This PDF file contains the front matter associated with SPIE Proceedings Volume 10478, including the Title Page, Copyright information, Table of Contents, and Conference Committee listing.

Full Field OCT (FFOCT) is a shot noise limited interferometric microscopy technique that uses incoherent light and has proved to be an effective diagnostic tool in terms of sensitivity and specificity. We have used the FFOCT setup built by LLTech for the analysis of various cancerous tissues corresponding to the following organs: breast, skin, prostate, lungs, ENT, bladder, brain etc. The scores obtained were found in the range between 80 and 98%. To do better and to provide informations that the histology does not carry we have studied, using the same setup, the temporal dependence of our signals which we found to be related to the cellular metabolism. We have used the new high speed and high full well capacity of the Adimec camera to achieve a time analysis ranging between 2 and a few thousands of ms. We thus obtain a new contrast which constitutes a biomarker at the sub-cellular scale. We monitor the characteristic frequencies and amplitude of the signal and display them on the images of the tissues using a new processing code of the time series. This metabolic contrast also reveal the evolution of the activity of cancer cells under treatments such as chemotherapy. We will illustrate this new approach through examples of cancer tissues that are planned to be used as intraoperative tools.

Despite improvements in surgical resection, 20-40% of patients undergoing breast conserving surgery require at least one additional re-excision. Leveraging the unique surface expression of heat shock protein 90 (Hsp90), a chaperone protein involved in several key hallmarks of cancer, in breast cancer provides an exciting opportunity to identify residual disease during surgery. We developed a completely non-destructive strategy using HS-27, a fluorescently-tethered Hsp90 inhibitor, to assay surface Hsp90 expression on intact tissue specimens using a fluorescence microendoscope with a field of view of 750 μm and subcellular resolution of 4 μm. HS-27 consists of an FDA approved Hsp90 inhibitor tethered to fluorescein isothiocyanate (EX 488nm, EM 525nm). Here, we optimized ex vivo HS-27 administration in pre-clinical breast cancer models and validated our approach on 21 patients undergoing standard of care ultrasound guided core needle biopsy. HS-27 administration time was fixed at 1- minute to minimize imaging impact on clinical workflow. HS-27 and HS-217 (non-specific control) doses were modulated from 1 μM up to 100 μM to identify the dose maximizing the ratio of specific uptake (HS-27 fluorescence) to non-specific uptake (HS-217 fluorescence). The specificity ratio was maximized at 100 μM and was significantly greater than all other doses (p<0.05). We applied our optimized imaging protocol to clinical samples and demonstrated significantly greater uptake of HS-27 by tumor than non-tumor tissue (p<0.05). The ubiquitous nature of HS-27 binding to all subtypes of breast cancer makes this technology attractive for assessing tumor margins, as one agent can be used for all subtypes.

Tumor-associated extracellular vesicles (TEVs), which represent a unique kind of inter-cellular communication carrier, have been found to play vital roles in directing the invasion and metastasis of tumor cells. However, because the human tumor microenvironment and TEVs significantly degrade or lose vitality over relatively brief periods of time after breast cancer surgical excision, lab-based studies with fresh human tissue specimens cannot provide accurate TEV information. By designing and building a portable label-free nonlinear imaging system, we have been able to conduct intraoperative imaging of fresh, unstained breast tissue specimens immediately after excision. Various features of the breast tumor microenvironment from multimodal nonlinear images were characterized to indicate tumor progression, invasiveness, and tumor grade, such as tumor-accommodating collagen structure visualized using second harmonic generation imaging, fibroblasts shown by two photon auto-fluorescence, and TEVs highlighted using third harmonic generation imaging. In particular, we found TEV count as a promising biomarker of tumor aggressiveness and margin distance. A decreasing trend of TEV counts with larger margin distance and lower cancer aggressiveness grades was revealed among 18 breast cancer cases. In addition, clear differences in TEV counts between images collected from breast cancer cases and healthy breast reduction cases, in another aspect, validate the potential of identifying TEVs using our imaging method. Acquisition and interpretation of these intraoperative image data not only provided assessment of the human tumor microenvironment, but also offered the potential to intraoperatively assess tumor margin distance and determine tumor aggressiveness.

