ORIGINAL RESEARCH published: 07 September 2020 doi: 10.3389/fonc.2020.01677 Immune-Related Adverse Events and Corticosteroid Use for Cancer-Related Symptoms Are Associated With Efficacy in Patients With Non-Small Cell Lung Cancer Receiving Anti-PD-(L)1 Blockade Agents Mariona Riudavets 1,2 , Joaquin Mosquera 3 , Rosario Garcia-Campelo 3 , Jorgina Serra 1 , Georgia Anguera 1 , Pablo Gallardo 1 , Ivana Sullivan 1 , Andrés Barba 1 , Luís del Carpio 1 , Agustí Barnadas 1 , Ignasi Gich 4,5,6 and Margarita Majem 1* Edited by: 1 Department of Medical Oncology, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain, 2 Department of Medicine, Lizza E. L. Hendriks, Universitat Autónoma de Barcelona (UAB), Barcelona, Spain, 3 Department of Medical Oncology, Hospital Universitario a Maastricht University Medical Coruña, a Coruña, Spain, 4 Department of Clinical Epidemiology and Public Health, Hospital de la Santa Creu i Sant Pau, Centre, Netherlands Barcelona, Spain, 5 CIBER Epidemiología y Salud Pública (CIBERESP), Madrid, Spain, 6 Sant Pau Biomedical Research Reviewed by: Institute (IIB Sant Pau), Barcelona, Spain Yuhchyau Chen, University of Rochester, United States Natasha Leighl, Background: Immune-related adverse events (irAEs) have been associated with University Health Network, Canada improved efficacy in advanced non-small cell lung cancer (NSCLC) patients receiving *Correspondence: anti-PD-(L)1 blockade agents, while the concurrent use of corticosteroids seems to Margarita Majem mmajem@santpau.cat worsen it. We evaluated outcomes in advanced NSCLC patients treated with anti-PD-(L)1 blockade agents in relation to the presence of irAEs and the reasons for using Specialty section: corticosteroids: whether for palliative cancer-related reasons or for the management This article was submitted to Thoracic Oncology, of irAEs. a section of the journal Methods: Clinical outcomes in advanced NSCLC patients treated with anti-PD-(L)1 Frontiers in Oncology blockade agents were calculated with regard to the presence of irAEs and the use of Received: 16 December 2019 Accepted: 29 July 2020 corticosteroids. A landmark analysis was performed to avoid immortal time bias due to Published: 07 September 2020 the time-dependent nature of irAEs. Citation: Results: Out of a total of 267 patients, the 56.9% of patients who experienced Riudavets M, Mosquera J, Garcia-Campelo R, Serra J, irAEs had significantly improved outcomes. In the landmark analysis, median Anguera G, Gallardo P, Sullivan I, progression-free survival (PFS) was 12.4 months for patients with irAEs vs. 4.1 Barba A, del Carpio L, Barnadas A, Gich I and Majem M (2020) months for patients without irAEs (p < 0.001), while median overall survival (OS) Immune-Related Adverse Events and was 28.2 vs. 12.5 months, respectively (p < 0.001). Likewise, objective response Corticosteroid Use for Cancer-Related and disease control rates were significantly higher in patients experiencing irAEs: Symptoms Are Associated With Efficacy in Patients With Non-Small 48.6 vs. 22.8% and 77.1 vs. 39.6% (p < 0.001), respectively. Median OS was Cell Lung Cancer Receiving significantly shorter for patients receiving ≥10 mg of prednisone equivalent daily Anti-PD-(L)1 Blockade Agents. Front. Oncol. 10:1677. for cancer-related symptoms than for the rest of patients (<10 mg prednisone doi: 10.3389/fonc.2020.01677 equivalent daily or for management of irAEs): 6 vs. 15.9 months (p < 0.001). Frontiers in Oncology | www.frontiersin.org 1 September 2020 | Volume 10 | Article 1677 Riudavets et al. irAEs, Corticosteroids and Immunotherapy in NSCLC Conclusions: IrAEs were associated with improved efficacy in advanced NSCLC patients when a landmark analysis was applied. Patients receiving corticosteroids had significantly poorer outcomes when they were used for cancer-related symptoms. Keywords: immune-related adverse events, immunotherapy, advanced NSCLC, corticosteroids, efficacy INTRODUCTION were reviewed. All patients starting an anti-PD-(L)1 blockade agent alone or in combination with chemotherapy or an anti- Immunotherapy has become established as a new standard- CTLA-4 (cytotoxic T-lymphocyte antigen 4) in any treatment line of-care for multiple solid malignancies, including non-small were included. cell lung cancer (NSCLC). Among other strategies, immune- The end of follow-up was December 31, 2018. The study was checkpoint blockade agents targeting the inhibitory pathways of approved by the local institutional review board. the immune cascade have resulted in an increase in the response Patients were evaluated for objective response rate (ORR), against tumor cells (1, 2). disease control rate (DCR), duration of response (DoR), Examples of these agents are nivolumab and pembrolizumab, progression-free survival (PFS), and overall survival (OS). Tumor monoclonal antibodies against the programmed cell death-1 responses were assessed as per clinical practice using the receptor (PD-1) and atezolizumab, durvalumab, and avelumab, Response Evaluation Criteria for Solid Tumors (RECIST) version against its ligand PD-L1. Most of these agents have been approved 1.1 every 8–12 weeks (27). in different settings for the treatment of patients with NSCLC IrAEs were defined as adverse events with a potential (3–8), which has led the scientific community to advance by immunologic basis that required close monitoring and/or exploring combinations of anti-PD-(L)1 blockade agents with potential intervention with immunosuppressives or hormone chemotherapy and/or other immune-checkpoint blockade to replacement. Patient symptoms and physical exploration and achieve better results (9–11). laboratory data were assessed at every cycle. Thyroid function was Due to their mechanism of action, anti-PD-(L)1 blockade evaluated at baseline and every 6 weeks thereafter. irAEs severity agents can induce inflammatory side effects known as immune- was graded according to the Common Terminology Criteria for related adverse events (irAEs), which are not triggered by Adverse Events (CTCAE) version 4.0, as part of routine clinical conventional cytotoxic anticancer agents. The most commonly practice (28). reported irAEs are those affecting the skin, the gastrointestinal Patient data were obtained from a unified database in tract and the thyroid gland, though any organ or system may be which clinical and pathological characteristics and toxicity were involved, including the lung, liver, and the hypophysis (12) IrAEs accurately recorded. are generally mild, but ∼10% of cases are severe and may require ORR data included patients with partial or complete response, immunosuppressors and/or treatment discontinuation (13, 14). and DCR data included partial