Publications
505 results found
Elson D, 2022, Surgical spectral sensing and imaging (invited), Photonics Europe: Clinical Biophotonics II
Elson D, 2022, Multispectral and polarization-resolved endoscopic surgical imaging (invited), Photon 2022
Nazarian S, Gkouzionis I, Kawka M, et al., 2021, Real-time tracking and classification of tumour and non-tumour tissue in upper gastrointestinal cancer specimens using diffuse reflectance spectroscopy, UGI Congress 2021, ISSN: 0007-1323
Gkouzionis I, Nazarian S, Anandakumar A, et al., 2021, Using diffuse reflectance spectroscopy probe tracking to identify non-tumour and tumour tissue in upper gastrointestinal specimens, Translational Biophotonics: Diagnostics and Therapeutics, Publisher: SPIE
The use of a diffuse reflectance spectroscopy probe and tracking system was successfully used in real-time for automated tissue classification in upper gastrointestinal surgery to aid resection margin assessment.
Cartucho J, Wang C, Huang B, et al., 2021, An enhanced marker pattern that achieves improved accuracy in surgical tool tracking, Computer Methods in Biomechanics and Biomedical Engineering: Imaging and Visualization, Vol: 10, Pages: 1-9, ISSN: 2168-1163
In computer assisted interventions (CAI), surgical tool tracking is crucial for applications such as surgical navigation, surgical skill assessment, visual servoing, and augmented reality. Tracking of cylindrical surgical tools can be achieved by printing and attaching a marker to their shaft. However, the tracking error of existing cylindrical markers is still in the millimetre range, which is too large for applications such as neurosurgery requiring sub-millimetre accuracy. To achieve tool tracking with sub-millimetre accuracy, we designed an enhanced marker pattern, which is captured on images from a monocular laparoscopic camera. The images are used as input for a tracking method which is described in this paper. Our tracking method was compared to the state-of-the-art, on simulation and ex vivo experiments. This comparison shows that our method outperforms the current state-of-the-art. Our marker achieves a mean absolute error of 0.28 [mm] and 0.45 [°] on ex vivo data, and 0.47 [mm] and 1.46 [°] on simulation. Our tracking method is real-time and runs at 55 frames per second for 720×576 image resolution.
Teh JJ, Cai W, Kedrzycki M, et al., 2021, 392 Magseed-guided wide local excision during the COVID-19 pandemic: a tenable solution to barriers in accessing elective breast cancer surgery, Association of Surgeons in Training, Publisher: British Journal of Surgery Society, ISSN: 0007-1323
Perrott C, Patil A, Elson D, et al., 2021, Novel methods of detecting tumour margins in gastrointestinal cancer surgery, 2021 Association of Surgeons in Training International Surgical Conference, Publisher: British Journal of Surgery Society, Pages: 1-1, ISSN: 0007-1323
AimGastrointestinal (GI) cancers account for 26% of global cancer incidence with prevalence projected to rise exponentially due to the ageing population and lifestyle choices. Surgical resection is the mainstay of treatment to remove the cancer in its entirety to achieve an R0 resection. Positive margins, when cancerous tissue has been left in situ, is associated with increased morbidity and mortality. Current margin assessment involves histopathological analysis, after resection of the specimen. Diffuse Reflectance Spectroscopy (DRS) and Hyperspectral Imaging (HSI) are novel imaging techniques that have the potential to provide real-time assessment of cancer margins intra-operatively to reduce the incidence of positive resection margins and improve patient outcomes. The aim of this review is to assess the current state of evidence for the use of novel imaging techniques in GI cancer margin assessment.MethodA literature review was conducted of studies using DRS and HSI in GI cancers in adult patients, published from inception to October 2020.ResultsA total of 15 studies were analysed, nine of which used DRS and six used HSI and the majority of studies were performed ex-vivo. Current image acquisition techniques and processing algorithms vary greatly. The sensitivity and specificity of DRS ranged from 0.90-0.98 and 0.88-0.95 respectively and for HSI 0.63-0.98 and 0.69-0.98, respectively across five types of GI cancers.ConclusionsDRS and HSI are novel imaging techniques, currently in their infancy but the outlook is promising. With further research focused on standardising methodology and in-vivo settings, DRS and HSI could transform intra-operative margin assessment in GI cancers.
