Prediction of burn healing time using artificial neural networks and reflectance spectrometer

Sensor technology and machine learning to guide clinical decision making in plastic surgery

Subjective clinical evaluations are deeply rooted in medical practice. Recent advances in sensor technology facilitate the acquisition of extensive amounts of objective physiological data that can serve as a surrogate for subjective assessments. Along with sensor technology, a branch of artificial intelligence, known as machine learning, has provided decisive advances in several areas of medicine due to its pattern recognition and outcome prediction abilities. The assimilation of machine learning algorithms into sensor technology can substantially improve our current diagnostic and treatment competencies. This review explores available data on the use of sensor technology and machine learning in areas of interest for plastic surgeons, updates current knowledge on the most recent technological advances, and provides a new perspective on the field.

Applications of Artificial Intelligence in Facial Plastic and Reconstructive Surgery: A Narrative Review

Importance: Artificial intelligence (AI) has made invaluable contributions to the technologic advancements across many fields. It is transforming health care and may have a role in improving patient outcomes in facial plastic and reconstructive surgery (FPRS). Observations: In recent years, new automated approaches to simulating and analyzing outcomes using AI have emerged. Advances in rhinoplasty, facelifts, orthognathic surgery, facial reanimation, and preoperative consultation are currently being developed in FPRS. Conclusions and Relevance: Applications of AI have been applied to assist facial plastic surgeons in the preoperative stage, intraoperative planning process, and objective assessment of postoperative outcomes. The application of AI provides avenues to improve postoperative outcomes, while also optimizing patient care.

Beauty Re-defined: A Comparative Analysis of Artificial Intelligence-Generated Ideals and Traditional Standards

Background Traditional methods for assessing facial beauty rely on subjective measures like averages or "golden ratios." However, artificial intelligence (AI) offers a data-driven approach to analyzing attractiveness. This study explores how AI-generated beauty criteria compare to established ideals, considering cultural influences and the evolving concept of beauty. Methods To explore how AI-generated beauty ideals compare to traditional standards, we used three AI text-to-image generation tools (Dezgo (Dezgo SAS LLC, France), Freepik (FreePik Company, Malaga, Spain), and ImagineArt (Vyro, Islamabad, Pakistan)) to create images from a specific prompt. The first four generated images for each gender that met our criteria were included in this study. A single researcher used MediaPipe Studio software to identify ten key facial landmarks on each image. Landmark distances were measured twice in Adobe Photoshop 2023 (Adobe, San Jose, California, United States) and averaged for each measurement. The average values were then used to calculate 23 facial proportion ratios based on established neoclassical canons and golden facial ratios. We then compared these AI-generated ratios to the ideal values using one-sample t-tests in IBM SPSS Statistics for Windows, Version 29 (Released 2023; IBM Corp., Armonk, New York, United States), p < 0.05 significance, to assess alignment with traditional beauty standards. Results AI-generated faces displayed statistically significant differences, p < 0.05, from established beauty standards in both neoclassical canons and golden ratios for both males and females. Differences were seen in facial width, upper and lower face proportions, and eye symmetry. Conclusion AI-generated faces deviated from traditional beauty standards of neoclassical canons and golden ratios for both genders. This suggests AI incorporates factors beyond established ideals, potentially reflecting contemporary preferences, cultural biases, or emerging trends.

Advancement of Generative Pre-trained Transformer Chatbots in Answering Clinical Questions in the Practical Rhinoplasty Guideline

BACKGROUND

The Generative Pre-trained Transformer (GPT) series, which includes ChatGPT, is an artificial large language model that provides human-like text dialogue. This study aimed to evaluate the performance of artificial intelligence chatbots in answering clinical questions based on practical rhinoplasty guidelines.

METHODS

Clinical questions (CQs) developed from the guidelines were used as question sources. For each question, we asked GPT-4 and GPT-3.5 (ChatGPT), developed by OpenAI, to provide answers for the CQs, Policy Level, Aggregate Evidence Quality, Level of Confidence in Evidence, and References. We compared the performance of the two types of artificial intelligence (AI) chatbots.

RESULTS

A total of 10 questions were included in the final analysis, and the AI chatbots correctly answered 90.0% of these. GPT-4 demonstrated a lower accuracy rate than GPT-3.5 in answering CQs, although without statistically significant difference (86.0% vs. 94.0%; p = 0.05), whereas GPT-4 showed significantly higher accuracy for the level of confidence in Evidence than GPT-3.5 (52.0% vs. 28.0%; p < 0.01). No statistical differences were observed in Policy Level, Aggregate Evidence Quality, and Reference Match. In addition, GPT-4 rated significantly higher in presenting existing references than GPT-3.5 (36.9% vs. 24.1%; p = 0.01).

CONCLUSIONS

The overall performance of GPT-4 was similar to that of GPT-3.5. However, GPT-4 provided existing references at a higher rate than GPT-3.5. GPT-4 has the potential to provide a more accurate reference in professional fields, including rhinoplasty.

LEVEL OF EVIDENCE V

This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266 .

Risk predictions of surgical wound complications based on a machine learning algorithm: A systematic review

Surgical wounds may arise due to harm inflicted upon soft tissue during surgical intervention, and many complications and injuries may accompany them. These complications can lead to prolonged hospitalization and poorer clinical outcomes. Also, Machine learning (ML) is a Section of artificial intelligence (AI) that has emerged in medical care and is increasingly used for diagnosis, complications, prognosis and recurrence prediction. This study aims to investigate surgical wound risk predictions and management using a ML algorithm by R programming language analysis. The systematic review, following PRISMA guidelines, spanned electronic databases using search terms like 'machine learning', 'surgical' and 'wound'. Inclusion criteria covered experimental studies from 1990 to the present on ML's application in surgical wound evaluation. Exclusion criteria included studies lacking full text, focusing on ML in all surgeries, neglecting wound assessment and duplications. Two authors rigorously assessed titles, abstracts and full texts, excluding reviews and guidelines. Ultimately, relevant articles were then analysed. The present study identified nine articles employing ML for surgical wound management. The analysis encompassed various surgical procedures, including Cardiothoracic, Caesarean total abdominal colectomy, Burn plastic surgery, facial plastic surgery, laparotomy, minimal invasive surgery, hernia repair and unspecified surgeries. ML was skillful in evaluating surgical site infections (SSI) in seven studies, while two extended its use to burn-grade diagnosis and wound classification. Support Vector Machine (SVM) and Convolutional Neural Network (CNN) were the most utilized algorithms. ANN achieved a 96% accuracy in facial plastic surgery wound management. CNN demonstrated commendable accuracies in various surgeries, and SVM exhibited high accuracy in multiple surgeries and burn plastic surgery. In sum, these findings underscore ML's potential for significant improvements in postoperative management and the development of enhanced care techniques, particularly in surgical wound management.