response, complete response, and However, immunotherapy raises several questions that remain stable disease. PFS and OS were measured as the time from unknown. First, a positive correlation between the presence of the start of anti-PD-(L)1 blockade agent to documented disease irAEs and the efficacy of immunotherapy has been postulated, progression or death owing to any cause (PFS) or to death (OS). suggesting that a proper management of such events might be Patients with no events were censored on the date of the last required to maximize the therapeutic effect of these drugs and to follow-up. Those patients who were not evaluable for response avoid treatment interruption (15–23). Second, the activity of anti- were not included in the ORR assessment but were included in PD-(L)1 blockade agents in patients receiving corticosteroids the PFS and OS evaluations. or antibiotics during immunotherapy is controversial, and their Corticosteroid usage within 1 month before the initiation use and safety in certain groups, such as patients with brain or during anti-PD-(L)1 therapy, and the administration of metastasis, needs to be defined (24–26). antibiotics from 3 months before the start of anti-PD-(L)1 We performed a retrospective study to investigate irAEs therapy to 3 months after finishing were also recorded. The profiles and their association with clinical activity in patients reason for corticosteroid use was specified, distinguishing with advanced NSCLC treated with anti-PD-(L)1 blockade agents between management of irAEs or the palliative treatment using landmark and multivariable analyses. Additionally, we of cancer-related symptoms. Types of corticosteroid and the evaluated the efficacy of immunotherapy with regard to the use prednisone equivalent daily dose, as well as type of antibiotics, of corticosteroids and antibiotics. were also collected. Transient corticosteroids given along with a chemotherapy combination were not registered. MATERIALS AND METHODS To analyze efficacy according to the dose of corticosteroids patient were divided into two groups: those receiving prednisone The medical records of all patients with advanced NSCLC equivalent ≥10 mg daily and those receiving prednisone treated with anti-PD-(L)1 blockade agents at two tertiary equivalent ≤10 mg daily (including patients that did not institutions in Spain between March 2013 and August 2018 receive corticosteroids). Frontiers in Oncology | www.frontiersin.org 2 September 2020 | Volume 10 | Article 1677 Riudavets et al. irAEs, Corticosteroids and Immunotherapy in NSCLC Statistical Analysis line), and second, to assess the influence of different types of Survival curves were estimated using the Kaplan-Meier method irAEs on OS. and compared with the log-rank test. Taking into account the To describe our population, numbers and percentages were immortal time bias due to the time-dependent nature of irAEs, used for qualitative variables, while medians and interquartile we performed tests at 2.4-months for PFS and 5.9-months ranges (IQR) were calculated for ordinal and quantitative for OS landmark analyses including only patients manifesting variables with an asymmetric distribution. disease control or those who were alive at these time points. All p-values were based on a two-sided hypothesis, and those Consequently, 92 patients were excluded from the PFS landmark <0.05 were considered statistically significant. Statistical analyses analysis (n = 175) and 100 patients were excluded from the OS were carried out using IBM SPSS Statistics for Windows, version analysis (n = 167). Landmark times were pre-defined before the 25.0 (IBM SPSS Inc., Chicago, IL, USA). start of data analysis and were determined from the median PFS and OS data of patients with no irAEs (29, 30). In addition, RESULTS 6 and 12-month landmark analyses were also performed as complementary evaluations. Patient Characteristics Odds ratios were used for ORR and DCR. Univariable and We included 267 consecutive patients with advanced multivariable Cox proportional hazard regression models were NSCLC treated with anti-PD-(L)1 blockade agents. Baseline adopted to determine hazard ratios (HR). characteristics of patients are described in Table 1. The median Two multivariable analyses were performed. First, to age was 66.1 years (range 26.7–85.2, IQR 14.3), 69.7% were determine the influence of clinical characteristics (such male and 30.3% female. The majority of patients were current as age, smoking status, Eastern Cooperative Oncology or former smokers (74.2%) and baseline ECOG PS was 0–1 in Group [ECOG] performance status [PS], brain, and liver 85% of patients. Non-squamous was the most common histology metastases, presence of irAEs, toxicity grade, use of prednisone (67.8%). Brain metastases were present in 15.7% of patients and equivalent ≥10 mg daily, use of antibiotics and treatment liver metastases in 15.4%. PD-L1 expression analysis in tumor TABLE 1 | Patients characteristics and comparison and the presence of irAEs. Category Total irAEs No irAEs P-value n = 267 (%) n = 152 (%) n = 115 (%) Gender Male 186 (69.7) 108 (71.1) 78 (67.8) 0.593 Female 81 (30.3) 44 (28.9) 37 (32.2) Age 66.1 years 66.4 years 65.8 years 0.362 median (range) (26.7–85.2, IQR 14.3) (26.7–85.2, IQR 15.3) (38.8–81.0, IQR 13.7) Smoking status Non- or light smoker 26 (9.7) 13 (8.6) 13 (11.3) 0.533 Current or former smoker 241 (90.3) 139 (91.4) 102 (88.7) ECOG PS 0–1 227 (85) 136 (89.5) 91 (79.1) 0.024 2 40 (15) 16 (10.5) 24 (20.9) Histological subtype Squamous 86 (32.2) 49 (32.2) 37 (32.2) 1.000 Non-squamous 181(67.8) 103 (67.8) 78 (67.8) Treatment line 1st line 81 (30.3) 60 (39.5) 21 (18.3) <0.001 ≥2nd line 186 (69.7) 92 (60.5) 94 (81.7) Immune-checkpoint blockade schedule Monotheraphy 209 (78.3) 112 (73.7) 97 (84.4) 0.082 Combination with chemotherapy 33 (12.3) 24 (15.8) 9 (7.8) Combination with anti-CTLA-4 25 (9.4) 16 (10.5) 9 (7.8) Treatment duration 2.75 m 4.8 m 1.8 m <0.001 median (range) (0.03–56.4, IQR 6.2) (0.03–56.4, IQR 9.8) (0.03–46.2, IQR 2.7) Prednisone equivalent ≥10 mg/day use No 133 (49.8) 66 (43.4) 67 (58.3) 0.019 Yes 134 (50.2) 86 (56.6) 48 (41.7) Antibiotics use No 126 (47.2) 63 (41.4) 63 (54.8) 0.036 Yes 141 (52.8) 89 (58.6) 52 (45.2) irAEs, immune-related adverse events; IQR, interquartile range; ECOG PS, Eastern Cooperative Oncology Group performance status; CTLA-4, cytotoxic T-lymphocyte antigen 4. Frontiers in Oncology | www.frontiersin.org 3 September 2020 | Volume 10 | Article 1677 Riudavets et al. irAEs, Corticosteroids and Immunotherapy in NSCLC samples was available from 135 patients (50.6%), and expression were the irAEs most frequently associated with treatment was low (PD-L1 expression 1–49%) in 52 (38.5%), high (PD-L1 discontinuation. Thirty of the 44 patients who stopped anti-PD- ≥ 50%) in 41 (30.4%), and negative (PD-L1 < 1%) in 42 (31.1%). (L)1 therapy due to toxicity presented irAEs grade ≥3 (68.2%). By the time of data analysis, 83.9% of irAEs had been resolved. Anti-PD-(L)1 Blockade Treatment and irAEs Characteristics Association Between irAEs and Treatment Eighty-one patients (30.3%) received anti-PD-(L)1 blockade Outcomes agents as first line treatment, 131 (49.1%) as second line At the time of data analysis, the median follow-up time was 8.5 treatment and 55 (20.6%) as third line treatment or beyond. months (range 0.3–56.4, IQR 10.6) and the median duration of Anti-PD-(L)1 blockade agents were given alone (78.3%) or in treatment with anti-PD-(L)1 blockade agents was 2.8 months combination with chemotherapy (12.3%) or with an anti-CTLA- (range 0.1–56.4, IQR 6.2). The median OS and PFS of the study 4 agent (9.4%). Nivolumab (44.2%) and pembrolizumab (25.6%) population were 12.4 months (95% confidence interval [CI], were the most commonly used types of anti-PD-(L)1 blockade 10.1–14.7) and 4.2 months (95% CI, 3.1–5.3), respectively. In agents, followed by atezolizumab (17.2%). first line setting, the median OS was 19.4 months (95% CI, 11.9– One hundred and fifty-two patients (56.9%) experienced a 27.0) and the median PFS was 9.8 months (95% CI, 5.4–14.2). As total of 255 irAEs. The median number of irAEs per patient was expected, patients receiving anti-PD-(L)1 therapy as second line one (range 0–5, IQR 1), and 64 patients (24%) experienced two or treatment or beyond had significantly poorer outcomes, with a more irAEs. The most common irAEs was skin toxicity (35.6%), median OS of 9.0 months (95% CI, 7.0–11.1) (HR 1.89; 95% CI, followed by diarrhea (16.5%) and hypothyroidism (10.2%). 1.29–2.79; p = 0.001) and a median PFS of 3.3 months (95% CI, According to the CTCAE terminology, 149 irAEs (58.4%) were 2.6–4.1) (HR 1.78; 95% CI, 1.27–2.51; p = 0.001). No differences grade 1, 65 (25.5%) were grade 2, 33 (12.9%) were grade 3 and were found between patients receiving anti-PD-(L)1 blockade three (1.2%) were grade 4. There were five treatment-related agents in monotherapy or combination with anti-CTLA-4 or deaths (2%): four due to pneumonitis and one due to hepatitis. chemotherapy in first line setting, neither in terms of OS (p = IrAEs were more frequent in patients receiving immunotherapy 0.177) nor in PFS (p = 0.343). in first line treatment (74.1%) than in second line treatment or The landmark analysis showed that PFS was significantly beyond (49.5%) (p < 0.001). No differences were observed in the longer in patients experiencing irAEs than in those without presence of grade ≥3 irAEs according to the treatment line (p = irAEs: 12.4 months (95% CI, 1.9–22.9) vs. 4.1 months (95% CI, 0.342). A trend to a higher rate of grade ≥3 irAEs was also seen 2.6–5.6), (HR 0.43; 95% CI, 0.28–0.64; p < 0.001). Similarly, OS in patients receiving immune blockade combination regimens among patients with irAEs was significantly higher: 28.2 months (50%) in contrast to single-agent anti-PD-1 and anti-PD-L1 (21.6 (95% CI, not achieved) vs. 12.5 months (95% CI, 10.8–14.2) (HR and 20%, respectively) (p = 0.058). 0.38; 95% CI, 0.24–0.59; p < 0.001) (Figures 1A,B). Endocrine toxicity was significantly higher with a Six and 12 months landmark analyses were also performed combination of immune blockade agents (36%) than with to provide complementary information. The median OS at both single-agent anti-PD-1 (8.6%) or anti-PD-L1 (8.9%) (p = time-points also favored patients with irAEs: 12.9 months (95% 0.034). A trend to a higher rate of pneumonitis was seen with CI, 11.3–14.5) vs. 28.2 months (95% CI not calculated) (HR 0.39; combinations of immune blockade agents (20% vs. 6.5 and 95% CI, 0.25–0.61; p < 0.001), and 19.6 months (95% CI, 15.2– 10.7%) (p = 0.101), and a greater number of cases of arthritis 23.9) vs. not reached (HR 0.33; 95% CI, 0.17–0.64; p = 0.001), was observed with anti-PD-1 blockade agents (11.3%) than respectively. Landmark analyses for PFS at 6 and 12 months could with anti-PD-L1 (3.6%) or than with a combination of immune not be calculated since no event happened after these time-points blockade agents (4%) (p = 0.099). Global median time to irAEs in the no-irAEs group. onset was 7.6 weeks (0.1–123.4, IQR 13.3). A description of irAEs Of note, the ORR was significantly higher in patients who and median onset time are detailed in Table 2. experienced irAEs than in those without irAEs: 48.6 vs. 22.8% The median duration of treatment with anti-PD-(L)1 (odds ratio [OR] 0.31; 95% CI, 0.18–0.55; p < 0.001). DCR was blockade agents was significantly longer in patients who also significantly better when irAEs were present: 77.1 vs. 39.6% experienced irAEs than in those who did not: 4.8 months (range (OR 0.20; 95% CI, 0.11–0.34; p < 0.001). 0.03–56.4, IQR 9.8) vs. 1.8 months (range 0.03–46.2, IQR 2.7) (p The landmark analysis was also applied when comparing ORR < 0.001). Comparisons between patients regarding the presence regarding the development of irAEs. Landmark analysis at 8 and of irAEs can be found in Table 1. 10 weeks showed that ORR was significantly higher in the irAEs Two hundred and eighteen patients (82%) discontinued group of patients. At 8 weeks, 54.8 vs. 28% (OR 0.32; 95% CI, treatment with anti-PD-(L)1 blockade agents. The most common 0.15–0.68; p = 0.004) and at 10 weeks, 62.3 vs. 36.7% (OR 0.35; reason given was progressive disease in 145 patients (66.5%), 95% CI, 0.14–0.85; p = 0.028). However, no differences were followed by the presence of irAEs in 44 patients (20.2%). detected when greater time-point were used, probably because Twenty-nine patients (13.3%) stopped treatment due to other the progressive decline in number of patients who show first causes, such as deterioration in their general condition or response later in time (Table S1). complications unrelated to disease progression. Pneumonitis Though median time to response was slightly shorter in the (34.1%), endocrine dysfunction (29.5%), and diarrhea (22.7%) no-irAEs group [8 weeks (1.3–122.6, IQR 5.3) vs. 9.8 weeks Frontiers in Oncology | www.frontiersin.org 4 September 2020 | Volume 10 | Article 1677 Riudavets et al. irAEs, Corticosteroids and Immunotherapy in NSCLC TABLE 2 | Description of immune-related adverse events. Types of irAEs All patients, n = 267 (%) Median onset time (range), weeks All grades Grade 3–5 irAEs requiring n = 255a n = 41 prednisone (95.5) (15.3) equivalent ≥10 mg/dd n = 63 (23.6) Cutaneous Rash 45 (17) 3 (1.1) 5 (2) 10.8 (0.3–145) Pruritus 46 (17.2) 0 3 (1.1) Diarrhea 42 (15.7) 6 (2.2) 8 (3) 8.9 (0.1–89.7) Endocrine