Predictive value of baseline metabolic tumor volume for non-small-cell lung cancer patients treated with immune checkpoint inhibitors: A meta-analysis.

Background: Immune checkpoint inhibitors (ICIs) have emerged as a promising treatment option for advanced non-small-cell lung cancer (NSCLC) patients, highlighting the need for biomarkers to identify responders and predict the outcome of ICIs. The purpose of this study was to evaluate the predictive value of baseline standardized uptake value (SUV), metabolic tumor volume (MTV) and total lesion glycolysis (TLG) derived from 18F-FDG-PET/CT in advanced NSCLC patients receiving ICIs.
Methods: PubMed and Web of Science databases were searched from January 1st, 2011 to July 18th, 2022, utilizing the search terms "non-small-cell lung cancer", "PET/CT", "standardized uptake value", "metabolic tumor volume", " total lesion glycolysis", and "immune checkpoint inhibitors". Studies that analyzed the association between PET/CT parameters and objective response, immune-related adverse events (irAEs) and prognosis of NSCLC patients treated with ICIs were included. We extracted the hazard ratio (HR) with a 95% confidence interval (CI) for progression-free survival (PFS) and overall survival (OS). We performed a meta-analysis of HR using Review Manager v.5.4.1.
Results: Sixteen studies were included for review and thirteen for meta-analysis covering 770 patients. As for objective response and irAEs after ICIs, more studies with consistent assessment methods are needed to determine their relationship with MTV. In the meta-analysis, low SUVmax corresponded to poor PFS with a pooled HR of 0.74 (95% CI, 0.57-0.96, P=0.02). And a high level of baseline MTV level was related to shorter PFS (HR=1.45, 95% CI, 1.11-1.89, P<0.01) and OS (HR, 2.72; 95% CI, 1.97-3.73, P<0.01) especially when the cut-off value was set between 50-100 cm3. SUVmean and TLG were not associated with the prognosis of NSCLC patients receiving ICIs. Conclusions: High level of baseline MTV corresponded to shorter PFS and OS, especially when the cut-off value was set between 50-100 cm3. MTV is a potential predictive value for the outcome of ICIs in NSCLC patients.

1 Introduction

Lung cancer is the most common cause of cancer-related deaths worldwide in 2020, accounting for 1.80 million deaths (1). Non-small-cell lung cancer (NSCLC), compromising 80-85% of the lung cancer cases (2), has raised significant public health concerns. NSCLC is mainly composed of squamous cell carcinoma and adenocarcinoma (3), and the 5-year survival rate is 25% (4).Clinically, more than 60% of NSCLC patients had locally progressed or metastatic diseases (stage III or IV) at the time of diagnosis, when the tumor can not be effectively treated by surgical treatment alone (5), and the median overall survival varies between 7.0 and 12.2 months (6).

For the treatment of advanced NSCLC, chemotherapy remains the primary conventional therapy. But the response rate of NSCLC patients to chemotherapy was only about 20% (7), and the adverse events such as vomiting and diarrhea had a significant impact on patients’ daily lives. The advent of immune checkpoint inhibitors (ICIs) targeting programmed cell death 1 (PD-1) or its ligand (PD-L1) has brought about a promising treatment option for the management of advanced NSCLC (8). A meta-analysis of 13 randomized controlled trials (RCTs) has proved that ICIs show better efficacy and result in fewer adverse events than chemotherapy as the treatment for advanced NSCLC (9). However, the benefits of ICIs remain limited to only 20% of advanced NSCLC patients (10). Thus it’s necessary to identify potential biomarkers to identify NSCLC patients who would benefit from ICIs treatment.

18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) monitors the uptake of 18F-FDG of tumor cells. It is a convenient imaging modality for the staging, treatment guidance, and response predicting in NSCLC patients and is more practical and noninvasive than abdominal ultrasound and mediastinoscopy (11).