Raman spectroscopy provides a wealth of diagnostic information to the surgeon with in situ cancer detection and label-free histopathology in intraoperative conditions. Raman spectroscopy is a promising optical technique which can analyze biological tissues with light scattering. The difference in frequencies between the incident light and the scattering light are called Raman shifts, which correspond to the vibrational energy of the molecular bonds. Raman spectrum gives information about the molecular structure and composition in biological specimens. We had been previously reported that Raman spectroscopy could distinguish various histological types of human lung cancer cells from normal cells in vitro, and also confirmed that Raman spectra obtained from cancer cells and their environment including other cells and extracellular matrix in xenograft models and spontaneous metastasis models were distinguishable using Raman spectroscopy combined with fluorescence microscopy and photoluminescence imaging. Malignancy can be characterized not only by the cancer cells but also by the environmental factors including immune cells, stroma cells, secretion vesicles and extracellular matrix, but to identify and detect cancer diagnostic biomarkers in vivo on Raman spectroscopy is still challenging. Here we investigate morphological and molecular dynamics in advanced cancer specimens obtained from patients. We are also constructing a customdesigned Raman spectral imaging system for both in vitro and in vivo assay of tumor tissues to reveal the metastasis process and to evaluate therapeutic effects of anti-cancer drugs and their drug delivery toward the clinical application of the technique.

Surgery is the first-line therapy for most solid cancers, but it’s effectiveness as a cancer treatment is reduced when all cancer cells are not detected during surgery. Instead, these residual cancer cells remain leading to recurrences that negatively impact survival. We present the development and validation of handheld multimodal optical spectroscopy imaging systems to guide surgical procedures in brain, prostate and lung cancer. These systems rely on statistical models using inelastically scattered light and intrinsic tissue fluorescence spectra for live tissue classification for interventional medicine applications including surgical resection guidance and molecularly-targeted biopsy collection. These highly sensitive optical molecular imaging tools can profoundly impact a wide range of surgery and oncology procedures by improving cancer detection capabilities, reducing cancer burden, and thereby improving survival and quality of life.

Mueller matrix polarimetry (MMP) can be utilized to determine optical anisotropy in birefringent materials. Many factors must be optimized to improve the quality of information collected from MMP of biological samples. As part of a study of pre-term birth (PTB) that relied on measurement of the orientation and distribution of collagen in the cervix, an optimal wavelength for MMp to allow more accurate characterization of collagen in cervical tissue was sought. To this end, we developed a multispectral Mueller matrix polarimeter and conducted experiments on ex-vivo porcine cervix samples preserved in paraffin. The Mueller matrices obtained with this system were decomposed to generate orientation and retardation images. Initial findings indicate that wavelengths below 560 nm offer a more accurate characterization of collagen anisotropy in the porcine cervix.

A great unmet need in oncologic surgery is the ability to accurately identify tumor-positive margins during surgical resections and to rapidly assess the margin status of resection specimens immediately following surgery. While the development of tumor-targeted fluorescent probes is a major area of investigation, it will be several years before these probes are realized for clinical use. We report the use of Indocyanine-green (ICG), a clinically approved, non-targeted dye, in conjunction with fluorescence lifetime detection to provide high accuracy for tumor detection in living mice. The improved performance relies on the distinct fluorescence lifetimes of ICG within tumors compared to tissue autofluorescence, and is further aided by the well-known enhanced permeability and retention of ICG in tumors and the clearance of ICG from normal tissue several hours after intravenous injection. Using in vivo models of human breast and brain tumors, we show that fluorescence lifetime contrast can provide a more than 98% sensitivity and specificity, and a 10-fold reduction in error rates compared to fluorescence intensity. Our studies suggest the significant potential of lifetime-contrast for accurate tumor detection using ICG and other targeted probes under development, both for intra-operative imaging and for ex-vivo margin assessment of surgical specimens

Surgery is the primary curative option for patients with cancer, with the overall objective of complete resection of all cancerous tissue while avoiding iatrogenic damage to healthy tissue. Simultaneous imaging of weak fluorescence signals from multiple targeted molecular markers under bright surgical illumination is an unmet goal with current intra operating instrument. In this talk, I will describe our recent efforts in solving this intraoperative challenge by drawing inspiration from the visual system of the mantis shrimp – a compact biological system optimized for multispectral imaging. We have successfully designed, tested and clinically translated our bio-inspired imagers by monolithically integrating vertically stacked photodetectors with pixelated interference filters. The sensor is capable of recording color and NIR fluorescence from three different molecular markers and display this information using augmented reality goggles. The sensor resolution is 1280 by 720 and operates at 30 frames per second and has been used to simultaneously image tumor targeted dye IR800 and nerve targeted dye, Oxazine-4. Displaying this information in the operating room is a challenging feat. We have used variety of augmented reality displays and will provide overview of both pre-clinical and clinical translation of this technology.