dysfunction Hypothyroidism 26 (9.7)e 1 (0.4) 1 (0.4) 16 (1.7–106) Hyperthyroidism 6 (2.2) 0 2 (0.7) Adrenal insufficiency 4 (1.5) 2 (0.7) 3 (1.1) Pneumonitis 23 (8.6) 12b (4.5) 19 (7.1) 16.7 (0.9–189.4) Hepatitis 14 (5.3) 4c (1.5) 7 (2.6) 5.7 (0.4–33) Mucositis 2 (0.7) 1 (0.4) 0 2.7 (1.6–3.9) Arthritis 24 (9) 0 4 (1.5) 11 (0.3–123.4) Others 23 (8.6) 12 (4.5) 11 (4.1) Not calculated Hemolytic anemia 1 1 1 Thrombocytopenia 2 2 1 Flu-like 4 0 0 Nephritis 4 1 2 Vitiligo 1 0 0 Pancreatitis 1 1 1 Myopericarditis 2 2 1 Myositis 3 1 1 Vasculitis 1 0 1 Aseptic meningitis 1 1 1 Encephalitis 2 2 2 Miasteniforme syndrome 1 1 0 irAEs, immune-related adverse events. a Total number of irAEs. b,c Four cases of pneumonitis and one case of hepatitis were grade 5. d High-dose steroid pulse therapy (metilprednisolone at 1g/day) for 3 days followed by metilprednisolone (1 to 2 mg/kg) treatment for several weeks was administered in one case of grade 3 colitis. No patient received other types of immunosupressors. e Eleven patients required thyroid hormone replacement therapy. (1.9–117.4, IQR 11.3) (p = 0.004)], DoR was significantly longer One hundred and forty-six patients (54.7%) received in patients with irAEs: 6.1 months (range 0.5–50, IQR 10.6) vs. corticosteroids during therapy with anti-PD-(L)1 blockade 2.6 months (range 0.2–51.9, IQR 3.8) (p < 0.001). As mentioned agents, of whom 134 patients (91.8%) required ≥10 mg of previously, 44 patients (22.2%) discontinued treatment due to prednisone equivalent per day: 59 patients (44%) for the irAEs. Within this group, 29 patients (65.9%) did not progress treatment of irAEs, and 75 patients (56%) for the management after stopping immunotherapy, in contrast to the 28.7% (64/223) of cancer-related symptoms, including asthenia (8.2%) and of patients in the group of patients who did not discontinue anorexia (6.3%), symptomatic bone metastases (13.4%), treatment due to toxicity (p < 0.001). symptomatic brain metastases (36.3%), dyspnea (14.8%), and chronic obstructive pulmonary disease (COPD) management (21%). No other chronic illness required steroid therapy in our Association Between the Use of study population. Only seven patients started corticosteroids Corticosteroids and Efficacy within the 30 days before immunotherapy initiation, and all The most commonly used types of corticosteroids were patients continued corticosteroids therapy during anti-PD- prednisone (39.7%) and dexamethasone (34.9%). The median (L)1 therapy (Figure S1). Patients receiving corticosteroids dose of prednisone equivalent was 50 mg daily (range 5– for cancer-related symptoms presented significant differences 1,250 mg, IQR 53.4). The median duration of corticosteroid compared to the rest of the population: there was a higher treatment was 59 days (range 0.5–83.0, IQR 159). No differences proportion of patients with ECOG PS 2 (24.1 vs. 10.9%; p = in corticosteroid usage were observed in patients receiving 0.009) receiving second line therapy or beyond (83.1 vs. 63.6%; p first line therapy (53.1%) vs. second line or beyond (55.4%) = 0.001) or as a single-agent instead of a combination regimen (p = 0.790). (90.4 vs. 72.8%; p 0.002). No differences were observed regarding Frontiers in Oncology | www.frontiersin.org 5 September 2020 | Volume 10 | Article 1677 Riudavets et al. irAEs, Corticosteroids and Immunotherapy in NSCLC CI, 4.4–7.5) vs. 15.9 months (95% CI, 11.2–20.7) (HR 2.28; 95% CI, 1.63–3.20; p < 0.001). No differences in terms of PFS were observed. No differences in OS were found between patients who started corticosteroids for cancer-related symptoms in the 30 days before starting immunotherapy or after starting anti-PD-(L)1 blockade: 5.2 months (95% CI, 0.3–4.6) vs. 6.4 months (95% CI, 1.1–4.3) (p = 0.898). It is important to highlight that no significant differences were observed in OS between patients who received no corticosteroids or <10 mg prednisone equivalent daily and those who received ≥10 mg for the management of irAEs (p = 0.314) (Figures 2B,C). However, when analyzing the outcomes of patients with irAEs, median OS was greater in patients receiving <10 mg of prednisone equivalent daily than those receiving ≥10 mg for toxicity management: not reached vs. 19.5 months (95% CI, 10.7–28.4) (HR 2.05; 95% CI, 1.13–3.73; p = 0.016). In our study, a duration of corticosteroids ≥30 days correlated with a better outcome in terms of OS: 12.1 months (95% CI, 8.3–15.8) vs. 4.6 months (95% CI, 2.4–6.7) (p = 0.001). The same happened when the cut-off point was changed to 15 days: 9.3 months (95% CI, 5.3–13.2) vs. 5.1 (95% CI, 2.4–7.9) (p = 0.007). No differences in terms of PFS were detected (p = 0.746 and p = 0.726 for a cut-off of 30 and 15 days, respectively). However, when analyzing results in terms of corticosteroid use, both OS and PFS were higher in patients receiving ≥10 mg prednisone equivalent for irAEs management. In terms of OS, patients treated with ≥10 mg of prednisone equivalent for irAEs management during ≥30 days presented the highest survival rates: 23 months (95% CI, 11.5–34.6) (p 0.001). Notably, no differences were detected when steroid therapy was given for cancer-related symptoms. FIGURE 1 | Landmark analysis according to the presence of irAEs. (A) Progression-free survival and the presence of irAEs (n = 175). (B) Overall Association Between the Use of Antibiotics survival and the presence of irAEs (n = 167). Kaplan-Meier curves with 2.4 months landmark analysis for progression-free survival (A) and 5.9 months for and Efficacy overall survival (B) in patients with or without irAEs. Abbreviations: PFS, One hundred and forty-one patients (52.8%) received antibiotics. progression-free survival; OS, overall survival; irAEs, immune-related adverse Quinolone (37.6%) and penicillin (33.3%) were the most events; CI, confidence interval; HR, hazard ratio; m, months. commonly used groups of antibiotics. Of note, the group of patients experiencing irAEs received significantly more antibiotics (58.6 vs. 45.2%, p = 0.036). However, no relation was found between the use of antibiotics and efficacy of anti-PD-(L)1 blockade agents, with a median the presence of liver or brain metastases. Interestingly, patients OS of 10.2 months (95% CI, 6.4–13.9) in patients receiving receiving corticosteroids for cancer-related symptoms presented antibiotics vs. 12.5 months (95% CI, 9.9–15.0) in patients not a lower incidence of irAEs (38.6 vs. 65.2%; p < 0.001) (Table 3). receiving antibiotics (p = 0.924), and a median PFS of 3.8 months Median OS was significantly longer in the group of patients (95% CI, 0.9–1.9) vs. 4.4 months (95% CI, 0.7–2.9), respectively that received <10 mg prednisone equivalent daily or no (p = 0.454). corticosteroids (n = 133) than in the group of patients that received ≥10 mg prednisone equivalent daily (n = 134): 14.7 Multivariable Analysis months (95% CI, 11.1–18.3) vs. 8.3 months (95% CI, 6.9– Multivariable analyses revealed that the presence of irAEs was 9.8) (HR 0.66; 95% CI, 0.48–0.90; p = 0.010) (Figure 2A). No the variable most strongly associated with a better response rate differences in PFS were observed. Median OS was significantly and OS (Table 4A and Table S2A), with an OR and HR of 0.36 shorter in patients receiving ≥10 mg prednisone equivalent daily and 0.32, respectively. In addition, in the multivariable analysis of for cancer-related symptoms (n = 75) than in the rest of the types of irAEs, cutaneous, endocrinological, and rheumatological study population (patients who did not receive corticosteroids irAEs were found to be significantly associated with increased or <10 mg prednisone equivalent daily and those who received ORR and OS (Table 4B and Table S2B). Pruritus and arthritis them for the management of irAEs, n = 192): 6 months (95% were the irAEs subtypes with the lowest HR. Frontiers in Oncology | www.frontiersin.org 6 September 2020 | Volume 10 | Article 1677 Riudavets et al. irAEs, Corticosteroids and Immunotherapy in NSCLC TABLE 3 | Comparison of patient characteristics between patients receiving corticosteroids for cancer-related symptoms and patients not receiving corticosteroids or received corticosteroids for management of irAEs* . Category Total Corticosteroids for Corticosteroids for irAEs or P-value n = 267 (%) cancer-related symptoms prednisone equivalent <10 mg/d * n = 83 (%) n = 184(%) Gender Male 186 (69.7) 60 (72.3) 126 (68.5) 0.568 Female 81 (30.3) 23 (27.7) 58 (31.5) Age 66.1 years 63.6 years 66.7 years 0.742 median (range) (26.7–85.2, (38.8–85.2, (26.7–83.7, IQR 14.3) IQR 13.3) IQR 14.2) Smoking status Non- or light smoker 26 (9.7) 9 (10.8) 17 (9.2) 0.662 Current or former smoker 241 (90.3) 74 (89.2) 167 (90.8) ECOG PS 0–1 227 (85) 63 (75.9) 164 (89.1) 0.009 2 40 (15) 20 (24.1) 20 (10.9) Histological subtype Squamous 86 (32.2) 25 (30.1) 61 (33.2) 0.673 Non-squamous 181(67.8) 58 (69.9) 123 (66.8) Treatment line 67 (36.4) 1st line 81 (30.3) 14 (16.9) 117 (63.6) 0.001 ≥2nd line 186 (69.7) 69 (83.1) Immune-checkpoint blockade schedule Monotherapy 209 (78.3) 75 (90.4) 134 (72.8) 0.002 Combination with chemotherapy 33 (12.3) 3 (3.6) 30 (16.3) Combination with anti-CTLA-4 25 (9.4) 5 (6) 20 (10.9) Presence of irAEs No 115 (43.1) 51 (61.4) 64 (34.8) <0.001 Yes 152 (56.9) 32 (38.6) 120 (65.2) Presence of brain metastases No 225 (84.3) 65 (78.3) 160 (87) 0.101 Yes 42 (15.7) 18 (21.7) 24 (13) Presence of liver metastases No 226 (84.6) 67 (80.7) 159 (86.4) 0.272 Yes 41 (15.4) 16 (19.3) 25 (13.6) Prednisone equivalent ≥10 mg/day use No 133 (49.8) 8 (9.6) 125 (67.9) <0.001 Yes 134 (50.2) 75 (90.4) 59 (32.1) Antibotic use No 126 (47.2) 32 (38.6) 94 (51.1) 0.064 Yes 141 (52.8) 51 (61.4) 90 (48.9) irAEs, immune-related adverse events; IQR, interquartile range; ECOG PS, Eastern Cooperative Oncology Group performance status; CTLA-4, cytotoxic T-lymphocyte antigen 4. *Includes patients that did not received corticosteroids. Variables such as ECOG PS ≥ 2, the presence of liver DISCUSSION metastasis, the use of corticosteroids, and receiving anti-PD-(L)1 therapy as a second line treatment or beyond were related to As far as we are aware, this is one of the largest studies to assess poorer outcomes, specially regarding to OS (Table 2A). The HR the association between the presence of irAEs and the efficacy for OS regarding corticosteroid use for cancer-related symptoms of anti-PD-(L)1 blockade agents in advanced NSCLC patients. was >2 (HR = 2.40), though it gave a result of 1.81 when Statistically significant differences were observed in terms of OS, corticosteroids were used for irAEs management. Notably, no PFS, ORR and DoR between patients experiencing irAEs and association was found between OS or ORR and the presence those who did not. Of note, a landmark analysis was performed of brain metastasis. No differences were found regarding the to minimize the immortal time bias potentially associated with grade of irAEs (grade 1–2 vs. grade 3–4 irAEs, excluding grade time-dependent factors such as the development of irAEs. 0), nor in terms of OS (p = 0.198) nor in regard to ORR In our study the incidence of irAEs was 56.9%, which is (p = 0.349). higher than previously reported (13–23). This could be explained Frontiers in Oncology | www.frontiersin.org 7 September 2020 | Volume 10 | Article 1677 Riudavets et al. irAEs, Corticosteroids and Immunotherapy in NSCLC FIGURE 2 | treated with anti-PD-(L)1 blockade agents on the basis of reported corticosteroid usage (≥10 mg of prednisone equivalent) in terms of OS (A). OS comparison in patients who did not received any corticosteroids, those who received them for irAEs treatment or cancer-related symptoms management (B). OS comparison according to the use of corticosteroids: for management of cancer-related symptoms and the rest of the population (C). a Use of ≥10 mg of prednisone equivalent daily for irAEs management. b Use of ≥10 mg of prednisone equivalent daily for cancer-related symptoms. TABLE 4 | Multivariable analysis of overall response rate according to clinical features (A) and type of irAEs (B). Variable Odds ratio 95% CI P-value (A). Overall response rate and clinical features (n = 267) Presence of irAEs No 0.36 0.20–0.65 0.001 Yes Treatment line 1st line 2.41 1.36–4.28 0.003 ≥2nd line (B). Overall response rate and type of irAEs (n = 267) Rash No 0.34 0.17–0.69 0.002 Yes Endocrine dysfunction No 0.38 0.18–0.81 0.012 Yes Arthritis No 0.28 0.11–0.72 0.008 Yes CI, confidence interval; irAEs, immune-related adverse events. mainly because this is a real-world data study and, in addition, 30.3% of patients received anti-PD-(L)1 blockade agents in a first line setting. Patients receiving anti-PD-(L)1 blockade agents as a second line therapy or beyond experienced significantly fewer irAEs than those treated in the first line. These findings are in line with those observed in KEYNOTE-010 and KEYNOTE-024 trials showing a higher rate of toxicity in treatment naïve patients who are probably less immunosuppressed than pre-treated patients (5, 8). A significantly higher rate of endocrinological toxicity was observed with the combination of immune blockade agents. This was also reported by a group of experts in endocrinopathies. In their review, hypophysitis was more common with anti-CTLA- 4 agents, whereas thyroid dysfunction was more frequent with anti-PD-1 agents. The combination of these agents appeared to increase the risk of immune-related endocrinopathies, which