Click on an contribution preview the abstract content.
Click on an contribution preview the abstract content.
Introductory Talk by Go van Dam - Groningen, The Netherlands
This talk provides an overview of state-of-the-art research and refers to the following presentations selected from abstract submissions.
Standardization of intra-operative fluorescence imaging systems using a multi-parameter solid phantom (#402)
D. Gorpas1, 2, M. Koch1, 2, M. Anastasopoulou1, 2, U. Klemm1, V. Ntziachristos1, 2
1 Helmholtz Zentrum München, Institute of Biological and Medical Imaging (IBMI), Neuherberg, Bavaria, Germany
Fluorescence guided surgery is emerging into a significant tool for oncology. Limitations in human vision and tactile feedback require post-surgical pathology assessment . Fluorescence imaging has shown promising results in clinical studies. Nevertheless, it has not yet been integrated into the standard clinical practice. This, in part, is due to the lack of standardization methods that enable repeatability of studies and comparison of systems . In this work, we propose a novel composite phantom that allows characterization of intra-operative fluorescence imaging systems.
We used transparent polyurethane for the phantom matrix, while organic quantum dots, TiO2 nanoparticles, and nigrosin in the phantom matrix and Hemin in the wells simulated fluorescence, scattering and absorption respectively. The phantom wells assess sensitivity, fluorescence and optical resolution, light cross-talk, and enable flat-fielding (Fig. 1a-b). To test the phantom, we employed a hybrid imaging system (Camera I), consisting of an electron multiplying CCD (EMCCD) for fluorescence detection and a CCD camera for color imaging . A second system (Camera II), without the color channel, was also employed to demonstrate the phantom application for benchmarking (Fig. 1c-d). Following phantom imaging we segmented the wells  and quantified the corresponding SNR and contrast.
The results demonstrate the phantom application to characterize an imaging system, as well as to benchmark systems of markedly different specifications. Specifically for Camera I, besides the quantification of the assessment metrics, imaging of the phantom further enables registration between fluorescence and color data, and augmentation of color images with fluorescence information. Images are also corrected for possible vignetting due to illumination inhomogeneities (Fig. 1e-f). The comparison of the two systems showed that Camera I can achieve ~30% higher SNR and ~45% higher contrast than Camera II for all wells assessing sensitivity (Fig. 2). Nevertheless, Camera II can perform equivalently to Camera I when its working distance is reduced by ~40% and 2× binning is applied (Fig. 2). To identify which of the acquisition settings of Camera II result into equivalent to Camera I performance, we adopted a least squares method between all metrics quantified through the phantom.
Lack of fluorescence imaging standardization is one of the major factors delaying clinical translation of this exciting technology [1, 5]. Most current approaches quantify one performance parameter . The study herein introduced a multi-parametric phantom for the characterization of fluorescence imaging systems. This procedure can also guide the configuration of different systems for comparable performance. The latter is essential for multicenter clinical trials. We envision that composite phantoms will become important assets for clinical translation of fluorescence molecular imaging.
 M. Koch, and V. Ntziachristos, 2016, "Advancing Surgical Vision with Fluorescence Imaging", Annu Rev Med, 67(1): 153-164.
The authors would like to thank Dr. Pilar Beatriz Garcia-Allende for her contribution during the phantom preparation. The research leading to these results has received funding by the Deutsche Forschungsgemeinschaft (DFG), Sonderforschungsbereich-824 (SFB-824), subproject A1 and from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No. 644373 (PRISAR).
Fig. 1. Application of composite standardization phantom for fluorescence camera assessment.
(a) Phantom. (b) Phantom wells. Sensitivity vs red-optical properties and blue-depth; purple-resolution; pink-cross-talk; green-illumination; cyan-matrix. (c) Camera I. (d) Camera II. D: diffuser; F: filter; DM: dichroic mirror; RL: relay lens. (e) Intensity of scattering wells (top left); light profile (top right); raw image (bottom left); corrected image (bottom right). (f) Augmented image.
Fig. 2 Quantitative comparisons between Camera I and Camera II.