Artificial Intelligence: Singularity Approaches

SUMMARY

Artificial intelligence (AI) has been a disruptive technology within health care, from the development of simple care algorithms to complex deep-learning models. AI has the potential to reduce the burden of administrative tasks, advance clinical decision-making, and improve patient outcomes. Unlocking the full potential of AI requires the analysis of vast quantities of clinical information. Although AI holds tremendous promise, widespread adoption within plastic surgery remains limited. Understanding the basics is essential for plastic surgeons to evaluate the potential uses of AI. This review provides an introduction of AI, including the history of AI, key concepts, applications of AI in plastic surgery, and future implications.

Past, present, and future of global research on artificial intelligence applications in dermatology: A bibliometric analysis

In recent decades, artificial intelligence (AI) has played an increasingly important role in medicine, including dermatology. Worldwide, numerous studies have reported on AI applications in dermatology, rapidly increasing interest in this field. However, no bibliometric studies have been conducted to evaluate the past, present, or future of this topic. This study aimed to illustrate past and present research and outline future directions for global research on AI applications in dermatology using bibliometric analysis. We conducted an online search of the Web of Science Core Collection database to identify scientific papers on AI applications in dermatology. The bibliometric metadata of each selected paper were extracted, analyzed, and visualized using VOS viewer and Cite Space. A total of 406 papers, comprising 8 randomized controlled trials and 20 prospective studies, were deemed eligible for inclusion. The United States had the highest number of papers (n = 166). The University of California System (n = 24) and Allan C. Halpern (n = 11) were the institution and author with the highest number of papers, respectively. Based on keyword co-occurrence analysis, the studies were categorized into 9 distinct clusters, with clusters 2, 3, and 7 containing keywords with the latest average publication year. Wound progression prediction using machine learning, the integration of AI into teledermatology, and applications of the algorithms in skin diseases, are the current research priorities and will remain future research aims in this field.

Understanding the Impact of Aging on Attractiveness Using a Machine Learning Model of Facial Age Progression

Background: Advances in machine learning age progression technology offer the unique opportunity to better understand the public's perception on the aging face. Objective: To compare how observers perceive attractiveness and traditional gender traits in faces created with a machine learning model. Methods: Eight surveys were developed, each with 10 sets of photographs that were progressively aged with a machine learning model. Respondents rated attractiveness and masculinity or femininity of each photograph using a sliding scale (range: 0-100). Mean attractiveness scores were calculated and compared between men and women as well as between age groups. Results: A total of 315 respondents (51% men, 49% women) completed the survey. Accuracy of the facial age progression model was 85%. Females were considered significantly less attractive (-10.43, p < 0.01) and less feminine (-7.59, p < 0.01) per decade with the greatest drop over age 40 years. Male attractiveness and masculinity were relatively preserved until age 50 years where attractiveness scores were significantly lower (-5.45, p = 0.39). Conclusions: In this study, observers were found to perceive attractiveness at older ages differently between men and women.

The Use of Machine Learning for Predicting Complications of Free-Flap Head and Neck Reconstruction

BACKGROUND

Machine learning has been increasingly used for surgical outcome prediction, yet applications in head and neck reconstruction are not well-described. In this study, we developed and evaluated the performance of ML algorithms in predicting postoperative complications in head and neck free-flap reconstruction.

METHODS

We conducted a comprehensive review of patients who underwent microvascular head and neck reconstruction between January 2005 and December 2018. Data were used to develop and evaluate nine supervised ML algorithms in predicting overall complications, major recipient-site complication, and total flap loss.

RESULTS

We identified 4000 patients who met inclusion criteria. Overall, 33.7% of patients experienced a complication, 26.5% experienced a major recipient-site complication, and 1.7% suffered total flap loss. The k-nearest neighbors algorithm demonstrated the best overall performance for predicting any complication (AUROC = 0.61, sensitivity = 0.60). Regularized regression had the best performance for predicting major recipient-site complications (AUROC = 0.68, sensitivity = 0.66), and decision trees were the best predictors of total flap loss (AUROC = 0.66, sensitivity = 0.50).

CONCLUSIONS

ML accurately identified patients at risk of experiencing postsurgical complications, including total flap loss. Predictions from ML models may provide insight in the perioperative setting and facilitate shared decision making.

Artificial intelligence in the management and treatment of burns: A systematic review and meta-analyses

INTRODUCTION AND AIM

Artificial Intelligence (AI) is already being successfully employed to aid the interpretation of multiple facets of burns care. In the light of the growing influence of AI, this systematic review and diagnostic test accuracy meta-analyses aim to appraise and summarise the current direction of research in this field.

METHOD

A systematic literature review was conducted of relevant studies published between 1990 and 2021, yielding 35 studies. Twelve studies were suitable for a Diagnostic Test Meta-Analyses.

RESULTS

The studies generally focussed on burn depth (Accuracy 68.9%-95.4%, Sensitivity 90.8% and Specificity 84.4%), burn segmentation (Accuracy 76.0%-99.4%, Sensitivity 97.9% and specificity 97.6%) and burn related mortality (Accuracy >90%-97.5% Sensitivity 92.9% and specificity 93.4%). Neural networks were the most common machine learning (ML) algorithm utilised in 69% of the studies. The QUADAS-2 tool identified significant heterogeneity between studies.

DISCUSSION

The potential application of AI in the management of burns patients is promising, especially given its propitious results across a spectrum of dimensions, including burn depth, size, mortality, related sepsis and acute kidney injuries. The accuracy of the results analysed within this study is comparable to current practices in burns care.

CONCLUSION

The application of AI in the treatment and management of burns patients, as a series of point of care diagnostic adjuncts, is promising. Whilst AI is a potentially valuable tool, a full evaluation of its current utility and potential is limited by significant variations in research methodology and reporting.

