TitleRole of Smart Image Segmentation in Infection Diagnosis and Monitoring of Acute Surgical Wounds

Honey is a viscous, supersaturated sugar solution derived from nectar gathered and modified by the honeybee, Apis mellifera. Honey has been used since ancient times as a remedy in wound care. Evidence from animal studies and some trials has suggested honey may accelerate wound healing. The objective was to determine whether honey increases the rate of healing in acute wounds (burns, lacerations and other traumatic wounds) and chronic wounds (venous ulcers, arterial ulcers, diabetic ulcers, pressure ulcers, infected surgical wounds). We searched the Cochrane Wounds Group Specialised Register (May 2008), CENTRAL (May 2008) and several other electronic databases (May 2008). Bibliographies were searched and manufacturers of dressing products were contacted for unpublished trials. Randomised and quasi randomised trials that evaluated honey as a treatment for any sort of acute or chronic wound were sought. There was no restriction in terms of source, date of publication or language. Wound healing was the primary endpoint. Data from eligible trials were extracted and summarised using a data extraction sheet by one author and independently verified by a second author. 19 trials (n=2554) were identified that met the inclusion criteria. In acute wounds, three trials evaluated the effect of honey in acute lacerations, abrasions or minor surgical wounds and nine trials evaluated the effect the honey in burns. In chronic wounds two trials evaluated the effect of honey in venous leg ulcers and one trial in pressure ulcers, infected post-operative wounds, and Fournier's gangrene respectively. Two trials recruited people with mixed groups of chronic or acute wounds. The poor quality of most of the trial reports means the results should be interpreted with caution, except in venous leg ulcers. In acute wounds, honey may reduce time to healing compared with some conventional dressings in partial thickness burns (WMD -4.68 days, 95%CI -4.28 to -5.09 days). All the included burns trials have originated from a single centre, which may have impact on replicability. In chronic wounds, honey in addition to compression bandaging does not significantly increase healing in venous leg ulcers (RR 1.15, 95%CI 0.96 to 1.38). There is insufficient evidence to determine the effect of honey compared with other treatments for burns or in other acute or chronic wound types. Honey may improve healing times in mild to moderate superficial and partial thickness burns compared with some conventional dressings. Honey dressings as an adjuvant to compression do not significantly increase leg ulcer healing at 12 weeks. There is insufficient evidence to guide clinical practice in other areas.

BackgroundDigital wound monitoring has become increasingly feasible with the widespread use of smartphones and mobile messaging platforms. Although most previous studies have focused on chronic wounds and demonstrated the clinical benefits of remote monitoring, little is known about how patients with acute wounds perceive and report wound-related changes after discharge; these factors may affect the accuracy and reliability of patient-facing digital health systems.ObjectiveThis study aimed to evaluate the diagnostic performance of patient-reported infection symptoms in predicting clinician-initiated callbacks within a chatbot-based wound monitoring platform. A secondary objective was to identify wound features most strongly associated with patient-reported infection and examine differences between acute and chronic wound populations.MethodsThis retrospective observational study was conducted at a tertiary medical center in Taipei, Taiwan, between June 30, 2022, and March 1, 2023, as part of an institutional digital health initiative. Within this program, adults with acute or chronic wounds voluntarily joined a chatbot-based monitoring system deployed through the Line messaging app using a bring-your-own-device model. Participants submitted daily symptom reports and wound photographs through the chatbot interface. For each submission, patient self-report of infection served as the primary predictor variable, while an independent review by a senior plastic surgeon determined the reference standard (callback vs no callback). Logistic regression and generalized estimating equation models were applied to account for within-subject correlation, with covariates including age, sex, and wound type. Analyses were performed separately for acute and chronic wounds.ResultsThis study included 159 patients; 88 (55.3%) had acute wounds and 71 (44.7%) had chronic wounds. Across the study period, 4764 wound photographs were submitted, with a median of 5 (IQR 2-18) photographs per patient. Diagnostic performance differed by wound type. For acute wounds, the area under the receiver operating characteristic curve was 0.702, with 52.6% sensitivity (95% CI 31.7-72.7) and 87.8% specificity (95% CI 84.7-90.3). For chronic wounds, the area under the receiver operating characteristic curve was 0.907, with 94.9% sensitivity (95% CI 93.3-96.2) and 86.4% specificity (95% CI 85.2-87.5). In symptom correlation analyses, redness was significantly associated with patient-reported infection in the acute wound subgroup (odds ratio [OR] 3.94, 95% CI 1.97-7.90; P<.001), whereas in the chronic wound subgroup, both redness (OR 86.35, 95% CI 57.11-130.56; P<.001) and skin darkening (OR 358.55, 95% CI 244.79-525.16; P<.001) showed significant associations (all P<.001).ConclusionsThis study highlights the differences in how patients perceive and report infection-related symptoms between populations with acute and chronic wounds. Lower diagnostic accuracy for acute wounds underscores the influence of limited experience and contextual constraints on patient self-assessment. These findings suggest that patient-facing digital wound monitoring systems should be tailored according to wound chronicity and patient experience, incorporating adaptive feedback and artificial intelligence–assisted screening to enhance patient-reported symptom interpretation.

