Combining Dual Checkpoint Immunotherapy with Ablative Radiation to All Sites of Oligometastatic Non-Small Cell Lung Cancer: Toxicity and Efficacy Results of a Phase Ib Trial

Abstract

Background:

Ablative local treatment of all radiographically detected metastatic sites in patients with oligometastatic NSCLC increases progression free survival (PFS) and overall survival (OS). Prior studies demonstrated safety of combining stereotactic body radiation therapy (SBRT) with single agent immunotherapy. We investigated the safety of combining SBRT to all metastatic tumor sites with dual checkpoint, anti-CTLA-4 and anti-PD-L1 immunotherapy for patients with oligometastatic NSCLC.

Methods:

We conducted a phase Ib clinical trial in patients with oligometastatic NSCLC with up to 6 sites of extracranial metastatic disease. All sites of disease were treated with stereotactic body radiation therapy (SBRT) to a dose of 30 – 50 Gy in 5 fractions. Dual checkpoint immunotherapy was started 7 days following completion of radiation utilizing anti-CTLA-4 (tremelimumab) and anti-PD-L1 (durvalumab) immunotherapy for a total of four cycles followed by durvalumab alone until progression or toxicity.

Results:

Seventeen patients were enrolled on study, with 15 patients receiving at least one dose of combination immunotherapy per protocol. The study was closed early (17 of planned 21 patients) due to slow accrual during COVID 19 pandemic. Grade 3+ treatment related adverse events were seen in 6 patients (40%). Of these, only one was felt possibly related to the addition of SBRT to immunotherapy. Median PFS was 42 months while median OS has not yet been reached.

Conclusions:

Delivering ablative SBRT to all sites of metastatic disease in combination with dual checkpoint immunotherapy did not result in excessive rates of toxicity compared to historical studies of dual checkpoint immunotherapy alone. While not powered for treatment efficacy results, durable PFS and OS results suggest potential therapeutic benefit compared to immunotherapy or radiation alone in this patient population.

Introduction:

Metastatic non-small cell lung cancer is the leading cause of cancer related death in the United States(1). Historically, cytotoxic chemotherapy was the standard treatment option, though outcomes were poor with a median survival of 10–12 months(2, 3). Significant progress in patient outcomes was achieved with the introduction of immune checkpoint inhibitors (ICI) with major positive landmark trials of ICI against programmed cell death protein 1 (PD-1) and programmed cell death ligand 1 (PD-L1). Tumor expression levels of PD-L1 may guide treatment decisions and are a predictive biomarker for patient outcomes(4). For patients with driver negative advanced NSCLC whose tumor express PD-L1 ≥50%, monotherapy with ICI is the preferred approach. The National Comprehensive Cancer Network (NCCN) guidelines currently recommend pembrolizumab, atezolizumab, or cemiplimab monotherapy as treatment options, regardless of histology(5). For patients whose tumor express lower levels of PD-L1 expression (≥1–49% or negative), combination of ICI and platinum-based chemotherapy is commonly used(5).

Dual checkpoint immunotherapy combines anti-PD-1/PD-L1 treatment with anti-cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) treatment with a goal of enhancing T cell mediated tumor response. Combination therapy has shown greater response rates in other cancer types like melanoma(6) and RCC(7). Several studies have investigated the addition of anti-PD-1/PD-L1 + anti- CTLA-4 agents in NSCLC. This approach, nivolumab plus ipilumumab and nivolumab plus ipilumumab in combination with platinum chemotherapy, was approved as an optional first-line treatment for patients with advanced NSCLC with tumor PD-L1 expression ≥1%, based on clinical data from the Phase 3 CheckMate 227 study(8) and in patients with advanced NSCLC, independent of tumor PD-L1 expression, based on the phase 3 CheckMate 9LA(9). Most recently, the phase III POSEIDON study led to the approval of tremelimumab plus durvalumab and chemotherapy in first-line metastatic NSCLC(10), with significant improval in overall survival (OS) and progression-free survival (PFS) compared to platinum-chemotherapy. Based on these findings, dual checkpoint immunotherapy now has category 1 recommendations for metastatic NSCLC and is an “other recommended” treatment option alternative to single checkpoint immunotherapy alone in NCCN guidelines (11). Despite these improvements with ICI therapy compared to standard chemotherapy alone, overall outcomes remain poor for the majority of these patients with systemic therapy alone. Thus further investigation combining systemic therapy with ablative local treatments is warranted.

