmark trials that defined the RT-ICI paradigm.
In this real-world stage IV setting, the factors that independently predict infield tumor control and outfield systemic disease control following thoracic RT combined with ICI remain poorly defined. Whether local tumor response mediates systemic disease control in this population has not been rigorously examined. This study therefore aimed to identify independent predictors of infield and outfield progression in patients with stage IV NSCLC receiving thoracic RT combined with ICI, and to evaluate the association between local tumor response and systemic disease control as evidence for a clinically detectable abscopal effect in this combination setting.
Materials and Methods
1. Study population
This study was a single-institution retrospective cohort study conducted at a tertiary referral center. A total of 128 patients with stage IV NSCLC who received thoracic RT in combination with ICI between January 2018 and December 2023 were identified. Patients were excluded if they had a history of prior thoracic RT (n = 17), lacked adequate follow-up imaging for response assessment (n = 9), or had a double primary cancer (n = 3). One patient was excluded for failing to complete the planned RT course. Final analysis included 98 patients.
2. Treatment
RT was delivered using intensity-modulated radiation therapy (IMRT) or three-dimensional conformal radiation therapy. All patients underwent four-dimensional (4D) computed tomography (CT) simulation along with abdomen compressor for motion control if needed. The internal target volume (ITV) encompassed all the gross tumor volume on the 4D CT images. Clinical target volume (CTV) was established by expanding the ITV by 5 mm, and the planning target volume (PTV) was established by expanding the CTV by 3 mm. Positron emission tomography (PET)–CT was used as a reference to aid in target delineation when available. The prescribed dose per fraction to the PTV ranged from 2.0 to 12.5 Gy, and the number of fractions ranged from 1 to 30, determined at the discretion of the treating physician. Doses were converted to equivalent doses in 2 Gy fractions (EQD2) using an α/β ratio of 10 [8]. ICI consisted of PD-1/PD-L1 inhibitors administered at standard doses. Most patients with PD-L1 expression greater than or equal to 50% received ICI monotherapy. The rest were treated with concurrent platinum-based doublet chemotherapy alongside ICI. Treatment sequencing was defined based on the chronological order of treatment initiation and treatment interval was defined as difference between the initiation dates. Patients were classified as "RT first" if ICI was administered during the course of RT or within two months of RT completion, and as "ICI first" if RT was administered during the course of ICI or within two months of ICI completion. Metastatic burden was categorized as oligometastatic (≤5 metastatic lesions) or polymetastatic (>5 metastatic lesions) per the European Society for Radiotherapy and Oncology–American Society for Radiation Oncology consensus definition [9]. RT intent was classified based on dose-fractionation criteria: schedules equivalent to or lower than 30 Gy in 10 fractions were classified as palliative and schedules exceeding this threshold were classified as salvage [10,11].
3. Follow-up
All patients received regular follow-up every 3 to 4 months with CT and with brain magnetic resonance imaging or PET-CT if needed. Treatment responses were evaluated using the Response Evaluation Criteria in Solid Tumors version 1.1. Infield objective response was defined as complete response or partial response (PR) of the irradiated target within six months following RT completion. For time-to-event analyses, infield progression was defined as progressive disease (PD) at the irradiated site on follow-up imaging, with time measured from the date of RT initiation. Outfield progression was defined in the same manner but outside of the irradiated field. Patients without an event were censored at the date of last follow-up. The starting date for all survival analyses was the date of RT initiation, applied uniformly across both sequence groups to ensure a consistent reference point and to limit differential immortal time.
4. Statistical analysis
Time-to-event analyses for infield and outfield progression were performed using a competing risk analysis, with death as a competing risk. Gray's non-parametric test was used to estimate and compare cumulative incidence functions between groups [12] and the Fine-Gray subdistribution hazard model was used for multivariable analysis [13].
PFS and OS were estimated using the Kaplan-Meier method. Univariable and multivariable Cox proportional hazards models were used to identify independent predictors [14]. The proportional hazards assumption was verified using Schoenfeld residual tests [15]. In this analysis, PFS was defined as time from RT initiation to first documented progression or death from any cause.
