Rescue by radiotherapy and anti-CTLA4/PD-1 after failure of anti-PD-1 therapy in patients with metastatic NSCLC: the phase II RECLAIM trial Authors: Ezgi B Ulas, Nora D Purcell, Ilias Houda, Sayed M S Hashemi, Joris Veltman, Johannes M A Daniels, Marieke F Fransen, Teodora Radonic, Linda R Beeloo, Peter S N van Rossum, Febe van Maldegem, Nicole Barlo, Maria Disselhorst, Marjolein van Laren, Marian A Tiemessen, Svitlana Tarasevych, Jan M van Haarst, Peter van Tilburg, Peter W A Kunst, Arifa Moons-Pasic, Daniela E Oprea-Lager, Lilian J Meijboom, Natalja Bouwhuis, Tanja D de Gruijl, Suresh Senan, Famke L Schneiders, Idris Bahce Published: 2026 ABSTRACT AbstractBackgroundResponse to anti-programmed cell death protein-1 (anti-PD-1) immunotherapy remains limited in patients with metastatic non-small cell lung cancer (NSCLC), whose tumors exhibit low or absent programmed death-ligand 1 (PD-L1) expression, and subsequent second-line therapy has poor efficacy. To address this limitation, we evaluated the efficacy and safety of combined ipilimumab and nivolumab (IPI/NIVO) with subablative radiotherapy (RT) in patients with metastatic NSCLC with negative or low PD-L1 expression, who had progressed on prior anti-PD-1 therapy.MethodsThis single-arm, prospective phase II trial aimed to enroll 30 evaluable patients with metastatic NSCLC exhibiting low (1–49%) or negative (<1%) PD-L1 tumor expression who had progressed after first-line anti-PD-1 therapy. Primary endpoints were safety, disease control rate (DCR), and objective response rate (ORR) at 6 and 12 weeks, assessed in non-irradiated tumor lesions. Treatment consisted of IPI 1 mg/kg every 6 weeks (Q6W) and NIVO 240 mg every 2 weeks for 6 weeks combined with subablative RT (3×8 Gy to 1–4 lesions). Thereafter, IPI 1 mg/kg Q6W and NIVO 360 mg every 3 weeks were continued.ResultsIn 31 patients of the intention-to-treat population, ORR was 7% and 10% at 6 and 12 weeks, and reached 29% as the best response. DCR was 58% and 39% at 6 and 12 weeks. Overall survival (OS) differed significantly by best response, with a median OS of 3.1, 13.5 and 22.5 months for progressive disease, stable disease and partial/complete response (p<0.001). Baseline sum of longest diameters, together with age, blood inflammatory markers and albumin levels, were prognostic of treatment response. All patients experienced treatment-related adverse events (AEs), with grade 3 as the highest severity in eight patients (26%). Immune-related AEs led to treatment discontinuation in five patients (16%). Early T-cell activation in peripheral blood samples (day 8) was detectable and more pronounced in responders than in progressors.ConclusionsIn patients with metastatic NSCLC and low or negative tumor PD-L1 expression, IPI/NIVO/RT was able to induce objective clinical responses in a subset of patients who had progressed after first-line anti-PD1 therapy. Treatment was associated with a strong T-cell activation, improved OS and an acceptable safety profile.Trial registration number2020–001097-29. WHAT IS ALREADY KNOWN ON THIS TOPIC Patients with metastatic NSCLC whose tumors express low or no programmed death-ligand 1 (PD-L1) have limited benefit from anti-programmed cell death protein-1 (anti-PD-1) immunotherapy, and effective strategies to restore responsiveness after progression on first-line anti-PD-1 therapy are lacking. Radiotherapy has been proposed as a means to potentiate antitumor immunity when combined with immune checkpoint inhibition, suggesting a potential therapeutic avenue for this population. Patients with metastatic NSCLC whose tumors express low or no programmed death-ligand 1 (PD-L1) have limited benefit from anti-programmed cell death protein-1 (anti-PD-1) immunotherapy, and effective strategies to restore responsiveness after progression on first-line anti-PD-1 therapy are lacking. Radiotherapy has been proposed as a means to potentiate antitumor immunity when combined with immune checkpoint inhibition, suggesting a potential therapeutic avenue for this population. WHAT THIS STUDY ADDS The phase II RECLAIM trial evaluated the combination of subablative radiotherapy (3×8 Gy) with dual checkpoint inhibition using ipilimumab and nivolumab in patients with metastatic NSCLC with low or negative PD-L1 expression who had progressed on prior anti-PD1 therapy, and showed objective response rates of 7% and 10% at 6 and 12 weeks, respectively, with a best overall response of 29% in non-irradiated lesions, and acceptable safety. Immune profiling of peripheral blood samples revealed distinct T-cell activation patterns following treatment, which differed in strength between clinical responders and non-responders. The phase II RECLAIM trial evaluated the combination of subablative radiotherapy (3×8 Gy) with dual checkpoint inhibition using ipilimumab and nivolumab in patients with metastatic NSCLC with low or negative PD-L1 expression who had progressed on prior anti-PD1 therapy, and showed objective response rates of 7% and 10% at 6 and 12 weeks, respectively, with a best overall response of 29% in non-irradiated lesions, and acceptable safety. Immune profiling of peripheral blood samples revealed distinct T-cell activation patterns following treatment, which differed in strength between clinical responders and non-responders. