Bacterial Fluorescence Signals Predict Dermal Regeneration Template Integration Outcomes

ABSTRACT

Acute trauma wounds with exposed bone or tendon present significant reconstructive challenges, with high rates of dermal regeneration template (DRT) failure, often due to missed infections. Standard‐of‐care wound assessment, which relies on subjective clinical signs and traditional swabbing methods, frequently fails to detect clinically significant bacterial loads (> 10⁴ CFU/g). This prospective observational study evaluated the ability of intraoperative bacterial fluorescence imaging to predict DRT outcomes in patients with complex wounds. A subset of 13 patients underwent a blinded fluorescence imaging procedure whereby modified Levine and fluorescence‐guided cultures were collected immediately prior to DRT placement. We found a 63% overall DRT complication rate, with 56% of patients experiencing partial DRT failure despite being clinically and microbiologically cleared of infection pre‐operatively. Fluorescence imaging demonstrated high predictive accuracy; DRTs placed on fluorescence‐negative areas had an 85.6% success rate, whereas those placed on fluorescence‐positive areas had a success rate of only 14.4%. Fluorescence‐guided cultures detected pathogens more often than traditional methods. These findings suggest that fluorescence imaging shows potential value as an adjunctive tool for real‐time, intraoperative bacterial detection, which can guide more effective wound bed preparation and significantly improve DRT integration rates.

Abbreviations

CFU

colony forming units

DRT

dermal regeneration template

FN

false negative

FP

false positive

IV

intravenous

IWII

International Wound Infection Institute

LOS

length of stay

NPV

negative predictive value

NPWT

negative pressure wound therapy

PPV

positive predictive value

RCT

randomized controlled trial

TN

true negative

TP

true positive

1. Introduction

Acute trauma wounds with exposed tendon and/or bone present significant reconstructive challenges due to poor vascularization, which increases the risk of infection, graft failure and amputation when split‐thickness skin grafts are used alone [1]. In such cases, a two‐stage surgical approach using dermal regeneration templates (DRTs) is often required to facilitate revascularization and neodermal formation prior to skin grafting. Adequate wound bed preparation is crucial for DRT success and, as with all skin substitute products, complete infection control must be achieved before placement. Standard of care wound management involves assessing for clinical signs and symptoms of infection based on the International Wound Infection Institute (IWII) guidelines and collecting microbiological wound swabs from suspicious regions to identify and treat bacterial contamination or infection. However, clinical signs and symptoms, such as delayed wound healing, pain and hypergranulation, are often covert and unreliable indicators of bacterial burden and infection [2]. Importantly, these symptoms reflect a host response to bacteria, but do not reliably indicate the location of elevated bacterial loads that contribute to delayed healing [3]. Standard of care practices also do not typically involve collecting wound swabs in the absence of overt infection. Moreover, traditional swabbing methods are limited to sampling the wound bed (Z technique) or rely on these subjective visual cues to select the sampling location (Levine technique) and therefore may miss bacteria located in ambiguous or discreet regions of the wound (e.g., periphery, deep structures, within biofilm). In a systematic review, superficial wound cultures demonstrated poor sensitivity (49%) and specificity (62%) compared with deep tissue cultures [4].

As a result, wounds with elevated bacterial burden may go unrecognized prior to reconstruction, leading to inappropriate DRT placement over contaminated tissue and increasing the risk of partial or complete DRT failure [5, 6, 7]. Considering the substantial costs associated with a single DRT placement [8], as well as the increased hospitalization and resource utilization associated with DRT failure, the economic burden is considerable. These factors highlight the importance of adequate wound bed preparation prior to DRT placement and the need for supportive tools to accurately identify high bacterial loads for targeted treatment.

