Immune checkpoint inhibitors in real-world treatment of older adults with non-small cell lung cancer

Abstract

Objective:

To evaluate the efficacy and toxicity of immune checkpoint inhibitors (ICIs) in older patients with advanced non-small cell lung cancer (NSCLC) seen in routine clinical practice.

Design:

Retrospective study.

Setting:

Single academic institution and its affiliated centers.

Participants:

Patients aged ≥70 years with advanced stage NSCLC seen between 4/1/2015 and 4/1/2017 and treated with ICIs.

Measurements:

Efficacy data included overall survival (OS) and time to treatment failure (TTF), stratified by age, comorbidities [Charlson comorbidity index (CCI)], and performance status, and estimated using the Kaplan-Meier method and log-rank test. Toxicity data included immune-related adverse events (irAEs), need for glucocorticoids, and hospitalization. The associations of toxicity with age, CCI, and Eastern Cooperative Group Performance Status (ECOG PS) were evaluated using the exact Chi-square test or Fisher’s exact test.

Results:

We included 75 patients (median age: 74 years, range 70-92); 53% had CCI ≥3 and 49% had ECOG PS ≥2. Median OS for the whole cohort was 8.2 months (ECOG PS 0-1 vs. ≥2: 13.7 vs. 3.8 months; p<0.01). Median TTF was 4.2 months (ECOG PS 0-1 vs. ≥2: 5.6 vs. 2.0 months; P=0.02). Overall, 37% of patients experienced irAE of any grade (a total of 37 events); 8% were Grade ≥3 (no ICI-related deaths). Of those who discontinued ICIs (N=64), 15% were due to irAEs. Of those who experienced irAEs, 64% required glucocorticoids. Hospitalizations during ICI treatment occurred in 72%. Toxicity generally did not differ by age, CCI, or ECOG PS.

Conclusions:

Outcomes in our cohort were driven by ECOG PS rather than chronological age or comorbidities. The relatively high rates of ICI discontinuation, glucocorticoids utilization, and hospitalization during ICI treatment in our study highlight the vulnerability of older adults with advanced NSCLC even in the immunotherapy era.

Introduction:

Non-small cell lung cancer (NSCLC) is predominantly a disease of older adults; 50% of all lung cancer diagnoses and NSCLC-related deaths occur in patients over the age of 70.^6,7 Although immune checkpoint inhibitors (ICIs) have changed the paradigm of oncologic care in multiple tumor types and in NSCLC in particular,^1,2,8-10 the efficacy of ICIs in older patients with advanced-stage NSCLC, alone or in combination with other agents, remains undefined, largely because of the under-representation of older adults in landmark clinical trials evaluating their use.^11,12 Patients aged >65 years made up 31.7% to 41.0% of clinical trial participants, despite the fact that 60.6% to 64.0% of patients with NSCLC receiving ICIs in clinical practice are over the age of 65.¹² In the CheckMate 017 and 057 trials, for example, only about 30% of the study population was aged 64-75 years and 8% was aged >75 years.^4,5 In addition, older adults who are enrolled on clinical trials are typically fitter than the general geriatric oncology population seen in routine clinical practice.

The Eastern Cooperative Oncology Group Performance Status (ECOG PS) is a standard scale used in oncology practices as a simple tool to assess performance status and commonly used to determine eligibility for clinical trials. The scale ranges from 0-5 (0: fully active, 1: restricted in physically strenuous activity, 2: up and about >50% of walking hours, 3: confined to bed or chair >50% of walking hours, 4: completely disabled, and 5: dead). Those with ECOG PS ≥2 are frequently excluded from clinical trials evaluating ICIs, even though more than 20% of such patients are treated with ICIs in real-world practice.^13-15 The number of older patients with cancer and ECOG PS ≥2 treated with ICIs in real-world practice may be as high as 30% if not higher.¹³

Immune checkpoint inhibitors block certain proteins [such as anti-programmed death-1 (PD-1), anti-programmed death-ligand 1 (PD-L1), and cytotoxic T-lymphocyte associated protein 4 (CTLA-4)] that are on cancer or immune cells. By blocking these proteins, they remove inhibitory signals of T-cell activation and subsequently allow tumor-reactive T cells to mount an antitumor response i.e. releasing the “brakes” on the immune system.^16-18 Whether the efficacy and toxicity profile of ICIs in older patients differs from that in younger patients remains unclear.¹⁹ Concerns have been raised about diminished efficacy of immunotherapy in the older patient population, perhaps secondary to immunosenescence, involving a complex physiologic interplay between both the aging innate and adaptive immune systems.^19-22 In addition, it is unclear if the efficacy and toxicity profile of ICIs differs by performance status. To address these knowledge gaps, we conducted this study to evaluate the efficacy and toxicity of ICIs in older patients with advanced NSCLC seen in routine clinical practice.

