Metabolic Tumor Volume Measured by ¹⁸F-FDG PET/CT is Associated with the Survival of Unresectable Hepatocellular Carcinoma Treated with PD-1/PD-L1 Inhibitors Plus Molecular Targeted Agents

Introduction

Immune checkpoint inhibitor therapy or targeted monotherapy has recently shown considerable efficacy in unresectable hepatocellular carcinoma (uHCC).^1,2 Moreover, clinical studies have shown that the combination of anti-PD-1/PD-L1 immunotherapy with targeted therapy encourages efficacy in HCC.^3–7 However, only 10.7–37.4% of patients achieve an objective response.⁸ Thus, there is a clear need for better patient selection before immunotherapy.^7,9

Despite significant efforts, only a few clinical, histological, and genetic parameters have been identified to reliably predict the efficacy of immunotherapy for uHCC.^9–12 In clinical practice, the utility of biomarkers based on tumor tissue is limited by the available tumor tissue due to spatial heterogeneity.¹⁸F-fluorodeoxyglucose positron emission tomography-computed tomography (¹⁸F-FDG PET/CT), a non-invasive tool, has been shown to be a promising method for screening surveillance after treatment, evaluation of therapy response, and predicting the outcome.^13–15 In patients with uHCC, the highest metabolic activity within the tumor in a two-dimensional region of interest (ROI), as well as the metabolic tumor volume (MTV) and total lesion glycolysis (TLG) in a three-dimensional ROI, are considered to provide valuable prognostic information.^16,17 However, there is limited data regarding the prognostic relevance of MTV and TLG in uHCC patients treated with immunotherapy.

Therefore, this study aimed to determine the potential prognostic value of baseline PET/CT parameters for predicting the efficacy of a combination of PD-1/PD-L1 inhibitors and targeted therapy for uHCC. We also examined the integration of PET/CT parameters with clinical characteristics to identify which patients could achieve longer survival.

Materials and Methods

Study Design and Patients

This retrospective study was conducted to investigate the value of standard ¹⁸F-FDG PET/CT in predicting outcomes in patients with uHCC undergoing a combination of PD-1/PD-L1 inhibitors and targeted therapy at Peking Union Medical College Hospital (PUMCH). Patients who were scheduled to begin immunotherapy from July 2018 to December 2021 as their first- or later-line systemic treatment for uHCC were retrospectively included in this study. This study was approved by the Ethics Committee of PUMCH, and informed consent was obtained from all participants. The inclusion criteria were as follows: (1) Patients with pathologically confirmed HCC or imaging confirmed as HCC according to the American Association for the Study of Liver Diseases (AASLD) or standard for the diagnosis and treatment of primary liver cancer in 2017 in China,¹⁸ (2) patients underwent the combination therapy of a PD-1 inhibitor with a targeted therapy (such as lenvatinib, sorafenib, and apatinib) or PD-L1 inhibitor atezolizumab plus bevacizumab; (3) Eastern Cooperative Oncology Group performance status (ECOG-PS) score of 0–2, and (4) ¹⁸F-FDG PET/CT scan pre-immunotherapy or within 3 months post-immunotherapy. The main exclusion criteria includes patients who has no PET/CT imaging during the therapies. This study is registered as NCT03892577.

¹⁸F-FDG PET/CT Examination

All ¹⁸F-FDG PET/CT scans were performed using a PoleStar m660 PET/CT scanner (SinoUnion Healthcare, Beijing, China) at PUMCH. CT transmission scans (120 kV, 260 effective mA) were conducted for attenuation correction and image fusion.¹⁹ Patients were instructed to fast for ≥6 hours before the intravenous injection of 3.70–5.55 MBq/kg of ¹⁸F-FDG. Low-dose attenuation CT acquisition was performed (120 kV, 260 mA, 1.8665 mm slice thickness) at 1 h post-administration, followed by a static 3D PET acquisition from vertex to mid-thigh with an image duration of 120 s per bed position, an axial field of view of 22 cm, and a matrix of 192×192 pixels. Three key PET/CT parameters (maximum standardized uptake values, SUV_max; MTV; TLG) were measured using an MIM workstation (MIM Software Inc., USA). Volumetric parameters were defined as described previously.²⁰ The SUV_max was calculated from a 10 mm diameter ROI placed on the highest uptake site of the tumors and for the most intense lesions. To assess tumor burden, MTV was defined as the sum of volumes enclosed by a 41% isocontour around each tumor lesion voxel with the maximum ¹⁸F-FDG uptake, as recommended by the European Association of Nuclear Medicine.²¹ TLG was defined as the average metabolic activity within the tumor multiplied by the tumor volume.

