Specific PET Imaging for Precision Management of Lung Cancer: Advances, Clinical Translation and Future Directions

Abstract

Lung cancer continues to be the leading cause of cancer-related deaths worldwide, primarily due to persistent challenges in early detection and the limited effectiveness of precision medicine. Although low-dose computed tomography (CT) has been widely implemented for lung cancer screening and has contributed to a measurable reduction in disease-specific mortality, its diagnostic accuracy is limited by its inability to reliably distinguish benign from malignant pulmonary nodules. Furthermore, the clinical standard for metabolic imaging¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography (PET)is significantly constrained by a high incidence of false-positive findings in patients with inflammatory conditions and frequent false-negative results in tumors exhibiting low glycolytic activity. To overcome these critical limitations, highly specific PET radiotracers have been developed to target immune checkpoints, cell surface receptors, and distinct features of the tumor microenvironment. These molecularly targeted probes offer improved biological specificity, enabling more precise visualization of tumor pathophysiology and supporting applications in early diagnosis, real-time treatment monitoring, and prognostic stratification. This review summarizes recent advances in targeted PET imaging and critically evaluates the potential of integrating these approaches with multimodal imaging, liquid biopsy, and artificial intelligence (AI)-driven radiomics to enhance diagnostic accuracy and inform therapeutic decision-making. By synthesizing key developments from both preclinical studies and clinical trials, we highlight the translational significance and future directions of target-specific PET tracers. Despite existing challenges related to tracer development, regulatory approval, and methodological standardization, the integration of next-generation whole-body PET systems with advanced artificial intelligence algorithms holds considerable promise for establishing molecular-targeted PET as a cornerstone of precision oncology in lung cancer management.

1. Introduction

Lung cancer remains the leading cause of cancer-related deaths globally, with a persistently low five-year survival rate despite significant advances in therapeutic strategies, including targeted therapy, immunotherapy, and combination treatments. − A substantial proportion of patients are diagnosed at advanced stages, at which curative interventions are frequently no longer feasible, and for whom treatment responses and recurrence risks exhibit considerable heterogeneity. As a result, precision medicine has become increasingly critical in the clinical management of lung cancer with imaging, molecular biomarkers, and multimodal diagnostic approaches now recognized as indispensable tools for guiding personalized treatment decisions.

Low-dose computed tomography (LDCT) is considered the cornerstone of lung cancer screening and early detection, however, its limitations are increasingly acknowledged in clinical practice. − As a structural imaging modality, LDCT provides minimal biological insight into tumor characteristics. , Repeated LDCT examinations contribute to cumulative radiation exposure, posing potential long-term health risks, particularly in asymptomatic individuals undergoing routine screening. Moreover, the diagnostic accuracy of LDCT in distinguishing benign from malignant lesions (especially in early stage, small, or subsolid nodules such as adenocarcinoma in situ and minimally invasive adenocarcinoma) remains suboptimal. − The high rate of false-positive findings further undermines its reliability by complicating the differentiation between malignant neoplasms and non-neoplastic pulmonary conditions, including infection, inflammatory changes, scarring, and fibrosis. , Moreover, CT cannot provide insights into functional or molecular characteristics such as metabolic activity, receptor expression, or the tumor microenvironment, thereby limiting its applicability in guiding precision treatment strategies and dynamically evaluating therapeutic response.

In contrast, positron emission tomography (PET), particularly integrated PET/CT, has emerged as a cornerstone of functional and molecular imaging in lung cancer. , ¹⁸F-fluorodeoxyglucose (FDG) PET/CT is extensively utilized for tumor staging, assessment of lymph node involvement and distant metastases, and prognostic stratification, and is firmly established in major clinical guidelines. − Moreover, the integration of radiomics and artificial intelligence has significantly enhanced the diagnostic and predictive performance of FDG PET/CT. , Nevertheless, FDG uptake reflects nonspecific glucose metabolism rather than tumor-specific biological characteristics, as heightened accumulation commonly occurs under inflammatory or infectious conditions. Furthermore, certain subtypes of lung adenocarcinoma display low glycolytic activity, increasing the likelihood of false-positive or false-negative findings. , Additionally, FDG PET provides minimal information on molecular targets, immune status, or tumor microenvironmental signals, thereby limiting its applicability in precision oncology. ,

These limitations have driven the development of targeted PET tracers specifically designed to probe the distinct molecular and cellular features of lung cancer. ImmunoPET, which employs radiolabeled antibodies or antibody fragments, facilitates the noninvasive imaging of key biomarkers (including programmed death-ligand 1 (PD-L1), epidermal growth factor receptor (EGFR), human epidermal growth factor receptor 2 (HER2), and members of the CEACAM family) with high binding affinity and precise biological specificity. Recent early phase clinical studies have reported favorable biodistribution, safety, and strong correlations between imaging signals and immunohistochemical PD-L1 expression for agents such as gallium-68 (⁶⁸Ga) labeled NOTA-WL12 in nonsmall cell lung cancer (NSCLC). In parallel, tracers targeting tumor stroma, immune components, the extracellular matrix, and metabolic pathways have expanded the scope of PET imaging. Fibroblast activation protein (FAP)-targeted radioligands, in particular, have shown high tumor-to-background contrast and the ability to visualize stromal-rich or FDG-low lesions, highlighting their potential for response assessment and treatment guidance. −

Technological advances in radiochemistry, detector sensitivity, reconstruction algorithms, and total-body PET systems have accelerated the clinical translation of these agents. When combined with CT or MR imaging and advanced radiomics or machine learning approaches, targeted PET offers new opportunities to link imaging phenotypes with tumor biology, treatment response, and clinical outcomes. , Recent findings suggest that radiomics-based models incorporating PET/CT data can achieve noninvasive, dynamic prediction of responses to immune checkpoint inhibitors in NSCLC. At the same time, the concept of theranostics has gained momentum in lung cancer, as the same molecular targets used for PET imaging can also serve as platforms for radioligand or radioimmunotherapy when paired with therapeutic radionuclides. Early studies of FAP-targeted radioligands have provided initial evidence supporting the therapeutic potential of such approaches. ,

Despite these promising advances, several critical challenges persist in limiting the widespread clinical implementation of targeted PET imaging. The translational pathway from probe discovery to clinical application demands rigorous optimization of multiple interdependent factors, including synthetic chemistry, radiolabeling efficiency, in vivo stability, metabolic behavior, safety profiling, and dosimetry evaluation. Achieving quantitative consistency across institutions remains a significant hurdle due to variability in imaging equipment, reconstruction algorithms, and timing of image acquisition (factors that collectively impede robust multicenter standardization). Moreover, tumor heterogeneity and therapy-induced dynamics in target expression can lead to heterogeneous tracer uptake patterns, thereby confounding image interpretation and biomarker reliability. To date, large-scale, prospective clinical trials remain insufficient, underscoring the need for comprehensive validation of targeted PET biomarkers and the establishment of statistically robust, clinically meaningful associations with patient outcomes such as survival and disease recurrence. In addition, the integration of PET with complementary data streams (including radiomics, CT or MR imaging, liquid biomarkers, and omics-level biological profiles) introduces substantial analytical complexity, necessitating systematic harmonization, reproducible quantification methods, and sophisticated modeling frameworks to derive actionable insights.

Given these considerations, targeted PET emerges as a highly promising and continually evolving modality in precision oncology for lung cancer, offering significant potential across multiple domains including early diagnosis, stratified therapeutic strategies, real-time treatment monitoring, guidance for radiotherapy and radioimmunotherapy, and prognostic evaluation. Continued progress will depend on overcoming the key barriers related to probe development, clinical standardization, multimodal integration, and prospective validation. This review therefore examines recent advances in targeted PET tracers, evaluates current clinical evidence, discusses opportunities arising from multimodal imaging and AI, and outlines future directions including theranostics, total-body PET, and multitarget imaging strategies.

2. Research Progress in Targeted Molecular PET Tracers for Lung Cancer

Recent progress in PET-based molecular imaging for lung cancer has focused on tracers directed at oncogenic driver-related receptors, tumor microenvironment (TME)-associated markers, and several probes under active development by our group including CEACAM6-targeting antibodies and integrin receptor-directed peptides. Owing to the marked molecular heterogeneity of lung cancer-particularly NSCLC subtypes characterized by EGFR mutations, MET amplification or exon 14 skipping, and GRPR overexpression-targeted PET tracers provide substantial value for early diagnosis, molecular stratification, treatment selection, and longitudinal disease monitoring (Figure ).

1.

