Artificial Intelligence in Surgery: The Future is Now.

Background Clinical Artificial intelligence (AI) has reached a critical inflection point. Advances in algorithmic science and increased understanding of operational considerations in AI deployment are opening the door to widespread clinical pathway transformation. For surgery in particular, the application of machine learning algorithms in fields such as computer vision and operative robotics are poised to radically change how we screen, diagnose, risk-stratify, treat and follow-up patients, in both pre- and post-operative stages, and within operating theatres. Summary In this paper, we summarise the current landscape of existing and emerging integrations within complex surgical care pathways. We investigate effective methods for practical use of AI throughout the patient pathway, from early screening and accurate diagnosis to intraoperative robotics, post-operative monitoring and follow-up. Horizon scanning of AI technologies in surgery is used to identify novel innovations that can enhance surgical practice today, with potential for paradigm shifts across core domains of surgical practice in the future. Any AI-driven future must be built on responsible and ethical usage, reinforced by effective oversight of data governance, and of risks to patient safety in deployment. Implementation is additionally bound to considerations of usability and pathway feasibility, and the need for robust healthcare technology assessment and evidence generation. While these factors are traditionally seen as barriers to translating AI into practice, we discuss how holistic implementation practices can create a solid foundation for scaling AI across pathways. Key Messages The next decade will see rapid translation of experimental development into real-world impact. AI will require evolution of work practices, but will also enhance patient safety, enhance surgical quality outcomes, and provide significant value for surgeons and health systems. Surgical practice has always sat on a bedrock of technological innovation. For those that follow this tradition, the future of AI in surgery starts now.

Harnessing Machine Learning to Personalize Web-Based Health Care Content.

Web-based health care content has emerged as a primary source for patients to access health information without direct guidance from health care providers. The benefit of this approach is dependent on the ability of patients to access engaging high-quality information, but significant variability in the quality of web-based information often forces patients to navigate large quantities of inaccurate, incomplete, irrelevant, or inaccessible content. Personalization positions the patient at the center of health care models by considering their needs, preferences, goals, and values. However, the traditional methods used thus far in health care to determine the factors of high-quality content for a particular user are insufficient. Machine learning (ML) uses algorithms to process and uncover patterns within large volumes of data to develop predictive models that automatically improve over time. The health care sector has lagged behind other industries in implementing ML to analyze user and content features, which can automate personalized content recommendations on a mass scale. With the advent of big data in health care, which builds comprehensive patient profiles drawn from several disparate sources, ML can be used to integrate structured and unstructured data from users and content to deliver content that is predicted to be effective and engaging for patients. This enables patients to engage in their health and support education, self-management, and positive behavior change as well as to enhance clinical outcomes.

Nance-Horan Syndrome-like 1 protein negatively regulates Scar/WAVE-Arp2/3 activity and inhibits lamellipodia stability and cell migration.

Cell migration is important for development and its aberrant regulation contributes to many diseases. The Scar/WAVE complex is essential for Arp2/3 mediated lamellipodia formation during mesenchymal cell migration and several coinciding signals activate it. However, so far, no direct negative regulators are known. Here we identify Nance-Horan Syndrome-like 1 protein (NHSL1) as a direct binding partner of the Scar/WAVE complex, which co-localise at protruding lamellipodia. This interaction is mediated by the Abi SH3 domain and two binding sites in NHSL1. Furthermore, active Rac binds to NHSL1 at two regions that mediate leading edge targeting of NHSL1. Surprisingly, NHSL1 inhibits cell migration through its interaction with the Scar/WAVE complex. Mechanistically, NHSL1 may reduce cell migration efficiency by impeding Arp2/3 activity, as measured in cells using a Arp2/3 FRET-FLIM biosensor, resulting in reduced F-actin density of lamellipodia, and consequently impairing the stability of lamellipodia protrusions.

Electrical stimulation as a therapeutic approach in obstructive sleep apnea - a meta-analysis.

