Introduction
Worldwide, non-small cell lung cancer (NSCLC) is the most common cause of cancer death [
1,
2], and more than 70% of patients are found as locally advanced or metastatic disease at the time of diagnosis. Immunotherapy has become a new therapeutic approach in NSCLC with the potential for prolonged benefits [
3]. Since 2015, three immune checkpoint inhibitors (nivolumab, pembrolizumab, and atezolizumab) have been approved by the US Food and Drug Administration (FDA) for the treatment of NSCLC [
4].
Selecting patients who will benefit before or at the early stage of immunotherapy is a major issue in clinical application [
5]. The expression level of programmed death ligand 1 (PD-L1) in NSCLC is chiefly used to screen patients for immunotherapy in clinical trials. However, the patients who cannot obtain sufficient specimens for subjective or objective reasons usually cannot complete the pathological test, even more impossible to gain specimens repeatedly to evaluate curative efficacy. Furthermore, the tumor response patterns of immunotherapy may differ compared with conventional chemotherapeutic agents or targeted therapies, and the accuracy of response assessment is radiologically challenging [
6‐
8].
18F-FDG PET-CT is the most useful tool for evaluating changes of lesion on molecular level. The mechanism of FDG uptake within tumor cells is concerned with the presence of glucose metabolism, hypoxia, and angiogenesis [
9‐
11]. The level of PD-L1 expression has been associated with that of glucose transporter 1 (Glut1) and hypoxia-inducible factor 1α (HIF-1α) in patients with NSCLC [
12,
13]. Lopci et al. [
14] found a direct association between SUV
max and SUV
mean with the expression of PD-1 (rho = 0.33;
p = 0.017 and rho = 0.36;
p = 0.009, respectively) in patients with NSCLC. In a recent study, FDG PET was considered to provide more useful information on assessing response of advanced NSCLC to immunotherapy than that of computed tomography (CT) [
15]. On the basis of these findings,
18F-FDG PET-CT in predicting immunotherapy response to NSCLC seems to be a valuable and important research area in clinical application. Sintilimab is a recombinant humanized anti-PD-1 monoclonal antibody injection that blocks interactions between PD-1 and its ligands and has been tested regarding the safety and activity in patients with advanced stage solid tumor and was approved for lymphoma by Chinese Center for Drug Evaluation in China in 2018 [
16]. Phase I/II development of sintilimab for use in solid tumors is underway in the USA, with the US FDA accepting an Investigational New Drug Application for sintilimab in January 2018 [
16]. The current study aims to evaluate the relationship between tumor metabolic parameters of
18F-FDG PET-CT and the surgical pathology of the neoadjuvant sintilimab in resectable NSCLC patients, and to investigate if PET-CT has the potential to predict the major pathologic response (MPR), which predict improved long-term patient outcome [
17,
18].
Discussion
Immunotherapy is one of the most exciting fields in NSCLC with the potential for prolonged benefits [
3]. Evaluation of this novel therapy is a major challenge, since immunotherapy radically differs from other strategies in relying on the reactivation of the immune system to recognize and kill cancer cells [
21]. The use of immunomodulatory monoclonal antibodies that directly enhance the function of components of the antitumor immune response, such as T cells, or block immunological checkpoints that would otherwise restrain effective antitumor immunity has recently been actively investigated in oncology [
21,
22]. Forde et al. [
23] reported that a major pathological response occurred in 45% of tumors after neoadjuvant administration of two doses of nivolumab in patients with early stage lung cancer, and two patients whose tumors had increased in size on presurgical CT scans (although the increase was less than RECIST-defined progression) were found to have minimal or no residual tumor in the surgical specimen. These findings represent pathological evidence supporting the possibility that some patients may derive clinical benefit from immunotherapy without initial radiographic tumor shrinkage. Conventional imaging criteria, either RECIST1.1 or iRECIST, has the above limitations for depending on morphologic changes [
6‐
8,
24]. PET-CT was considered to overcome such limitations and more suitable for assessment of therapeutic effect, because it can reflect on tumor metabolic level before morphological changes. In 2009, Wahl et al. [
20] proposed the PET Response Criteria in Solid Tumors. The major concepts of PERCIST were the use of SUL for the tumor response assessment, and the identification of a measurable target lesion SUL is at least 1.5-fold greater than liver SUL
mean + 2 SDs (in 3 cm spherical ROI in normal right lobe of liver) [
20]. PERCIST proposed to use the percentage change in SUL
peak (or sum of lesion SULs) between the pre- and post-treatment scans for assessing therapy response. The mechanism of FDG uptake within tumor cells is concerned with the presence of glucose metabolism, hypoxia, and angiogenesis [
9‐
11]. The level of PD-L1 expression is associated with Glut1 and HIF-1α in patients with NSCLC [
12,
13]. Therefore, some studies attempt to search whether baseline
18F-FDG PET-CT can provide useful information on the expression of checkpoint inhibitors in patients with NSCLC, in order to distinguish patients from the potential for prolonged benefits. Indeed, there is a statistically significant association between tumor metabolic parameters on
18F-FDG PET-CT and PD1/PD-L1 expression in resected tumor specimens [
12‐
14,
25]. Grizzi et al. [
26] found that almost all patients (
n = 27) with SUV
max ≤ 17.1 or SUV
mean ≤ 8.3 on baseline PET had fast progression after 8 weeks immunotherapy.
