Confocal Laser Endomicroscopy Images Research Articles

Intraoperative in vivo confocal endomicroscopy of the glioma margin: performance assessment of image interpretation by neurosurgeon users.

Confocal laser endomicroscopy (CLE) is an intraoperative real-time cellular resolution imaging technology that images brain tumor histoarchitecture. Previously, we demonstrated that CLE images may be interpreted by neuropathologists to determine the presence of tumor infiltration at glioma margins. In this study, we assessed neurosurgeons' ability to interpret CLE images from glioma margins and compared their assessments to those of neuropathologists. In vivo CLE images acquired at the glioma margins that were previously reviewed by CLE-experienced neuropathologists were interpreted by four CLE-experienced neurosurgeons. A numerical scoring system from 0 to 5 and a dichotomous scoring system based on pathological features were used. Scores from assessments of hematoxylin and eosin (H&E)-stained sections and CLE images by neuropathologists from a previous study were used for comparison. Neurosurgeons' scores were compared to the H&E findings. The inter-rater agreement and diagnostic performance based on neurosurgeons' scores were calculated. The concordance between dichotomous and numerical scores was determined. In all, 4275 images from 56 glioma margin regions of interest (ROIs) were included in the analysis. With the numerical scoring system, the inter-rater agreement for neurosurgeons interpreting CLE images was moderate for all ROIs (mean agreement, 61%), which was significantly better than the inter-rater agreement for the neuropathologists (mean agreement, 48%) (p < 0.01). The inter-rater agreement for neurosurgeons using the dichotomous scoring system was 83%. The concordance between the numerical and dichotomous scoring systems was 93%. The overall sensitivity, specificity, positive predictive value, and negative predictive value were 78%, 32%, 62%, and 50%, respectively, using the numerical scoring system and 80%, 27%, 61%, and 48%, respectively, using the dichotomous scoring system. No statistically significant differences in diagnostic performance were found between the neurosurgeons and neuropathologists. Neurosurgeons' performance in interpreting CLE images was comparable to that of neuropathologists. These results suggest that CLE could be used as an intraoperative guidance tool with neurosurgeons interpreting the images with or without assistance of the neuropathologists. The dichotomous scoring system is robust yet simple and may streamline rapid, simultaneous interpretation of CLE images during imaging.

Clustering-Guided Twin Contrastive Learning for Endomicroscopy Image Classification.

Learning better representations is essential in medical image analysis for computer-aided diagnosis. However, learning discriminative semantic features is a major challenge due to the lack of large-scale well-annotated datasets. Thus, how can we learn a well-structured categorizable embedding space in limited-scale and unlabeled datasets? In this paper, we proposed a novel clustering-guided twin-contrastive learning framework (CTCL) that learns the discriminative representations of probe-based confocal laser endomicroscopy (pCLE) images for gastrointestinal (GI) tumor classification. Compared with traditional contrastive learning, in which only two randomly augmented views of the same instance are considered, the proposed CTCL aligns more semantically related and class-consistent samples by clustering, which improved intra-class tightness and inter-class variability to produce more informative representations. Furthermore, based on the inherent properties of CLE (geometric invariance and intrinsic noise), we proposed to regard CLE images with any angle rotation and CLE images with different noises as the same instance, respectively, for increased variability and diversity of samples. By optimizing CTCL in an end-to-end expectation-maximization framework, comprehensive experimental results demonstrated that CTCL-based visual representations achieved competitive performance on each downstream task as well as more robustness and transferability compared with existing state-of-the-art SSL and supervised methods. Notably, CTCL achieved 75.60%/78.45% and 64.12%/77.37% top-1 accuracy on the linear evaluation protocol and few-shot classification downstream tasks, respectively, which outperformed the previous best results by 1.27%/1.63% and 0.5%/3%, respectively. The proposed method holds great potential to assist pathologists in achieving an automated, fast, and high-precision diagnosis of GI tumors and accurately determining different stages of tumor development based on CLE images.

Diagnosis of malignancy in oropharyngeal confocal laser endomicroscopy using GPT 4.0 with vision.

Confocal Laser Endomicroscopy (CLE) is an imaging tool, that has demonstrated potential for intraoperative, real-time, non-invasive, microscopical assessment of surgical margins of oropharyngeal squamous cell carcinoma (OPSCC). However, interpreting CLE images remains challenging. This study investigates the application of OpenAI's Generative Pretrained Transformer (GPT) 4.0 with Vision capabilities for automated classification of CLE images in OPSCC. CLE Images of histological confirmed SCC or healthy mucosa from a database of 12 809 CLE images from 5 patients with OPSCC were retrieved and anonymized. Using a training data set of 16 images, a validation set of 139 images, comprising SCC (83 images, 59.7%) and healthy normal mucosa (56 images, 40.3%) was classified using the application programming interface (API) of GPT4.0. The same set of images was also classified by CLE experts (two surgeons and one pathologist), who were blinded to the histology. Diagnostic metrics, the reliability of GPT and inter-rater reliability were assessed. Overall accuracy of the GPT model was 71.2%, the intra-rater agreement was κ = 0.837, indicating an almost perfect agreement across the three runs of GPT-generated results. Human experts achieved an accuracy of 88.5% with a substantial level of agreement (κ = 0.773). Though limited to a specific clinical framework, patient and image set, this study sheds light on some previously unexplored diagnostic capabilities of large language models using few-shot prompting. It suggests the model`s ability to extrapolate information and classify CLE images with minimal example data. Whether future versions of the model can achieve clinically relevant diagnostic accuracy, especially in uncurated data sets, remains to be investigated.