All multivariable models included variables with p < 0.10 on univariable analysis. All statistical tests were two-sided, with a significance threshold of p < 0.05. Statistical analyses were performed using R version 4.5.2 (R Foundation for Statistical Computing, Vienna, Austria).
Results
1. Patient characteristics
Among 98 patients, the median age was 66 years (range, 34 to 88). Nonsquamous pathology was present in 57 patients (58.2%), and 65 patients (66.3%) had PD-L1 expression below 50%. RT target site was either lung alone (n = 50, 51.0%) or lung with mediastinal lymph nodes (n = 48, 49.0%). Six patients (6.1%) received RT to bone metastases outside the thoracic field in addition to thoracic RT. The median total dose and number of fractions were 30.0 Gy (range, 5.0 to 69.0) and 5 fractions (range, 1 to 30), respectively. The median EQD2 was 40.0 Gy (range, 6.3 to 99.2). ICI monotherapy was administered in 56 patients (57.1%), with the remainder receiving concurrent platinum-based doublet chemotherapy. Most of the patients were treated with IMRT (n = 85, 86.7%). ICI was administered before RT in 64 patients (65.3%) and RT was initiated before ICI in 34 patients (34.7%). At the time of RT referral, 46 patients (46.9%) had de novo metastatic disease and 52 patients (53.1%) had disease that had progressed on prior systemic therapy. Fifty-nine patients (60.2%) had oligometastatic disease and 39 patients (39.8%) had polymetastatic disease. The median follow-up was 11.2 months (range, 1.4 to 84.4). Total baseline characteristics are summarized in Table 1.
Table 1.
Baseline patient and treatment characteristics (n = 98)
| Characteristic | Value |
|---|---|
| Age (year) | 66.0 (34.0–88.0) |
| Sex | |
| Female | 13 (13.3) |
| Male | 85 (86.7) |
| ECOG PS | |
| 0 | 3 (3.1) |
| 1 | 88 (89.8) |
| 2 | 7 (7.1) |
| Smoking history | |
| Never | 18 (18.4) |
| Ex | 38 (38.8) |
| Current | 42 (42.9) |
| Histology | |
| Nonsquamous | 57 (58.2) |
| Squamous | 40 (40.8) |
| Disease status at RT referral | |
| De novo | 46 (46.9) |
| Progression | 52 (53.1) |
| Metastatic burden | |
| Oligometastatic | 59 (60.2) |
| Polymetastatic | 39 (39.8) |
| PD-L1 expression (%) | |
| <50 | 65 (66.3) |
| ≥50 | 31 (31.6) |
| RT site | |
| Lung | 50 (51.0) |
| Lung and mediastinal LNs | 48 (49.0) |
| RT modality | |
| IMRT | 85 (86.7) |
| 3D-CRT | 13 (13.3) |
| RT intent | |
| Palliative | 21 (21.4) |
| Salvage | 77 (78.6) |
| EQD2 (Gy) | 40.0 (6.3–99.2) |
| <40 | 34 (34.7) |
| ≥40 | 64 (65.3) |
| Dose per fraction (Gy/fraction) | 5 (2–12.5) |
| ≤5 | 61 (62.2) |
| >5 | 37 (37.8) |
| No. of fractions | 5 (1–30) |
| 5 or 10 | 63 (64.3) |
| Other | 35 (35.7) |
| Systemic agent | |
| ICI only | 56 (57.1) |
| ICI + CTx | 42 (42.9) |
| Treatment sequence | |
| RT first | 34 (34.7) |
| ICI first | 64 (65.3) |
| Treatment interval (day) | |
| RT to ICI | 10.5 (0.0–87.0) |
| ICI to RT | 56.5 (1.0–756.0) |
| ICI cycles received | |
| ≤6 cycles | 41 (41.8) |
| >6 cycles | 57 (58.2) |
Values are presented as median (range) or number (%). Missing data for histology (n = 1) and PD-L1 expression (n = 2) were due to insufficient tissue available for pathological analysis.