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY These findings support the feasibility and antitumor activity of integrating subablative radiotherapy with dual checkpoint blockade in a second-line setting for patients with otherwise limited therapeutic options. The results highlight the need for further research to refine patient selection and optimize multimodal treatment strategies aimed at overcoming resistance to immunotherapy. These findings support the feasibility and antitumor activity of integrating subablative radiotherapy with dual checkpoint blockade in a second-line setting for patients with otherwise limited therapeutic options. The results highlight the need for further research to refine patient selection and optimize multimodal treatment strategies aimed at overcoming resistance to immunotherapy. Introduction Non-small cell lung cancer (NSCLC) is among the most prevalent cancers worldwide, and is frequently diagnosed at an advanced, metastatic stage.1 2 For patients with metastatic NSCLC and tumors expressing low (1–49%) or negative (<1%) levels of programmed death-ligand 1 (PD-L1), current guidelines recommend first-line therapy consisting of chemotherapy combined with an immune checkpoint inhibitor (ICI) targeting the PD-L1 and programmed cell death protein-1 (PD-1) axis.3 However, a substantial proportion of patients either fail to respond or experience only brief responses, primarily due to either primary or acquired resistance against immunotherapy.4 5 Currently, there are no established follow-up treatment strategies to (re-)sensitize tumors to immunotherapy or effectively overcome this immunotherapy resistance. Instead, second-line chemotherapy (eg, docetaxel or pemetrexed) is generally recommended, but is limited by poor overall response rates of only 7–11%.6 7 The combination of radiotherapy (RT) with anti-PD-1 therapy has been proposed as a promising strategy to augment antitumor immunity through two primary mechanisms. First, RT induces tumor cell death and subsequent tumor antigen release, potentially facilitating enhanced immune recognition and priming.810 In addition, RT can activate the cyclic GMP-AMP synthase (cGAS)-stimulator of interferon genes (STING) pathway, promoting the production of proinflammatory cytokines and shifting the tumor microenvironment from an immunosuppressive to a more immune-infiltrated and immunologically active state. Furthermore, combining RT with anti-cytotoxic T-lymphocyte-associated protein 4 (anti-CTLA4) therapy may further augment these effects due to the distinct and complementary mechanisms of CTLA4 inhibition, which include enhanced activation of antigen-presenting cells, reduction of suppressive immune populations such as regulatory T cells (Tregs) and myeloid-derived suppressor cells, and increased priming and activation of effector T cells.11 Thus, concurrent use of RT and anti-CTLA4 therapy could mutually enforce T-cell priming and boosting. While there is evidence supporting the benefits of RT combined separately with either anti-PD-1 or anti-CTLA4 agents, limited data currently exist on the simultaneous combination of anti-PD-1, anti-CTLA4, and RT, highlighting an important area for further investigation.1215 We hypothesized that combining anti-PD-1 with anti-CTLA4 immunotherapy and subablative RT could synergistically overcome resistance to anti-PD-1 therapy and elicit robust antitumor activity in metastatic NSCLC with low or negative PD-L1 expression. Accordingly, this prospective phase II trial aimed to evaluate the efficacy and safety of ipilimumab and nivolumab (IPI/NIVO) administered with subablative RT in patients with negative or low PD-L1-expressing metastatic NSCLC who progressed on first-line anti-PD-1 therapy. Materials and methods Study population This single-center, single-arm prospective phase-II study (RECLAIM trial, EudraCT number 2020–001097-29) enrolled patients with metastatic (stage IV) NSCLC who had progressed on first-line chemotherapy plus anti-PD-1 therapy, between September 2020 and February 2022. The study was approved by the medical ethics committee (METc VUmc number 2020.336) and all enrolled patients signed informed consent for study participation. Eligible patients were required to have at least one lesion suitable for RT delivered at three fractions of 8 Gy and at least one measurable, non-irradiated lesion according to Response Evaluation Criteria in Solid Tumors (RECIST) V.1.1.16 Baseline tumor biopsies showing negative (<1%) or low (1–49%) PD-L1 expression were mandatory. Additional inclusion criteria included a European Cooperative Oncology Group performance score of 0–2, while patients with actionable genomic alterations such as EGFR and ALK were excluded. All cases were reviewed by a multidisciplinary team (MDT) prior to enrollment. Comprehensive inclusion and exclusion criteria are provided in the online supplemental material. Study design Figure 1 illustrates the RECLAIM study flowchart. Study therapy was delivered in two phases. Phase 1 (day 1–42), consisted of a single dose of IPI (1 mg/kg intravenous) on day 1, and NIVO (240 mg intravenous) on days 1, 15, and 