may be related to a more frequent association between CTLA-4 polymorphisms and autoimmune endocrinopathies in FIGURE 2 | Survival analysis with regard to corticosteroid use (≥10 mg of comparison with PD-1 polymorphisms (31). prednisone equivalent daily). (A) Overall survival and the use of ≥10 mg of prednisone equivalent daily (n = 267). (B) Overall survival and the reason for The landmark analysis showed longer PFS and OS, and more corticosteroid use (n = 267). (C) Overall survival and corticosteroid use (n = importantly, a greater ORR, in patients experiencing irAEs, 267). Abbreviations: PFS, progression-free survival; OS, overall survival; irAEs, which corroborates the influence exerted by the development immune-related adverse events; PDNe, prednisone equivalent; CI, confidence of immune-mediated toxicity on immunotherapy efficacy. The interval; HR, hazard ratio; m, months. Kaplan-Meier curves of patients (Continued) longer duration of treatment in patients experiencing irAEs might be explained by a lower percentage of treatment Frontiers in Oncology | www.frontiersin.org 8 September 2020 | Volume 10 | Article 1677 Riudavets et al. irAEs, Corticosteroids and Immunotherapy in NSCLC discontinuation in this group (75 vs. 91%, p = 0.001). This our study showed a greater median OS in patients receiving raises the question of whether the survival advantage attributed <10 mg of prednisone equivalent than those receiving ≥10 mg to the presence of irAEs is a reflection of the increased toxicity for toxicity management. The median duration of ≥10 mg associated with a longer duration of treatment or a direct prednisone equivalent had a significant positive impact on the result of the irAEs themselves. Recently, both prospective and efficacy of anti-PD-(L)1 blockade agents when given for irAEs retrospective data have suggested that this relation is independent management, but this finding might be explained by irAEs itself of guarantee-time bias, mainly because the majority of patients playing a role as a confounding factor. developed irAEs within the first 2–8 weeks after treatment Regarding the implications of the use of antibiotics, no initiation, supporting the predictive value of irAEs over treatment relation was found between the use of antibiotics (3 months duration (22, 23). before, during or 3 months after the end of anti-PD-(L)1 therapy) The presence of irAEs itself was the strongest variable and immunotherapy efficacy. These results differ from those associated with better outcomes in a multivariable analysis, observed in a cohort of 249 advanced NSCLC, renal cell and both according to ORR (OR 0.36; 95% CI, 0.20–0.65; p = urothelial carcinoma patients, in which 28% of patients received 0.001) and OS (HR 0.32; 95% CI, 0.22–0.46; p < 0.001). antibiotics within 2 months before or 1 month after the initiation Additionally, endocrine dysfunction, rash/pruritus, and arthritis of anti-PD-(L)1 blockade agents, for whom both PFS and OS were significantly associated with increased ORR and OS. were significantly shorter (34). However, those results were based Pruritus and arthritis presented the most favorable HR. on 69 out of 249 patients, so although informative, they are Consistent with our findings, two studies have also suggested that insufficient to draw any firm conclusion. In addition, the different thyroiditis and skin toxicity are related to longer OS (32, 33). time period over which the use of antibiotics was analyzed makes Of note, no association was found between OS or ORR and the it difficult to compare results. The prospective analysis of the severity of irAEs. effect of antibiotics on the efficacy of anti-PD-(L)1 therapy might Corticosteroids are the mainstay in the management of irAEs, help to understand the relation between them and the period of though they are also a common symptomatic treatment in time when the use of antibiotics might be discouraged. advanced NSCLC patients. Prednisone equivalent ≥10 mg daily Anti-PD-(L)1 therapy interruption due to irAEs is an issue for the symptomatic treatment of cancer-related symptoms at the that concerns us all. In our study, the 65.9% of patients who time of initiation or during anti-PD-(L)1 blockade treatment was discontinued treatment due to irAEs and did not progress associated with significantly poorer outcomes than for patients contrasts with the 28.7% of patients who did not interrupt who did not receive corticosteroids or those who received them immunotherapy due to toxicity and did not progress. These to manage irAEs. These results are in line with those reported results are in line with a post-hoc analysis of the Checkmate- recently in a study carried out on 640 patients with advanced 067 trial in patients with advanced melanoma, in which NSCLC receiving single-agent anti-PD-(L)1 blockade treatment both PFS and OS were similar after 4 years regardless of (25). In that publication, a multivariable analysis including discontinuation of treatment due to irAEs (35). Taken together, smoking status, ECOG, and history of brain metastases showed these data suggest that treatment interruption due to irAEs that baseline corticosteroid use was significantly associated with does not seem to compromise the efficacy of anti-PD-(L)1 decreased ORR, PFS, and OS. Our multivariable analysis assessed blockade agents. similar prognostic factors, but also included the treatment line This study has several limitations that could be addressed as a variable that could influence outcomes. Corticosteroid use in future research. First, its retrospective design and the for palliation of cancer-related symptoms and anti-PD-(L)1 need for a longer follow-up period to fully assess long-term therapy in the second line or beyond were strong, independent outcomes. Second, the heterogeneity of treatment strategies variables associated with a poorer outcome. Patients receiving included in this study, which may influence the efficacy of corticosteroids for cancer-related symptoms showed a higher immunotherapy and the frequency of irAEs. Third, the low percentage of patients with an ECOG PS of 2 and treatment frequency of some irAEs subtypes may limit the evaluation of in the second line or beyond with single-agent. On the whole, their relationship with efficacy. A prospective study including a these findings suggest that baseline corticosteroid use may simply larger cohort of patients with the same treatment strategy would identify a group of patients with a higher volume or more help to overcome those limitations and adequately assess the aggressive disease or with basal illnesses that worsen cancer real impact of corticosteroids use by accounting for potential prognosis. The same inference has