This PDF file contains the front matter associated with SPIE Proceedings Volume 12825, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

Surgical excision is the primary treatment for solid tumors in oral squamous cell carcinomas, where achieving a healthy tissue margin of >5 mm is the goal. However, current clinical methods of assessing surgical margins cannot provide assessment of the whole margins intraoperatively (while the patient is still on the operating table) and while recent intraoperative fluorescence-guided surgery approaches have shown promise for detected “positive” inadequate margins (<1 mm), they have had limited success in the detection of “close” inadequate margins (1-5 mm), in patients injected with cetuximab-IRDye 800CW prior to surgery. Here, a dual aperture fluorescence ratio (dAFR) approach presented previously by our group is expanded upon, where herein we present a version of the analysis where the measurements are normalized by a background signal. We compare this additional approach directly against a single aperture view fluorescence (sAF) and pathology measurements of margin thickness in specimens from five patients and a total 14 margin locations (1 positive, 7 close, and 6 clear margins). The area under the curve of the receiver operating characteristic, representing the ability to detect close compared to clear margins was found to be 1.0 and 0.6 using dAFR and sAF, respectively, with the improvements in dAFR being statistically significant (p < 0.01). We demonstrate that the addition of a background normalization can account for noise and low signal in narrow aperture images.

Mucins are a family of glycoproteins that have recently been the focus of intense investigation for fluorescence-guided surgery of cancer due to their overexpression in many malignancies. In pancreatic cancer alone there are at least twelve mucins that are upregulated or uniquely expressed. In our group, we have been developing a near infrared fluorescent labelled mAb, NIRF-AR9.6, that targets MUC16, which is overexpressed in 60-80% of pancreatic cancers. We have previously reported that in an orthotopic xenograft model of pancreatic cancer, NIRF-AR9.6 resulted in a 3.7-fold enhancement of tumor compared to surrounding pancreas, while the isotype IgG control resulted in a 2-fold increase. We also demonstrated that NIRF-AR9.6 could enhance PDAC xenografts, even with lower levels of MUC16. Moreover, initial studies showed that NIRF-AR9.6 could enhance a PDX pancreatic cancer model, consistent with MUC16 staining throughout the tumor. We are now investigating the feasibility for NIRF-AR9.6 for fluorescence-guided surgery after neoadjuvant therapy, implementing orthotopic PDX models of varying MUC16 expression, and investigating targeting multiple mucins to account for tumor heterogeneity.

Fluorescence cryo-imaging is a high-resolution optical imaging technique that produces 3-D whole-body biodistributions of fluorescent molecules within an animal specimen. To accomplish this, animal specimens are administered a fluorescent molecule or reporter and are frozen to be autonomously sectioned and imaged at a temperature of -20°C or below. Thus, to apply this technique effectively, administered fluorescent molecules should be relatively invariant to low temperature conditions for cryo-imaging and ideally the fluorescence intensity should be stable and consistent in both physiological and cryo-imaging conditions. Herein, we assessed the mean fluorescence intensity of 11 fluorescent contrast agents as they are frozen in a tissue-simulating phantom experiment and show an example of a tested fluorescent contrast agent in a cryo-imaged whole pig brain. Most fluorescent contrast agents were stable within ~25% except for FITC and PEGylated FITC derivatives, which showed a dramatic decrease in fluorescence intensity when frozen.

High-grade gliomas (HGG) are the most common and most aggressive primary tumors of the brain. Despite recent advances in neuro-oncology survival of these tumors remains around 12-15 months. The first-line of treatment is surgical resection. However, due to its infiltrative nature maximal safe resection leaves residual invasive cancer cells that lead to disease recurrence. The vast majority of recurrences are in or near the resection cavity. Advances in optical imaging techniques might aid better delineation of the invasive margin intraoperatively. However as of yet most of these technologies have failed to do so consistently. Recently, alterations in fatty acid metabolism have been linked to the initiation, progression, and recurrence of gliomas. These alterations might provide a novel target for better differentiation of glioma cells and healthy brain tissue. To exploit this, we introduce and test a novel, near-infrared, fluorescent dye, fatty-acid indocyanine green (FA-ICG) in vitro and in vivo. We hypothesize that the combination of targeting mechanism, and the near-infrared properties make FA-ICG a promising candidate for further clinical translation in fluorescence-guided surgery.

Guided surgery has demonstrated significant improvements in patient outcomes in some disease processes. Interest in this field has led to substantial growth in the technologies under investigation. Most likely no single technology will prove to be “best,” and combinations of macro- and microscale guidance— using radiological imaging navigation, probes (activatable, perfusion, and molecular-targeted; large- and small-molecule), autofluorescence, tissue intrinsic optical properties, bioimpedance, and other characteristics—will offer patients and surgeons the greatest opportunity for high-success/low-morbidity medical interventions. Problems are arising, however, from the lack of valid testing formats; surgical training simulators suffer the same problems. Small animal models do not accurately recreate human anatomy, especially in terms of tissue volume. Large animal models are expensive and have difficulty replicating many pathological states, particularly when molecular specificity for individual cancers is required. Furthermore, the sheer number of technologies and the potential for synergistic combination leads to exponential growth of testing requirements that is unrealistic for in vivo testing. Therefore, critical need exists to expand the ex vivo/in vitro testing platforms available to investigators and, once validated, a need to increase the acceptance of these methods for funding and regulatory endpoints. Herein is a review of the available ex vivo/in vitro testing formats for guided surgery, a review of their advantages/disadvantages, and consideration for how our field may safely and more swiftly move forward through stronger adoption of these testing and validation methods.

