This PDF file contains the FM associated with SPIE Proceedings Volume 11362, including the Title Page, Copyright information, table of Contents, Author and Committee lists.

Collagen is the major component of connective tissues, where it assembles into fibrils that exhibit various sizes and form various 3D structures depending on the observed tissue. Any disruption of this microstructure is associated with tissue malfunction and defective biomechanical properties. This study thus aims to investigate the relationship between the microstructure and the macroscopic mechanical response of typical connective tissues: skin (disordered) and cornea (highly ordered). To overcome technical issues in providing experimental multiscale data in intact tissues, we have implemented an original setup combining mechanical assays at tissue scale and Second Harmonic Generation (SHG) imaging of collagen reorganization. This multiphoton imaging modality represents an effective structural probe of the micrometer-scale collagen organization in unstained tissues. 3D SHG images were acquired in dermis from ex vivo murine skin biopsies during controlled stretching until rupture, and in ex vivo Human cornea during inflation assays. Specific image processing was implemented to quantify the reorganization of tissue microstructure and correlate it with stress/stretch relationship at macroscopic scale. In murine skin, we showed that the collagen fibers continuously aligned with stretch, generating the observed increase in mechanical stress, which challenges the usual theoretical explanation of the microstructural origin of the skin macroscopic mechanical response. Moreover, dermis from transgenic mice with defective collagen microstructure exhibited altered collagen reorganization upon traction, which could be linked to the microstructural modifications. In Human cornea, our results showed no reorganization of the collagen fibrils at sub-micrometer scale within the stromal lamellae, while the lamellae at micrometer scale reorganized in a more balanced way along 2 main perpendicular directions, in line with the deformation observed at the macroscopic scale. In conclusion, our approach provides an efficient tool to investigate the biomechanics of collagen-rich tissues in normal and pathological context and to guide tissue engineering with appropriate biomechanical responses.

The purpose of the study was in vivo assessment of the vulvar blood and lymphatic vessels characteristics in norm and lichen sclerosus by multimodal optical coherence tomography (OCT). The study was performed using a multimodal OCT setup developed at the Institute of Applied Physics RAS (Russia). OCT angiography and OCT lymphangiography are based on the analysis of speckle structure. Visualization of blood and lymph vessels does not require the use of exogenous contrast agents. A histological study of vulvar biopsy samples from two points was carried out for 10 patients without vulvar pathology and 12 patients with lichen sclerosus. 3D OCT data was obtained from six vulvar points in each patient. OCT images were verified by histological examination. It was shown that normally the vulvar mucosa has good blood supply and a well-developed network of lymphatic vessels. In the case of lichen sclerosus, the density of blood vessels in the area of hyalinosis significantly reduced and amounted to 2.5 ± 0.79% versus 3.9 ± 0.23% in norm (p = 0.0003). OCT lymphangiographic images also show a significant decrease in the number of lymphatic vessels, their density was 1.7 ± 0.75% versus 3.7 ± 0.54% in norm (p = 0.02). A direct relationship between the state of the blood and lymph vessels from the condition of the connective tissue of the vulva was shown. The number of blood and lymph vessels is sharply reduced in the area of hyalinosis and sclerosis of collagen fibers.

The combination of molecular (hyperspectral imaging) and morphological (optical coherent tomography imaging) optical technologies helps in the assessment of biological tissue both in pathological diagnosis and in the follow-up treatments. The co-registration of both imaging features allows quantifying the presence of chromophores and the subsurface structure of tissue. This work proposes the fusion of two optical imaging technologies for the characterization of different types of tissues where the attenuation coefficient calculated from OCT imaging serves to track the presence of anomalies in the distribution of chromophores over the sample and therefore to diagnose pathological conditions. The performance of two customized hyperspectral imaging systems working in two complementary spectral ranges (VisNIR from 400 to 1000 nm, and SWIR 1000 to 1700 nm) and one commercial OCT system working at 1325 nm reveals the presence of fibrosis, collagen alterations and lipid content in cardiovascular tissues such as aortic walls (to assess on aneurysmal conditions) or tendinous chords (to diagnose the integrity of the valvular system) or in muscular diseases prone to fibrotic changes and inflammation.