(a) SNR, and (b) contrast from the nine wells with different optical properties. (c) SNR, and (d) contrast from the nine wells with different depths from the phantom’s top surface. In all panels, x -axis labeling corresponds to the labeling of phantom elements shown in Fig. 1(b).
Keywords: Intra-operative fluorescence imaging, standardization, composite phantom, benchmarking, calibration
[18F]fluoroethyltyrosine-induced Cerenkov luminescence improves image-guided surgical resection of glioma (#93)
D. Lewis1, 2, R. Mair1, A. Wright1, K. Allinson4, S. Lyons1, T. Booth1, J. Jones1, R. Bielik1, D. Soloviev1, 2, K. M. Brindle1, 3
1 University of Cambridge, Cancer Research UK - Cambridge Institute, Cambridge, United Kingdom
The extent of surgical resection is significantly correlated with outcome in glioma 1, however current intraoperative navigational tools such as fluorescence imaging with 5‑aminolevulinic acid (5-ALA) are effective only in a subset of patients 2,3. We show here that a new optical intraoperative technique, Cerenkov luminescence imaging (CLI) following intravenous injection of O‑(2-[18F]fluoroethyl)-L-tyrosine (FET), can be used to accurately delineate glioma margins, performing better than the current standard of fluorescence imaging with 5-aminolevulinic acid (5-ALA).
Rats implanted orthotopically with U87, F98 and C6 glioblastoma cells were injected with FET and 5-aminolevulinic acid (5-ALA). Positive and negative tumor regions on histopathology were compared with CL and fluorescence images. The capability of FET CLI and 5-ALA fluorescence imaging to detect tumor was assessed using receptor operator characteristic curves and optimal thresholds (CLIOptROC and 5-ALAOptROC) that separated tumor from healthy brain tissue were determined. These thresholds were used to guide prospective tumor resections, where the presence of tumor cells in the resected material and in the remaining brain were assessed by Ki-67 staining.
FET CLI signal was correlated (y = 1.06x – 0.01; p < 0.0001; Fig 1a) with signal in preoperative PET/MR images (Fig 1b,c) and with expression of the amino acid transporter SLC7A5 (LAT1). FET CLI (AUC = 97%) discriminated between glioblastoma and normal brain in human and rat orthografts more accurately than 5-ALA fluorescence (AUC = 91%), with a sensitivity >92% and specificity >91% (Fig 1d-i), and resulted in a more complete tumor resection (Fig 2).
FET CLI can be used to accurately delineate glioblastoma tumor margins intraoperatively, performing better than the current standard of fluorescence imaging following 5-ALA administration, and is therefore a promising technique for clinical translation.
1. Hervey-Jumper, S.L. & Berger, M.S. Maximizing safe resection of low- and high-grade glioma. Journal of neuro-oncology 130, 269-282 (2016).
2. Lau, D., et al. A prospective Phase II clinical trial of 5-aminolevulinic acid to assess the correlation of intraoperative fluorescence intensity and degree of histologic cellularity during resection of high-grade gliomas. Journal of Neurosurgery 124, 1300-1309 (2016).
3. Jaber, M., et al. The Value of 5-Aminolevulinic Acid in Low-grade Gliomas and High-grade Gliomas Lacking Glioblastoma Imaging Features: An Analysis Based on Fluorescence, Magnetic Resonance Imaging, 18F-Fluoroethyl Tyrosine Positron Emission Tomography, and Tumor Molecular Factors. Neurosurgery 78, 401-411 (2016).
CLI with [18F]FET more accurately delineates GBM from brain tissue than 5-ALA imaging. (a) FET CLI in tumor slices correlates with PET signal. (b) CE T1-w MRI of a U87 GBM and (c) FET PET/ T2-w MRI. H&E of a U87 GBM (d), with FET CLI (e), 5-ALA (g) and segmentation. (h) ROC curves for FET CLI (blue) and 5-ALA (red) (n≥5). (i) AUC of the ROC, sensitivity and specificity with FET CLI and 5-ALA (n≥5)
Image-guided surgical resection is improved using CLI of [18F]-FET (a) compared to fluorescence imaging with 5-ALA (b). Images acquired before and after two surgical resections of the tumor. (c) Ki-67 staining of the resected tumor specimens (#) and (##), showing positive staining for tumor cells and negative Ki-67 in the remaining brain (###).