Clinical Characteristics

The following clinical parameters were recorded: age at treatment initiation, gender, immunotherapy start date, prior line systemic target treatments, prior local treatment, current treatment, alpha-fetoprotein (AFP), ECOG-PS, Child-Pugh grade, Barcelona Clinic Liver Cancer (BCLC) stage, numbers of involved organ, extrahepatic spread, presence of bone metastasis, macrovascular invasion, and cirrhosis. Patients were followed up for more than 6 months with regular clinical evaluations. Overall survival (OS) was defined as the time between the start of anti-PD-1/PD-L1 treatment and death due to any cause. Progression-free survival (PFS) was defined as the time from the start of anti-PD-1/PD-L1 treatment to the first documented disease progression or death from any cause based on the Response Evaluation Criteria in Solid Tumors v1.1. Patients were censored at the last follow-up or 1 March 2022, whichever came first.

Statistical Analysis

The normal distribution was assessed using the Kolmogorov–Smirnov test, and continuous variables with a normal distribution are presented as mean ± standard deviation, whereas those with a non-normal distribution are presented as median (interquartile range). For continuous variables, statistical significance was determined using the Student’s t-test or Wilcoxon’s test. A p-value <0.05 for two-sided tests was considered significant. For statistical analysis, all continuous variables were divided into two groups, and specific cutoff values were determined based on median values. Survival functions of patient subgroups defined by SUV_max, MTV, and TLG were estimated using the Kaplan-Meier method and compared using the Log rank test. Univariate analysis was used to identify factors associated with PFS and OS. Prognostic factors identified as significant in the univariate analysis (P<0.1) were then entered into a Cox multivariate regression analysis model. Forward stepwise multivariate regression analysis was performed to identify factors correlated with OS and PFS. In each step, variables with a p-value <0.05 were inputted. The χ² test was used to analyze the correlations between MTV and clinical characteristics. Statistical analyses were performed using SPSS version 19.0 and R version 4.1.0.

Results

Patient Characteristics

A flow diagram for the inclusion of potentially eligible patients is shown in Figure 1. Overall, 77 patients with uHCC (median age 55 years, range 21–76) were finally included in this retrospective study. Table 1 shows the patients’ clinical and biological characteristics. They were mainly men (84.4%), had cirrhosis (68.8%), and received PD-1 inhibitors (90.9%). Of the 77 patients, 50 patients were treated with local-regional therapy. Baseline ¹⁸F-FDG PET/CT scans were also obtained. The median follow-up was 16.8 (5.8–25.4) months. During the follow-up period, 81.8% and 55.8% of patients experienced progression and death, respectively.

Table 1 Characteristics of Patients with Unresectable Hepatocellular Carcinoma

Figure 1 Flow chart of the data selection process. Abbreviations: HCC, hepatocellular carcinoma; SUVmax, maximum standardized uptake value; MTV, metabolic tumor volume; TLG, total lesion glycolysis; ECOG-PS, Eastern Cooperative Oncology Group Performance Status; PFS, progression-free survival, OS; overall survival.

Association of ¹⁸F-FDG PET/CT Parameters with Survival and Clinical Parameters

In our study, the cutoff values for the median of SUV_max, MTV, and TLG were 7.12, 39.65 cm³, and 159.2 g, respectively. MTV was significantly associated with OS. The median OS in the high MTV group was 12.8 months (3.7–19.2), while it was 22.8 months (9.6–30.7) in the low MTV group (P=0.044, HR 1.862, 95% CI 1.009–3.437; Figure 2). The prognostic value of TLG had a trend, but it was not significant. The median OS in the high TLG group was 12.8 months (3.73–19.37) compared with 21.5 months (9.58–30.70) in the low TLG group (P=0.056, HR 1.342, 95% CI 0.989–1.823). SUV_max was not significantly associated with OS (16.1 months vs 17.6 months, P=0.37, HR 1.148, 95% CI 0.849–1.553). Moreover, for PFS, the SUV_max, MTV, and TLG were not significantly different (P>0.05, Supplemental Table S1 and Figure S1).

Figure 2 Kaplan–Meier curves for overall survival in relation to (A) SUVmax, (B) MTV, and (C) TLG (p values for Log rank test). Abbreviations: MTV, metabolic tumor volume; SUVmax, maximum standardized uptake value; TLG, total lesion glycolysis; OS, overall survival.

In addition, the clinical features of high and low MTV were investigated. Notably, patients with a high MTV were characterized by being old age (Supplemental Table S2, P<0.05).