As metabolic parameters on PET/CT, SUV is associated with 18F-FDG uptake of the tumor; MTV combines the information of 18F-FDG uptake and tumor volume; TLG is the product of MTV and SUVmean, and is related to both tumor volume and tumor glycolytic activity. They reflect both tumor burden and aggressiveness (12). Takada et al. found that the accumulation of 18F-FDG as SUVmax and SUVmean in tumor cells was significantly associated with PD-L1 expression in NSCLC patients (13). In addition, MTV and TLG have been potential prognostic factors in NSCLC patients treated with surgery (14) and chemotherapy (15). Thus SUV, MTV and TLG are expected to evaluate the efficacy of ICIs in advanced NSCLC patients. However, relevant studies showed inconclusive results. Monaco et al. have demonstrated that NSCLC patients with MTV and TLG values lower than the median values had improved outcomes of ICIs compared to those with higher values (16). No significant relationship was found between MTV, TLG, and ICIs response in studies conducted by Yamaguchi et al. (17) and Castello et al. (18).

Thus, we conducted this meta-analysis to assess the predictive value of SUV, MTV and TLG for advanced NSCLC patients receiving ICIs.

2 Material and methods

2.1 Data search and study selection

From January 1st, 2011, to July 18th, 2022, We searched comprehensively English language publications from PubMed and Web of Science using the terms “non-small-cell lung cancer”, “PET/CT”, “Standardized uptake value”, “metabolic tumor volume”, “ total lesion glycolysis”, and “immune checkpoint inhibitors”. We extracted data from the full-text articles that met the following inclusion criteria: studies limited to NSCLC; ICIs administered alone for the patients; 18F-FDG PET/CT completed before ICIs initiation; studies reported objective response, immune-related adverse events (irAEs), survival data, including progression-free survival (PFS) or overall survival (OS); hazard ratio (HR) with 95% CI was provided for PFS or OS. Reviews, meeting abstracts, and editorial material were excluded. Two authors conducted the searches and screening independently. A consensus resolved any discrepancies.

2.2 Data extraction

Data were extracted from the publications independently by two reviewers (YC and CC), and the following information was recorded: first author’s name, year of the paper published, country, types of ICIs, median follow-up, number of patients, median age of patients, median values of MTV and TLG, HR and p-value for PFS and OS. The data were collected and organized in a standardized data extraction table for analysis. We also formed a table including median values of MTV or numbers of patients in different objective response groups and the related p-value to demonstrate the relationship between MTV and objective response. When there was uncertainty in the inclusion of data, a third researcher assisted with confirming the data.

2.3 Quality assessment

We used ROBINS-I (Risk Of Bias In Non-randomised Studies - of Interventions) to assess the quality of included articles from seven bias domains, including confounding bias, selection bias, bias due to classification of interventions, bias from intended interventions, bias due to missing data, bias in outcomes measure, and bias due to selection reporting result. We classified each article as low, moderate, or high risk according to detailed guidance from ROBINS-I (19).

2.4 Statistics analysis

We performed all statistical analyses using Review Manager v.5.4.1 and pooled the hazard ratio (HR) and its 95% confidence index (CI) of PFS and OS using the inverse variance method. An HR greater than 1 indicated worse survival for patients with high SUV, MTV or TLG, while an HR less than 1 indicated a better survival for patients with a high SUV, MTV or TLG. Chi-square test and I statistics were used to detect heterogeneity between studies. I values of more than 50% were considered high heterogeneity. If high heterogeneity was found between primary studies, a random effect model would be used for meta-analysis. Otherwise, a fixed effect model would be applied. P values less than 0.05 were considered statistically significant.

3 Results

3.1 Literature search

Eight hundred and sixteen studies were retrieved from the systematic search of PubMed and Web of Science from January 1st, 2011, to July 18th, 2022. We excluded 134 duplicate studies and further screened the remaining 682 using titles and abstracts. 641 studies did not meet the inclusion criteria and thus were excluded. The full texts of the 41 potentially eligible studies were evaluated. Then 25 studies were excluded for the following reasons: not single ICIs as treatment (n=3), no available data (n = 19), and overlapped data (n=3). Ultimately, sixteen studies were included for review and thirteen studies assessing the predictive value of SUV, MTV and TLG in NSCLC patients receiving ICIs were included in this meta-analysis. shows the flowchart diagram.