29. Subablative radiation was administered as 3×8 Gy on days 8, 10, and 12 to 1–4 tumor sites, while some patients also received 1×8 Gy to additional lesions when required for pain control. Selection of tumor sites for RT was determined by the MDT decision based on limiting anticipated toxicity as a key determinant. Priority was given to treat sites of symptomatic disease. Phase 2 (day 43 and onward) included IPI (1 mg/kg intravenous every 6 weeks) and NIVO (360 mg intravenous every 3 weeks). Radiologic responses were assessed using CT by RECIST V.1.1 (as progressive disease (PD), stable disease (SD), partial response (PR) or complete response (CR)), based on unirradiated lesions. CT scans were planned at 6 and 12 weeks in phase 1 and every 12 weeks in phase 2. Treatment continued until disease progression, unacceptable toxicity, or completion of 2 years of IPI/NIVO therapy. The complete study protocol is provided in online supplemental file 2. Overview of the RECLAIM study design. The figure depicts treatment initiation and the scheduled time points of PBMC collection and biopsy retrieval. In phase 2, IPI (Q6W) and NIVO (Q3W) were continued until disease progression, unacceptable toxicity or completion of 2 years of therapy. anti-PD-L1, anti-programmed death-ligand 1; DCR, disease control rate; ECOG PS, European Cooperative Oncology Group performance status; IPI, ipilimumab; NIVO, nivolumab; NSCLC, non-small cell lung cancer; ORR, objective response rate; PBMC, peripheral blood mononuclear cells; Q3W, every 3 weeks; Q6W, every 6 weeks; RECIST, Response Evaluation Criteria in Solid Tumors; RT, radiotherapy. Clinical outcomes Primary endpoints included (1) objective response rate (ORR) and disease control rate (DCR) at 6 and 12 weeks after the start of IPI/NIVO/RT, and (2) safety defined by adverse events (AEs) graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events V.517 with classification of immune-related and serious AEs, assessed from the time of signed informed consent through 100 days following the last IPI/NIVO. Secondary endpoints included (1) progression-free survival (PFS) defined by the time from the start of IPI/NIVO/RT to the emergence of new lesions, progression of pre-existing lesions according to RECIST V.1.1, death, or censoring at last follow-up, whichever occurred first, (2) overall survival (OS) defined by the time between the start of IPI/NIVO/RT until the time of death of any cause or censoring at last follow-up, and (3) DCR and ORR differences between low and negative PD-L1 expression. The last follow-up date for censoring was October 15, 2025. Exploratory objectives comprised immune profiling of the tumor microenvironment by immunohistochemistry (IHC) on tumor biopsies taken after progression on first-line therapy and immune monitoring by means of flow cytometry in peripheral blood samples. Detailed definitions of all endpoints are provided in the online supplemental material. Tumor tissue and peripheral blood analyses Tumor biopsies obtained at baseline (after first-line therapy) were collected and used for IHC stains to profile the tumor microenvironment (IDO1, VISTA, LAG3, TIM3, CD8, FOXP3, CD20, HLA1, PDL1/CD68, Ki67). Slides were scanned by the Olympus VS200 research scanner (Evident Scientific, Tokyo, Japan) and analyzed by the QuPath algorithm.18 Peripheral blood mononuclear cell (PBMC) samples were obtained at baseline, on day 8 (start of RT and 1 week after the first IPI/NIVO; time point (TP) 1), on day 15 (1 week after the start of RT and 2 weeks after the first IPI/NIVO, TP2) and on day 42 (1 month after RT, at the end of phase 1 and moment of response evaluation, TP3). The study flowchart in figure 1 depicts the PBMC and tumor biopsy sampling schedule in relation to treatment administration. For both the baseline biopsies and PBMC samples, the results were compared between PR and PD patients defined by the best response achieved. Detailed procedures for pathology and PBMC analyses are provided in the online supplemental material, including the specific markers and antibodies used (online supplemental tables 1 and 2). Statistical analysis Given the historical ORR of ∼10% for standard second-line chemotherapy after progression to first-line anti-PD-1 therapy, we aimed to demonstrate superior efficacy. The Simon’s two-stage design was used to test the null hypothesis that the true ORR is ≤10%.19 One-sided significance level was set at 5%. In the first stage, 22 patients were enrolled. At least one response was observed and the study proceeded to a second stage with an additional 8 patients (to a total of 30). The null hypothesis would be rejected if two or more responses were observed among the 30 patients in phase 1 of the study, providing 80% power to detect a true response rate of >10%. Descriptive statistics were used to summarize frequencies and percentages for categorical variables, and median with ranges for continuous variables. Kaplan-Meier curves were used to estimate survival outcomes, which were assessed using log-rank tests, with comparisons between PD, SD and PR as best achieved response. Differences in baseline characteristics between patients with PR or PD as best response were assessed by χ2 