Clinical and computational development of a patient-calibrated ICGFA bowel transection recommender.

INTRODUCTION: Intraoperative indocyanine green fluorescence angiography (ICGFA) aims to reduce colorectal anastomotic complications. However, signal interpretation is inconsistent and confounded by patient physiology and system behaviours. Here, we demonstrate a proof of concept of a novel clinical and computational method for patient calibrated quantitative ICGFA (QICGFA) bowel transection recommendation. METHODS: Patients undergoing elective colorectal resection had colonic ICGFA both immediately after operative commencement prior to any dissection and again, as usual, just before anastomotic construction. Video recordings of both ICGFA acquisitions were blindly quantified post hoc across selected colonic regions of interest (ROIs) using tracking-quantification software and computationally compared with satisfactory perfusion assumed in second time-point ROIs, demonstrating 85% agreement with baseline ICGFA. ROI quantification outputs detailing projected perfusion sufficiency-insufficiency zones were compared to the actual surgeon-selected transection/anastomotic construction site for left/right-sided resections, respectively. Anastomotic outcomes were recorded, and tissue lactate was also measured in the devascularised colonic segment in a subgroup of patients. The novel perfusion zone projections were developed as full-screen recommendations via overlay heatmaps. RESULTS: No patient suffered intra- or early postoperative anastomotic complications. Following computational development (n = 14) the software recommended zone (ROI) contained the expert surgical site of transection in almost all cases (Jaccard similarity index 0.91) of the nine patient validation series. Previously published ICGFA time-series milestone descriptors correlated moderately well, but lactate measurements did not. High resolution augmented reality heatmaps presenting recommendations from all pixels of the bowel ICGFA were generated for all cases. CONCLUSIONS: By benchmarking to the patient's own baseline perfusion, this novel QICGFA method could allow the deployment of algorithmic personalised NIR bowel transection point recommendation in a way fitting existing clinical workflow.

Geotemporal Fluorophore Biodistribution Mapping of Colorectal Cancer: Micro and Macroscopic Insights.

Fluorescence-guided oncology promises to improve both the detection and treatment of malignancy. We sought to investigate the temporal distribution of indocyanine green (ICG), an exogenous fluorophore in human colorectal cancer. This analysis aims to enhance our understanding of ICG's effectiveness in current tumour detection and inform potential future diagnostic and therapeutic enhancements. METHODS: Fifty consenting patients undergoing treatment for suspected/confirmed colorectal neoplasia provided near infrared (NIR) video and imagery of transanally recorded and ex vivo resected rectal lesions following intravenous ICG administration (0.25 mg/kg), with a subgroup providing tissue samples for microscopic (including near infrared) analysis. Computer vision techniques detailed macroscopic 'early' (<15 min post ICG administration) and 'late' (>2 h) tissue fluorescence appearances from surgical imagery with digital NIR scanning (Licor, Lincoln, NE, USA) and from microscopic analysis (Nikon, Tokyo, Japan) undertaken by a consultant pathologist detailing tissue-level fluorescence distribution over the same time. RESULTS: Significant intra-tumoural fluorescence heterogeneity was seen 'early' in malignant versus benign lesions. In all 'early' samples, fluorescence was predominantly within the tissue stroma, with uptake within plasma cells, blood vessels and lymphatics, but not within malignant or healthy glands. At 'late' stage observation, fluorescence was visualised non-uniformly within the intracellular cytoplasm of malignant tissue but not retained in benign glands. Fluorescence also accumulated within any present peritumoural inflammatory tissue. CONCLUSION: This study demonstrates the time course diffusion patterns of ICG through both benign and malignant tumours in vivo in human patients at both macroscopic and microscopic levels, demonstrating important cellular drivers and features of geolocalisation and how they differ longitudinally after exposure to ICG.

Impact of standardising indocyanine green fluorescence angiography technique for visual and quantitative interpretation on interuser variability in colorectal surgery.