With the rise of antibiotic resistance, phage therapy is seen as a promising alternative to cure infection to multiresistant bacteria strains. However, phage susceptibility tests currently carried out are time-consuming and are not compliant with the automated environment of hospital laboratories. In this work, we present a method for phage susceptibility testing through optical density measurement with the use of a lensless imaging technique. Fluid assays containing bacteria and phages are loaded in the wells of a 5mm-thick custom-made microfluidic card. The card is put on a 3.3 cm2 CMOS imaging sensor taken from a CANON dslr camera. It is illuminated by a screen paired with a 560 nm spectral filter to provide a homogeneous monochromatic lighting over the whole sensor area. Thanks to the large imaging area of the CMOS sensor, it is possible to simultaneously monitor the level of light transmitted through the well of the microfluidic card and hence to compute the optical density of a dozen sample without the need of mechanical elements. We thus monitor the decay or increase of optical density to determine respectively the lysis or growth of the bacteria under test. This method provides a reliable result of optical phage susceptibility testing in less than 4 hours. The prototype shown here is compact, inexpensive (<1 k€) and is compliant with automated environment of hospital laboratories. Moreover, it is versatile and can be used for other application such as lysis plaque imaging to provide a fast measurement of a viral titer of a bacteriophage suspension.

The Laser-Induced Breakdown Spectroscopy (LIBS) technique has been utilized in several studies for the identification of pathogenic bacterial strains based on their characteristic spectral fingerprint. Currently used LIBS techniques for discrimination of bacterial strains belonging to the same species require sophisticated methodology and expensive instrumentation. In this study, we present the strategies adopted to achieve this goal using a low-cost LIBS methodology. Time-resolved LIBS experiments were carried out using a nanosecond pulsed Nd: YAG laser (1064 nm, 6 ns, 10 Hz) for ablating the bacterial samples, and the resulting emission spectra were recorded using a portable and non-intensified Charge Coupled Device (CCD) detector. The bacterial strains used for this study were two clinical isolates of Escherichia coli (E. coli) - a pathogen causing severe infections in humans. The individual bacterial strains were cultured using a standard optimized protocol, and their respective chemical fingerprints were captured using the LIBS technique. We also investigate the efficacy of standardizing the growth environment and its role in modulating the chemical composition of the bacterial strains. A bacterial growth study was performed to assess the influence of regulating growth environment (concentration of sodium and potassium in the nutrient media), on the growth phases of the two bacterial strains. The spectral lines corresponding to sodium (589.5 nm), and potassium (766.5 nm, 769.9 nm) were found to be significant among the characteristic LIBS emission of the bacterial strains. The sodium to potassium ratio (Na/K), calculated from the elemental line intensities in the LIBS spectrum of bacteria, was found to be a highly significant feature for discrimination of bacterial strains with a classification accuracy >90%.

Personal wearable medical devices and sensors have become more popular which results in better resources optimization and a better systematic monitoring or health condition. The ability to extract quantitative information in living cells without actual biopsy (disturbance in the structure of the cell) can be used to study and monitor morphological and physiological changes such as pre-cancerous or cancerous conditions. Basically, light scattered from a cell depends on the size of the cell, structure of the cell and the properties of the incident light. It is also termed as “optical biopsy”. Such Mie scattering based Lab-on-a-Chip (LOC) sensor device if implemented can be used for early diagnosis of terminal diseases such as cancer. Gold nano particle acts as a sensor on this wearable device. Mie scattering based nonlinear optical phenomenon is used for cancer detection. This work involves modelling, simulation and analysis of a nano sensor using Discontinuous Gallerkin Time Domain method (DGTD). Light is used from arbitrary shaped objects at various incident angles. The Mie solution to Maxwell’s equations describes the scattering of an electromagnetic plane wave by a homogeneous sphere. Mie scattering theory has been used to determine whether scattered light from tissue corresponds to healthy or cancerous cell nuclei using angle resolved low coherence interferometry. Gold nanoparticle has been used in biological applications and it is necessary to know the co-efficient of scattering and co-efficient of absorption. Mie scattering has no upper limits with respect to size of the particle.