Research progress of targeted molecular PET tracers in lung cancer. This schematic overview summarizes the major categories of targeted molecular PET tracers developed for lung cancer imaging, highlighting tumor-intrinsic targets, tumor microenvironment-associated biomarkers, and emerging tracers under active investigation. Tumor receptor-targeting PET probes enable noninvasive, whole-body assessment of key oncogenic drivers and surface antigens, including EGFR, MET, and GRPR in NSCLC, as well as B7H3, DLL3, and Trop2 in SCLC, supporting molecular stratification, therapy selection, and treatment monitoring. In parallel, PET tracers targeting the tumor microenvironment provide functional insights into immune checkpoint expression (PD-L1), immune effector cell infiltration (CD8⁺ T cells), stromal and extracellular matrix components, hypoxia, and metabolic reprogramming, thereby addressing spatial and temporal heterogeneity beyond the limits of tissue biopsy. The lower panel highlights representative innovative tracers developed or evaluated by our group, including CEACAM6-targeted immunoPET and peptide-based probes for subcentimeter pulmonary nodule detection, integrin α_vβ₃-targeted imaging of tumor angiogenesis and metastases, and exploratory PET imaging of the immune checkpoint FGL1. Collectively, these targeted PET strategies extend conventional ¹⁸F-FDG imaging by enabling biologically specific, dynamic, and lesion-level characterization of lung cancer, with broad implications for early detection, personalized therapy guidance, and longitudinal disease surveillance.

2.1. Tumor-Associated Receptor-Targeting Tracers

EGFR is among the most prevalent oncogenic drivers of lung adenocarcinoma. Activating mutations such as L858R and exon 19 deletions lead to constitutive activation of downstream signaling pathways that promote tumor growth. , EGFR-directed PET imaging enables noninvasive, whole-body assessment of receptor expression and has the potential to overcome the spatial and temporal limitations of tissue biopsy. Early studies using small-molecule inhibitor-based tracers, including ¹¹C-PD153035, demonstrated correlations between tracer uptake and EGFR mutation status. ImmunoPET using ⁸⁹Zr labeled anti-EGFR antibodies such as ⁸⁹Zr-cetuximab and ⁸⁹Zr-panitumumab have also been applied to visualize EGFR overexpressing tumors and is suitable for companion diagnostics due to its prolonged circulation time. , To further enhance imaging resolution and improve tumor penetration, researchers have developed nanobodies and antibody fragments labeled with isotopes including ⁶⁴Cu and ⁶⁸Ga. These developments provide EGFR mutant patients with a comprehensive, noninvasive, whole-body map of receptor expression, which may support tyrosine kinase inhibitor (TKI) selection, monitoring of treatment response, and early detection of resistance. Despite these advances, clinical implementation remains limited due to intratumoral heterogeneity of EGFR expression, challenges related to radiotracer pharmacokinetics, and the extended imaging windows required for antibody-based probes. Future studies must address standardization, optimize dosing strategies, and integrate dynamic imaging parameters into predictive models to enhance their clinical applicability and methodological rigor.

Mesenchymal-epithelial transition factor (MET) exon 14 skipping mutations and MET amplification are recognized actionable alterations in NSCLC. Because MET expression often varies across tumor regions, conventional tissue biopsies may not fully capture its heterogeneity. , PET imaging that targets MET provides a noninvasive, whole-body approach for identifying patients with MET exon14 skipping or MET amplification and for predicting treatment efficacy. The peptide- or ligand-based tracer ⁶⁸Ga-EMP-100, which specifically binds c-MET, has been used to visualize MET exon14 positive foci. Data from early clinical and preclinical studies demonstrate a clear association between its uptake and the response to MET TKIs such as capmatinib and tepotinib. Additional probes including ⁶⁸Ga-NOTA-PFCM01 and ⁸⁹Zr-1E7-Fc nanobodies have been created for rapid imaging to facilitate patient selection and early therapeutic evaluation. Overall, MET-targeted PET shows significant potential for improving patient stratification and supporting personalized therapeutic strategies in NSCLC.

Gastrin-releasing peptide receptor (GRPR) is highly expressed in many lung adenocarcinomas while showing minimal expression in normal lung tissue. GRPR expression is associated with tumor aggressiveness, proliferative capacity, and metastatic behavior. GRPR-directed PET tracers such as the antagonist ⁶⁸Ga-RM2 have demonstrated high tumor-to-background ratios and superior tissue specificity compared to ¹⁸F-FDG PET. To improve in vivo stability and enhance pharmacokinetic performance, second-generation peptides such as ⁶⁸Ga-NeoBOMB1 and ⁶⁸Ga-NeoB have been developed and exhibit promising characteristics for future clinical translation.

In small-cell lung cancer (SCLC), several targets with direct therapeutic relevance have emerged. B7H3 (CD276), delta-like protein 3 (DLL3), and trophoblast cell surface antigen-2 (Trop2) are frequently overexpressed and can be visualized using PET. , A novel tracer, ⁶⁸Ga–B7H3-BCH, has demonstrated a high affinity and significant tumor uptake in preclinical and early clinical studies. This agent enables noninvasive, whole-body mapping of B7H3 distribution, thereby improving the precision of pretreatment evaluation and facilitating the monitoring of responses and potential resistance to B7H3-targeted therapies. DLL3, an inhibitory ligand in the Notch signaling pathway expressed on the surface of over 80% of SCLC tumors, can be robustly visualized using immunoPET probes such as ⁸⁹Zr-DFO-SC16-SS. This tracer not only clearly delineates both orthotopic and metastatic SCLC lesions but also quantifies heterogeneous DLL3 expression with high sensitivity, offering a strong foundation for rational patient stratification and longitudinal treatment monitoring. Meanwhile, Trop2, a transmembrane glycoprotein highly expressed in epithelial cancers, is being developed for imaging using nanobody-based radioligands like ⁶⁸Ga-MY6349 in multiple types of cancer. Such tools are valuable for predicting and assessing outcomes of Trop2-directed antibody-drug conjugates and bispecific therapies. Collectively, these approaches provide lesion-level quantification of antigen expression beyond the limits of biopsy, which is particularly valuable in aggressive diseases such as SCLC.

2.2. Tumor Microenvironment (TME)-Targeting PET Tracers

The TME, encompassing immune cells, fibroblasts, vasculature, and metabolic conditions such as hypoxia, plays a central role in tumor progression and therapy resistance. , PET tracers targeting these processes have therefore become an important focus. With the widespread use of immune checkpoint inhibitors, tools capable of assessing the immune status are increasingly needed. Unlike immunohistochemistry, PET enables a dynamic whole-body evaluation.

PD-L1 targeted PET tracers illustrate this potential. PD-L1 expression did not differ significantly before and after concurrent chemoradiotherapy (CRT) treatment. To improve clinical feasibility and reduce the imaging time window, peptide or small-molecule PD-L1 tracers including ⁶⁸Ga-NOTA-WL12 have been designed and tested in first-in-human studies. A prospective study in operable NSCLC patients receiving neoadjuvant immunochemotherapy found that baseline ⁶⁸Ga-NOTA-WL12 SUVmax showed a strong positive correlation with major pathologic response, whereas ¹⁸F-FDG PET/CT primarily reflected size changes and lacked predictive value. These findings suggest that PD-L1 PET may serve as a superior companion diagnostic tool for guiding immunotherapy. PET tracers targeting T-cell infiltration, such as the anti-CD8 alpha antibody fragment ⁸⁹Zr-IAB22M2C, are also under investigation for monitoring T-cell dynamics during immunotherapy. We also developed a PET imaging agent, ⁶⁸Ga-NODAGA-SNA006, that targets CD8⁺T cells for in vivo quantification of CD8 expression with a favorable safety profile. It holds potential for dynamically tracking CD8⁺T cells and monitoring responses to personalized cancer immunotherapy. , Combining tumor-targeted PET with PD-L1 and CD8 imaging may improve the discrimination between true progression and pseudoprogression.

Stromal elements such as cancer-associated fibroblasts, tumor angiogenesis, and extracellular matrix remodeling are key drivers of invasion, metastasis, and therapeutic resistance. PET tracers targeting these processes, including probes directed at fibronectin, integrins, FAP, and extracellular matrix molecules, are actively under investigation. The integrin αvβ6-targeted tracer ⁶⁸Ga-Trivehexin and FAP-target tracer ¹⁸F-FAPI were recently reported to outperform FDG PET in sensitivity and specificity for preoperative lymph node staging in lung cancer. A prospective head-to-head comparison involving 34 patients with advanced lung cancer showed that ⁶⁸Ga-FAPI PET/CT detected significantly more suspected metastatic lesions than ¹⁸F-FDG PET/CT, including lymph nodes (356 vs 320), brain (23 vs 10), bone (109 vs 91), and pleura (66 vs 35), highlighting its superior performance for metastatic detection and accurate M-staging in lung cancer. Metabolic and hypoxia tracers further expand the functional assessment. Agents such as ¹⁸F-FSPG reflect glutamine metabolism and may be informative in KRAS mutant tumors, , while hypoxia tracers including ¹⁸F-FMISO and ¹⁸F-FAZA can guide radiotherapy planning.