Our study has revealed the clinical significance of
18F-FDG PET-CT as a promising biomarker for predicting early phase clinical outcomes of PD-1 blockade therapy in NSCLC patients. Specially, PMR (100%, 13/13) according to PERCIST showed excellent prediction capabilities to distinguish patients with MPR. Despite metabolic parameters of baseline
18F-FDG PET-CT, including SUL
max, SUL
peak, MTV, TLG, cannot distinguish patients with MPR (
p > 0.05), which may be due to small-sized sample, SUL
max and SUL
peak of baseline were positively correlated to the degree of pathological regression (SUL
max,
p = 0.036; SUL
peak,
p = 0.058). The result may indirectly support the hypothesis that metabolic characteristics of tumor on baseline may be part of a larger panel of predictive factors of response to immunotherapy of NSCLC [
22,
27].
Although little is known about PERCIST criteria with respect to response to immunotherapy of NSCLC, there are a few studies or case reports describing its role in evaluating response to immune checkpoint inhibitors [
15,
25,
28‐
30]. In a recent study, 24 patients treated with PD-1 blockade (nivolumab) were investigated at baseline and 1 month after the start of treatment [
15]. Response was determined using both morphological (RECIST 1.1) and PERCIST criteria. The value of PET in predicting PR (partial response) and PD (progressive disease) was significantly higher than that of CT. The multivariate analysis confirmed FDG uptake after administration of nivolumab was an independent prognostic factor in predicting progression free survival (PFS) (HR 3.624,
p < 0.001) and overall survival (OS) (HR 2.461,
p = 0.012) [
15]. Another study assessing response of NSCLC to immunotherapy, 103 patients treated with anti-PD-L1 agent (atezolizumab) were evaluated the potential of FDG PET-CT for assessing response [
30]. Patients with metabolic response on 6-week scans had a higher ORR (objective response rate) than metabolic non-responders (73.9% (in 17 of 23 patients) vs. 6.3% (in 5 of 80 patients)) [
30]. The reports above have noted that PET-CT is a useful tool for immune monitoring, nevertheless, the optimal time for evaluating the appropriate efficacy remains uncertain. The results of our study suggest that 4 weeks after the first dose of sintilimab maybe an opportune moment, and the percentage changes of the metabolic parameters (SUL
peak%, ΔSUL
max%) could correctly predict MPR. The specificity, sensitivity, and accuracy were 100%, 100%, and 100% (
p = 0.000). The metabolic parameters of scan-2 also showed good prediction capabilities to distinguish patients with MPR. All metabolic parameters of scan-2, including SUL
max, SUL
peak, MTV, TLG, were negatively correlated with the degree of pathological regression. SUL
peak of scan-2 has the best differentiation ability. By setting threshold of SUL
peak to 6.7, the specificity, sensitivity, and accuracy were 92.3%, 81.8%, and 86.1%, with AUC of 0.90 (
p = 0.000). This preliminary result of our study demonstrated the clinical significance of the follow-up scan at 4 weeks after the first dose of sintilimab treatment and may provide useful information for selecting patients who had benefit at the early stage of immunotherapy.
In addition, we observed an interesting phenomenon in this study. One patient (Fig.
3) in this clinical trial was evaluated as PMD before surgery. He had a remarkable enlargement in the size of tumor (4.1 cm vs. 5.7 cm) on preoperative PET-CT. Despite ΔSUL
peak% (32.1%) had a significant increase, both ΔMTV% (− 60.1%) and ΔTLG% (− 50%) of the primary tumor decreased markedly. The postoperative pathology showed the primary tumor had 60% of pathological regression and large numbers of macrophages and infiltrating lymphocytes. This interesting phenomenon may help to explain the pathological basis of pseudoprogression [
22,
24]. Besides, the deviation of metabolic parameters (e.g., increased ΔSUL
peak% vs. reduced ΔMTV% and ΔTLG%) may help to differentiate pseudoprogression from PMD. However, there was only one patient with pathologically confirmed “pseudoprogression” according to PERCIST criteria. Further studies are necessarily needed to explore the efficiency of the combined application of multiple metabolic parameters for distinguishing “pseudoprogression” from PMD.
There are several limitations in this study. Firstly, our study is a preliminary study and includes a small sample size. There was only one patient with histopathologically confirmed PD. Thus, we did not analyze the potential of metabolic parameters to predict patients who cannot benefit from immunotherapy before surgery. Further studies including larger numbers of patients are necessary to validate these results. Secondly, this study mainly focuses on the metabolic response of
18F-FDG PET-CT for predicting the major pathologic response to the neoadjuvant PD-1 blockade. We did not analyze the relationship between biomarkers such as the tumor mutational burden and PD-L1 expression and metabolic parameters, for they are not available yet until we submit the manuscript. We also did not analyze the immune-related side effects in this study. Thirdly, previous studies indicated that EGFR mutations were associated with low response rates to PD-1 blockade treatment among patients with NSCLC; in some cases, inhibition of checkpoint blockade even increased the rate of tumor growth considerably [
31‐
33]. Therefore, we excluded the patients with the presence of EGFR-sensitive gene mutation in tumor tissue. However, EGFR mutation rate are very high (51.8%) in Chinese lung adenocarcinoma population [
34], and quite a few (56%) adenocarcinoma were ground glass opacity (GGO) which were not suitable for the study [
35]. Therefore, most of the patients with adenocarcinoma were excluded, and the large majority of the patients had squamous cell carcinoma subtype in this trial, which may bias the results in this study. Fourthly, we did not evaluate clinical end points such as OS rate or PFS, as our study focused on MPR, which strongly associates with improved survival of neoadjuvant therapy [
17,
18]. Long-term follow-up is necessary to confirm the prognostic significance of OS using
18F-FDG PET-CT.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.