Oligometastatic non-small cell lung cancer represents an intermediate stage between locally advanced disease and widely disseminated metastatic disease(12). Ablative local treatment to all identifiable metastatic sites of disease in patients with various oligometastatic cancers increases progression free survival and overall survival.(13, 14) Durable long term local control of all sites of metastatic disease leads to improved progression free survival(15). However, most patients fail distantly at new sites of disease. Thus, there is growing interest in combining ablative local treatment of radiographically detectable tumors with systemic therapy to further improve the control of radiographically occult tumor sites. Local therapy to sites of gross disease may also allow improved response rates to systemic immunotherapy by reducing overall disease burden and overcoming suppressive immune mechanisms in established tumor microenvironments(16–22). Prior studies have shown the safety of combining high dose stereotactic body radiation therapy (SBRT) with single agent anti-PD1/PD-L1 therapy(23–25). Recent consensus guidelines have provided some insight on combining SBRT with dual checkpoint immunotherapy, based on 37 patients identified in the consensus process (26). However, the safety and efficacy of SBRT to all sites of metastatic disease combined with dual checkpoint therapy is less well established.

Here, we performed a phase Ib study investigating the safety of SBRT combined with dual checkpoint durvalumab and tremelimumab therapy in patients with oligometastatic NSCLC with up to 6 sites of metastatic disease.

Methods

This was a Phase Ib study conducted at a single institution between March 2018 – March 2020, under an IRB approved protocol (Anonymyized for Review). Written informed consent was obtained from all participants prior to enrollment in the trial. The study was conducted in compliance with the Declaration of Helsinki.

Patient Eligibility:

Patients age 18 and over with pathologically confirmed metastatic NSCLC were enrolled on trial. Eligibility criteria included Eastern Cooperative Oncology Group (ECOG) performance status 0–1, with up to 6 sites of extracranial metastases safely amenable to SBRT radiation. Patients could not have prior immunotherapy treatment. Patients must have tumors that lacked sensitizing EGFR mutation or ALK rearrangement. All patients with adenocarcinoma histology were required to have ALK and EGFR testing prior to enrollment and were ineligible if actionable targets were identified. If a patient had squamous histology, then EGFR and ALK testing was not required. PD-L1 expression or testing was not required for study enrollment. Patients with untreated brain metastases or spinal cord compression were not eligible for enrollment. A treatment washout period of 14 days for prior surgery, radiation, or chemotherapy was required prior to starting treatment. Patients did not need to have stable disease prior to enrollment and patients with newly diagnosed metastatic disease were eligible. Full inclusion and exclusion criteria are found in supplementary table 1.

Treatment sequence:

SBRT was delivered starting on day 0 for a total of 5 treatments, 2–3 treatments per week. Dual checkpoint inhibitor treatment started 7 days after final radiation treatment and consisted of 4 cycles of tremelimumab (75 mg) and durvalumab (1500 mg) every 28 days followed by durvalumab (1500 mg) every 28 days until disease progression (Supplemental Figure 1).

