On February 2, 2024, the latest “Global Cancer Burden Growing and Mounting Need for Services” (1)released by the International Agency for Research on Cancer (IARC) of the World Health Organization showed that lung cancer is the most common type of cancer in the world, and ranks first in cancer “death list”. Non-Small Cell Lung Cancer (NSCLC) accounts for about 85% of all types of lung cancer (2) In recent years, immune checkpoint inhibitors (ICI), especially programmed death receptor 1 (PD-1) and its ligand 1 (PD-L1), have been widely used.PD-1/PD-L1 inhibitors have significantly improved the treatment of patients with advanced NSCLC (3). However, in view of the occurrence of pseudoprogression, hyperprogression, and withdrawal of immune-related adverse reactions in clinical practice, as well as the heterogeneity of response to immunotherapy among different NSCLC patients, we should choose an effective indicator to predict the survival prognosis of NSCLC patients before treatment. Given the occurrence of phenomena such as pseudo-progression, super-progression, and discontinuation of immunotherapy due to immune dysfunction in clinical practice, as well as the heterogeneity of responses to immunotherapy among different NSCLC patients. Selecting an effective indicator for predicting the survival prognosis of NSCLC patients before treatment is of vital importance for screening out the sensitive NSCLC patients for immunotherapy and improving the overall treatment efficacy. This is because it can help to identify patients who are more likely to respond to immunotherapy and thus enhance the effectiveness of the treatment (4).Furthermore, even among those patients who initially responded to ICIs, eventually, they will develop resistance to ICIs. This may require an effective combination with other treatments to effectively remodel the immune microenvironment into a mode more suitable for immunotherapy (5). A large number of studies have confirmed that immunotherapy combined with chemotherapy can play a synergistic effect, which can benefit patients more than chemotherapy alone (6, 7).

PD-L1, measured using immunohistochemistry, is the only approved biomarker used to select patients who can benefit from immunotherapy (8). However, a large number of studies have found Tumor Mutational Burden (TMB) (9, 10), derived neutrophil to lymphocyte ratio (dNLR) (4), serum tumor markers (such as encompass carcinoembryonic antigen (CEA), neuron-specific enolase (NSE), cytokeratin fragment 21-1 (Cyfra21-1) and carbohydrate antigen (CA19-9) (11)have predictive significance. Among them, dNLR includes monocytes and other granulocyte subsets, which can capture systemic inflammatory responses and changes in peripheral blood white blood cells. But these parameters provide limited prognostic information at the individual level. Therefore, effective biomarkers for predicting the efficacy of ICIs are still being explored. ¹⁸F-FDG PET/CT imaging, as an important clinical imaging examination method, has been widely applied in clinical practice for cancer patients. It plays a significant role in the clinical diagnosis, staging, treatment evaluation, and prognosis of NSCLC patients. At the same time, relevant studies have also confirmed that ¹⁸F-FDG PET/CT metabolic parameters can reflect the level of tumor glucose metabolism and the state of the tumor immune microenvironment to a certain extent (12).

This study will focus on discussing the application value of PET metabolic parameters and the aforementioned related hematological indicators in the prognosis of survival of advanced NSCLC patients who received PD-1/PD-L1 ICIs in combination or without chemotherapy. It aims to seek an effective PET/CT imaging indicator for NSCLC patients to screen out the sensitive population before treatment.

This research plan has been reviewed and approved by the Institutional Review Board (IRB) of Tianjin Medical University Cancer Hospital (Approval Number: [bc20240732]). Given that this study is a retrospective analysis, all data have been de-identified and the identities of individual patients cannot be identified. Moreover, the research risk is no greater than the minimal risk. The committee has approved the exemption from obtaining informed consent from the patients.