Intraoperative in vivo confocal laser endomicroscopy imaging at glioma margins: can we detect tumor infiltration?

Confocal laser endomicroscopy (CLE) is a US Food and Drug Administration-cleared intraoperative real-time fluorescence-based cellular resolution imaging technology that has been shown to image brain tumor histoarchitecture rapidly in vivo during neuro-oncological surgical procedures. An important goal for successful intraoperative implementation is in vivo use at the margins of infiltrating gliomas. However, CLE use at glioma margins has not been well studied. Matching in vivo CLE images and tissue biopsies acquired at glioma margin regions of interest (ROIs) were collected from 2 institutions. All images were reviewed by 4 neuropathologists experienced in CLE. A scoring system based on the pathological features was implemented to score CLE and H&E images from each ROI on a scale from 0 to 5. Based on the H&E scores, all ROIs were divided into a low tumor probability (LTP) group (scores 0-2) and a high tumor probability (HTP) group (scores 3-5). The concordance between CLE and H&E scores regarding tumor probability was determined. The intraclass correlation coefficient (ICC) and diagnostic performance were calculated. Fifty-six glioma margin ROIs were included for analysis. Interrater reliability of the scoring system was excellent when used for H&E images (ICC [95% CI] 0.91 [0.86-0.94]) and moderate when used for CLE images (ICC [95% CI] 0.69 [0.40-0.83]). The ICCs (95% CIs) of the LTP group (0.68 [0.40-0.83]) and HTP group (0.68 [0.39-0.83]) did not differ significantly. The concordance between CLE and H&E scores was 61.6%. The sensitivity and specificity values of the scoring system were 79% and 37%. The positive predictive value (PPV) and negative predictive value were 65% and 53%, respectively. Concordance, sensitivity, and PPV were greater in the HTP group than in the LTP group. Specificity was higher in the newly diagnosed group than in the recurrent group. CLE may detect tumor infiltration at glioma margins. However, it is not currently dependable, especially in scenarios where low probability of tumor infiltration is expected. The proposed scoring system has excellent intrinsic interrater reliability, but its interrater reliability is only moderate when used with CLE images. These results suggest that this technology requires further exploration as a method for consistent actionable intraoperative guidance with high dependability across the range of tumor margin scenarios. Specific-binding and/or tumor-specific fluorophores, a CLE image atlas, and a consensus guideline for image interpretation may help with the translational utility of CLE.

In vivo evaluation of liver function by multimodal imaging in an alcohol-induced liver injury model.

Visually evaluating liver function is a hot topic in hepatology research. There are few reliable and practical visualization methods for evaluating the liver function in vivo in experimental studies. In this study, we established a multimodal imaging approach for in vivo liver function evaluation and compared healthy mice with chronic alcoholic liver injury (cALI) model mice to explore its potential applicability in experimental research. In vivo fluorescence imaging (IVFI) technology was utilized to visually represent the clearance of indocyanine green from the liver of both healthy mice and mice with cALI. The reserve liver function was evaluated via IVFI using the Cy5.5-galactosylated polylysine probe, which targets the asialoglycoprotein receptor of hepatocytes. Hepatic microcirculation was assessed through laser speckle perfusion imaging of hepatic blood perfusion. The liver microstructure was then investigated by in vivo confocal laser endomicroscopy imaging. Finally, hepatic asialoglycoprotein receptor expression, histology, and the levels of serum alanine aminotransferase and aspartate aminotransferase were measured. In vivo multimodal imaging results intuitively and dynamically showed that indocyanine green clearance [mean ± standard deviation (SD): 30.83±14.71, 95% confidence interval (CI): 20.3 to 41.35], the fluorescence signal intensity (mean ± SD: 1,217.92±117.63; 95% CI: 1,148.38 to 1,290.84) and fluorescence aggregation area (mean ± SD: 5,855.80±1,271.81; 95% CI: 5,051.57 to 6,653.88) of Cy5.5-galactosylated polylysine targeting the asialoglycoprotein receptor, and hepatic blood perfusion (mean ± SD: 1,494.86±299.33; 95% CI: 1,316.98 to 1,690.16) in model mice were significantly lower than those in healthy mice (all P<0.001). Compared to healthy mice, the model mice exhibited a significant decline in liver asialoglycoprotein receptor expression (mean ± SD: 219.03±16.34; 95% CI: 208.97 to 230.69; P<0.001), increased serum alanine aminotransferase (mean ± SD: 149.70±47.89 U/L; 95% CI: 81.75 to 128.89; P=0.01) and aspartate aminotransferase levels (mean ± SD: 106.30±36.13 U/L; 95% CI: 122.01 to 180.17; P=0.021), hepatocyte swelling and deformation, disappearance of the hepatic cord structure, partial necrosis, and disintegration of hepatocytes. The imaging features of fluorescence signals in liver regions, hepatic blood perfusion and microstructure were biologically related to hepatic asialoglycoprotein receptor expression, serum indices of liver function, and histopathology in model mice. Utilizing in vivo multimodal imaging technology to assess liver function is a viable approach for experimental research, providing dynamic and intuitive visual evaluations in a rapid manner.