In the United States, the annual incidence of oral squamous cell carcinoma (OSCC) exceeds 50,000 cases. Primary tumor resection remains the first line of treatment in these patients yet follow up neck dissection and chemoradiation treatment may be indicated if cancer has spread to tumor draining lymph nodes. There is a push to minimize morbidity from neck dissection by sentinel lymph node biopsy, where only the first lymph node(s) draining the primary tumor are excised and evaluated for cancer spread. However, with current pathology methods, results are not available to surgeons until patients have been sent home. In response, we are developing a method to rapidly stain and image whole excised lymph nodes in less than 30 min, so surgeons can react to positive cases while patients are still on the operating table. Here we present a human head and neck cancer spheroid model implanted in porcine lymph nodes as a means evaluating the potential for our staining and imaging protocols to rapidly identify cancer burden in lymph nodes.

Fluorescence guided surgery (FGS) has entered clinical practice over the past fifteen years and has established specific clinical niches of real patient benefit. The next phase will broaden the clinical application spectrum and harness breakthrough advances in computer vision and artificial intelligence methods to optimise interpretation especially of dynamic signals that will allow better usage of and impact from new agents and imaging methods. With a focus on tissue characterisation regarding perfusion sufficiency and pathological classification (in particular in situ digital cancer discrimination) in colorectal surgery, the current and near future state of the art is here described in the context of determinant clinical trials, experience and failures since the inception of FGS with a focus on work from our collaborative translational research group and on the remaining next steps for standard of care designation.

To assess the agent penetration in tissue, we developed two methodologies of image analysis for quantifying the extravasation distances of fluorescent imaging agents: Perpendicular Line Intensity Profiling (PLIP) and Iterative Radial Dilation Profiling (IRDP). Utilizing automated vessel identification and fluorescence microscopy, these methods provide a robust framework for analyzing agent distribution in tumor tissues. Our findings demonstrate that the IRDP method significantly mitigates noise, offering a smoother intensity decay profile compared to PLIP. This advantage suggests that IRDP could improve the reliability of assessing agent delivery and penetration, which are critical for providing effective imaging contrast in FGS. By comparing agent penetration profiles, we advocate for the broader adoption of the IRDP method in future quantitative fluorescence imaging analyses to enhance assessment accuracy and consistency.

Imaging of indocyanine green (ICG) can reveal vascular permeability, and it has been previously demonstrated in pancreatic adenocarcinoma tumors [1]. The relevance of this to clinical use has remained speculative, although it is likely that these vascular permeability measurements could be used for resection guidance or could also be predictive of drug retention or immune infiltration in dysplastic tissues in non-surgical tumors. Second-Window Indocyanine Green (SWIG) imaging, in which a high ICG dose is administered followed by imaging at hours or days post-injection, has been shown to have potential in several oncologic indications [2,3], and is dependent upon dysplastic tissues having a higher degree of bulk tissue retention. Herein, we evaluated the capability of early phase vascular permeability estimates within minutes after ICG injection, and how they may be related to the degree of ICG retention in SWIG imaging.

Pancreatic cell lines, AsPC1 or BxPC3 were grown into tumors in nude mice, providing models that display different capillary network morphologies. Using a clinical surgical fluorescence imaging system, mice were imaged for 10 minutes following bolus IV injection of 4mg/kg ICG. Mice were subsequently imaged 24 hours after the initial injection to measure the intensity of the tumor relative to a muscle tissue reference for SWIG images. The temporal slope of tissue uptake within the first few minutes was used to estimate vascular permeability.

Initial vascular permeability estimates from flow kinetics imaging were not predictive of the ICG retention in SWIG imaging. This would indicate that lymphatics or other factors likely play a larger role in determining retention.

Cherenkov-Excited Luminescence Scanned Tomography (CELST) involves a system of coupled continuous wavedomain diffusion equations for modeling. The excitation field quantized by the Complex Cosine (CC) method effectively simulates the forward light process in these equations. However, when considering x-ray-induced Cherenkov light within biological tissue, the CC-based excitation field lacks precision, instead requires the use of stochastic Monte Carlo (MC) methods. To accurately describe the radiation-induced light transport in biological tissue and CELST image reconstruction, in this paper, we develop a MC-based method for CELST, named sheet Monte Carlo (sMC). Experiments show that the sMC field can achieve 11.47 on contrast-to-noise ratio (CNR) and 0.74 on Pearson correlation (PC), while 7.25 and 0.57 for the CC initialization field under 4% noise level. Furthermore, our results highlight that the proposed excitation field exhibits superior reconstruction performance, especially when dealing with low ratios of fluorescent targets.