Keywords: Cerenkov, glioblastoma, intraoperative, FET, CLI
PET/CT and near-infrared fluorescence imaging of colorectal cancer using a single injection of a dual-labeled cRGD-based tracer (#547)
B. Sibinga Mulder1, H. Handgraaf1, D. Vugts2, C. Sewing2, M. Stammes3, L. de Geus-Oei4, A. Windhorst2, M. Bordo5, J. Mieog1, C. van de Velde1, J. Frangioni5, A. Vahrmeijer1
1 LUMC, Surgery, Leiden, Netherlands
Hybrid tracers, allowing both positron emission tomography (PET) and near infrared fluorescence (NIRF) imaging, can aid in accurate preoperative surgical planning and intraoperative real-time NIRF detection of the tumor. cRGD peptide targets integrins associated with angiogenesis (e.g. αvβ3) and has been used successfully for both PET and NIRF imaging. This study evaluates the hybrid tracer ZW800F-cRGD-[89Zr]Zr-DFO for PET and NIR fluorescence imaging in human colorectal xenografts.
An in vitro binding assay was performed and subsequently 10 nmol ZW800F-cRGD-Zr-DFO was injected in mice (n=7) bearing orthotopic colorectal tumors (HT29-luc2). NIRF imaging for detection of the tumor specific signals was performed at 4 and 24 h, the biodistribution was determined. For assessment of the stability of the signal, NIRF imaging was performed up to 168 h in mice bearing subcutaneous tumors. Subsequently, ZW800F-cRGD-Zr-DFO was synthesized and labeled with 89Zr. Finally, 10 nmol ZW800F-cRGD-[89Zr]Zr-DFO (3 MBq) was injected to mice (n=8) bearing orthotopic colorectal tumors (HT29-luc2). PET/CT was performed at 1, 4 and 24 h post injection. Biodistribution was performed at 4 and 24 h post injection.
The in vitro binding assay demonstrated an almost linear increase in fluorescence intensity with increasing concentrations. Sufficient fluorescent signals were measured in the tumors of the mice injected with ZW800F-cRGD-Zr-DFO (emission peak ~800nm). The fluorescence signal of ZW800F-cRGD-[89Zr]Zr-DFO remained stable after labelling with zirconium-89 and PET/CT at 4 h allowed clear visualization of the colorectal tumors. Biodistribution at 4 h showed the highest uptake of the tracer in kidneys and sufficient uptake in the tumor. At 24h the uptake in the tumor was still sufficient.
This study shows the feasibility of PET and fluorescence imaging of tumors with a single injection of ZW800F-cRGD-[89Zr]Zr-DFO, allowing preoperative surgical planning followed by intra-operative imaging.
Keywords: Near-infrared fluorescence imaging, PET imaging, cRGD peptide, Hybrid tracers
Development and characterization of a multi-modality anti-PSMA targeting agent for imaging, surgical guidance, and targeted photodynamic therapy of PSMA-expressing tumors (#75)
S. Lütje1, 2, S. Heskamp1, G. M. Franssen1, C. Frielink1, A. Kip1, G. Fracasso3, K. Herrmann2, M. Gotthardt1, M. Rijpkema1
1 Radboudumc, Radiology and Nuclear Medicine, Nijmegen, Netherlands
Prostate cancer (PCa) recurrences after surgery frequently occur. To improve management of PCa, the multi-modality anti-PSMA targeting agent 111In-DTPA-D2B-IRDye700DX was developed and characterized in vivo. This agent can be used for both pre- and intra-operative tumor localization, image-guided surgery, and eradication of (residual) tumor tissue by PSMA-targeted photodynamic therapy (tPDT), which is a highly selective cancer treatment based on targeting molecules conjugated to photosensitizers that can induce cell destruction upon exposure to near-infrared (NIR) light.