Context-aware dynamic filtering network for confocal laser endomicroscopy image denoising

Objective. As an emerging diagnosis technology for gastrointestinal diseases, confocal laser endomicroscopy (CLE) is limited by the physical structure of the fiber bundle, leading to the inevitable production of various forms of noise during the imaging process. However, existing denoising methods based on hand-crafted features inefficiently deal with realistic noise in CLE images. To alleviate this challenge, we proposed context-aware kernel estimation and multi-scale dynamic fusion modules to remove realistic noise in CLE images, including multiplicative and additive white noise. Approach. Specifically, a realistic noise statistics model with random noise specific to CLE data is constructed and further used to develop a self-supervised denoised model without the participation of clean images. Secondly, context-aware kernel estimation, which improves the representation of features by similar learnable region weights, addresses the problem of the non-uniform distribution of noises in CLE images and proposes a lightweight denoised model (CLENet). Thirdly, we have developed a multi-scale dynamic fusion module that decouples and recalibrates features, providing a precise and contextually enriched representation of features. Finally, we integrated two developed modules into a U-shaped backbone to build an efficient denoising network named U-CLENet. Main Results. Both proposed methods achieve comparable or better performance with low computational complexity on two gastrointestinal disease CLE image datasets using the same training benchmark. Significance. The proposed approaches improve the visual quality of unclear CLE images for various stages of tumor development, helping to reduce the rate of misdiagnosis in clinical decision-making and achieve computer graphics-assisted diagnosis.

Intraoperative confocal laser endomicroscopy for interpretation of a sellar hemangioblastoma: illustrative case.

Intraoperative frozen sections play a critical role in surgical strategy because of their ability to provide rapid histopathological information. In cases in which intraoperative biopsy carries a significant risk of bleeding, intraoperative confocal laser endomicroscopy (CLE) can assist in decision-making. The authors present a rare case of a large sellar hemangioblastoma. Preoperative radiographic imaging and normal pituitary function suggested a differential diagnosis that included hemangioblastoma. The patient underwent partial preoperative embolization and a right-sided pterional craniotomy for resection of the lesion. Gross intraoperative examination revealed a highly vascular sellar lesion requiring circumferential dissection to minimize blood loss. The serious vascularity precluded intraoperative frozen section analysis, and CLE imaging was performed. CLE imaging provided excellent visualization of the remarkable vascular structure and characteristic histoarchitecture with microvasculature, intracytoplasmic vacuoles, and atypical cells consistent with hemangioblastoma. Resection and decompression of the chiasm was accomplished, and the patient was discharged with improved vision. The final histopathological diagnosis was hemangioblastoma. When the benefits of obtaining intraoperative frozen sections greatly outweigh the associated risks, CLE imaging can aid in decision-making. CLE imaging offers real-time, on-the-fly evaluation of intraoperative tissue without the need to biopsy a vascular lesion.

In vivo differentiation of adenoma and carcinoma in CRC progression by PAR2-M/DiD ratiometric fluorescence and its microstructures

Colorectal cancer (CRC) is the third leading cause of morbidity and mortality for malignant cancer globally. Accurate differentiation of colorectal lesions can provide information for precise treatment and minimize postoperative complications. Specific biomarkers for adenoma and carcinoma classification were compellingly needed. TCGA datasets were applied for identifying differential genes. Particularly, protease activated receptor 2 (PAR2) coded by F2RL1 gene displayed significant difference between clinical colorectal normal, adenoma and carcinoma tissues by IHC. Based on PAR2 targeting agonist, high affinity NIR probe PAR2-M was designed and validated both in vitro and in vivo, with membrane tag DiD as internal reference. In vivo fluorescence ratio of PAR2-M/DiD was absolutely higher in carcinoma than adenoma on CRC mouse models and confocal laser endomicroscopy (CLE) images exhibited their microstructures along with different PAR2 expression. The conclusive differentiation of adenoma and carcinoma was firmly supported by Convolutional Neural Networks (CNN) algorithm, with specificity of 95%.