AIM/BACKGROUND: Intra-operative colonic perfusion assessment via indocyanine green fluorescence angiography (ICGFA) aims to address malperfusion-related anastomotic complications; however, its interpretation suffers interuser variability (IUV), especially early in ICGFA experience. This work assesses the impact of a protocol developed for both operator-based judgement and computational development on interpretation consistency, focusing on senior surgeons yet to start using ICGFA. METHODS: Experienced and junior gastrointestinal surgeons were invited to complete an ICGFA-experience questionnaire. They subsequently interpreted nine operative ICGFA videos regarding perfusion sufficiency of a surgically prepared distal colon during laparoscopic anterior resection by indicating their preferred site of proximal transection using an online annotation platform (mindstamp.com). Six ICGFA videos had been prepared with a clinical standardisation protocol controlling camera and patient positioning of which three each had monochrome near infrared (NIR) and overlay display. Three others were non-standardised controls with synchronous NIR and overlay picture-in-picture display. Differences in transection level between different cohorts were assessed for intraclass correlation coefficient (ICC) via ImageJ and IBM SPSS. RESULTS: 58 clinicians (12 ICGFA experts, 46 ICGFA inexperienced of whom 23 were either finished or within one year of finishing training and 23 were junior trainees) participated as per power calculations. 63% felt that ICGFA should be routinely deployed with 57% believing interpretative competence requires 11-50 cases. Transection level concordance was generally good (ICC = 0.869) across all videos and levels of expertise (0.833-0.915). However, poor agreement was evident with the standardised protocol videos for overlay presentation (0.208-0.345). Similarly, poor agreement was seen for the monochrome display (0.392-0.517), except for those who were trained but ICG inexperienced (0.877) although even here agreement was less than with unstandardised videos (0.943). CONCLUSION: Colorectal ICGFA acquisition and display standardisation impacts IUV with this specific protocol tending to diminish surgeon interpretation consistency. ICGFA video recording for computational development may require dedicated protocols.

Evaluating clinical near-infrared surgical camera systems with a view to optimizing operator and computational signal analysis.

Significance: As clinical evidence on the colorectal application of indocyanine green (ICG) perfusion angiography accrues, there is also interest in computerizing decision support. However, user interpretation and software development may be impacted by system factors affecting the displayed near-infrared (NIR) signal. Aim: We aim to assess the impact of camera positioning on the displayed NIR signal across different open and laparoscopic camera systems. Approach: The effects of distance, movement, and target location (center versus periphery) on the displayed fluorescence signal of different systems were measured under electromagnetic stereotactic guidance from an ICG-albumin model and in vivo during surgery. Results: Systems displayed distinct fluorescence performances with variance apparent with scope optical lens configuration (0 deg versus 30 deg), movement, target positioning, and distance. Laparoscopic system readings fitted inverse square function distance-intensity curves with one device and demonstrated a direction dependent sigmoid curve. Laparoscopic cameras presented central targets as brighter than peripheral ones, and laparoscopes with angled optical lens configurations had a diminished field of view. One handheld open system also showed a distance-intensity relationship, whereas the other maintained a consistent signal despite distance, but both presented peripheral targets brighter than central ones. Conclusions: Optimal clinical use and signal computational development requires detailed appreciation of system behaviors.

Clinical application of machine learning and computer vision to indocyanine green quantification for dynamic intraoperative tissue characterisation: how to do it.

INTRODUCTION: Indocyanine green (ICG) quantification and assessment by machine learning (ML) could discriminate tissue types through perfusion characterisation, including delineation of malignancy. Here, we detail the important challenges overcome before effective clinical validation of such capability in a prospective patient series of quantitative fluorescence angiograms regarding primary and secondary colorectal neoplasia. METHODS: ICG perfusion videos from 50 patients (37 with benign (13) and malignant (24) rectal tumours and 13 with colorectal liver metastases) of between 2- and 15-min duration following intravenously administered ICG were formally studied (clinicaltrials.gov: NCT04220242). Video quality with respect to interpretative ML reliability was studied observing practical, technical and technological aspects of fluorescence signal acquisition. Investigated parameters included ICG dosing and administration, distance-intensity fluorescent signal variation, tissue and camera movement (including real-time camera tracking) as well as sampling issues with user-selected digital tissue biopsy. Attenuating strategies for the identified problems were developed, applied and evaluated. ML methods to classify extracted data, including datasets with interrupted time-series lengths with inference simulated data were also evaluated. RESULTS: Definable, remediable challenges arose across both rectal and liver cohorts. Varying ICG dose by tissue type was identified as an important feature of real-time fluorescence quantification. Multi-region sampling within a lesion mitigated representation issues whilst distance-intensity relationships, as well as movement-instability issues, were demonstrated and ameliorated with post-processing techniques including normalisation and smoothing of extracted time-fluorescence curves. ML methods (automated feature extraction and classification) enabled ML algorithms glean excellent pathological categorisation results (AUC-ROC > 0.9, 37 rectal lesions) with imputation proving a robust method of compensation for interrupted time-series data with duration discrepancies. CONCLUSION: Purposeful clinical and data-processing protocols enable powerful pathological characterisation with existing clinical systems. Video analysis as shown can inform iterative and definitive clinical validation studies on how to close the translation gap between research applications and real-world, real-time clinical utility.