ECOG PS, Eastern Cooperative Oncology Group performance status; PD-L1, programmed cell death ligand 1; RT, radiation therapy; LNs, lymph nodes; IMRT, intensity modulated radiation therapy; 3D-CRT, 3-dimensional conformal radiation therapy; EQD2, equivalent dose in 2 Gy fractions; ICI, immune checkpoint inhibitors; CTx, chemotherapy.
2. Treatment outcomes
For the entire cohort, the median OS was 24.9 months (95% confidence interval [CI], 18.4 to 31.3 months), with 1-year and 2-year OS rates of 71.6% (95% CI, 61.5% to 83.4%) and 55.1% (95% CI, 43.0% to 70.7%), respectively. The median PFS was 4.5 months (95% CI, 3.5 to 6.7 months), with 1-year and 2-year PFS rate of 33.3% (95% CI, 24.6% to 45.0%) and 15.8% (95% CI, 9.2% to 27.2%), respectively (Fig. 1A, 1B). Disease progression occurred in 46 patients (46.9%) with infield progression occurring in 21 patients (21.4%), outfield progression occurring in 19 patients (19.4%), and progression in both fields occurring in six patients (6.1%). The best infield tumor response within 6 months after RT was PR in 76 patients (77.6%), stable disease in 17 patients (17.3%), and PD in five patients (5.1%).
Fig. 1.
Overall survival (A) and progression-free survival (B) in all patients. CI, confidence interval.
3. Predictors of infield progression
Univariable analysis showed that EQD2 (<40 Gy vs. ≥40 Gy), pathology (nonsquamous vs. squamous), and systemic agent (ICI only vs. ICI and chemotherapy) were associated with infield progression (Table 2). In multivariable analysis, EQD2 ≥ 40 Gy (subdistribution hazard ratio [SHR], 0.32; 95% CI, 0.15 to 0.65; p = 0.002) (Fig. 2), concurrent chemotherapy with ICI (SHR, 0.30; 95% CI, 0.13 to 0.69; p = 0.005), and squamous histology (SHR, 2.33; 95% CI, 1.11 to 4.87; p = 0.025) were identified as significant prognostic factors of infield progression. In sensitivity analysis, disease status at RT referral, metastatic burden, and RT intent did not meet the pre-specified p < 0.10 entry criterion for either the infield or outfield multivariable models (p = 0.650, p = 0.570, and p = 0.290, respectively). Dose per fraction marginally met the entry criterion for the infield model (SHR, 0.48; p = 0.084) but was excluded from the multivariable model due to collinearity with EQD2 (Cramér's V = 0.369) and model parsimony constraints.
Table 2.
Univariable and multivariable analyses of prognostic factors for infield progression
| Characteristic | Univariable |
Multivariable |
||||
|---|---|---|---|---|---|---|
| SHR | 95% CI | p-value | SHR | 95% CI | p-value | |
| EQD2 (Gy) | ||||||
| <40 | - | - | - | - | ||
| ≥40 | 0.42 | 0.20–0.89 | 0.022 | 0.32 | 0.15–0.65 | 0.002 |
| Treatment sequence | ||||||
| RT first | - | - | - | - | ||
| ICI first | 1.00 | 0.44–2.27 | >0.999 | - | - | - |
| PD-L1 (%) | ||||||
| <50 | - | - | - | - | ||
| ≥50 | 0.78 | 0.35–1.74 | 0.540 | - | - | - |
| Pathology | ||||||
| Nonsquamous | - | - | - | - | ||
| Squamous | 2.02 | 0.96–4.26 | 0.064 | 2.33 | 1.11–4.87 | 0.025 |
| Systemic agent | ||||||
| ICI only | - | - | - | - | ||
| ICI + CTx | 0.39 | 0.17–0.91 | 0.029 | 0.30 | 0.13–0.69 | 0.005 |
| Age | 1.02 | 0.98–1.07 | 0.330 | - | - | - |
| ECOG PS | ||||||
| 0–1 | - | - | - | - | ||
| 2 | 1.95 | 0.60–6.34 | 0.270 | - | - | - |
| Smoking history | ||||||
| Never | - | - | - | - | ||
| Ever | 0.58 | 0.26–1.31 | 0.190 | - | - | - |
| RT site | ||||||
| Lung | - | - | - | - | ||
| Lung and mediastinal LNs | 1.10 | 0.52–2.33 | 0.800 | - | - | - |
SHR, subdistribution hazard ratio; CI, confidence interval; EQD2, equivalent dose in 2 Gy fractions; RT, radiation therapy; ICI, immune checkpoint inhibitors; PD-L1, programmed cell death ligand 1; CTx, chemotherapy; ECOG PS, Eastern Cooperative Oncology Group performance status; LNs, lymph nodes.