PURPOSE: Electrical stimulation of the upper airway dilator muscles is an emerging treatment for obstructive sleep apnea (OSA). Invasive hypoglossal nerve stimulation (HNS) has been accepted as treatment alternative to continuous positive airway pressure (CPAP) for selected patients, while transcutaneous electrical stimulation (TES) of the upper airway is being investigated as non-invasive alternative. METHODS: A meta-analysis (CRD42017074674) on the effects of both HNS and TES on the apnea-hypopnea index (AHI) and the Epworth Sleepiness Scale (ESS) in OSA was conducted including published evidence up to May 2018. Random-effects models were used. Heterogeneity and between-study variance were assessed by I2 and τ2, respectively. RESULTS: Of 41 identified clinical trials, 20 interventional trials (n = 895) could be pooled in a meta-analysis (15 HNS [n = 808], 5 TES [n = 87]). Middle-aged (mean ± SD 56.9 ± 5.5 years) and overweight (body mass index 29.1 ± 1.5 kg/m2) patients with severe OSA (AHI 37.5 ± 7.0/h) were followed-up for 6.9 ± 4.0 months (HNS) and 0.2 ± 0.4 months (TES), respectively. The AHI improved by - 24.9 h-1 [95%CI - 28.5, - 21.2] in HNS (χ2 79%, I2 82%) and by - 16.5 h-1 [95%CI - 25.1, - 7.8] in TES (χ2 7%, I2 43%; both p < 0.001). The ESS was reduced by - 5.0 (95%CI - 5.9, - 4.1) (p < 0.001). CONCLUSION: Both invasive and transcutaneous electrical stimulation reduce OSA severity by a clinically relevant margin. HNS results in a clinically relevant improvement of symptoms. While HNS represents an invasive treatment for selected patients with moderate to severe OSA, TES should be further investigated as potential non-invasive approach for OSA.

A methodologic assessment of lymphedema clinical practice guidelines.

OBJECTIVE: The objective of this review was to determine the methodologic quality of current lymphedema clinical practice guidelines (CPGs) to assist health care professionals in selecting accessible, high-quality guidance and to identify areas for improvement in future CPGs. METHODS: MEDLINE, Embase, online CPG databases, and reference lists of included guidelines were searched up to January 31, 2020. Full-text CPGs reporting on evidence-based recommendations in lymphedema diagnosis or management in English were included. CPGs based on expert consensus, CPG summaries, or CPGs that were not freely available were excluded. Two reviewers identified eligible CPGs, extracted data, and assessed their quality independently using the Appraisal of Guidelines for Research and Evaluation II instrument. Significant scoring discrepancies were discussed with a third reviewer. An overall scaled quality score of ≥80% was the threshold to recommend guideline use. RESULTS: Six relevant CPGs were identified. One was subsequently excluded as its full text could not be obtained. Overall, there was very good inter-reviewer reliability of scores with intraclass correlation coefficient of 0.952 (95% confidence interval, 0.921-0.974). No single CPG scored highest in all domains, with methodologic heterogeneity observed. Poor performance was noted in domain 5 (mean scaled score, 23.8% ± 17.1%) and domain 6 (22.9% ± 26.7%). No CPG achieved an overall scaled quality score of ≥80%, with the top CPG scoring 79.2%. CONCLUSIONS: According to the defined threshold, no lymphedema CPG was considered adequate for use in clinical practice. All current lymphedema CPGs have areas for improvement with elements of methodologic quality lacking, particularly with respect to rigor of development. A structured approach, guided by the use of CPG creation tools and checklists such as the Appraisal of Guidelines for Research and Evaluation II instrument, should help CPG development groups in improving the quality of future CPGs. This is of particular importance in a complex, multidisciplinary condition such as lymphedema.

Development of a technical checklist for the assessment of suturing in robotic surgery.

BACKGROUND: With the increased use of simulation for surgical training, there is a need for objective forms of assessment to evaluate trainees. The Global Evaluative Assessment of Robotic Skills (GEARS) is widely used for assessing skills in robotic surgery, but there are no recognised checklist scoring systems. This study aimed to develop a checklist for suturing in robotic surgery. METHODS: A suturing checklist for needle driving and knot tying was constructed following evaluation of participants performing urethrovesical anastomoses. Key procedural steps were identified from expert videos, while assessing novice videos allowed identification of common technical errors. 22 novice and 13 expert videos were marked on needle driving, while 18 novices and 10 experts were assessed on knot tying. Validation of the finalised checklist was performed with the assessment of 39 separate novices by an expert surgeon and compared to GEARS scoring. RESULTS: The internal consistency of the preliminary checklist was high (Cronbach's alpha = 0.870 for needle driving items; 0.736 for knot tying items), and after removal of poorly correlating items, the final checklist contained 23 steps. Both the needle driving and knot tying categories discriminated between novices and experts, p < 0.005. While the GEARS score demonstrated construct validity for needle driving, it could not significantly differentiate between novices and experts for knot tying, p = 0.286. The needle driving category significantly correlated with the corresponding GEARS scores (rs = 0.613, p < 0.005), but the correlation for knot tying was insignificant (rs = 0.296, p = 0.127). The pilot data indicates the checklist significantly correlated with the GEARS score (p < 0.005). CONCLUSION: This study reports the development of a valid assessment tool for suturing in robotic surgery. Given that checklists are simple to use, there is significant scope for this checklist to be used in surgical training.