Real-time administration of indocyanine green in combination with computer vision and artificial intelligence for the identification and delineation of colorectal liver metastases.

Introduction: Fluorescence guided surgery for the identification of colorectal liver metastases (CRLM) can be better with low specificity and antecedent dosing impracticalities limiting indocyanine green (ICG) usefulness currently. We investigated the application of artificial intelligence methods (AIM) to demonstrate and characterise CLRMs based on dynamic signalling immediately following intraoperative ICG administration. Methods: Twenty-five patients with liver surface lesions (24 CRLM and 1 benign cyst) undergoing open/laparoscopic/robotic procedures were studied. ICG (0.05 mg/kg) was administered with near-infrared recording of fluorescence perfusion. User-selected region-of-interest (ROI) perfusion profiles were generated, milestones relating to ICG inflow/outflow extracted and used to train a machine learning (ML) classifier. 2D heatmaps were constructed in a subset using AIM to depict whole screen imaging based on dynamic tissue-ICG interaction. Fluorescence appearances were also assessed microscopically (using H&E and fresh-frozen preparations) to provide tissue-level explainability of such methods. Results: The ML algorithm correctly classified 97.2 % of CRLM ROIs (n = 132) and all benign lesion ROIs (n = 6) within 90-s of ICG administration following initial mathematical curve analysis identifying ICG inflow/outflow differentials between healthy liver and CRLMs. Time-fluorescence plots extracted for each pixel in 10 lesions enabled creation of 2D characterising heatmaps using flow parameters and through unsupervised ML. Microscopy confirmed statistically less CLRM fluorescence vs adjacent liver (mean ± std deviation signal/area 2.46 ± 9.56 vs 507.43 ± 160.82 respectively p < 0.001) with H&E diminishing ICG signal (n = 4). Conclusion: ML accurately identifies CRLMs from surrounding liver tissue enabling representative 2D mapping of such lesions from their fluorescence perfusion patterns using AIM. This may assist in reducing positive margin rates at metastatectomy and in identifying unexpected/occult malignancies.

Implementation of an on-site simulation programme during COVID-19 and the assessment of its impact on medical students' competence.

BACKGROUND: COVID-19 has greatly impacted medical students' clinical education. This study evaluates the usefulness of a rapidly implemented on-site simulation programme deployed to supplement our disrupted curriculum. METHODS: Students on surgical rotations received 4-hour tutor-led simulated patient sessions (involving mannikins with remote audio-visual observation) respecting hospital and public health protocols. Attitudes were questionnaire-assessed before and after. Independent, blinded, nonacademic clinicians scored students' clinical competencies by observing real patient interactions using the surgical ward assessment tool in a representative sample versus those completing same duration medicine clinical rotations without simulation (Mann-Whitney U testing, p < 0.05 denoting significance) with all students receiving the same surgical e-learning resources and didactic teaching. RESULTS: A total of 220 students underwent simulation training, comprising 96 hours of scheduled direct teaching. Prior to commencement, 15 students (7% of 191 completing the survey) admitted anxiety, mainly due to clinical inexperience, with only two (1%) anxious re on-site spreading/contracting of COVID-19. A total of 66 students (30%, 38 females and 29 graduate entrants) underwent formal competency assessment by clinicians from ten specialties at two clinical sites. Those who received simulation training (n = 35) were judged significantly better at history taking (p = 0.004) and test ordering (p = 0.01) but not clinical examination, patient drug chart assessment, or differential diagnosis formulation. Of 75 students providing subsequent feedback, 88% stated simulation beneficial (notably for history taking and physical examination skills in 63%) with 83% advocating for more. CONCLUSION: Our rapidly implemented simulation programme for undergraduate medical students helped mitigate pandemic restrictions, enabling improved competence despite necessarily reduced clinical activity encouraging further development.

Deconstructing mastery in colorectal fluorescence angiography interpretation.