The anti-PSMA monoclonal antibody D2B was conjugated with IRDye700DX and DTPA and subsequently radiolabeled with 111In. To determine the optimal time point for tPDT, BALB/c nude mice with PSMA+ s.c. LS174T-PSMA xenografts received 30 µg of the conjugate intravenously (8 MBq/mouse) followed by µSPECT/CT, near-infrared fluorescence imaging, and biodistribution at 24, 48, 72, and 168 h p.i.. Tumor lesions were resected under image-guidance using intraoperative NIR fluorescence imaging. Tumor growth of LS174T-PSMA xenografts and overall survival of mice treated with 80 µg of the conjugate followed by 1-3 times of NIR light irradiation (100 or 150 J/cm2) at 24 h p.i. was compared to control mice that received either NIR light or PBS alone.
Highest specific tumor uptake was observed at conjugate doses of 80 µg/mouse. Biodistribution revealed no significant difference in tumor uptake in mice at 24, 48, 72, and 168 h p.i.. PSMA+ tumors were clearly visualized with both µSPECT/CT and NIR fluorescence imaging. Median survival of mice treated with 80 µg of DTPA-D2B-IRDye700DX and 3x150 J/cm2, 1x150 J/cm2, 3x100 J/cm2 of NIR light at 24 h p.i. was significantly improved compared to both control groups (70, 61, and 36 days vs. 16 days for both controls, respectively, p=0.0009). Treatment with 3x150J/cm2 resulted in significantly prolonged survival compared to treatment with 3x100J/cm2 (p=0.0224). No significant difference was observed between mice that receive 3x150J/cm2 or 1x150J/cm2 (p=0.6714).
Proof-of-principle was provided that 111In-DTPA-D2B-IRDye700DX can be used for pre- and intra-operative detection of PSMA+ tumors with radionuclide and fluorescence imaging, surgical guidance, and PSMA-targeted PDT. The optimal conjugate dose for PSMA-tPDT in this setting is 80 µg/mouse, the optimal time point for NIR light irradiation is 24 h p.i.. In vivo, PSMA-tPDT resulted in significant prolongation of median survival.
Imaging and targeted photodynamic therapy of PSMA-expressing tumors
1) Near-infrared fluorescence image (left) and microSPECT/CT image of a mouse with a s.c. PSMA+ LS174T-PSMA tumor performed at 24h after injection of 111In-DTPA-D2B-IRDye700DX (80μg/mouse, 8MBq). 2) Kaplan-Meier survival curve of mice treated with 80μg of DTPA-D2B-IRDye700DX and different NIR light irradiation intensities.
Keywords: Prostate cancer, PSMA, targeted photodynamic therapy, image-guided surgery
Image-Guided Pathology for Evaluation of Resection Margins in Locally Advanced Rectal Cancer using the Near-Infrared Fluorescent Tracer Bevacizumab-800CW (#496)
S. J. de Jongh1, J. J. J. Tjalma1, M. Koller2, M. D. Linssen3, A. Jorritsma-Smit3, A. Karrenbeld4, K. Havenga2, P. H. J. Hemmer2, E. G. E. de Vries5, G. A. P. Hospers5, B. van Etten2, V. Ntziachristos6, G. M. van Dam2, W. B. Nagengast1
1 University Medical Center Groningen, Gastroenterology and Hepatology, Groningen, Netherlands
Negative circumferential resection margins (CRM) are the cornerstone for curative treatment of patients with locally advanced rectal cancer (LARC). Unfortunately, perioperative techniques for evaluation of resection margins are lacking, whereas standard histopathological examination is time-consuming. In this study, we evaluated the feasibility of optical molecular imaging as a tool for evaluation of resection margins at the surgical theater, i.e. Image-Guided Pathology (IGP), to improve clinical decision making.
Fluorescence imaging data of fresh surgical specimens and subsequent bread-loaf slices from patients with LARC (NCT01972373) were analyzed as a side study. All patients were administered intravenously with 4.5 mg of the fluorescent tracer bevacizumab-800CW 2-3 days prior to surgery. Seven patients met the inclusion criteria for correlation of fluorescence intensities in fresh surgical specimens with histology, to evaluate resection margins. For analysis of bevacizumab-800CW localization in bread-loaf slices, sufficient data was available from 17 patients. A receiver operating characteristics (ROC) curve was plotted to determine the mean fluorescence intensity (MFI) cut-off value for tumor detection.