Fig. 2.
Cumulative incidence of infield progression according to radiation therapy dose (equivalent doses in 2 Gy fractions). SHR, subdistribution hazard ratio; CI, confidence interval.
4. Predictors of outfield progression
Univariable analysis showed that infield response (no response vs. response), treatment sequence (RT first vs. ICI first), and Eastern Cooperative Oncology Group (ECOG) performance status (PS) (0–1 vs. 2) were associated with outfield progression (Table 3). In multivariable analysis, infield response (SHR, 0.28; 95% CI, 0.13 to 0.60; p = 0.001) (Fig. 3), ICI-first sequence (SHR, 0.27; 95% CI, 0.12 to 0.62; p = 0.002), and ECOG PS 2 (SHR, 5.32; 95% CI, 2.08 to 13.6; p < 0.001) were identified as significant prognostic factors of outfield progression.
Table 3.
Univariable and multivariable analyses of prognostic factors for outfield progression
| Characteristic | Univariable |
Multivariable |
||||
|---|---|---|---|---|---|---|
| SHR | 95% CI | p-value | SHR | 95% CI | p-value | |
| EQD2 (Gy) | ||||||
| <40 | - | - | - | - | ||
| ≥40 | 0.60 | 0.27–1.30 | 0.190 | - | - | - |
| Infield response | ||||||
| No response | - | - | - | - | ||
| Response | 0.34 | 0.15–0.75 | 0.008 | 0.28 | 0.13–0.60 | 0.001 |
| Treatment sequence | ||||||
| RT first | - | - | - | - | ||
| ICI first | 0.31 | 0.14–0.70 | 0.005 | 0.27 | 0.12–0.62 | 0.002 |
| PD-L1 (%) | ||||||
| <50 | - | - | - | - | ||
| ≥50 | 0.64 | 0.26–1.58 | 0.330 | - | - | - |
| Pathology | ||||||
| Nonsquamous | - | - | - | - | ||
| Squamous | 0.62 | 0.27–1.42 | 0.260 | - | - | - |
| Systemic agent | ||||||
| ICI only | - | - | - | - | ||
| ICI + CTx | 0.71 | 0.32–1.60 | 0.410 | - | - | - |
| Age | 1.01 | 0.97–1.06 | 0.640 | - | - | - |
| ECOG PS | ||||||
| 0–1 | - | - | - | - | ||
| 2 | 3.49 | 1.14–10.70 | 0.028 | 5.32 | 2.08–13.60 | <0.001 |
| Smoking history | ||||||
| Never | - | - | - | - | ||
| Ever | 1.74 | 0.52–5.82 | 0.370 | - | - | - |
| RT site | ||||||
| Lung | - | - | - | - | ||
| Lung and mediastinal LNs | 0.96 | 0.44–2.07 | 0.910 | - | - | - |
SHR, subdistribution hazard ratio; CI, confidence interval; EQD2, equivalent dose in 2 Gy fractions; RT, radiation therapy; ICI, immune checkpoint inhibitors; PD-L1, programmed cell death ligand 1; CTx, chemotherapy; ECOG PS, Eastern Cooperative Oncology Group performance status; LNs, lymph nodes.
Fig. 3.
Cumulative incidence of outfield progression according to infield response. SHR, subdistribution hazard ratio; CI, confidence interval.
5. Predictors of OS and PFS
In univariable analysis, infield response (no response vs. response), pathology (nonsquamous vs. squamous), and age were associated with OS. In the subsequent multivariable analysis, no significant prognostic factors of OS were identified (Supplementary Table S1).