2.3. Research Progress of Other Types of Targeted PET Tracers in Lung Cancer

Using machine learning approaches, our group identified CEACAM6 as a biomarker distinguishing malignant from benign pulmonary nodules in lung adenocarcinoma. CEACAM6 overexpression correlated with disease stage and prognosis. − Based on this discovery, we developed a novel immunoPET probe using an ⁸⁹Zr labeled CEACAM6 antibody (⁸⁹Zr-DFO-Tinurilimab). Cell uptake assays confirmed its selective uptake in CEACAM6 positive A549 cells. PET imaging in tumor-bearing mice showed strong in vivo targeting with significant accumulation in CEACAM6 positive tumors and uptake proportional to CEACAM6 expression. The tumor-to-background ratio reached 9.93, markedly higher than that of ¹⁸F-FDG, and the probe enabled detection of malignant nodules as small as approximately 2 mm⁸⁵. These results demonstrate the capability of CEACAM6 targeted PET to overcome the limited sensitivity of FDG PET for subcentimeter nodules. To meet the need for rapid diagnosis, our group has also developed a CEACAM6-targeted peptide probe, ⁶⁸Ga-NODA-P3, which achieved a tumor-to-background ratio of 7.68 in xenograft models, indicating promising translational potential.

Regarding tumor neovascularization, our team created the integrin α_vβ₃ targeted tracer ¹⁸F-Alfatide. Clinical studies in lung cancer patients showed that ¹⁸F-Alfatide clearly delineated tumor angiogenesis, with imaging results closely matching pathological findings. Compared with ¹⁸F-FDG, ¹⁸F-Alfatide exhibited superior specificity and positive predictive value for differentiating benign from malignant lesions. It also significantly improved the accuracy of detecting brain and bone metastases. In addition, our team has initiated the development of a PET probe targeting immune checkpoint fibrinogen-like protein 1 (FGL1). The tracer ⁶⁸Ga-NODAGA-FGLP21 demonstrated excellent specificity in FGL1 positive tumor models, with a tumor-to-background ratio of 8.2. Further preclinical studies of the lung cancer immune microenvironment are ongoing.

Targeted molecular PET tracers for lung cancer now span a wide range of biological processes, including driver receptors such as EGFR and MET, tumor-associated targets such as GRPR, CEACAM6, DLL3, B7H3, and Trop2, TME-related markers, and metabolic pathways. Antibody-based tracers offer high specificity but require longer imaging windows, whereas peptide and small-molecule tracers enable rapid imaging with a lower immunogenicity. Microenvironmental and metabolic tracers provide additional insights into heterogeneity, resistance, and recurrence. Given that a substantial body of preclinical and clinical studies has already comprehensively evaluated these PET tracers across multiple biological dimensions (including targeting specificity, pharmacokinetics, safety, and imaging performance), the following section shifts focus toward their concrete clinical manifestations. Building on this progress, the next section will provide a comprehensive examination of the application of these tracers in key clinical contexts, encompassing early detection, disease staging, therapy selection, immunotherapy guidance, radiotherapy planning, and recurrence monitoring.

3. Clinical Applications of Targeted Molecular PET in Lung Cancer

As precision medicine continues to reshape lung cancer care, PET molecular imaging has progressed from conventional ¹⁸F-FDG glucose metabolism imaging to a broad suite of targeted radiotracers. These agents are designed to interrogate biologically relevant features, including tumor driver receptors such as EGFR, MET, and GRPR, immune checkpoints including PD-L1, infiltrating immune effector cells such as CD8⁺ T cells, hypoxia signatures, and key metabolic pathways. By enabling in situ, whole-body, and dynamic characterization of tumor biology, targeted PET has demonstrated clinical relevance across multiple applications including early detection, pathological subtyping, selection of targeted and immune therapies, monitoring of treatment response, evaluation of radiotherapy and chemotherapy effects, and prognosis prediction. This section summarizes current evidence, clinical advantages, and remaining challenges of targeted PET in major lung cancer scenarios (Figure ).

2.

Clinical applications of targeted molecular PET imaging in lung cancer. (A) In high-risk populations, targeted PET tracers overcome the limited specificity of low-dose CT and the reduced sensitivity of ¹⁸F-FDG PET for small or low-metabolic lesions. GRPR-targeted PET, FAPI PET, ¹⁸F-FSPG and CEACAM6-directed immunoPET enable improved detection and characterization of indeterminate pulmonary nodules, supporting molecular triage and earlier clinical intervention. (B) Whole-body PET imaging of tumor-associated receptors and stromal components enables noninvasive assessment of intratumoral and interlesional heterogeneity beyond biopsy sampling. EGFR-, MET-, GRPR-, and FAP-targeted PET tracers support accurate molecular subtyping, patient stratification, and selection of targeted therapies, including EGFR and MET inhibitors. (C) Immuno-PET probes targeting PD-L1 visualize spatially heterogeneous immune checkpoint expression, whereas CD8⁺ T-cell PET and granzyme B PET provide complementary information on immune cell infiltration and effector function. Together, these mechanistically distinct biomarkers enable early prediction of response, differentiation of true progression from pseudoprogression, and dynamic monitoring during immune checkpoint inhibitor therapy. (D) Mechanism-based PET tracers, including hypoxia (¹⁸F-FMISO, ¹⁸F-FAZA), proliferation (¹⁸F-FLT), angiogenesis (integrin α_vβ₃ and VEGF-related probes), stromal remodeling (FAPI), and metabolic pathway imaging (¹⁸F-FSPG), allow early evaluation of radiotherapy, chemotherapy, and antiangiogenic treatment effects. Targeted PET further enhances detection of recurrence and micrometastatic disease and enables integration of quantitative imaging metrics with radiomics and circulating tumor DNA (ctDNA) to support adaptive treatment strategies and longitudinal disease surveillance.

3.1. Early Diagnosis and Screening in High-Risk Populations

LDCT has been widely adopted as a first-line screening tool for high-risk individuals, but its limited specificity and high false positive rate of 20 to 40% compromise its clinical utility and often lead to unnecessary follow-up procedures. , Furthermore, its sensitivity for small or low-metabolic lesions such as adenocarcinoma in situ and minimally invasive adenocarcinoma remains suboptimal. Although ¹⁸F-FDG PET provides a metabolic assessment, its sensitivity is insufficient for lesions with low glycolytic activity, and differentiating benign from malignant findings remains challenging.

Targeted molecular PET addresses these limitations by exploiting disease-specific biological features. Receptors such as GRPR are highly expressed in lung adenocarcinoma, and GRPR-targeted PET probes including ⁶⁸Ga-RM2 and ⁶⁸Ga-NeoBOMB1 achieve high tumor-to-background contrast in small pulmonary nodules, improving detection of lesions ≤ 1 cm and offering a complementary option for FDG-negative or low-uptake nodules. , Fibroblast activation protein inhibitor (FAPI) PET/CT exhibits superior diagnostic performance over ¹⁸F-FDG PET/CT in evaluating indeterminate pulmonary nodules and staging NSCLC, particularly in distinguishing benign from malignant lesions and identifying metastatic disease. FAPI PET/CT demonstrated a sensitivity of 99% (95% CI: 0.90–1.00) for detecting metastatic lesions in lung cancer (including lymph node and distant metastases), markedly superior to the 77% (95% CI: 0.66–0.85) achieved by ¹⁸F-FDG PET/CT. Similarly, ¹⁸F-FSPG PET, reflecting glutamate and oxidative stress pathways, has demonstrated diagnostic accuracy comparable to FDG, with trends toward improved sensitivity and negative predictive value. For detecting high-grade malignant tumors, ¹⁸F-FDG PET showed a sensitivity of 75%, specificity of 100%, positive predictive value (PPV) of 100%, and negative predictive value (NPV) of 33%, while ¹⁸F-FSPG PET performed better with a sensitivity of 88%, maintaining perfect specificity and PPV of 100%, and raising the NPV to 50%.

Preclinical studies from our group demonstrated that CEACAM6-targeted immuno-PET can identify malignant nodules as small as 2 mm in orthotopic lung cancer models with greater sensitivity than FDG. In urethane-treated mice bearing small malignant nodules (mean diameter 1.39 mm), the tumor-to-background ratio (TBR) for ¹⁸F-FDG was markedly lower than that seen with ⁸⁹Zr-Df-tinurilimab (2.68 ± 1.19 vs 9.93 ± 1.59, p < 0.001), highlighting the superior targeted uptake of the CEACAM6-directed tracer in these lesions. EGFR-targeted small-molecule PET tracers, such as ¹⁸F-MPG, have also shown strong correlations between tracer uptake and activating EGFR mutations, supporting their potential role as a molecular screening layer following LDCT to guide biopsy or surveillance decisions. Additional targets, including CXCR4 imaged with ⁶⁸Ga-pentixafor, may capture invasive or premalignant features not evident on conventional imaging, although clinical data in lung cancer remain limited. ,

For individuals with known genetic risk, high-risk exposures, or familial predisposition, a longitudinal targeted PET assessment may enable early detection of molecular changes preceding anatomical progression. Combined with genomic testing, serial PET evaluation could support earlier intervention and more proactive disease management.