Figure 1

Flowchart diagram for the literature search.

3.2 Characteristics of included studies

The thirteen articles, including 770 patients, were analyzed in this meta-analysis. Characteristics of the included studies are summarized in . Four studies were conducted in France (20, 22, 26, 27), followed by three in Italy (16, 18, 21) and three in Japan (16, 18, 21). We also identified a single study in the United States (19), Israel (24) and Belgium (28). Two studies were of a prospective design (18, 20). SUV, MTV and TLG were measured in four studies (18, 20, 25, 28) and MTV alone was measured in five studies (17, 19, 21, 22, 24).

Table 1

Characteristics and results of included studies.

Studies	Year	Country	Study design	Types of ICIs	Median follow-up	No. of patients	Median Age	Median values as cut-offs	Outcome
									HR (95% CI) for PFS	p value	HR (95% CI) for OS	p value
Andraos et al. (19)	2022	USA	R	–	17.0 months	124	67	MTV: 87.8(cut-offs: 88.0)	1.36 (0.91-2.01)	0.131	2.23 (1.35-3.69)	0.002
Castello et al. (18)	2021	Italy	P	Nivolumab/pembrolizumab/atezolizumab	12.4 months	50	73	SUVmax: 13.6	0.9 (0.5-1.8)	0.75	0.9 (0.4-2.0)	0.75
								SUVmean: 5.9	0.9 (1.0-1.7)	0.75	0.8 (0.4-1.9)	0.71
								MTV: 63.7	2.5 (1.2-4.8)	0.01	2.3 (1.0-5.3)	0.04
								TLG: 330.1	1.8 (0.9-3.6)	0.08	1.5 (0.7-3.6)	0.27
Chardin et al. (20)	2020	France	P	Nivolumab/pembrolizumab	12.3 months	79	64	SUVmax: 13.4	–	–	1.31 (0.63-2.75)	0.5
								SUVpeak: 9.7	–	–	1.15 (0.55-2.40)	0.7
								MTV:36.5	–	–	5.37 (2.17-13.3)	<0.0001
								TLG: 267.0	–	–	5.05 (2.05-12.5)	0.0001
Dall’Olio et al. (21)	2021	Italy	R	Pembrolizumab	20.3 months	34	66.6	MTV: 75.0	–	–	5.37 (1.72-16.77)	0.004
Eude et al. (22)	2022	France	R	Pembrolizumab	–	65	64.1	MTV: 188.3	–	–	1.314	0.012
Hashimoto et al. (23)	2020	Japan	R	Nivolumab/pembrolizumab	–	85	–	MTV: 17.8(cut-offs: 5.0)	1.28 (0.97-1.73)	0.07	1.59 (1.09-2.45)	0.001
								TLG: 75.4(cut-offs: 20.0)	1.21 (0.92-1.63)	0.16	1.47 (1.03-2.21)	0.03
Icht et al. (24)	2020	Israel	R	Nivolumab/pembrolizumab	–	58	65	MTV:12.95	1.1 (0.87-1.4)	0.4	1.2 (0.86-1.73)	0.26
Kitajima et al. (25)	2021	Japan	R	Nivolumab/pembrolizumab	36.8 months	40	69.1	SUVmax: 8.57	1.04 (0.49-2.18)	0.92	1.56 (0.67-3.69)	0.3
								MTV: 15.5	2.15 (1.03-4.73)	0.042	2.15 (1.03-4.73)	0.042
								TLG: 87.7	1.15 (0.55-2.42)	0.7	1.35 (0.59-3.13)	0.47
Monaco et al. (16)	2021	Italy	R	Nivolumab/pembrolizumab/atezolizumab	–	92	70	SUVmean: 4.9	0.365 (0.150-0.890)	0.027	0.261 (0.084-0.808)	0.02
								MTV: 94.9	1.139 (0.989-1.311)	0.07	1.221 (1.063-1.402)	0.005
Seban et al. (26)	2019	France	R	Nivolumab/pembrolizumab/atezolizumab	11.6 months	80	61.9	SUVmax: 12.8	0.8 (0.5-1.3)	0.35	0.9 (0.5-1.5)	–
								MTV: 75.0	1.0 (0.9-1.1)	0.25	3.1 (1.7-5.7)	0.0001
Seban et al. (27)	2020	France	R	Pembrolizumab	13.4 months	63	65	SUVmax: 18	0.6 (0.3-1.1)	0.11	0.6 (0.2-1.6)	0.31
								SUVmean: 10.1	0.5 (0.3-1.1)	0.04	0.8 (0.3-1.9)	0.56
								MTV: 84.0	2.1 (1.1-4.3)	0.02	3.1 (1.1-8.3)	0.03
Vekens et al. (28)	2021	Belgium	R	Pembrolizumab	20 months	30	67	SUVmax: 15.7	0.62 (0.39-0.98)	0.04	0.54 (0.29-1.01)	0.06
								SUVpeak: 10.2	1.43 (0.97-2.11)	0.07	1.71 (0.97-3.03)	0.06
								SUVmean: 6.06	1.76 (0.54-5.79)	0.35	1.51 (0.46-4.93)	0.5
								MTV: 123.9	1.01 (0.99-1.03)	0.25	1.01 (0.99-1.02)	0.29
								TLG: 802.6	0.99 (0.99-1.00)	0.29	0.99 (0.99-1.00)	0.42
Yamaguchi et al. (17)	2020	Japan	R	Pembrolizumab	346 days	48	69	MTV: 112.0(cut-offs: 268.0)	1.49 (0.77-3.24)	0.32	1.57 (0.98-2.41)	0.04