The most common type of cancer are carcinomas: cancers which originate from the epithelial tissue lining the outer surfaces of organs. To detect carcinomas at an early stage, techniques are required with small sampling volumes. Single Fiber Reflectance spectroscopy (SFR) is a promising technique to detect early-stage carcinomas since it has a measurement volume in the order of hundreds of microns. SFR uses a single fiber to emit and collect broadband light. The model from Kanick et al. to relate SFR reflectance to tissue optical properties provided accurate results for tissue with a modified Henyey Greenstein phase function only. However, in many tissues other types of phase functions have been measured. We have developed a new model to relate SFR reflectance to the scattering and absorption properties of tissue, which provides accurate results for a large range of tissue phase functions. SFR measurements fall into the sub-diffuse regime. We, therefore, describe the SFR reflectance as a diffuse plus a semi-ballistic component. An accurate description of the diffuse SFR signal requires double integration of spatially resolved reflectance over the fiber surface. We use approaches from Geometric Probability and have derived the first analytic solution for the diffuse contribution to SFR. For the semi-ballistic contribution to the SFR signal we introduce a new phase function dependent parameter, psb, to describe the semi-ballistic part of the SFR signal. We will use the model to derive optical properties from SFR measurements performed endoscopically in patients with Barrett’s esophagus. These patients are at an increased risk to develop esophageal adenocarcinoma and, therefore, undergo regular endoscopic surveillance. When detected at an early stage, endoscopic treatment is possible, thereby avoiding extensive surgery. Based on the concept of field cancerization, we investigated whether SFR could be used as a tool to identify which patients are developing esophageal adenocarcinoma.

Near-Infrared Spectroscopy (NIRS) instruments most often measures the reflected NIR attenuation at a couple of wavelengths, to quantify the concentration of the oxygenated and deoxygenated hemoglobin ([HbO2], [HHb]) and provide information about the brain oxygen levels. Of particular interest are the changes in brain oxygenation due to neuronal activity as they can provide us with an indirect measurement of brain function. For several years now we have been developing technology that extend NIRS instrumentation by allowing measuring hundreds of NIR wavelengths. The technique is called broadband near-infrared spectroscopy (or bNIRS). The bNIRS system measures changes in light attenuation, reflected back from the head, over 308 near-infrared (NIR) wavelengths (610nm-918nm). This allow us to quantify the changes in brain tissue [HbO2], [HHb] and the concentration changes in the oxidation state of cerebral cytochrome-c-oxidase ([oxCCO]). CCO is an enzyme in the electron transport chain of the mitochondrial catalyzing more than 95% of oxygen to produce ATP. In my talk I will discuss how we have been using this technology both in our preclinical and clinical investigations in perinatal hypoxic-ischemia. In particular, I will introduce our clinical study in newborns with Hypoxic-Ischemic (HI) injury undergoing therapeutic hypothermia. In 24 neonates, 54 episodes of spontaneous decreases in peripheral oxygen saturation (desaturations) were recorded between 6h and 81h after birth. As determined by magnetic resonance spectroscopy derived lactate/N-acetyl-aspartate (MRS-measured Lac/NAA) 8 newborns had unfavorable and 16 newborns had favorable outcomes after HI. We demonstrated that a strong relationship between cerebral metabolism (bNIRS-measured oxCCO) and oxygenation was associated with unfavorable outcome; this is likely to be due to a lower cerebral metabolic rate and mitochondrial dysfunction in severe encephalopathy. In addition, uniquely we have obtained these results using a non-invasive bedside technique, bNIRS, in the first 4 days of life.

