Sarcopenia and Visceral Adiposity Did Not Affect Efficacy of Immune-Checkpoint Inhibitor Monotherapy for Pretreated Patients With Advanced Non-Small Cell Lung Cancer.

BACKGROUND: This study aimed to investigate the association of computed tomography (CT)-assessed sarcopenia and visceral adiposity with efficacy and prognosis of immune-checkpoint inhibitor (ICI) therapy for pretreated non-small cell lung cancer (NSCLC).
METHODS: We retrospectively collected 74 patients with pretreated NSCLC who had initiated programmed cell death protein 1 (PD-1) or programmed cell death ligand 1 (PD-L1) inhibitor monotherapy between December 2015 and November 2018 at our hospital. As CT-assessed pretreatment markers, we used psoas muscle index (PMI), intramuscular adipose tissue content (IMAC), visceral-to-subcutaneous ratio (VSR) and visceral fat area (VFA) at lumbar vertebra L3 level. We divided 74 patients into high and low groups according to each Japanese sex-specific cut-off value. Using Kaplan-Meier curves and log-rank tests, we compared overall survival (OS) and progression-free survival (PFS). Adjusted by serum albumin, neutrophil-to-lymphocyte ratio, performance status and driver mutations, multivariate Cox proportional hazard analyses evaluated various variables as independent prognostic factors of OS and PFS.
RESULTS: We could not find significant difference in response rate (RR) and disease control rate (DCR) between low and high groups according to any factors. The OS of patients with body mass index (BMI) < 18.5 was significantly shorter than that of patients with BMI ≥ 18.5 (median 3.3 vs. 15.8 months, P < 0.01), while there was no significant difference in OS and PFS according to PMI, IMAC, VSR and VFA. Multivariate analyses detected no significant prognostic factor in OS and PFS, except for low IMAC (hazard ratio 0.43, 95% confidence interval 0.18 - 0.998, P = 0.0496) as a favorable prognostic factor of longer OS.
CONCLUSIONS: Neither PMI nor VSR, VFA might be a significant prognostic factor of PFS and OS of ICI monotherapy for pretreated NSCLC. According to our multivariate analyses, IMAC was a significant prognostic factor of OS, but not of PFS. Thus, neither sarcopenia nor visceral adiposity may be associated with the efficacy of ICI therapy. Copyright 2020, Minami et al.

Introduction

Sarcopenia is characterized by progressively decreased mass, strength and function of general skeletal muscle. This condition is a well-known prognostic factor of poor outcome in various solid cancers [1]. On the other hand, obesity also has adverse effects on cancer development, progression and prognosis [2]. Adipose tissue is distributed in the visceral fat area (VFA) and the subcutaneous fat area (SFA), which have different structural and functional characteristics. Owing to insulin metabolism disruptions, growth factors, sex hormones and chronic inflammation, excessive visceral adiposity is a well-established risk factor of tumorigenesis and cancer progression [3], while reduced subcutaneous adipose tissue is also independently associated with increased mortality and shorter survival in cancer patients [4]. In patients with metastatic melanoma, higher ratio of visceral to subcutaneous fat is associated with poorer survival benefits [5]. Thus, both sarcopenia and abnormal body fat distribution are significant prognostic factors for cancer patients. There are currently various methods for evaluating muscle mass and visceral fat accumulation, but most of them have not been standardized. Computed tomography (CT) and readily available software for image analysis have made it simplified to assess skeletal muscle and visceral fat. Among various CT-assessed muscle indexes, both psoas muscle index (PMI) and intramuscular adipose tissue content (IMAC) have been frequently used as parameters of skeletal muscle quantity and quality, respectively. For precise estimation of intra-abdominal fat distribution, quantitative CT determines visceral adiposity by measuring VFA or VFA/SFA ratio (VSR).