BackgroundPopulation based surveillance of surgical site infections (SSIs) requires precise case-finding strategies. We sought to develop and validate machine learning models to automate the process of complex (deep incisional/organ space) SSIs case detection.MethodsThis retrospective cohort study included adult patients (age ≥ 18 years) admitted to Calgary, Canada acute care hospitals who underwent primary total elective hip (THA) or knee (TKA) arthroplasty between Jan 1st, 2013 and Aug 31st, 2020. True SSI conditions were judged by the Alberta Health Services Infection Prevention and Control (IPC) program staff. Using the IPC cases as labels, we developed and validated nine XGBoost models to identify deep incisional SSIs, organ space SSIs and complex SSIs using administrative data, electronic medical records (EMR) free text data, and both. The performance of machine learning models was assessed by sensitivity, specificity, positive predictive value, negative predictive value, F1 score, the area under the receiver operating characteristic curve (ROC AUC) and the area under the precision–recall curve (PR AUC). In addition, a bootstrap 95% confidence interval (95% CI) was calculated.ResultsThere were 22,059 unique patients with 27,360 hospital admissions resulting in 88,351 days of hospital stay. This included 16,561 (60.5%) TKA and 10,799 (39.5%) THA procedures. There were 235 ascertained SSIs. Of them, 77 (32.8%) were superficial incisional SSIs, 57 (24.3%) were deep incisional SSIs, and 101 (42.9%) were organ space SSIs. The incidence rates were 0.37 for superficial incisional SSIs, 0.21 for deep incisional SSIs, 0.37 for organ space and 0.58 for complex SSIs per 100 surgical procedures, respectively. The optimal XGBoost models using administrative data and text data combined achieved a ROC AUC of 0.906 (95% CI 0.835–0.978), PR AUC of 0.637 (95% CI 0.528–0.746), and F1 score of 0.79 (0.67–0.90).ConclusionsOur findings suggest machine learning models derived from administrative data and EMR text data achieved high performance and can be used to automate the detection of complex SSIs.

Applying Machine Learning Across Sites: External Validation of a Surgical Site Infection Detection Algorithm

To assess the differences in proteolytic activity of acute and chronic wound environments, wound fluids were collected from acute surgical wounds (22 samples) and chronic wounds (25 samples) of various etiologies, including mixed vessel disease ulcers, decubiti and diabetic foot ulcers. Matrix metalloproteinase (MMP) activity measured using the Azocoll assay was significantly elevated by 30 fold in chronic wounds (median 22.8 microg MMP Eq/ml) compared to acute wounds (median 0.76 microg MMP Eq/ml) (p < 0.001). The addition of the matrix metalloproteinase inhibitor Illomostat decreased the matrix metalloproteinase activity by approximately 90% in all samples, confirming that the majority of the activity measured was due to matrix metalloproteinases. Gelatin zymograms indicated predominantly elevated matrix metalloproteinase-9 with smaller elevations of matrix metalloproteinase-2. In addition tissue inhibitor of metalloproteinase-1 levels were analyzed in a small subset of acute and chronic wounds. When tissue inhibitor of metalloproteinase-1 levels were compared to protease levels there was an inverse correlation (p = 0.02, r = - 0.78). In vitro degradation of epidermal growth factor was measured by addition of 125I labelled epidermal growth factor to acute and chronic wound fluid samples. There was significantly higher degradation of epidermal growth factor in chronic wound fluid samples (mean 28.1%) compared to acute samples (mean 0.6%). This also correlated to the epidermal growth factor activity of these wound fluid samples (p < 0. 001, r = 0.64). Additionally, the levels of proteases were assayed in wound fluid collected from 15 venous leg ulcers during a nonhealing and healing phase using a unique model of chronic wound healing in humans. Patients with nonhealing venous leg ulcers were admitted to the hospital for bed rest and wound fluid samples were collected on admission (nonhealing phase) and after 2 weeks (healing phase) when the ulcers had begun to heal as evidenced by a reduction in size (median 12%). These data showed that the elevated levels of matrix metalloproteinase activity decreased significantly as healing occurs in chronic leg ulcers (p < 0.01). This parallels the processes observed in normally healing acute wounds. This data also supports the case for the addition of protease inhibitors in chronic wounds in conjunction with any treatments using growth factors.