Methods:

Study design, setting, and sample:

We retrospectively analyzed data on consecutive patients aged ≥70 years with advanced-stage NSCLC who were treated with any ICI [i.e., anti-PD1/anti-PD-L1 antibodies] at the University of Rochester Medical Center/Wilmot Cancer Institute and its affiliated satellite medical oncology clinics between April 1, 2015, and April 1, 2017. The data cut-off was on December 1, 2017, and data were censored at patients' last documented clinic visit. Inclusion criteria were: 1) A histologically confirmed diagnosis of NSCLC (adenocarcinoma, squamous cell carcinoma, or mixed histology); 2) Age ≥70 years at ICI initiation; 3) Receipt of at least one cycle of ICI (e.g., nivolumab, pembrolizumab, or atezoluzumab). The project was approved by the University of Rochester Research Subjects Review Board.

Data collection:

All patient data were extracted from the electronic medical record. Data were entered into the REDCap database utilizing electronic data capture forms to ensure consistency during data abstraction, input, and analysis.²³ Data captured included age (at initiation of ICI), sex, comorbidities [Charlson Comorbidity Index (CCI)²⁴ excluding NSCLC], ECOG PS, cancer stage, tumor histology, and line of ICI therapy (e.g., first line, second line). CCI was manually abstracted rather than calculated using the International Classification of Diseases Codes as the agreement between the two was previously shown to be good.²⁵ Tumor PD-L1 expression was able to be evaluated only among patients who received ICI in the first-line setting.

Study Aim Definitions:

For the primary aim of efficacy, overall survival (OS) and time to treatment failure (TTF) were measured. OS was defined as the time from start of ICI treatment to death or censored at the last clinic visit. TTF was measured from the start of ICI treatment to the time ICI treatment was discontinued due to any reason including recurrence or disease progression, toxicity, unrelated hospitalization, patient withdrawal, death, or censored at last clinic visit. Best response was determined on consensus review of radiographic and clinical findings by the study investigators rather than formal Response Evaluation Criteria in Solid Tumors (RECIST) criteria, which were not captured.

For the secondary aim of toxicity, data were collected during receipt of ICI therapy and for up to 30 days post-ICI discontinuation. Adverse events (AEs) were graded using Common Terminology Criteria for Adverse Events (CTCAE) v4.0. In addition we collected the following information: 1) frequency and severity of immune-related AEs (irAEs); 2) the need for glucocorticoids; 3) dose delays or interruptions, hospitalizations, and admissions to the intensive care unit during therapy; 4) number of falls as captured in the medical record during treatment; and 5) referrals to physical/occupational therapy or skilled nursing facilities (SNF) during treatment. Additional variables of interest included referrals to palliative care and hospice at any point during cancer care.

Statistical Analysis:

Descriptive analyses were used to summarize study sample characteristics, toxicity, and efficacy data. Event-time distributions were estimated with the use of the Kaplan-Meier method. The log-rank test was used to compare OS and TTF by patient characteristics determined at the start of ICI treatment. Similar to the log-rank test, baseline characteristics were assessed singly in an unadjusted Cox regression model, and then jointly in an adjusted Cox multiple regression model. Hazard ratios for OS and TTF are presented with 95% confidence intervals (CIs) along with p-values from these models. Although not a guarantee to address the inter-relationships between these characteristics (confounding), the adjusted Cox model provides estimates for each factor holding the other factors constant and is a better evaluation of the independent association with the outcomes. The associations between categorical outcomes of interest, namely best response and toxicity, and clinical factors including age, comorbidities (CCI), and ECOG PS were evaluated using the exact Chi-square test or Fisher’s exact test where appropriate. All reported P-values were two-sided, and a P-value of <0.05 was considered statistically significant. All data analyses were performed using SAS software, version 9.4 (SAS Institute, Inc., Cary, NC).