Regarding PFS, PD-L1 ≥ 50% was the only significant factor associated with PFS in univariable and multivariable analysis (hazard ratio [HR], 0.53; 95% CI, 0.32 to 0.89; p = 0.017) (Supplementary Table S2).
Discussion and Conclusion
This study evaluated predictors of infield and outfield disease control in patients with stage IV NSCLC receiving thoracic RT combined with ICI. Two principal findings emerged. First, higher RT dose (EQD2 ≥40 Gy) independently reduced the cumulative incidence of infield progression, supporting a dose-dependent relationship between RT and local disease control in this combination setting. Second, infield response was an independent predictor of outfield disease control, consistent with a potential dose-dependent abscopal association [6].
The independent association between higher RT dose and reduced risk of infield progression (SHR, 0.32) is consistent with the established dose-response relationship for local tumor control with RT. Tonneau et al. [16] reported in a competing risk analysis of early-stage NSCLC treated with stereotactic body radiation therapy (SBRT) that a higher biologically effective dose was associated with reduced local recurrence. Luke et al. [17] demonstrated in a multisite SBRT and pembrolizumab trial that local tumor control was dose-dependent, with higher prescribed doses associated with better local control rates. The study also showed that local tumor response independently correlated with OS, suggesting that the response of irradiated lesions has consequences beyond the irradiated field. The PEMBRO-RT trial similarly reported improved objective response rates and a trend towards improved OS with SBRT followed by pembrolizumab compared with pembrolizumab alone [18]. Preclinical data further suggest that the immunostimulatory effects of RT are also dose-dependent, with high doses promoting greater immunogenic cell death and downstream systemic immune activation [19]. These findings provide biological plausibility for the dose threshold and abscopal effect observed in our cohort.
These findings contrast with several studies that failed to demonstrate a significant abscopal benefit from RT combined with ICI. The CHEERS trial, which evaluated subablative multisite SBRT combined with ICI in a diverse polymetastatic population, found no clinically meaningful abscopal effect [20]. The MDACC trial similarly found no significant improvement in overall out-of-field response or PFS when analyses were not stratified by RT schedule, although a pooled analysis of the MDACC trial and PEMBRO-RT trial revealed positive results [21]. Several factors may explain these discrepant findings and reasons why a local-to-systemic association was detectable in our cohort when other studies have not. Understanding these factors is important because the current evidence base does not yet permit definitive conclusions about the conditions required for a clinically meaningful abscopal response.
First, RT dose may contribute to the difference between studies. Our study used a median EQD2 of 40 Gy, and the delivery of a dose equal to or greater than this threshold was associated with reduced risk of infield progression (SHR, 0.32). However, the CHEERS trial used an EQD2 of 36 Gy across multiple peripheral sites. Schoenfeld et al. [22] also used an EQD2 of 36 Gy (hypofractionated arm) or 7 Gy (low-dose arm) in a study of ICI-refractory NSCLC patients and failed to demonstrate benefits of additional RT. Even though the PEMBRO-RT trial also used an EQD2 of 36 Gy, it focused on a single site—mostly thoracic—and showed a positive trend. Although there is emerging evidence that fraction size itself can modulate the likelihood of triggering an abscopal effect [23], dose per fraction did not show a significant association in our cohort's sensitivity analysis.
The site of irradiation may also be an important determinant of the magnitude of abscopal response. In our study, RT was delivered mainly to thoracic targets (only 6 patients received RT to bone metastases), which may have preferentially engaged immunologically active tumor-draining pathways. Ziegler et al. [24] reported in a retrospective analysis of metastatic lung cancer patients receiving concurrent RT and ICI that RT to the primary tumor was the strongest independent predictor of OS. In another analysis of advanced NSCLC patients receiving RT combined with ICI, Wu et al. [25] found that lung irradiation was associated with a greater objective response rate than irradiation to bone metastases. In contrast, the aforementioned CHEERS trial irradiated multiple peripheral metastatic sites including bone, soft tissue, and visceral organs, which may have been less effective at triggering the focused immunogenic stimulus that thoracic RT can provide.