INTRODUCTION: Indocyanine green fluorescence angiography (ICGFA) is commonly used in colorectal anastomotic practice with limited pre-training. Recent work has shown that there is considerable inconsistency in signal interpretation between surgeons with minimal or no experience versus those consciously invested in mastery of the technique. Here, we deconstruct the fluorescence signal patterns of expert-annotated surgical ICGFA videos to understand better their correlation and combine this with structured interviews to ascertain whether such interpretative capability is conscious or unconscious. METHODS: For fluorescence signal analysis, expert-annotated ICGFA videos (n = 24) were quantitatively interrogated using a boutique intensity tracker (IBM Research) to generate signal time plots. Such fluorescence intensity data were examined for inter-observer correlation (Intraclass Correlation Coefficients, ICC) at specific curve milestones: the maximum fluorescence signal (Fmax), the times to both achieve this maximum (Tmax), as well as half this maximum (T1/2max) and the ratio between these (T1/2/Tmax). Formal tele-interview with contributing experts (n = 6) was conducted with the narrative transcripts being thematically mapped, plotted, and qualitatively analyzed. RESULTS: Correlation by mathematical measures was excellent (ICC0.9-1.0) for Fmax, Tmax, and T1/2max (0.95, 0.938, and 0.925, respectively) and moderate (0.5-0.75) for T1/2/Tmax (0.729). While all experts narrated a deliberate viewing strategy, their specific dynamic signal appreciation differed in the manner of description. CONCLUSION: Expert ICGFA users demonstrate high correlation in mathematical measures of their signal interpretation although do so tacitly. Computational quantification of expert behavior can help develop the necessary lexicon and training sets as well as computer vision methodology to better exploit ICGFA technology.

Impact of intra-abdominal insufflation pressure on gas leakage occurring during laparoscopy.

INTRODUCTION: The advent of the COVID-19 pandemic led to recommendations aimed at minimizing the risk of gas leaks at laparoscopy. As this has continuing relevance including regarding operating room pollution, we empirically quantified carbon dioxide (CO2) leak jet velocity (important for particle propulsion) occurring with different instruments inserted into differing trocars repeated across a range of intra-abdominal pressures (IAPs) and modern insufflators in an experimental model. METHOD: Laparoscopic gas plume leak velocity (metres/second) was computationally enumerated from schlieren optical flow videography on a porcine cadaveric laparoscopic model with IAPs of 4-5, 7-8, 12-15 and 24-25 mmHg (repeated with 5 different insufflators) during simulated operative use of laparoscopic clip appliers, scissors, energy device, camera and staplers as well as Veres needle (positive control) and trocar obturator (negative control) in fresh 5 mm and 12 mm ports. RESULTS: Close-fitting solid instruments (i.e. cameras and obturators) demonstrated slower gas leak velocities in both the 5 mm and 12 mm ports (p = 0.02 and less than 0.001) when compared to slimmer instruments, however, hollow instrument designs were seen to defy this pattern with the endoscopic linear stapler visibly inducing multiple rapid jests even when compared to similarly sized clip appliers (p = 0.03). However, on a per device basis the operating instrumentation displayed plume speeds which did not vary significantly when challenged with varying post size, IAP and a range of insufflators. CONCLUSION: In general, surgeon's selection of instrument, port or pressure does not usefully mitigate trocar CO2 leak velocity. Instead better trocar design is needed, helped by a fuller understanding of trocar valve mechanics via computational fluid dynamics informed by relevant surgical modelling.

Evaluation of inter-user variability in indocyanine green fluorescence angiography to assess gastric conduit perfusion in esophageal cancer surgery.

Indocyanine Green Fluorescence Angiography (ICGFA) has been deployed to tackle malperfusion-related anastomotic complications. This study assesses variations in operator interpretation of pre-anastomotic ICGFA inflow in the gastric conduit. Utilizing an innovative online interactive multimedia platform (Mindstamp), esophageal surgeons completed a baseline opinion-practice questionnaire and proceeded to interpret, and then digitally assign, a distal transection point on 8 ICGFA videos of esophageal resections (6 Ivor Lewis, 2 McKeown). Annotations regarding gastric conduit transection by ICGFA were compared between expert users versus non-expert participants using ImageJ to delineate longitudinal distances with Shapiro Wilk and t-tests to ascertain significance. Expert versus non-expert correlation was assessed via Intraclass Correlation Coefficients (ICC). Thirty participants (13 consultants, 6 ICGFA experts) completed the study in all aspects. Of these, a high majority (29 participants) stated ICGFA should be used routinely with most (21, including 5/6 experts) stating that 11-50 cases were needed for competency in interpretation. Among users, there were wide variations in dosing (0.05-3 mg/kg) and practice impact. Agreement regarding ICGFA video interpretation concerning transection level among experts was 'moderate' (ICC = 0.717) overall but 'good' (ICC = 0.871) among seven videos with Leave One Out (LOO) exclusion of the video with highest disagreement. Agreement among non-experts was moderate (ICC = 0.641) overall and in every subgroup including among consultants (ICC = 0.626). Experts choose levels that preserved more gastric conduit length versus non-experts in all but one video (P = 0.02). Considerable variability exists with ICGFA interpretation and indeed impact. Even adept users may be challenged in specific cases. Standardized training and/or computerized quantitative fluorescence may help better usage.