Goblet cells segmentation from confocal laser endomicroscopy with an improved U-Net

Gastric intestinal metaplasia (GIM) is regarded as a remarkable precursor for the development of intestinal-type stomach cancer. Goblet cell (GC) segmentation is the crucial step for assessing the degree of GIM by confocal laser endomicroscopy (CLE). However, GC segmentation by hand is difficult, unreliable, and time-consuming. Meanwhile, due to the high resolution and noise interference of CLE images, existing segmentation approaches perform poorly on this task. To tackle those issues, we collected 343 confocal laser endomicroscopy images of 62 patients from a Grade-A tertiary hospital. Each CLE image is manually annotated and then verified three times by skilled medical specialists. Then, U-Net is improved by incorporating the pixel gradient attention mechanism, which focuses on color gradient information around GC and captures color gradient features to direct feature maps in the skip connection layer. At last, the model output vector is used to calculate the possibility map and generate the final segmentation area. Compared with mainstream models, our proposed GC segmentation method from CLE with an improved U-Net (GCSCLE) performs the better segmentation result when tested on our CLE dataset and achieved an IOU of 87.95% and a DICE of 86.64%. Our result shows, the performance of the GCSCLE can be compared with the manual CLE image processing in clinical settings, and it can improve segmentation accuracy and save time and costs.

Confocal laser endomicroscopic imaging of equine endometrium: evaluating image quality

A heathy uterus is a key piece of the fertility jigsaw and assessment of the endometrium is an essential component of the breeding soundness examination in mares. The current gold standard for evaluating uterine health is histopathological assessment of an endometrial biopsy. The Kenney-Doig (KD) grading scale is most used and scores both fibrotic and inflammatory changes present within the endometrium. The KD grade provides a prognostic indicator of the probability a mare can carry a pregnancy to full-term. Inconsistencies may arise when a single blinded biopsy is used, especially when uterine pathologic changes don't have a widespread distribution pattern. Confocal laser endomicroscopy (CLE) is a novel imaging modality that enables real-time in-vivo imaging of the endometrium. Multiple subcellular endometrial images (“virtual biopsies”) can be captured in quick succession in multiple locations within the uterine lumen noninvasively. Image depth in the z-axis can be controlled (0-100µm). Image quality was noted to be variable between mares in a previous pilot study (Gallacher.et.al. Journal of Equine Veterinary Science. 2018;66:106). The objective was to determine if physiological endometrial oedema present during oestrus has a negative effect on image quality. Twenty-four mares were examined with CLE (ViewnVivo, Optiscan, Australia) on two separate occasions, once during oestrus and once in dioestrus. A total of 4294 CLE images were captured, witha mean of 89.5 images saved per exam. A 4-point grading scale (1=poor and 4=excellent) was developed for quality assessment. A subset of ∼10 representative images from each of the 48 examinations were selected. Four examiners were blinded to the phase of the oestrous cycle and graded the images using the quality criteria assigned to the 4-point grading scale. An intraclass correlation (ICC) was estimated to determine reliability between the 4 scorers. The ICC coefficient was estimated using the irr package in R. Moderate agreement (ICCC(C,4) = 0.69) between scorers was observed while using the two-way random effect models and average rater where p<0.001. Quality data were analysed using a mixed model in asreml-R. The fixed factors included in the model were depth (1-luminal epithelium, 2-subluminal), cycle (oestrus and dioestrus) with brightness included as a covariate. Scorer and horse were included as random terms. There was a difference (p<0.001) between depth 1 and 2 (2.51 vs 2.71 respectively), and cycle (p<0.001) dioestrus = 2.71 oestrus = 2.51. The brightness of the image was also related to quality score (p<0.001) where a 100unit increase in brightness was associated with a reduction in quality score of ∼0.3. Image quality appears to be influenced by the phase of the oestrous cycle and this may guide the clinician to perform CLE examination during dioestrus.

Dynamic real-time optical microscopy of oral mucosal lesions using confocal laser endomicroscopy.