Paired-Agent Imaging (PAI) is a quantitative fluorescence imaging technique that estimates the drug target concentration. It involves the co-administration of a targeted and an untargeted imaging probe to correct for nonspecific uptake and to quantify the available receptor concentration, known as the binding potential. PAI has been demonstrated in a pre-clinical setting using a 1:1 molar concentration of the targeted, ABY-029, and untargeted, IRDYE 680LT, imaging agent. However, the effects of different molar concentrations of imaging agents on the binding potential have not been studies thus far. In this study, we examined this relationship in tissue-mimicking liquid phantoms with varying molar concentration ratios. The phantom fluorescence was measured using the Pearl Imaging system and then the binding potential was quantified using MATLAB. We determined that the binding potential remains stable across concentration but increases for dye ratios where the targeted dye is higher.

Fluorescence-guided surgery systems employed during oral cancer resection help detect the lateral margin yet fail to quantify the deep margins of the tumor prior to resection. Without comprehensive quantification of three-dimensional tumor margins, complete resection remains challenging. While interoperative techniques to assess the deep margin exist, they are limited in precision, leaving an unmet need for a system that can quantify depth. Our group is developing a deep learning (DL)-enabled fluorescence spatial frequency domain imaging (SFDI) system to address this limitation. The SFDI system captures fluorescence (F) and reflectance (R) images that contain information on tissue optical properties (OP) and depth sensitivity across spatial frequencies. Coupling DL with SFDI imaging allows for the near-real time construction of depth and concentration maps. Here, we compare three DL architectures that use SFDI images as inputs: i) F+OP, where OP (absorption and scattering) are obtained analytically from reflectance images; ii) F+R; iii) F/R. Training the three models required 10,000 tumor samples; synthetic tumors derived from composite spherical harmonics circumvented the need for patient data. The synthetic tumors were passed to a diffusion-theory light propagation model to generate a dataset of artificial SFDI images for DL training. Two oral cancer models derived from MRI of patient tongue tumors are used to evaluate DL performance in: i) in silico SFDI images ii) optical phantoms. These studies evaluate how system performance is affected by the SFDI input data and DL architectures. Future studies are required to assess system performance in vivo.

Pre-operative MRI with gadolinium-based contrast agents (Gd-MRI) is a central feature in surgical planning and intra-surgical navigation of glioma, yet brain movement during the surgical procedure can degrade the accuracy of these pre-operative images. Fluorescence guided neurosurgery is a technique which can complement MRI guidance by providing direct visualization of the tumor during surgery, and several agents either used routinely or under clinical development have shown effective tumor discrimination and impact on surgical outcomes. We have built a multi-spectral kinetic imaging system to acquire behavior of fluorophores overtime in animal models. Here, we exhibit this fluorescence kinetic imaging system and report its performance with tissue-simulating phantoms with multiple fluorophores. Also reported is our first experience with multiple fluorescent contrast agents in a novel oncopig model.

Fluorescence optical tomography systems have been invented to directly measure several deep tissue features. While the highly scattered nature of the signal degrades spatial resolution, this scattering process also dramatically increases the optical path length and thereby amplifies the signal sensitivity to features such as capillaries and cells, suppressing the dominance of major blood vessels. The potential for high optical contrast with centimeter-level penetration into tissue motivates sampling (i) capillary leakage with a temporal sampling of indocyanine green (ICG); (ii) mitochondrial activity with protoporphyrin IX (PpIX) intensity, and (iii) oxygen metabolism sensing with delayed fluorescence of PpIX. In this work, the single-channel prototype of an optical fiber-based tomography system was developed for these purposes. The system was comprised of two sources of laser diodes, at 633 nm and 780 nm wavelengths, two avalanche photodiode detectors, a DAQ I/O card as a control and data collection unit, and fiber cables with filter blocks, all run via LabView control. The detection fiber channels included a notch filter in each arm, at 633nm and 750nm wavelengths that act as a band stop filter for these wavelengths. Fiber cables deliver and receive light from the tissue, enabling a closed-loop passive switching system. This design has a dual-purpose channel fiber system that makes it possible to just electronically switch between sequential measurement of indocyanine green (ICG) excitation (785 nm) and protoporphyrin IX (PpIX) excitation (635 nm) without any optical component movements, while transmission of the excitation signal is implicitly measured at the opposite detector. The fluorescence-to-transmission ratio data eliminates issues of fiber coupling or tissue transmission variations. The system was validated for linearity of response in relevant biological concentration ranges for each fluorophore. Also, in vivo animal studies were carried out on a mouse to see the reliability of the system at capturing the temporal ICG kinetics. This single design of dual-channel detection is a prototype for what will be a multi-fiber tomography system that can measure the intrinsic properties of tissue coupled with ultrasound imaging.