Innovative technologies such as fluorescence imaging systems have emerged as a valuable adjunctive tool for detecting elevated bacterial loads in wounds. These point‐of‐care devices offer real‐time visualization of most Gram‐positive and Gram‐negative bacterial species, including Staphylococcus aureus , Pseudomonas aeruginosa and Enterobacteriaceae species, at moderate‐to‐heavy loads exceeding 10⁴ CFU/g [9, 10, 11]. By providing objective, accurate and location‐specific INFORMATION on bacterial burden in or around the wound, real‐time fluorescence imaging enhances clinical decision‐making with high sensitivity and specificity [10, 12]. Its effectiveness in managing chronic wounds is well‐established [10, 13, 14, 15], and growing evidence supports its intraoperative application, particularly in wounds requiring skin substitutes or grafting [16, 17, 18]. We previously demonstrated that fluorescence imaging predicted skin graft outcomes in burn wounds with 86% sensitivity and 98% specificity, significantly outperforming the standard Levine technique (30% sensitivity, 76% specificity) [19]. These findings demonstrate how fluorescence imaging can support more informed clinical decisions regarding graft placement.

Herein, we describe the clinical outcomes of DRTs for 32 patients with complex trauma wounds with exposed bone and/or tendon. Given the limitations of traditional bacterial detection methods in these wounds, we aimed to evaluate whether intraoperative fluorescence imaging could predict the success or failure of DRT integration.

2. Materials and Methods

2.1. Study Setting and Patient Enrollment

This single‐center prospective observational study was conducted at Hospital General Dr. Ruben Leñero (Mexico City, Mexico) between January 2022 and September 2025. This study complies with the ethical rules for human experimentation as stated in the 1975 Declaration of Helsinki. The study protocol was approved by the institutional ethics committee (registration #2050103225). Initial patient enrollment included patients 18 years and older with a trauma wound exhibiting significant bone and/or tendon exposure. Patients with a previously documented infection, defined by at least one prior positive wound swab that had been successfully treated (i.e., lack of clinical signs of infection and negative wound swab culture), were considered candidates for DRT application prior to skin grafting and eligible to participate in this study. To meet inclusion criteria, patients were required to demonstrate no local or systemic signs or symptoms of infection and have at least one negative wound culture prior to DRT application. Informed consent was obtained from all enrolled patients. Patients were excluded if they were unwilling or not able to consent or if they presented with comorbidities that may confound study outcomes (e.g., chronic steroid use, chronic immunodeficiency). Of note, all patients with type 2 diabetes included in this study were under pharmacologic treatment and maintained glycemic levels within the normal range. A total of 37 patients were initially enrolled. Five patients were withdrawn due to nonadherence to protocols (n = 3), study withdrawal on request (n = 1), and death (n = 1). Thus, 32 patients were included in this study. All patients received standard of care wound management and DRT application. In 10 cases, two additional wound cultures (modified Levine and fluorescence‐guided) were taken intraoperatively in a blinded manner immediately prior to DRT application, as described below.

2.2. Standard of Care Preoperative Wound Management

The standard of care preoperative wound management protocol involved cleansing, debridement and coverage with sterile silver sulfadiazine‐impregnated and sterile compresses, cotton and bandages. In this study, all patients had a previously documented infection based on clinical assessment and positive wound culture results treated, as described above, with IV antibiotics. Preoperative wound cultures were collected using a modified Levine technique [19, 20], whereby the surface area swabbed covered as much of the wound as possible. Subsequent treatment adjustments were made based on wound culture results, tailoring the approach to the identified susceptibility. Adjuvant therapies, such as silver antimicrobial dressings and negative pressure wound therapy (NPWT), were applied if medically indicated.