In the last decade, non-small cell lung cancer (NSCLC) has taken advantages of new treatment opportunities to improve survival benefits dramatically. For advanced NSCLC, in addition to conventional cytotoxic chemotherapy, new molecular targeted drugs and cancer immunotherapy have revolutionized treatment. Sarcopenia is still controversial as a prognostic factor of cytotoxic chemotherapy for patients with advanced NSCLC [6-9]. Skeletal muscle index (SMI), but not skeletal muscle radiodensity (SMD), was a significant prognostic factor in an Italian study [9]. On the contrary, SMD, but not SMI, was independently prognostic in a Norwegian study [7]. Thus, neither skeletal muscle quantity nor quality has been confirmed as a prognostic marker of cytotoxic chemotherapy. On the other hand, there was only one study that had investigated the association of sarcopenia with molecular-targeted therapy for advanced NSCLC. The Italian study failed to detect sarcopenia as a significant prognostic factor of epidermal growth factor receptor-tyrosine kinase inhibitor (EGFR-TKI), gefitinib [10]. Regarding cancer immunotherapy, little has been known about association of sarcopenia with immune-checkpoint inhibitors (ICI). Obesity, stratified by body mass index (BMI) based on World Health Organization criteria, may be a favorable prognostic factor of improved survival for patients with definitively treated locally advanced NSCLC [11]. On the other hand, another retrospective combined study of three Eastern Cooperative Oncology Group (ECOG) trials indicated a time effect while undergoing treatment. Compared with normal weight and overweight, obesity had superior outcomes earlier, but later increased risk [12]. Regarding body fat distribution, there is no study investigating contribution of visceral adiposity to survival disadvantage in NSCLC.

This study aimed to investigate the association of CT-assessed sarcopenia and visceral adiposity with efficacy and prognosis of ICI therapy.

Patients and Methods

Patients and study design

This study design was single-centered and retrospective. The inclusion criteria included: 1) pathological confirmation of NSCLC diagnosis; 2) pretreated and already advanced; 3) initiation of nivolumab, pembrolizumab or atezolizumab between December 2015 and November 2018 at our hospital; 4) CT scan covering L3 level within 90 days prior to the introduction of ICI; 5) pretreatment peripheral venous blood test within 2 weeks prior to the first day of ICI; 5) for patients harboring driver mutation or rearrangement of EGFR, anaplastic lymphoma kinase (ALK) and c-ros oncogene 1 (ROS1). Each specific TKI had already be administered until ICI therapy. The exclusion criteria included the patients with tumor proportion score (TPS) ≥ 50% who had initiated pembrolizumab in the first-line setting. Comparing KEYNOTE-024 trial [13] with KEYNOTE-010 trial [14] for patients with TPS ≥ 50%, the first-line pembrolizumab appeared to be superior in response to the second or later line pembrolizumab. LSI Medience Cooperation (Tokyo, Japan) examined EGFR mutation status by the peptide nucleic acid-locked nucleic acid PCR clamp method or EGFR gene mutation analysis COBAS version 2. The Department of Pathology at our hospital examined TPS of programmed cell death ligand 1 (PD-L1), using PD-L1 immunohistochemistry (IHC) 22C3 pharmDx test. We collected data of pretreatment backgrounds, including sex, age, BMI, smoking, histology, PD-L1 TPS, EGFR mutation, ALK and ROS1 rearrangement, ECOG performance status (PS) and number of metastatic sites, and of baseline blood examinations, including serum albumin concentration, the numbers of neutrophils and lymphocytes, and of ICI treatment, including regimen, response, progression-free survival (PFS) and overall survival (OS). In four patients, the mean value of the other patients complemented the missing values of serum albumin. None had missing values of neutrophils and lymphocytes. PFS was the interval between the first administration of ICI and documented progressive disease (PD) or death. OS was the interval between the first day of ICI regimen and death due to any causes. According to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1., response to ICI was determined. Response rate (RR) and disease control rate (DCR) were defined as the proportion of complete response (CR) + partial response (PR) in all patients, and as that of CR + PR + stable disease (SD) in all patients. The neutrophil-to-lymphocyte ratio (NLR) was formed by dividing absolute neutrophil count by lymphocyte count. The data were cut-off on February 8, 2019.

This study was conducted in agreement with the Declaration of Helsinki, and was approved by the Osaka Police Hospital Ethics Committee with waiver of the written informed consents in view of the retrospective and anonymous characteristics.

