In this study, the GNRI was found to be a predictor of the prognosis of patients with metastatic or unresectable clear cell RCC who were treated with first-line ICI combination therapy. However, the GNRI was not identified as an independent prognostic factor, but multivariate analysis showed that KPS < 80 was associated with prognosis. The rates of response to therapy with dual ICIs and ICI with TKI were higher among patients with good nutritional status than among patients with poor nutritional status. The incidence of adverse events did not differ according to nutritional status.

The present study showed that the GNRI was able to predict the prognosis of patients with RCC who received ICI combination therapy with statistical significance. This result was consistent with those of previous studies on the usefulness of GNRI in predicting the OS of patients with metastatic renal cancer on first-line TKI treatment or second-line nivolumab monotherapy [12, 16]. Several studies have reported that albumin, a component of the GNRI, is associated with cancer immunity, as it suppresses excessive neutrophil inflammatory responses or extracellular traps involved in tumor metastasis [17,18,19,20,21]. Low BMI is also associated with poor prognosis of patients with malignant diseases [22]. These results support the idea that poor nutritional status, based on the combination of low albumin level and low BMI, is an effective indicator of cancer survival. In contrast, high BMI and obesity are associated with malignant tumor mortality [23]. In this study, the GNRI formula was 14.89 * serum albumin level + 41.7 * 1 if the BMI was higher than 22. Therefore, this formula is robust for obesity and allows for the evaluation of poor nutrition.

In this study, the KPS score alone and not the GNRI was associated with the OS and CSS of patients on ICI therapy, in line with the results of a previous report discussing the usefulness of the GNRI in predicting the CSS of urothelial carcinoma patients on second-line ICI monotherapy [10]. In contrast, a study of patients on second-line ICI monotherapy for lung cancer or renal cancer reported that the GNRI is a stronger predictor of prognosis than the KPS score [11, 12]. Comparisons of these data are difficult given that drug indications and patient characteristics vary across different types of cancers and treatment lines. Patients who received previous treatment might have poor nutritional status due to cachexia caused by disease progression or low albumin values which have been occasionally observed as adverse events of TKIs. It appears that the usefulness of GNRI and KPS is interchangeable depending on the type of cancer and line of treatment. The combined efficacy of KPS and GNRI in predicting prognosis has been reported among patients on second-line nivolumab monotherapy for renal cancer. However, in this study, subgroup analysis performed only for patients with favorable KPS scores showed no statistically significant results.

Furthermore, in our study, low GNRIs were not associated with the PFS, in contrast with the results of a previous study on renal or lung cancer patients on second-line ICI monotherapy [11, 12]. Some previous reports present evidence supporting the notion that albumin and BMI are associated with PFS, e.g., albumin enhances PD-L1 antitumor immunity via regulatory T cells by reducing oxidative stress and that patients with BMI > 30 on ICI treatment have high PFS rates [18, 21, 24,25,26]. It is possible that the response rate for first-line ICI combination treatment is higher than that for second-line nivolumab monotherapy, which is one of the reasons for the insignificant difference in PFS during the follow-up period.

The rates of response to treatment were better in the group with GNRI ≥ 98 than in the GNRI < 98 group for both dual ICI therapy and ICI with TKI therapy. The incidence of high-grade adverse events was similar between the two groups. However, we could not determine the indicators of the preferred treatment choice: dual ICI therapy or ICI with TKI. Some patients respond well to treatment with ICI and have a durable response over a long duration [27, 28]. The predictors of durable responses are still unknown. The IMDC risk model was used to select the treatment, but it did not predict durable response. The identification of useful biomarkers for treatment selection remains challenging.

This study had a few limitations. The data were collected from multiple institutions; therefore, we were not able to obtain homogenous data. The details of adverse events were not available. Unfortunately, there was one death due to immune-related myocarditis, and its association with the GNRI was unclear. We used a cutoff GNRI of 98, while previous studies used a cutoff of 92 [11, 12]. However, the cutoff point of 92 did not reveal any new significant results. In the present study, we found that poor nutritional status led to poor prognosis after ICI combination therapy for metastatic or unresectable RCC. Although this is a retrospective study, the findings provide valuable new insights into the prognoses of patients with RCC who are treated with ICIs.