Taken together, the positive findings in our cohort are most plausibly explained by the convergence of adequate dose and focal thoracic irradiation—conditions that distinguish our real-world cohort from several trials that failed to demonstrate systemic benefit. However, the heterogeneity of findings across the literature means that the relative contribution of each factor remains uncertain, and prospective studies with pre-specified RT parameters are needed to distinguish their individual effects.
Among other predictors of infield progression, squamous histology was associated with higher infield progression rates (SHR, 2.33), consistent with prior studies demonstrating that squamous cell carcinoma carries a higher risk of local failure following RT-based treatment in NSCLC [26,27]. Concurrent platinum-based chemotherapy alongside ICI independently reduced infield progression (SHR, 0.30), consistent with the known radiosensitizing properties of platinum compounds and their additive cytotoxic effects on the irradiated tumor [28]. PD-L1 ≥50% was the sole independent predictor of PFS (HR, 0.53), consistent with its established role as a predictive biomarker for ICI response [3]. ECOG PS 2 strongly predicted outfield progression (SHR, 5.32), but as this subgroup was comprised of only seven patients, the large effect estimate should be interpreted with caution given the inherent statistical imprecision. Notably, ECOG PS 2 patients in this cohort received a median of only two ICI cycles (range, 1 to 4), suggesting that limited systemic immune exposure from curtailed ICI treatment may have substantially contributed to the high rate of outfield progression observed in this subgroup. ICI-first sequencing appeared to reduce outfield progression in the multivariable model (SHR, 0.27), but this association was not observed for PFS and should be interpreted with caution due to the effect of immortal time bias. Because all analyses were anchored to the date of RT initiation, patients in the ICI-first group had by definition survived and remained progression-free from their initial ICI start until RT—a median of 56.5 days—introducing an inherent selection advantage that cannot be eliminated by uniform baseline assignment alone.
This study has several limitations. The small cohort size limits statistical power, and the precision of estimates for low-frequency subgroups is inherently constrained as in the case of ECOG PS 2 subgroup. The single-institution retrospective design introduces potential selection bias and limits generalizability, as treatment decisions were made at physician discretion rather than per a pre-specified protocol. The relatively short median follow-up limits the maturity of OS and long-term disease control data. Additionally, the wide range of dose per fraction (2.0–12.5 Gy) reflects heterogeneous clinical intent and may confound the EQD2-based dose-response relationship observed for infield progression. Finally, the association between infield response and outfield disease control—while consistent with an abscopal effect—may alternatively reflect inherent ICI sensitivity. Patients with favorable tumor biology or robust systemic immune responses may achieve concurrent local and systemic disease control independently of any RT-mediated immune mechanism. Despite these limitations, this study provides clinically meaningful evidence from a real-world stage IV NSCLC population that is systematically excluded from landmark RT-ICI trials. The findings generate specific, testable hypotheses regarding RT dose thresholds and site of irradiation that can inform the design of prospective studies.
In conclusion, this study provides clinical evidence that higher RT dose is independently associated with reduced infield progression, and that infield response drives outfield disease control in stage IV NSCLC treated with thoracic RT and ICI, consistent with a potential dose-dependent abscopal association. Prospective evaluation of RT dose and site optimization strategies is warranted, and future studies should investigate whether the local-to-systemic immune association described here extends to other metastatic sites beyond the thorax in the metastatic NSCLC setting.
Footnotes
Statement of Ethics
The study protocol was reviewed and approved by the Institutional Review Board of Yonsei University Health System, Severance Hospital (approval no. 4-2026-0317). Written informed consent was not required, as decided by the Yonsei University Health System, Severance Hospital, Institutional Review Board committee.
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Funding
None.
Author Contributions
Conceptualization, JL, KHK; Data collection and anonymization, JL, SHM, YJH, HC; Statistical analysis, JL; Supervision, KHK; Writing of the original draft, JL; Writing of the review and editing, JL, KHK; Approval of final manuscript, all authors.
