Radiotherapy combined with PD‐1/PD‐L1 inhibitors in NSCLC brain metastases treatment: The mechanisms, advances, opportunities, and challenges

Abstract At present, whole‐brain radiation therapy/stereotactic radiosurgery is one of the main local treatments for brain metastasis of non‐small‐cell lung cancer (NSCLC). Currently, it has been proved that radiotherapy (RT) can regulate the immune response, and small‐sample studies have shown that patients with NSCLC brain metastases (BMs) can benefit from RT combined with immunotherapy (IO). However, the efficacy and safety of the combination treatment have not been deeply elaborated. Notably, as a challenge that is still being explored, the timing of RT combined with IO is likely to be an important factor affecting efficacy and prognosis. This article reviews the current application and challenges of RT combined with IO from the perspectives of molecular mechanism, combination timing, safety, and efficacy. The purpose is to provide information on clinical evidence‐based medicine of combination between RT with IO. For further investigation, we also discuss the major challenges and prospects of RT combined with IO in NSCLC BMs.

has been recommended as a first-line treatment for advanced NSCLC patients with high PD-L1 expression. Despite these achievements, only a part of patients benefited from PD-1/PD-L1 inhibitors. Additionally, PD-1/ PD-L1 inhibitors resistance might eventually develop. To improve the prognosis and overcome resistance, some clinical studies have also suggested that PD-1 and PD-L1 inhibitors combined with radiotherapy (RT) may be expected to improve the anti-cancer efficacy. [3][4][5] Preclinical studies have shown that RT can regulate the immune system, including inducing the abscopal effect, increasing the infiltration and activity of tumor-related lymphocytes, and elevating the number of tumor neoantigens. 6 The combination of PD-1/PD-L1 inhibitors with WBRT/SRS should be evaluated in the era of immunotherapy (IO). In this review, we will discuss the mechanisms, advances, opportunities, and challenges of the combination of RT and PD-1/PD-L1 inhibitors for BM in patients with NSCLC.

MECHANISMS OF RT AND PD -1/PD -L1 INHIBITORS IN ANTI-TUMOR IMMUNITY
The role of RT in immunomodulation has gained extensive attention. WBRT/SRS could cause immunogenic cell death, promote the release of tumor-associated antigens, increase the number of cytotoxic T lymphocytes (CTLs) 7,8 to improve the immunocompetence of patients. The immunogenic cell death triggers the extracellular release of high mobility group box-1 protein and transfer of calreticulin to the cell surface, which activates CTLs, dendritic cells, and other immune cells. [9][10][11] Moreover, RT can also increase the permeability of the BBB. In addition, RT activates microglia and induces microglia polarization to M1 type accompany by secreting miR-9 to regulate mesenchymal-to-epithelial transition (MET) of tumor cells through Cadherin 1 and reduce the colonization of BMs. 12 It is observed that M1-type microglia also recruit various immune cells by secreting pro-inflammatory factors, such as natural killer cells and CTLs. 13 Radiation also increases the MHC class I receptors on various antigen-presenting cells and enhances the reactivity of T-cell immune response. 14 Thus, WBRT or SRS provides a more active immune microenvironment for PD-1/PD-L1 inhibitor treatment. 8 It is generally acknowledged that PD-L1 expression was an important prognostic indicator of NSCLC patients. 15 The increased PD-L1 expression induced by radiation could make patients more sensitive to PD-1/PD-L1 inhibitors, which gain a better response rate and extends overall survival (OS). [16][17][18] Current evidence suggests that the combination of WBRT with αPD-1 is involved in a variety of immunomodulatory processes and plays a significant role in anti-tumor immunity. 19 The combination of WBRT or SRS and PD-1/PD-L1 inhibitors reduces myeloid-derived suppressor cells and regulatory T cells (Tregs) to inhibit the immunosuppressive microenvironment 20,21 and induce the expression of vascular cell adhesion molecule-1 and intercellular cell adhesion molecule-1 on the endothelial cells of BBB. 22,23 These adhesion molecules are associated with T-cell recruitment. 24 In conclusion, RT combined with immune checkpoint inhibitors (ICIs) plays a collaborative role in the treatment of BMs by affecting the tumor microenvironment and improving immune function within the brain ( Figure 1). Summarily, based on the immunomodulatory effect and abscopal response of RT, the combination of RT with IO appears to be one of the most promising treatment protocols for NSCLC.