ICIs, immune checkpoint inhibitors; SUV, standardized uptake value; MTV, metabolic tumor volume; TLG, total lesion glycolysisl; R, retrospective; P, prospective; PFS, progression-free survival; OS, overall survival.

Regarding types of ICIs, nine studies (17, 20–25, 27, 28) reported using PD-1 inhibitors, while three used PD-1 and PD-L1 inhibitors (16, 18, 26). Patients were divided into high or low SUV/MTV/TLG groups in each study based on the cut-off values, and their PFS/OS were analyzed. And eleven of the thirteen studies used median MTV/TLG as cut-off values (16, 18–22, 24–28). The left two used log-rank test (23) and receiver operating characteristic (ROC) curve analysis (17) to determine cut-off values, respectively.

3.3 Quality assessment

We used the Cochrane collaboration tool to assess the risk of bias in included studies. The risks of the selected studies are shown in . As shown, the overall risk of bias was relatively low, and the overall quality met the requirements of the meta-analysis.

Figure 2

Results of quality assessment.

3.4 Outcomes of included studies

3.4.1 PET/CT parameters and response assessment

Eight studies discussed whether PET/CT parameters including MTV, TLG, SUVmax, SUVmean, and SUVpeak, can predict the response of ICIs in different patients. All of them classified responses to ICIs as complete remission (CR), partial response (PR), stable disease (SD), and progression of disease (PD) based on the Response Evaluation Criteria In Solid Tumors (RECIST) 1.1. Four articles demonstrated that none of the PET/CT parameters significantly correlated with ICIs response (17, 23, 28, 29). However, the other four studies showed that NSCLC patients who achieved CR, PR, or SD after ICIs treatment had significantly lower median MTV values than those with PD (16, 26, 27, 30). The detailed data were shown in . In addition, Seban et al. found that SUVmean was significantly higher in patients who achieved long-term benefit (LTB, defined as CR, PR or SD maintained ±12 months) compared to those without LTB (27), while Polveri et al. concluded that TLG was significantly associated with progressive vs non-progressive disease status (30).

Table 2

MTV values and objective response (RECIST 1.1).