Data Availability Statement
The data that support the findings of this study are not publicly available because the information could compromise the privacy of research participants but are available from the corresponding author upon reasonable request.
Supplementary Materials
Supplementary materials can be found via https://doi.org/10.3857/roj.2026.00549.
Predictors of overall survival
Predictors of progression-free survival
References
- 1.Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71:209–49. doi: 10.3322/caac.21660. [DOI] [PubMed] [Google Scholar]
- 2.Schiller JH, Harrington D, Belani CP, et al. Comparison of four chemotherapy regimens for advanced non-small-cell lung cancer. N Engl J Med. 2002;346:92–8. doi: 10.1056/nejmoa011954. [DOI] [PubMed] [Google Scholar]
- 3.Reck M, Rodriguez-Abreu D, Robinson AG, et al. Pembrolizumab versus chemotherapy for PD-L1-positive non-small-cell lung cancer. N Engl J Med. 2016;375:1823–33. doi: 10.1056/nejmoa1606774. [DOI] [PubMed] [Google Scholar]
- 4.Gandhi L, Rodriguez-Abreu D, Gadgeel S, et al. Pembrolizumab plus chemotherapy in metastatic non-small-cell lung cancer. N Engl J Med. 2018;378:2078–92. doi: 10.1056/nejmoa1801005. [DOI] [PubMed] [Google Scholar]
- 5.Dovedi SJ, Cheadle EJ, Popple AL, et al. Fractionated radiation therapy stimulates antitumor immunity mediated by both resident and infiltrating polyclonal T-cell populations when combined with PD-1 blockade. Clin Cancer Res. 2017;23:5514–26. doi: 10.1158/1078-0432.CCR-16-1673. [DOI] [PubMed] [Google Scholar]
- 6.Ngwa W, Irabor OC, Schoenfeld JD, Hesser J, Demaria S, Formenti SC. Using immunotherapy to boost the abscopal effect. Nat Rev Cancer. 2018;18:313–22. doi: 10.1038/nrc.2018.6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Antonia SJ, Villegas A, Daniel D, et al. Durvalumab after chemoradiotherapy in stage III non-small-cell lung cancer. N Engl J Med. 2017;377:1919–29. doi: 10.1056/NEJMoa1709937. [DOI] [PubMed] [Google Scholar]
- 8.Fowler JF. The linear-quadratic formula and progress in fractionated radiotherapy. Br J Radiol. 1989;62:679–94. doi: 10.1259/0007-1285-62-740-679. [DOI] [PubMed] [Google Scholar]
- 9.Lievens Y, Guckenberger M, Gomez D, et al. Defining oligometastatic disease from a radiation oncology perspective: an ESTRO-ASTRO consensus document. Radiother Oncol. 2020;148:157–66. doi: 10.1016/j.radonc.2020.04.003. [DOI] [PubMed] [Google Scholar]
- 10.Rodrigues G, Videtic GM, Sur R, et al. Palliative thoracic radiotherapy in lung cancer: an American Society for Radiation Oncology evidence-based clinical practice guideline. Pract Radiat Oncol. 2011;1:60–71. doi: 10.1016/j.prro.2011.01.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Fairchild A, Harris K, Barnes E, et al. Palliative thoracic radiotherapy for lung cancer: a systematic review. J Clin Oncol. 2008;26:4001–11. doi: 10.1200/jco.2007.15.3312. [DOI] [PubMed] [Google Scholar]
- 12.Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat. 1988;16:1141–54. doi: 10.1214/aos/1176350951. [DOI] [Google Scholar]
- 13.Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496–509. doi: 10.1080/01621459.1999.10474144. [DOI] [Google Scholar]
- 14.Cox DR. Regression models and life-tables. J R Stat Soc Series B Methodol. 1972;34:187–220. doi: 10.1111/j.2517-6161.1972.tb00899.x. [DOI] [Google Scholar]
- 15.Schoenfeld D. Partial residuals for the proportional hazards regression model. Biometrika. 1982;69:239–41. doi: 10.1093/biomet/69.1.239. [DOI] [Google Scholar]
- 16.Tonneau M, Richard C, Routy B, et al. A competing risk analysis of the patterns and risk factors of recurrence in early-stage non-small cell lung cancer treated with stereotactic ablative radiotherapy. Radiother Oncol. 2023;185:109697. doi: 10.1016/j.radonc.2023.109697. [DOI] [PubMed] [Google Scholar]