Background and AimsMinimally adherent silver dressings (SILVER MASD) are antimicrobial, nonirritating, provide a moist wound healing environment, and low cost. The purpose of this pilot, single‐center, non‐blinded randomized controlled trial was to quantify the outcomes of acute surgical wounds treated with MASD versus standard of care (SoC) dressings.MethodsThirty‐two patients with acute wounds were randomized 1:1 to be treated with MASD once weekly or SoC following surgical excision of skin and/or subcutaneous tissue between September 13, 2016 and November 28, 2017. The outcome variables included clinical infection, time to wound closure, and pain scores at dressing changes. Two independent, one‐sided sample t‐tests were performed to assess statistical significance.ResultsThere was no difference in wound healing between SILVER MASD and SoC. Dressing changes were less painful for wounds managed with MASD silver dressings.ConclusionsThe results of this study suggest that MASD are not less effective in wound healing compared to SoC while also providing the benefit of decreased pain at dressing changes. Therefore, minimally adherent silver dressings can and should be considered a viable option in the management of acute surgical wounds.

The ability to predict surgical site infections (SSIs) early would be advantageous. Previous studies have investigated the use of inflammatory factors in fluids drained from surgical sites to predict SSI, but the diagnostic efficacy of this method requires improvement. Baseline levels of inflammatory factors vary between individuals, but this variation tends to differ in patients with and without SSIs. Therefore, we standardized subsequently acquired concentrations of interleukin 6 and C-reactive protein in fluids drained from surgical sites by dividing them by the concentrations determined at day 1 to preclude the confounding effects of differences in baseline levels. The standardized concentrations had higher predictive efficacy than the absolute concentrations. Standardizing the data rendered SSI prediction more precise and practical in a diverse group of real patients. This translational study suggests that inflammatory factors in fluid drained from injury sites are promising tools for the prediction of SSI in the clinic.

Honey is a viscous, supersaturated sugar solution derived from nectar gathered and modified by the honeybee, Apis mellifera. Honey has been used since ancient times as a remedy in wound care. Evidence from animal studies and some trials has suggested that honey may accelerate wound healing. The objective was to determine whether honey increases the rate of healing in acute wounds (e.g. burns, lacerations) and chronic wounds (e.g. skin ulcers, infected surgical wounds). For this first update of the review we searched the Cochrane Wounds Group Specialised Register (searched 13 June 2012); The Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2012, Issue 5); Ovid MEDLINE (2008 to May Week 5 2012); Ovid MEDLINE (In-Process & Other Non-Indexed Citations 12 June 2012); Ovid EMBASE (2008 to 2012 Week 23); and EBSCO CINAHL (2008 to 8 June 2012). Randomised and quasi-randomised trials that evaluated honey as a treatment for any sort of acute or chronic wound were sought. There was no restriction in terms of source, date of publication or language. Wound healing was the primary endpoint. Data from eligible trials were extracted and summarised by one review author, using a data extraction sheet, and independently verified by a second review author. We identified 25 trials (with a total of 2987 participants) that met the inclusion criteria, including six new trials that were added to this update. In acute wounds, three trials evaluated the effect of honey in acute lacerations, abrasions or minor surgical wounds and 12 trials evaluated the effect of honey in burns. In chronic wounds, two trials evaluated the effect of honey in venous leg ulcers, and single trials investigated its effect in infected post-operative wounds, pressure injuries, cutaneous Lieshmaniasis, diabetic foot ulcers and Fournier's gangrene. Three trials recruited people into mixed groups of chronic or acute wounds. Most trials were at high or unclear risk of bias. In acute wounds, specifically partial-thickness burns, honey might reduce time to healing compared with some conventional dressings (WMD -4.68 days, 95%CI -4.28 to -5.09 days), but, when compared with early excision and grafting, honey delays healing in partial- and full-thickness burns (WMD 13.6 days, 95% CI 10.02 to 17.18 days). In chronic wounds, honey does not significantly increase healing in venous leg ulcers when used as an adjuvant to compression (RR 1.15, 95% CI 0.96 to 1.38), and may delay healing in cutaneous Leishmaniasis when used as an adjuvant to meglumine antimoniate compared to meglumine antimoniate alone (RR 0.72, 95% CI 0.51 to 1.01). Honey dressings do not increase rates of healing significantly in venous leg ulcers when used as an adjuvant to compression. Honey may delay healing in partial- and full-thickness burns in comparison to early excision and grafting, and in cutaneous Leishmaniasis when used as an adjuvant with meglumine antimoniate. Honey might be superior to some conventional dressing materials, but there is considerable uncertainty about the replicability and applicability of this evidence. There is insufficient evidence to guide clinical practice in other types of wounds, and purchasers should refrain from providing honey dressings for routine use until sufficient evidence of effect is available.