tests for categorical variables and t-tests for continuous variables, for which equality of variances was evaluated by Levene’s test. Treatment-induced fold changes of immune cell subset rates in PBMC samples were evaluated using a ranked linear mixed model adjusted with a post hoc Bonferroni test,20 applied both for time point comparisons to baseline and for comparisons between partial responders and PD patients. IHC results were analyzed by Mann-Whitney U tests and visually displayed using GraphPad Prism (V.10.6.0; GraphPad Software, San Diego, California, USA). Statistical analyses were performed using R (V.4.2.1; R Foundation for Statistical Computing, Vienna, Austria). A p value <0.05 was considered statistically significant. Results Patient disposition and characteristics A total of 32 patients were enrolled into the study, of which 31 were evaluable at baseline (97%). One patient started study treatment but was excluded from further analysis due to the lack of a measurable lesion. Three patients discontinued treatment before the first study evaluation point at 6 weeks due to rapid clinical progression. Of the 28 patients who continued treatment, 8 discontinued prior to the 12-week assessment due to radiologic progression (n=6), toxicity (n=1), or withdrawal of consent (n=1). The patient inclusion and study assessment flowchart is presented in figure 2. Patient disposition of the RECLAIM trial. *=One patient was excluded due to lack of a measurable lesion following radiotherapy. Table 1 summarizes the baseline patient characteristics. The mean age was 66.3±9.9 years. Most patients were people who smoke currently or people who smoked previously (97%), had an adenocarcinoma (74%), and received pembrolizumab as first-line anti-PD-1 therapy combined with platinum doublet chemotherapy (88%). Four patients received another anti-PD(L)1 therapy (NIVO, atezolizumab, durvalumab), due to clinical considerations, including pembrolizumab-related toxicity, histology-specific choice (large cell neuroendocrine carcinoma) or as part of a chemoradiotherapy sequence. At baseline, 55% of tumors exhibited a negative (<1%) PD-L1 expression and 42% a low PD-L1 expression (1–49%). After inclusion, one patient (3%) was later found to have a tumor with high PD-L1 expression (≥50%), despite initial assessments indicating negative or low expression, which was identified only after initiation of phase 1. Most patients (n=22, 71%) received 3×8 Gy to a single lesion, predominantly on lung metastases or intrathoracic lymph nodes, while a subset received 3×8 Gy to multiple lesions (n=9, 29%) or a combination of 3×8 Gy to a single lesion with additional lesion(s) receiving 1×8 Gy (n=14, 45%). 17 patients (55%) received 3×8 Gy on a single lesion without additional 1×8 Gy to other lesions. A complete overview of all irradiated lesions is provided in online supplemental table 3. Table 1Baseline patient characteristicsAll patientsN=31 (%)Progressors*N=13 (%)Responders†N=9 (%)P value‡Age in years (mean, ±SD)66.3±9.969.9±9.259.6±5.7 0.007 Days between last anti-PD(L)1 and start RECLAIM (median, range)61 (14–714)69 (14–164)49 (31–714)0.502bSLD (mm)57 (15–143)76 (30–143)48 (30–70) 0.046 SexMale18 (58.1)9 (69.2)4 (44.4)0.245Female13 (41.9)4 (30.8)5 (55.6)Smoking statusNever1 (3.2)0 (0.0)0 (0.0)0.899Former20 (64.5)9 (69.2)6 (66.7)Current10 (32.2)4 (30.8)3 (33.3)ECOG PS012 (38.7)3 (23.1)5 (55.6)0.201117 (54.8)8 (61.5)4 (44.4)22 (6.5)2 (15.4)0 (0.0)Laboratory (median, range)CRP, mg/L19.0 (1.0–267.0)64.0 (1.8–267.0)11.0 (1.0–51.0) 0.010 Hemoglobin, g/L124 (71–150)118 (71–147)124 (106–148)0.235Thrombocytes, ×109/L269 (119–612)284 (170–621)250 (119–432)0.475Leukocytes, ×109/L7.5 (3.9–16.2)10.1 (3.9–16.2)7.0 (4.3–11.7) 0.041 Neutrophils, ×109/L5.32 (1.8–13.6)7.11 (1.8–13.6)3.53 (2.6–7.5) 0.017 Lymphocytes, ×109/L1.34 (0.6–3.0)1.32 (0.6–3.0)1.28 (0.8–3.6)0.853NLR3.51 (1.36–10.27)4.70 (1.36–10.27)3.10 (1.36–3.74) 0.014 Albumin, g/L36 (21–43)34 (21–37)37 (30–43) 0.027 Creatinine, umol/L79 (57–164)78 (63–164)79 (65–134)0.710ALAT, U/L20 (6–76)21 (6–76)22 (10–29)0.111LDH, U/L239 (129–982)315 (160–982)268 (129–780)0.590TSH, mU/L1.20 (0.11–63.0)1.10 (0.11–63.00)1.30 (0.26–2.20)0.233T-stage§T01 (3.2)0 (0.0)1 (11.1)0.535T17 (22.6)3 (23.1)4 (44.4)T23 (9.7)2 (15.4)1 (11.1)T38 (25.8)3 (23.1)1 (11.1)T411 (35.5)5 (38.5)2 (22.2)Tx1 (3.2)0 (0.0)0 (0.0)N-stage§N03 (9.7)2 (15.4)0 (0.0)0.466N11 (3.2)0 (0.0)0 (0.0)N211 (35.5)5 (38.5)4 (44.4)N316 (51.6)6 (46.2)5 (55.6)M-stage§M1a6 (19.4)1 (7.7)2 (22.2)0.331M1b2 (6.5)2 (15.4)0 (0.0)M1c23 (74.2)10 (76.9)7 (77.8)HistologyAdenocarcinoma23 (74.2)9 (69.2)7 (77.8)0.127Squamous4 (12.9)3 (23.1)0 (0.0)NOS2 (6.5)1 (7.7)0 (0.0)LCNEC2 (6.5)0 (0.0)2 (22.2)TumorPD-L1<1%17 (54.8)7 (53.8)6 (66.7)0.6441–49%13 (41.9)5 (38.5)3 (33.3)≥50%1 (3.2)1 (7.7)0 (0.0)ICI of prior linePembrolizumab27 (87.1)11 (84.6)8 (88.9)0.680Nivolumab2 (6.5)1 (7.7)1 (11.1)Atezolizumab1 (6.3)0 (0.0)0 (0.0)Durvalumab1 (6.3)1 (7.7)0 (0.0)P-values in bold are statistically significant.*Defined as radiological or clinical progressive disease as best response achieved (by RECIST V.1.1).†Defined as either complete response or partial response as best response achieved (by RECIST V.1.1).