CT image analysis

Using pretreatment and cross-sectional CT at the level of transverse process of lumbar vertebra L3, the skeletal muscle and adipose tissue areas were investigated by SYNAPSE VINCENT software (Fujifilm Medical, Tokyo, Japan). Using the CT attenuation values, the bilateral psoas muscle area, subcutaneous and visceral adipose tissue areas were automatically identified (Fig. 1a). PMI (cm2/m2) was defined by normalizing psoas muscle area (cm2) for the square of the patient’s height (m2) [15]. VSR was calculated by dividing VFA by SFA [16]. The multifidus muscles area was estimated by manual tracing method (Fig. 1b). IMAC was calculated by dividing the mean CT attenuation value (HU) of the bilateral multifidus muscles by that (HU) of four points of subcutaneous fat away from major vessels [17]. Unlike the previous studies [17, 18], in which the mean CT value of four small circles on subcutaneous fat was used, we adopted the mean CT value of four points on subcutaneous fat, because even small circles could not be placed in some thin patients’ subcutaneous regions. Based on the previous studies, as Japanese sex-specific cut-offs for PMI, IMAC and VSR, we used PMI cut-offs of 6.36 cm2/m2 for men and 3.92 cm2/m2 for women [15], IMAC cut-offs of -0.358 for men and -0.229 for women [19] and VSR cut-offs of 1.33 in male and 0.93 in female [16]. Based on the Japanese criteria of obesity disease [20] and the cut-off value of obesity-related cardiovascular risk factor [21], we pre-defined VFA cut-off of 100 cm2, irrespective of gender, age and BMI.

Figure 1

Cross-sectional CT images at the third lumber vertebra level of a 75-year-old male patient. (a) The areas of bilateral psoas muscles (green area), visceral (red) and subcutaneous (blue) adipose tissue area were identified. (b) The multifidus muscle (yellow area) was precisely traced. Four points (yellow dots) were placed on subcutaneous fat away from major vessels. CT: computed tomography.

Data analysis

Continuous, categorical and survival data are presented as median with interquartile range (IQR), frequencies and median months with 95% confidential intervals (CIs), respectively. These variables were compared using non-parametrically the Mann-Whitney U test, Fisher’s exact test and Kaplan-Meier methods with log-rank test, respectively. According to the numbers of events in our study and the findings of previous studies, multivariate Cox proportional hazards analyses adjusted the sarcopenia-related variables of our interests by the following explanatory co-variables: serum albumin concentration (< 3.5 vs. ≥ 3.5 g/dL), NLR (< 5 vs. ≥ 5), ECOG-PS (0 - 1 vs. 2 - 4) and PD-L1 TPS (> 1% vs. 0% or not tested). The results of multivariate analyses were shown as hazard ratios (HRs) with 95% CI. A P value less than 0.05 was considered significant. All statistical analyses were performed with EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan) [22], which is a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria).

Results

We collected 74 NSCLC patients, and then divided them into sarcopenic and non-sarcopenic groups according to BMI, PMI, IMAC, VSR and VFA. Except for two patients who had been treated only with EGFR-TKIs as the front-line regimen, 72 patients had received platinum-based regimen before the ICI monotherapy. The ICI was administered as the second-line regimen in 37 patients, third-line in 11, fourth-line in eight and fifth or later line regimen in 18. Until the data cut-off, we confirmed 50 deaths, 19 survival, five lost to follow-up and 68 discontinuation of ICI therapy. Six patients still continued ICI therapy. The reasons of discontinuation of ICI therapy were PD in 50 patients, adverse effects in 10, deteriorated general conditions and complications in seven, respectively. One patient continued ICI even after documented PD. Our study included 11patients with positive EGFR mutation and one with ALK rearrangement. None had ROS1 rearrangement. All the patients with any driver mutation received ICI therapy after PD of TKI therapy. Tables 1-5 describe backgrounds, treatment and pretreatment laboratory data according to BMI, PMI, IMAC, VSR and VFA, respectively. We could not find significant difference in RR and DCR between low and high groups according to any factors.