EFFECTS OF RT AND PD -1/PD -L1 INHIBITORS FOR NSCLC BMS TREATMENT
RT and IO have synergistic effects in clinical (Table 1). In a comparative analysis of survival with intracranial RT plus IO versus no IO in 13,998 patients with BMs from the US National Cancer Database, the median OS with ICIs was 13.1 months (13.1 vs. 9.7 months) and the 3-year OS rate was 17% (17% vs. 12%), higher than that of non-recipients. Paired multivariate comparisons showed that ICIs are an independent predictor of increased OS in NSCLC patients with intracranial metastases. 25 In addition, the singlecenter secondary analysis of the KEYNOTE-001 showed that median progression-free survival (PFS) was longer in patients who received RT (cerebral or extracranial) in combination with pembrolizumab than in patients who received pembrolizumab alone. 26 Furthermore, a retrospective analysis of SRS combined with nivolumab or durvalumab revealed that the Kaplan-Meier rates of OS at 6 and 12 months were 48/41% from the date of SRS, and 81/51% from the date of cranial metastases diagnosis, respectively. 27 Similarly, Chen et al. reported that the median survival time was 12.9 months (RT alone), 14.5 months (non-concurrent), and 24.7 months (concurrent) from 260 patients (157 NSCLC) receiving SRS with/without ICI. 28 Additionally, a retrospective multicenter study showed that prolonged median survival was 192 days for patients with NSCLC BMs who were treated by WBRT/SRS combined with PD-1 inhibitors, compared with historical controls. 29 In summary, these results suggested that NSCLC patients benefited from the combination of RT and IO, ongoing trials will further help us obtain a comprehensive understanding of the combination treatment, thus guiding clinical practice to achieve superior survival benefits.

RT COMBINED WITH ICIS
Currently, the optimal timing for combining RT and PD-1/ PD-L1 inhibitors remains to be elucidated and might have a significant effect on the generation of a potent and durable anti-tumor effect. The conclusions of the researches conducted by various research cohorts are different ( Table 2). Li et al. demonstrated that the median time of intracranial PFS was 9.7 months, and the intracranial PFS rate was 75% at 4 months by concurrent treatment of SRS combined with nivolumab/ipilimumab. 30 Moreover, better OS (1-year OS: 70.2%) and Regional control (1-year local progression-free interval: 78.9%) were demonstrated in the "concurrent SRT and IO" group. 31 Nonetheless, some studies have been performed to evaluate concurrent or sequential (nivolumab or pembrolizumab) and RT in patients with stage IV NSCLC. PFS and median OS for patients who received RT prior to, concurrently, or post-IO were 7.8/17.5, 9.2/23.4, and 5.7/14.4 months, respectively. The data showed that PFS and OS in the concurrent arm were better than the sequential arm. 32 Similarly, a retrospective study conducted by Chen et al. (2015) evaluate the efficacy and safety of ipilimumab/nivolumab plus SRS for patients with intracranial diseases. The results showed that concurrent SRS with ipilimumab/nivolumab could reduce the incidence of new intracranial diseases and tumor burden. 33 Another retrospective study performed by Srivastava and Huang showed that concurrent RT with ICIs significantly improved the 6-month local control rate (76% vs. 100%) and 6-month brain control rate (41% vs. 71%) in patients with NSCLC BMs, compared to the non-concurrent group. 34 Related to this, a preclinical study showed that compared to receiving sequential ICIs and RT, concurrent RT and ICIs had a better anti-tumor effect. 35 For the optimal timing of the combination, there is no certain consensus on whether PD-1/PD-L1 inhibitors should be administrated concurrent, before or after the initiation of RT. Imber et al. observed the median time of distant brain failure (defined as the presence of a new brain metastatic tumor) in patients with NSCLC BMs who received PD-(L)1 based IO sequentially or concurrently combined with SRS/hypofractionated radiation therapy were 4.9 and 3.9 months, respectively. 36 Additionally, Shepard et al. carried out a retrospective paired cohort study regarding combined ICIs (nivolumab, pembrolizumab, or atezolizumab) with SRS, which proved that compared to the ICI-naive group, the BMs shrinkage time was shorter without increasing the incidence of peritumoral edema or radiation necrosis in the concurrent group. However, the comparison of OS and PFS between the two groups was not statistically different. 37 Moreover, a survey of 462 European doctors showed that 49.4% thought ICIs should not be used during WBRT, while 44.6% thought ICIs should not be used during SRS. 38 The RTOG 3505 study showed that palliative RT required at least a 14-day window period F I G U R E 1 The theoretical basis of radiotherapy combined with immunotherapy. (A) PD-1 inhibitors can reverse the immunosuppressive microenvironment induced by radiotherapy. (B) Radiotherapy combined with immunotherapy sequentially or concurrently. ɑ-PD-1, ɑ-programmed death-1; BBB, blood-brain barrier; CDH1, cadherin 1; CRT, calreticulin; CTL, cytotoxic T lymphocyte; DC, dendritic cell; HMGB1, high mobility group box-1 protein; ICAM-1, intercellular cell adhesion molecule-1; MDSCs, myeloid-derived suppressor cells; MET, mesenchymal-epithelial transition; miR9, microRNA 9; NK, natural killer cell; PD-L1, programmed death-ligand 1; RT, radiotherapy; SBS, stereotactic radiosurgery; TAAs, tumor-associated antigens; Tregs, regulatory T cells; VCAM-1, vascular cell adhesion molecule-1; WBRT, whole-brain radiotherapy.