Radiation Technique:

Patients with 6 or less extracranial sites of metastases were treated with SBRT between 30 – 50 Gy in 5 fractions (BED₁₀ 48 – 100 Gy). All sites were treated concurrently, with 5 fractions delivered to each site. A site may have multiple tumor lesions within it as long as the combined maximal dimension of gross tumor volume (GTV) of the site is 8 cm or less and the targets could be encompassed in a single isocenter. Therefore, multiple mediastinal lymph nodes or a central pulmonary lesion and mediastinal lymph node may be counted as a single site if meeting these criteria. All areas of gross disease (visible via imaging) were targeted with radiation, as allowable per normal tissue constraints (Supplemental Table 2). Normal tissue constraints were prioritized over target coverage based on supplemental table 2 and targets were undercovered to meet constraints. Overlap with prior radiation treatment fields was not allowed. Up to six sites of disease were allowed based on prior clinical experience(27) and to allow for treatment of primary disease plus up to 5 sites of metastatic disease.

Toxicity Analysis:

The primary study endpoint was safety of combined radiation and dual checkpoint inhibitor treatment. Adverse events (AEs) and Dose Limiting Toxicities (DLTs) were monitored and graded by investigators using the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE) version 4.03. A DLT was defined as any Grade 3 or higher toxicity that occurs from day of first SBRT until 28 days post completion of first durvalumab and tremelimumab infusion. Treatment related AEs were determined by investigator assessment and scored as definitely related, probably related, possibly related, unlikely related, or unrelated. All AEs were reviewed by each co-investigator to determine relatedness to therapy, with final consensus determined by majority. AEs considered definitely, probably, or possibly related to treatment were considered treatment related AEs. Immune related AEs were assessed longitudinally throughout follow up. DLTs were defined as outlined in supplemental table 3.

Survival Outcomes:

The secondary study endpoints were PFS and OS. Progressive disease of irradiated lesions was defined by RECIST version 1.1 criteria (28). Progressive disease outside of irradiated targets required pathological confirmation to be called progressive disease. Unplanned exploratory analysis of PFS and OS was performed following patient stratification based on prior history of brain metastases.

Follow Up:

Patients were followed every 8 weeks for the first year after completing SBRT and dual checkpoint therapy (not including maintenance therapy) and every 12 weeks thereafter until time of disease progression. Patients were censored at time of last follow up.

Sample Size:

The planned sample size for this trial was a total of 21 patients. This sample size chosen was sufficient to provide accuracy in estimating the toxicity rates. Specifically, with a planned sample size of 21, toxicity rates would have been estimated with a standard error of less than ^~10% and the corresponding 95% confidence interval would have been no wider than 40%, assuming that the true toxicity rates were 33% or less. Due to a slow accrual rate during COVID-19 pandemic, the study was terminated early after a total of 17 patients were accrued.

Statistical Analysis:

The number of adverse events with an attritubiton of at least possibly related were summarized in tabular format, stratrified by type and severity. Toxicity rates were calculated as the proportion of patients experience toxicities divided by the total number of patients in the safety population. The safety population for this study was defined as all patients receiving SBRT and at least a single immunotherapy treatment. Efficacy analyses were contucted using both the intent-to-treat and per-protocol populations. PFS was defined as the number for months from the start day of SBRT to the date of documented disease progression. If a patient did not experience a progression event, then PFS time was censored at the last disease assessment date. OS was defined as the number of months from the start date of SBRT to the date of death or last known survival (censored). If a patient died without evidence of progression, PFS time was censored at date of last imaging based disease assessment and treated as a censored observation event. PFS and OS were analzyed using the Kaplan-Meier method. In an exploratory analysis, PFS was compared between patients with brain metastases metasteses vs. patients without brain metastases using a log-rank test. The corresponding hazard ratio and two-sided 95% confidence interval was calculated using Cox-proprtional hazard regression analysis. All reported P-values are two-sided. Statistical analyses were conducted using SAS software (SAS Institute, Cary NC) version 9.2 and GraphPad Prism software (GraphPad Software, Inc).