‡Two-sided p values measured via χ² and independent samples t-test, equal variances assessed by Levene’s test.§According to the eighth IASLC TNM staging.ALAT, alanine transaminase; bSLD, baseline sum of longest diameters; CRP, C-reactive protein; ECOG, Eastern Cooperative Oncology Group; ICI, immune checkpoint inhibitor; LCNEC, large cell neuroendocrine carcinoma; LDH, lactate dehydrogenase; NLR, neutrophil-lymphocyte ratio; NOS, not otherwise specified; PD-L1, programmed death-ligand 1; PS, performance score; RECIST, Response Evaluation Criteria in Solid Tumors; TNM, tumor, node, metastases; TSH, thyroid-stimulating hormone. Baseline patient characteristics P-values in bold are statistically significant. Defined as radiological or clinical progressive disease as best response achieved (by RECIST V.1.1). Defined as either complete response or partial response as best response achieved (by RECIST V.1.1). Two-sided p values measured via χ² and independent samples t-test, equal variances assessed by Levene’s test. According to the eighth IASLC TNM staging. ALAT, alanine transaminase; bSLD, baseline sum of longest diameters; CRP, C-reactive protein; ECOG, Eastern Cooperative Oncology Group; ICI, immune checkpoint inhibitor; LCNEC, large cell neuroendocrine carcinoma; LDH, lactate dehydrogenase; NLR, neutrophil-lymphocyte ratio; NOS, not otherwise specified; PD-L1, programmed death-ligand 1; PS, performance score; RECIST, Response Evaluation Criteria in Solid Tumors; TNM, tumor, node, metastases; TSH, thyroid-stimulating hormone. Response rates and survival outcomes In the intention-to-treat (ITT) population, the ORR was 7% at 6 weeks, 10% at 12 weeks, and 29% when assessed as best overall response. The first response in those nine patients (29%) was achieved after a median of 4.13 months (range 1.4–7.0). The median duration of response was 13.3 months (range 3.1–42.9), with maximal sum of longest diameters (SLD) reduction observed beyond 4 months in six out of nine patients (67%), consistent with delayed tumor regression in a subset of patients. Corresponding DCRs were 58% at 6 weeks and 39% at 12 weeks. No significant difference in response was observed according to tumor PD-L1 expression, with ORR of 35%, 23% and 0% in the PD-L1 <1% (n=17), PD-L1 1–49% (n=13) and PD-L1 ≥50% (n=1) groups, respectively (p=0.665, online supplemental table 4). Irradiated lesions showed heterogeneous changes in diameter at 6 and 12 weeks in response to therapy (online supplemental figure 1, tables 5 and 6). The median OS in the whole group was 10.1 months (online supplemental figure 2). When compared between treatment responses, as depicted in figure 3, the median OS was 3.1 months (95% CI 1.6 to 4.6) for 13 patients with PD, 13.5 months (95% CI 8.8 to 18.3) for 9 with SD, and 22.5 months (95% CI 16.1 to 28.8) for 9 patients achieving a response, including 8 with PR and 1 with CR (p<0.001). The Kaplan-Meier curves of PFS outcomes are presented in online supplemental figures 2 and 3. The median PFS was 3.0 months and differed significantly by best response (median PFS 1.4, 4.0, and 9.7 months for PD, SD, and PR/CR, respectively, p<0.001). According to PD-L1 expression, median PFS and OS were 4.2 and 10.1 months in the PD-L1 negative (<1%) group and 2.9 and 8.6 months in the PD-L1 positive (≥1%) group, respectively (online supplemental figure 4). Radiologic tumor response and OS outcome. (A) Spider plot showing changes in SLD per RECIST V.1.1, categorized by best overall response (PD in red, SD in orange, PR in light green and CR in dark green). (B) Kaplan-Meier curve of OS according to best response (PD in red, SD in orange and PR+CR in green). CR, complete response; mOS, median overall survival; OS, overall survival; PD, progressive disease; PR, partial response; RECIST, Response Evaluation Criteria in Solid Tumors; SD, stable disease; SLD, sum of longest diameter. Baseline characteristics were compared between the patients with PR/CR (n=9) and those with PD (n=13) as best response. Baseline SLD (bSLD) per RECIST V.1.1, age, inflammatory markers, and albumin levels differed significantly between these two groups (table 1). When the overall number of irradiated lesions was compared between progressors and responders, no significant difference was observed (table 1). However, progressors showed a trend toward a higher number of lesions treated with additional 1×8 Gy, although this did not reach statistical significance (online supplemental table 3). Adverse events AEs were reported for all patients of the ITT population. All patients experienced at least one treatment-emergent AE (TEAE), and 61% experienced a grade 3 TEAE (online supplemental table 7). No grade 4 TEAE was observed, while the observed grade 5 TEAEs were not related to study treatment. Treatment-related AEs (TRAEs) were also reported in all 31 patients, with 8 (26%) developing a grade 3 TRAE (table 2). No grade 4 or 5 TRAE was observed. Online supplemental table 8 shows an overview of all the reported TRAEs, including those occurring in less than 10% of patients. 