Clinical application of functional near-infrared spectroscopy for burn assessment

Significance: Early assessment of local tissue oxygen saturation is essential for clinicians to determine the burn wound severity. Background: We assessed the burn extent and depth in the skin of the extremities using a custom-built 36-channel functional near-infrared spectroscopy system in patients with burns. Methods: A total of nine patients with burns were analyzed in this study. All second-degree burns were categorized as superficial, intermediate, and deep burns; non-burned skin on the burned side; and healthy skin on the contralateral non-burned side. Hemodynamic tissue signals from functional near-infrared spectroscopy attached to the burn site were measured during fNIRS using a blood pressure cuff. A nerve conduction study was conducted to check for nerve damage. Results: All second-degree burns were categorized into superficial, intermediate, and deep burns; non-burned skin on the burned side and healthy skin on the contralateral non-burned side showed a significant difference distinguishable using functional near-infrared spectroscopy. Hemodynamic measurements using functional near-infrared spectroscopy were more consistent with the diagnosis of burns 1 week later than that of the degree of burns diagnosed visually at the time of admission. Conclusion: Functional near-infrared spectroscopy may help with the early judgment of burn extent and depth by reflecting differences in the oxygen saturation levels in the skin.

Simulation and Artificial Intelligence in Rhinoplasty: A Systematic Review

BACKGROUND

Rhinoplasty is one of the most popular cosmetic procedures. The complexity of the nasal structure and the substantial aesthetic and functional impact of the operation make rhinoplasty very challenging. The past few years have witnessed an increasing implementation of artificial intelligence (AI) and simulation systems into plastic surgery practice. This review explores the potential uses of AI and simulation models in rhinoplasty.

METHODS

Five electronic databases were searched: PubMed, CINAHL, EMBASE, Scopus, and Web of Science. We used the Preferred Reporting Items for Systematic Reviews and Meta-Analysis as our basis of organization.

RESULTS

Several simulation models were described to predict the nasal shape that aesthetically matches the patient's face, indicate the implant size in augmentation rhinoplasty and construct three-dimensional (3D) facial images from two-dimensional images. Machine learning was used to learn surgeons' rhinoplasty styles and accurately simulate the outcomes. Deep learning was used to predict rhinoplasty status accurately and analyze the factors associated with increased facial attractiveness after rhinoplasty. Finally, a deep learning model was used to predict patients' age before and after rhinoplasty proving that the procedure made the patients look younger.

CONCLUSION

3D simulation models and AI models can revolutionalize the practice of functional and aesthetic rhinoplasty. Simulation systems can be beneficial in preoperative planning, intra-operative decision making, and postoperative evaluation. In addition, AI models can be trained to carry out tasks that are either challenging or time-consuming for surgeons.

LEVEL OF EVIDENCE III

This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266 .

The Use of Artificial Intelligence and Machine Learning in Surgery: A Comprehensive Literature Review

Interest in the use of artificial intelligence (AI) and machine learning (ML) in medicine has grown exponentially over the last few years. With its ability to enhance speed, precision, and efficiency, AI has immense potential, especially in the field of surgery. This article aims to provide a comprehensive literature review of artificial intelligence as it applies to surgery and discuss practical examples, current applications, and challenges to the adoption of this technology. Furthermore, we elaborate on the utility of natural language processing and computer vision in improving surgical outcomes, research, and patient care.

Artificial intelligence in the management and treatment of burns: a systematic review

BACKGROUND

Artificial intelligence (AI) is an innovative field with potential for improving burn care. This article provides an updated review on machine learning in burn care and discusses future challenges and the role of healthcare professionals in the successful implementation of AI technologies.

METHODS

A systematic search was carried out on MEDLINE, Embase and PubMed databases for English-language articles studying machine learning in burns. Articles were reviewed quantitatively and qualitatively for clinical applications, key features, algorithms, outcomes and validation methods.

RESULTS

A total of 46 observational studies were included for review. Assessment of burn depth (n = 26), support vector machines (n = 19) and 10-fold cross-validation (n = 11) were the most common application, algorithm and validation tool used, respectively.

CONCLUSION

AI should be incorporated into clinical practice as an adjunct to the experienced burns provider once direct comparative analysis to current gold standards outlining its benefits and risks have been studied. Future considerations must include the development of a burn-specific common framework. Authors should use common validation tools to allow for effective comparisons. Level I/II evidence is required to produce robust proof about clinical and economic impacts.

Evaluation of non-invasive bioforensic techniques for determining the age of hot-iron brand burn scars in cattle

Hot-iron branding is a traditional form of permanent cattle identification in the United States. There is a need for science-based determination of cattle brand age. Near infrared reflectance spectroscopy (NIRS) has been used to obtain information about animal tissues and healing processes. Height-width allometry and NIRS were applied to hot-iron cattle brand scars to determine if either or both of these methods can be used to non-invasively establish the interval sincethe application of hot-iron cattle brands. Length and width of a brand routinely applied to calves (~30–60 d old) were established and then the same measurements were recorded on 378 calfhood branded cattle of known age ranging from 0.5 to > 6.5 yr-of-age. Brand width and height increased over the original measurements by > 100% between calfhood application and 2.5 yr-of-age (P < 0.001). Brand size did not change dramatically between 2.5 and > 6.5 yr, however, both width and height were (P < 0.05) greater at maturity than at weaning. Near infrared spectra were collected from a) branded skin b) non-clipped (hair), non-branded skin, and c) hair clipped, non-branded skin on Bos taurus cross calves. Individual trial calibrations yielded high R² and low SE of calibration values as well as similar cross validation performance (P < 0.001). Numerically lower but still strong performance (P < 0.001) resulted from combined data set calibrations. Cross-trial prediction of brand age was unsuccessful. One single year calibration underpredicted (P < 0.001) brand age of an independent validation set by 2.83 d, and another single year calibration underpredicted (P < 0.001) the same validation set by 9.91 d. When combined, these two datasets resulted in a calibration that overpredicted brand age in the validation set by 6.9 d (P < 0.02). Discriminant analyses for identification of skin surface type yielded success rates of 90% for branded, 99% for non-clipped, non-branded, and 96% for clipped, non-branded (P < 0.01). Discriminant analyses were also performed on samples grouped into a) less than 33 d, b) 141–153 d, and c) 169 d categories. All group membership identifications were successful at greater than 90% (P < 0.01). Preliminary results indicate that brand size could be used to indicate brand age and that NIRS can predict brand age as well as discriminate between broad brand age groups in cattle. More work will need to be done before these techniques can be used in real-world forensic applications.