Table 1

Baseline Characteristics According to BMI

	BMI		P
	≥ 18.5	< 18.5
N	60	14
Backgrounds
Sex (N)
Male/female	40/20	8/6	0.54a
Age (years)
Median (IQR)	70 (65.3 - 75.0)	69 (61.3 - 71.5)	0.38b
< 70/≥ 70 years (N)	28/32	9/5	0.37a
Smoking status
NS/Ex or CS	12/48	3/11	1.00a
Histology
Non-SQ/SQ	47/13	8/6	0.17a
EGFR mutation
(+)/(-) or NE	11/49	0/14	0.11a
PD-L1 TPS
≥ 50%/1-49%/< 1%/NA	11/11/7/31	2/4/1/7	0.86a
ECOG-PS
0 - 1/2/3	45/12/3	8/4/2	0.21a
Metastatic sites
< 3/≥ 3	26/34	9/5	0.24a
BMI
Median (IQR)	23.1 (21.9 - 25.3)	17.0 (15.7 - 17.8)	< 0.01b
Treatment
ICI regimen (N)
Nivo/Pem/Atezo	45/10/5	12/2/0	0.76a
Previous treatment (N)
Anti-angiogenic drugs	29	5	0.55a
Radiotherapy	14	5	0.33a
Further line treatment (N)	33	4	0.14a
ICI efficacy
CR/PR/SD/PD/NE	2/7/13/36/2	1/1/2/7/3
ORR (%) (95% CI)	15.0 (7.1 - 26.6)	14.3 (1.8 - 42.8)	1.00a
DCR (%) (95% CI)	36.7 (24.6 - 50.1)	28.6 (8.4 - 58.1)	0.76a
Laboratory data
NLR
Median (IQR)	2.81 (1.96 - 4.08)	4.96 (3.69 - 5.97)	< 0.01b
≤ 5/> 5 (N)	50/10	7/7	0.01a
Albumin (g/dL)
Median (IQR)	3.6 (3.3 - 3.9)	3.2 (2.8 - 3.2)	< 0.01b
≥ 3.5/< 3.5 g/dL (N)	42/18	3/11	< 0.01a

aFisher’s exact test. bMann-Whitney U test. Atezo: atezolizumab; BMI: body mass index; CI: confidence interval; CR: complete response; CS: current smoker; DCR: disease control rate; ECOG-PS: Eastern Cooperative Oncology Group performance status; EGFR: epidermal growth factor receptor; Ex: ex-smoker; ICI: immune-checkpoint inhibitors; IQR: interquartile range; NE: not evaluated; Nivo: nivolumab; NLR: neutrophil-to-lymphocyte ratio; NS: non-smoker; ORR: overall response rate; PD: progressive disease; PD-L1: programmed cell death ligand 1; Pem: pembrolizumab; PR: partial response; SD: stable disease; SQ: squamous cell carcinoma; TPS: tumor proportion score.

Table 2

Baseline Characteristics According to PMI

	PMI		P
	High	Low
N	21	53
Backgrounds
Sex (N)
Male/female	13/8	35/18	0.79a
Age (years)
Median (IQR)	70 (61 - 73)	69 (63 - 74)	0.61b
< 70/≥ 70 years (N)	10/11	27/26	1.00a
Smoking status
NS/Ex, CS	5/16	10/43	0.75a
Histology
Non-SQ/SQ	18/3	37/16	0.24a
EGFR mutation
(+)/(-) or NA	3/18	8/45	1.00a
PD-L1 status
≥ 50%/1-49%/< 1%/NA	3/5/0/13	10/10/8/25	0.25a
ECOG-PS
0 - 1/2/3	15/4/2	38/12/3	0.82a
Metastatic sites
< 3/≥ 3	8/13	27/26	0.44a
BMI
Median (IQR)	24.5 (21.5 - 27.7)	22.0 (19.8 - 23.8)	0.02a
PMI
Median (IQR)
Male	7.73 (6.82 - 7.85)	4.29 (3.49 - 5.60)	< 0.01b
Female	4.61 (4.22 - 4.88)	3.27 (2.99 - 3.49)	< 0.01b
Treatment
ICI regimen (N)
Nivo/Pem/Atezo	17/3/1	40/9/4	1.00a
Previous treatment (N)
Anti-angiogenic drug	12	22	0.30a
Radiotherapy	5	14	1.00a
Further line treatment (N)	13	24	0.30a
ICI efficacy
CR/PR/SD/PD/NE	1/0/5/15/0	2/8/10/28/5
ORR (%) (95% CI)	4.8 (0.1 - 23.8)	18.9 (9.4 - 32.0)	0.16a
DCR (%) (95% CI)	28.6 (11.3 - 52.2)	37.7 (24.8 - 52.1)	0.59a
Laboratory data
NLR
Median (IQR)	2.79 (2.00 - 3.66)	3.57 (2.10 - 4.88)	0.36b
≤ 5/> 5 (N)	17/4	40/13	0.76a
Albumin (g/dL)
Median (IQR)	3.6 (3.2 - 3.9)	3.5 (3.2 - 3.9)	0.67b
≥ 3.5/< 3.5 g/dL (N)	13/8	32/21	1.00a