Background. The aim of our study was to estimate the association of lipid fractions and skin autofluorescence (AF), measured with experimental sample of fiber optic spectrometer FOS-1, in patients with diabetes mellitus (DM). Methods. We included in the study 44 type 1 DM patients (median age 29.5 (24.0; 36.0) and 23 type 2 DM patients, median age 58.0 (47.0; 61.0). Blood lipids and glycosylated hemoglobin (HbA1c) were analyzed. Skin AF was measured with fluorescence–reflection spectrometer FOS-1 at a wavelength of 460 nm with excitation of 365 nm. Ratio of fluorescent signal to signal of reflection in excitation region was used as measured parameter to reduce the effect of skin pigmentation. The measurements were carried out at 5 points on the skin area of forearms. Results. Type 1 DM group was divided to 13 dyslipidemic and 31 normolipidemic patients, subgroups were comparable by age, DM duration and HbA1c level. AF level in dyslipidemic and normolipidemic patients was 0.95 (0.82; 1.16) arb. units and 0.86 (0.79; 0.91) arb. units respectively, p=0.016. In type 1 DM group we found significant positive correlation between skin AF and total cholesterol level (R=0.45, р<0.05) and triglycerides level (R=0.57, p<0.05). In type 2 DM group no lipid fraction showed significant positive correlation with skin AF level. Conclusion. The spectrometer FOS-1 can be used for additional risk evaluation in young and middle-aged type 1 DM patients, but simultaneous presence of several risk factors in older group of type 2 DM patients obstruct the use of the method.

Raman spectroscopy is an optical technique that can assess a sample’s molecular content by probing its vibrational modes and has been used over the last decades to diagnose multiple types of cancer. The standard method used to build the classification models, based on machine learning algorithms, is the source of two majors limitations: the small size of the collected training datasets and the issue of portability of statistical models across imaging systems and medical centers. Model portability can be adressed by using a spectrum processing method that totally removes the hardware influence from the processed Raman measurements. We focus here on the results of two experiments conducted to evaluate the reproductibility of Raman measurements made with nine different point-probe systems. For the first experiment, we used a nylon phantom to assess inter-systems differences and applied the data processing method which lowered the inter-systems deviation for the processed nylon peaks under 3%. Furthermore, system #1 was used in vivo in a human brain surgery to acquire 15 Raman measurements from normal and tumor tissue. We evaluated the deviation between classes and found that it was superior to the 3% inter-systems reproductibility for 10 Raman peaks associated with proteins, lipids and nucleic acids. The second experiment was done with the system #1 as a master system and systems #2 to #9 as slave systems. The master system was used to build a Support Vector Machine classification model to discriminate white matter from grey matter on fixed ex vivo monkey brain slices. The model was exported from master to slaves performing a diagnosis accuracy consistently over 95%. The reported results indicate the possibility to succesfully export statistical model from one system to another and to greatly increase the size of dataset using multiple imaging systems.

Laser ablation (LA) is a minimally invasive procedure based on light/tissue interaction aimed to induce a controlled tumor necrosis by increasing tissue temperature. Given the relationship between tissue damage and produced heat, LA needs a fine control of evolving thermal effects in order to evaluate and control procedure outcome. This study relies on biomedical optics principles for non-invasive diagnostic tools development, and presents a contactless approach based on hyperspectral imaging (HSI) to monitor thermal damage during in vivo porcine LA. By collecting relative pixel-by-pixel reflectance/absorbance of a wide range spectrum (500-1000nm), HSI can track molecular structure modifications caused by the thermoablative procedure. Indeed, these modifications alter tissue light scattering and absorption. In order to investigate tissue spectrum change by increasing temperature, HSI was collected at fixed maximum temperatures (37, 60, 70, 80, 90, 100, 110 °C) and immediately after LA (1, 2, 3, 4, and 5 minutes). Tissue spectral response for two tests was analyzed also relying on the ablated area considered. Regions of Interest of different dimensions (16, 77, and 170 pixels) were placed in the images after applying a motion correction. Obtained spectra show noticeable variations once a specific temperature threshold has been reached (80-100 °C). Specifically, the measured absorbance variation for selected wavelengths (630, 760, 960nm, for methemoglobin, deoxyhemoglobin, and water respectively) confirms tissue optical behavior dependence with its thermal state. This preliminary investigation discloses the potential of HSI measurement to characterize LA damage, encouraging future studies to standardize this novel technique.