T A B L E 3 Safety and tolerance of RT combined with ICIs
Author (  before using nivolumab. 39 Although the addition of RT may enhance the immune response, the combination of IO with RT requires more effort to optimize the dose and schedule. 40 Based on the current data, RT concomitant with PD-(L)1-based IO has shown considerable survival benefits, but a few researchers have argued otherwise. Therefore, prospective head-to-head clinical trials are warranted to provide more evidence and further guide clinical research and practice.

RT COMBINED WITH ICIS
Most small samples and retrospective clinical studies have shown that RT combined with ICIs was safe ( Table 3). All toxic side reactions were below Grade 4 during treatment of SRS combined nivolumab, atezolizumab, or pembrolizumab on NSCLC BMs. 41 A retrospective study conducted by Hubbeling et al. showed that the rate of adverse events (AEs) at all grades in the ICIs treatment group and the ICIs-naive treatment group was comparable to any type of in patients with NSCLC BMs. The most frequently observed in the Grade 3 (G3) or worse AEs of ICIs-treated queue were headaches, anorexia, and cognitive impairment. 4 Li et al. reported that except for only one patient with a dose limit-related toxic reaction, three patients (25%) developed treatment-related G3 events, and one patient developed G4 events in the Phase I/II single-center trial about concurrent SRS combined nivolumab/ipilimumab. 30 The most common AE we concerned about is treatment-related neurotoxicity. Although there was a strong correlation between symptomatic radioactive necrosis (RN) and ICIs in patients with NSCLC treated with SRS-SRT and newly diagnosed BMs, 42 the AEs are less obvious. Many studies have shown that RT combined with ICIs does not increase the risk of neurotoxicity, 28,43 and even this combination rarely presents neurotoxicity 27,44 that requires intervention with drugs. The timing of RT combined with ICIs may not be a factor for the increase in toxicity. 45 However, the study had not yet reported the results of the evaluation of cognitive function. 44 Imber et al. (2017) found no acute Grade 3-4 AEs or new intracranial bleeding caused by brain RT in neither concurrent group nor sequential group. This study showed that IO combined with high-dose brain RT was safe as well. 36,40 A multi-center retrospective study showed no correlation between an increased risk of severe or accidental toxic side effects in a combination of and nivolumab treatment for NSCLC patients with metastasis. 46 Kotecha et al. found that the timing of RT combined with ICI did not affect the risk of toxicity. In patients treated with chemo-radiotherapy (CRT) and immediate

months
Abbreviations: CRT, chemoradiotherapy; hRT, hypofractionated radiotherapy; ICIs, immune checkpoint inhibitors; IO, immunotherapy; NSCLC, non-small-cell lung cancer; PBI, partial brain irradiation; PD-1, Programmed death-1; PD-L1, programmed death-ligand 1; q2w, 2 weeks using a; q4w, 4 weeks using a; RN, radionecrosis; RT, radiotherapy; SRS, stereotactic radiosurgery; SRT, stereotactic radiotherapy; WBRT, wholebrain radiation therapy. (±1 half-life) ICIs, none required surgery even though some of them had symptomatic RN. 47 In conclusion, the occurrence of AEs would not increase between the different treatment sequences of ICIs and CRT. Longer and rigorous follow-up data need to be obtained for a comprehensive understanding of the toxicities from the combination treatment.