Wide-field fluorescent imaging for fluorescence molecular guidance has become a promising technique for use in imaging guided surgical navigation, but quick and intuitive microscopic inspection of fluorescent hot spots is still needed to confirm local disease states of tissues. To address this unmet need, we have developed a clinically translatable dual-modality handheld surgical microscope that incorporates both, wide-field (mesoscopic) fluorescence imaging and high-resolution (microscopic) horizontal optical-sectioning. This is accomplished by integrating a commercially available wide-field fiberscope, modified for two-color (660nm and 785nm) fluorescent detection, into a compact package (5.5 mm dia.) which also contains a dual-axis confocal (DAC) microscope. DAC microscopy is a high-sensitivity, high-resolution fluorescent imaging technology that benefits from the specificity of molecular probes, and enables interrogation of deeper regions of tissue by performing optical-sectioning of tissue. The DAC microscope has been designed with custom catadioptric micro-lenses to provide broadband multispectral capability for fluorescence imaging of multiple fluorophores over a broad spectral range (VIS to NIR), and also uses a novel MEMS-based scanning system for horizontal sectioning, and thus enables access to deeper regions of tissue at resolutions comparable to histological analysis. Large field-of-view (mm scale) is further provided by image mosaicing. The instrument thus provides simultaneous mesoscopic and microscopic fluorescence imaging over a broad spectral range for intuitively performing fast in-vivo search and microscopic confirmation of optical molecular markers in tissue, which is a capability that will become increasingly important for precise tumor resection in oncology as more optical molecular markers become approved for human use.

Image-guided surgery (IGS) can improve the patient’s outcome by providing meaningful real-time information about the location of cancerous tumors and surrounding tissue, aiding in the elimination of positive tumor margins and reducing iatrogenic damage. However, the clinical need for imaging systems that can provide real-time feedback under real operating room settings remains unmet. State-of-the-art imaging systems for near-infrared fluorescence IGS rely on a series of complex optics and several imaging sensors. As a result, these systems are bulky and expensive, and their architecture lacks the versatility to simultaneously image multiple fluorophores, effectively making them cumbersome when merged into the current surgical workflow. To address these shortcomings, we have designed a multi-spectral imager capable of spatially co-registered hexachromatic vision: three spectral channels in the visible spectrum for the identification of anatomical features in color and three spectral channels in the near-infrared spectrum for the simultaneous identification of multiple near-infrared fluorescence dyes used in IGS. Our single-chip imaging sensor combines the vertically stacked photodetectors technology with pixelated interference filters to create a multi-spectral imager that can help surgeons make clinically relevant decisions in real time, with an effective resolution of 1280x720x3 photodiodes and a frame rate of 24 FPS. Our imager has the ability to identify different shades of near-infrared fluorescent light, allowing the surgeon to use and differentiate multiple fluorophores as molecular probes with high sensitivity. Pre-clinical data is shown where simultaneous imaging of anatomical features in color, and identification of nerves and cancerous tumors, are achieved using multiple near-infrared fluorescent agents.