With an adequate tissue dataset, supervised classification of tissue optical properties can be achieved in SFDI images of breast cancer lumpectomies with deep convolutional networks. Nevertheless, the use of a black-box classifier in current ex vivo setups provides output diagnostic images that are inevitably bound to show misclassified areas due to inter- and intra-patient variability that could potentially be misinterpreted in a real clinical setting. This work proposes the use of a novel architecture, the self-introspective classifier, where part of the model is dedicated to estimating its own expected classification error. The model can be used to generate metrics of self-confidence for a given classification problem, which can then be employed to show how much the network is familiar with the new incoming data. A heterogenous ensemble of four deep convolutional models with self-confidence, each sensitive to a different spatial scale of features, is tested on a cohort of 70 specimens, achieving a global leave-one-out cross-validation accuracy of up to 81%, while being able to explain where in the output classification image the system is most confident.

Fluorescent dyes that emit in the near-infrared (NIR) region gained increasing attention in medical imaging and fluorescence guided surgery. [1,2] Optical imaging allows the surgeons to visualize the normal and diseased tissues in real-time.[2] Particularly interesting in this respect is NIR fluorescent coatings on medical devices such as implants, catheters and stents etc., which are extensively used in various treatments and surgical procedures. However, most of the existing coating materials are composed of either indocyanine green (ICG) or other visible/blue light active fluorophores. ICG-based coatings suffer from limited brightness and poor photostability whereas blue/visible light active coatings are not compatible with well-established laparoscopic systems. [3] Here, we introduce new NIR coating material composed of cyanine dye salts and a biocompatible polymer. [4] The dye is encapsulated into the polymer coating with help of bulky counterions, which, according to our studies on polymeric nanoparticles, [5] prevents dye leakage and minimizes self-quenching. A series of dyes based on cyanine 7.5 family were synthesized and tested. In comparison to ICG-based coatings, our new NIR coatings exhibit superior brightness and photostability Moreover, these coating materials showed good adhesion properties and useful to coat medical devices, including metallic fiducials and catheters. Such coated medical devices show high fluorescence brightness and exceptional stability in air and buffer medium for at least 150 days. Moreover, our coatings are biocompatible and thus, appear promising for biomedical use. We have demonstrated the various possible medical applications of the NIR coated devices (fiducials and catheters) in ex-vivo and in-vivo porcine models. Funding: This work was supported by “NICE” grant of SATT Conectus, Alsace. References: [1] A. H. Ashoka, et. al. J Phys Chem Lett, 2019, 10, 2414. [2] S. Hernot et. al. Lancet Oncol, 2019, 20, e354. [3] Y. Choi et. al. Surg Endosc 2011, 25, 2372. [4] European patent application filed. [5] Reisch, et al., Nature Commun., 2014, 5, 4089.

Near infrared photoimmunotherapy (NIR-PIT) is a recently developed cancer-targeted theranostic technology that produces a therapeutic immune response. Conventional immunotherapies, such as immune-activating cytokine therapy, checkpoint inhibition, engineered T cells (e.g. chimeric antigen receptor or CAR-T cells) and suppressor cell depletion do not directly destroy cancer cells, but rely exclusively on activating the immune system. NIR-PIT not only selectively destroys cancer cells but also activates anti-cancer immune reactions that kill cells that escape direct killing. NIR-PIT can be applied to a wide variety of cancers either as monotherapy or in combination with other immune therapies to further activate anti-cancer immunity. The first NIR-PIT Phase 3 clinical trial targeting EGFR uses the antibody-photoabsorber conjugate (APC), cetuximab-IR700 (RM1929/ASP1929) and is now underway for recurrent and advanced head and neck cancer in North America, Asia and Europe (https://clinicaltrials.gov/ct2/show/NCT03769506). NIR-PIT has been given fast track status by regulators in the US and Japan. A variety of imaging methods can be used to monitor NIR-PIT therapy including optical methods such as direct IR700 fluorescence imaging before and during therapy with fluorescent endoscopy systems or a special camera using the therapeutic light as excitation. Other optical, PET and MR methods may be useful as well. NIR-PIT is a versatile method of treating cancers and will likely acquire specific roles for treating various cancers particularly those presenting with localized or locally advanced disease even with distant metastasis.