Results:

Patient Characteristics:

Seventy-five patients with advanced-stage NSCLC age ≥70 years received ICI therapy during the study period (Table 1). Median age at initiation of ICI was 74 years (range, 70-92). Mean CCI was 2.7 (standard deviation, 1.9; range 0-8); 53% had CCI of ≥3. Almost half (49%) had ECOG PS ≥2. Most patients (87%) received nivolumab and received ICI therapy in the second-line setting (69%).

Table 1:

Baseline patient characteristics

Patient Characteristics		N=75
Age at initiation of ICI, mean in years (SD, range)		74 (5.4, 70-92)
Age, N (%)	<80	58 (77.3)
	≥80	17 (22.7)
Male, N (%)		39 (52.0)
Charlson Comorbidity Index, N (%)	<3	35 (46.7)
	≥3	40 (53.3)
ECOG PS, N (%)	0	3 (4.0)
	1	35 (46.7)
	2	34 (45.3)
	3	3 (4.0)
Cancer stage, N (%)	IIIb	2 (2.7)
	IV	73 (97.3)
Histology, N (%)	Adenocarcinoma	51 (68.0)
	Other	24 (32.0)
Immune checkpoint inhibitor, N (%)	Nivolumab	65 (86.7)
	Pembrolizumab	6 (8.0)
	Other	4 (5.3)
Line of therapy with immune checkpoint inhibitor, N (%)	1	12 (16.0)
	2	52 (69.3)
	3	8 (10.7)
	4	3 (4.0)

Abbreviation: ECOG PS, Eastern Cooperative Group Performance Status; ICI, Immune checkpoint inhibitor; SD, standard deviation

By December 1, 2017, almost three-quarters (71%) of evaluated patients had died. Median OS for the entire cohort was 8.2 months [95% Confidence Interval (CI) 4.8-13.6]. Patients with ECOG PS 0-1 had a higher median OS than patients with ECOG PS ≥2: 13.7 months (95% CI 10.0-25.1) vs. 3.8 months (95% CI 1.6-7.8; p<0.01) (Figure 1). One-year survival was 60% for ECOG PS 0-1 and 21% for ECOG PS ≥2. The OS difference by ECOG PS was statistically significant for second-line treatment (P=0.02), with only a trend towards significance for first-line of ICI treatment (P=0.07). In the multivariate Cox proportional hazards regression model, there was a higher risk of death for patients with ECOG PS ≥2 [Hazard Ratio (HR) 3.0, 95% CI 1.6-5.7, p<0.01) after adjusting for age, sex, CCI, histology, or line of therapy (Table 2). Other factors were not statistically associated with OS.

Figure 1:

Overall survival (left) and time to treatment failure (right) stratified by Eastern Cooperative Group Performance Status

Table 2:

Unadjusted and adjusted cox regressions models evaluating the associations of age, sex, Eastern Cooperative Group Performance Status, Charlson Comorbidity Index, histology, and line of therapy with overall survival and treatment failure

	Unadjusted Cox Regression Models			Adjusted Cox Regression Models
	Hazard Ratio	95% CI	P-value	Hazard Ratio	95% CI	P-value
Overall Survival:
ECOG PS (≥2 vs. <2)	2.45	1.40-4.28	<0.01	3.03	1.61-5.71	<0.01
Age (≥80 vs. <80)	0.92	0.48-1.74	0.79	0.83	0.43-1.63	0.59
Male (yes vs. no)	1.06	0.62-1.82	0.83	1.01	0.52-1.97	0.97
CCI (≥3 vs. <3)	0.94	0.54-1.61	0.81	0.62	0.30-1.27	0.19
Adenocarcinoma (yes vs. no)	0.87	0.49-1.55	0.65	0.78	0.42-1.46	0.44
Front line (yes vs. no)	0.62	0.26-1.45	0.27	0.71	0.29-1.72	0.45
Time to Treatment Failure:
ECOG PS (≥2 vs. <2)	1.82	1.10-3.02	0.02	2.11	1.21-3.68	<0.01
Age (≥80 vs. <80)	0.77	0.42-1.39	0.38	0.70	0.38-1.32	0.27
Male (yes vs. no)	0.86	0.53-1.41	0.55	0.75	0.41-1.37	0.35
CCI (≥3 vs. <3)	0.89	0.54-1.45	0.63	0.78	0.41-1.45	0.43
Adenocarcinoma (yes vs. no)	0.87	0.52-1.48	0.62	0.77	0.42-1.39	0.38
Front line (yes vs. no)	0.71	0.36-1.40	0.3190	0.68	0.33-1.40	0.30