Two fundamental and unsolved problems facing bioimaging and nanomedicine are nonspecific uptake of intravenously administered diagnostic and/or therapeutic agents by normal tissues and organs, and incomplete elimination of unbound targeted agents from the body. To solve these problems, we have synthesized a series of indocyanine near-infrared (NIR) fluorophores that varied systematically in net charge, conformational shape, hydrophilicity/lipophilicity, and charge distribution. Using 3D molecular modeling and optical fluorescence imaging, we have defined the relationship among the key independent variables that dictate biodistribution and tissue-specific targeting such as lung and sentinel lymph nodes (Nat Biotechnol. 2010), human prostate cancers (Nat Nanotechnol. 2010), and human melanomas (Nat Biotechnol. 2013). Recently, we have developed new pharmacophore design strategy “structure-inherent targeting,” where tissue- and/or organ-specific targeting is engineered directly into the non-resonant structure of a NIR fluorophore, thus creating the most compact possible optical contrast agent for bioimaging and nanomedicine (Angew Chem. 2015, Nat Med. 2015). The biodistribution and targeting of these compounds vary with dependence on their unique physicochemical descriptors and cellular receptors, which permit 1) selective binding to the target tissue/organ, 2) visualization of the target specifically and selectively, and 3) provide curing options such as image-guided surgery or photo dynamic therapy. Our study solves two fundamental problems associated with fluorescence image-guided surgery and lays the foundation for additional targeted agents with optimal optical and in vivo performance.

Skull base tumors are particularly difficult to visualize and access for surgeons because of the crowded environment and close proximity of vital structures, such as cranial nerves. As a result, accidental nerve damage is a significant concern and the likelihood of tumor recurrence is increased because of more conservative resections that attempt to avoid injuring these structures. In this study, a paired-agent imaging method with direct administration of fluorophores is applied to enhance cranial nerve identification. Here, a control imaging agent (ICG) accounts for non-specific uptake of the nerve-targeting agent (Oxazine 4), and ratiometric data analysis is employed to approximate binding potential (BP, a surrogate of targeted biomolecule concentration). For clinical relevance, animal experiments and simulations were conducted to identify parameters for an optimized stain and rinse protocol using the developed paired-agent method. Numerical methods were used to model the diffusive and kinetic behavior of the imaging agents in tissue, and simulation results revealed that there are various combinations of stain time and rinse number that provide improved contrast of cranial nerves, as suggested by optimal measures of BP and contrast-to-noise ratio.

Nerve damage plagues surgical outcomes, significantly affecting post-surgical quality of life. Surprisingly, no method exists to enhance direct nerve visualization in the operating room, and nerve detection is completed through a combination of palpation and visualization when possible. Fluorescence image-guided surgery offers a potential means of enhanced nerve identification and preservation, however a clinically approved nerve-specific contrast agent does not yet exist. Several classes of nerve-specific fluorophores have recently been demonstrated including the distyrylbenzenes (DSB), select oxazines (oxazine 4 perchlorate), and certain cyanines (3,3’-diethylthiatricarbocyanine iodine), which could provide intraoperative guidance. The nerve-sparing radical prostatectomy is a surgical procedure that could benefit from fluorescence image-guided nerve identification. Although the nerve-sparing surgical technique was developed over 30 years ago, nerve damage following radical prostatectomy is reported in some form in up to 60% of patients one to two years post-surgery. To facilitate clinical translation of fluorescence image guided surgery to the nerve sparing prostatectomy, a direct administration methodology was developed that allows selective nerve highlighting with a significantly lower fluorophore dose than systemic administration, where large animal studies have confirmed the technique’s translatability. Tissue penetration will be critical for clinical feasibility of the direct administration methodology and novel formulation strategies have been explored to enhance tissue penetration for identifying buried nerves. In addition, several biomolecular targets of Oxazine 4, a promising candidate for nerve-specific fluorophore development into a near-infrared (NIR) agent, have been identified, providing insight into the mechanism of nerve-specificity. Fluorophore development has made progress towards the goal of creating a NIR nerve-specific fluorophore and determining the structure-activity relationship responsible for nerve binding.

Molecular guided surgery—employing injectable fluorescent imaging agents targeted to cancer-specific biomolecules—is advancing rapidly. However, cancer physiology, specifically blood flow, vascular permeability, and lymphatic drainage can significantly influence and even dominate how imaging agents are taken up and retained in cancerous tissue. Moreover, each of these physiological parameters can vary extensively from tumor-type to tumor-type, from patient-to-patient, and even within an individual’s cancer, making it impossible to know if tumor contrast is achieved through molecular targeting or enhanced permeability and retention (EPR) effect. Alternatively, if there is no contrast, there is no guarantee that the tissue is non-cancerous, only that there was no accumulation of imaging agent, which could be attributable to low blood flow or vascular permeability. This talk will explore the potential effects of cancer physiology on imaging agent uptake and retention in cancerous tissue and discuss methods such as paired-agent imaging, which work to account for tumor physiology variability to achieve true molecular guided surgery.