Urothelial carcinoma (UC) is the most common type of bladder cancer, and its treatment depends from both tumour invasiveness (stage) and aggressiveness (grade). The gold standard for detecting UC is white-light cystoscopy, followed by tissue biopsy and histopathological examination; however, such process is invasive, time-consuming, operatordependent and prone to sampling errors. In this framework, optical spectroscopy techniques could be a promising solution for fast and label-free diagnosis of bladder tissues and for early detection of UC. Thus, we combined autofluorescence, diffuse reflectance and Raman spectroscopy in a compact and transportable setup based on an optical fibrebundle probe. This experimental setup was used for studying fresh biopsies of urothelial tumour (140 samples) and healthy bladder (50 samples) collected from 90 patients undergoing Transurethral Resection of Bladder Tumours (TURBT). The aim of this study was to develop an automated classification of the examined tissues based on the intrinsic spectral information provided by all three techniques. We found that healthy and diseased tissues showed significant spectral differences for each technique, resulting in high accuracy (up to 90%) from a Linear Discriminant Analysis (LDA) routine. In particular, fluorescence spectroscopy – excited either with blue or UV light – provided very good results in detecting UC. However, tumour grading and staging proved to be more challenging tasks, for which no single spectroscopic technique could provide sufficient sensitivity and specificity. Therefore, we found that a multimodal approach can improve significantly the diagnosis of UC stages and grades.

Recently, in vivo fluorescence imaging using indocyanine green (ICG) has actively been applied to hepatobiliary and pancreatic surgery in clinical settings. 1) Fluorescence cholangiography: fluorescence images of the extrahepatic bile ducts can be obtained by intrabiliary injection of ICG solution (0.025 mg/mL) or preoperative intravenous injection of ICG (2.5 mg). The latter technique begins to be used worldwide for confirmation of the bile duct anatomy during minimally-invasive cholecystectomy. 2) Identification of hepatic tumors: Following preoperative intravenous injection of ICG (0.5 mg/kg), it can accumulate in hepatocellular carcinoma tissues and in non-cancerous hepatic parenchyma surrounding liver metastasis, enabling intraoperative identification of subcapsular hepatic tumors by fluorescence imaging. 3) Hepatic segmentation: ICG solution (0.25 mg/5 mL) is injected into a tumor-bearing portal branch under ultrasound guidance (positive staining). ICG can also be administered intravenously following closure of a corresponding portal pedicle (negative staining). These techniques enables delineation of hepatic segmental boundaries throughout surgical procedures. 4) Assessment of blood perfusion: Fluorescence imaging following intraoperative bolus injection of ICG (2.5mg) visualizes arterial/portal blood flows and perfusion in the surrounding organs during surgeries requiring resection/reconstruction of the major vessels. for intraoperative visualization of biological structures and perfusion assessment. In addition, we have developed a novel fluorophore (glutaryl-phenylalanine hydroxymethyl rhodamine green) activated by pancreatic chymotrypsin for real-time identification of pancreatic juice leakage.