2.3. Fluorescence Imaging System

The MolecuLight i:X fluorescence imaging system (MolecuLight Inc., Toronto Canada) is a handheld, point‐of‐care device that emits a safe, violet light (405 nm) onto a wound, causing the excitation of specific endogenous components of bacteria and tissues to fluoresce various colours without the need for contrast dyes [21]. Certain aerobic and anaerobic bacterial species containing porphyrins, such as S. aureus and Escherichia coli , produce a red or blush fluorescence while P. aeruginosa exclusively produces a cyan fluorescence due to the presence of pyoverdines [9, 11, 22]. Native tissue components (e.g., collagen, fibrin, elastin) fluoresce green, providing anatomical context for accurate image interpretation. The real‐time detection of red and/or cyan fluorescent signals by the device indicates an elevated, clinically relevant bacterial load exceeding 10⁴ CFU/g that impairs healing and requires intervention [3]. The device captures standard and fluorescence images and videos of wounds that can be interpreted at the point‐of‐care based on the clinical judgement of an experienced healthcare provider and recorded in the patient's medical record for later assessment. In this study, a blinded, independent clinical investigator (CDTG) was solely responsible for image capture and interpretation. The blinded investigator received extensive training on device use and image interpretation prior to study commencement in accordance with the manufacturer's guidelines.

2.4. Study Workflow

Enrolled study subjects who underwent successful standard of care preoperative wound management and were deemed negative for infection were taken to the operating room for DRT placement. All relevant data (e.g., demographics, type of injury, wound characteristics, including measurements, documented infections and previously used antibiotics) were recorded prior to the procedure. First, bandages and dressings were removed, the wound was cleansed with a saline (0.9%) solution lavage to remove surface debris, then dried with sterile gauze. Surgical debridement was performed to excise any remaining nonviable tissues followed by complete haemostasis (Figure 1a).

FIGURE 1.

(a) Preoperative wound management followed standard of care, including thorough cleansing and debridement to confirm the absence of infection. (b) Bacterial fluorescence imaging was performed by an independent, blinded investigator in the absence of the surgical team in a subset of patients. (c) The surgical team subsequently proceeded with DRT placement.

Bacterial fluorescence imaging was performed in a subset of patients for which the surgical team consented to intraoperative imaging (Demographics, Table 1; Figure 1b). To ensure adequate blinding, the surgical team exited the room during this procedure, leaving only the blinded, independent investigator to first obtain an intraoperative wound culture using the modified Levine technique. Next, a standard white‐light image was taken with the MolecuLight i:X device from a distance of 8–12 cm from the wound before the room was appropriately darkened to capture fluorescence images. Serial fluorescence images or videos of the wound were captured if the wound could not be comprehensively captured in a single image. During this time, an additional wound culture was obtained, guided by fluorescence‐positive areas demonstrating red or cyan signals.

TABLE 1.

Demographics summary of patients.

	Patient cohort (n = 32)	Patient subset: fluorescence imaging procedure (n = 13)
Age (years)	43.5 ± 16.5	47.5 ± 16.9
Gender (%)
Male	26 (81%)	11 (85%)
Female	6 (19%)	2 (15%)
BMI	24.7 ± 2.30	23.4 ± 1.34
Comorbidities (n, %)
Type 2 Diabetes	10 (32%)	6 (19%)
Hypertension	5 (16%)	4 (31%)
Injury Etiology (n, %)
Crushing	13 (41%)	5 (38%)
Laceration	19 (59%)	8 (62%)
Wound Characteristics
Anatomic Region (n, %)
Right Upper Limb	13 (41%)	4 (31%)
Right Lower Limb	3 (9.4%)	0 (0%)
Left Upper Limb	10 (32%)	6 (46%)
Left Lower Limb	6 (19%)	3 (23%)
Extent of Injury (cm2)	93.4 ± 129	99.8 ± 78.0
Extent of Tendon/Bone Exposed (cm2)	19.4 ± 25.7	26.8 ± 22.7
Tendon/Bone exposure/Extent of Injury (%)	32.3 ± 17.3	31.0 ± 17.4
Length of Stay (LOS, days)	18.4 ± 14.9	23.5 ± 15.5

Note: Values are calculated as mean ± SD unless otherwise noted. Values were rounded to the nearest number and may not equate to 100%.