Complete removal of malignant tissue during primary breast cancer resection is a critical prognostic indicator of local recurrence and overall patient survival, making intraoperative tumor margin assessment essential. Positive margin status following breast-conserving surgery (BCS) is a common difficulty reported in 20-60% of patients, with re-excision rates >55%. Re-excision increases the risk of morbidity and delays the use of adjuvant therapy, thus significant efforts are underway to develop successful intraoperative margin assessment strategies to eliminate repeat surgery. One novel strategy uses topical application of dual probe staining and a fluorescence imaging strategy termed dual probe difference specimen imaging (DDSI) where a receptor-targeted fluorescent probe and an untargeted, spectrally-distinct fluorescent probe are topically applied to the fresh resected specimen. While conceptually simple, resected specimen staining is dominated by non-specific uptake of fluorescent probes in normal tissue, requiring the use of a dual probe strategy for accuracy. DDSI permits fluorescence images from both the targeted and untargeted probes to be used to calculate a normalized difference image, facilitating quantitative identification of targeted probe tumor distribution in the specimen. While previous reports suggested this approach is a promising new tool for surgical guidance, advancing the approach into the clinic requires methodical protocol optimization and validation. Current work is focused on development of targeted and untargeted small molecule affinity tags, facilitating access to intracellular breast cancer biomarkers and quantitative assessment of DDSI signal in the context of varied biomarker expression level.

Intraoperative tumor/surgical margin assessment is required to achieve higher tumor resection rate in breast-conserving surgery. Though current histology provides incomparable accuracy in margin assessment, thin tissue sectioning and the limited field of view of microscopy makes histology too time-consuming for intraoperative applications. If thick tissue, wide-field imaging can provide an acceptable assessment of tumor cells at the surface of resected tissues, an intraoperative protocol can be developed to guide the surgery and provide immediate feedback for surgeons. Topical staining of margins with cancer-targeted molecular imaging agents has the potential to provide the sensitivity needed to see microscopic cancer on a wide-field image; however, diffusion and nonspecific retention of imaging agents in thick tissue can significantly diminish tumor contrast with conventional methods. Here, we present a mathematical model to accurately simulate nonspecific retention, binding, and diffusion of imaging agents in thick tissue topical staining to guide and optimize future thick tissue staining and imaging protocol. In order to verify the accuracy and applicability of the model, diffusion profiles of cancer targeted and untargeted (control) nanoparticles at different staining times in A431 tumor xenografts were acquired for model comparison and tuning. The initial findings suggest the existence of nonspecific retention in the tissue, especially at the tissue surface. The simulator can be used to compare the effect of nonspecific retention, receptor binding and diffusion under various conditions (tissue type, imaging agent) and provides optimal staining and imaging protocols for targeted and control imaging agent.

Specific tumor targeting can result in selective labeling of cancer in vivo for surgical navigation. In the present study, we show that the use of an anti-CEA antibody conjugated to the near-infrared (NIR) fluorescent dye, IRDye800CW, can selectively target and label pancreatic cancer and its metastases in a clinically relevant patient derived xenograft mouse model.

Total thyroidectomy (TT) is responsible for transient postoperative hypocalcemia (PH) in 20-30% of patients. This complication results from surgery-induced parathyroid damage, vascular impairment or unintentional parathyroid resection of these small glands, which can be difficult to identify visually. Fluoptics introduced lately a Near-infrared (NIR) Fluorescence imaging device able to visualize auto-fluorescence of parathyroid gland during thyroid surgery without injection of any fluorescent dye. In order to assess the impact of this technology on the post-thyroidectomy transient hypocalcemia rate, we performed a before-after study where we compared all consecutive patients who underwent TT with the intraoperative use of the Fluobeam® NIR camera (NIR+ group) and those operated on without NIR (NIR- group), by the same surgeon. Parathyroid glands were identified via NIR camera before they were visualized by the surgeon in 68% of patients in the NIR+group. In the NIR+group, postoperative hypocalcemia rate was significantly lower (5.2% vs 20.9%, p<0.05), mean number of identified parathyroid glands was significantly higher (3.1±0.9 vs 2.6±1.0 per patient, p<0.05) and parathyroid auto-transplantation rates was significantly lower (2.1% vs 15% of patients, p<0.05) than in the NIR- group, although no difference was observed in inadvertent resection rates. This trial seems to demonstrate that auto-fluorescence-based visualization of the parathyroid glands using NIR light during total thyroidectomy can significantly reduce postoperative hypocalcemia rates and improve identification and preservation of parathyroid glands.