Nazarian S, Gkouzionis I, Anandakumar A, et al., 2021, Using diffuse reflectance spectroscopy (DRS) to identify tumour and non-tumour tissue in upper gastrointestinal specimens, Association of Surgeons of Great Britain and Ireland Virtual Congress, Publisher: British Journal of Surgery Society, Pages: 41-41, ISSN: 0007-1323
AimCancers of the upper gastrointestinal (GI) tract remain a major contributor to the global cancer risk. Surgery aims to completely resect tumour with clear margins, whilst preserving as much surrounding tissue. Diffuse reflectance spectroscopy (DRS) is a novel technique that presents a promising advancement in cancer diagnosis. We have developed a novel DRS system with tracking capability. Our aim is to classify tumour and non-tumour GI tissue in real-time using this device to aid intra-operative analysis of resection margins.MethodAn ex-vivo study was undertaken in which data was collected from consecutive patients undergoing upper GI cancer resection surgery between August 2020- January 2021. A hand-held DRS probe and tracking system was used on normal and cancerous tissue to obtain spectral information. After acquisition of all spectra, a classification system using histopathology results was created. A user interface was developed using Python 3.6 and Qt5. A support vector machine was used to classify the results.ResultsThe data included 4974 normal spectra and 2108 tumour spectra. The overall accuracy of the DRS probe in differentiating normal versus tumour tissue was 88.08% for the stomach (sensitivity 84.8%, specificity 89.3%), and 91.42% for the oesophagus (sensitivity 76.3%, specificity 98.9%).ConclusionWe have developed a successful DRS system with tracking capability, able to process thousands of spectra in a small timeframe, which can be used in real-time to distinguish tumour and non-tumour tissue. This can be used for intra-operative decision-making during upper GI cancer surgery to help select the best resection plane.
Huang B, Zheng J-Q, Nguyen A, et al., 2021, Self-supervised generative adverrsarial network for depth estimation in laparoscopic images, International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), Publisher: Springer, Pages: 227-237
Dense depth estimation and 3D reconstruction of a surgical scene are crucial steps in computer assisted surgery. Recent work has shown that depth estimation from a stereo image pair could be solved with convolutional neural networks. However, most recent depth estimation models were trained on datasets with per-pixel ground truth. Such data is especially rare for laparoscopic imaging, making it hard to apply supervised depth estimation to real surgical applications. To overcome this limitation, we propose SADepth, a new self-supervised depth estimation method based on Generative Adversarial Networks. It consists of an encoder-decoder generator and a discriminator to incorporate geometry constraints during training. Multi-scale outputs from the generator help to solve the local minima caused by the photometric reprojection loss, while the adversarial learning improves the framework generation quality. Extensive experiments on two public datasets show that SADepth outperforms recent state-of-the-art unsupervised methods by a large margin, and reduces the gap between supervised and unsupervised depth estimation in laparoscopic images.
Ahmad OF, Mori Y, Misawa M, et al., 2021, Establishing key research questions for the implementation of artificial intelligence in colonoscopy - a modified Delphi method., Endoscopy, Vol: 53, Pages: 893-901, ISSN: 0013-726X
Background and Aims Artificial intelligence (AI) research in colonoscopy is progressing rapidly but widespread clinical implementation is not yet a reality. We aimed to identify the top implementation research priorities. Methods An established modified Delphi approach for research priority setting was used. Fifteen international experts, including endoscopists and translational computer scientists/engineers from 9 countries participated in an online survey over 9 months. Questions related to AI implementation in colonoscopy were generated as a long-list in the first round, and then scored in two subsequent rounds to identify the top 10 research questions. Results The top 10 ranked questions were categorised into 5 themes. Theme 1: Clinical trial design/end points (4 questions), related to optimum trial designs for polyp detection and characterisation, determining the optimal end-points for evaluation of AI and demonstrating impact on interval cancer rates. Theme 2: Technological Developments (3 questions), including improving detection of more challenging and advanced lesions, reduction of false positive rates and minimising latency. Theme 3: Clinical adoption/Integration (1 question) concerning effective combination of detection and characterisation into one workflow. Theme 4: Data access/annotation (1 question) concerning more efficient or automated data annotation methods to reduce the burden on human experts. Theme 5: Regulatory Approval (1 question) related to making regulatory approval processes more efficient. Conclusions This is the first reported international research priority setting exercise for AI in colonoscopy. The study findings should be used as a framework to guide future research with key stakeholders to accelerate the clinical implementation of AI in endoscopy.