Confocal laser endomicroscopy (CLE) is a novel non-invasive point-of-care optical biopsy technology that enables real-time in vivo microscopic visualisation of cellular and tissue architecture. In this study, we assessed the diagnostic accuracy of a hand-held fluorescence single-fibre distal-scanning CLE (fsdCLE) platform for diagnosing oral epithelial dysplasia (OED) and oral squamous cell carcinoma (OSCC). Forty-seven patients presenting with 63 distinct oral mucosal lesions were subjected to optical biopsy using a miniaturised fsdCLE system (ViewnVivo®, Optiscan Imaging Ltd) and topical exogenous acriflavine hydrochloride contrast agent before undergoing tissue biopsy and histopathological consensus review by four pathologists. CLE images were captured in vivo in real-time during clinical examination and assessed on-the-fly for the presence of cellular and architectural features of OED/OSCC offering an instantaneous diagnosis. Predicted optical diagnoses were compared to definitive consensus tissue histopathology. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy were calculated for the presence/absence of dysplasia/malignancy on optical biopsy. Percentage agreement, Fleiss' kappa, and intraclass correlation coefficient (ICC) were calculated for each assessment stage during the consensus histopathology process. Diagnostic accuracy was extremely high at 88.9%. Other metrics were sensitivity 86.8%, specificity 92%, PPV 94.3% and NPV 82.1%. One hundred percent of carcinoma cases were detected accurately using CLE in the clinic. fsdCLE is a highly accurate, easy-to-use, rapid and slide-free point-of-care in vivo optical technology for diagnosing OED/OSCC and discriminating between dysplastic and non-dysplastic pathology. It demonstrates near-perfect agreement with traditional consensus histopathology without the need for physical tissue biopsy.

384 Characterization of Ex Vivo and In Vivo Intraoperative Neurosurgical Confocal Laser Endomicroscopy Imaging

INTRODUCTION: The new FDA-cleared fluorescein sodium (FNa)-based confocal laser endomicroscopy (CLE) imaging system allows for intraoperative on-the-fly cellular level imaging. Two feasibility studies have been completed with intraoperative use of this CLE system in ex vivo and in vivo modalities. METHODS: Images acquired from two prospective CLE clinical studies, one ex vivo and one in vivo, were analyzed quantitatively. Two image quality parameters – brightness and contrast – were measured using Fiji software and compared between ex vivo and in vivo images for imaging timing from FNa dose and in glioma, meningioma, and intracranial metastatic tumor cases. The diagnostic performance of the two studies was compared. RESULTS: Overall, the in vivo images have higher brightness and contrast than the ex vivo images (p &lt; 0.001). A weak negative correlation exists between image quality and timing of imaging after FNa dose for the ex vivo images, but not the in vivo images. In vivo images have higher brightness and contrast than ex vivo images acquired after FNa redosing. In vivo images have higher image quality than ex vivo images (p &lt; 0.001) in glioma, meningioma, and intracranial metastatic tumor cases. In vivo imaging yielded higher sensitivity (90% vs. 72%) and negative predictive value (81% vs 38%) than ex vivo imaging. CONCLUSIONS: In our setting, in vivo CLE optical biopsy outperforms ex vivo CLE by producing higher quality images and less image deterioration, leading to better diagnostic performance. These results support the in vivo modality as the modality of choice for intraoperative CLE imaging.

Cellular density and variability in laryngeal and pharyngeal squamous cell carcinoma using confocal laser endomicroscopy.

Confocal laser endomicroscopy (CLE) allows the visualization of epithelium in a thousand-fold magnification. This study analyzes the architectural differences at the cellular level of the mucosa and squamous cell carcinoma (SCC). A total of 60 CLE sequences recorded in 5 patients with SCC undergoing laryngectomy between October 2020 and February 2021 were analyzed. The corresponding histologic sample derived from H&E staining was assigned to each sequence, capturing CLE images of the tumor and healthy mucosa. In addition, the cellular structure analysis was performed to diagnose SCC by measuring the total number of cells and cell size in 60 sequences in a fixed field of view (FOV) with 240 μm in diameter (45,239 μm2). Out of 3,600 images, 1,620 (45%) showed benign mucosa and 1,980 (55%) SCC. The automated analysis yielded a difference in cell size, with healthy epithelial cells being 171.9±82.0 μm2 smaller than SCC cells, which were 246.3±171.9 μm2 and showed greater variability in size (p=0.037). In addition, due to the probe's fixed FOV, there was a difference in cell count with a total of 188.7±38.3 and 124.8±38.6 cells in images of normal epithelium and SCC (p<0.001), respectively. Regarding cell density as a criterion for the differentiation of benign/malign, using a cut-off value of 145.5 cells/FOV, we obtained sensitivity and specificity of 88.0% and 71.9%, respectively. SCC reveals marked differences at a cellular level compared to the healthy epithelium. Our results further support the importance of this feature for identifying SCC during CLE imaging.

Boosting few-shot confocal endomicroscopy image recognition with feature-level MixSiam.