Abbreviations: CCI, Charlson Comorbidity Index; CI, Confidence Interval; ECOG PS, Eastern Cooperative Group Performance Status; ICI, Immune checkpoint inhibitor

Median TTF for the entire cohort was 4.2 months (95% CI 2.0-5.8). Median TTF for patients with ECOG PS 0-1 and PS ≥2 was 5.6 months (95% CI 3.2-7.3) and 2.0 months (95% CI 0.9-5.0; P=0.02) (Figure 1). In the multivariate Cox proportional hazards regression model, patients with PS ≥2 had increased risk of treatment failure (HR 2.1, 95% CI 1.2-3.7, p<0.01) after adjusting for age, sex, CCI, histology, and line of therapy (Table 2). Other factors were not statistically associated with TTF.

In terms of best response as deemed by clinical and radiographic assessments, we observed no complete response, and 36% of patients attained a partial response (Table 3). Non-responders included almost half (48%) who had progressive disease and 16% who had stable disease. There was no correlation demonstrated between responders and non-responders in terms of age, sex, CCI, ECOG PS, histology, or line of therapy (Table 3).

Table 3:

Associations of age, Eastern Cooperative Group Performance Status, histology, and line of therapy with best response

Patient Characteristics		Best Response to ICI
	N	Respondera	Non-responderb	P-value c
	75	27 (36.0)	48 (64.0)
Age at initiation of ICI (years)
<80	58	21 (36.2)	37 (63.8)	1
≥80	17	6 (35.3)	11 (64.7)
Sex
Male	39	14 (35.9)	25 (64.1)	1
Female	36	13 (36.1)	23 (63.9)
CCI
<3	35	13 (48.2)	22 (45.8)	1
≥3	40	14 (51.9)	26 (54.2)
ECOG PS
0-1	38	16 (42.1)	22 (57.9)	0.34
≥2	37	11 (29.7)	26 (70.3)
Histology
Adenocarcinoma	51	20 (39.2)	31 (60.8)	0.33
Squamous cell carcinoma	18	4 (22.2)	14 (77.8)
Mixed/Other	6	3 (50.0)	3 (50.0)
Line
Front line	12	7 (58.3)	5 (41.7)	0.10
Subsequent line	63	20 (31.8)	43 (68.3)

Abbreviations: CCI, Charlson Comorbidity Index; ECOG PS, Eastern Cooperative Group Performance Status; ICI, Immune checkpoint inhibitor

^aComplete or partial response

^bStable disease or progression

^cFisher’s exact test

Safety:

By December 1, 2017, the majority (85%) of patients had discontinued ICI. Of these patients, 15% (N=10) discontinued secondary to irAEs, whereas 60% discontinued due to disease progression. Nearly half (47%) experienced a dose delay or interruption. The overall frequency of irAEs is shown and stratified by age, sex, CCI, ECOG PS, histology, and line of therapy (Table 4). Twenty-eight patients (37%) experienced at least one irAE, 6 of whom (8%) had at least one Grade ≥3 toxicity, including 4 patients who had Grade ≥3 pneumonitis (Table 4). The most common irAEs of any grade were pneumonitis and thyroiditis (12% each), followed by colitis and dermatitis (9% each). The rate of colitis in patients age ≥80 and <80 years were 24% and 5% (P=0.04), respectively. The rate of pneumonitis in patients with adenocarcinoma and other histologies were 5.9% and 0% (P=0.03), respectively. The rate of thyroiditis in patients who received an ICI in the front line and subsequent line were 33.3% and 7.9% (P=0.03), respectively. No patient had documented evidence of transaminitis, myocarditis, or adrenal insufficiency. Among those who experienced an irAE, almost two-thirds (64%) required systemic glucocorticoids (Supplemental Table 1). Approximately three-quarters (72%, N=54/75) of the cohort experienced at least one hospitalization during ICI treatment (7 patients were hospitalized due to irAEs); 17 patients (32%) had ≥ 2 hospitalizations, 6 patients (8%) had ≥3 hospitalizations, and 10 patients (13%) were admitted to the intensive care unit (ICU). During ICI treatment, 12 patients (16%) sustained a fall, 10 patients (13%) had a physical or occupational therapy referrals, and 8 patients (11%) were referred to a SNF. From the time of their initial cancer diagnosis, only 55% of patients had a palliative care referral. None of these variables were statistically associated with age, sex, CCI, ECOG PS, histology, or line of therapy except for palliative care referrals. Palliative care referrals for patients treated with ICI in the subsequent line and front line were 60% and 25% (P=0.03), respectively.

