Click on an contribution to preview the abstract content.
Click on an contribution to preview the abstract content.
|1:00 p.m.||PL 09-11a-01||
Image guided surgery for generic tumor detection in solid tumors using the PH activated micelle-based imaging agent ONM-100 (#96)
Floris J Voskuil1, Pieter J Steinkamp2, Marjory Koller2, Bert van der Vegt3, Jan J Doff3, Tian Zhao4, Jeffrey Hartung6, Yalia Jayalakshmi4, Baran D Sumer5, Jinming Gao7, Max J H Witjes1, Gooitzen M van Dam2
1 University Medical Center Groningen, Oral & Maxillofacial Surgery, Groningen, Netherlands
ONM-100, an ICG labeled micelle-based polymer imaging agent with an exquisitely pH-sensitive binary on/off mechanism, can be used for intra-operative tumor detection. Micelles dissociate in acidic environments resulting in fluorescent activation of ICG. As most solid cancer types are acidotic due to anaerobic glycolysis, the so-called Warburg effect, ONM-100 acts as a tumor agnostic imaging agent targeting a broad range of tumors. This first in-human study investigates the safety and feasibility of ONM-100 as an imaging agent for intra-operative fluorescent imaging of various solid tumors.
In this ongoing clinical Phase I study, the pH-activated near-infrared fluorescent imaging agent ONM-100 was IV administered 24±8h prior to surgery in a dose escalation scheme (0.1-1.2mg/kg), and the optimal dose was verified in a subsequent cohort. Patients with histopathologically proven breast cancer (BC), head and neck squamous cell carcinoma (HNSCC), colorectal cancer (CRC) and esophageal cancer (EC) were included. Blood was drawn to assess safety data and pharmacokinetics. Intra-operative images were collected before and after tumor excision and from the wound bed. After excision, fluorescence images were obtained from serially sliced specimens and formalin fixated paraffin embedded tissue blocks and correlated with standard histopathological assessment.
Currently, 28 patients (10 BC, 13 HNSCC, 3 EC, 2 CRC) were enrolled between March and November 2018. No tracer related (serious) adverse events were observed. A strong and sharply demarcated fluorescent signal was observed in all 27 patients with vital tumor tissue with in- and ex vivo imaging (median Contrast to Noise Ratio 3.5; IQR 4.0) which correlated with tumor on final histopathology. Absence of fluorescence in 1 EC patient was confirmed to be a complete pathological response after neo-adjuvant treatment. Pharmacokinetics showed increased tumor fluorescence in a dose and plasma concentration-dependent manner. HNSCC and superficially located BC could be clearly visualized in vivo during surgery. In 3 patients (BC and HNSCC), perioperative and otherwise unnoticed tumor by the surgeon was detected on the margin or wound bed using fluorescence imaging. Additionally, 2 BC tumor lesions were detected which were missed by conventional pre-operative imaging and pathological assessment.
This study shows that ONM-100 appears to be safe and allows fluorescent tumor visualization both in- and ex vivo. Here, we provide the first in-human data that this pH-sensitive tumor agnostic imaging agent can be used for image guided surgery, detection of occult disease and margin assessment. Analysis on microscopic biodistribution of ONM-100 is currently being performed and applications for metastatic lymph node detection will be explored.
F.J. Voskuil and P.J. Steinkamp contributed equally to this study
ONM-100 visualizes tumor both in- and ex vivo
Overview of two representative patients who received ONM-100 24±8h prior to surgery. Top: Intra-operative sharply demarcated fluorescence tumor visualisation of HNSCC of the tongue. Bottom: Postoperative sharply demarcated fluorescence tumor visualization of a mastectomy specimen during pathology analysis showing a high correlation with histopathology (H/E). Legend: + Vital tumor tissue.* Fat and necrosis.
Keywords: Image Guided Surgery, Generic Imaging Agent, Near Infrared, Optical Imaging
|1:15 p.m.||PL 09-11a-02||
Precision brain tumor theranostics using [123I]I-PARPi, an Auger radiation emitter (#322)
Giacomo Pirovano1, Stephen A. Jannetti1, 2, Ahmad Sadique1, Susanne Kossatz1, Navjot Guru1, Lukas M. Carter1, Jason S. Lewis1, John L. Humm1, Thomas Reiner1, 3
1 Memorial Sloan Kettering Cancer Center, Radiology, New York, New York, United States of America
In the USA, glioblastoma (GBM) has a median survival of only 12 - 15 months and a 5-years survival of 5% [1, 2]. In an effort to respond to this problem, we developed and characterized a theranostic small molecule, [123I]I-PARPi, which simultaneously leverages the highly-localized energy deposition of the Auger electron emitter 123I, with the specific targeting ability of the PARPi scaffold. [123I]I-PARPi allows to specifically deliver lethal DNA damage to targeted cancer cells while conjointly allowing SPECT imaging for dosimetry, pre-therapy dose planning, and monitoring disease progression.