Machine Learning Demonstrates High Accuracy for Disease Diagnosis and Prognosis in Plastic Surgery

INTRODUCTION

Machine learning (ML) is a set of models and methods that can detect patterns in vast amounts of data and use this information to perform various kinds of decision-making under uncertain conditions. This review explores the current role of this technology in plastic surgery by outlining the applications in clinical practice, diagnostic and prognostic accuracies, and proposed future direction for clinical applications and research.

METHODS

EMBASE, MEDLINE, CENTRAL and ClinicalTrials.gov were searched from 1990 to 2020. Any clinical studies (including case reports) which present the diagnostic and prognostic accuracies of machine learning models in the clinical setting of plastic surgery were included. Data collected were clinical indication, model utilised, reported accuracies, and comparison with clinical evaluation.

RESULTS

The database identified 1181 articles, of which 51 articles were included in this review. The clinical utility of these algorithms was to assist clinicians in diagnosis prediction (n=22), outcome prediction (n=21) and pre-operative planning (n=8). The mean accuracy is 88.80%, 86.11% and 80.28% respectively. The most commonly used models were neural networks (n=31), support vector machines (n=13), decision trees/random forests (n=10) and logistic regression (n=9).

CONCLUSIONS

ML has demonstrated high accuracies in diagnosis and prognostication of burn patients, congenital or acquired facial deformities, and in cosmetic surgery. There are no studies comparing ML to clinician's performance. Future research can be enhanced using larger datasets or utilising data augmentation, employing novel deep learning models, and applying these to other subspecialties of plastic surgery.

An Artificial Intelligence Tool for Image Simulation in Rhinoplasty

During rhinoplasty consultations, surgeons typically create a computer simulation of the expected result. An artificial intelligence model (AIM) can learn a surgeon's style and criteria and generate the simulation automatically. The objective of this study is to determine if an AIM is capable of imitating a surgeon's criteria to generate simulated images of an aesthetic rhinoplasty surgery. This is a cross-sectional survey study of resident and specialist doctors in otolaryngology conducted in the month of November 2019 during a rhinoplasty conference. Sequential images of rhinoplasty simulations created by a surgeon and by an AIM were shown at random. Participants used a seven-point Likert scale to evaluate their level of agreement with the simulation images they were shown, with 1 indicating total disagreement and 7 total agreement. Ninety-seven of 122 doctors agreed to participate in the survey. The median level of agreement between the participant and the surgeon was 6 (interquartile range or IQR 5-7); between the participant and the AIM it was 5 (IQR 4-6), p-value < 0.0001. The evaluators were in total or partial agreement with the results of the AIM's simulation 68.4% of the time (95% confidence interval or CI 64.9-71.7). They were in total or partial agreement with the surgeon's simulation 77.3% of the time (95% CI 74.2-80.3). An AIM can emulate a surgeon's aesthetic criteria to generate a computer-simulated image of rhinoplasty. This can allow patients to have a realistic approximation of the possible results of a rhinoplasty ahead of an in-person consultation. The level of evidence of the study is 4.

Predicting Discharge Disposition Following Meningioma Resection Using a Multi-Institutional Natural Language Processing Model

BACKGROUND

Machine learning (ML)-based predictive models are increasingly common in neurosurgery, but typically require large databases of discrete variables for training. Natural language processing (NLP) can extract meaningful data from unstructured text.

OBJECTIVE

To present an NLP model that predicts nonhome discharge and a point-of-care implementation.

METHODS

We retrospectively collected age, preoperative notes, and radiology reports from 595 adults who underwent meningioma resection in an academic center from 1995 to 2015. A total of 32 algorithms were trained with the data; the 3 best performing algorithms were combined to form an ensemble. Predictive ability, assessed by area under the receiver operating characteristic curve (AUC) and calibration, was compared to a previously published model utilizing 52 neurosurgeon-selected variables. We then built a multi-institutional model by incorporating notes from 693 patients at another center into algorithm training. Permutation importance was used to analyze the relative importance of each input to model performance. Word clouds and non-negative matrix factorization were used to analyze predictive features of text.

RESULTS

The single-institution NLP model predicted nonhome discharge with AUC of 0.80 (95% CI = 0.74-0.86) on internal and 0.76 on holdout validation compared to AUC of 0.77 (95% CI = 0.73-0.81) and 0.74 for the 52-variable ensemble. The multi-institutional model performed similarly well with AUC = 0.78 (95% CI = 0.74-0.81) on internal and 0.76 on holdout validation. Preoperative notes most influenced predictions. The model is available at http://nlp-home.insds.org.

CONCLUSION

ML and NLP are underutilized in neurosurgery. Here, we construct a multi-institutional NLP model that predicts nonhome discharge.

Making the Most of Big Data in Plastic Surgery: Improving Outcomes, Protecting Patients, Informing Service Providers

ABSTRACT

In medicine, "big data" refers to the interdisciplinary analysis of high-volume, diverse clinical and lifestyle information on large patient populations. Recent advancements in data storage and electronic record keeping have enabled the expansion of research in this field. In the United Kingdom, Big data has been highlighted as one of the government's "8 Great Technologies," and the Medical Research Council has invested more than £100 million since 2012 in developing the Health Data Research UK infrastructure. The recent Royal College of Surgeons Commission of the Future of Surgery concluded that analysis of big data is one of the 4 most likely avenues to bring some of the most innovative changes to surgical practice in the 21st century.In this article, we provide an overview of the nascent field of big data analytics in plastic and highlight how it has the potential to improve outcomes, increase safety, and aid service planning.We outline the current resources available, the emerging role of big data within the subspecialties of burns, microsurgery, skin and breast cancer, and how these data can be used. We critically review the limitations and considerations raised with big data, offer suggestions regarding database optimization, and suggest future directions for research in this exciting field.