Ovarian cancer is the most lethal gynecologic malignancy in women. Due to the absence of early symptoms and the lack of effective screening methods, most women are diagnosed at an advanced stage. Large tumors, which are readily seen by surgeons, require no detection assistance, but metastatic tumors, particularly the intraperitoneal-spread tumors, are difficult to identify beaause of their small size, variable morphology, unpredictable location, and similarity to normal tissues. A highly sensitive and specific sprayable fluorogenic probe was thus developed to visualize these hard-to-detect micro tumors. The rapid fluorogenic signal development will allow for a real-time guidance in the operating room.

BACKGROUND: Presence of lymph node (LN) metastasis is considered the most important prognostic factor in patients with head and neck cancer, yet intraoperative identification of metastatic LNs is considered challenging. We propose the near-infrared fluorescently labeled epidermal growth factor receptor (EGFR) antibody panitumumab-IRDye800 for intraoperative metastatic LN identification. METHODS: Patients were injected 2-5 days before surgery with panitumumab-IRDye800 (0.5 or 1.0 mg/kg). On the day of surgery, (excised) LN samples were evaluated on high sensitivity fluorescence imaging systems (SurgVision (SurgOptix), PINPOINT (Novadaq), and Pearl imager and Odyssey CLx (LI-COR Biosciences). Location and intensity of the fluorescence signal was correlated to the location of tumor as defined on the hematoxylin and eosin staining by the pathologist, and the EGFR expression pattern. We calculated the sensitivity, specificity, positive and negative predictive values of panitumumab-IRdye800 for metastatic LN identification. RESULTS: We thus far included 9/27 patients in our ongoing phase I trial. 244 LNs were removed intraoperatively of which 8 were tumor-positive. Fluorescence imaging of panitumumab-IRdye800 revealed 236 true-negative nodes (not fluorescent, not tumor-positive), 8 true-positive nodes (fluorescent, tumor-positive), 0 false-positive nodes (fluorescent, not tumor-positive) and 0 false-negative nodes (not fluorescent, tumor-positive) resulting in a sensitivity of 100%, a specificity of 100%, and a positive and negative predictive value of 100% and 100%, respectively. CONCLUSION: Preliminary results from our ongoing study suggest panitumumab-IRDye800 can identify metastatic LNs. Upon trial progression, if findings remain constant, it can open a whole new era for intraoperative metastatic LN identification.

This talk will review the current clinical trial results using antibody-based optical imaging to guide surgical resections in head and neck, brain, skin and pancreas tumors. Each tumor type presents unique challenges that dramatically alters how the optical imaging data can provide clinical utility to the surgical team and pathologists. Specific cases and synthesized data will be presented and regulatory hurdles discussed.

The purpose of our study is to evaluate patient safety parameters, pharmacokinetics, and ureteral fluorescence/ease of visualization of the compound IS-001, an intravenously administered, renally cleared indigo-cyanine dye in patients undergoing robotic total laparoscopic hysterectomy using the da Vinci Xi Firefly® fluorescent imaging system. Our study design is a Phase I Open Label Case Series Trial. Our intervention is the IV administration of increasing doses (10 mg, 20mg and 40mg) of IS-001 (2mg/mL) in each eight patient cohort. Pharmacokinetic data was collected at predefined intraoperative set points and at 2, 4, and 6 hours post-op to determine maximum blood concentration, total drug exposure, and time course of excretion. Subjective surgeon assessment of ureteral fluorescence was compared to computer calculations of fluorescence intensity to evaluate ease and duration of ureteral visualization using Firefly®. IS-001 appears safe in human subjects. Pharmacokinetic data is consistent with preclinical findings in animal subjects. Subjective surgeon assessment of ureteral fluorescence indicates rapid onset of ureteral fluorescence that persists to a clinically useful degree for 30+ minutes.