Results:

Demographics:

Between March 2018 and March 2020, 17 patients enrolled on the trial. Two patients were noted to have progressive metastatic disease following enrollment, after completion of SBRT, and did not receive dual checkpoint inhibitor therapy on study (both with new intracranial lesions following SRS treatment of brain lesions prior to study enrollment). Neither of these patients had toxicity related to radiation. Patient Characteristics are shown in Table 1 for all patients enrolled on study. The median age was 68 years old; 76% of patients were male. Median follow up was 35 months. The median number of extracranial metastatic sites was 2. Seven patients had 3 or more sites of extracranial disease, including the two patients with progression prior to starting immunotherapy. The majority of patients had adenocarcinoma histology (76%). PD-L1 expression testing was not required for study enrollment; 3 patients had high tumor PD-L1 expression (PD-L1>50%), 7 patient had low expression (PD-L1 1–49%), and 4 patients had negative PD-L1 expression. Three patients did not have PD-L1 testing available. Thirteen of the seventeen patients had no prior lung cancer directed therapy, while 3 patients had prior SBRT and 1 patient had prior surgical resection for initial early stage lung cancer. No patients had prior chemotherapy or immunotherapy.

Table 1:

Patient demographics for all patients enrolled on study.

	N	% of total enrollees
Age (yrs)
Median, range	68	42–81
Number of Extracranial Sites
Median, range	2	1–6
Gender
Female	4	24
Male	13	76
Race
Black/African American	2	12
White	15	88
T Stage (at diagnosis)
T1a	1	6
T1b	4	24
T1c	4	24
T2a	4	24
T2b	2	12
T3	1	6
T4	1	6
N Stage (at diagnosis)
N0	11	65
N1	2	12
N2	4	24
M Stage (at diagnosis)
M0	2	12
M1a	2	12
M1b	7	40
M1c	6	36
Histology
Adenocarcinoma	13	76
Squamous Cell Carcinoma	2	12
Large Cell Carcinoma	1	6
Non-Small Cell Carcinoma NOS	1	6
PD-L1 Expression
Absent	4	23
Low	7	41
High	3	18
Unknown	3	18
History of Brain Metastasis
No	11	65
Yes	6	35
Previous therapy for Lung Cancer
No	13	76
Yes	4	24
Type of prior therapy
Lobectomy	1	6
Definitive SBRT	3	18

Radiation Delivery:

The most commonly treated anatomical sites were metastatic pulmonary lesions (21), thoracic lymph nodes (7), or osseous lesions (5) (Supplemental Table 4). Most patients had 2 sites of disease treated with SBRT (7 patients), while only 3 patients had 4 or more sites of metastatic disease treated (Supplemental Table 5). Twelve patients were treated to a maximum SBRT dose of 50 Gy in 5 fractions (Supplemental Table 6). SBRT dose delivered to each anatomical subsite is shown in supplemental table 7. Only 5 individual targets were undercovered (95% of volume receiving less than 95% of dose) to meet adjacent organ at risk constraints. All sites were treated simultaneously over 2 weeks.

Systemic therapy delivery

The median number of cycles of dual checkpoint immunotherapy completed in patients treated per protocol was 4 (range 1–4). Eight patients completed all 4 cycles of dual checkpoint treatment while 5 patients discontinued treatment due to toxicity and 2 patients discontinued due to disease progression (brain) before completing 4 cycles of dual checkpoint therapy. Two patients developed disease progression after completion of durvalumab and tremelimumab but prior to starting maintenance durvalumab. The median number of cycles of maintenance durvalumab in patients started on maintenance treatment was 24 (range 7 – 27). All patients who started maintenance durvalumab were transitioned to surveillance before having disease progression or toxicity leading to treatment discontinuation.