20 patients (65%) experienced a serious AE, 13 of which were grade 5 events related to disease progression in 12 patients and a colonic perforation in 1 patient, all of which were not related to study treatment. Immune-related AEs (irAEs) were observed in 21 (68%) patients, with a single grade 3 irAE as the highest severity. Table 2Treatment-related adverse eventsTRAEsAny graden (%)Grade 1nGrade 2nGrade 3nAny31 (100)31268Electrolyte disorders22 (71.0)1921Dermatitis/rash14 (45.2)671Anemia13 (41.9)661Transaminitis/hepatitis13 (41.9)1210Appetite loss/nausea/vomiting11 (35.4)182Alk Phos increased10 (32.3)901Amylase/lipase increased10 (32.3)901Fatigue10 (32.3)550GGT increased8 (25.8)701Pain8 (25.8)431Constipation/diarrhea7 (22.6)340Thyroid disorders5 (16.1)410Lymphopenia4 (12.9)310Cough3 (9.7)210Eye disorders3 (9.7)210Pneumonitis3 (9.7)030Renal function decreased3 (9.7)210irAEsAny21 (67.7)10121Leading to treatment discontinuation5 (16.1)050Dermatitis/rash12 (38.7)561Thyroid disorders5 (16.1)410Pneumonitis3 (9.7)030Transaminitis/hepatitis1 (3.2)010Colitis1 (3.2)010Arthritis/myositis1 (3.2)100TRAEs monitored in all patients that started with phase 1, in the period from the start of therapy until 100 days until last immunotherapy dose received. In the bottom part, irAEs are shown.GGT, gamma glutamyl transferase; irAEs, immune-related adverse events; TRAEs, treatment-related adverse events. Treatment-related adverse events TRAEs monitored in all patients that started with phase 1, in the period from the start of therapy until 100 days until last immunotherapy dose received. In the bottom part, irAEs are shown. GGT, gamma glutamyl transferase; irAEs, immune-related adverse events; TRAEs, treatment-related adverse events. Immunohistochemistry and peripheral blood immune monitoring IHC performed on 20 available baseline biopsies revealed that IDO1 expression in the total tissue area was significantly higher in patients with PR as best overall response compared with patients with PD, while other immune markers showed no significant differences between these groups (online supplemental figure 5 and table 9). Online supplemental figure 6 depicts an example of the difference in IDO1 expression in a PR versus PD patient. Peripheral blood immune monitoring was performed on samples from 24 patients with sufficient PBMCs available at baseline and three time points on treatment. Flow cytometry data of T-cell subsets at baseline did not reveal any significant association with treatment response. In general, T-cell activation was evident at the first assessed time point (1 week after initiation of IPI/NIVO), remained detectable at day 15 (1 week after start of RT), and gradually returned toward baseline levels by day 42 despite ongoing NIVO dosing. These changes were most pronounced in patients who achieved a PR. Figure 4 summarizes treatment-related changes in the frequency, proliferation and activation state of circulating T-cell populations. Treatment generally induced an immediate increase in effector memory (EM) CD4+ T-helper (Th) cells, which remained elevated throughout the treatment course (figure 4A). One week after RT, a significant decrease in naïve CD8+ T cells was observed, accompanied by a shift toward EM CD8+ T cells (online supplemental figure 7). Tregs generally showed a modest but consistent increase at all post-treatment time points compared with baseline, together with transient elevations in proliferation (Ki67) and activation (HLA-DR) (figure 4B). Across both central memory (CM) and EM Th and CD8+ T-cell subsets, treatment generally induced significant but transient increases in proliferation (Ki67) and activation (HLA-DR) (figure 4C,D). Expression of CD39 and CTLA-4 on Th and CD8+ T cells also rose during treatment (online supplemental figure 7). No significant differences in immune populations were observed between patients who received irradiation on single lesions versus multiple lesions, or between those with irradiated lymph nodes versus other irradiated sites (online supplemental figures 8 and 9). Longitudinal fold changes in PBMC blood samples for participants with sufficient samples (n=24), out of the total study population (n=31), measured at BL, start RT/1 week post IPI-NIVO (TP1), 1 week post RT/2 weeks post IPI-NIVO (TP2) and 1 month post RT at the moment of response evaluation (TP3), analyzed by mixed-effects model with post hoc Bonferroni. The graphs depict changes in (A) frequencies of CD4 EM of T helper cells, (B) frequencies, proliferation and activation of Tregs, (C) proliferation and activation of CM and EM T helper cells, (D) proliferation and activation of CD8 CM and EM T cells. Mean percentages are depicted at BL below the dotted line. Adjusted p value (two-sided): *≤0.05, **≤0.01, ***≤0.001. BL, baseline; CM, central memory; EM, effector memory; IPI, ipilimumab; NIVO, nivolumab; PBMC, peripheral blood mononuclear cells; TP, time point; Treg, regulatory T cell. Figure 5 shows pretreatment and on-treatment flow cytometric T-cell data stratified by patients with PR/CR (n=9) and PD (n=9) as best clinical response. Partial responders showed generally more pronounced, though transient, on-treatment increases in CD39 expression on Th cells, together with stronger transient elevations in proliferation (Ki67) and activation markers (HLA-DR/CTLA-4) within both CM and EM Th and CD8+ T cell subsets (figure 5A,B). Transient increases