With 2500 scanners installed and 2 million scans performed last year in the United States, positron emission tomography (PET) has emerged as one of the most important imaging tools in oncology. The widespread availability of PET imaging and radiochemistry facilities has stirred the development of new PET tracers. The current pipeline for radiotracer development uses analytical tools that assume that the targeted cell populations are homogeneous. For instance, in vivo imaging tests reduce results to a single quantitative metrics, the standardized uptake value, which represents the average behavior of millions of cells. In reality, tumors and other organs are formed by clonally heterogeneous collections of different cell types; the average of these cells is not necessarily representative of the individual cells that make up the population. To give an example, the PET signal measured in a tumor is often ascribed to cancer cells alone when in fact other cell types such as immune cells and stromal cell contribute significantly to the average signal. These considerations have led to the development of a novel method, radioluminescence microscopy (RLM), that can characterize radiotracer uptake heterogeneity at the level of single cells. RLM combines optical and ionizing radiation to yield high-resolution microscopic images of cells and their interactions with radiotracers. The fundamental principles of RLM will be explained and its application to radiotracer discovery and validation will be presented.

Cerenkov luminescence (CL), optical radiation induced by PET radiotracers, has shown promise as a means to visualize tumor margins during surgery. However, detecting this faint optical signal under ambient lighting conditions represents a major challenge. We have developed an ambient light CL imaging system that uses a sensitive imaging detector, custom electronic control board, and an LED illumination array. By gating both LED illumination and imaging detector, we have demonstrated that is possible to image faint Cerenkov-emitting sources in a perceptually well-lit room, without harm to the sensitive detector. System performance was characterized by imaging 18F radionuclide solution contained in 10 mm well plates, ranging in activity from >1 MBq to <1 kBq, under visible light conditions with irradiances ranging from 0 to >30 µW/cm2. Both detector and LED illumination were gated at 30 Hz with 10 ms active duty cycles. Contrast-to-noise ratio (CNR) was computed from ROIs drawn over activity-containing wells. Though CNR decreased with increasing illumination levels, an activity of 240 kBq, was unambiguously detectable with gated illumination of 37 µW/cm2 (equivalent to typical indoor light levels) and an activity of <24 kBq was unambiguously detectable with gated illumination of 2 µW/cm2. Furthermore, we have characterized sources of noise for the imaging system, which have provided insight into strategies for optimization in anticipation of use in an intraoperative setting.

Autoradiography potentially offers high molecular sensitivity and spatial resolution for tumor margin estimation. However, conventional autoradiography requires sectioning the sample which is destructive and labor-intensive. Here we describe a novel autoradiography technique that uses a flexible ultra-thin scintillator which conforms to the sample surface. Imaging with the flexible scintillator enables direct, high-resolution and high-sensitivity imaging of beta particle emissions from targeted radiotracers. The technique has the potential to identify positive tumor margins in fresh unsectioned samples during surgery, eliminating the processing time demands of conventional autoradiography. We demonstrate the feasibility of the flexible autoradiography approach to directly image the beta emissions from radiopharmaceuticals using lab experiments and GEANT-4 simulations to determine i) the specificity for ¹⁸F compared to ^99mTc-labeled tracers ii) the sensitivity to detect signal from various depths within the tissue. We found that an image resolution of 1.5 mm was achievable with a scattering background and we estimate a minimum detectable activity concentration of 0.9 kBq/ml for ¹⁸F. We show that the flexible autoradiography approach has high potential as a technique for molecular imaging of tumor margins using ¹⁸F-FDG in a tumor xenograft mouse model imaged with a radiation-shielded EMCCD camera. Due to the advantage of conforming to the specimen, the flexible scintillator showed significantly better image quality in terms of tumor signal to whole-body background noise compared to rigid and optimally thick CaF₂:Eu and BC400. The sensitivity of the technique means it is suitable for clinical translation.