aFisher’s exact test. bMann-Whitney U test. Atezo: atezolizumab; BMI: body mass index; CI: confidence interval; CR: complete response; CS: current smoker; DCR: disease control rate; ECOG-PS: Eastern Cooperative Oncology Group performance status; EGFR: epidermal growth factor receptor; Ex: ex-smoker; ICI: immune-checkpoint inhibitors; IQR: interquartile range; NE: not evaluated; Nivo: nivolumab; NLR: neutrophil-to-lymphocyte ratio; NS: non-smoker; ORR: overall response rate; PD: progressive disease; PD-L1: Programmed cell Death ligand 1; Pem: pembrolizumab; PMI: psoas muscle index; PR: partial response; SD: stable disease; SQ: squamous cell carcinoma; TPS: tumor proportion score.

Table 3

Baseline Characteristics According to IMAC

	IMAC		P
	High	Low
N	11	63
Backgrounds
Sex (N)
Male/female	6/5	42/21	0.50a
Age (years)
Median (IQR)	75 (70.5 - 77.5)	68 (62 - 73)	< 0.01b
< 70/≥ 70 years (N)	2/9	35/28	0.046a
Smoking status
NS/Ex, CS	3/8	12/51	0.68a
Histology
Non-SQ/SQ	7/4	48/15	0.46a
EGFR mutation
(+)/(-) or NA	2/9	9/54	0.66a
PD-L1 status
≥ 50%/1-49%/< 1%/NA	1/3/3/4	12/12/5/34	0.21a
ECOG-PS
0 - 1/2/3	6/3/2	47/13/3	0.17a
Metastatic sites
< 3/≥3	1/10	34/29	< 0.01a
BMI
Median (IQR)	23.4 (22.9 - 26.5)	22.0 (19.6 - 24.1)	0.05b
IMAC
Median (IQR) (minus data)
Male	0.33 (0.34 - 0.29)	0.54 (0.65 - 0.48)	< 0.01b
emale	0.16 (0.18 - 0.15)	0.41 (0.53 - 0.32)	< 0.01b
Treatment
ICI regimen (N)
Nivo/Pem/Atezo	9/1/1	48/11/4	0.73a
Previous treatment (N)
Anti-angiogenic drug	5	29	1.00a
Radiotherapy	5	14	0.14a
Further line treatment (N)	4	33	0.52a
ICI efficacy
CR/PR/SD/PD/NE	0/1/2/6/2	3/7/13/37/3
ORR (%) (95% CI)	9.1 (0.2 - 41.3)	15.9 (7.9 - 27.3)	1.00a
DCR (%) (95% CI)	27.3 (6.0 - 61.0)	36.5 (24.7 - 49.6)	0.74a
Laboratory data
NLR
Median (IQR)	3.89 (2.93 - 4.89)	2.98 (2.01 - 4.51)	0.41b
≤ 5/> 5 (N)	8/3	49/14	0.71a
Albumin (g/dL)
Median (IQR)	3.5 (3.1 - 4.0)	3.5 (3.2 - 3.9)	0.69b
≥ 3.5/< 3.5 g/dL (N)	6/5	39/24	0.74a

aFisher’s exact test. bMann-Whitney U test. Atezo: atezolizumab; BMI: body mass index; CI: confidence interval; CR: complete response; CS: current smoker; DCR: disease control rate; ECOG-PS: Eastern Cooperative Oncology Group performance status; EGFR: epidermal growth factor receptor; Ex: ex-smoker; ICI: immune-checkpoint inhibitors; IMAC: intramuscular adipose tissue content; IQR: interquartile range; NE: not evaluated; Nivo: nivolumab; NLR: neutrophil-to-lymphocyte ratio; NS: non-smoker; ORR: overall response rate; PD: progressive disease; PD-L1: Programmed cell Death ligand 1; Pem: pembrolizumab; PR: partial response; SD: stable disease; SQ: squamous cell carcinoma; TPS: tumor proportion score.