Table 4:

Immune mediated adverse events stratified by age, sex, Charlson Comorbidity Index, Eastern Cooperative Group Performance Status, histology, and line of therapy

Event	Overall (N=75)		Age at initiation of ICI (years)a		Sexa		CCIa		ECOG PSa		Adenocarcinomaa		Front linea
	Any irAEb	Grade 3,4, and 5 irAE	<80 (N=58)	≥80 (N=17)	Female (N=36	Male (N=39)	<3 (N=35)	≥3 (N=40)	0-1 (N=38)	≥2 (N=37)	Yes (N=51)	No (N=24)	Yes (N=12)	No (N=63)
Colitis, N (%)	7 (9.3)	0	3 (5.2)c	4 (23.5)c	5 (13.9)	2 (5.1)	5 (14.3)	2 (5.0)	5 (13.2)	2 (5.4)	5 (9.8)	2 (8.3)	0	7 (11.1)
Dermatitis, N (%)	7 (9.3)	1 (1.3)	5 (8.6)	2 (11.8)	4 (11.1)	3 (7.7)	5 (14.3)	2 (5.0)	4 (10.5)	3 (8.1)	4 (7.8)	3 (12.5)	1 (16.7)	5 (7.9)
Hypophysitis, N (%)	1 (1.3)	1 (1.3)	1 (1.7)	0	0	1 (2.6)	0	1 (2.5)	0	1 (2.7)	1 (2.0)	0	0	1 (1.6)
Musculoskeletal, N (%)	2 (2.7)	0	2 (3.4)	0	1 (2.8)	1 (2.6)	2 (5.7)	0	2 (5.3)	0	1 (2.0)	0	0	2 (3.2)
Nephritis, N (%)	2 (2.7)	0	1 (1.7)	1 (5.9)	0	2 (5.1)	0	2 (5.0)	2 (5.3)	0	2 (3.9)	1 (4.2)	0	2 (3.2)
Pneumonitis, N (%)	9 (12.0)	4 (5.3)	8 (13.8)	1 (5.9)	2 (5.6)	7 (18.0)	5 (14.3)	4 (10.0)	6 (15.8)	3 (8.1)	3 (5.9)d	0d	1 (8.3)	8 (12.7)
Thyroiditis, N (%)	9 (12.0)	0	9 (15.5)	0	6 (16.7)	3 (7.7)	3 (8.6)	6 (15.0)	6 (15.8)	3 (8.1)	8 (15.7)	6 (25.0)	4 (33.3)	5 (7.9)e
Myocarditis, N (%)	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Hepatitis, N (%)	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Adrenal insufficiency, N (%)	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Total events, N	37	6	29	8	18	19	20	17	25	12	24	13	7	30

Abbreviations: CCI, Charlson Comorbidity Index; ECOG PS, Eastern Cooperative Group Performance Status; ICI, immune checkpoint inhibitors; irAE, immune-related adverse events

^aUnless otherwise stated, all comparisons were not statistically significant

^b28 patients (37%) experienced a total of 37 irAEs

^cRates of colitis were significantly different in patients age <80 and ≥80 years (P=0.04)

^dRates of pneumonitis were significantly different in patients with adenocarcinoma vs other histologies (P=0.03)

^eRates of thyroiditis were significantly different in patients being treated front line vs subsequent line (P=0.03)