PET/CT imaging, as a fusion imaging of CT and PET, can provide anatomical structure information and molecular metabolic information of lesions at the same time, so as to find metastatic lesions more sensitively and evaluate the disease more comprehensively. Previous studies have confirmed that tumor burden and metabolic changes are associated with the prognosis of patients receiving PD-1 inhibitors combined with chemotherapy. The aim of this study is to explore the predictive value of ¹⁸F-FDG PET metabolic parameters, including PET metabolic parameters of primary lesions and systemic lesions, in the survival prognosis of advanced NSCLC patients treated with PD-1/PD-L1 inhibitors with/without chemotherapy. The results show that patients with low MTVwb are more likely to benefit from the treatment and achieve longer PFS and OS. The advantage of this study lies in the extremely high completeness of data for all variables, with no missing values. This avoids potential biases that may arise from data interpolation and enhances the strength of the statistical conclusions. Furthermore, we found that the treatment method does not affect the prognostic effect of MTVwb. MTVwb is a reliable prognostic biomarker, and its value is not influenced by the changes in the treatment strategies covered by this analysis. Spearman correlation analysis in this study demonstrated a very strong positive correlation between MTVwb and TLGwb (rs = 0.956, P < 0.001). However, the multivariate model ultimately confirmed that only MTVwb was an independent prognostic factor. Therefore, we did not conduct principal component analysis to preserve the clear clinical interpretability of MTVwb. In future studies involving multiple significant imaging parameters, it is advisable to consider using dimensionality reduction techniques such as PCA to construct composite indicators. A retrospective study by Monaco Let al. (20) included 92 lung cancer patients treated with PD-1/PD-L1 monotherapy and found that the baseline PET metabolic parameter MTVwb was an independent predictor of ICIs treatment response. In some other studies, which metabolic parameters have higher specificity and sensitivity and higher predictive value still need to be explored by prospective studies with large sample sizes. Some studies (18, 21, 22) also hold that PET metabolic parameters MTVwb/TMTV or primary tumor metabolic volume MTV and total lactate production (TLG) are valuable indicators for predicting the outcome and survival period of immunotherapy in NSCLC patients. Some studies also suggest that SUVmax can also predict the prognosis of lung cancer immunotherapy (23), but in recent years, many studies on lung cancer immunotherapy have shown that tumor metabolic volume and total glycolysis seem to be better than semi-quantitative parameters (such as SUVmax) to reflect the prognosis. However, which metabolic parameters have higher specificity and sensitivity and higher predictive value still need to be explored in prospective studies with large sample sizes. At present, immunotherapy combined with chemotherapy has become an important strategy for the first-line treatment of advanced NSCLC. This study also included patients who received immunotherapy as a single-agent treatment and those who received immunotherapy in combination. The results show that the combined treatment group performed significantly better than the single-drug treatment group in terms of prognosis. This discovery provides new evidence-based basis and clinical ideas for formulating more precise and personalized treatment plans for patients with advanced NSCLC. This might be related to the mechanism of action of chemotherapy. Not only can it inhibit the growth and proliferation of tumor cells through multiple mechanisms, but it can also act on various immune cells in the body (24). By blocking the activity of immunosuppressive cells, such as regulatory T cells, myeloid suppressor cells, and tumor-associated macrophages, or directly inducing the upregulation of MHC1 class molecules expression and the maturation of dendritic cells, it can promote the anti-tumor immune response of the body, thereby achieving a synergistic effect in tumor immunotherapy. In a meta-analysis by Zhou, Yixin et (25), 5 clinical trials with a total of 1289 patients were included for subgroup analysis of PD-L1≥50%. They also agreed that the objective response rate (ORR) and PFS of immunotherapy plus chemotherapy were superior to immunotherapy alone, and there was no significant difference in OS. At the same time, another meta-analysis (26)including 11 randomized controlled trials with a total of 6731 patients, analyzed the subgroup of patients with PD-L1≥50%. They also agreed that immunotherapy plus chemotherapy was superior to immunotherapy alone in terms of PFS (HR = 0.54, 95%CI: 0.35-0.82), but OS (HR = 0.86,95%CI: 0.65-1.14) was not significant. The results of these two studies were similar to those of the present study, but subgroup analysis of PD-L1 expression was not performed due to the small sample size of PD-L1 gene expression in this article. This outcome leads to an increased risk of committing a Type II error. Meanwhile, patients who undergo PD-L1 testing may be influenced by the specific clinical criteria of this research center or the judgment of the doctors. This may lead to selection bias. In the future, we will continue to increase the sample size and conduct multi-center data collection, in order to obtain more widely applicable research results. At present, there are more and more studies on PET/CT parametric radiomics to predict the efficacy of immunotherapy. Mu, Wei et (27) used PET/CT radiomics to predict the durable clinical benefit of immunotherapy, which proved the good prospect of multi-parameter radiomics features (mpRS) in predicting the durable clinical benefit. With the emergence of many related studies on imaging parameters and radiomics, the number of image features extracted from PET/CT parameters is increasing, which may cause redundancy and lack of reproducibility of image features (28). Therefore, how to select image feature parameters more suitable for clinical practice becomes very important.

dNLR encompasses monocytes and other granulocyte subsets and can capture systemic inflammatory responses as well as changes in peripheral blood leukocytes. Zhang Lei et al. (29). However, in this study, the hematological parameter dNLR did not show any statistically significant difference related to prognosis. Taking into account that this study and the studies by Zhang Lei et (29) have potential sample selection errors due to the inclusion of advanced NSCLC patients, and that the PD-1 inhibitors used and the chemotherapy regimens are different. In this study, the same sample size was used to conduct parallel analyses of hematological and PET metabolic parameters. A scientific and rigorous head-to-head comparison was employed to ensure the fairness and reliability of the research results. This method has significantly promoted the development of efficacy prediction research for ICIs toward greater precision, practicality and cost-effectiveness. But it is necessary to conduct multi-center large-sample prospective related studies in the future to deeply analyze this issue.