in Treg proliferative and activation state were similar between patients with PR or PD (figure 5C). Online supplemental tables 10 and 11 provide an overview of T cell–related flow cytometric PBMC analyses. Longitudinal fold changes in PBMC blood samples of patients with partial/complete response (n=9, in green) and progressive disease (n=9, in red) as best achieved response, measured at BL, start RT/1 week post IPI-NIVO (TP1), 1 week post RT/2 weeks post IPI-NIVO (TP2) and 1 month post RT at the moment of response evaluation (TP3), analyzed by mixed-effects model with post hoc Bonferroni. The graphs depict changes in (A) T helper CD39 expression, and the proliferation, activation and CTLA4 expression in CM and EM T helper cells, (B) CD39 expression, proliferation, activation and CTLA4 expression in CD8 CM and EM T cells, (C) CD39 expression, proliferation, activation and CTLA expression in Tregs. Mean percentages are depicted at BL below the dotted line. Adjusted p value (two-sided): *≤0.05, **≤0.01, ***≤0.001. BL, baseline; CM, central memory; CTLA4, cytotoxic T-lymphocyte-associated protein 4; EM, effector memory; IPI, ipilimumab; NIVO, nivolumab; PBMC, peripheral blood mononuclear cells; TP, time point; Treg, regulatory T cell. Discussion The phase II RECLAIM trial is, to our knowledge, the first prospective clinical trial to investigate the use of dual PD-1 and CTLA-4 blockade in combination with RT in patients with metastatic PD-L1 low or negative NSCLC who progressed following first-line anti-PD-1-based therapy. We observed an ORR of 7% at 6 weeks, 10% at 12 weeks and 29% based on best response in the ITT population. Corresponding DCRs were 58% and 39% at 6 and 12 weeks. As ORR and DCR at 6 and 12 weeks were the protocol-defined primary endpoints, best overall response should be interpreted as a supportive measure of delayed antitumor activity. Response to study treatment was associated with improved OS outcomes and a manageable toxicity profile, with no significant differences between PD-L1 expression groups. Currently, patients progressing on first-line immunotherapy combined with chemotherapy typically receive second-line single-agent chemotherapy, predominantly taxane-based regimens. Historical data suggest response rates ranging from 7–23%, with a recent multicenter cohort reporting an ORR of 16%, DCR of 42%, and median OS of 8.1 months.21 22 In contrast, our study yielded nearly a twofold higher ORR of 29% and improved DCR, suggesting enhanced efficacy in this immunotherapy-refractory population. Consistent with prior immunotherapy studies, responses evolved gradually, with initial tumor shrinkage typically observed at the first radiographic assessment (∼6 weeks), while best responses with maximal SLD reduction most frequently occurred ≥4 months after treatment initiation. This finding underlines the importance of an adequate follow-up duration to capture delayed responses and, more importantly, suggests that patients with more indolent tumor biology may derive the greatest benefit from this treatment strategy. In addition to responders (n=9, 29%), a subset of patients achieved SD (n=9, 29%), whereas a third subgroup of 13 patients experienced rapid disease progression, accounting for 42% of the study population. These rapid disease progressors had a median OS of only 3.1 months and were characterized by higher age, elevated inflammatory markers (ie, CRP, neutrophils, leukocytes and neutrophil-to-lymphocyte ratio), low albumin levels, and increased tumor burden at baseline (as reflected by bSLD and number of lesions treated with palliative dose of 1×8 Gy). These findings suggest that systemic inflammation and tumor load may serve as prognostic markers and may aid in refining patient selection in future trials. In terms of toxicity, TRAEs occurred in all patients, most commonly electrolyte abnormalities, rash, anemia, and transaminitis. Grade 3 events occurred in 26% of patients and primarily involved electrolyte disturbances. irAEs were reported in 68% of patients, mainly rash and thyroid dysfunction, and led to treatment discontinuation in 16%. Most AEs were low grade, and the overall safety profile was consistent with that previously reported for dual immunotherapy,15 23 supporting the tolerability of this multimodal regimen. The clinical activity observed in RECLAIM is supported by an expanding body of evidence demonstrating synergistic effects between PD-1 and CTLA-4 inhibition.42426 CTLA-4 blockade facilitates the priming of naive T-cells in tumor-draining lymph nodes, while PD-1 inhibition restores exhausted T-cell effector function within the tumor microenvironment.27 The CheckMate 227 trial demonstrated a survival benefit of first-line IPI plus NIVO over chemotherapy in metastatic NSCLC, independent of PD-L1 status, with an ORR of 35.9% and a median OS of 17.1 months (vs 14.9 months with chemotherapy; HR 0.79, 95% CI 0.65 to 0.96). Notably, the safety profile observed in CheckMate 227 with grade 3–4 TRAEs in 32.8% of patients, was in line with our findings.15 A recent meta-analysis confirmed a survival benefit when immunotherapy was combined with RT in lung cancer.28 The authors identified 19 articles including a total of 5109 