Authors	Published year	CR+PR+SD group		PD group		P value
		value	Number of patients	value	Number of patients
Median value of MTV
Ferrari et al. (29)	2021	203.0	15	–	13	0.387
Monaco et al. (16)	2021	77	61	160.2	31	0.039
Polvari et al. (30)	2020	57.4	27	124.4	30	0.028
Seban et al. (26)	2019	55.4	32	83.4	48	0.04
Seban et al. (27)	2020	59.4	17	90.5	46	0.05
Vekens et al. (28)	2021	192.8	23	119.8	7	0.17
Number of patients in high/low MTV group
Hashimoto et al. (23)	2020	High MTV: 36Low MTV: 17		High MTV: 18Low MTV: 9		>0.99
Yamaguchi et al. (17)	2020	High MTV: 3Low MTV: 20		High MTV: 7Low MTV: 15		0.16

CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; MTV, metabolic tumor volume; Bold means statistically significant.

3.4.2 PET/CT parameters and immune-related adverse events (irAEs)

Two studies discussed the relationship between PET/CT parameters and irAEs. In the analysis of Mu et al. (31), SUVmax and MTV were not correlated with irAEs, with the odds ratio of 0.95 (95%CI, 0.87-1.05, P=0.34) and 0.99 (95%CI, 0.98-1.00, P=0.27), respectively. However, Hashimoto et al. (23) reported that the frequency of irAE was significantly higher in patients with low values of SUVmax, MTV, and TLG than in those with high values, inconsistent with the result of Mu et al.

3.4.3 PET/CT parameters and NSCLC survival

3.4.3.1 SUVmax and NSCLC survival

Six studies (18, 20, 25–28) analyzed the relationship between SUVmax and PFS/OS, as shown in . The cut-off values of SUVmax ranged from 8.57 to 18 cm3. Five studies analyzing PFS showed a pooled HR of 0.74 (95% CI, 0.57-0.96, P=0.02). However, SUVmax was not significantly associated with OS (HR, 0.89; 95% CI, 0.64-1.23, P=0.48). There was no significant heterogeneity between studies in both PFS (I2 = 0%, P=0.72) and OS group (I2 = 13%, P=0.33).

Figure 3

Forest plots of hazard ratios comparing progression free survival (A) or overall survival (B) of patients with high level versus low level max standardized uptake value treating with immune checkpoint inhibitors.

3.4.3.2 SUVmean and NSCLC survival

We performed SUVmean and survival analysis based on four studies (16, 18, 27, 28) with cut-off values between 4.9 and 10.1cm3 ( ). SUVmean was not associated with either PFS (HR, 0.67; 95% CI, 0.39-1.16, P=0.15) or OS (HR, 1.11; 95% CI, 0.65-1.19, P=0.69). The heterogeneity test didn’t show significant heterogeneity in PFS (I2 = 53%, P=0.1) and OS group (I2 = 19%, P=0.29).

Figure 4

Forest plots of hazard ratios comparing progression free survival (A) or overall survival (B) of patients with high level versus low level mean standardized uptake value treating with immune checkpoint inhibitors.

3.4.3.3 MTV and NSCLC survival

Thirteen studies (16–28) analyzed the relationship between MTV and PFS/OS, as shown in . The cut-off values of MTV ranged from 5.0 to 268.0cm3, so we performed a subgroup analysis based on the cut-off values, dividing them into three groups: MTV < 50cm3, MTV between 50-100cm3, and MTV >100cm3.

Figure 5

Forest plots of hazard ratios comparing progression free survival (A) or overall survival (B) of patients with high level versus low level metabolic tumor volume treating with immune checkpoint inhibitors.

In eleven studies analyzing PFS, a pooled HR of 1.21 (95% CI, 1.06-1.36, P<0.01) was shown. There was statistically significant heterogeneity between studies, with an I2 of 79.4% (P<0.01). It is also demonstrated that patients with higher MTV would have shorter PFS (HR=1.45, 95% CI, 1.11-1.89, P<0.01) when the cut-off values was set at 50-100cm3. There was no evidence of a significant association between MTV and PFS in the other two subgroups.