Artificial Intelligence in Plastic Surgery: Current Applications, Future Directions, and Ethical Implications

Background:

Artificial intelligence (AI) in healthcare delivery has become an important area of research due to the rapid progression of technology, which has allowed the growth of many processes historically reliant upon human input. AI has become particularly important in plastic surgery in a variety of settings. This article highlights current applications of AI in plastic surgery and discusses future implications. We further detail ethical issues that may arise in the implementation of AI in plastic surgery.

Methods:

We conducted a systematic literature review of all electronically available publications in the PubMed, Scopus, and Web of Science databases as of February 5, 2020. All returned publications regarding the application of AI in plastic surgery were considered for inclusion.

Results:

Of the 89 novel articles returned, 14 satisfied inclusion and exclusion criteria. Articles procured from the references of those of the database search and those pertaining to historical and ethical implications were summarized when relevant.

Conclusions:

Numerous applications of AI exist in plastic surgery. Big data, machine learning, deep learning, natural language processing, and facial recognition are examples of AI-based technology that plastic surgeons may utilize to advance their surgical practice. Like any evolving technology, however, the use of AI in healthcare raises important ethical issues, including patient autonomy and informed consent, confidentiality, and appropriate data use. Such considerations are significant, as high ethical standards are key to appropriate and longstanding implementation of AI.

Artificial Intelligence in plastic surgery: What is it? Where are we now? What is on the horizon?

INTRODUCTION

An increasing quantity of data is required to guide precision medicine and advance future healthcare practices, but current analytical methods often become overwhelmed. Artificial intelligence (AI) provides a promising solution. Plastic surgery is an innovative surgical specialty expected to implement AI into current and future practices. It is important for all plastic surgeons to understand how AI may affect current and future practice, and to recognise its potential limitations.

METHODS

Peer-reviewed published literature and online content were comprehensively reviewed. We report current applications of AI in plastic surgery and possible future applications based on published literature and continuing scientific studies, and detail its potential limitations and ethical considerations.

FINDINGS

Current machine learning models using convolutional neural networks can evaluate breast mammography and differentiate benign and malignant tumours as accurately as specialist doctors, and motion sensor surgical instruments can collate real-time data to advise intraoperative technical adjustments. Centralised big data portals are expected to collate large datasets to accelerate understanding of disease pathogeneses and best practices. Information obtained using computer vision could guide intraoperative surgical decisions in unprecedented detail and semi-autonomous surgical systems guided by AI algorithms may enable improved surgical outcomes in low- and middle-income countries. Surgeons must collaborate with computer scientists to ensure that AI algorithms inform clinically relevant health objectives and are interpretable. Ethical concerns such as systematic biases causing non-representative conclusions for under-represented patient groups, patient confidentiality and the limitations of AI based on the quality of data input suggests that AI will accompany the plastic surgeon, rather than replace them.

Use of Machine Learning for Prediction of Patient Risk of Postoperative Complications After Liver, Pancreatic, and Colorectal Surgery

BACKGROUND

Surgical resection is the only potentially curative treatment for patients with colorectal, liver, and pancreatic cancers. Although these procedures are performed with low mortality, rates of complications remain relatively high following hepatopancreatic and colorectal surgery.

METHODS

The American College of Surgeons (ACS) National Surgical Quality Improvement Program was utilized to identify patients undergoing liver, pancreatic and colorectal surgery from 2014 to 2016. Decision tree models were utilized to predict the occurrence of any complication, as well as specific complications. To assess the variability of the performance of the classification trees, bootstrapping was performed on 50% of the sample.

RESULTS

Algorithms were derived from a total of 15,657 patients who met inclusion criteria. The algorithm had a good predictive ability for the occurrence of any complication, with a C-statistic of 0.74, outperforming the ASA (C-statistic 0.58) and ACS-Surgical Risk Calculator (C-statistic 0.71). The algorithm was able to predict with high accuracy thirteen out of the seventeen complications analyzed. The best performance was in the prediction of stroke (C-statistic 0.98), followed by wound dehiscence, cardiac arrest, and progressive renal failure (all C-statistic 0.96). The algorithm had a good predictive ability for superficial SSI (C-statistic 0.76), organ space SSI (C-statistic 0.76), sepsis (C-statistic 0.79), and bleeding requiring transfusion (C-statistic 0.79).

CONCLUSION

Machine learning was used to develop an algorithm that accurately predicted patient risk of developing complications following liver, pancreatic, or colorectal surgery. The algorithm had very good predictive ability to predict specific complications and demonstrated superiority over other established methods.

Prevalence of Machine Learning in Craniofacial Surgery

Machine learning (ML) revolves around the concept of using experience to teach computer-based programs to reliably perform specific tasks. Healthcare setting is an ideal environment for adaptation of ML applications given the multiple specific tasks that could be allocated to computer programs to perform. There have been several scoping reviews published in literature looking at the general acceptance and adaptability of surgical specialities to ML applications, but very few focusing on the application towards craniofacial surgery. This study aims to present a detailed scoping review regarding the use of ML applications in craniofacial surgery.

The applications of machine learning in plastic and reconstructive surgery: protocol of a systematic review

BACKGROUND

Machine learning, a subset of artificial intelligence, is a set of models and methods that can automatically detect patterns in vast amounts of data, extract information and use it to perform various kinds of decision-making under uncertain conditions. This can assist surgeons in clinical decision-making by identifying patient cohorts that will benefit from surgery prior to treatment. The aim of this review is to evaluate the applications of machine learning in plastic and reconstructive surgery.

METHODS

A literature review will be undertaken of EMBASE, MEDLINE and CENTRAL (1990 up to September 2019) to identify studies relevant for the review. Studies in which machine learning has been employed in the clinical setting of plastic surgery will be included. Primary outcomes will be the evaluation of the accuracy of machine learning models in predicting a clinical diagnosis and post-surgical outcomes. Secondary outcomes will include a cost analysis of those models. This protocol has been prepared using the Preferred Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) guidelines.

DISCUSSION

This will be the first systematic review in available literature that summarises the published work on the applications of machine learning in plastic surgery. Our findings will provide the basis of future research in developing artificial intelligence interventions in the specialty.

SYSTEMATIC REVIEW REGISTRATION

PROSPERO CRD42019140924.