Kedrzycki MS, Leiloglou M, Chalau V, et al., 2021, ASO visual abstract: the impact of temporal variation in indocyanine green administration on tumor identification during fluorescence-guided breast surgery, Annals of Surgical Oncology, Vol: 28, Pages: 650-651, ISSN: 1068-9265
Kedrzycki MS, Leiloglou M, Chalau V, et al., 2021, The impact of temporal variation in indocyanine green administration on tumor identification during fluorescence guided breast surgery., Annals of Surgical Oncology, Vol: 28, Pages: 5617-5625, ISSN: 1068-9265
BACKGROUND: On average, 21% of women in the USA treated with Breast Conserving Surgery (BCS) undergo a second operation because of close positive margins. Tumor identification with fluorescence imaging could improve positive margin rates through demarcating location, size, and invasiveness of tumors. We investigated the technique's diagnostic accuracy in detecting tumors during BCS using intravenous indocyanine green (ICG) and a custom-built fluorescence camera system. METHODS: In this single-center prospective clinical study, 40 recruited BCS patients were sub-categorized into two cohorts. In the first 'enhanced permeability and retention' (EPR) cohort, 0.25 mg/kg ICG was injected ~ 25 min prior to tumor excision, and in the second 'angiography' cohort, ~ 5 min prior to tumor excision. Subsequently, an in-house imaging system was used to image the tumor in situ prior to resection, ex vivo following resection, the resection bed, and during grossing in the histopathology laboratory to compare the technique's diagnostic accuracy between the cohorts. RESULTS: The two cohorts were matched in patient and tumor characteristics. The majority of patients had invasive ductal carcinoma with concomitant ductal carcinoma in situ. Tumor-to-background ratio (TBR) in the angiography cohort was superior to the EPR cohort (TBR = 3.18 ± 1.74 vs 2.10 ± 0.92 respectively, p = 0.023). Tumor detection reached sensitivity and specificity scores of 0.82 and 0.93 for the angiography cohort and 0.66 and 0.90 for the EPR cohort, respectively (p = 0.1051 and p = 0.9099). DISCUSSION: ICG administration timing during the angiography phase compared with the EPR phase improved TBR and diagnostic accuracy. Future work will focus on image pattern analysis and adaptation of the camera system to targeting fluorophores specific to breast cancer.
Kedrzycki MS, Leiloglou M, Ashrafian H, et al., 2021, Meta-analysis comparing fluorescence imaging with radioisotope and blue dye-guided sentinel node identification for breast cancer surgery., Annals of Surgical Oncology, Vol: 28, Pages: 3738-3748, ISSN: 1068-9265
INTRODUCTION: Conventional methods for axillary sentinel lymph node biopsy (SLNB) are fraught with complications such as allergic reactions, skin tattooing, radiation, and limitations on infrastructure. A novel technique has been developed for lymphatic mapping utilizing fluorescence imaging. This meta-analysis aims to compare the gold standard blue dye and radioisotope (BD-RI) technique with fluorescence-guided SLNB using indocyanine green (ICG). METHODS: This study was registered with PROSPERO (CRD42019129224). The MEDLINE, EMBASE, Scopus, and Web of Science databases were searched using the Medical Subject Heading (MESH) terms 'Surgery' AND 'Lymph node' AND 'Near infrared fluorescence' AND 'Indocyanine green'. Studies containing raw data on the sentinel node identification rate in breast cancer surgery were included. A heterogeneity test (using Cochran's Q) determined the use of fixed- or random-effects models for pooled odds ratios (OR). RESULTS: Overall, 1748 studies were screened, of which 10 met the inclusion criteria for meta-analysis. ICG was equivalent to radioisotope (RI) at sentinel node identification (OR 2.58, 95% confidence interval [CI] 0.35-19.08, p < 0.05) but superior to blue dye (BD) (OR 9.07, 95% CI 6.73-12.23, p < 0.05). Furthermore, ICG was superior to the gold standard BD-RI technique (OR 4.22, 95% CI 2.17-8.20, p < 0.001). CONCLUSION: Fluorescence imaging for axillary sentinel node identification with ICG is equivalent to the single technique using RI, and superior to the dual technique (RI-BD) and single technique with BD. Hospitals using RI and/or BD could consider changing their practice to ICG given the comparable efficacy and improved safety profile, as well as the lesser burden on hospital infrastructure.
