BOB Ped Surg 2023 - Bernadette Jeremiasse, IPSO - Presentation

Hi, I'm Bernadette, and I'm going to present you our research which is called Multiplex organoid-based 3D Life Imaging Platform to screen probes for fluorescence guided surgery. Achieving complete tumor resections without major complications remains challenging. Some tumors, it's difficult to distinguish tumorous tissue from healthy, and in other tumors, the margin is very narrow. Fluorescence guided surgery with tumor-specific probes can visually assist the surgeon in discriminating tumors from healthy tissue. However, tumor-specific probes are currently screened one by one and per tumor type. Here we present an imaging platform that is able to screen fluorescent guided surgery probes in a tumor-specific and even patient-specific way. For this platform, we combine tumor organoid biobanks with 3D live imaging. The main advantages of organoids is that they maintain molecular tumor characteristics and can be held in culture for a long time. Which shows for both an adult and pediatric tumor organoid biobank, namely neuroblastoma and breast cancer. Control organoids were chosen based on the tissue surrounding the tumor. So for neuroblastoma, kidney organoids, and for breast cancer, healthy breast organoids. For the sake of time, I will focus on neuroblastoma in the rest of the presentation. Surface markers that are up regulated in neuroblastoma or breast cancer were identified based on RNA profiling, the human protein atlas, and literature review. Probes targeting these surface markers were conjugate to six different fluorophores ranging between 488 and 647 nanometers. Together with the general marker E fluor, they were used for seven colored 3D live imaging using confocal microscopy on living organoids. We use segmentation analysis by uh the staple pipeline as developed in our lab, and the fluorescent signals were quantified. This high throughput in vitro screening was validated by testing the three most promising probes for neuroblastoma in vivo using a mouse enograph model with tumors originating from neuroblastoma organoids. This all with the ultimate end goal to generate a multiplex organoid-based 3D live imaging platform to screen these probes and already validate some of the most promising probes for clinical trials. Here you can see the different neuroblastoma organoid lines that we screened and also the healthy kidney on the left top. It is very clear that the probes do work, but their fluorescence varies between patients and even within a patient. Also, the colors are very different from the healthy kidney, meaning that other probes work on tumor organoids with respect to the healthy control. After microscopy, we quantified the fluorescence using the staples pipeline, which is a 3D segmentation pipeline developed in our lab. We first segmented the memranal and nuclear signals, generated cells, and these cells together formed single organoids. After stapled segmentation, we determined the fluorescent intensity of the tumor organoids and of the healthy control organoids. We determined which organoid was positive based on the following. The intensity of the tumor organoid must be at least twice the fluorescence of the healthy control organoid, and in this way we quantified thousands of organoids. Quantification shows us the percentage of positive organoids per organoid line and per marker. So for neuroblastoma, you can see here the control line, and it is very clear that these 4 markers are promising for neuroblastoma because they have high percentages and high fluorescent intensities. Because we screen 6 probes simultaneously, we can also predict which combinations of probes are most promising, and here you can see a heat map with on the right side neuroblastoma organoid lines, and here all probes and probe combinations, and it is very clear which group of probes and probe combinations works works best or is most promising, this group. Our high throughput in vitro screen was validated by testing three most promising probes in viO GD2, L1 chem, and NCchem. These probes resulted in TBRs, tumor to background ratios above 2, indicating good intraoperative visibility, and therefore, these probes could be tested for clinical purposes. Furthermore, we even found that in vitro imaging patterns resemble the efficacy in FIFO. For example, L1 chem is higher than TD2 in vitro, and this is also in FIFO, and also the order of the lines was similar, so neuroblastoma, line 129 being highest and 67 being lowest. We are now very happy to announce that one of the probes that we tested is going into a clinical trial this year. And with this, I would like to thank you and all my colleagues, of course, for their hard work.