Table 4

Baseline Characteristics According to VSR

	VSR		P
	High	Low
N	20	54
Backgrounds
Sex (N)
Male/female	15/5	33/21	0.41a
Age (years)
Median (IQR)	74 (69.8 - 77.0)	68 (61.3 - 71.8)	< 0.01b
< 70/≥ 70 years (N)	5/15	32/22	0.02a
Smoking status
NS/Ex, CS	3/17	12/42	0.75a
Histology
Non-SQ/SQ	12/8	43/11	0.13a
EGFR mutation
(+)/(-) or NA	3/17	8/46	1.00a
PD-L1 status
≥ 50%/1-49%/< 1%/NA	1/5/2/12	12/10/6/26	0.38a
ECOG-PS
0 - 1/2/3	11/6/3	42/10/2	0.08a
Metastatic sites
< 3/≥ 3	11/9	24/30	0.45a
BMI
Median (IQR)	22.5 (20.4 - 24.0)	22.6 (20.1 - 25.2)	0.84a
VSR
Median (IQR)
Male	1.87 (1.48 - 3.00)	0.80 (0.52 - 1.07)	< 0.01b
Female	1.07 (1.01 - 1.15)	0.35 (0.19 - 0.56)	< 0.01b
Treatment
ICI regimen (N)
Nivo/Pem/Atezo	17/1/2	40/11/3	0.20a
Previous treatment (N)
Anti-angiogenic drug	9	25	1.00a
Radiotherapy	7	12	0.37a
Further line treatment (N)	9	28	0.79a
ICI efficacy
CR/PR/SD/PD/NE	0/1/3/13/3	3/7/12/30/2
ORR (%) (95% CI)	5.0 (0.1 - 24.9)	18.5 (9.3 - 31.4)	0.27a
DCR (%) (95% CI)	20.0 (5.7 - 43.7)	40.7 (27.6 - 55.0)	0.11a
Laboratory data
NLR
Median (IQR)	4.37 (2.39 - 5.93)	3.00 (2.01 - 3.97)	0.13b
≤ 5/> 5 (N)	12/8	45/9	0.06a
Albumin (g/dL)
Median (IQR)	3.5 (3.2 - 3.8)	3.6 (3.2 - 4.0)	0.25b
≥ 3.5/< 3.5 g/dL (N)	11/9	34/20	0.60a

aFisher’s exact test. bMann-Whitney U test. Atezo: atezolizumab; BMI: body mass index; CI: confidence interval; CR: complete response; CS: current smoker; DCR: disease control rate; ECOG-PS: Eastern Cooperative Oncology Group performance status; EGFR: epidermal growth factor receptor; Ex: ex-smoker; ICI: immune-checkpoint inhibitors; IQR: interquartile range; NE: not evaluated; Nivo: nivolumab; NLR: neutrophil-to-lymphocyte ratio; NS: non-smoker; ORR: overall response rate; PD: progressive disease; PD-L1: programmed cell death ligand 1; Pem: pembrolizumab; PR: partial response; SD: stable disease; SQ: squamous cell carcinoma; TPS: tumor proportion score; VSR: visceral to subcutaneous adipose tissue area ratio.

Table 5

Baseline Characteristics According to VFA

	VFA		P
	< 100 cm2	≥ 100 cm2
N	44	30
Backgrounds
Sex (N)
Male/female	18/26	8/22	0.23a
Age (years)
Median (IQR)	68.5 (61.8 - 72.3)	71.0 (66.5 - 76.5)	0.04b
< 70/≥ 70 years (N)	26/18	11/19	0.10a
Smoking status
NS/Ex, CS	10/34	5/25	0.57a
Histology
Non-SQ/SQ	33/11	22/8	1.00a
EGFR mutation
(+)/(-) or NA	6/38	6/24	0.53a
PD-L1 status
≥ 50%/1-49%/< 1%/NA	7/9/3/25	6/6/5/13	0.49a
ECOG-PS
0 - 1/2/3	33/9/2	20/7/3	0.57a
Metastatic sites
< 3/≥ 3	24/20	11/19	0.16a
BMI
Median (IQR)	20.8 (18.0 - 22.8)	24.8 (23.0 - 27.3)	< 0.01a
VFA (cm2)
Median (IQR)	36.4 (18.1 - 57.3)	139.5 (119.5 - 165.9)	< 0.01a
Treatment
ICI regimen (N)
Nivo/Pem/Atezo	37/7/3	23/5/2	1.00a
Previous treatment (N)
Anti-angiogenic drug	23	21	0.24a
Radiotherapy	8	11	0.10a
Further line treatment (N)	19	18	0.24a
ICI efficacy
CR/PR/SD/PD/NE	3/6/6/26/3	0/2/9/17/1
ORR (%) (95%CI)	20.5 (9.8 - 35.3)	6.7 (0.8 - 22.1)	0.18a
DCR (%) (95%CI)	34.1 (20.5 - 49.9)	36.7 (19.9 - 56.1)	0.81a
Laboratory data
NLR
Median (IQR)	3.13 (2.28 - 4.92)	3.00 (1.88 - 4.42)	0.72b
≤ 5/> 5 (N)	33/11	24/6	0.78a
Albumin (g/dL)
Median (IQR)	3.5 (3.2 - 3.9)	3.6 (3.2 - 3.9)	0.68b
≥ 3.5/< 3.5 g/dL (N)	25/19	20/10	0.47a