The authors report the use of nanocrystalline silver (Acticoat 7, Smith and Nephew, London, UK) in an acute surgical wound to prevent localized skin necrosis due to infection, thereby avoiding skin grafting as a secondary procedure. Two patients were successfully treated with Acticoat 7 dressings without using systemic antimicrobials after developing methicillin-resistant Staphylococcus aureus infection in the surgical site. Despite a history of smoking and incision through scar tissues, the wound did not progress into deep infection nor was there recurrence of infection at 2 years follow-up. The intention was to use this particular dressing to assess the effect of silver on infected keratinocytes in an acute wound environment. It is possible that the use of Acticoat 7 may reduce the bacterial loading at the wound site, thereby decreasing a propensity for skin necrosis caused by the infective process. This case report demonstrates that the acute surgical wound with impending cutaneous necrosis due to localized infection may be treated without oral antimicrobials.

Activated platelets release a rich broth of growth factors involved in wound healing. One way to deliver activated platelets to wounds is in the form of platelet-rich plasma (PRP) harvested by centrifuging the patient's venous blood after activating the platelets with collagen or calcium chloride and/or autologous thrombin, then delicately removing the supernatant, called platelet-poor plasma (PPP). Platelet-rich plasma is usually injected into the lesion and/or applied topically, then sealed in or over the wound using a moisture-retentive dressing. Platelet-rich plasma (often with PPP) has been applied at different times, depths, and frequencies to chronic and acute wounds using various PRP doses and vehicles to achieve widely differing results. Meta-analyses have reported that PRP improved healing rates of open diabetic foot ulcers and venous ulcers and may reduce pain and surgical site infection (SSI) incidence in open and closed acute surgical wounds. However, inconsistency in study methods and outcome measures limited consistency of pain and SSI results. No consistent effect on healing or deep SSI rates was reported as a result of adding 1 intraoperative dose of PRP in the surgical site before closing elective foot and ankle surgery incisions of 250 patients as compared with 250 similar patients receiving the same procedure without PRP. After decades of research, ideal parameters of PRP delivery and use on each type of wound remain unclear for improving SSI, acute wound pain, and healing outcomes. This installment of the Evidence Corner reviews 2 surgical studies that may provide clues about optimal PRP use. One triple-blind randomized clinical trial (RCT) focused on irrigation of freshly closed carpal ligament surgical incisions with PRP as compared with PPP. Another non-blind RCT explored the effect of injecting PRP into open pilonidal sinus excisions 4 days and 12 days after surgery.

Chronic wounds are a challenge to treatment. In this retrospective study, the effect of transdermal CO2 application on wound healing in chronic ulcers was investigated and compared to the effect of CO2 on acute surgical wounds. Eighty-six patients (52 females and 34 males) with chronic wounds of different origin except arterial occlusive disease were included. In addition, 17 patients (5 females, 12 males) with wide excision wounds after surgical therapy of acne inversa were considered. The indication for CO2 application was a wound at risk for infection. Treatment was performed with a Carboflow device once daily for 30 to 60 minutes. There was clinical evidence of improvement of granulation and reduction of discharge and malodor within 1 week of treatment in both chronic and acute wounds. Only 9 patients, all diabetics, needed an additional systemic antibiosis. The treatment was well tolerated. No adverse effects have been noted. Transdermal CO2 application is a useful method to reduce the risk of infection and improve wound healing in both chronic and certain acute wounds. Systematic prospective trials are needed.

Background The use of Artificial Intelligence (AI) in healthcare has the potential to transform medical diagnosis, including the detection of surgical site infections (SSI). Real-time monitoring of surgical sites and early detection of infection through AI can lead to prompt treatment, improving patient outcomes and reducing healthcare costs and spread of infections in healthcare facilities. Aim This study aimed to determine the use of AI in the early detection of surgical site infections and its impact on patient recovery. Materials and methods The study analysed patients admitted to the hospital over the past year with SSIs. A novel AI device was used to image the wound and detect presence of infection through fluorescence lights that identify bacteria through bioluminescence. Targeted tissue cultures were taken, antibiotics were administered. Some wounds required targeted debridement. Daily dressings were done followed by secondary suturing. Results A total of 34 patients were included in the study, with a mean age of 52.4 years and 14 female and 20 male patients. 25 patients reported pus from the wound, 9 had serous discharge, and 19 had slough present. In 85.2% of patients, the AI device's gram stain for bacteria in the wound was similar to the tissue culture report. The mean time for wound healing was 39.52 days. Conclusion Our study suggests that AI devices can accurately identify the gram stain of bacteria in a wound, leading to prompt antibiotic treatment for patients with SSIs. This could have significant benefits in improving patient outcomes and healthcare costs.