3.2. Clinical Subtyping, Stratification, and Targeted Therapy Decision-Making

Precision therapy in NSCLC depends upon accurate pathological and molecular subtyping, yet tissue biopsies are constrained by sampling bias and intratumoral heterogeneity. Targeted PET overcomes these limitations by noninvasively assessing receptor and pathway expression across the entire tumor burden, thereby enhancing the accuracy of treatment decisions.

EGFR remains a critical driver in NSCLC, and PET probes targeting EGFR such as ¹⁸F-MPG, ¹¹C-PD153035, and antibody or nanobody based constructs have shown promise in both preclinical and early clinical studies for predicting response to EGFR tyrosine kinase inhibitors. ¹⁸F-MPG PET has been shown to differentiate activating EGFR mutations from wild-type tumors and to map the spatial distribution of resistant subclones such as those harboring T790 M or C797S mutations. This capability is especially valuable for patients with limited access to biopsy procedures as it allows for a comprehensive assessment of tumor heterogeneity and provides critical support in guiding individualized treatment strategies.

Alterations in MET, including exon 14 skipping and gene amplification, represent important actionable variants in NSCLC. Targeted MET inhibitors such as capmatinib and tepotinib have demonstrated meaningful clinical benefit in these subgroups, and PET probes targeting MET (such as ⁸⁹Zr-onartuzumab) have shown uptake patterns consistent with MET status, supporting their use in identifying patients likely to benefit from MET targeted therapy. Although available clinical data are limited, early studies suggest that MET PET tracers can both stratify patients and predict therapeutic responses, and ongoing trials are evaluating these agents in larger cohorts.

GRPR expression is elevated in many lung adenocarcinomas, and GRPR-targeted PET delivers imaging contrast distinct from FDG PET. ⁶⁸Ga-NeoBOMB1 targets GRPR-overexpressing subtypes (e.g., aggressive adenocarcinoma) for visualization and potential theranostic stratification, though clinical data remain scarce. Additionally, ⁶⁸Ga-FAPI or ¹⁸F-FAPI PET/CT, which targets cancer-associated fibroblasts, outperforms ¹⁸F-FDG in detecting metastases (lymph node, brain, bone) and staging low FDG-avidity subtypes like adenocarcinoma. , This supports microenvironment assessment and potential theranostic stratification, especially for distinguishing NSCLC from SCLC.

3.3. Immunotherapy Response Prediction and Personalized Assessment

Immune checkpoint inhibitors have become standard treatments for many patients with NSCLC, but PD-L1 immunohistochemistry remains limited by sampling variability and intratumoral heterogeneity. Targeted immuno-PET enables real-time, whole-body assessment of immune checkpoint expression and the broader immune milieu. PD-L1 targeted PET probes include antibody labeled formats such as ⁸⁹Zr-atezolizumab and ⁸⁹Zr-durvalumab and smaller nanobody or peptide-based agents such as ⁶⁸Ga-THP-APN09 and ⁶⁸Ga-WL12. Early clinical studies have demonstrated that PD-L1 PET can localize expression with high specificity and may correlate with immunotherapy outcomes more effectively than tissue-based assays alone, suggesting a role for PET as a dynamic biomarker in immunotherapy selection. For instance, ¹⁸F-AlF-APN09 can visualize PD-L1 expression in NSCLC and positively correlated with PD-1 treatment response.

Monitoring of tumor-infiltrating CD8⁺T cells using PET imaging with CD8 specific probes provides complementary information about the immune response. The ⁸⁹Zr-Df-IAB22M2C minibody demonstrated safety and specific uptake in CD8-rich tissues including spleen, lymph nodes, and tumor lesions in a phase I first-in-human study of patients with metastatic solid tumors including NSCLC, where increased tracer uptake post-ICI was observed in responders. Subsequent clinical PET/MRI experiences confirmed feasibility of assessing CD8⁺ T-cell distribution and revealed interpatient variability in lymphoid and metastatic uptake that may relate to immune response patterns. , Phase II trials of ⁸⁹Zr-Df-IAB22M2C are underway to evaluate its predictive value in immunotherapy contexts (NCT03802123). Alternative formats such as nanobody- and small-molecule-based CD8 tracers with faster kinetics and improved signal-to-background ratios are also progressing toward clinical evaluation.

In addition to the cellular presence, imaging of effector function has gained traction. PET tracers targeting granzyme B (GZB), a serine protease released by activated cytotoxic lymphocytes at the tumor interface, provide a mechanistic readout of immune killing. Preclinical and early translational work demonstrates that granzyme B PET (such as ⁶⁸Ga-NOTA-GZP) can discriminate responders from nonresponders earlier than conventional biomarkers by directly reporting effector activity, with ongoing reviews highlighting its potential advantages in stratifying treatment outcomes and differentiating pseudoprogression.

Collectively, these targeted PET probes (PD-L1, CD8 and granzyme B) offer a spectrum of mechanistically relevant biomarkers encompassing immune checkpoint expression, effector cell infiltration, and functional activation. Critical challenges include harmonizing quantification standards, confirming prospective outcome correlations in large NSCLC cohorts, optimizing imaging timing relative to therapy, and demonstrating that PET-guided decisions improve clinical outcomes. Multicenter trials and standardized protocols will be essential to transition these targeted PET biomarkers into routine precision immunotherapy workflows.

3.4. Efficacy Assessment for Radiotherapy, Chemotherapy, and Antiangiogenic Therapy

Tumor hypoxia is a well-defined mechanism of resistance to radiotherapy, and hypoxia PET probes such as ¹⁸F-FMISO and ¹⁸F-FAZA enable noninvasive visualization of hypoxic regions. , These tools support radiotherapy planning by identifying radioresistant subvolumes and correlating them with local control outcomes. For instance, dose escalation of intensity-modulated radiotherapy for lung cancer can be guided by ¹⁸F-FMISO PET. Proliferation imaging with ¹⁸F-FLT offers a mechanism-specific readout of DNA synthesis, and prospective pilot studies in lung cancer have demonstrated that reductions in FLT uptake within days to weeks of cytotoxic chemotherapy or concurrent chemoradiation correlate with pathological response and progression-free survival, supporting FLT as an early adaptive biomarker for chemotherapy/radiotherapy scheduling. For antiangiogenic strategies, integrin α_vβ₃ RGD-based tracers (such as ¹⁸F-Alfatide) and VEGF/VEGFR-targeting radioligands (including exploratory ⁶⁴Cu-NOTA-RamAb imaging) permit noninvasive measurement of vascular target expression and vascular normalization dynamics. Translational and small clinical studies report that changes in RGD or VEGF tracer uptake mirror pharmacodynamic effects of antiangiogenic agents and may predict subsequent radiographic control. In addition, FAPI provide high-contrast imaging of tumor stroma and several recent clinical series show that early changes in ⁶⁸Ga-FAPI uptake after radiotherapy or chemoradiation reflect stromal remodelling and associate with local control and fibrosis risk. Fan, Min et al. reported that prediction of radioligand therapy in patients with NSCLC using ⁶⁸Ga-FAP-2286 PET/CT. Gao, Haiyan et al. reported a dual-targeted heterodimer tracer (Al¹⁸F-LNC1007) targeting FAP and integrin α_vβ₃, which had better uptake than ¹⁸F-FDG and Al¹⁸F-FAPI-04 in both primary and metastatic tumor lesions. However, remaining barriers to routine clinical implementation include the need for larger prospective, histology-stratified trials, standardized quantification and imaging-timing protocols, tracer production scalability, and the demonstration that tracer-guided adaptations improve patient-level outcomes.

Metabolic PET tracers such as ¹⁸F-FSPG, which reflects glutamate and oxidative stress pathways has shown potential to detect metabolic responses before volumetric changes occur, complementing conventional assessment methods and enabling earlier evaluation of treatment effect. , In addition, quantitative parameters derived from targeted PET, including standardized uptake values, receptor density estimates, and dynamic changes over time, are closely associated with clinical outcomes in lung cancer. Integrating PET imaging data with radiomics, circulating tumor DNA, and multiomics profiling facilitates construction of molecular imaging fingerprints that enhance individualized survival prediction models.

In summary, a growing body of clinical evidence supports the feasibility and potential clinical value of targeted molecular PET imaging across the entire lung cancer care continuum. Larger prospective studies, standardized protocols, and clear demonstration of measurable outcome benefits are now essential to definitively establish its role in routine clinical practice and strengthen the implementation of PET-guided precision oncology strategies.

4. Multimodal Imaging: PET with CT/MRI and Circulating Biomarkers

While individual imaging modalities offer distinct advantages, their clinical utility is often limited when applied in isolation. By integrating complementary structural, functional, and molecular data, multimodal imaging effectively addresses these limitations, thereby enhancing diagnostic accuracy, enabling more precise treatment response assessment, and improving prognostic stratification. , This section systematically reviews recent advances in the integration of positron emission tomography (PET) with computed tomography (CT) and magnetic resonance imaging (MRI), and evaluates the emerging role of combining PET with peripheral blood-based biomarkers (such as circulating tumor DNA (ctDNA), circulating tumor cells (CTCs), and proteomic or metabolomic profiles) in facilitating personalized and precision-driven management of lung cancer.