As an emerging early diagnostic technology for gastrointestinal diseases, confocal laser endomicroscopy lacks large-scale perfect annotated data, leading to a major challenge in learning discriminative semantic features. So, how should we learn representations without labels or a few labels? In this paper, we proposed a feature-level MixSiam method based on the traditional Siamese network that learns the discriminative features of probe-based confocal laser endomicroscopy (pCLE) images for gastrointestinal (GI) tumor classification. The proposed method is divided into two stages: self-supervised learning (SSL) and few-shot learning (FS). First, in the self-supervised learning stage, the novel feature-level-based feature mixing approach introduced more task-relevant information via regularization, facilitating the traditional Siamese structure can adapt to the large intra-class variance of the pCLE dataset. Then, in the few-shot learning stage, we adopted the pre-trained model obtained through self-supervised learning as the base learner in the few-shot learning pipeline, enabling the feature extractor to learn richer and more transferable visual representations for rapid generalization to other pCLE classification tasks when labeled data are limited. On two disjoint pCLE gastrointestinal image datasets, the proposed method is evaluated. With the linear evaluation protocol, feature-level MixSiam outperforms the baseline by 6% (Top-1) and the supervised model by 2% (Top1), which demonstrates the effectiveness of the proposed feature-level-based feature mixing method. Furthermore, the proposed method outperforms the previous baseline method for the few-shot classification task, which can help improve the classification of pCLE images lacking large-scale annotated data for different stages of tumor development.

Clinical feasibility of miniaturized Lissajous scanning confocal laser endomicroscopy for indocyanine green-enhanced brain tumor diagnosis.

Intraoperative real-time confocal laser endomicroscopy (CLE) is an alternative modality for frozen tissue histology that enables visualization of the cytoarchitecture of living tissues with spatial resolution at the cellular level. We developed a new CLE with a "Lissajous scanning pattern" and conducted a study to identify its feasibility for fluorescence-guided brain tumor diagnosis. Conventional hematoxylin and eosin (H&E) histological images were compared with indocyanine green (ICG)-enhanced CLE images in two settings (1): experimental study with in vitro tumor cells and ex vivo glial tumors of mice, and (2) clinical evaluation with surgically resected human brain tumors. First, CLE images were obtained from cultured U87 and GL261 glioma cells. Then, U87 and GL261 tumor cells were implanted into the mouse brain, and H&E staining was compared with CLE images of normal and tumor tissues ex vivo. To determine the invasion of the normal brain, two types of patient-derived glioma cells (CSC2 and X01) were used for orthotopic intracranial tumor formation and compared using two methods (CLE vs. H&E staining). Second, in human brain tumors, tissue specimens from 69 patients were prospectively obtained after elective surgical resection and were also compared using two methods, namely, CLE and H&E staining. The comparison was performed by an experienced neuropathologist. When ICG was incubated in vitro, U87 and GL261 cell morphologies were well-defined in the CLE images and depended on dimethyl sulfoxide. Ex vivo examination of xenograft glioma tissues revealed dense and heterogeneous glioma cell cores and peritumoral necrosis using both methods. CLE images also detected invasive tumor cell clusters in the normal brain of the patient-derived glioma xenograft model, which corresponded to H&E staining. In human tissue specimens, CLE images effectively visualized the cytoarchitecture of the normal brain and tumors. In addition, pathognomonic microstructures according to tumor subtype were also clearly observed. Interestingly, in gliomas, the cellularity of the tumor and the density of streak-like patterns were significantly associated with tumor grade in the CLE images. Finally, panoramic view reconstruction was successfully conducted for visualizing a gross tissue morphology. In conclusion, the newly developed CLE with Lissajous laser scanning can be a helpful intraoperative device for the diagnosis, detection of tumor-free margins, and maximal safe resection of brain tumors.

PMON155 Intraoperative, confocal laser endomicroscopy in pituitary surgery

Abstract Introduction Fluorescence imaging, such as confocal laser endomicroscopy (CLE) is a fast modality of image acquisition, which offers the possibility of early intraoperatively diagnosis and may be used to improve surgical decision making. Objective Within a 3-month prospective trail we used the Zeiss CONVIVO CLE in transspehnoidal surgery for different sellar pathologies. CLE is a miniaturized portable, handheld, fiber optic-based, confocal laser scanning microscope used in conjunction with a 488-nm excitable fluorophore. During the trial period we investigated feasibility, diagnostic accuracy and usability in terms of differential diagnosis in transsphenoidal surgeries of various pituitary tumours (15 pituitary adenomas, 2 craniopharyngiomas, 1 sellar metastasis). Method: Patients were randomly selected. Inclusion criteria were a pituitary macroadenoma or a sellar mass lesion exceeding 10 mm in diameter. A transsphenoidal approach was used and fluorescin was injected upon opening the sellar floor. CLE was carried out immediately after dural opening and during tumour resection, whenever possible. Additionally, the microprobe was used ex vivo with the resected tumor material. Histopathologic features of CLE images and H&amp;E-stained permanent sections were assessed and compared postoperatively. Result: We were able to obtain CLE in-vivo images in all cases. Using the microscopic approach, no problems occurred regarding space for the microprobe. To obtain in-vivo images we had to rest the probe within a mount. Wiggling of the surgeon's hand, pulse synchronous movements of the mass and mild intraoperative venous bleeding led to inexact images. To overcome these problems, we used CLE as early as possible at the beginning of tumor resection to avoid bleeding interferences. Additionally we investigated resected tumor specimens ex-vivo within the OR. Since no bleeding and no interference with movements occurred image quality and acquisition time improved. Histopathologic features of CLE images and H&amp;E-stained permanent sections showed, that except for the sellar metastasis, a diagnosis could be made on the basis of the intraoperative CLE images, whereas the ex-vivo images were more precise. Conclusion CLE is a potential imaging modality for transnasal assessment of skull base and pituitary lesions. It may help to establish an early diagnosis, to identify pituitary microadenomas, especially in cushing´s disease and to distinguish the various sellar pathologies. CLE is a fascinating possibility to verify a suspected diagnoses already intraoperatively. Presentation: Monday, June 13, 2022 12:30 p.m. - 2:30 p.m.