In this study, patients with high expression of the PD-L1 gene also showed a significant increase in TLGwb. Evangelista et al. (31) suggested that the high expression of PD-L1 was also associated with TLGwb, SUVmax, and MTV. Moreover, the sample size of this study was relatively large, and the results were quite convincing. In NSCLC, PD-L1 gene expression is currently the only biomarker approved by FAD for the specific PD-1/PD-L1 inhibitor. However, in clinical practice, there are technical issues in PD-L1 detection (such as batch effects of staining itself) and biological reasons for PD-L1 (such as sampling errors caused by uneven distribution). Moreover, some patients are unable to obtain sufficient tumor tissues for PD-L1 staining. Therefore, some studies have further investigated the relationship between tumor glycolysis and PD-L1 expression and explored more favorable biomarkers.

This study is a retrospective single-center study with a relatively small sample size and a short follow-up period. Inevitably, the completeness of medical records and the accuracy of data are limited, which may lead to information bias. Although we have endeavored to minimize this bias through a rigorous data verification process, we will still be committed to conducting prospective studies in the future to confirm the reliability of this result. The use of glucocorticoids at baseline has been associated with adverse outcomes with immunotherapy in some studies (30, 31), but these patients were not excluded because of the small number of patients who received glucocorticoids in our study. The ongoing debate about the optimal TMTV measurement hinders comparability across studies. In this study, there are individual differences in the use of PD-1/PD-L1 inhibitors. These drugs have slightly different mechanisms of action in the microenvironment, but the clinical efficacy differences remain inconclusive for now. Although the iRECIST efficacy assessment criteria have been recommended for use in clinical trials. The RECIST1.1 standard may not be able to accurately identify all patients who benefit from immunotherapy. The study is a retrospective study; the follow-up period is quite long. At present, in the NCCN guidelines, the RECIST 1.1 standard is still regarded as the standard tool for evaluating the therapeutic efficacy of most solid tumors. the consensus guideline iRECIST was not used (31).

Our research further substantiates the ability of tumor metabolism to evaluate the therapeutic efficacy of ICIs treatment. Immune checkpoint inhibitors can correct the escape inhibition of T cells by the tumor microenvironment (TME) and restore T cells to a normal activated state. MTVwb is a powerful prognostic biomarker, which lays the foundation for its future clinical application. In the early stage of diagnosis, MTVwb can be used to identify patients with a high tumor burden and poor prognosis. However, its ability to predict the prognosis of immunotherapy can provide an early and precise diagnostic plan for these patients. One of the major challenges in truly transforming MTVwb into a clinical tool lies in the standardization and validation of its threshold. In this study, the cutoff values (such as the median or the optimal value determined by the ROC curve) were data-driven and originated from the analysis of the current retrospective cohort. Although it has significant statistical significance, it should be cautious to directly apply this result to other populations, as there is a risk of overfitting. Therefore, the results of this study urgently need to be verified through prospective, multi-center clinical trials. This is the necessary path to bring MTVwb into clinical application. The verification study we envision should include: 1. Prospective design, where the treatment methods for patients and the timing of PET/CT examinations are pre-defined, and data are collected in large, multi-center cohorts. 2. Standardized procedures: All participating centers are required to follow the unified standard operating procedures for PET/CT scans and image analysis to minimize technical variations. 3. Clinical efficacy verification: The ultimate goal of this study is not merely to reconfirm the prognostic value of MTVwb, but to construct an integrated model and create a new stratification index.

High MTVwb of baseline PET metabolic parameters is associated with poor PFS, OS, and non-DCB in advanced NSCLC patients treated with PD-1/PD-L1 inhibitors. Patients treated with ICIs combined with chemotherapy are more likely to achieve DCB and better PFS than those treated with ICIs alone, but it is not associated with OS.¹⁸F-FDG uptake may be a promising predictor of the response to PD-1/PD-L1 inhibitor therapy, and it may be useful for individualized treatment planning in clinical practice (31).