Following the fluorescence imaging procedure, the independent investigator exited the operating room, allowing the surgical team to continue with DRT placement procedure (Figure 1c). For the patients that did not undergo the fluorescence imaging procedure (n = 19), the blinded investigator was not involved and the surgical team proceeded directly to the DRT placement procedure. The surgical team, blinded by the fluorescence imaging results, performed DRT placement (Integra, Plainsboro, NJ) in all patients. The bilayer Integra DRT was shaped to accurately fit the excised wound margins and placed according to the manufacturer's instructions. Once placed, DRT integration or failure was assessed weekly for 3 weeks by the surgical team, with evaluations focusing on colour changes to rule out seroma formation or clinical signs of infection. DRT integration was achieved following an average of 24 ± 5.1 days post‐application. DRT failure was defined as any area where there was no DRT integration and/or granulation tissue at the wound bed. The failed DRT may have been completely detached (floating) or disintegrated (lysis).

2.4.1. Data and Image Interpretation

Standard and fluorescence wound images and videos captured by the device were transferred to a secure, centralized database for interpretation and further data analysis. The clinical findings and wound culture results were linked to the deidentified patient information. Image interpretation was performed by the same independent investigator who conducted the intraoperative fluorescence imaging procedure. To quantify the wound and fluorescence signal distribution, a 2D standard image of the wound taken prior to DRT application was divided into 1cm² sections to determine the total wound area (Figure 2a). Fluorescence images were then overlaid to identify and map 1 cm² regions exhibiting red or cyan signals, labelled as ‘positive red’ or ‘positive cyan’, respectively (Figure 2b). Following DRT application, a standard image captured during follow‐up was used to assess DRT integration or failure on a per‐square‐centimetre basis (Figure 2c). These data, comprised of the total wound area, fluorescence positive and negative areas regions and areas of DRT integration or failure, were used to evaluate the diagnostic accuracy of fluorescence imaging in predicting DRT outcomes.

FIGURE 2.

(a) A standard preoperative image of the wound was divided into 1 cm² sections to quantify total wound area. (b) Fluorescence imaging was then overlaid on the wound map, highlighting 1 cm² regions with red or cyan signals. (c) A standard image taken approximately 3 weeks post‐DRT application was then used to calculate diagnostic accuracy measurements based on true negative (yellow), true positive (black), false negative (grey) and false positive (orange) based on the correlations between FL signals and DRT outcomes (i.e., integration or failure).

2.5. Statistical Analyses

Demographic data and wound characteristics were reported using descriptive analyses. To assess the diagnostic accuracy of fluorescence imaging in predicting DRT integration or failure, we calculated the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). These calculations were based on the segmented (per‐square‐centimetre) 2D wound schematics. For each 1 cm² section, the presence of fluorescence signals (positive or negative) and DRT outcomes (integration or failure) were recorded. Fluorescence‐positive regions that later corresponded to DRT failure were classified as true positive (TP), while fluorescence‐positive regions corresponding to successful DRT integration were classified as false positives (FP). Conversely, fluorescence‐negative regions that corresponded to areas of successfully integrated DRT were classified as true negatives (TN) and fluorescence‐negative regions that subsequently failed were classified as false negatives (FN). With respect to DRT failure, sensitivity and specificity were calculated as TP/(TP + FN) and TN/(TN + FP), respectively, while PPV and NPV were calculated as TP/(TP + FP) and TN/(TN + FN), respectively.

These diagnostic accuracy metrics were reported in three ways. First, we identified the raw burden of tissue that was accurately or inaccurately classified as at‐risk or not for DRT failure by fluorescence imaging. The absolute area (cm²) was calculated as the sum of positive or negative fluorescence regions corresponding to each DRT outcome (lysis or integration, respectively). Second, the absolute area value was expressed as a proportion over the total wound area (%) to adjust for differences in wound size and express the diagnostic accuracy relative to the overall wound surface. Lastly, to determine how accurately positive fluorescence signals could predict DRT failure within positive regions, and conversely how negative fluorescence signals could predict DRT integration within negative regions, each diagnostic output was further normalized by the total fluorescence‐positive (TP and FP) or fluorescence‐negative (TN and FN) area to the corresponding wound. This adjustment accounts for differences in wound size and distribution of fluorescence signals (e.g., large, contiguous regions vs. small, scattered patches of fluorescence‐positive signals), ultimately assessing how accurately fluorescence imaging can predict outcomes specifically within the regions where the fluorescence signal (positive or negative) was present. This multi‐tiered approach allowed for a comprehensive assessment of the diagnostic utility of fluorescence imaging at the wound bed level.