S3358 Exploring Probe-Based Confocal Laser Endomicroscopy as a Novel Tool in Graft Monitoring After Small Bowel Transplantation

Introduction: The most common and severe complication early in Small Bowel Transplant (SBT) is graft rejection making timely detection key to graft preservation. Endoscopic surveillance with protocol biopsy is the gold standard for graft monitoring. Confocal Laser Endomicroscopy (CLE) has emerged as a modality for real-time diagnosis at a histological scale. However, its role in SBT is not known. Herein, we report 3 cases of SBT patients and compare CLE results with standard protocols. Case Description/Methods: Case 1: A 73-year-old man underwent SBT for sclerosing mesenteritis and total enterectomy. CLE and ileoscopy was performed 15 times post-transplant without findings of acute cellular rejection (ACR). Patient expired of sepsis on day 30 post-transplant. Case 2: A 17-year-old male with familial adenomatous polyposis complicated by total colectomy, hepatoblastoma s/p resection and chemotherapy underwent SBT for large desmoid tumor requiring total enterectomy. CLE and ileoscopy was performed 14 times post-transplant without findings of ACR. Case 3: A 23-year-old female underwent SBT for intestinal neuronal dysplasia after chronic rejection from 1st transplant 10 years ago. Ileoscopy done at 20 days post-transplant showed congested and ulcerated mucosa in the proximal ileum. CLE images and histopathology showed findings consistent with ACR. Immunosuppression was increased and reassessment with CLE and ileoscopy done. Discussion: Cases 1 and 2 show normal CLE assessment. Normal villus architecture, non-edematous, and adequate microcirculation suggesting low suspicion for ACR. This correlated with normal pathology characterized by less than 6 apoptotic bodies per 10 crypts, no intravascular neutrophils, and negative viral cytopathology. However, in case 3, CLE images showed significant villus congestion and decreased microcirculation (Figure) indicative of ACR. Biopsies confirmed ACR. Rejection protocols were rapidly initiated.Repeat CLE and ileoscopy showed recovery of mucosal congestion and return of microcirculation. This correlated with complete resolution of ACR on pathology. In the absence of endoscopic criteria for small bowel graft rejection, biopsies are the gold standard. Enhanced technology for in vivo visualization of grafted bowel is needed and would allow for examination of a greater surface area of the bowel than limited biopsies. CLE is a promising tool that could significantly improve early transplant graft monitoring. Further data and prospective studies are needed.Figure 1.: CLE images showing significant villus congestion, and decreased microcirculation.

P13.05.A Image annotation guideline for invivo confocal laser endomicroscopy, interrater reliability and how to learn from medical consensus for machine learning algorithms

Abstract Background Intraoperative confocal laser endomicroscopy (CLE) is an in vivo imaging technique increasingly studied in neurosurgery and neuropathology. It can be affected by artifacts introduced by the CLE device or related to the intraoperative setting. We developed and evaluated an image annotation guideline (AGL) to detect and eliminate images bearing no valuable information as a result of such artifacts. Images ware classified into good and bad quality, based on defined technical criteria, which are also considered relevant by clinical experts. Material and Methods Datasets were created from intraoperative CLE in vivo specimens of patients resected for brain tumors. The process from data collection to development of the ML algorithm followed 7 steps: data quality specification, image and metadata collection, AGL development, annotation, data allocation for clinical validation, clinical validation, and, optionally, algorithm development. Final diagnoses were obtained by pathological analysis. Artifacts were grouped into three categories: diminished signal-to-noise-ratio (dSNR), optical distortions (movement/perturbations), and contrast/brightness artifacts. Images were annotated by 4 medical data annotators (T4). For clinical validation, 500 images were excluded from the training data and additionally annotated by 3 board certified neuropathologists (NPs 1-3) with experience in CLE imaging, to determine the medical consensus on good and bad images. All raters (NPs) were compared against each other and against T4; T4 was also compared against the medical consensus. Cohen’s Kappa and overall percentage agreement (OPA) were used to evaluate inter-rater reliability. Positive percent agreement (PPA) and negative percentage agreement (NPA) were also used to evaluate agreement between medical consensus and T4. Results 21,616 CLE images and corresponding clinical metadata were collected from 94 patients and annotated. For each case between 27 and 815 CLE images were acquired over the course of the surgery (mean=175 images per case, SD=170.6). 11% and 13% of images were labeled as dSNR and distortion, respectively, and 34% as class contrast. 42% of the images represented the good quality images. Interrater agreement between the 3 NPs ranged between 0.30 and 0.59. Agreement between T4 and the medical consensus was substantial (Cohen’s Kappa &amp;gt;=0.61). OPA between T4 and the medical consensus was 80.60%, PPA 72.34% and NPA 87.92%. Conclusion Annotations according to a well-structured and expertly curated AGL show higher values for Cohen’s Kappa and Overall Percent Agreement (OPA) with the medical consensus, than that of individual experts among one another. Such an AGL can be considered appropriate and produces on par results with annotations by a group of experts in the field and can be further employed for training machine learning (ML) algorithms.