Artificial Intelligence in Global Health -A Framework and Strategy for Adoption and Sustainability

Artificial Intelligence (AI) applications in medicine have grown considerably in recent years. AI in the forms of Machine Learning, Natural Language Processing, Expert Systems, Planning and Logistics methods, and Image Processing networks provide great analytical aptitude. While AI methods were first conceptualized for radiology, investigations today are established across all medical specialties. The necessity for proper infrastructure, skilled labor, and access to large, well-organized data sets has kept the majority of medical AI applications in higher-income countries. However, critical technological improvements, such as cloud computing and the near-ubiquity of smartphones, have paved the way for use of medical AI applications in resource-poor areas. Global health initiatives (GHI) have already begun to explore ways to leverage medical AI technologies to detect and mitigate public health inequities. For example, AI tools can help optimize vaccine delivery and community healthcare worker routes, thus enabling limited resources to have a maximal impact. Other promising AI tools have demonstrated an ability to: predict burn healing time from smartphone photos; track regions of socioeconomic disparity combined with environmental trends to predict communicable disease outbreaks; and accurately predict pregnancy complications such as birth asphyxia in low resource settings with limited patient clinical data. In this commentary, we discuss the current state of AI-driven GHI and explore relevant lessons from past technology-centered GHI. Additionally, we propose a conceptual framework to guide the development of sustainable strategies for AI-driven GHI, and we outline areas for future research.

Identifying Ear Abnormality from 2D Photographs Using Convolutional Neural Networks

Quantifying ear deformity using linear measurements and mathematical modeling is difficult due to the ear's complex shape. Machine learning techniques, such as convolutional neural networks (CNNs), are well-suited for this role. CNNs are deep learning methods capable of finding complex patterns from medical images, automatically building solution models capable of machine diagnosis. In this study, we applied CNN to automatically identify ear deformity from 2D photographs. Institutional review board (IRB) approval was obtained for this retrospective study to train and test the CNNs. Photographs of patients with and without ear deformity were obtained as standard of care in our photography studio. Profile photographs were obtained for one or both ears. A total of 671 profile pictures were used in this study including: 457 photographs of patients with ear deformity and 214 photographs of patients with normal ears. Photographs were cropped to the ear boundary and randomly divided into training (60%), validation (20%), and testing (20%) datasets. We modified the softmax classifier in the last layer in GoogLeNet, a deep CNN, to generate an ear deformity detection model in Matlab. All images were deemed of high quality and usable for training and testing. It took about 2 hours to train the system and the training accuracy reached almost 100%. The test accuracy was about 94.1%. We demonstrate that deep learning has a great potential in identifying ear deformity. These machine learning techniques hold the promise in being used in the future to evaluate treatment outcomes.

Predicting Mortality in the Surgical Intensive Care Unit Using Artificial Intelligence and Natural Language Processing of Physician Documentation

The purpose of this study was to use natural language processing of physician documentation to predict mortality in patients admitted to the surgical intensive care unit (SICU). The Multiparameter Intelligent Monitoring in Intensive Care III database was used to obtain SICU stays with six different severity of illness scores. Natural language processing was performed on the physician notes. Classifiers for predicting mortality were created. One classifier used only the physician notes, one used only the severity of illness scores, and one used the physician notes with severity of injury scores. There were 3838 SICU stays identified during the study period and 5.4 per cent ended with mortality. The classifier trained with physician notes with severity of injury scores performed with the highest area under the curve (0.88 ± 0.05) and accuracy (94.6 ± 1.1%). The most important variable was the Oxford Acute Severity of Illness Score (16.0%). The most important terms were “dilated” (4.3%) and “hemorrhage” (3.7%). This study demonstrates the novel use of artificial intelligence to process physician documentation to predict mortality in the SICU. The classifiers were able to detect the subtle nuances in physician vernacular that predict mortality. These nuances provided improved performance in predicting mortality over physiologic parameters alone.

A systematic review on the quality of measurement techniques for the assessment of burn wound depth or healing potential

PURPOSE

Reliable and valid assessment of burn wound depth or healing potential is essential to treatment decision-making, to provide a prognosis, and to compare studies evaluating different treatment modalities. The aim of this review was to critically appraise, compare and summarize the quality of relevant measurement properties of techniques that aim to assess burn wound depth or healing potential.

METHODS

A systematic literature search was performed using PubMed, EMBASE and Cochrane Library. Two reviewers independently evaluated the methodological quality of included articles using an adapted version of the Consensus-based Standards for the selection of health Measurement INstruments (COSMIN) checklist. A synthesis of evidence was performed to rate the measurement properties for each technique and to draw an overall conclusion on quality of the techniques.

RESULTS

Thirty-six articles were included, evaluating various techniques, classified as (1) laser Doppler techniques; (2) thermography or thermal imaging; (3) other measurement techniques. Strong evidence was found for adequate construct validity of laser Doppler imaging (LDI). Moderate evidence was found for adequate construct validity of thermography, videomicroscopy, and spatial frequency domain imaging (SFDI). Only two studies reported on the measurement property reliability. Furthermore, considerable variation was observed among comparator instruments.

CONCLUSIONS

Considering the evidence available, it appears that LDI is currently the most favorable technique; thereby assessing burn wound healing potential. Additional research is needed into thermography, videomicroscopy, and SFDI to evaluate their full potential. Future studies should focus on reliability and measurement error, and provide a precise description of which construct is aimed to measure.

Using artificial intelligence to predict prolonged mechanical ventilation and tracheostomy placement

BACKGROUND

Early identification of critically ill patients who will require prolonged mechanical ventilation (PMV) has proven to be difficult. The purpose of this study was to use machine learning to identify patients at risk for PMV and tracheostomy placement.

MATERIALS AND METHODS

The Multiparameter Intelligent Monitoring in Intensive Care III database was queried for all intensive care unit (ICU) stays with mechanical ventilation. PMV was defined as ventilation >7 d. Classifiers with a gradient-boosted decision trees algorithm were created for the outcomes of PMV and tracheostomy placement. The variables used were six different severity-of-illness scores calculated on the first day of ICU admission including their components and 30 comorbidities. Mean receiver operating characteristic curves were calculated for the outcomes, and variable importance was quantified.