Treatment related toxicity:

Toxicity analysis was performed on the 15 patients who received at least one cycle of dual checkpoint immunotherapy per protocol. Two patients were excluded due to progressive intracranial disease following SBRT prior to receiving immunotherapy. Notably, these patients had no toxicity related to SBRT. As the primary endpoint was safety of the combined modality, the patients were not felt appropriate to add negative toxicity data without receiving dual checkpoint immunotherapy. Table 2 shows total cumulative number of adverse events as well as number of patients experiencing either treatment related or immune related adverse events. Overall rates and type of Grade 3+ adverse events scored at least possibly related to therapy are shown in table 3. Cumulative incidence of highest graded toxicity per individual patient seen in follow up period (out of 15 total patients) are shown in supplemental table 8. The most common grade 1–2 treatment related adverse events was fatigue (73% of patients). No other grade 1–2 toxicity was observed in more than 50% of patients. The most common grade 3–4 treatment related adverse events were elevation in liver transaminase levels (20% of patients). Notably, none of these patients received liver SBRT. No Grade 5 treatment related adverse events were observed. Two patients experienced dose limiting toxicity, both for grade 3 liver transaminase elevation >8 times the upper normal limit related to immunotherapy. No dose limiting toxicity was observed as a result of radiation. Overall, 21 individual grade 3 or greater adverse events were noted in a total of seven individual patients. Sixteen of the twenty-one adverse events were felt to be at least possibly related to treatment in 6 individual patients (40%). Grade 3+ adverse events were classified as immune related adverse events in 5 of the 6 patients (33.3% overall). Four out of these five patients discontinued immunotherapy due to toxicity. A seventh patient had grade 3 cerebral edema and radiation necrosis from prior brain metastasis treatment determined to be unrelated to radiation treatment delivered on this trial.

Table 2:

Summary table of total number of adverse events and number of patients experiencing grade 3 or greater treatment related and immune related adverse events.

Total Adverse Events (AEs)	460
Total Grade 3+ AEs	21
Total Immune related AEs (IRAEs)	54
Total Immune related Grade 3+ AEs	10
Total patients with IRAEs	66.7%
Total patients with G3+ IRAEs	33.3%
Total patients discontinued Tx due to IRAE	33.3%
Total Treatment Related AEs (TRAE)	191
Total Treatment Related G3+ AEs	16
Percent Patients with TRAE AEs	86.7%
Percent Patients with TRAE G3+ AEs	40.0%

Table 3:

Cumulative incidence and type of Grade 3+ adverse events scored at least possibly related to therapy.

Toxicity	Grade 3	Grade 4
ALT increased	2
AST increased	3
Colitis	1
Creatinine Phosphokinase increased		1
Dyspnea	1
Fatigue	1
GGT Increased	1
Hypoxia	1
Lipase increased	1
Maculopapular Rash	1
Pneumonitis	1
Serum amylase increase	1

One patient experienced grade 3 dyspnea, hypoxia, and fatigue felt to be possibly related to SBRT versus immunotherapy. The patient was treated with SBRT to two right upper lobe lesions and a thoracic spine vertebral metastasis. Three patients experienced grade 2 toxicity (cough, fatigue, and hypoalbuminemia) felt to be at least possibly related to SBRT (Supplemental Table 9). No additional radiation related toxicity was noted in the two patients who did not receive immunotherapy on trial due to progressive disease.

Treatment Response:

Treatment efficacy was assessed using both intent to treat analysis including the two patients who progressed prior to receiving immunotherapy as well as a per-protocol analysis. In both cases, the median PFS was 42 months (Figure 1a and Supplemental Table 10). At time of analysis, 9 patients are alive without evidence of progressive disease. Median OS has not yet been reached in either per-protocol (Figure 1b) or intent to treat analysis (Supplemental Table 10). The most common site of progression was brain (4 patients), followed by bone (2 patients), lung (2 patients), and liver (1 patient). Three of these patients had simultaneous progression in multiple sites of disease. One patient with multifocal progression did have progression at the SBRT treatment margin as well as new sites of disease at time of first progression. This was a vertebral body lesion treated to 30 Gy in 5 fractions, BED₁₀ 48 Gy, to the entire vertebral body. Progression occurred in bilateral pedicles outside of high dose radiation field. No other patients had in-field or marginal progression.