Discussion:

Our study describes a “real world” evaluation of efficacy and toxicity of ICIs in older patients age ≥70 years with advanced-stage NSCLC, treated mostly in the second-line setting. Our results demonstrated that older patients derive similar benefits from ICIs, in particular those with ECOG PS 0-1. On the other hand, those with ECOG PS ≥2 had poor outcomes despite receipt of ICIs, which raises the question of whether ICIs should be offered to patients with poor PS, who were not captured in the seminal trials of ICIs leading to their approval. Our findings also demonstrated that neither chronological age nor comorbidity measured by CCI was associated with worse OS or TTF. Overall, ICIs were generally well-tolerated, with 37% of patients experiencing any irAE and 7% experiencing Grade ≥3 irAE.

The median OS in our study cohort (8.2 months) was lower than that previously reported in the CheckMate 017 and CheckMate 057 trials (12.2 and 9.2 months, respectively).^4,5 These results are not surprising given the higher number of patients with ECOG PS ≥2 in our study; moreover, when we focused only on patients with ECOG PS 0-1, the median OS (13.7 months) was similar to that reported in the landmark trials. In a separate study that included 30 patients age ≥70 years treated for advanced NSCLC at a single institution similar to the patient population in our study, median OS was 7.1 months.²⁶

In both CheckMate 017 and CheckMate 057, subgroup analyses by age revealed consistent benefits except for patients age ≥75 years in CheckMate 057 (Hazard ratio 1.76, 95% CI 0.77-4.05).^4,5 However, given the small number of patients age ≥75 years (N=29), the clinical implication of this finding is unclear. CheckMate 153 was a phase 3b/4 study that evaluated the safety of nivolumab in patients with previously treated NSCLC.²⁷ Of the 1308 patients included, 520 were age ≥70 years, and 8% of the total cohort had ECOG PS of 2. The efficacy of nivolumab in patients age <70 and ≥70 years was comparable, and the difference in outcomes was driven primarily by ECOG PS. Although our study included only patients age ≥70 years, we similarly found no difference in efficacy in patients age <80 compared to those ≥80 years. The similar outcomes regardless of age has also been demonstrated in other cancer types.²⁸ Overall, these data indicate that chronological age alone does not diminish the efficacy of ICI treatment in advanced stage NSCLC among older adults; rather, physiologic or functional age may be more useful to assess in decision-making regarding ICI treatment.

Approximately 37% of patients experienced any irAEs in our cohort, which was higher than that seen in the second-line pembrolizumab trial (Keynote-010; 20%), front line pembrolizumab monotherapy trial (Keynote-024; 29%), and front line pembrolizumab and chemotherapy trial (Keynote-189; 23%).^1,8,29 In two other studies of patients with NSCLC treated with nivolumab mostly in the second-line setting, incidences of irAEs of any grade were approximately 50% and Grade ≥3 irAEs were <10%.^26,30 Unfortunately, adverse events collected in landmark trials for nivolumab were not categorized based on irAEs.^4,5 The reported incidence of treatment-related adverse events of any grade was between 58-69%, and Grade ≥3 treatment-related adverse events were between 7-10% in CheckMate 017 and 057 trials.^4,5 We observed a higher rate of pneumonitis (12%) than that seen in the landmark trials (4-7%), which was also seen in a prior retrospective study of older patients, suggesting that older patients might be particularly susceptible to immune-related pneumonitis.^1,8,26,29,31 We also showed that the rates of irAEs generally did not differ significantly by age, CCI, and ECOG PS, and neither did rates of hospitalization, ICU admission, or falls during treatment. Prior studies have shown that, in general, the incidence of irAEs in patients treated with ICIs for NSCLC and other cancer types was no different or only slightly higher in older patients than in younger patients, although older patients may experience higher grades of irAEs.^28,31 Older patients may also be more likely to experience irAEs such as arthritis, endocrine toxicity, colitis, nephritis, and rash.^28,31