RESULTS

There were 20,262 ICU stays identified. PMV was required in 13.6%, and tracheostomy was performed in 6.6% of patients. The classifier for predicting PMV was able to achieve a mean area under the curve (AUC) of 0.820 ± 0.016, and tracheostomy was predicted with an AUC of 0.830 ± 0.011. There were 60.7% patients admitted to a surgical ICU, and the classifiers for these patients predicted PMV with an AUC of 0.852 ± 0.017 and tracheostomy with an AUC of 0.869 ± 0.015. The variable with the highest importance for predicting PMV was the logistic organ dysfunction score pulmonary component (13%), and the most important comorbidity in predicting tracheostomy was cardiac arrhythmia (12%).

CONCLUSIONS

This study demonstrates the use of artificial intelligence through machine-learning classifiers for the early identification of patients at risk for PMV and tracheostomy. Application of these identification techniques could lead to improved outcomes by allowing for early intervention.

Predicting central line-associated bloodstream infections and mortality using supervised machine learning

PURPOSE

The purpose of this study was to compare machine learning techniques for predicting central line-associated bloodstream infection (CLABSI).

MATERIALS AND METHODS

The Multiparameter Intelligent Monitoring in Intensive Care III database was queried for all ICU admissions. The variables included six different severities of illness scores calculated on the first day of ICU admission with their components and comorbidities. The outcomes of interest were in-hospital mortality, central line placement, and CLABSI. Predictive models were created for these outcomes using classifiers with different algorithms: logistic regression, gradient boosted trees, and deep learning.

RESULTS

There were 57,786 total hospital admissions and the mortality rate was 10.1%. There were 38.4% patients with a central line and the rate of CLABSI was 1.5%. The classifiers using deep learning performed with the highest AUC for mortality, 0.885±0.010 (p<0.01) and central line placement, 0.816±0.006 (p<0.01). The classifier using logistic regression for predicting CLABSI performed with an AUC of 0.722±0.048 (p<0.01).

CONCLUSIONS

This study demonstrates models for identifying patients who will develop CLABSI. Early identification of these patients has implications for quality, cost, and outcome improvements.

Machine learning to identify multigland disease in primary hyperparathyroidism

BACKGROUND

20%-25% of patients with primary hyperparathyroidism will have multigland disease (MGD). Preoperatative imaging can be inaccurate or unnecessary in MGD. Identification of MGD could direct the need for imaging and inform operative approach. The purpose of this study is to use machine learning (ML) methods to predict MGD.

METHODS

Retrospective review of a prospective database. The ML platform, Waikato Environment for Knowledge Analysis, was used, and we selected models for (1) overall accuracy and (2) preferential identification of MGD. A review of imaging studies was performed on a cohort predicted to have MGD.

RESULTS

2010 patients met inclusion criteria: 1532 patients had single adenoma (SA) (76%) and 478 had MGD (24%). After testing many algorithms, we selected two different models for potential integration as clinical decision-support tools. The best overall accuracy was achieved using a boosted tree classifier, RandomTree: 94.1% accuracy; 94.1% sensitivity, 83.8% specificity, 94.1% positive predictive value, and 0.984 area under the receiver operating characteristics curve. To maximize positive predictive value of MGD prediction, a rule-based classifier, JRip, with cost-sensitive learning was used and achieved 100% positive predictive value for MGD. Imaging reviewed from the cohort of 34 patients predicted to have MGD by the cost-sensitive model revealed 39 total studies performed: 28 sestamibi scans and 11 ultrasounds. Only 8 (29%) sestamibi scans and 4 (36%) ultrasounds were correct.

CONCLUSIONS

ML methods can help distinguish MGD early in the clinical evaluation of primary hyperparathyroidism, guiding further workup and surgical planning.

Spectrophotometric intracutaneous analysis for the assessment of burn wounds: A service evaluation of its clinical application in 50 burn wounds

INTRODUCTION

The assessment of burn depth can be challenging even to the experienced burn clinician. Clinical assessment is most widely used to determine burn depth. Because of this subjective nature, various imaging modalities have been invented. The use of photospectometry as a novel technique in burn wound depth analysis has been previously described but the literature is very limited.

METHODOLOGY

We carried out a single blinded non-randomized comparative study of healing potential of 50 burn wounds between tissue spectrophotometry analysis versus clinical evaluation.

RESULTS

ScanOSkin™ technology has an overall sensitivity of 75% and specificity of 86% in predicting healing potential of wounds. Analysis of Inter Rater Agreement (IRA) using Kappa calculations showed strengths of agreement varied from fair to moderate in perfusion and burn depth. IRA for assessing pigmentation however, was poor and this was reflected in user feedback.

CONCLUSION

There is a potential role for ScanOSkin™ tissue spectrophotometric analysis in burn depth assessment. Future studies comparing several imaging modalities with ScanOSkin^®, taking into account costs comparison may be useful for future health resources planning.

Imaging Techniques for Clinical Burn Assessment with a Focus on Multispectral Imaging

Significance: Burn assessments, including extent and severity, are some of the most critical diagnoses in burn care, and many recently developed imaging techniques may have the potential to improve the accuracy of these evaluations. Recent Advances: Optical devices, telemedicine, and high-frequency ultrasound are among the highlights in recent burn imaging advancements. We present another promising technology, multispectral imaging (MSI), which also has the potential to impact current medical practice in burn care, among a variety of other specialties. Critical Issues: At this time, it is still a matter of debate as to why there is no consensus on the use of technology to assist burn assessments in the United States. Fortunately, the availability of techniques does not appear to be a limitation. However, the selection of appropriate imaging technology to augment the provision of burn care can be difficult for clinicians to navigate. There are many technologies available, but a comprehensive review summarizing the tissue characteristics measured by each technology in light of aiding clinicians in selecting the proper device is missing. This would be especially valuable for the nonburn specialists who encounter burn injuries. Future Directions: The questions of when burn assessment devices are useful to the burn team, how the various imaging devices work, and where the various burn imaging technologies fit into the spectrum of burn care will continue to be addressed. Technologies that can image a large surface area quickly, such as thermography or laser speckle imaging, may be suitable for initial burn assessment and triage. In the setting of presurgical planning, ultrasound or optical microscopy techniques, including optical coherence tomography, may prove useful. MSI, which actually has origins in burn care, may ultimately meet a high number of requirements for burn assessment in routine clinical use.