Figure 1:

Progression free and overall survival analysis of all patients receiving at least one cycle of immunotherapy per protocol. A. Progression free survival is shown via Kaplan Meier analysis. Median PFS is 42 months. B. Overall survival shown via Kaplan Meier analysis. Median overall survival not yet met.

In an unplanned exploratory analysis, evaluating treatment response based on presence of brain metastases prior to enrollment showed significant changes in treatment efficacy. On intent to treat analysis, median PFS was 42 months in patients without brain metastases compared to 2 months in patients with brain metastases (HR 8.6 (95% CI 1.6–45.2) p=0.0027) (Supplemental Table 11). Similar results were seen on per protocol analysis, with median PFS of 42 months in patients without brain metastases versus 4 months in patients with brain metastases (HR 6.1 (95% CI 1.0 – 37.0) p=0.0248) (Figure 2a and Supplemental Table 11). Median OS was not reached in either the intent to treat or per protocol analysis for patients with or without brain metastases at time of enrollment (Figure 2b and Supplemental Table 11).

Figure 2:

Progression free and overall survival analysis of patients with (n=4) or without (n=11) prior history of brain metastasis receiving at least one cycle of immunotherapy per protocol. A. Significantly longer PFS observed in patients without prior brain metastasis treated per protocol (42 vs 4 months, HR 6.1, p=0.0248). B. No significant difference observed between patients with or without prior brain metastasis treated per protocol (Median OS not reached for either group).

Discussion:

In this phase Ib study for patients with oligometastatic NSCLC, we observed that a combination of locally ablative SBRT delivered to all radiographically detected sites of disease followed by dual checkpoint immunotherapy was safe with rates of treatment related toxicities similar to those reported with dual checkpoint immunotherapy alone. Median PFS was 42 months and median OS has not yet been reached. While not powered to assess for treatment efficacy given the small number of patients, the duration of PFS and OS seen in patients treated per protocol is encouraging and compares favorably to prior studies of ablative SBRT plus single agent anti PD-L1 immunotherapy(23, 25).

To our knowledge, this is the first prospective study to evaluate the safety of combining high dose per fraction ablative radiotherapy with durvalumab and tremelimumab for treatment of oligometastatic NSCLC. Grade 3 or greater treatment related toxicity in our study was seen in 40% of patients, compared to the 23% rate of grade 3+ treatment related toxicity for dual checkpoint inhibitor treatment on the recently reported MYSTIC study combining durvalumab and tremelimumab alone. However, overall rate of any grade 3+ toxicity was similar between our trial (7/15, 46%) and the MYSTIC trial (52%)(29). These rates of overall grade 3+ toxicity for dual-agent immunotherapy are similar to the reported rate of 42% in a recent meta-analysis of 28 trials(30). Notably, only one patient in our study experienced a grade 3 toxicity felt possibly related to SBRT (grade 3 hypoxia, dyspnea, and fatigue) suggesting the addition of radiation to dual checkpoint inhibition is not associated with excessive increased risk of toxicity. While the two therapies were not given concurrently, immunotherapy was initiated immediately following radiation, suggesting rapid sequential local and systemic therapy is safe with similar toxicity compared to systemic therapy alone in these patients. These findings are similar to another phase I trial combining SBRT and dual checkpoint nivolumab and ipilimumab in polymetastatic disease, which reported grade 3+ toxicity related to SBRT in 8 out of 37 patients with 2 patients experiencing radiation related DLT(31). Addtionally, on our study, 33% of patients discontinued immunotherapy due to toxicity, but none of these toxicities were related to radiation therapy suggesting that the addition of radiation did not lead to markedly higher rates of treatment discontinuation compared to dual checkpoint therapy alone(29). Interestingly, while the rate of toxicity seen in our study is similar to reported toxicity of dual checkpoint therapy alone, the rate is higher than prior reports of ablative SBRT with single agent anti-PD-L1 therapy. SBRT to all sites of disease followed by pembrolizumab showed low rates of grade 3+ toxicity, with the highest reported toxicity being pneumonitis (6%)(25). Thus, the addition of anti-CTLA4 therapy is associated with higher rates of acute grade 3+ toxicity but the addition of ablative radiation therapy is not associated with increased toxicity compared to dual checkpoint therapy alone.