OS was analyzed in thirteen MTV studies. The pooled HR was 1.67 (95% CI, 1.36-2.06, P<0.01) with statistically significant heterogeneity between studies (I2 = 84%, P<0.01). High MTV was significantly associated with poor OS, with an HR of 1.90 (95% CI, 1.15-3.15, P=0.01) and 2.35 (95% CI, 1.43-3.87, P<0.01) when the cut-off value was set below 50cm3 and 50-100cm3, respectively. The left subgroup showed no evidence of significant association.

3.4.3.4 TLG and NSCLC survival

TLG and survival analysis was performed based on five studies (18, 20, 23, 25, 28) with cut-off values between 20 and 802.6 ( ). TLG was not associated with either PFS (HR, 1.10; 95% CI, 0.91-1.33, P=0.34) or OS (HR, 1.52; 95% CI, 0.98-2.34, P=0.06). The heterogeneity test showed high heterogeneity in OS (I2 = 79%, P<0.01) and no significant results in PFS (I2 = 41%, P=0.17).

Figure 6

Forest plots of hazard ratios comparing progression free survival (A) or overall survival (B) of patients with high level versus low level total lesion glycolysis treating with immune checkpoint inhibitors.

4 Discussion

This study evaluated the predictive values of PET/CT parameters including SUVmax, SUVmean, MTV and TLG in NSCLC patients receiving ICIs. The cut-off values categorized patients into high or low-level parameter groups in the included studies.

Firstly, we analyzed the relationship between PET/CT parameters and the objective response of ICIs. Eight studies assessed the objective response based on RECIST 1.1. Four studies showed NSCLC patients who achieved CR, PR, or SD after ICIs treatment had significantly lower median MTV values than those with PD (16, 26, 27, 30), while four demonstrated no significant correlation (17, 23, 28, 29). More studies with consistent response assessments are needed to determine whether MTV is associated with the objective response of ICIs. SUVmean (27) and TLG (30) were also said to have a significant relationship with disease status in a single study, respectively.

ICIs may alter the physiological homeostasis of the immune response, thus leading to the development of irAEs. Two studies discussed the relationship between PET/CT parameters and irAEs (23, 31). However, no consistent results could be yet concluded.

We also discussed whether PET/CT parameters could predict NSCLC survival by PFS and OS after ICIs. We found that lower SUVmax corresponded to shorter PFS. Lopci et al. found a positive association between SUVmax and CD8-tumor infiltrating lymphocytes and PD-1 expression (32). SUVmax were also independent predictors of PD-L1 positivity by Takada et al. (13). However, the predictive role of baseline SUVmax is still under discussion since only one of the five included studies about SUVmax showed significant results.

In terms of MTV, we found that a high baseline MTV level was significantly associated with shorter PFS and OS than a low MTV level for patients treated with ICIs.

MTV refers to the metabolically active volume of tumors segmented using FDG PET (33), reflecting tumor burden and the metabolic status. Regarding tumor burden, Kim et al. concluded that larger-size tumors are more immunosuppressive than smaller-size tumors, which negatively affects the immune responses induced by immunotherapy (34).The experiments in mice also verified that PD-L1 blocker is less effective in mice bearing larger lung squamous cell tumors (35). On the cell level, Wang et al. analyzed one hundred twenty-two NSCLC tumor specimens by immunohistochemistry and found a significantly positive correlation between MTV and CD163-TAM, Foxp3-Tregs (36). CD163-TAMs were tumor-promoting M2 macrophages (37), and Foxp3-Tregs were a kind of immune regulatory cells (35), both of which are immunosuppressive cells. Therefore, we hypothesize that patients with a higher MTV would have a worse prognosis when treated with ICIs than those with a lower MTV, since a higher MTV would result in a more immunosuppressive tumor microenvironment.