4.1. PET/CT and PET/MRI with Molecular Tracers: Improved Biological Resolution

The core advantage of multimodal imaging resides in the integration of PET-derived molecular and metabolic data with the high-resolution structural and histological contrast provided by CT and MRI, which significantly improves the detection and staging accuracy for primary tumors, lymph node involvement, and distant metastases. In radiotherapy planning, the incorporation of metabolically active tumor volumes (MTV) and total lesion glycolysis (TLG, delineated by PET) as weighted targets has become a key component of dose-escalation strategies and experimental protocols designed to minimize radiation exposure to surrounding healthy tissues. While some studies have demonstrated a reduction in radiation-induced toxicity, evidence regarding overall survival benefit continues to accumulate.

Hybrid PET/CT and PET/MRI systems provide an anatomical scaffold that allows tracer-specific signals to be interpreted within a clinically meaningful spatial context. While FDG PET/CT remains widely used, multiple lung cancer-relevant tracers now demonstrate distinct and complementary advantages when integrated with CT or MRI. , FAPI represents one of the most prominent examples. Clinical studies using ⁶⁸Ga-FAPI-04 and ⁶⁸Ga-FAPI-46 PET/CT have shown markedly higher tumor-to-background ratios than FDG in both primary NSCLC and metastatic lesions, particularly in adenocarcinoma and low-FDG-avid tumors. ,, When fused with CT, FAPI PET improves lymph node staging accuracy and refines gross tumor volume delineation for radiotherapy, highlighting the contribution of tumor stroma to the disease extent and potential treatment resistance.

Proliferation-targeted imaging with ¹⁸F-FLT PET/CT provides complementary information to FDG by reflecting thymidine kinase-1 activity and DNA synthesis. In NSCLC patients undergoing radiotherapy or chemoradiotherapy, FLT uptake has been shown to decrease earlier than FDG uptake, suggesting value for early treatment response assessment when interpreted alongside CT-based anatomical changes. , Hypoxia-specific tracers, including ¹⁸F-FMISO, ¹⁸F-FAZA, and ¹⁸F-HX4, have been evaluated in NSCLC to map intratumoral hypoxic subvolumes associated with radioresistance and poor prognosis. , When coregistered with CT or MRI, hypoxia PET enables biologically guided radiotherapy planning, including dose escalation to resistant regions, an approach not achievable with anatomical imaging alone.

Beyond these, receptor-targeted tracers also demonstrate an added value in hybrid imaging. ⁶⁸Ga-DOTATATE PET/CT is indispensable in pulmonary neuroendocrine tumors, where it outperforms FDG in well-differentiated lesions and directly guides peptide receptor radionuclide therapy. , Similarly, early clinical studies using EGFR-targeted PET tracers (e.g., ¹¹C-erlotinib, ¹¹C-PD153035) illustrate how hybrid imaging can visualize drug-target engagement and spatial heterogeneity of receptor expression within CT-defined lesions.

Collectively, these examples illustrate that hybrid imaging is not merely a localization tool but a platform through which tracer-specific biological signals gain spatial and clinical relevance.

4.2. Integration of Targeted PET Imaging with Circulating Biomarkers: Spatial Molecular Complementarity

While PET/CT or PET/MRI provides spatially resolved information, circulating biomarkers capture dynamic molecular changes at the systemic level. The integration of specific PET probes with ctDNA and CTC analysis addresses the fundamental limitations of each modality alone. Multiple studies indicate that changes in ctDNA levels during the early weeks of systemic therapy or chemoradiotherapy can signal biological response or early progression sooner than imaging findings. Combining ctDNA with PET metabolic parameters (SUV, MTV, and TLG) improves the ability to predict treatment failure or recurrence risk. For instance, a representative cohort of 141 patients from the RELEVANCE phase 3 trial showed that only combined ctDNA and PET identify patients at high risk of early relapse (24 months of diagnosis, POD24), linked to poor survival. In a small multicenter prospective cohort of advanced NSCLC, integrated changes in ctDNA and FDG-PET metrics within the first 1–3 weeks of treatment identified patients with treatment failure ahead of conventional imaging. For minimal residual disease (MRD) monitoring after curative treatment, the concurrent presence of ctDNA positivity and volumetric or metabolic abnormalities on follow-up PET/CT provides a more reliable prediction of recurrence. In clinical practice, PET/CT, liquid biopsy, and AI provide complementary and quantifiable value through multimodal integration. In a multicenter cohort of 394 patients with NSCLC, PET/CT habitat imaging stratified patients into three recurrence risk groups with significant prognostic separation independent of clinicopathologic factors. High risk imaging subtypes showed persistent ctDNA positivity in approximately 67% of patients, whereas 62% of low-risk patients were ctDNA negative. By integration of PET/CT habitat features, ctDNA status, tumor volume, and clinical variables, an AI driven Cox model achieved a concordance index of 0.82 for recurrence free survival prediction, outperforming single modality models. This workflow illustrates how spatial imaging, molecular dynamics, and AI based integration jointly enable precise risk stratification. , In addition, in terms of specific PET tracer, PD-L1 targeted PET imaging, using tracers such as ⁸⁹Zr-atezolizumab, has revealed marked inter- and intrapatient heterogeneity of PD-L1 expression across tumor lesions in NSCLC. When interpreted together with ctDNA mutational burden or immune-related gene alterations, PD-L1 PET may offer a whole-body, noninvasive assessment of immunotherapy targets that complements tissue biopsy and liquid profiling. ,,

These multimodal strategies highlight how specific PET tracers provide spatial biological context, while circulating biomarkers offer molecular sensitivity, together enabling earlier detection of treatment failure and minimal residual disease than either approach alone.

4.3. Radiomics and Radiogenomics with Targeted PET Tracers: From Imaging Phenotypes to Molecular Programs

Radiomics and radiogenomics represent a higher-order integration of multimodal data, translating imaging patterns into biologically meaningful signatures, effectively bridging noninvasive imaging with tumor genotype, gene expression programs, and microenvironmental phenotypes. Within lung cancer research, radiomics pipelines applied to PET/CT or PET/MRI now extend beyond conventional ¹⁸F-FDG uptake metrics to include multiple targeted PET tracers that capture diverse biological processes. Integration of these tracer-specific imaging phenotypes with genomic data deepens understanding of oncogenic programs, enables noninvasive molecular stratification, and supports precision therapy selection.

In NSCLC, radiomic analysis of ¹⁸F-FDG PET/CT has been extensively studied for mutation prediction and clinical outcome modeling. For example, retrospective radiomics models have successfully predicted EGFR mutation status from ¹⁸F-FDG PET/CT images using machine learning, achieving area under the curve (AUC) values exceeding 0.80 in some cohorts and demonstrating the ability to noninvasively infer molecular driver status. This links metabolic heterogeneity with underlying genotype and suggests a role for radiomics in stratifying patients for EGFR-targeted therapies without repeat biopsy. Beyond FDG, the pilot radiomics analysis comparing ⁶⁸Ga-DOTA FAPI-46 and ¹⁸F-FDG PET/CT in NSCLC demonstrated that stromal-targeted imaging features differ significantly from conventional metabolic metrics, revealing complementary information about tumor microenvironment and intratumoral heterogeneity. , These differences were quantified across intensity and texture features, suggesting that FAPI radiomics could capture the activation of cancer-associated fibroblasts and related signaling pathways not reflected by glucose metabolism alone.

Radiogenomic integration also linked PET-derived radiomic signatures with transcriptomic profiles. A large NSCLC study developed a prognostic radiomic risk score from ¹⁸F-FDG PET features and correlated radiomic clusters with distinct oncogenic pathways and immune cell signatures. Patients stratified into high-risk radiomic groups exhibited transcriptomic hallmarks of immune suppression and aggressive behavior, validating that imaging phenotypes can reflect underlying biological programs. Additionally, PET/CT radiomic signatures have been used to predict mutations in cancer pathways such as KEAP1/NFE2L2, which are associated with radiation resistance and relapse patterns, suggesting a role for imaging in guiding adaptive therapy strategies.

Radiomic predictions have also been applied to the expression of immune checkpoints such as PD-L1. Radiomic models derived from ¹⁸F-FDG PET/CT demonstrated moderate ability to differentiate PD-L1 positive versus negative tumors, offering a noninvasive approach to immunotherapy target identification when tissue is limited or heterogeneous. Systematic reviews of PET-based radiogenomics in oncology show that most studies (∼75% focused on lung cancer) find statistically significant associations between PET radiomic features and key oncogenic alterations including EGFR, KRAS, TGFβ, and immune pathways, although standardization and multicenter validation remain challenges. Radiomics and radiogenomics integrating tracer-specific PET features with molecular data have shifted imaging from descriptive uptake measures to interpretable biomarkers reflective of molecular programs, supporting noninvasive tumor profiling, risk stratification, and individualized therapy selection in lung cancer. Recent AI advances have enabled multimodal PET/CT models to improve outcome prediction in lung cancer. PET/CT radiomics combined with machine learning can predict immunotherapy response and long-term survival more accurately than conventional metrics. Prognostic models integrating PET features with clinical variables achieve AUCs > 0.85¹⁴⁷. Moreover, thorax-encompassing deep learning models using whole-chest PET/CT and clinical data outperform TNM staging in predicting recurrence-free survival. These findings highlight the value of AI-driven PET/CT fusion for precision prognostication in lung cancer.

