Immunotherapy has revolutionized the treatment landscape for numerous cancers, demonstrating remarkable efficacy in diseases previously associated with poor prognoses, thereby ushering in a new era of therapeutic possibilities [1]. Growing evidence suggests that tumors exploit immune checkpoints, crucial mechanisms for evading antitumor immune responses [2]. Activation of the Programmed death protein-1/programmed cell death ligand-1 (PD-1/PD-L1) pathway hampers the proliferation of CD4+ T cells and CD8+ T cells, diminishing the capacity of T cells to eradicate tumor cells [3, 4]. Consequently, inhibitors of PD-1/PD-L1, such as nivolumab, pembrolizumab, atezolizumab, and durvalumab, have become focal points in tumor immunotherapy research. These inhibitors are currently being used as monotherapy or in combination with chemotherapies, in both first- or second-line treatments, as well as in neoadjuvant or adjuvant settings, for approximately 50 types of cancer, showing encouraging results [5,6,7].

Despite the broad range of activity shown by checkpoint inhibitors, not all patients benefit from this therapy, and some experience immune-related adverse events. The critical question remains: “Will the treatment be effective?” New challenges arise in individualizing the treatment for patients who could benefit from anti-PD-1/PD-L1 treatment and dynamically monitoring their response during therapy [8,9,10]. The National Comprehensive Cancer Network (NCCN) guidelines recommend treating patients with immune checkpoint therapeutics targeting PD-L1 based on its expression levels, as determined by immunohistochemistry (IHC) on a histological tumor sample. However, IHC cannot assess the dynamic nature of PD-L1 expression in tumors, and multiple and serial biopsies are impractical and present additional risks to patients. This may be a factor in why some tumors deemed “expressers” ultimately do not benefit from immunotherapy [11, 12]. Several studies have evaluated [18F]FDG-PET/CT as a predictive imaging biomarker for the response to immune checkpoint inhibitor therapy in patients [13,14,15]. Conventional [18F]-FDG-PET scans have failed to detect the antitumor activity exerted by anti-PD-1/PD-L1 therapy. The tumor volume and metabolism may increase after treatment, affecting the accuracy of evaluating the clinical immunotherapy efficacy [16, 17].

These aspects, combined with the complex imaging challenges in evaluating responses to anti-PD-1/PD-L1 therapy, have steered scientific research towards the pursuit of novel predictive biomarkers. There is growing anticipation that advanced molecular-based technologies could potentially overcome the current impediments in identifying patients who may benefit from such treatment. In this context, molecular imaging through PET has emerged as a noninvasive method to visually, and quantitatively, detect PD-L1 expression at the molecular level. This approach holds promise in overcoming the inherent limitations of PD-L1 IHC assays, offering insights into the dynamic changes in PD-L1 expression during cancer therapy. PET imaging thus presents a compelling avenue for assessing patients’ responsiveness to antibody drugs and has been explored in both preclinical and clinical settings for immune checkpoint imaging in tumors. Its utility is particularly pronounced in scenarios of immunotherapy resistance, offering a personalized therapeutic approach [17,18,19,20].

Currently, PD-L1-targeting probes used for immuno-imaging are typically categorized into antibodies, nanobodies, and other small PD-L1-binding proteins (such as protein scaffolds and peptides), based on their molecular weight [21]. A drawback of drug monoclonal antibody radiotracers lies in their large molecular size, which necessitates several hours or days for adequate biodistribution and optimal tumor-to-background contrast for imaging. Consequently, small-molecule PD-L1 probes with favorable pharmacokinetic profiles have garnered increasing interest [22]. Notably, peptide-based PD-L1 tracers such as [68Ga]Ga-NOTA-WL12, [18F]NOTA-WL12, [18F]DK222, and [18F]NOTA-NF12 have emerged as viable candidates [23,24,25,26]. Peptides are readily accessible and can be optimized to exhibit high affinity and specificity for a broad spectrum of targets. Tracers based on peptides not only retain superior imaging capabilities but also demonstrate higher affinity and specificity for tumor tissues, coupled with a shorter blood circulation time—typically in the favorable range of minutes to hours—underscoring their potential for clinical applications [27,28,29].

The development of a same-day PET imaging agent capable of measuring PD-L1 status in tumors holds significant importance for optimizing PD-1 and PD-L1 treatments. In a recent issue of the European Journal of Nuclear Medicine and Molecular Imaging (EJNMMI), Dr. Donnelly and colleagues from Small Molecule Drug Discovery-PET Radiochemical Synthesis, Bristol Myers Squibb Research and Early Development, USA, present a meticulously designed and executed study on the discovery and evaluation of a novel fluorine-18 labeled macrocyclic peptide-based PET ligand, [18F]BMS-986229, for imaging PD-L1 [30]. This ligand was designed based on BMS986189, a potent macrocyclic peptide-derived PD-L1 antagonist with picomolar PD-L1 affinity. The authors chose to incorporate a propargyl glycine moiety into the peptide’s solvent-exposed portion, taking advantage of the copper-catalyzed azide-alkyne cycloaddition to rapidly label it with fluorine-18, resulting in a PD-L1 PET ligand with picomolar affinity [31, 32]. The kinetic parameters determined through surface plasmon resonance (SPR) analysis suggest that all these PD-L1 binding vectors (mAb, adnectin, peptides) have very similar SPR results. BMS-986229 demonstrated no appreciable difference in binding parameters compared to its unmodified counterpart BMS-986189.