Objective:To develop an artificial intelligence–based pipeline to assess and triage patient-submitted postoperative wound images.Background:The rise of outpatient surgeries, remote monitoring, and patient-submitted wound images via online portals has contributed to a growing administrative workload on clinicians. Early identification of surgical site infection (SSI) is essential for reducing postoperative morbidity.Methods:Patients ≥18 years old who underwent surgery at 9 affiliated Mayo Clinic hospitals (2019-2022) and were captured by the National Surgical Quality Improvement Program (NSQIP) were included. Eligibility required submission of one image via the patient portal within 30 days after surgery. Images were independently screened in duplicate to determine the presence of an incision. SSI outcomes were obtained from NSQIP. The developed model consisted of 2 stages: incision detection and SSI detection in images with incisions. Four pretrained architectures were evaluated using 10-fold cross-validation, with upsampling and data augmentation to mitigate class imbalance. An end-to-end pipeline, image quality assessment and sensitivity analysis stratified by race were also performed.Results:Among 6060 patients, the median age was 54 years (interquartile range: 40–65), 61.4% (n=3805) were female, and 92.5% (n=5731) identified as White. SSIs were confirmed in 6.2% (n=386) images. Vision Transformer outperformed all others, achieving an incision detection accuracy of 0.94 (area under the curve=0.98) and an SSI detection accuracy of 0.73 (area under the curve=0.81). In addition, it demonstrated strong performance in assessing image quality. Sensitivity analysis revealed comparable performance across racial subgroups.Conclusion:This artificial intelligence pipeline demonstrates promising performance in automating wound image assessment and SSI detection, reducing clinical workload and improving postoperative care.

Aim Surgical site infection (SSI) is one of the most common post-operative complications and is associated with significant morbidity and mortality. This study aims to provide an overview of whether artificial intelligence can accurately identify a surgical site infection on postoperative wound images. Methods The PubMed, Embase, CINAHL, Cochrane Library, Scopus, Web of Science, IEEE Explore, ArXiv and ACM Digital Library databases were searched up to November 12th 2024, and eligible studies were included. Risk of bias was assessed in the different studies using Prediction Model Risk of Bias Assessment Tool (PROBAST). A systematic review was conducted by combining data from the separate trials and analysing multiple area under the receiver operating curve (AUC) values. A meta-analysis to combine different AUC values was performed. Results We identified 12 eligible studies for inclusion. Out of 15 941 images from nine different surgical subspecialties, 1995 (12.5%) images were suspected for SSI. SSI labelling was either done by physical exam by a general practitioner or photographic data was judged by clinical experts. All included studies utilized both traditional machine learning or deep learning approaches. Overall risk of bias varied from low to high. Nine studies were found eligible for inclusion in the meta-analysis and analysis of pooled data from the nine included studies demonstrated a combined AUC of 0.92 (95% c.i. 0.68–0.99). Conclusion The use of AI models for SSI detection on images could prove beneficial in automating post-discharge surgical wound follow-up, possibly reducing the need for face to face evaluation and, in consequence, costs. However, current models do not yet meet the performance standard on sensitivity needed for them to be useful in postoperative wound monitoring. In this context, international data sharing and collaboration is key, both to increase generalizability and to improve model robustness with more images. Moreover, future studies should strive to incorporate external validation to evaluate model performance.

The aim of the study was to describe the prevalence and general characteristics of acute and chronic wounds in 2018 in Alentejo (Portugal) continuing care units. In order to look at associations, wound characteristics studied were location, type, place of acquisition, number, and duration, and patient characteristics were sex, age, and presence of risk factors. During the first 2 weeks of February 2018, a total of 770 patients were assessed at continuing care units of Alentejo. Of these, 135 exhibited wounds, a prevalence of 17.5%. Almost two out of three patients (63%) had arterial hypertension, slightly more than one in three (37%) had a stroke and/or immobility and 30% had diabetes. Of the total wounds identified, 18% were acute wounds and 82% were chronic wounds. Of the 24 acute wounds, traumatic wounds (76%), and surgical wounds (22%) were the most prevalent. The four types of pressure ulcers represented 80% of the chronic wounds. The median duration of the pressure ulcers was 5.5 months and 25% had duration over 10 months.