Using IGP, in one patient a histologically confirmed tumor-positive CRM was predicted correctly at the surgical theater (Figure 1). Tumor-negative CRMs were predicted correctly in four patients using IGP. One tumor-positive CRM could not be detected; however, this positive margin was based on the presence of only an isolated microscopic tumor deposit in the CRM. One close CRM (1.4 mm) was identified as tumor-positive. Optical imaging enabled a clear differentiation between tumor and surrounding tissue in the bread-loaf slices (n=42) of all 17 patients ex vivo. In our limited sample size, an optimal MFI cut-off value of 5085 was determined based on the ROC curve, with a sensitivity and specificity of 98.2% and 76.8% respectively.
We demonstrate for the first time the potential of IGP for identification of positive resection margins directly after surgery in patients with LARC. Clearly, this might change current peri-operative decision making with regard to additional targeted resections or intraoperative brachytherapy. Based on the initial results from this study, a standardized methodology was developed to confirm these findings in a subsequent larger IGP study.
Keywords: Image-Guided Pathology, Locally Advanced Rectal Cancer, Bevacizumab-800CW, Near-Infrared Fluorescence
Tumor-Specific Uptake of the Near-Infrared Fluorescent Anti-EGFR Antibody Panitumumab-IRDye800 in Patients with Head and Neck Squamous Cell Carcinoma: Safety and Feasibility Results from our Phase I Study (#310)
N. S. van den Berg1, N. Teraphongphom1, R. W. Gao1, S. Hong1, B. Martin2, V. Divi1, M. J. Kaplan1, R. Ertsey1, N. J. Oberhelman4, G. Lu1, C. S. Kong2, A. D. Colevas3, E. L. Rosenthal1
1 Stanford University, Otolaryngology - Head and Neck Surgery, Stanford, California, United States of America
Surgery remains the cornerstone of cancer treatment. Yet, radical tumor resection cannot always be achieved, and often healthy tissue has to be resected as well, leading to high morbidity. Real-time intraoperative fluorescence imaging may provide a solution. In the current study panitumumab, an anti-epidermal growth factor receptor (EGFR) antibody, was coupled to the near-infrared (NIR) dye IRDye800. Besides safety evaluation, we evaluated panitumumab-IRdye800 for intraoperative real-time fluorescence-based navigation to primary and metastatic head-and-neck squamous cell carcinoma (HNSCC).
We conducted an open-label, dose escalation clinical trial of panitumumab-IRdye800 in 21 HNSCC patients. Cohort 1 (n=3) received a microdose (0.06mg/kg). Cohort 2A (n=5), 2B (n=7) and 2C (n=6) received 0.5mg/kg, 1.0mg/kg or 50mg (flat dose) of panitumumab-IRDye800, respectively. Ccohort 2A and 2B patients also received 100mg unlabeled panitumumab. Patients were followed for 30 days after infusion, and adverse events were recorded. Surgery will be performed 2-5 days post-panitumumab-IRDye800 infusion. Prior ot standard of care surgery, real-time intraoperative NIR fluorescence imaging was used to evaluate tumor and lymph node (LN) visibility. Specimen fluorescence intensities were correlated with histology and immunohistochemistry for amongst others EGFR.
No dose-limiting toxicities occurred. One grade 1 adverse event was observed. Intraoperative imaging showed clear demarcation between cancerous and normal tissue, and ex vivo imaging findings correlated well with intraoperative findings. Cohort 1 was not analyzed, as it was primarily to establish safety. The average tumor-to-background ratio (TBR) was 5.40±0.68 for cohort 2A, 5.44±0.70 for cohort 2B and 6.53±1.25 for cohort 2C. Correlation of fluorescence intensities with tumor location resulted in a sensitivity and specificity of 92.03% and 78.07% for cohort 2A, and 91.86% and 91.17% for cohort 2B and 97.44% and 87.97% for cohort 2C. Positive predictive value were 68.32%, 85.51% and 87.85%, and negative predictive value found were 95.02%, 95.63%, and 97.96% for cohort 2A, 2B and 2C, respectively. EGFR expression positively correlated with fluorescence intensity.