Intraoperative guidance using targeted near-infrared (NIR) fluorescent tracers can provide surgeons with real-time feedback on the presence of residual tumour tissue. To overcome the still limited depth penetration of NIR light, and limit potentially missing residual occult or deeper lying lesions, the combination of fluorescence with nuclear imaging is proposed. We describe the design and preclinical validation of the anti-HER2 nanobody 2Rs15d, conjugated with a ‘multifunctional single attachment point’ (MSAP), which integrates a Cy5 fluorophore and diethylenetriaminepentaacetic acid (DTPA) chelator into a single label. After random conjugation to primary amines in the nanobody, functionality of the tracer and stability after 111In labelling were evaluated in vitro. Using SKOV3 (HER2+) and MDA-MB-435S (HER2-) xenografted mice, the in vivo biodistribution of 2Rs15d-MSAP.111In was determined by SPECT/CT (1h post-injection) and fluorescence imaging (1h30 post-injection). Ensuing, the ex vivo biodistribution was determined 2h (both xenograft models) and 24h post-injection (SKOV3 only). The tracer retained its affinity after conjugation of the MSAP and remained stable over 24h in both PBS and human serum after 111In labelling. The in vivo SPECT/CT and fluorescence images corresponded well, showing the expected biodistribution pattern for nanobody tracers, meaning low background except for high renal uptake due to clearance, and specific tumour uptake in HER2-overexpressing tumours. Ex vivo biodistribution data revealed a SKOV3 tumour-specific uptake of 7.0 ± 2.5 %ID/g after 2h, significantly higher than 1.1 ± 1.2 %ID/g for control tumours. The tumour-to-blood ratio was 47.6± 25.4, tumour-to-muscle ratio 23.2 ± 11.6, and tumour-to-liver ratio 6.9 ± 3.7. After 24h SKOV3 tumour uptake was 5.6 ± 1.9 %ID/g, tumour-to-blood ratio 229.1 ± 85.1, tumour-to-muscle ratio 16.8 ± 8.0, and tumour-to-liver ratio 5.1 ± 1.9. In conclusion, functional bimodal nuclear/fluorescent nanobody-tracers can be conveniently generated by conjugation of a single-molecule MSAP-reagent carrying both fluorophore and a chelator.

Post-operative pancreatic fistula is the most dreadful complication of pancreatic resections. Tissue perfusion is a recognized risk factor of fistula in gastrointestinal anastomosis, but has been poorly studied in case of pancreatic surgery. Organ perfusion can be estimated by several intraoperative optical imaging modalities, including exogenous fluorescence imaging with Indocyanine Green (ICG). A limitation of ICG fluorescence angiography is the lack of a quantitative analytic method. Our group has developed and validated a real-time computational imaging analysis of tissue perfusion, based on the slope of the time of fluorescence peak, defined fluorescence-based enhance reality (FLER). This consists in encoding tissue hemodynamics into a virtual perfusion cartography that is superimposed on intraoperative images in real time, in order to provide the perfusion information to surgeons. Another emerging technology is hyperspectral imaging (HSI), which combines a spectrometer and a camera. HSI obtains spectral tissue curves pixel by pixel in a wide wavelength range and can provide absolute values of concentration and/or oxygen saturation of hemoglobin. The aim of this study was to investigate the efficacy of both imaging modalities, FLER and HSI, to estimate pancreatic perfusion. Methods Twelve pigs were involved and randomly assigned to interventional group (n=6) and a control group (n =6). After a median laparotomy the pancreas was fully exposed. In the interventional group, under radiographic guidance, a segmental ischemia of the pancreas was induced by means of coil embolization in small splenic arterial branches, via a femoral artery approach. No ischemia was induced in the control group. HSI images were obtained using the TIVITA camera (Diaspective Vision, Germany). FLER was obtained after injecting 0.2 mg/kg of ICG and image processing using a dedicated software (ER PERFUSION, IRCAD, France). The augmented reality of color-coded images derived from both FLER and HSI were overlapped on the video image. Local capillary lactates (LCL) were sampled in different regions of interests (ROIs) of the pancreas. LCL were correlated to the absolute values provided by the imaging analyses (slope of time-to-peak and StO2, respectively for fluorescence and HSI). Results In all pigs from the interventional group, the segmental ischemic areas were successfully created. The mean slope was slower; the mean StO2 was lower; and mean LCL were lower in the ischemic zone than in the transition and vital zones. The scatter plot between the slope and StO2, the slope and LCL, and StO2 and LCL in all 12 pigs showed statistically significant correlation. In addition, LCL can be predicted from FLER and StO2. The prediction from HSI StO2 (median error: 0.67 mmol/L) is significantly more accurate than FLER (median error: 0.8 mmol/L) (P=0.02). Conclusions FLER and HSI, enable both to precisely quantify and visualize real-time perfusion of the pancreas in this porcine model of pancreas ischemia. HSI showed a greater accuracy in predicting the value of capillary local lactates when compared to quantitative exogenous fluorescence analysis.