aFisher’s exact test. bMann-Whitney U test. Atezo: atezolizumab; BMI: body mass index; CI: confidence interval; CR: complete response; CS: current smoker; DCR: disease control rate; ECOG-PS: Eastern Cooperative Oncology Group performance status; EGFR: epidermal growth factor receptor; Ex: ex-smoker; ICI: immune-checkpoint inhibitors; IQR: interquartile range; NE: not evaluated; Nivo: nivolumab; NLR: neutrophil-to-lymphocyte ratio; NS: non-smoker; ORR: overall response rate; PD: progressive disease; PD-L1: programmed cell death ligand 1; Pem: pembrolizumab; PR: partial response; SD: stable disease; SQ: squamous cell carcinoma; TPS: tumor proportion score; VSR: visceral to subcutaneous adipose tissue area ratio.

Comparisons of OS and PFS according to BMI, PMI, IMAC, VSR and VFA are shown in Figures 2 and 3, respectively. The OS of patients with BMI < 18.5 was significantly shorter than that of patients with BMI ≥ 18.5 (median 3.3 vs. 15.8 months, P < 0.01), while there was no significant difference according to PMI, IMAC, VSR and VFA (Fig. 2). No significant difference was observed in PFS according to any variables (Fig. 3). Excluding four non-smokers with EGFR mutation (N = 3) or ALK rearrangement (N = 1), we could not find any significant differences in OS and PFS according to BMI, PMI, IMAC, VSR and VFA, except for OS according to BMI (Supplementary Materials 1 and 2, www.wjon.org). Multivariate Cox proportional hazard analyses detected low IMAC (HR 0.43, 95% CI 0.18 - 0.998, P = 0.0496) as a favorable prognostic factor of longer OS, while no significant prognostic factor was detected in PFS analyses (Table 6). Excluding four non-smokers with any driver mutations, we could not find any significant variables as a prognostic factor in the multivariate analyses (Supplementary Material 3, www.wjon.org).

Figure 2

Kaplan-Meier curves of overall survival according to BMI (a), PMI (b), IMAC (c), VSR (d) and VFA (e). BMI: body mass index; PMI: psoas muscle index; IMAC: intramuscular adipose tissue content; VSR: visceral to subcutaneous adipose tissue area ratio; VFA: visceral fat area.

Figure 3

Kaplan-Meier curves of progression-free survival according to BMI (a), PMI (b), IMAC (c), VSR (d) and VFA (e). BMI: body mass index; PMI: psoas muscle index; IMAC: intramuscular adipose tissue content; VSR: visceral to subcutaneous adipose tissue area ratio; VFA: visceral fat area.

Table 6

Adjusted Hazard Ratios of Sarcopenic Factors for Overall Survival and Progression-Free Survival by Multivariate Cox Proportional Hazard Analyses

Variables	Progression-free survival		Overall survival
	HR (95% CI)	P	HR (95% CI)	P
BMI (kg/m2)
≥ 18.5	1 (Reference)		1(Reference)
< 18.5	1.11 (0.55 - 2.24)	0.77	1.83 (0.79 - 4.21)	0.16
PMI (cm2/m2)
High	1 (Reference)		1(Reference)
Low	0.84 (0.48 - 1.48)	0.56	1.05 (0.54 - 2.03)	0.88
IMAC
High	1 (Reference)		1(Reference)
Low	0.79 (0.38 - 1.67)	0.54	0.43 (0.18 - 0.998)	0.0496
VSR
High	1 (Reference)		1(Reference)
Low	0.78 (0.42 - 1.46)	0.44	0.72 (0.35 - 1.49)	0.37
VFA
< 100 cm2	1 (Reference)		1 (Reference)
≥ 100 cm2	1.09 (0.64 - 1.84)	0.75	1.22 (0.67 - 2.20)	0.52

Multivariate adjustment for serum albumin (< 3.5 vs. ≥ 3.5 g/dL), neutrophil to lymphocyte ratio (< 5 vs. ≥ 5), ECOG-PS (0 - 1 vs. 2 - 4) and driver mutation of EGFR or ALK (wild-type vs. mutated). BMI: body mass index; CI: confidence interval; HR: hazard ratio; IMAC: intramuscular adipose tissue content; PMI: psoas muscle index; VFA: visceral fat area; VSR: visceral to subcutaneous adipose tissue area ratio; ECOG-PS: Eastern Cooperative Oncology Group performance status; EGFR: epidermal growth factor receptor; ALK: anaplastic lymphoma kinase.