The current study was not powered for treatment efficacy. However, the long PFS and OS observed certainly are encouraging in a non-PD-L1 selected population with a mix of high, low and no PD-L1 expression. No direct comparison for oligometastatic NSCLC patients treated with dual ICI alone has been reported. The median OS for newly diagnosed patients treated with durvalumab and tremelimumab on the MYSTIC trial was 11.9 months(29). However, the overall burden of metastatic disease was not reported on that study, and likely included oligo and polymetastatic patients. In our study, patients with extracranial metastatic disease involving up to 6 sites of disease were eligible and thus the results are not applicable to polymetastatic patients. Aggressive local therapy to oligometastatic disease has been shown to improve overall disease outcomes. Long term follow up of the SABR-COMET trial evaluating SBRT in oligometastatic disease showed a 5 year OS of 42% in patients treated with SBRT(15). In NSCLC specifically, SBRT to 1–3 sites of oligometastatic disease improved PFS from 4.4 months to 14.2 months(13). Additionally, the phase II PEMBRO-RT study combined SBRT with pembrolizumab and reported a doubling of median OS from 7.6 to 15.9 months(23). A closer comparison is seen in the study by Bauml and colleagues, who targeted all sites of metastatic disease with ablative therapy followed by single agent pembrolizumab and showed a median PFS of 19.1 months(25). These studies suggest efficacy of aggressive local therapy with systemic treatment in selected patients. In our study, the median survival has yet to be reached in both our per-protocol and intent-to-treat analysis and the median PFS is 42 months, exceeding the results seen in these prior studies for immunotherapy or single agent immunotherapy with radiation. This improvement in PFS seen in our study may suggest the critical addition of anti-CTLA4 treatment to upfront, ablative radiotherapy for oligometastatic patients. The combination and timing of immunotherapy with ablative therapy to all sites of disease warrants future investigation.

Aggressive local treatment to reduce the overall burden of metastatic disease may improve the efficacy of systemic immunotherapy by allowing the immune system to target microscopic disease following ablative treatment(12, 17, 32). Moreover, radiation may enhance antigen cross presentation enabling greater recognition of tumor specific neoantigens and this has been observed in NSCLC patients treated with a combination of SBRT and anti-CTLA-4 inhibition(33). Additionally, radiation may alter immunosuppressive mechanisms in the tumor microenvironment that affect systemic responses(19–22, 33). In the latter cases, the ability to proceed immediately to immunotherapy following ablative radiation is likely a key component of treatment response. In our study, tumor PD-L1 status was not an enrollment criteria and patients had variable levels of PD-L1 expression at time of diagnosis. Notably, in the PEMBRO-RT study, subset analysis suggested radiation may boost the response rate to anti-PD-L1 treatment and thus be most beneficial in PD-L1 low tumors(23). Exploratory analysis including pre and post treatment biopsy and circulating tumor DNA is underway in our patient cohort to investigate changes in immune expression following radiation treatment and dual checkpoint immunotherapy and will be reported in future publications.