While the irAE profile in this study was comparable to that reported in previously published trials, we identified other toxicity variables that revealed the lower reserve and diminished resilience of older adults who undergo ICI treatment. Seventy-two percent of patients experienced at least one hospitalization while on ICI therapy (9% were due to irAEs), half experienced dose delay or interruption during treatment, and 16% discontinued treatment due to toxicity from ICI treatment compared to the 3-10% reported in the landmark trials.^5,32 Furthermore, a disproportionate number of older patients required medical intervention to manage their irAEs relative to what has been observed in previous clinical trials.⁴ The higher rate of treatment discontinuation in older patients that we observed is supported by a prior analysis showing a discontinuation rate of 17% in patients age ≥65 treated with ICIs across cancer types, which could have significant cost-effectiveness and treatment decision-making implications.³¹ Although none of our patients experienced fatal toxic effects from ICIs, a systematic review and meta-analysis did show that fatal toxic effects from ICIs increase with age.³³

Older adults undergoing cancer treatment have distinctive challenges that must be considered prior to proceeding with treatment, including poor functional status, changes in cognition, unique psychological and social factors, comorbidities, and reduced physiologic reserve.^34,35 The role of geriatric assessment (GA) has been gradually incorporated into clinical practice since introduction by the International Society of Geriatric Oncology (SIOG) in 2005.³⁶ GA has been shown to be instrumental in chemotherapy treatment decision-making for older adults with NSCLC.^9,37,38 In addition, GA may be able to better define older patients’ functional status (“physiologic age”) rather than reliance merely on age cut-offs (“chronologic age”).³⁹ Currently, GA is recommended to aid in such decision-making for older adults with cancer by multiple organizations including the SIOG, American Society of Clinical Oncology (ASCO), and National Comprehensive Cancer Network (NCCN).^36,40,41 Given the established utility of the GA and the increased risk of toxicity as demonstrated by our study as well as others, further study on GA in the setting of ICIs is needed to identify patients with low physiologic reserve who are at risk of experiencing functional decline and death from ICIs. There is also a need to increase enrollment of older patients onto randomized clinical trials, which are the gold standard of evidence-based medicine, and ASCO has developed recommendations to improve the evidence base for older patients with cancer.⁴² ASCO and Friends of Cancer Research have also developed recommendations on eligibility criteria with the goal of increasing participation of adults with comorbidities (which are more prevalent in older adults) without compromising safety.⁴³

Our study has several limitations. First, this is a retrospective study conducted in a single institution and therefore may not be generalizable. Nevertheless, our findings are generally similar to those previously reported in the literature and add information on the efficacy and toxicity of ICIs in older patients with NSCLC. Second, our small sample size also limited our ability to identify differences in toxicity or other outcomes in subgroup analysis, and thus our findings are largely hypothesis generating. Third, we may have underestimated the rates of irAEs and hospitalizations due to irAEs given the retrospective nature of our study. Finally, we did not collect GA in our patient population. Despite these weaknesses, our study has several strengths. To the authors’ knowledge, this is the largest retrospective cohort study reporting efficacy and toxicity data specifically in older patients treated with ICIs for advanced stage NSCLC in a “real-world” setting. In addition, we also evaluate outcome data not usually reported in clinical trials (e.g., falls and hospitalization rates during treatment).

In conclusion, our study demonstrated that older patients with advanced NSCLC derived similar benefits from ICI as reported in landmark trials evaluating relatively younger, fitter patients. Based on our results, overall survival in this patient population appears to be driven primarily by ECOG PS rather than chronological age or comorbidities. ICI treatment appears generally well tolerated in older patients, with a possible exception of a higher rate of pneumonitis and seemingly higher rates of glucocorticoids utilization for irAEs and treatment discontinuation due to toxicity. Prospective studies are needed to better capture the efficacy and toxicity of ICIs among older patients with advanced NSCLC and other cancer types. Such studies will better inform ICI-based treatment decision-making and supportive care interventions for these more vulnerable patients where the role of a geriatric assessment could be further defined.

Supplementary Material

Suppl Table1

Supplemental Table 1: Therapy-related outcomes stratified by age, sex, Charlson Comorbidity Index, Eastern Cooperative Group Performance Status, histology, and line of therapy

Impact statement:

We certify that this work is confirmatory of recent novel clinical research.^1-5 Our work confirms the efficacy of immune checkpoint inhibitors in older patients with Eastern Cooperative Group Performance Status 0-1, but also highlights the potential vulnerability of older adults with advanced NSCLC even in the immunotherapy era.