Synthesis of [123I]I-PARPi has been previously described . A stannylated benzoic acid is radiolabeled with Iodine-123, purified, dried, and conjugated to a PARPi scaffold to create [123I]I-PARPi at a clinically relevant specific activity of 3.93 GBq/μmol (Fig. 1A). We performed in vitro assays to show blockable uptake and efficacy in cancer cell killing. We tested the molecule in vivo in a subcutaneous patient-derived xenograft model (PDX). SPECT/CT images were acquired to show specificity and pharmacokinetics. We quantified biodistribution to show tumor uptake. We developed an orthotopic GBM model to monitor survival after intratumoral delivery of the [123I]I-PARPi. Different therapeutic regimens were tested including pre-irradiated brains and osmotic pump continuous delivery.
In vitro data showed ~10% cellular uptake of [123I]I-PARPi which could be blocked with a 100-fold excess dose of Olaparib. Viability assay showed efficacy of the compound with an EC50 of ~100 nM, which is not observed when treating cells with the unlabeled PARP1 inhibitor (Fig. 1). In vivo we performed SPECT/CT imaging of subcutaneous and orthotopic PDX models (Fig. 2A, B). These, and biodistribution studies, showed tumor binding of [123I]I-PARPi. Survival was monitored in orthotopic-tumor bearing mice. A [123I]I-PARPi injection at week 3 after implantation improved median survival from 5.6 to 8.3 weeks (p-value=0.009). This was increased (p-value=0.002) by inducing overexpression of PARP1 by pre-irradiating the brain (2 Gy external beam radiation) (Fig. 2C). MRI images of treated mice brains confirmed tumor regression (Fig. 2D, E). To further improve survival we implanted osmotic pumps on the back of tumor bearing mice for a continuous drug delivery (1 μL/h) (Fig. 2F).
We developed and characterized the first stable, tumor-specific, Auger-emitting PARP1 inhibitor. [123I]I-PARPi can be used for precision therapy capable of selectively kill cancer cells while reducing overall radiation burden in comparison to traditional radionuclide therapies. The ability to image tumors through SPECT also offers insight into disease regression during therapy, providing valuable feedback to both the clinician and the patient.
[123I]I-PARPi structure, binding, and efficacy in vitro.
A. Crystal structure of PARP1 bound by PARP inhibitor. PDB ID# 4OPX. B. Structure of Olaparib (top) and [123I]I-PARPi (bottom). C. Cellular uptake of [123I]I-PARPi in U251 GBM cell line. Blocking performed with 100 fold incubation of Olaparib prior to [123I]I-PARPi treatment. Sodium Iodine-123 represented by dashed line. D. Comparison of [123I]I-PARPi with Olaparib at equimolar concentrations.
Treatment of TS543 xenograft mouse model with [123I]I-PARPi
A. SPECT/CT showing tumor accumulation. B. Tumor accumulation of [123I]I-PARPi in orthotopic TS543 brain tumor model. C. Survival data showing vehicle (median survival=5.6 weeks, n=12), [123I]I-PARPi single injection (370-1110 kBq) (median survival=8.3 weeks, n=11), and pre-irradiated (2 Gy external beam) (n=9). D. MRI and H&E. E. Treated mouse MRI. F. Monte Carlo dose modeling of an osmotic pump.
Keywords: glioblastoma, SPECT, Auger, theranostics, Iodine-123
|1:30 p.m.||PL 09-11a-03||
Non-invasive Translational Molecular Imaging in Duchenne Muscular Dystrophy (#20)
Adrian P. Regensburger1, Lina Fonteyne2, Jörg Jüngert1, Matthias Qurashi3, Markus F. Neurath3, 4, Nikolai Klymiuk2, Wolfgang Rascher1, Regina Trollmann1, Eckhard Wolf2, Maximilian J. Waldner3, 5, Ferdinand Knieling1
1 Friedrich-Alexander-University of Erlangen-Nuremberg, Department of Pediatrics and Adolescent Medicine, Erlangen, Bavaria, Germany
Duchenne muscular dystrophy (DMD) is the most common X-linked lethal genetic muscular disease in newborns. It is caused by loss-of-function mutations in the dystrophin gene producing a dysfunctional protein leading to muscular degeneration and inflammation, followed by fibro-fatty transformation with loss of physical muscle function. Non-invasive imaging with multispectral optoacoustic tomography (MSOT) in the extended near-infrared range was used in a bench-to-bedside approach to assess subcellular components of affected muscles in a porcine DMD model and in pediatric DMD patients in vivo.