P05.01.B Distribution of the staining agent sodium-fluorescein in cerebral neoplasia - a comparison between intraoperative confocal laser endomicroscopy and in vitro fluorescence microscopy

Abstract Background While conventional frozen sections and other ex vivo microscopy always come along with artefacts, in vivo confocal laser endomicroscopy (CLE) offers less invasive, less manipulated imaging of living tissues. Understanding the distribution of the staining agent sodium-fluorescein (NaFl) plays a key role in establishing in vivo CLE as a new opportunity for real-time transmitted, in vivo imaging. By comparing the distribution of NaFl in in vivo and ex vivo CLE to conventional fluorescence microscopy of NaFl incubated tumor cell cultures we gain a better understanding of the staining mechanism. Material and Methods Sodium-fluorescein is the water-soluble sodium salt of fluorescein and is used as a fluorescent tracer in neurosurgery. The staining agent was applied intravenously at the beginning of the surgical procedure. In vivo CLE of the lesion was performed 30 to 50 minutes later and compared to ex vivo CLE imaging and conventional fluorescence microscopy. In addition, different tumor cell lines derived from malignant gliomas and carcinomas, respectively, were incubated with NaFl in vitro and the uptake of the fluorescent dye was monitored over time. Results From initial results, the intraoperative images showed specific fluorescein distribution depending on the architecture of the tumor entity. In most cases, glial tumors demonstrated higher accumulation of the staining agent in the extracellular tumor matrix, whereas the cells of carcinoma metastases appeared to take up NaFl intracellularly. These results were corroborated by NaFl uptake in cell culture experiments. Compared to ex vivo CLE, in vivo imaging offered a faster assessment of the tissue, brighter images and higher staining levels. Due to movement artefacts and the narrow intraoperative imaging time frame, in vivo CLE images were sometimes impaired by lower image quality. Conclusion The specific distribution of the fluorescent agent NaFl allowed for a discrimination between the different neoplastic entities. Images from in vivo and ex vivo confocal laser endomicroscopy showed NaFl uptake in concordance with the results of NaFl incubated cell cultures. Intraoperative, in vivo confocal laser endomicroscopy shows promising first results in the understanding of brain tumor histomorphology in situ. Being faster and less manipulated by artefacts than ex vivo investigations, it opens up wider opportunities for research and intraoperative diagnostics.

Bronchoscopic Journey of in vivo Real-Time Microscopic Imaging in ILD: A Case Series

Background: Patients with interstitial lung diseases (ILDs) frequently present with nondiagnostic high-resolution CT (HRCT) scan and bronchoalveolar lavage (BAL) results, resulting in the need for invasive surgical or cryo-lung biopsy that is associated with significant morbidity. Confocal laser endomicroscopy (CLE) and optical coherence tomography (OCT) are high-resolution laser and light-based techniques that provide real-time imaging of the alveolar compartment during bronchoscopy with a different depth and field of view. Objectives: The aim of the study was to correlate OCT and CLE imaging to HRCT imaging in ILD. Methods: This is a retrospective case series of 20 ILD patients who underwent alveolar CLE and OCT imaging during a standard bronchoscopy with BAL, followed by a lung biopsy when indicated. CLE and OCT imaging were compared to four main HRCT patterns and histology. The final diagnosis was based on the multidisciplinary discussion diagnosis. Results: Bronchoscopic CLE and OCT imaging were feasible and safe and provided additional high-detailed anatomical information compared to the HRCT. Bronchoscopic real-time CLE was capable of identification of “alveolar cells” (ground glass opacities) and lung fibrosis (increased alveolar elastin fibers). Bronchoscopic real-time OCT allowed for visualization of “patchy fibrotic disease”, “honeycombing” (microcysts), and mucosal granulomas in the airways. Conclusions: Bronchoscopic CLE and OCT of the alveolar compartment is feasible and safe and enables minimally invasive, high-resolution detection of specific ILD features with the potential to improve ILD diagnostics and monitoring and decrease the need for surgical or cryo-lung biopsies.