TRUE PROGRESSION DURING RT COMBINED WITH ICIS
For patients receiving ICIs and RT, the pseudoprogression (Psp) of BMs should be taken into consideration. Psp meant that the tumor lesions initially expanded and then the tumor improves spontaneously. It is a common phenomenon that may be necrosis or inflammatory cell infiltration after IO. Biopsy, imaging follow-up, and evaluation of patient's functional status are standard procedures to distinguish between pseudo and true progression. 48  Furthermore, lots of clinicians support this combination treatment in oligometastases or advanced stage in NSCLC BMs, but the evidence-based researches and clinical data about this combination are scarce. In this review, we concluded that the mechanism of RT combined with ICIs mainly focused on affecting immunogenic cell death, 9 MET, 12  Abbreviations: BMs, brain metastases; FSRT, fractionated stereotactic radiotherapy; HFSRT, hypofractionated stereotactic radiotherapy; ICIs, immune checkpoint inhibitors; IO, immunotherapy; NSCLC, non-smallcell lung cancer; RS, radiosurgery; RT, radiotherapy; SBRT, stereotactic body radiotherapy; SRS, stereotactic radiosurgery; SRT, stereotactic radiotherapy; WBRT, whole-brain radiation therapy.

T A B L E 4 (Continued)
Nowadays, the main combination regimen in NSCLC BMs is the combination of brain RT and ICIs. In terms of the data in our review, concurrent RT combined with ICIs has a relatively high control rate of intracranial progression 33 and a better prognosis. 32,37 However, some studies support the practice of RT before IO. 39,40 Both concurrent and sequential RT combined with IO have different efficacy, but most studies have advocated that RT concurrently combined with PD-(L)1-based IO has a better survival benefit in NSCLC BMs, compared with the sequential mode. Interestingly, according to the researches included in our paper, the mode of RT and timing of RT combined with IO may not be factors for adverse reactions. Due to the short follow-up period, there is a limitation of not observing the time of late adverse reactions. Therefore, to determine a more beneficial combined RT regimen and the timing of the combination of IO and RT, more head-to-head, prospective, and multi-center clinical researches are needed to be carried out. At the same time, future studies need to extend the follow-up time to avoid missing late AEs.
In the exploration of the anti-tumor efficacy of RT combined with ICIs, it is inevitable to encounter the situation of Psp. Currently, the reliability of identifying progression is usually increased by biopsy, clinical status, imaging follow-up, and other means. 48 However, there is no clear standard for the diagnosis of Psp in clinical practice. Therefore, it is expected that relatively accurate and non-invasive technologies can be developed or relevant statistics of large data researches can help clinical diagnosis of the true progression of tumors.
There are some limitations in this review. The first part of the results of RT combined with IO is the OS data of BMs in a variety of tumors, although including NSCLC, these studies did not independently explore the survival data of BMs in NSCLC. Secondly, some studies included WBRT and SRS as the same intervention to collect the outcome data, so that they could not distinguish the differences and advantages of WBRT and SRS combined with IO, respectively. In addition, pretreatment in some studies is not described, which may ignore the effect of pretreatment on results. Finally, some of these studies are only in abstract form, which limits understanding of the specific design scheme of these studies, and descriptions of the results are also in abstract form. However, the overall direction is supportive of RT combined with IO for BMs in NSCLC.
Most of the current clinical evidence on the safety and importance of ICIs combined with RT to NSCLC BMs is retrospective. For patients with NSCLC BMs, the efficacy of other ICIs remain unknown, whose targets are T-cell immunoglobulin-3 (Tim-3), OX40 (CD134), 4-1BB (CD137), pyridine 2, 3-dioxygenase-1, killer cell immunoglobulin-like receptors. To date, clinical studies on RT combined with PD-1/L1 inhibitors for NSCLC BMs are ongoing (Table 4), such as NCT04889066, NCT04787185, NCT04650490, etc. Therefore, further translational medicine research and large-scale prospective studies are necessary to better reveal the interactions and combinations of RT and IO to optimize treatment options for the patients with NSCLC BMs.

AUTHOR CONTRIBUTIONS
Xiao-Xia Zhu and Xiao-Tong Duan carried out conceptualization. Zi-Ying Chen and Xiao-Tong Duan carried out article writing and editing. Xiao-Xia Zhu and Si-Miao Qiao carried out supervision. All authors read and approved the final manuscript. Zi-Ying Chen and Xiao-Tong Duan contributed equally to the manuscript.