- 17.Luke JJ, Onderdonk BE, Bhave SR, et al. Improved survival associated with local tumor response following multisite radiotherapy and pembrolizumab: secondary analysis of a phase I trial. Clin Cancer Res. 2020;26:6437–44. doi: 10.1158/1078-0432.ccr-20-1790. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Theelen WS, Peulen HM, Lalezari F, et al. Effect of pembrolizumab after stereotactic body radiotherapy vs pembrolizumab alone on tumor response in patients with advanced non-small cell lung cancer: results of the PEMBRO-RT phase 2 randomized clinical trial. JAMA Oncol. 2019;5:1276–82. doi: 10.1001/jamaoncol.2019.1478. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Deng L, Liang H, Burnette B, et al. Irradiation and anti-PD-L1 treatment synergistically promote antitumor immunity in mice. J Clin Invest. 2014;124:687–95. doi: 10.1172/jci67313. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Spaas M, Sundahl N, Kruse V, et al. Checkpoint inhibitors in combination with stereotactic body radiotherapy in patients with advanced solid tumors: the CHEERS phase 2 randomized clinical trial. JAMA Oncol. 2023;9:1205–13. doi: 10.1001/jamaoncol.2023.2132. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Theelen WS, Chen D, Verma V, et al. Pembrolizumab with or without radiotherapy for metastatic non-small-cell lung cancer: a pooled analysis of two randomised trials. Lancet Respir Med. 2021;9:467–75. doi: 10.1016/s2213-2600(20)30391-x. [DOI] [PubMed] [Google Scholar]
- 22.Schoenfeld JD, Giobbie-Hurder A, Ranasinghe S, et al. Durvalumab plus tremelimumab alone or in combination with low-dose or hypofractionated radiotherapy in metastatic non-small-cell lung cancer refractory to previous PD(L)-1 therapy: an open-label, multicentre, randomised, phase 2 trial. Lancet Oncol. 2022;23:279–91. doi: 10.1016/s1470-2045(21)00658-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Vanpouille-Box C, Alard A, Aryankalayil MJ, et al. DNA exonuclease Trex1 regulates radiotherapy-induced tumour immunogenicity. Nat Commun. 2017;8:15618. doi: 10.1038/ncomms15618. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ziegler DA, Ambold M, Schirmer MA, et al. The abscopal effect in metastatic lung cancer: a retrospective analysis of combined radiotherapy and immunotherapy. Clin Exp Metastasis. 2025;42:56. doi: 10.1007/s10585-025-10375-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Wu M, Liu J, Wu S, et al. Systemic immune activation and responses of irradiation to different metastatic sites combined with immunotherapy in advanced non-small cell lung cancer. Front Immunol. 2021;12:803247. doi: 10.3389/fimmu.2021.803247. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zhou R, Xu T, Nguyen QN, et al. Radiation dose, local disease progression, and overall survival in patients with inoperable non-small cell lung cancer after concurrent chemoradiation therapy. Int J Radiat Oncol Biol Phys. 2018;100:452–61. doi: 10.1016/j.ijrobp.2017.10.003. [DOI] [PubMed] [Google Scholar]
- 27.Spigel DR, Faivre-Finn C, Gray JE, et al. Five-year survival outcomes from the PACIFIC trial: durvalumab after chemoradiotherapy in stage III non-small-cell lung cancer. J Clin Oncol. 2022;40:1301–11. doi: 10.1200/jco.21.01308. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Dewit L. Combined treatment of radiation and cisdiamminedichloroplatinum (II): a review of experimental and clinical data. Int J Radiat Oncol Biol Phys. 1987;13:403–26. doi: 10.1016/0360-3016(87)90015-0. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Predictors of overall survival
Predictors of progression-free survival