Univariate and Multivariate Analyses for PFS and OS

The median OS of 77 patients with uHCC was 16.4 (5.8–25.4) months. In the univariate analysis, high ECOG-PS (≥1), Child-Pugh grade (B-C), the presence of bone metastasis, and a high MTV (≥39.65 cm³) were significantly associated with poor OS (P=0.047; Table 2). Despite a clear trend in the univariate analysis, high TLG (≥159.2 g) was not significantly correlated with poor OS (P=0.059). In the multivariate analysis, a high ECOG-PS, the poor Child-Pugh grade, the presence of bone metastasis, and high MTV remained as independent significant prognostic factors for OS (HR 1.371, 95% CI 1.001–1.878; HR 3.73, 95% CI 1.925–7.231; HR 15.384, 95% CI 14.292–55.556; and HR 2.994, 95% CI 1.425–6.289, respectively; Table 3).

Table 2 Univariate Analysis of Prognostic Factors Associated with Overall Survival in Patients with Unresectable Hepatocellular Carcinoma

Table 3 Multivariate Analysis of Prognostic Factors Associated with Overall Survival in Patients with Unresectable Hepatocellular Carcinoma

In this cohort, the median PFS was 5.1 (2.6–10.7) months. In the univariate analysis, high ECOG-PS (≥1), the Child-Pugh grade (B-C), and the presence of bone metastasis were significantly associated with poor PFS (P<0.1; Table S2). In the multivariate analysis, a high ECOG-PS and the Child-Pugh grade remained as independent significant prognostic factors for PFS (HR 1.558, 95% CI 1.066–2.276 and HR 2.315, 95% CI 1.377–3.891, respectively; Table S3). However, all PET/CT parameters were not correlated with PFS.

Integrating PET Parameters and Clinical Factors: A Metabolism-Based Prognostic Model

For the multivariate analysis, we developed a prognostic model combining MTV with the ECOG-PS, Child-Pugh grade, and presence of bone metastasis, which were the independent factors for OS. Combining these important clinical factors with MTV provided further patient stratification. Specifically, the three following risk groups were identified: (1) low risk (patients with a low MTV and favorable clinical risk factors); (2) moderate risk (low MTV and unfavorable clinical risk factors or high MTV and favorable clinical risk factors); and (3) high risk (high MTV and unfavorable clinical risk factors). The Kaplan–Meier survival curves of these three groups are shown in Figure 3. The OS in the high-risk group was significantly different from that in all other groups; this combined approach to risk stratification differentiated patients according to survival better than MTV or clinical parameters alone. The median OS in all patients with a high MTV was 12.8 months. For MTV together with ECOG-PS, the median OS of the low, moderate, and high-risk groups were more than 25.5, 15.8, and 4.6 months, respectively. The OS in the high-risk group was significantly shorter than that in the low-risk group (P<0.001) and the moderate-risk group (P<0.001). The OS of the moderate-risk group was not significantly different from that of the low-risk group (P=0.258). Similarly, for MTV with the Child-Pugh grade, the median OS of the low-, moderate-, and high-risk groups were more than 23.4, 12.8, and 2.0 months, respectively. The OS in the high-risk group was significantly shorter than that in the low-risk group (P<0.001) and the moderate-risk group (P<0.001). The OS of the moderate-risk group was significantly shorter than that of the low-risk group (P=0.03). Moreover, for MTV with the presence of bone metastasis, the median OS of the low-, moderate-, and high-risk groups were more than 24.0, 15.6, and 3.4 months, respectively. The OS in the high-risk group was significantly shorter than that in the low-risk group (P<0.001) and the moderate-risk group (P=0.011). The OS of the moderate-risk group was not significantly different from that of the low-risk group (P=0.174).

Figure 3 Kaplan–Meier curves for overall survival in three risk groups stratified according to MTV combined with the ECOG-PS (A), Child-Pugh grade (B), and bone metastasis (C). Illustration of high, moderate, and low risk using maximal intensity projection on 18F-FDG PET images of three patients (D).

Discussion

To the best of our knowledge, this is the largest retrospective study to indicate that a high baseline MTV on ¹⁸F-FDG PET/CT is predictive of adverse OS in patients with uHCC aiming for treatment with a combination of immunotherapy and targeted therapy. This discriminative and prognostic application of MTV has been used in the past to predict the outcomes of other HCC treatments, including before surgery, transarterial radioembolization, and targeted therapy.^22–24 Our study expands these findings to patients with HCC receiving a combination of PD-1/PD-L1 inhibitors and targeted therapy. We propose that it should be taken into consideration when screening appropriate candidates for PD-1/PD-L1 inhibitor-based therapy in future studies.