The presence of brain metastases is a poor prognostic sign in metastatic lung cancer, with an average overall survival of 8–17 months with aggressive therapy(34). Interestingly, while our numbers are small, we did observe differences in response to therapy for patients with a history of intracranial brain metastases compared to patients without brain disease. Patients with a history of brain metastases had a shorter progression free survival, mostly due to progressive intracranial disease. However, many of these patients were able to be salvaged with local therapy at time of progression followed by continued systemic therapy and the median OS for this cohort has not been met. This was also seen in the two patients excluded from the trial due to new intracranial disease prior to starting immunotherapy. Both patients underwent treatment to the brain followed by off trial immunotherapy, with one surviving 10 months and the other surviving 4.2 years. Together, these findings suggest the difference in oligometastatic and oligoprogressive diseases states. However, it does show the potential for synergistic treatment with aggressive local therapy combined with systemic immunotherapy even in patients with brain metastases to improve long term outcomes.

The current study does have multiple limitations. Overall, a small number of patients were enrolled and the study was not powered for survival outcomes. The trial was closed early due to slow accrual during COVID-19 pandemic, though 17 of 21 patients were enrolled. The patient population included patients with and without prior treatment as well as a mixture of patients with and without prior brain metastases. All patients were oligometastatic, though more than 5 metastatic lesions were allowed. All patients included in toxicity analysis were treated with SBRT and dual ICI, though 7 of the 15 patients did not complete four planned cycles of dual ICI due to toxicity of immunotherapy or progressive disease. All sites of disease were included in radiation treatment fields but patients had a heterogeneous distribution of metastatic burden, including additional lung lesions, osseous lesions, and distant solid organ lesions. Prognosis is likely widely variable across the patient population at baseline given these differences. Moreover, PD-L1 status was not included as an enrollment criteria. Finally, causality of toxicity between SBRT and immunotherapy can be difficult to determine objectively. However, within these limitations, the study did show the safety and tolerability of aggressive SBRT treatment to all sites of oligometastatic disease combined with dual ICI with favorable survival outcomes. Additionally, a recent consensus guideline identified only 37 patients and 62 sites of disease treated with combined SBRT and dual checkpoint immunotherapy(26). Thus, this study will add substantially to this limited knowledge and futher guide combining SBRT with dual checkpoint immunotherapy.

In conclusion, this phase Ib study showed combining ablative radiation therapy to all sites of oligometastatic disease followed by dual checkpoint immunotherapy was safe with similar rates of grade 3 toxicity as observed on dual checkpoint immunotherapy alone trials. Secondary endpoints investigating treatment efficacy showed excellent PFS and OS in this small cohort of patients. Further clinical studies are warranted to evaluate the efficacy of ablative radiation in combination with dual immune checkpoint blocakde for patients with oligometastatic NSCLC.

Supplementary Material

Bassetti et al., IJROBP, 2023 Supplement

Supplemental Figure 1: Treatment schema. Patients with up to 6 sites of extracranial metastatic disease were treated with 5 fraction SBRT followed by dual checkpoint immunotherapy started within 7 days of radiation completion.

Supplemental Table 1: Inclusion and exclusion criteria for trial enrollment.

Supplemental Table 2: Dosimetric planning objectives for target volumes and organs at risk.

Supplemental Table 3: Dose limiting toxicities as defined on trial.

Supplemental Table 4: Frequency of anatomical site treated with SBRT.

Supplemental Table 5: Number of sites treated with SBRT per individual patient.

Supplemental Table 6: Highest SBRT dose prescribed per individual patient

Supplemental Table 7: SBRT dose delivered to each anatomical subsite lesion.

Supplemental Table 8: Cumulative incidence of highest graded toxicity per individual patient seen in follow up period.

Supplemental Table 9: Incidence of Grade 3+ toxicity at least possibly related to SBRT treatment.

Supplemental Table 10: Progression free and overall survival analysis in patients treated per protocol and on intent to treat analysis.

Supplemental Table 11: Progression free and overall survival analysis in patients with or without prior brain metastases treated per protocol and on intent to treat analysis.

[Data Availability Statement for this Work]

• “Research data are stored in an institutional repository and will be shared upon request to the corresponding author.”