Lin J, Clancy NT, Qi J, et al., 2021, Corrigendum to Dual-modality endoscopic probe for tissue surface shape reconstruction and hyperspectral imaging enabled by deep neural networks [Medical Image Analysis 48 (2018) 162-176/2018.06.004]., Medical Image Analysis, Vol: 72, Pages: 1-1, ISSN: 1361-8415
The first version of this article neglected to mention that this work was additionally supported by ERC award 637960. This has now been corrected online. The authors would like to apologise for any inconvenience caused.
Leiloglou M, Chalau V, Kedrzycki MS, et al., 2021, Tissue texture extraction in indocyanine green fluorescence imaging for breast-conserving surgery, Journal of Physics D: Applied Physics, Vol: 54, ISSN: 0022-3727
A two-camera fluorescence system for indocyanine green (ICG) signal detection has been developed and tested in a clinical feasibility trial of ten patients, with a resolution in the submillimetre scale. Immediately after systemic ICG injection, the two-camera system can detect ICG signals in vivo (~2.5 mg ${{\text{l}}^{ - 1}}$ or 3.2 × ${10^{ - 6}}{ }$ M). Qualitative assessment has shown that the fluorescence signal does not always correlate with the cancer location in the surgical scene. Conversely, fluorescence image texture metrics when used with the logistic regression model yields good accuracy scores in detecting cancer. We have demonstrated that intraoperative fluorescence imaging for resection guidance is a feasible solution to tackle the current challenge of positive resection margins in breast conserving surgery.
Kedrzycki M, Teh J, Cai W, et al., 2021, P053. Prospective single-centre qualitative service evaluation on magseed for wide local excision, Association of Breast Surgery Conference 2021, Publisher: Elsevier, Pages: e310-e310, ISSN: 0748-7983
Kedrzycki M, Leiloglou M, Thiruchelvam P, et al., 2021, P051. Fluorescence guided surgery in breast cancer: A systematic review of the literature, Association of Breast Surgery Conference 2021, Publisher: Elsevier, Pages: e309-e309, ISSN: 0748-7983
Kedrzycki MS, Leiloglou M, Chalau V, et al., 2021, Guiding light to optimize wide local excisions: the "GLOW" study, Volume XXII 2021 Annual Meeting Scientific Session, Publisher: Springer, Pages: S199-S200, ISSN: 1068-9265
Kedrzycki M, Leiloglou M, Leff D, et al., 2021, Versatility in fluorescence guided surgery with the GLOW camera system, Surgical Life: The Journal of the Association of Surgeons of Great Britain and Ireland, Vol: 59
Collins JW, Marcus HJ, Ghazi A, et al., 2021, Ethical implications of AI in robotic surgical training: A Delphi consensus statement, European Urology Focus, ISSN: 2405-4569
ContextAs the role of AI in healthcare continues to expand there is increasing awareness of the potential pitfalls of AI and the need for guidance to avoid them.ObjectivesTo provide ethical guidance on developing narrow AI applications for surgical training curricula. We define standardised approaches to developing AI driven applications in surgical training that address current recognised ethical implications of utilising AI on surgical data. We aim to describe an ethical approach based on the current evidence, understanding of AI and available technologies, by seeking consensus from an expert committee.Evidence acquisitionThe project was carried out in 3 phases: (1) A steering group was formed to review the literature and summarize current evidence. (2) A larger expert panel convened and discussed the ethical implications of AI application based on the current evidence. A survey was created, with input from panel members. (3) Thirdly, panel-based consensus findings were determined using an online Delphi process to formulate guidance. 30 experts in AI implementation and/or training including clinicians, academics and industry contributed. The Delphi process underwent 3 rounds. Additions to the second and third-round surveys were formulated based on the answers and comments from previous rounds. Consensus opinion was defined as ≥ 80% agreement.Evidence synthesisThere was 100% response from all 3 rounds. The resulting formulated guidance showed good internal consistency, with a Cronbach alpha of >0.8. There was 100% consensus that there is currently a lack of guidance on the utilisation of AI in the setting of robotic surgical training. Consensus was reached in multiple areas, including: 1. Data protection and privacy; 2. Reproducibility and transparency; 3. Predictive analytics; 4. Inherent biases; 5. Areas of training most likely to benefit from AI.ConclusionsUsing the Delphi methodology, we achieved international consensus among experts to develop and reach