patients with lung cancer, and observed longer PFS (HR 0.68, 95% CI 0.62 to 0.75, p<0.00001) and OS (HR 0.70, 95% CI 0.62 to 0.79, p<0.00001) in the ICI plus RT group compared with ICI alone without affecting AEs (OR 1.09, 95% CI 0.80 to 1.50, p>0.05). This observation aligns with preclinical data suggesting that RT induces immunogenic cell death and enhances antigen presentation, allowing for synergy when administered sequentially.12 29 Dual CTLA-4/PD-1 blockade combined with RT in this study was associated with systemic immune activation. Early, transient increases in proliferation (Ki67) and activation (HLA-DR) were observed across circulating CD4+ and CD8+ CM and EM T cell subsets, accompanied by modest activation of Tregs. These treatment-related effects are consistent with the expected response to dual ICI, and the shift from naive to EM CD8+ T cells following RT aligns with the possibility of enhanced antigen release and peripheral T-cell differentiation. Patients who responded to treatment exhibited more pronounced on-treatment increases in CD39+ CD4+ T helper cells and CD39+ CD8+ central and EM T cells. This is consistent with prior studies demonstrating that CD39 expression potentially identifies tumor-reactive CD8+ T-cell populations and is associated with clinical benefit from immune checkpoint blockade.16 30 31 Enhanced proliferation, activation and CTLA4 expression in EM and CM CD4+ and CD8+ T-cell subsets further support a favorable antitumor immune response.3235 Although dual IPI/NIVO therapy induced an early increase in activated and proliferating Tregs at week 1, the effector T cell responses remained robust, with sustained activation predominantly within CD8+ CM and EM populations, an effect possibly enhanced by the addition of RT, which can potentiate CD8+-mediated antitumor immunity.36 The observed increases in CTLA-4 expression within memory subsets may reflect previously described compensatory upregulation following IPI.37 Furthermore, although prior research suggested that multi-site radiation may promote neoantigen release and that lymph node irradiation should be avoided to preserve immune priming,3841 no differences in immune activation were observed between patients treated at a single versus multiple lesions, or between those with and without lymph node irradiation. Finally, higher IDO1 expression within tumors of responding patients may indicate an immunosuppressive microenvironment that was more effectively countered by the combination of CTLA-4/PD-1 blockade and RT.42 Overall, these results indicate that dual ICI plus RT reliably induces systemic T-cell activation in this patient population, but the magnitude and quality of this response differed by clinical outcome. Responding patients tended to develop a more robust and sustained activation of antigen-experienced memory T-cell populations, while patients with PD showed more limited changes in response to therapy. While this trial demonstrated improved tumor responses in immunotherapy-refractory negative and low PD-L1 expressing tumors, several limitations should be acknowledged. This study was limited by its small sample size and single-arm design without a control arm. Furthermore, the optimal strategy for integrating RT with immune checkpoint inhibition remains to be defined. Emerging evidence suggests that RT delivered prior to PD-1 blockade provides a more favorable immune-priming window than RT administered after PD-1 therapy,43 raising the possibility that our sequencing (RT following PD-1 exposure) may not have fully leveraged the immunogenic potential of RT. Future studies should systematically evaluate the timing of RT relative to ICI, as well as key RT parameters such as dose, fractionation, and number of treated sites, to better define how to maximize therapeutic synergy and clinical benefit.13 44 Besides, the study population remained heterogeneous with respect to PD-L1 expression, disease burden, number of irradiated lesions, and prior treatment exposure, all of which may influence immune dynamics and clinical outcome. These factors should be considered when interpreting the findings. The significant differences in baseline systemic inflammation and tumor load reflected by the bSLD may also help to better identify patients most likely to benefit from this regimen and warrant prospective evaluation in future stratified studies. Conclusion This prospective phase II trial showed that the combination of IPI and NIVO with subablative RT yielded favorable response rates in patients with metastatic NSCLC with PD-L1 low or negative tumors who had progressed on first-line anti-PD1 therapy. The regimen was well tolerated, with a safety profile consistent with prior dual ICI studies. Durable clinical responses were marked by enhanced T-cell activation in peripheral blood in response to treatment, especially in patients achieving PR or CR. These results support the feasibility of this multimodal strategy in a population with limited therapeutic options. Prospective studies, including randomized controlled trials, are warranted to confirm these results, refine patient selection, RT dosing, and treatment sequencing, and enable broader clinical adoption. Supplementary material