In respective of tumor glycolysis, a higher MTV indicates a larger metabolically active volume of glucose uptake by the tumor (38). Different from normal cells, tumor cells can uptake a large amount of glucose at a rapid rate, consuming most of nutrients from the surrounding environment, and metabolizing glucose into lactic acid (Warburg effect) (39). Tumors with higher MTV would have worse response to ICIs by affecting T cells responsiveness by the following possible ways.

Firstly, in tumor microenvironment (TME), tumor cells and T cells compete for glucose as their primary energy source (40). Tumors with higher MTV would consume more glucose and lead to glucose deprivation of T cells, decreasing T cells’ ability to produce effector cytokines like interferon gamma (IFN-γ), which has impact on the function of tumor infiltrating CD8+ T cells (41).

In contrast, Harley et al. found that melanoma tumors with less glycolysis would provide more glucose for infiltrating T cells and are associated with increased antigen presentation and better response to anti-PD-1 ICIs (42). Secondly, the accumulation of lactate in the TME will inhibit CD8+T cell proliferation and activation by preventing lactic acid export from CD8+T cell (43) or inhibiting CD25 expression, a T cells activation marker (44). More studies are still needed to explain why MTV could predict the outcome of immunotherapy in patients with NSCLC.

Since the cut-off values of MTV ranged from 5.0 to 268.0 cm3 in different studies, we also did a subgroup analysis to determine the impact of cut-off values on outcome assessment. Our result showed that a high baseline MTV level was significantly associated with shorter PFS when setting the cut-off values of MTV below 50cm3 and shorter OS in the groups with cut-off values lower than 50 cm3 or between 50cm3 and 100 cm3.

The baseline MTV level didn’t show any predictive value when the cut-off values were more than 100 cm3. Thus further studies with a larger sample size should focus on cut-off values of MTV between 50 and 100 cm3 and try to figure out a more precise cut-off value to improve the efficacy of MTV prediction on response assessment to ICIs in NSCLC patients.

Although SUVmean and TLG were potential prognostic markers of NSCLC (45), our pooled results showed that they were not significantly associated with PFS and OS in NSCLC patients receiving ICIs.

In addition to NSCLC, PET/CT parameters also played potential predictive roles in other cancers treated with ICIs, supporting our findings. Zhang et al. reported that total SUVmax ≥12.5 was associated with worse PFS in head and neck squamous cell carcinoma (46). And according to a systematic review and meta-analysis of metastatic melenoma (47), MTV and TLG were promising predictors of OS for metastatic melanoma patients who received ICIs.

18F-FDG PET/CT is a convenient and noninvasive imaging modality, and SUVmax and MTV are easily obtained. Since our study proved that SUVmax and MTV have the potential predictive value for ICIs in NSCLC patients, further studies are needed to define the role of SUVmax and MTV in providing individualized treatments for advanced NSCLC patients. Early identification of NSCLC patients for ICIs can improve the efficacy of ICIs in responders and avoid the side effects and high costs of ICIs in non-responders, allowing them to initiate other treatments timely.

Our study also has several limitations. Firstly, majority of the included studies are retrospective studies. Potential selection bias may exist and impact the reliability of this meta-analysis. Secondly, the methods of PET/CT were not consistent between different studies. A golden method should be defined to ensure the homogeneity of studies. Thirdly, cut-off values of SUV, MTV and TLG ranged widely and were determined by different methods, including median values, log-rank test and ROC curve analysis. Thus the pooled results may show some risk of bias.

In conclusion, our study showed that high baseline MTV levels correspond to shorter PFS and OS compared with low baseline MTV levels especially when the cut-off value was set between 50-100 cm3. MTV is a potential predictor of ICI outcomes in NSCLC patients.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

Author contributions

CJ, TG and KZ conceived and designed research. Data collection was performed by DS, JW and ZC. Data extraction was performed by YC and CC and verified by KZ. Statistical analysis was performed by LC, RZ and KZ. KZ, DS, JW, ZC, YC, LC, CC, XB and CJ participated in drafting article. All authors gave final approval to the version submitted.

Funding

This work was funded by the Guangdong Pharmaceutical Association Scientific Research Fund (No: 2021ZX03).

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.