3. Results

3.1. Demographics and Wound Characteristics

A total of 32 patients were enrolled in this study. Demographics and wound characteristics are summarized in Table 1. The mean patient age was 43.5 ± 14.5 years, and the majority (81%) were male. Twelve patients had comorbidities: 10 (31%) patients had type 2 diabetes and 5 (16%) patients had hypertension. The majority of patients presented with laceration injuries (n = 19, 59%), while the remaining patients presented with crushing injuries (n = 13; 41%). The most common anatomic region of injury was the right upper limb (n = 13; 41%), followed by the left upper limb (n = 10; 31%), the left lower limb (n = 6; 19%) and the right lower limb (n = 3; 9%). The mean extent of injury was 93.4 ± 129 cm². All patients' wounds had either exposed tendon (n = 20; 63%), bone (n = 4; 13%), or tendon and bone (n = 8; 25%), which comprised a mean of 19.4 ± 25.7 cm² (32.3% ± 17.3%) of the total wound.

All patients had a previously documented infection, confirmed by the presence of clinical signs and symptoms and at least one positive microbiological wound culture. Patients were treated with antibiotics tailored to the results of the wound culture (e.g., Cefalexina, clindamycin, levofloxacino). Adjunct therapies (e.g., NPWT, silver sulfadiazine) and blood transfusions were administered, when required, to stabilze patients and achieve infection control. Prior to DRT application, all patients were deemed negative for clinical signs and symptoms of infection and had at least one negative wound culture. The mean duration between the last negative culture and DRT application was 6.9 days (SD, 4.7 days; range, 0–24 days).

3.2. Fluorescence Imaging and Wound Cultures

13 of the 32 study participants underwent intraoperative blinded fluorescence imaging immediately prior to DRT application, during which two wound cultures per patient were collected (Table 2). Fluorescence signal detection thresholds were applied consistently across all cases, and imaging was conducted under standardized low‐light conditions in accordance with manufacturer‐recommended guidelines to minimize variability related to ambient lighting. Among these 13 patients, we assessed the relationship between the fluorescence signals and the results of the modified Levine and fluorescence‐guided wound cultures. Of the nine fluorescence‐positive wounds, the modified Levine culture detected clinically significant bacterial loads in one wound (11%), while fluorescence‐guided cultures detected significant pathogens in six (67%) wounds. The mean positive fluorescence area and corresponding proportion of fluorescence positive signals over the total wound area was 14.4 ± 20.5 cm² and 19.0% ± 17.4%, respectively. Conversely, the mean negative fluorescence area and corresponding proportion to the total wound area was 59.9 ± 52.5 cm² and 81.0% ± 17.4%, respectively. All fluorescence‐positive wounds exhibited red fluorescence signals (n = 9; 69%), indicative of the presence of clinically significant bacterial loads (> 10⁴ CFU/g). Five of these wounds displayed red and cyan fluorescence signals. Three (33%) fluorescence‐guided wound cultures detected the presence of Gram‐positive bacterial loads, predominantly E. faecalis , while just one (7.7%) modified Levine culture confirmed the presence of this bacterial species. Additionally, three (33%) fluorescence‐guided cultures detected P. aeruginosa . The remaining cultures reported no significant growth (33%; < 10⁴ CFU/g). All modified Levine cultures collected from cyan fluorescence‐positive wounds yielded no growth. No pathogens were detected in wounds without fluorescence, and no fluorescence‐guided cultures were collected in these cases.