Digital dynamic discrimination of primary colorectal cancer using systemic indocyanine green with near-infrared endoscopy.

As indocyanine green (ICG) with near-infrared (NIR) endoscopy enhances real-time intraoperative tissue microperfusion appreciation, it may also dynamically reveal neoplasia distinctively from normal tissue especially with video software fluorescence analysis. Colorectal tumours of patients were imaged mucosally following ICG administration (0.25 mg/kg i.v.) using an endo-laparoscopic NIR system (PINPOINT Endoscopic Fluorescence System, Stryker) including immediate, continuous in situ visualization of rectal lesions transanally for up to 20 min. Spot and dynamic temporal fluorescence intensities (FI) were quantified using ImageJ (including videos at one frame/second, fps) and by a bespoke MATLAB® application that provided digitalized video tracking and signal logging at 30fps (Fluorescence Tracker App downloadable via MATLAB® file exchange). Statistical analysis of FI-time plots compared tumours (benign and malignant) against control during FI curve rise, peak and decline from apex. Early kinetic FI signal measurement delineated discriminative temporal signatures from tumours (n = 20, 9 cancers) offering rich data for analysis versus delayed spot measurement (n = 10 cancers). Malignant lesion dynamic curves peaked significantly later with a shallower gradient than normal tissue while benign lesions showed significantly greater and faster intensity drop from apex versus cancer. Automated tracker quantification efficiently expanded manual results and provided algorithmic KNN clustering. Photobleaching appeared clinically irrelevant. Analysis of a continuous stream of intraoperatively acquired early ICG fluorescence data can act as an in situ tumour-identifier with greater detail than later snapshot observation alone. Software quantification of such kinetic signatures may distinguish invasive from non-invasive neoplasia with potential for real-time in silico diagnosis.

Inter-user variation in the interpretation of near infrared perfusion imaging using indocyanine green in colorectal surgery.

INTRODUCTION: Despite increasing endorsement of near-infrared perfusion assessment using indocyanine green (ICG) during colorectal surgery, little work has yet been done regarding learning curve and interobserver variation most especially on surgical video reflective of real-world usage. METHODS: Surgeons with established expertise in ICG usage were invited to participate in the study along with others without such experience including trainees. All participants completed an opinion questionnaire and interpreted video presentations of fluorescence angiograms in a variety of colorectal case scenarios. An interactive video platform (Mindstamp) enabled dynamic annotation. Statistical analysis of data was performed using Kruskal-Wallis and Mann-Whitney testing as well as Intraclass Correlation Coefficients and Fleiss Multi-rater Kappa Scoring. RESULTS: Forty participants (six experts) completed questionnaire data and provided judgement of 14 videos (nine showing proximal colonic transection site perfusion, four showing completed anastomoses and one an acutely strangulated bowel). 70% felt > 10 cases were needed for competency in use with the majority of experts advocating > 50 (p < 0.05). Overall agreement among experts was "good" for videos showing colonic transection perfusion (versus "moderate" among in-experts) with experts clustering more distally. In contrast, there was no interpretation concordance among experts or in-experts when judging ICG perfusion sufficiency on a yes/no basis. CONCLUSION: Significant experience is recommended before reliance on ICG perfusion angiograms. ICG fluorescence assessment is prone to variable interpretation and influenced by experience and, perhaps, knowledge of preassessment operative steps suggesting a role for objective flow analysis with artificial intelligence methods as the next phase of this technology.

Practical Perfusion Quantification in Multispectral Endoscopic Video: Using the Minutes after ICG Administration to Assess Tissue Pathology.

The wide availability of near infrared light sources in interventional medical imaging stacks enables non-invasive quantification of perfusion by using fluorescent dyes, typically Indocyanine Green (ICG). Due to their often leaky and chaotic vasculatures, intravenously administered ICG perfuses through cancerous tissues differently. We investigate here how a few characteristic values derived from the time series of fluorescence can be used in simple machine learning algorithms to distinguish benign lesions from cancers. These features capture the initial uptake of ICG in the colon, its peak fluorescence, and its early wash-out. By using simple, explainable algorithms we demonstrate, in clinical cases, that sensitivity (specificity) rates of over 95% (95%) for cancer classification can be achieved.