Results from this first-in-human trial using panitumumab-IRDye800 suggest that it is a safe, and highly specific and sensitive optical imaging agent to aid in real-time intraoperative detection and surgical resection of both primary and metastatic HNSCC.
Schematic overview of the clinical trial
Eligible patients with HNSCC will receive a systemic infusion with panitumumab-IRdye800. 2-5 days later, the patient will undergo surgery whereby intraoperative NIR fluorescence imaging is performed to look at tumor and lymph node fluorescence. Following imaging, standard-of-care surgery will be performed. Fluorescence imaging findings will be correlated to histology and immunohistochemistry.
Keywords: fluorescence, head and neck cancer, lymph node, antibody, near-infrared
An instantaneous staining approach for tumor delineation in freshly excised biospecimens (#323)
S. Kossatz1, A. Strome1, W. Weber1, 3, S. Patel2, T. Reiner1, 3
1 Memorial Sloan Kettering Cancer Center, Radiology, New York, New York, United States of America
Molecularly specific delineation of tumors in the diagnostic and intraoperative setting remains an unmet clinical need with the potential to identify malignant growths faster and with higher accuracy than standard practice. Here, we present PARPi-FL, a fluorescent inhibitor of the DNA repair enzyme PARP1, as ultrafast topical contrast agent for whole tissues and sections to differentiate tumor vs. non-tumor tissues. We tested its ability to instantaneously stain tumor cells in xenografts of oral squamous cell carcinoma (OSCC), esophageal adenocarcinoma (EAC), and clinical biospecimens of OSCC.
PARPi-FL is a cell permeable, fluorescently labelled PARP inhibitor (excitation/emission max.: 503 nm/515 nm). We optimized the topical staining protocol on cryosections and fresh tissue samples of FaDu (OSCC) and OE19 (EAC) xenografts towards nuclear staining with low cytoplasmic background using confocal microscopy. Different staining times (1-10 min), concentrations (50-1000 nM) and washing protocols were tested. The topical staining outcome was validated against systemic PARPi-FL injection. We translated the optimized staining to human biospecimens of OSCC and evaluated the ability to outline tumor margins via microscopic and macroscopic imaging. We confirmed presence of tumor foci in tissues by H&E histopathology and PARP1 expression of cells by immunohistochemistry (IHC).
Optimized topical PARPi-FL staining in xenografts resulted in staining of tumor cell nuclei, but not the cytoplasm or stromal cells, allowing clear identification of tumor areas in cryosections and fresh tissues. Total staining time was 5 min for cryosections and 15 min for fresh tissue using 100 nM PARPi-FL. Higher concentrations increased cytoplasmic uptake. Topical staining quality and intratumoral distribution were comparable to intravenous PARPi-FL injection in FaDu and OE19 xenografts. The staining protocol was effectively translated to human biospecimens of OSCC without loss of quality (n=5 completed, ongoing). We identified tumor foci in both fresh tissue samples and cryosections, based on PARPi-FL accumulation in tumor cell nuclei, while uptake in normal epithelium and mucosa was low (Figure 1). According to the histopathological evaluation, 100% of SCC regions in the studied cohort were visualized by PARPi-FL staining and showed a positive IHC staining of PARP1.
We show that PARPi-FL topical staining is suitable for instantaneous tumor cell identification in live patient biospecimens. Widespread expression of PARP1 in tumors was confirmed in our patient cohort. This technology can be used in a diagnostic setting, as well as on excised tissue during surgery to identify positive margins. Importantly, the quick turnaround time for results allows for an immediate feedback to the physician, reducing diagnostic delays and repeated patient visits.
Rapid PARPi-FL staining of fresh tissue or cryosections in a patient sample of OSCC allowed for identification of tumor foci due to accumulation in tumor cell nuclei, but not surrounding normal cells. The biopsy was split in three parts to carry out staining of fresh tissue, cryosections and histopathology. Images were acquired on a confocal microscope. Green: PARPi-FL, blue: Hoechst DNA stain.
Keywords: optical imaging, PARP1, fluorescence imaging, topical staining, early detection, intraoperative imaging