In summary, multimodal integration plays a pivotal role in advancing precision imaging and therapeutic decision-making in lung cancer. PET/CT has emerged as a clinical standard for the integration of structural and functional information, and the AI-driven fusion of multimodal features further augments diagnostic accuracy and predictive performance. PET/MRI provides distinct advantages in soft tissue characterization and in clinical contexts where minimizing radiation exposure is critical, although several technical challenges remain to be resolved. The integration of PET imaging with peripheral blood biomarkers represents a promising approach for achieving deeper synergy between imaging and molecular data, thereby facilitating a transition toward truly multidimensional, precision-based diagnosis and treatment in lung cancer care (Figure ).

3.

Multimodal integration of targeted PET imaging with CT/MRI and circulating biomarkers in lung cancer. This figure illustrates the synergistic value of multimodal imaging for precision management of lung cancer. Hybrid PET/CT and PET/MRI integrate tracer-specific molecular information with high-resolution anatomical context, enabling accurate tumor detection, staging, and biologically guided radiotherapy using metabolic and functional parameters such as SUV, MTV, TLG, and hypoxic subvolumes. The combination of targeted PET imaging with circulating biomarkers, including ctDNA, circulating tumor cells, and proteomic or metabolomic markers, provides complementary spatial and systemic molecular information, allowing earlier assessment of treatment response, improved prediction of recurrence, and sensitive detection of minimal residual disease. Radiomics and radiogenomics further integrate multimodal imaging features with genomic and transcriptomic data, translating imaging phenotypes into biologically meaningful signatures that support noninvasive molecular stratification, prognostic assessment, and individualized therapy selection in lung cancer.

5. Technical Challenges and Clinical Translation Barriers

Although targeted PET molecular imaging has emerged as a powerful tool for the precision management of lung cancer, its integration into routine clinical practice is still limited by various technical, developmental, and regulatory challenges. This section presents a systematic evaluation of these barriers, with a focus on the inherent heterogeneity of target expression, the complexities associated with radiotracer design and translational development, and the operational and regulatory factors that impede widespread clinical adoption. Through representative examples, it highlights how these challenges influence image interpretation, probe optimization, and real-world deployment and outlines potential strategies to facilitate more efficient and reliable clinical translation.

5.1. Heterogeneity in Specific Target Expression

The clinical accuracy and interpretability of targeted PET depend fundamentally on the presence and distribution of the molecular target within the tumor burden. Two related but distinct problems substantially limit the reliability of single time point, single tracer PET in clinical decision making. The first problem is spatial heterogeneity, which reflects variable expression of the imaging target between different regions of a primary lesion and among separate metastatic deposits. , Multiple early clinical experiences and translational studies with immune checkpoint and other receptor-targeted tracers have documented marked intra- and interlesional variability in target abundance, such that a single PET scan may overestimate or underestimate the fraction of tumor that expresses a given biomarker. For example, recent translational PD-L1 imaging work using peptide and antibody tracers has highlighted substantial lesion-level discordance in PD-L1 signal across a patient’s disease sites, demonstrating that spatial heterogeneity can materially affect treatment assignment if imaging is interpreted without context. −

The second problem is temporal heterogeneity, which describes dynamic changes in target expression over time and in response to therapy. Target abundance and accessibility may change rapidly during systemic therapy or after local treatments, such as radiotherapy. Immune checkpoint molecules, receptor tyrosine kinases, and stromal markers have all been shown to upregulate or downregulate at intervals that are short relative to conventional imaging follow up. Temporal shifts may therefore produce false impressions of early response or treatment failure when PET signals are read without a serial context. Serial imaging studies and prospective immuno-positron emission tomography (immuno-PET) investigations are increasingly emphasizing the value of multitime point acquisition to capture these dynamics and to avoid misleading single-time point interpretations. , Radiomic quantification of intratumoral heterogeneity, including texture metrics and distributional analyses of SUV values, provides a practical approach to summarize spatial heterogeneity from PET images and has been integrated into predictive models in several recent studies. ,,, These radiomic features, when combined with serial imaging, improve the characterization of spatial and temporal heterogeneity beyond conventional summary metrics.

To mitigate the impact of heterogeneity in clinical workflows, practical strategies include whole-body PET acquisition with systematic multiregion region of interest analysis, protocolized multitime point imaging for key decision windows, and incorporation of heterogeneity metrics into multidisciplinary molecular tumor boards. In parallel, hybrid strategies that combine targeted PET with circulating biomarkers that reflect the total tumor biology can reduce reliance on any single imaging snapshot and improve confidence in therapeutic decisions.

5.2. Technical Challenges in Probe Development and Translation

Development of clinically useful targeted PET tracers requires close alignment among ligand design, radiochemistry, regulatory compliance, and manufacturing capacity. Radiotracer development differs from conventional small molecule drug discovery because tracers must satisfy tight constraints on rapid radiolabeling, metabolic stability during the imaging window, specific activity, and predictable pharmacokinetics so that signal reflects target binding rather than circulating fraction or nonspecific retention. , Common PET radionuclides each impose characteristic trade-offs. ¹¹C enables near-tracer-level chemistry but has a very short half-life that limits its distribution to centers with on-site cyclotrons. ¹⁸F offers favorable decay characteristics for centralized production and distribution but may require complex synthesis routes for certain ligands. ⁶⁸Ga facilitates kit-style labeling from generators and is attractive for peptides, while ⁸⁹Zr enables imaging of antibody formats because of its long half-life but increases cumulative radiation exposure and requires careful chelator chemistry to avoid in vivo demetalation and off-target activity. , The complexity and safety profile of zirconium radiochemistry have been the subject of recent methodologic reviews and good practice recommendations. ,

From a production and regulatory perspective, translation to routine clinical use requires cGMP compliant synthesis, robust quality control workflows, and often on-demand manufacturing for short-lived isotopes. Regulatory frameworks and guidance for PET drugs emphasize adherence to current good manufacturing practices and compel sponsors to validate manufacturing and release testing processes, which raises the bar for academic investigators seeking clinical translation. Recent workshops and regulatory summaries have highlighted the need for standardized manufacturing pathways and early engagement with regulatory authorities to accelerate safe translation. Differences in national regulatory requirements and the technical burden of establishing radiopharmacy capacity represent nontrivial barriers for broad adoption. −

Probe format choices drive distinct technical profiles and clinical trade-offs. Peptide tracers benefit from rapid blood clearance and favorable lesion penetration but often show renal excretion patterns that complicate abdominal imaging. Antibody based probes provide high target specificity and retention but require long circulation times that delay optimal imaging windows and increase radiation dose. Engineered formats such as nanobodies and affibodies attempt to combine rapid kinetics with high affinity and have shown promising preclinical and early clinical results for several targets, although manufacturing scale up and stability optimization remain active research areas. , In addition, ensuring in vivo stability of the ligand-chelator-radionuclide complex is a persistent technical requirement to avoid nonspecific background and to preserve quantitative interpretability.

To accelerate translation, the field is adopting platform strategies that standardize chelation chemistry, automated synthesis modules, and centralized quality assurance templates. Collaborative networks that link radiochemistry expertise, clinical imaging sites, and regulatory affairs can reduce the duplication of effort and facilitate multicenter evaluation. Nonetheless, overcoming manufacturing, regulatory, and radiochemistry constraints remains one of the most time-consuming and resource intensive components of bringing new targeted PET probes into clinical practice. Recent reviews of translational pathways and translational bottlenecks underscore the need for multidisciplinary project teams and early planning for cGMP scale up. ,

5.3. Barriers to Clinical Adoption

In lung cancer, barriers to the clinical adoption of targeted PET imaging are closely linked to the disease’s established biomarker-driven management framework and vary according to the biological role and clinical positioning of individual targets. Although tracers targeting PD-L1, EGFR, FAP, and DLL3 have demonstrated feasibility and biological relevance, their integration into routine workflows remains limited by infrastructure, regulatory, and evidence-generation constraints.