Cherenkov imaging during radiation therapy has been developed as a tool for dosimetry, which could have applications in patient delivery verification or in regular quality audit. The cameras used are intensified imaging sensors, either ICCD or ICMOS cameras, which allow important features of imaging, including: (1) nanosecond time gating, (2) amplification by 10³-10⁴, which together allow for imaging which has (1) real time capture at 10-30 frames per second, (2) sensitivity at the level of single photon event level, and (3) ability to suppress background light from the ambient room. However, the capability to achieve single photon imaging has not been fully analyzed to date, and as such was the focus of this study. The ability to quantitatively characterize how a single photon event appears in amplified camera imaging from the Cherenkov images was analyzed with image processing. The signal seen at normal gain levels appears to be a blur of about 90 counts in the CCD detector, after going through the chain of photocathode detection, amplification through a microchannel plate PMT, excitation onto a phosphor screen and then imaged on the CCD. The analysis of single photon events requires careful interpretation of the fixed pattern noise, statistical quantum noise distributions, and the spatial spread of each pulse through the ICCD.

Cherenkov-excited luminescence scanned imaging (CELSI) has been proposed for radiation-dose determination in medical physics due to its high spatial-resolution over centimeters of tissue. However, dense line-scanning illumination in typical CELSI is time-cost owing to the mechanical movement of the leaves in multi leaf collimator (MLC), resulting into increased radiation exposure. As a result, a scanningless Cherenkov luminescence imaging modality is herein proposed through structuring epi-illumination with MLC-based Hadamard-patterns, which utilizes a reduced radiation does by limiting illumination patterns, extremely shortening the sampling process. In order to effectively reconstruct unknowns from the resultant underdetermined linear system with sparse samplings, a compressed sensing-based reconstruction methodology with l1-norm regularization is adopted. Numerical and phantom experiments show that the proposed methodology achieves the same image quality as the traditional CELSI does.

Visible light generated as the result of interaction of ionizing radiation with matter can be used for radiation therapy quality assurance. In this work, we characterized the visible light observed during proton irradiation of poly(methyl methacrylate) (PMMA) and silica glass fiber materials by performing luminescence spectroscopy. The spectra of the luminescence signal from PMMA and silica glass fibers during proton irradiation showed continuous spectra whose shape were different from that expected from Čerenkov radiation, indicating that Čerenkov radiation cannot be the responsible radioluminescence signal. The luminescence signal from each material showed a Bragg peak pattern and their corresponding proton ranges are in agreement with measurements performed by a standard ion chamber. The spectrum of the silica showed two peaks at 460 and 650 nm stem from the point defects of the silica: oxygen deficiency centers (ODC) and non-bridging oxygen hole centers (NBOHC), respectively. The spectrum of the PMMA fiber showed a continuous spectrum with a peak at 410 nm whose origin is connected with the fluorescence of the PMMA material. Our results are of interest for various applications based on imaging radioluminescent signal in proton therapy and will inform on the design of high-resolution fiber probes for proton therapy dosimetry.

Total Skin Electron Therapy (TSET) utilizes high-energy electrons to treat cancers on the entire body surface. The otherwise invisible radiation beam can be observed via the optical Cherenkov photons emitted from interaction between the high-energy electron beam and tissue. Using a specialized camera-system, the Cherenkov emission can thus be used to evaluate the dose uniformity on the surface of the patient in real-time. Each patient was also monitored during TSET via in-vivo detectors (IVD) in nine locations. Patients undergoing TSET in various conditions (whole body and half body) were imaged and analyzed, and the viability of the system to provide clinical feedback was established.

The fluorescent properties of the reduced coenzyme NADH and its phosphorylated derivative (NADPH) have been explored in order to assess their potential as an intrinsic probe for cancer surgery. NADPH production is increased in cancer cells to quench reactive oxygen species and meet higher demands for biosynthesis, and has attractive fluorescent properties such as emission towards the visible part of the spectrum and a relatively long fluorescence lifetime upon binding to enzymes (~ 1 – 6.5 ns) that helps discriminate against other endogenous species. Different environmental effects on NAD(P)H fluorescence are reported here, including an increase in lifetime upon oxygen removal, an ability to retain its fluorescent properties in a complex medium (a silica phantom) and its fluorescence lifetime also being distinguishable in a cell environment. In addition, the development of a miniaturized liquid light guide filter-based timecorrelated single photon counting fluorescence lifetime system is reported as a step towards time-resolved visual imaging in cancer surgery. This system has been demonstrated as being capable of accurately measuring NAD(P)H fluorescence lifetimes in both simple solvent and cellular environments.