been made in a recently confounding factors. published study in which the authors concluded that the worst In conclusion, our results indicate that the presence of outcome associated with ≥10 mg of prednisone equivalent daily irAEs is associated with anti-PD-(L)1 blockade agents efficacy at the time of immunotherapy seemed to be driven by a poor- in patients with advanced NSCLC. This is one of the most prognosis subgroup of patients who received corticosteroids extensive studies to date to reveal an association between as a palliative treatment (26). Moreover, previous data suggest the presence of irAEs and the efficacy of immunotherapy in that the positive correlation between the presence of irAEs and advanced NSCLC when landmark and multivariable analyses efficacy of anti-PD-(L)1 blockade agents in advanced NSCLC is are applied. Corticosteroid use of ≥10 mg of prednisone not hampered by the use of corticosteroids for the treatment equivalent daily was associated with significantly poorer of irAEs, concluding that its use in this context should not be outcomes when given for patients’ cancer-related symptoms. restricted for fear of loss of any outcome advantage. In contrast, No significant differences were observed in terms of efficacy Frontiers in Oncology | www.frontiersin.org 9 September 2020 | Volume 10 | Article 1677 Riudavets et al. irAEs, Corticosteroids and Immunotherapy in NSCLC between patients that did not receive corticosteroids or who analysis and interpretation of the data. JS, GA, PG, IS, ABarb, received <10 mg prednisone equivalent daily and those who and ABarn contributed to the acquisition of the data. All authors received ≥10 mg for the management of irAEs. No relation reviewed and approved the final version of the manuscript. was found between antibiotics and outcomes of anti-PD-(L)1 All authors contributed to the article and approved the blockade agents. submitted version. DATA AVAILABILITY STATEMENT FUNDING The datasets generated for this study are available on request to This research did not receive any specific grant from funding the corresponding author. agencies in the public, commercial, or not-for-profit sectors. ETHICS STATEMENT ACKNOWLEDGMENTS The studies involving human participants were reviewed and approved by Hospital de la Santa Creu i Sant Pau. The We would like to thank all the patients and medical staff at patients/participants provided their written informed consent to the participating institutions that contributed to this study. participate in this study. Additionally, the authors would also like to acknowledge writing support from Hye Park and Constance Pell for their contribution AUTHOR CONTRIBUTIONS during the development of this manuscript. MR and MM contributed to the conception and design of the SUPPLEMENTARY MATERIAL study, data acquisition, statistical analysis, interpretation of the data, and writing of the manuscript. JM and RG-C contributed The Supplementary Material for this article can be found to the acquisition and interpretation of the data and revision online at: https://www.frontiersin.org/articles/10.3389/fonc. of the manuscript. IG and LC contributed to the statistical 2020.01677/full#supplementary-material REFERENCES phase 2 cohort of the open-label KEYNOTE-021 study. Lancet Oncol. (2016) 17:1497–508. doi: 10.1016/S1470-2045(16)30498-3 1. Chen DS, Mellman I. Oncology meets immunology: the cancer-immunity 10. Paz-Ares L, Luft A, Vicente D, Tafreshi A, Gümüş M, Mazières J, cycle. Immunity. (2013) 39:1–10. doi: 10.1016/j.immuni.2013.07.012 et al. Pembrolizumab plus chemotherapy for squamous non–small-cell 2. Naidoo J, Page DB, Wolchok JD. Immune checkpoint blockade. Hematol lung cancer. N Engl J Med. (2018) 379:2040–51. doi: 10.1056/NEJMoa Oncol Clin North Am. (2014) 28:585–600. doi: 10.1016/j.hoc.2014.02.002 1810865 3. Brahmer J, Reckamp KL, Baas P, Crinò L, Eberhardt W, Poddubskaya E, et al. 11. Gandhi L, Rodríguez-Abreu D, Gadgeel S, Esteban E, Felip E, De Angelis F, Nivolumab versus docetaxel in advanced squamous-cell non-small-cell lung et al. Pembrolizumab plus chemotherapy in metastatic non–small-cell lung cancer. N Engl J Med. (2015) 373:123–35. doi: 10.1056/NEJMoa1504627 cancer. N Engl J Med. (2018) 378:2078–92. doi: 10.1056/NEJMoa1801005 4. Borghaei H, Paz-Ares L, Horn L, Spiegel DR, Steins M, Ready NE, et al. 12. Hellmann MD, Ciuleanu T-E, Pluzanski A, Lee JS, Otterson GA, Audigier- Nivolumab versus docetaxel in advanced nonsquamous non-small-cell lung Valette C, et al. Nivolumab plus Ipilimumab in lung cancer with cancer. N Engl J Med. (2015) 373:1627–39. doi: 10.1056/NEJMoa1507643 a high tumor mutational burden. N Engl J Med. (2018) 378:2093– 5. Herbst RS, Baas P, Kim DW, Felip E, Pérez-Garcia JL, Han JY, 104. doi: 10.1056/NEJMoa1801946 et al. Pembrolizumab versus docetaxel for previously treated, 13. Postow MA, Sidlow R, Hellmann MD. Immune-related adverse events PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE- associated with immune checkpoint blockade. N Engl J Med. (2018) 378:158– 010): a randomised controlled trial. Lancet. (2016) 387:1540– 68. doi: 10.1056/NEJMra1703481 50. doi: 10.1016/S0140-6736(15)01281-7 14. Champiat S, Lambotte O, Barreau E, Belkhir R, Berdelou A, Carbonnel 6. Rittmeyer A, Barlesi F, Waterkamp D, Park K, Ciardiello F, von F, et al. Management of immune checkpoint blockade dysimmune Pawel J, et al. Atezolizumab versus docetaxel in patients with toxicities: a collaborative position paper. Ann Oncol. (2016) 27:559– previously treated non-small-cell lung cancer (OAK): a phase 3, 74. doi: 10.1093/annonc/mdv623 open-label, multicentre randomised controlled trial. Lancet. (2017) 15. Weber JS, Hodi FS, Wolchok JD, Topalian SL, Schadendorf D, Larkin 389:255–65. doi: 10.1016/S0140-6736(16)32517-X J, et al. Safety profile of nivolumab monotherapy: a pooled analysis 7. Reck M, Rodríguez-Abreu D, Robinson AG, Hui R, Csöszi T, Fülöp A, of patients with advanced melanoma. J Clin Oncol. (2017) 35:785– et al. Pembrolizumab versus chemotherapy for PD-L1-positive non-small- 92. doi: 10.1200/JCO.2015.66.1389 cell lung cancer. N Engl J Med. (2016) 375:1823–33. doi: 10.1056/NEJMoa160 16. Haratani K, Hayashi H, Chiba Y, Kudo K, Yonesaka K, Kato R, 6774 et al. Association of immune-related adverse events with nivolumab 8. Reck M, Rodríguez-Abreu D, Robinson AG, Hui R, Csoszi T, Fülöp A, efficacy in non-small cell lung cancer. JAMA Oncol. (2018) 4:374– et al. Updated analysis of KEYNOTE-024: pembrolizumab versus platinum- 8. doi: 10.1001/jamaoncol.2017.2925 based chemotherapy for advanced non-small-cell lung cancer with PD-L1 17. Sato K, Akamatsu H, Murakami E, Sasaki S, Kanai K, Hayata A, et al. tumor proportion score of 50% or greater. J