The metabolic tumor burden is an important prognostic factor for patients treated with PD-1/PD-L1 inhibitor-based immunotherapy. Recently, Ito et al found that melanoma patients with a high baseline MTV of greater than 26.85 cm³ had a significantly shorter OS than those with a low MTV.²⁵ Another cohort with advanced non-small cell lung cancer treated with immunotherapy showed that a high baseline MTV greater than a median of 75 cm³ was correlated with poor OS.²⁶ It was confirmed in our cohort of uHCC patients that those with high MTV (above a cutoff value of a median of 39.65 cm³) had poorer OS. The discrepancy in the above results among the studies is likely associated with the different tumor types, various baseline patient characteristics, and methodological differences. Thus, these findings indicate that MTV should be considered as a predictive biomarker in patients treated with immunotherapy. Wang et al also found that baseline metabolic parameters from ¹⁸F-FDG PET could predict the pathological response of 20 HCC patients treated with PD-1 inhibitors and Lenvatinib.¹⁵ Emerging evidence has demonstrated that novel methods like granzyme B PET imaging have a higher specificity with immune-related cells such as cytotoxic T cells, which has recently been found to be associated with clinical outcome treated with immunotherapy to immunotherapy via its role in adaptive immunity.^27,28 However, the development of the tracer is unfeasible in most hospitals and more multi-center trials are needed to prove its benefits in clinic. ¹⁸F-FDG PET has much more clinical applications globally and provides high-sensitivity signals, but the SUV changes may be affected by many factors such as tumor volume, tumor progression, and new detectable lesions, leading to more bias in monitoring immunotherapies.

Our findings suggest a greater impact for prognosticating OS when combining MTV with ECOG-PS and the Child-Pugh grade, which is commonly used to assess liver function in patients with liver disease.²⁹ Many studies have shown that a high Child-Pugh grade is an important factor for poor long-term prognosis in uHCC patients receiving immunotherapy.³⁰ In a cohort of 203 patients with uHCC treated with immunotherapy, the objective response rate was lower in patients with Child-Pugh B grade than in those with A grade (2.8% vs 15.9%; P=0.010), and the median OS was also shorter in Child-Pugh B patients (2.8 vs 10.7 months; HR=2.10; P<0.001).³¹ The physical condition of the ECOG-PS is also an extremely important risk factor for HCC. Kuo et al reported that a low ECOG-PS was the most important favorable prognostic indicator of OS in multivariate analysis.³² In addition, bone metastasis in HCC may cause significant complications such as spinal cord compression, intractable pain, and pathological fractures, which ultimately lead to a severe decrease in quality of life and a poor outcome.³³ We propose that all the above clinical factors could offer timely and direct information to clinical physicians on whether to continue with PD-1 inhibitors and potentially maximize treatment benefits from the combination of a PD-1/PD-L1 inhibitor and targeted therapy in HCC.

Based on these findings, a combination of multiple biomarkers could be an innovative way to improve both prediction and prognostic accuracy.²⁶ We also developed a model combining the ECOG-PS, Child-Pugh grade, and the presence of bone metastasis with MTV at baseline, stratifying the population into three prognostic groups. Patients with poor clinical prognostic factors and a high MTV were characterized by very low OS rates. In our cohort, patients with high clinical prognostic factors (ECOG-PS, Child-Pugh, and bone metastasis) and a high MTV had a short median OS of 2.0–4.6 months; showing they are not suitable for combination immunotherapy considering the cost-effectiveness. Of course, these hypotheses for future applications of combinations of biomarkers in clinical practice and research must be properly tested in prospective randomized clinical trials.

Our study has several potential limitations. First, even if all patients received PD-1/PD-L1-based therapy, there was some heterogeneity in treatment modalities. Moreover, previous studies on the role of ¹⁸F-FDG PET/CT in the evaluation of patients with HCC have indicated that ¹⁸F-FDG PET/CT metabolic values may depend on the degree of differentiation.^{34 18}F-FDG uptake in well-differentiated HCC is similar to that in normal liver tissue because of the high rate of gluconeogenesis in well-differentiated HCC.³⁵ Thus, some patients with well-differentiated HCC lesions may have a high tumor burden but low metabolic burden from the PET scan, which may influence the results. Finally, this MTV-based prognostic model was constructed based on a small number of uHCC patients; therefore, further larger independent multi-center studies including different ethnicities are necessary to validate the clinical value of our model.

Conclusion

Baseline high MTV on ¹⁸F-FDG PET/CT was associated with poor OS following a combination of PD-1/PD-L1 inhibitors and targeted therapy in patients with uHCC. An integrating model combined with clinical prognostic factors, such as the ECOG-PS, Child-Pugh grade, and presence of bone metastasis, may improve the selection of candidate uHCC patients for immunotherapy.