For optoacoustic imaging eleven wavelengths (680-1100nm) were used to quantitatively assess deoxygenated (HbR), oxygenated hemoglobin (HbO2), lipids, mean (CMEAN), and maximum collagen (CMAX) contents in vivo by spectral unmixing. MSOT was performed in n=7 transgenic DMD (deleted DMD exon 52) piglets and n=10 corresponding wildtypes (WT) on day 1-3 after birth. Resultswere compared to corresponding ex vivotissues. In parallel, a first-in-pediatric monocentric, open-label, parallelized clinical trial was performed with n=10 ambulant DMD patients and n=10 matched healthy volunteers (HV). Diagnostic performance was gauged by clinical standard physical examinations (timed function tests, e.g. 6-minute walk test) and B-Mode ultrasound.
2D MSOT detected fibrotic degeneration in DMD piglets by means of increased CMEAN (14.22±1.96a.u. vs. 22.68±3.48a.u., p<0.001) and CMAX (27.69±1.68a.u. vs. 40.84±4.88a.u., p<0.001) when compared to WT piglets (Fig. 1A). Ex vivo histopathology confirmed muscular dystrophy and a qualitative increase in collagen formation in DMD piglets (Fig. 1A). Similarly, increased CMEAN (15.29±2.35a.u. vs 25.05±2.66a.u., p<0.001) and CMAX (27.62±3.06a.u. vs 40.17±3.18a.u., p<0.001) was detected in all muscle regions of DMD patients when compared to HV (Fig. 1B). 3D MOST demonstrated significant increase in CMEAN and CMAX (all p<0.001) and decreased HbR(9.13±2.38a.u. vs 5.86±1.90a.u., p=0.008) and HbO2(15.70±3.63a.u. vs 7.47±1.18a.u., p<0.001) in DMD patients compared to HV (Fig. 2A). The degree of collagenMEAN/MAX corresponded well with physical performance (6-MWT vs. C2D/3D-MEAN/MAXr=-0.75-(-0.73), all p<0.001) and was independent from age (r=-0.13-0.01, p=0.58-0.98, Fig. 2B).
This study demonstrates the potential of 2D/3D MSOT imaging to visualize collagen in vivoand suggests its application as a non-invasive, age independent biomarker for the assessment of disease progression in DMD patients. (Animal Welfare Authority District Government of Upper Bavaria, Reference Number 55.2-1-54-2532-163-2014; clinicaltrial.gov ID: NCT03490214)
2D MSOT imaging of newborn piglets and young children
A:Exemplary images of a WT (upper row) and a DMD piglet (bottom row). The merged image visualizes the collagen distribution. Representative Hematoxylin & Eosin (H&E), Masson’s Trichrome (TriC), and dystrophin (Dys1) stainings from imaged piglet musculature. Pooled mean and maximal collagen signals of both groups demonstrating statistically significant signal differences between WT and DMD piglets.
B:Exemplary MSOT images of transversal scans from every anatomical region of a HV (left panels) compared to a DMD patient (right panels). Merged images show MSOT signals for Hb (red) and collagen (turquoises) as color-coded maps. Pooled CMEANand CMAXsignals (a.u.) indicating significant differences when comparing HV and DMD patients. Normalized optoacoustic spectra of collagen (collagen peak indicated at 1000nm). ***p<0.001
3D MSOT imaging and correlation to clinical standard examination
A:3D MSOT images of the same two boys from Figure 1. Maximum projection images of the gastrocnemius muscle in two axes (XZ and YZ) and a 3D volumetric (volume) area are depicted with color-coded maps of HbTin red, CMEANin turquoise and lipid in yellow. Quantification of 3D MSOT parameters. Correlation between collagen and deoxygenated (HbR)/oxygenated hemoglobin (HbO2).
B:Correlation matrix of physical examinations and MSOT parameters. Dark blue fields indicate highly positive and dark red highly negative correlation (Pearson’s correlation coefficient given in numbers). Correlation between 2D collagenMEANand 6-MWT/Sit to stand-test/age.
Red triangles represent DMD patients and black dots HV.**=p<0.01, ***=p<0.001.
Keywords: MSOT, Duchenne Muscular Dystrophy, Biomarker, Pediatric imaging, Optoacoustic