Clin Oncol. (2019) 37:537– Correlation between immune-related adverse events and efficacy in non- 46. doi: 10.1200/JCO.18.00149 small cell lung cancer treated with nivolumab. Lung Cancer. (2018) 115:71– 9. Langer CJ, Gadgeel SM, Borghaei H, Papadimitrakopoulou VA, Patnaik A, 4. doi: 10.1016/j.lungcan.2017.11.019 Powell SF, et al. Carboplatin and pemetrexed with or without pembrolizumab 18. Teraoka S, Fujimoto D, Morimoto T, Kawachi H, Ito M, Sato Y, for advanced, non-squamous non-small-cell lung cancer: a randomised, et al. Early immune-related adverse events and association with Frontiers in Oncology | www.frontiersin.org 10 September 2020 | Volume 10 | Article 1677 Riudavets et al. irAEs, Corticosteroids and Immunotherapy in NSCLC outcome in advanced non–small cell lung cancer patients treated 30. Dafni U. Landmark analysis at the 25-year landmark point. Circ Cardiovasc with nivolumab: a prospective cohort study. J Thorac Oncol. (2017) Qual Outcomes. (2011) 4:363–71. doi: 10.1161/CIRCOUTCOMES.110.957951 12:1798–805. doi: 10.1016/j.jtho.2017.08.022 31. Chang LS, Barroso-Sousa R, Tolaney SM, Hodi FS, Kaiser UB, Min L. 19. Toi Y, Sugawara S, Kawashima Y, Aiba T, Kawana S, Saito R, et al. Association Endocrine toxicity of cancer immunotherapy targeting immune checkpoints. of immune-related adverse events with clinical benefit in patients with Endocrine Rev. (2019) 40:17–65. doi: 10.1210/er.2018-00006 advanced non-small-cell lung cancer treated with nivolumab. Oncologist. 32. Osorio JC, Ni A, Chaft JE, Pollina R, Kasler MK, Stephens D, et al. (2018) 23:1358–65. doi: 10.1634/theoncologist.2017-0384 Antibody-mediated thyroid dysfunction during t-cell checkpoint blockade 20. Ricciuti B, Genova C, De Giglio A, Bassanelli M, Dal Bello MG, Metro in patients with non-small cell lung cancer. Ann Oncol. (2017) 28:583– G, et al. Impact of immune-related adverse events on survival in patients 89. doi: 10.1093/annonc/mdw640 with advanced non-small cell lung cancer treated with nivolumab: long-term 33. Hasan-Ali O, Diem S, Markert E, Jochum W, Katrin K, French LE, outcomes from a multi-institutional analysis. J Cancer Res Clin Oncol. (2019) et al. Characterization of nivolumab associated skin reactions in patients 145:479–85. doi: 10.1007/s00432-018-2805-3 with metastatic non-small cell lung cancer. Oncoimmunology. (2016) 21. Shafqat H, Gourdin T, Sion A. Immune-related adverse events 5:e1231292. doi: 10.1080/2162402X.2016.1231292 are linked with improved progression-free survival in patients 34. Routy B, Le Chatelier E, Derosa L, Duong CPM, Alou MT, Daillère R, et al. receiving anti-PD-1/PD-L1 therapy. Semin Oncol. (2018) 45:156– Gut microbiome influences efficacy of PD-1-based immunotherapy against 63. doi: 10.1053/j.seminoncol.2018.07.003 epithelial tumors. Science. (2018) 359:91–7. doi: 10.1126/science.aan3706 22. Lisberg A, Tucker DA, Goldman JW, Wolf B, Carroll J, Hardy A, et al. 35. Hodi FS, Chiarion-Sileni V, Gonzalez R, Grob JJ, Rutkowski P, Cowey CL, Treatment-related adverse events predict improved clinical outcome in et al. Nivolumab plus ipilimumab or nivolumab alone versus ipilimumab NSCLC patients on KEYNOTE-001 at a single center. Cancer Immunol Res. alone in advanced melanoma (CheckMate 067): 4-year outcomes of a (2018) 6:288–94. doi: 10.1158/2326-6066.CIR-17-0063 multicentre, randomised, phase 3 trial. Lancet Oncol. (2018) 19:1480– 23. Toi Y, Sugawara S, Sugisaka J, Ono H, Kawashima Y, Aiba T, et al. 92. doi: 10.1016/S1470-2045(18)30700-9 Profiling preexisting antibodies in patients treated with anti-PD-1 therapy for advanced non-small cell lung cancer. JAMA Oncol. (2019) 5:376– Conflict of Interest: RG-C reports personal fees and other from Roche, BMS, 83. doi: 10.1001/jamaoncol.2018.5860 MSD, Pfizer, Lilly, AstraZeneca, Takeda, Novartis, Boeringher Ingelheim, during 24. Garant A, Guilbault C, Ekmekjian T, Greenwald Z, Murgoi P, Vuong T. the conduct of the study. ABarb reports non-financial support from MSD, Roche, Concomitant use of corticosteroids and immune checkpoint inhibitors in AstraZeneca, BMS, outside the submitted work. ABarn reports grants from Roche, patients with hematologic or solid neoplasms: a systematic review. Crit Rev Novartis, Pfizer, BMS, AstraZeneca, Lilly, during the conduct of the study, personal Oncol Hematol. (2017) 120:86–92. doi: 10.1016/j.critrevonc.2017.10.009 fees from Novartis, Roche, Pfizer, Genomic Health, Pierre Fabre, AstraZeneca, 25. Arbour KC, Mezquita L, Long N, Rizvi H, Auclin E, Ni A, et al. Impact of EISAI, outside the submitted work. MM reports personal fees and other from baseline steroids on efficacy of programmed cell death-1 and programmed Viforpharma, Roche, BMS, MSD, Pfizer, Lilly, AstraZeneca, Takeda, Novartis, death-ligand 1 blockade in patients with non-small-cell lung cancer. J Clin Boeringher Ingelheim, during the conduct of the study. Oncol. (2018) 36:2872–8. doi: 10.1200/JCO.2018.79.0006 26. Ricciuti B, Dahlberg SE, Adeni A, Sholl LM, Nishino M, Awad MM. The remaining authors declare that the research was conducted in the absence of Immune checkpoint inhibitor outcomes for patients with non–small-cell lung any commercial or financial relationships that could be construed as a potential cancer receiving baseline corticosteroids for palliative versus nonpalliative conflict of interest. indications. J Clin Oncol. (2019) 37:1927–34. doi: 10.1200/JCO.19. 00189 The handling editor declared a past co-authorship with several of the authors IS 27. Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, et al. and MM. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer. (2009) 45:228–47. doi: 10.1016/j.ejca.2008. Copyright © 2020 Riudavets, Mosquera, Garcia-Campelo, Serra, Anguera, Gallardo, 10.026 Sullivan, Barba, del Carpio, Barnadas, Gich and Majem. This is an open-access 28. National Cancer Institute. Common Terminology Criteria for Adverse Events. article distributed under the terms of the Creative Commons Attribution License (CC (2009). Available online at: https://evs.nci.nih.gov/ftp1/CTCAE/CTCAE_4. BY). The use, distribution or reproduction in other forums is permitted, provided 03/CTCAE_4.03_2010-06-14_QuickReference_5x7.pdf the original author(s) and the copyright owner(s) are credited and that the original 29. Anderson JR, Cain KC, Gelber RD. Analysis of survival by tumor response publication in this journal is cited, in accordance with accepted academic practice. and other comparisons of time-to-event by outcome variables. J Clin Oncol. No use, distribution or reproduction is permitted which does not comply with these (2008) 26:3913–5. doi: 10.1200/JCO.2008.16.1000 terms. Frontiers in Oncology | www.frontiersin.org 11 September 2020 | Volume 10 | Article 1677