This study was meticulously designed to demonstrate the novel 18F-labeled macrocyclic peptide radioligand’s utility for PET imaging of PD-L1 expressing tissues, exhibited several advantages in a nonhuman primate model when compared directly to adnectin- or mAb-based ligands. The key findings are summarized below: (1) The macrocyclic peptide ligand demonstrated high affinity for PD-L1, with KD values in the picomolar range, and a long dissociation off-rate from PD-L1 (in the range of ≤ 1.0 × 10−5 s−1). Additionally, it showed rapid clearance from the blood and non-PD-L1 tissues, with a blood half-life of less than 2 h. (2) The authors used their recently described novel azide prosthetic group [18F]BMT-187144 to rapidly label this macrocyclic peptide with fluorine-18, which is matched to a blood half-life, resulted in reduced radiation burden on the patient and increased convenience for clinical application. (3) Using copper-mediated click chemistry conducted under mild conditions, the authors isolated [18F]BMS-986229 in high radiochemical purity. (4) In vivo imaging results demonstrated rapid delivery of the ligand to PD-L1 expressing tissues (both tumors and non-human primate spleen) and rapid clearance from non-PD-L1 expressing tissues (specific radioligand binding > 90%), providing amazing tumor contrast. Besides, ex vivo autoradiography also showed the differentiation of PDL1 binding by imaging findings, suggesting potential application for ex vivo test for pathological specimen. (5) [18F]BMS-98229 PET imaging was able to measure target engagement of PD-L1 inhibitors (peptide or mAb) and confirm an in vivo off-rate of a PD-L1 inhibitor as high as 97%. This information can help guide dosing recommendations for drug candidates. (6) Clinical studies with [18F]BMS-986229 are currently underway to better understand the checkpoint pathway and PD-L1 expression in human tumors [33, 34].

This work is significant as it marks the successful combination of copper-mediated click chemistry with a macrocyclic peptide radioligand, which has demonstrated targeted and specific PD-L1 binding. Previous studies by the authors’ team have shown that adnectin PD-L1 PET radioligands, [18F]BMS-986192 and [68Ga]BMS-986192, serve as useful tools for serial imaging of PD-L1 expression in tumors [27, 32, 35, 36]. However, the poor molecular stability of the peptide poses a significant challenge to the application of molecular probes. Additionally, the radiofluorination of BMS-986192 showed moderate radiochemical yields and challenging synthesis, limiting the broad clinical application of PD-L1 adnectin-based PET imaging. The predictive role of PD-L1 expression on tumor cells remains an ongoing area of research. Zhou et al. [20] have developed a novel PD-L1 PET imaging tracer, [68Ga]Ga-AUNP-12. They successfully visualized the B16F10 tumor, achieving a tumor uptake of 6.86 ± 0.71% ID/g and a tumor-to-muscle ratio of 6.83 ± 0.36 at 60 min post-injection of [68Ga]Ga-AUNP-12. However, it is important to note that 68Ga has a relatively higher positron energy (Emean = 0.83 MeV) compared to 18F (Emean = 0.25 MeV), resulting in a relatively longer positron range (Rmean = 3.5 mm), which may affect spatial resolution [37]. On the contrary, this study achieved higher image quality due to its utilization of positron emission and a high theoretical specific activity of 63 TBq/mmol. Importantly, the authors further performed non-human primate experiments to validate their results. Zhu et al. [10] have developed novel peptide-based radiotracers, [68Ga]/[18F]AlF-NOTA-IMB, for the accurate detection of PD-L1 expression. Both tracers exhibited high affinity for human and murine PD-L1, enabling sensitive and specific identification of tumors with varying levels of PD-L1 expression. However, in contrast to the macrocyclic peptide, IMB, being a straight-chain peptide, led to a high uptake of [68Ga]Ga-NOTA-IMB in the liver, which affected the contrast of PET images. Furthermore, a key requirement for clinical translation is the ease of implementation of the radiotracer in the workflow. The macrocyclic peptides utilized in this investigation exhibit greater structural stability compared to straight-chain peptide, displaying enhanced conformational robustness and a propensity for sustaining their bioactivity. Employing copper-mediated click chemistry, the author achieved a threefold increase in isolated yield of [18F]BMS-986229 and a 4.7-fold improvement in molar activity when contrasted with the original synthesis of [18F]BMS-986192. These results suggest that this ligand has the potential to facilitate a broader clinical application of PD-L1 PET imaging.

Immunotherapy represents an evolving field poised to revolutionize cancer therapy. Moving forward, molecular imaging serves as a powerful visualization tool, offering the potential to predict and monitor response to immunotherapy, thus transforming cancer care and improving patient quality of life. Furthermore, it is expected to play a critical role in drug development and trial design [38]. While radiolabeled antibodies provide information on the blockade of mAbs or PD-L1 expression, they fall short in delivering real-time evaluations and personalized diagnoses due to their prolonged circulation time. As the next generation of immuno-therapeutics shifts towards small molecules, we see a huge potential for imaging applications. As the technology behind molecular imaging continues to advance, and with fluorine-18 boasting superior imaging properties, the ongoing development of new radiofluorination strategies can aid in transforming therapeutic molecules into valuable tools for immunoimaging. This could pave the way for offering easily accessible imaging tools, even for smaller medical centers. Consequently, the development of small molecule-based PD-L1 radioligands merits increased attention from radiopharmaceutical scientists and clinicians.

In summary, Donnelly et al. elaborated on the small molecule radiotracer [18F]BMS-986229, showcasing high specificity and stability, which demonstrated several advantages within a nonhuman primate model compared to adnectin- or mAb-based ligands. These findings provide clinical support for the future development of immunoimaging technology. We anticipate that the results of [18F]BMS-986229 in human trials will contribute to improved clinical decision-making. Additionally, we believe that novel radiopharmaceuticals based on small molecules are likely to complement, rather than replace, existing biomarkers, and will be integral to the opportunities that may emerge from radiomics and artificial intelligence in imaging in the future.