For immune checkpoint imaging, PD-L1 targeted PET exemplifies both the promise and the adoption challenges of molecular imaging in lung cancer. Early clinical studies have shown that immuno-PET can capture spatial heterogeneity of PD-L1 expression and may correlate with response to immune checkpoint inhibitors, particularly in NSCLC. However, PD-L1 imaging is positioned as a complement rather than a substitute for immunohistochemistry or circulating biomarkers, which complicates its value proposition to regulators and payers. The lack of standardized uptake thresholds and the absence of prospective evidence demonstrating that PD-L1 PET-guided decisions lead to improved outcomes compared with existing patient selection strategies have restricted its use to exploratory or early phase clinical studies. EGFR-targeted PET encounters distinct adoption challenges. Although tracers designed to image EGFR expression or mutant EGFR signaling target a clinically actionable pathway in NSCLC, their added clinical value relative to established genomic testing remains uncertain. While EGFR PET has been proposed for assessing spatial heterogeneity, detecting resistant clones, or monitoring pharmacodynamic responses to tyrosine kinase inhibitors, these potential applications must demonstrate tangible impacts on clinical management beyond what can be achieved with tissue or liquid biopsy-based genotyping. In the absence of clear evidence that imaging-derived data influence therapeutic decisions or timing, widespread clinical adoption remains limited.

FAP-targeted PET has garnered substantial attention owing to its high tumor-to-background contrast and widespread expression across lung cancer subtypes. Despite rapid clinical adoption in certain regions (particularly for staging and radiotherapy planning), its regulatory approval status and integration into clinical pathways remain inconsistent across healthcare systems. The absence of harmonized indications and established reimbursement mechanisms, coupled with limited prospective evidence linking FAP PET findings to improved patient outcomes or treatment modifications, continues to hinder its formal inclusion in lung cancer management guidelines. , In SCLC, DLL3-targeted PET offers a highly specific imaging strategy aligned with a lineage-specific marker of therapeutic relevance. Early clinical studies have demonstrated feasibility and target specificity, supporting its potential for patient selection and theranostic applications. However, the relatively low prevalence of eligible patient populations and the reliance on concomitant development of targeted therapies limit commercial viability and slow regulatory advancement, underscoring the broader challenge of translating niche imaging tracers into routine clinical practice.

Across these targets, common adoption barriers emerge. Targeted PET tracers must compete with well-established diagnostic standards (such as histopathology and molecular assays) within increasingly complex lung cancer care pathways. Regulatory frameworks and reimbursement models remain inadequately adapted to diagnostic agents that offer incremental or complementary value rather than enable binary clinical decisions. Moreover, the lack of consensus protocols and validated clinical end points continues to undermine clinician confidence and deter institutional investment. Despite these challenges, targeted PET provides a unique capability to noninvasively assess whole-body molecular heterogeneity in lung cancer. Broader clinical adoption will require target-specific evidence demonstrating tangible impacts on patient management, alignment between regulatory expectations and the risk profiles of diagnostic agents, and consensus guidance defining the specific contexts in which each tracer adds value within established lung cancer workflows. This alignment is essential to ensure that advances in targeted PET translate into durable clinical utility rather than remaining limited to proof-of-concept studies (Figure ).

4.

Key problems and corresponding solutions for the clinical translation of targeted PET imaging in lung cancer. This figure summarizes the major barriers limiting the routine clinical implementation of targeted molecular PET and the strategies proposed to address them using a problem-solution framework. Upper panels depict critical challenges, including spatial and temporal heterogeneity of target expression that undermines single-tracer, single-time point interpretation, technical and regulatory complexity in radiotracer development and cGMP translation, and barriers to clinical adoption arising from competition with established tissue and liquid biopsy-based biomarkers and the lack of standardized clinical evidence. Lower panels illustrate matched mitigation strategies, encompassing integrated whole-body and multitime point PET assessment combined with radiomics and circulating biomarkers to capture global tumor biology, platform-based and standardized approaches to probe synthesis and regulatory translation, and target-specific clinical positioning supported by consensus protocols and management-impact end points. Together, these aligned solutions outline a practical roadmap to enhance the robustness, interpretability, and clinical utility of targeted PET imaging in lung cancer.

6. Future Directions: From Precision Diagnosis to Precision Therapy

Targeted molecular PET imaging has shown considerable promise in the diagnosis of lung cancer and monitoring the treatment response. Despite these advances, realizing true precision medicine necessitates moving beyond diagnostic applications toward predictive models of therapeutic efficacy, individualized treatment strategies, and integrated theranostic frameworks. This section outlines future development pathways across three core dimensions: the discovery of novel targets and advanced imaging platforms, expanded use of PET for treatment prediction and personalization, and the advancement of theranostic approaches that seamlessly integrate diagnosis with therapy.

6.1. Development of Novel Targets and Probe Platforms

Traditional single-target PET probes are limited by heterogeneous expression and insufficient specificity across tumor types and microenvironments. Consequently, innovation in carrier design and target selection constitutes a critical direction for future PET probe development.

Nanoparticle-based probes leverage high surface-to-volume ratios and prolonged circulation times to enhance targeted accumulation and improve pharmacokinetic profiles, potentially increasing lesion detectability and enabling more efficient delivery of therapeutic payloads. Bispecific antibodies engineered to simultaneously engage two distinct tumor antigens can improve tumor binding specificity and prolong retention, thereby enhancing imaging signal-to-noise ratios. High-throughput peptide library screening technologies, such as phage display and cyclic peptide libraries, , continue to facilitate rapid identification of short peptide PET probes with high affinity and favorable immunogenicity profiles.

The dual-target heterodimeric imaging probe ⁶⁸Ga-FAPI-RGD exemplifies this strategy. By concurrently recognizing fibroblast activation protein (FAP) and integrin α_vβ₃, both of which are highly expressed in tumor stroma and angiogenic regions, this probe has shown markedly improved primary tumor detection and mediastinal lymph node assessment compared to conventional metabolic imaging (¹⁸F-FDG) in lung cancer (Figure A,B). Moreover, ⁶⁸Ga-FAPI-RGD also demonstrates markedly better performance than ⁶⁸Ga-FAPI-46 alone in the diagnosis of lung cancer (Figure C). Such combinations illustrate how multitarget engagement can overcome limitations of single-target tracers and provide richer biological insight.

5.

Uptake patterns of ⁶⁸Ga-FAPI-RGD, ¹⁸F-FDG, and ⁶⁸Ga-FAPI-46 in lung cancer lesions. (A) Maximum intensity projection PET/CT images obtained with ¹⁸F-FDG and ⁶⁸Ga-FAPI-RGD in patients diagnosed with lung cancer. Ca: cancer; NPC: nasopharyngeal carcinoma. (B) Representative ⁶⁸Ga-FAPI-RGD and ¹⁸F-FDG PET/CT images from a patient with metastatic lung adenocarcinoma. Immunohistochemistry of the left rib metastasis (arrow) showed positive staining for FAP and integrin α_vβ₃. (C) PET/CT images using ⁶⁸Ga-FAPI-RGD and ⁶⁸Ga-FAPI-46, along with corresponding immunohistochemical results, in patients with metastatic small cell lung cancer and nasopharyngeal carcinoma. The biopsy location is marked by an arrow. Reproduced from ref . CC BY 4.0.

In addition to stromal and angiogenesis markers, several emerging molecular targets are entering PET research. CEACAM6 is overexpressed in many lung adenocarcinomas and correlates with aggressive behavior. Preclinical PET tracers targeting CEACAM6 have shown high target specificity and tumor-to-background ratios in xenograft models, and theranostic constructs combining diagnostic and therapeutic isotopes have demonstrated both imaging specificity and antitumor efficacy in early work. Although established in neuroendocrine tumor imaging and peptide receptor radionuclide therapy, somatostatin receptor subtype 2 (SSTR2) represents a proof-of-concept for PET-based theranostics that could be adapted for lung cancer subtypes with neuroendocrine features. Moreover, in the field of small cell lung cancer, emerging therapeutic targets such as B7H3, DLL3, and TROP2 have garnered significant attention with preclinical and clinical studies currently underway to develop related radiopharmaceutical agents.

Beyond biological targets, the choice of radionuclide influences the imaging window and probe utility. Conventional PET isotopes such as ¹⁸F and ⁶⁸Ga are limited by relatively short half-lives, which constrain imaging flexibility for larger molecules such as antibodies. Emerging radionuclides such as ⁶⁴Cu, with a half-life of approximately 12.7 h, provide flexibility for delayed imaging and support the development of both imaging and therapeutic applications from a single compound. In the therapeutic domain, terbium-161 (¹⁶¹Tb) represents a next-generation beta and conversion electron emitter with potentially greater efficacy than lutetium-177 (¹⁷⁷Lu), while alpha-emitters such as astatine-211 (²¹¹At) and actinium-225 (²²⁵Ac) are under active investigation for targeted radiotherapy due to their high linear energy transfer and potent cytotoxicity against micrometastatic disease. , These emerging radionuclides collectively enhance the capabilities of PET-based theranostics and enable the closer integration of diagnostic imaging with targeted therapy.

6.2. In-Depth Application of PET in Efficacy Prediction and Personalized Therapy

Precision diagnosis must be integrated with predictive modeling and individualized therapeutic strategies to establish PET as an integral component of clinical decision making. Immune checkpoint inhibitor (ICI) therapy responses are highly heterogeneous across patients, and the early identification of responders is critical for effective management. Targeted PET probes for PD-L1, CD8⁺ T cells, FAP, and other immune-related markers offer dynamic insights into both systemic and local immune environments, thereby enabling more accurate prediction of immunotherapy efficacy.

