Molecular imaging of patient-derived tumor organoids with clinical radiotracers

Introduction:: Patient-derived tumor organoids (tumoroids) offer new opportunities for translational cancer research, but unlike animal models, their broader use is hindered by the lack of clinically relevant imaging endpoints. Several biological endpoints are used for organoid studies, for instance, viability assays or cell death assays, but these bear little relevance to clinical biomarkers and do not easily translate in vivo. For instance, one gold standard in cancer clinics for molecular assessment of normal and diseased tissues is positron emission tomography (PET). 18F-fluorodeoxyglucose (18F-FDG) is the most common PET biomarker, which maps glucose metabolism in tumors, drastically improving the diagnosis, staging, and subsequent treatment of cancers in the last two decades. Other radiotracers, such as I-124 are commonly used for thyroid scans. Many other targeted radiotracers are in clinical trials—such as bombesin receptors, prostate-specific membrane antigen (PSMA), somatostatin receptor, αvβ6 integrin, fibroblast activation protein inhibitors (FAPI), or T-cell activation (Ara-G). Therefore, the ability to screen patient-derived organoids using clinical PET-radiotracers could open the door to new applications, including co-clinical trials.

Materials and Methods:: In vitro, radiotracer uptake assays are conventionally performed by gamma-counting. This method is quantitative but cannot provide any spatial information on tumoroids. Here, we present two new strategies for directly imaging tumoroids using clinical radiotracers and compare these images with the PET scan of the patient’s tumor of origin. First, we developed a method called ‘positron-emission microscopy of organoids (oPEM)’ with spatial resolution 100-fold better than clinical positron emission tomography (PET) [1]. As a method, oPEM draws from our previous work developing radioluminescence imaging with a wide-field microscope equipped with a scintillator plate and a single photon-sensitive electron-multiplying charge-coupled device (EMCCD) [2]. In contrast to single-cell imaging, the high count rate from tumoroids creates enough scintillation signal for fast imaging without image reconstruction. We are currently developing an alternate strategy for performing high-throughput drug screening with clinical radioisotopes. In this approach, named organoPET, we use a new generation small-animal µPET scanner (SOFIE GNEXT) with a very high spatial resolution to image tumoroids arrayed in a stack of culture plates (up to n=5). As proof of concept, the uptake of FDG will be compared with standard viability assays after various treatments (drug+/radiation) of tumoroids. Finally, when available, these results will be compared with the original tumors in the patients’ PET/CT scans from the clinic.

Results, Conclusions, and Discussions:: We imaged 18F-FDG influx with oPEM to show that head-and-neck squamous cell carcinoma (HNSCC) tumoroids (n=13) recapitulate (Pearson’s r = 0.756, P = 0.0032) the glycolytic activity of the tumor of origin in patients (n=4), and thus, could be used to predict therapeutic response in vitro. Tumoroids derived from two metastatic papillary thyroid carcinoma nodules of a single patient retained the metabolic heterogeneity with differential (P = 0.002) 18F-FDG avidity. Similar oPEM imaging was performed with kidney, colorectal, and liver tumoroids. In addition, in a preliminary study, 18F-AraG was explored to image immune activation in tumoroids cultured at the air-liquid interface (ALI). Although oPEM can image tumoroids in very high resolution, one bottleneck of the current iteration of this method is lower throughput. Our preliminary study with the SOFIE GNEXT µPET scanner shows the ability to image individual HNSCC tumoroids from 24-well organoid plates, thus promising measurements of up to 120 organoid cultures in a single 5 min scan. Another pilot study with cell lines has shown the feasibility of FDG-PET to measure dose-dependent drug response in 96-well plates.

In conclusion, we present molecular imaging in live tumoroids and show that these models recapitulate specific metabolic features of the tumors of origin, making them a promising tool for personalized cancer medicine. Moreover, oPEM and organoPET can complement each other to provide unified image-derived biomarkers to integrate, in real-time, the data from the preclinical and clinical arms of the co-clinical trial. The development of small-animal PET scanners 15 years ago has demonstrated the utility of developing preclinical imaging tools that have strong relevance to the clinical workflow. Accordingly, oPEM and oPET could become the standard technology for imaging tumor and normal tissue organoids using PET tracers, thereby expanding the use of organoid models for translational research.

Acknowledgements (Optional): :

References (Optional): : [1] Khan et al, Nat commun. 12, 5883 (2021)

[2] Kim et al, Nat. Protoc. 12, 1055 (2017).