Gkouzionis I, Avila-Rencoret F, Murphy J, et al., 2021, Real-time optical tracking of a diffuse reflectance spectroscopy probe for gastrointestinal tissue analysis, Advanced Biomedical and Clinical Diagnostic and Surgical Guidance Systems XIX, Publisher: SPIE
Wang D, Qi J, Huang B, et al., 2021, A polarization-based smoke removal method for surgical images, Polarized light and Optical Angular Momentum for biomedical diagnostics, Publisher: SPIE
Gkouzionis I, Nazarian S, Kawka M, et al., 2021, The use of machine learning for real-time detection of oesophageal and gastric cancer based on diffuse reflectance spectroscopy: a validation study, Joint Congress of the European Society for Diseases of the Esophagus and the International Gastric Cancer Association European Chapter
Chalau V, Kedrzycki M, Leiloglou M, et al., 2021, Variability of indocyanine green (ICG) near-infrared maximum emission wavelength in breast tissues, European Molecular Imaging Meeting
Vieira Cartucho J, Wang C, Huang B, et al., 2021, An Enhanced Marker Pattern that Achieves Improved Accuracy in Surgical Tool Tracking, Joint MICCAI 2021 Workshop on Augmented Environments for Computer-Assisted Interventions (AE-CAI), Computer-Assisted Endoscopy (CARE) and Context-Aware Operating Theatres 2.0 (OR2.0), Publisher: Taylor and Francis, ISSN: 2168-1163
Leiloglou M, Gkouzionis I, Kedrzycki MS, et al., 2021, Real-time spectral tracking routine for fluorescence hyperspectral guidance in breast conserving surgery
Fast spectral tracking routine, using simultaneous analysis of color and monochrome images, was developed and tested in phantoms. This routine could improve the efficiency of fluorescence hyperspectral imaging for breast conserving surgery guidance.
Nazarian S, Gkouzionis I, Kawka M, et al., 2021, The use of a diffuse reflectance spectroscopy probe and tracking system to classify tumour and non-tumour tissue in upper gastrointestinal cancer specimens to aid margin assessment, London Surgical Symposium
Cartucho J, Tukra S, Li Y, et al., 2021, VisionBlender: a tool to efficiently generate computer vision datasets for robotic surgery, Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, Vol: 9, Pages: 331-338, ISSN: 2168-1163
Surgical robots rely on robust and efficient computer vision algorithms to be able to intervene in real-time. The main problem, however, is that the training or testing of such algorithms, especially when using deep learning techniques, requires large endoscopic datasets which are challenging to obtain, since they require expensive hardware, ethical approvals, patient consent and access to hospitals. This paper presents VisionBlender, a solution to efficiently generate large and accurate endoscopic datasets for validating surgical vision algorithms. VisionBlender is a synthetic dataset generator that adds a user interface to Blender, allowing users to generate realistic video sequences with ground truth maps of depth, disparity, segmentation masks, surface normals, optical flow, object pose, and camera parameters. VisionBlender was built with special focus on robotic surgery, and examples of endoscopic data that can be generated using this tool are presented. Possible applications are also discussed, and here we present one of those applications where the generated data has been used to train and evaluate state-of-the-art 3D reconstruction algorithms. Being able to generate realistic endoscopic datasets efficiently, VisionBlender promises an exciting step forward in robotic surgery.
Kedrzycki MS, Elson DS, Leff DR, 2020, ASO author reflections: fluorescence-guided sentinel node biopsy for breast cancer, Annals of Surgical Oncology, Vol: 28, Pages: 3749-3750, ISSN: 1068-9265
Denham TLDO, Cleary K, Baena FRY, et al., 2020, Guest editorial medical robotics: surgery and beyond, IEEE Transactions on Medical Robotics and Bionics, Vol: 2, Pages: 509-510, ISSN: 2576-3202
The IEEE Transactions on Medical Robotics and Bionics (T-MRB) is an initiative shared by the two IEEE Societies of Robotics and Automation—RAS—and Engineering in Medicine and Biology—EMBS.
This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.