Discussion

Our study investigated whether pretreatment sarcopenia and visceral adiposity were practically predictive or prognostic markers of efficacy and survival benefit of PD-1/PD-L1 inhibitor monotherapy for pretreated NSCLC patients. As a result, we failed to find a significant association of sarcopenia and visceral adiposity with the benefits of PD-1 or PD-L1 inhibitor therapy.

The most important finding of our study was that neither PMI nor IMAC was a significant predictive marker of ICI monotherapy for pretreated NSCLC patients. This finding was contrary to that of the previous study [23]. To our knowledge, only two small sample-sized studies by Shiroyama (N = 42) [23] and by Cortellini (N = 23) [24] evaluated PMI or SMI at the third lumber vertebra (L3) with PFS of PD-1 inhibitors for pretreated and advanced NSCLC, respectively. In the former Japanese study, the comparison between sarcopenic (low PMI) and non-sarcopenic (high PMI) patients detected significant differences in PFS, overall response rate and 1-year PFS rate [23]. However, probably owing to the small sample size, the latter Italian hypothesis-generating preliminary report failed to demonstrate significant differences in PFS and OS between low and high SMI [24]. Unlike these two studies, our study included five patients treated with PD-L1 inhibitor, atezolizumab, and more sarcopenic patients with low skeletal muscle quantity (72% of low PMI in our study vs. 52% of low PMI in Shiroyama’s study, and 34.6% of low SMM in Cortellini’s study). In our study, neither skeletal quantity nor quality was significantly associated with PFS of ICI therapy. However, our multivariate analysis detected skeletal selected quality of IMAC as a significant prognostic factor of OS. Thus, our study suggested that skeletal quality might be more important for ICI therapy than skeletal quantity.

Interestingly, our study failed to show a significant association of visceral adiposity shown by VSR and VFA with efficacy and survival benefits of ICI therapy. To our knowledge, this was the first study that had investigated the association of VSR and VFA with ICI therapy. It requires further investigations whether visceral adiposity is a predictive or prognostic marker of ICI therapy.

There were some study limitations in our study. First, our study design was retrospective and single-institutional. Our sample size might be too small to detect any sarcopenic variables as significant factors of outcomes and survival benefits of ICI therapy. Thus, we could not deny bias and decreased validity in our results. Second, our CT scans were plain in some patients and enhanced in the other patients. In the previous Japanese studies of patients with non-alcoholic fatty liver disease [18] and patients undergoing living donor liver transplantation [17], IMAC was measured only by plain abdominal CT scan. Thus, our IMAC might be different from the conventional and standard IMAC. Unless contraindicated, enhanced CT scan is more frequently taken than plain CT scan in the management and follow-up of advanced cancer patients. In practice, we have to use different CT imaging method according to conditions of each cancer patient.

Conclusion

Neither PMI nor VSR, VFA might be a significant prognostic factor of PFS and OS of ICI monotherapy for pretreated NSCLC. According to our multivariate analyses, IMAC was a significant prognostic factor of OS, but not of PFS. Thus, neither sarcopenia nor visceral adiposity may be associated with the efficacy of ICI therapy.

Kaplan-Meier curves of overall survival according to BMI (a), PMI (b), IMAC (c), VSR (d) and VFA (e) in 70 ex-, current or unknown smokers without driver mutation (N = 70).

Click here for additional data file.

Kaplan-Meier curves of progression-free survival according to BMI (a), PMI (c), IMAC (c), VSR (d) and VFA (e) in 70 ex-, current or unknown smokers without driver mutation (N = 70).

Click here for additional data file.

Adjusted Hazard Ratios of Sarcopenic Factors for Overall Survival and Progression-Free Survival by Multivariate Cox Proportional Hazard Analyses, When Four Non-Smokers With Driver Mutations Were Excluded (N = 70).

Click here for additional data file.