PD-L1 PET imaging enables whole-body assessment of PD-L1 expression, which may correlate with clinical responses to ICI. Dynamic PET visualization of CD8⁺ T cell infiltration can capture immune engagement and differentiate inflammatory pseudoprogression from true disease progression. Imaging of FAP expression provides indirect insights into stromal activity and the immunosuppressive tumor microenvironment, aiding in the refinement of immunotherapy strategies in specific clinical settings. These applications are supported by recent theranostics reviews that highlight the potential of immuno-PET to guide personalized treatment.

In radiotherapy, PET has evolved beyond conventional staging to enable biology-driven treatment planning. Defining a biological target volume based on metabolic activity from PET allows for adaptive radiation strategies that concentrate dose escalation on regions with the highest metabolic activity while sparing surrounding normal tissues. These approaches aim to improve local tumor control without increasing treatment-related toxicity.

The integration of PET imaging with ctDNA for minimal residual disease (MRD) detection represents another advancement in personalized cancer surveillance. PET/CT is first performed at baseline and post treatment to provide whole body spatial assessment of tumor burden and biological heterogeneity, while serial ctDNA sampling is used to detect minimal residual disease and molecular recurrence during follow up. For example, Sujit et al. demonstrated that combining PET/CT habitat imaging with longitudinal ctDNA analysis significantly improved recurrence prediction in NSCLC compared with either modality alone. Combining PET readouts of metabolic and molecular target expression with ctDNA kinetics enhances the sensitivity and specificity of MRD monitoring, supporting earlier clinical intervention.

6.3. Theranostics

Theranostics unites diagnostic imaging with therapeutic application by using structurally related agents labeled with diagnostic and therapeutic radionuclides, respectively. This concept has been clinically validated in prostate cancer with PSMA directed PET and radioligand therapy (such as ⁶⁸Ga-PSMA-11/¹⁷⁷Lu-PSMA-617), and in neuroendocrine tumors with somatostatin receptor targeting (such as ⁶⁸Ga/¹⁷⁷Lu-DOTATATE). Such successes provide a blueprint for translating theranostic strategies to lung cancer.

Potential avenues for lung cancer theranostics include CEACAM6-targeted approaches, where diagnostic PET probes and therapeutic agent (¹³¹I) are developed against the same target. In preclinical studies, the ⁸⁹Zr-labeled anti-CEACAM6 antibody (⁸⁹Zr-DFO-Tinurilimab) outperformed the standard clinical tracer ¹⁸F-FDG in detecting and tracking the progression of malignant pulmonary nodules (Figure A,B). SPECT/CT imaging further showed that tumor uptake of ¹³¹I-tinurilimab, whether administered at high or low doses, steadily increased in A549 xenografts from day 1 to day 10 postinjection, unlike the rapid clearance seen in animals receiving free ¹³¹I (Figure C). To assess treatment response, ¹⁸F-FDG PET was employed to evaluate the residual tumor metabolic activity. As illustrated in Figure D,E, ¹⁸F-FDG uptake was substantially lower in xenografts from mice treated with ¹³¹I-tinurilimab compared to those receiving PBS, unlabeled tinurilimab, or free ¹³¹I alone. Consistent with these findings, ex vivo tumor measurements confirmed that ¹³¹I-tinurilimab therapy led to a marked reduction in the A549 xenograft size (Figure F). FAP-targeted radiotracers, such as ⁶⁸Ga-FAPI for PET/CT imaging (which demonstrates superior performance compared to FDG in lung cancer staging and detection) and ¹⁷⁷Lu-FAPI for therapy, which has shown feasibility in advanced cases, play a pivotal role in theranostics by enabling integrated noninvasive diagnosis, assessment of the tumor microenvironment, and targeted radioligand therapy. Another example is GRPR targeting, where PET imaging with GRPR antagonists can guide subsequent therapeutic delivery using beta- or alpha-emitting radionuclides (e.g., ⁶⁸Ga-NeoBOMB1). Parallel developments in radioligand therapy, such as ¹⁷⁷Lu-NeoBOMB1 targeting GRPR in other malignancies, demonstrate the feasibility of matched diagnostic and therapeutic platforms that could be adapted for lung cancer subtypes expressing GRPR.

6.

Preclinical assessment of the CEACAM6-targeted theranostic agent ⁸⁹Zr/¹³¹I-Tinurilimab in lung adenocarcinoma models. (A) Overview of the experimental design for detecting early malignant lung nodules in urethane-induced lung adenocarcinoma, together with the ¹³¹I radiolabeling approach for Tinurilimab-based radioimmunotherapy. (B) Representative PET/MR images showing a comparison between ¹⁸F-FDG (acquired 50 min postinjection) and ⁸⁹Zr-Df-Tinurilimab (acquired 72 h postinjection) in lung nodules from urethane-treated mice versus untreated controls. (C) Serial SPECT images illustrating the biodistribution of ¹³¹I-Tinurilimab in A549 xenograft-bearing mice at various time points after administration. (D, E) Representative ¹⁸F-FDG PET images, along with quantitative analysis of tracer uptake, in A549 tumor-bearing mice following treatment with ¹³¹I-Tinurilimab. (F) Ex vivo measurement of tumor volumes in A549 xenografts after ¹³¹I-Tinurilimab therapy. Tin: Tinurilimab. Reproduced from ref . CC BY 4.0.

Realizing high-precision theranostics requires systematic patient selection using robust imaging biomarkers, the development of matched therapeutic agents that target the same molecular pathway, and comprehensive dynamic monitoring of treatment response through integrated imaging and clinical end points. Enabling technologies such as physiologically based pharmacokinetic modeling and AI-driven digital twin frameworks will further enhance the predictive accuracy and individualized response forecasting. In addition, long-axial field-of-view PET systems provide a technical foundation for dynamic whole-body imaging and personalized dosimetry, supporting advanced theranostic workflows.

Collectively, these future directions position targeted molecular PET not only as a diagnostic modality but also as a cornerstone of personalized and theranostic cancer care, with rapidly evolving tracer designs, multimodal integration, and therapy strategies poised to improve clinical outcomes.

7. Conclusion

Since the clinical adoption of ¹⁸F-FDG PET, molecular imaging in lung cancer has evolved from assessing glucose metabolism to enabling direct visualization of tumor biology. While FDG PET/CT remains indispensable for staging, diagnosis, and treatment response assessment, its limited specificity in inflammatory conditions or low-glycolytic tumors restricts its performance in certain clinical scenarios. These limitations have driven the development of targeted PET tracers (including biomarker-specific agents, immuno-PET probes, and peptide-based radiotracers) that offer enhanced biological specificity and broader clinical utility.

Targeted PET enables noninvasive interrogation of key molecular features, including oncogenic drivers such as EGFR and MET, receptor expression targets such as GRPR and CEACAM6, and components of the tumor microenvironment such as FAP. Among these, FAP-targeted tracers have consistently demonstrated high lesion detectability and favorable tumor-to-background contrast in lung cancer, particularly in stroma-rich or FDG-avid lesions. These characteristics support their integration into precision staging and their emerging role in radioligand-based therapeutic strategies. Immuno-PET has become increasingly important for monitoring and predicting responses to immune checkpoint inhibitors. Because immune marker expression is spatially heterogeneous and temporally dynamic, tissue-based assays often yield incomplete assessments. Whole-body immuno-PET allows for longitudinal evaluation of immune targets such as PD-L1, as well as visualization of effector cell populations including CD8⁺ T cells. These capabilities improve discrimination between true progression and pseudoprogression and enhance response prediction when integrated with conventional clinical metrics.

The integration of PET imaging with liquid biomarkers, including circulating tumor DNA and proteomic signatures, further enhances the tumor characterization. Combined PET and ctDNA analyses enable earlier detection of recurrence, improved assessment of minimal residual disease, and more sensitive identification of emerging resistance compared to either modality alone. Artificial intelligence is accelerating these advances by enabling high-dimensional image analysis, radiomics-based phenotypic extraction, and automated image interpretation. AI-driven models that integrate PET with clinical data have demonstrated strong predictive performance for immune microenvironment features and treatment outcomes, while also supporting standardized workflows and reduced interobserver variability.

Looking ahead, precision lung cancer care will increasingly depend on the integration of targeted PET probes, immune and microenvironment imaging, liquid biomarkers, and AI-driven analytics. Theranostic paradigms exemplified by FAP-directed imaging and therapy highlight the potential for closed-loop systems that link diagnosis, treatment selection, and response monitoring. Continued efforts in standardization, multicenter validation, and translational collaboration will be essential to fully realize the role of targeted PET as a central platform in precision lung cancer medicine.

C.C. drafted the manuscript and created the figures. D.P., X.W., Y.X., J.Y., and L.W. helped with figure modification. M.Y. revised the manuscript. All of the authors read and approved the final manuscript.

The authors declare no competing financial interest.