Burn depth analysis using multidimensional scaling applied to psychophysical experiment data

In this paper a psychophysical experiment and a multidimensional scaling (MDS) analysis are undergone to determine the physical characteristics that physicians employ to diagnose a burn depth. Subsequently, these characteristics are translated into mathematical features, correlated with these physical characteristics analysis. Finally, a study to verify the ability of these mathematical features to classify burns is performed. In this study, a space with axes correlated with the MDS axes has been developed. 74 images have been represented in this space and a k-nearest neighbor classifier has been used to classify these 74 images. A success rate of 66.2% was obtained when classifying burns into three burn depths and a success rate of 83.8% was obtained when burns were classified as those which needed grafts and those which did not. Additional studies have been performed comparing our system with a principal component analysis and a support vector machine classifier. Results validate the ability of the mathematical features extracted from the psychophysical experiment to classify burns into their depths. In addition, the method has been compared with another state-of-the-art method and the same database.

Improving wound score classification with limited remission spectra

The classification of wounds into healing states depending on their absorption spectrum of visible and near infrared light remains an important task in dermatology. Moreover, a reduction of the spectrum that is used in the classification task to fewer but important wavelengths is desirable, as each measured wavelength increases the examination costs without necessarily providing further information to the classification of wound healing states. This paper addresses two aspects: First the improvement of the classification of wounds into healing states and second, a cost reduction by choosing only important wavelengths. Standard Data Mining methods are evaluated for their classification accuracy (CA) and compared to their performance when applying feature selection techniques that are used to reduce the amount of necessary wavelengths. The results indicate that the 1-nearest-neighbor approach (IB1 algorithm) comes up with the best CA, while only relying on a fraction (4%) of the standard wavelength spectrum.

Beneficial Effects of Chlorella-11 Peptide on Blocking LPS-Induced Macrophage Activation and Alleviating Thermal Injury-Induced Inflammation in Rats

Chlorella possesses various remarkable biological activities. One component, Val-Glu-Cys-Tyr-Gly-Pro-Asn-Arg-Pro-Gln-Phe ( Chlorella-11 peptide) was found to be able to suppress LPS-induced NO production and inflammation. However, the molecular mechanism behind these findings and the consistency between in vitro and in vivo data have not been investigated. LPS-activated RAW 264.7 macrophages were used to study in vitro molecular anti-inflammatory effects of Chlorella-11 peptide. After activation, NO production and the expression of iNOS and NF-κB proteins as well as iNOS mRNA were measured using Griess colorimetric assay, Western blotting and RT-PCR, respectively. Alterations in PGE2 and TNF-α contents were also monitored by ELISA. For in vivo studies, thermal injury Wistar rats were used and inflammatory indications e.g. serum malondialdehyde (MDA), TNF-α levels and skin erythema were evaluated 48 h after injury implementation. In vitro results showed that Chlorella-11 peptide produced a dose- and time-dependent inhibition on NO production. The effective inhibition could remain for at least 6 h after LPS activation. It was also found that the expression of LPS-induced iNOS mRNA, iNOS and NF-κB proteins were diminished by the peptide treatment. Concurrently, the levels on TNF-α and PGE2 production after LPS activation were also inhibited. These findings are in agreement with the in vivo data that animal serum MDA and TNF-α levels and skin erythema in rats were considerably reduced compared to the control group (saline-treated). The significance of this study sheds light on the effectiveness of Chlorella-11 peptide in preventing inflammation progression in vitro and in vivo and its potential for clinical applications.

Noninvasive assessment of diabetic foot ulcers with diffuse photon density wave methodology: pilot human study

A pilot human study is conducted to evaluate the potential of using diffuse photon density wave (DPDW) methodology at near-infrared (NIR) wavelengths (685 to 830 nm) to monitor changes in tissue hemoglobin concentration in diabetic foot ulcers. Hemoglobin concentration is measured by DPDW in 12 human wounds for a period ranging from 10 to 61 weeks. In all wounds that healed completely, gradual decreases in optical absorption coefficient, oxygenated hemoglobin concentration, and total hemoglobin concentration are observed between the first and last measurements. In nonhealing wounds, the rates of change of these properties are nearly zero or slightly positive, and a statistically significant difference (p<0.05) is observed in the rates of change between healing and nonhealing wounds. Differences in the variability of DPDW measurements over time are observed between healing and nonhealing wounds, and this variance may also be a useful indicator of nonhealing wounds. Our results demonstrate that DPDW methodology with a frequency domain NIR device can differentiate healing from nonhealing diabetic foot ulcers, and indicate that it may have clinical utility in the evaluation of wound healing potential.

Enhanced surgical imaging: laparoscopic vessel identification and assessment of tissue oxygenation

BACKGROUND

Inherent to minimally invasive procedures are loss of tactile feedback and loss of three-dimensional assessment. Tasks such as vessel identification and dissection are not trivial for the inexperienced laparoscopic surgeon. Advanced surgical imaging, such as 3-charge-coupled device (3-CCD) image enhancement, can be used to assist with these more challenging tasks and, in addition, offers a method to noninvasively monitor tissue oxygenation during operations.

STUDY DESIGN

In this study, 3-CCD image enhancement is used for identification of vessels in 25 laparoscopic donor and partial nephrectomy patients. The algorithm is then applied to two laparoscopic nephrectomy patients involving multiple renal arteries. We also use the 3-CCD camera to qualitatively monitor renal parenchymal oxygenation during 10 laparoscopic donor nephrectomies (LDNs).

RESULTS

The mean region of interest (ROI) intensity values obtained for the renal artery and vein (68.40 +/- 8.44 and 45.96 +/- 8.65, respectively) are used to calculate a threshold intensity value (59.00) that allows for objective vessel differentiation. In addition, we examined the renal parenchyma during LDNs. Mean ROI intensity values were calculated for the renal parenchyma at two distinct time points: before vessel stapling (nonischemic) and just before extraction from the abdomen (ischemic). The nonischemic mean ROI intensity values are statistically different from the ischemic mean ROI intensity values (p < 0.05), even with short ischemia times.

CONCLUSIONS

We have developed a technique, 3-CCD image enhancement, for identification of vasculature and monitoring of parenchymal oxygenation. This technique requires no additional laparoscopic operating room equipment and has real-time video capability.