TABLE 2.

Fluorescence imaging and wound culture results.

	n = 13
Modified Levine wound cultures *
Number of positive cultures (n, %)	1 (7.7%)
Species
Enterococcus faecalis	1 (7.7%)
No growth	12 (92%)
Fluorescence imaging
Negative fluorescence signal (n, %)	4 (31%)
Fluorescence‐negative area (cm2)	59.9 ± 52.5
Proportion of fluorescence‐negative regions over total wound area (%)	81.0 ± 17.4
Positive fluorescence signal (n, %)	9 (69%)
Red	9 (69%)
Cyan	5 (38%)
Red and Cyan	5 (38%)
Fluorescence‐positive area (cm2)	14.4 ± 20.5
Proportion of fluorescence‐positive regions over total wound area (%)	19.0 ± 17.4
Fluorescence‐guided wound cultures
Number of positive cultures (n, %)	6 (41%)
Species
Enterococcus faecalis	3 (33%)
Pseudomonas aeruginosa	3 (33%)
No growth	3 (33%)

Note: Values are calculated as mean ± SD unless otherwise noted. Values were rounded to the nearest number and may not equate to 100%.

^*Modified Levine wound cultures were taken prior to and blinded from fluorescence‐guided cultures.

3.3. DRT Outcomes

Of the entire patient cohort, the mean duration between DRT application and removal of the silicone layer after neodermal formation was 24.1 ± 5.1 days. Complete DRT integration was achieved in 14 (44%) patients, while 18 (56%) patients experienced partial DRT failure. The mean area and proportion of integrated DRT over the total wound area was 64.7 ± 109.5 cm² and 83.9% ± 20.9%, respectively. Conversely, a mean of 10.3 ± 17.2 cm² of DRT failed to integrate (lysis) into the open wound bed, corresponding to a failure rate of 16.1% ± 20.8%. Of the 32 patients, 20 (63%) developed DRT complications, including nine (28%) with infection and 11 (34%) with seroma.

Of the subset of 13 patients who underwent the fluorescence imaging procedure, 10 (77%) patients experienced partial DRT failure. Fluorescence‐positive signals were detected intraoperatively in nine (90%) of 10 wounds that experienced partial DRT failure, and no fluorescence signals were detected in the patients that achieved complete DRT integration.

3.4. Diagnostic Accuracy

The diagnostic accuracy measures of fluorescence imaging to predict DRT failure are summarized in Table 3. When calculating diagnostic accuracy at the level of absolute wound area per cm², fluorescence imaging demonstrated a sensitivity of 58.9%, specificity of 97.0%, PPV of 88.5%, NPV of 85.6% and an accuracy of 86.1% for predicting DRT failure. This indicates that while fluorescence imaging was highly specific in ruling out DRT failure when negative signals were present, its sensitivity for detecting all regions that eventually failed was more limited. When expressed as a proportion of the total wound area, diagnostic performance was similar, yielding a sensitivity of 53.0%, specificity of 95.0%, PPV of 81.0%, NPV of 83.4% and an accuracy of 82.9%. This approach adjusts for wound size and confirms that the overall trend of high specificity, but lower sensitivity, persists when scaled to the total wound surface.

TABLE 3.

Summary of diagnostic accuracy metrics for fluorescence imaging to predict DRT failure (n = 13), providing interpretive context for fluorescence‐positive and fluorescence‐negative wound regions during intraoperative assessment.

	Sensitivity	Specificity	PPV	NPV
Calculated using absolute area of fluorescence regions corresponding to DRT failure (cm2)	58.9%	97.0%	88.5%	85.6%
Calculated using proportion of fluorescence regions over total wound area (%)	53.0%	95.0%	81.0%	83.4%
Calculated using proportion of fluorescence regions over total fluorescence‐positive regions (%)	74.5%	84.5%	78.2%	81.7%

Finally, when diagnostic accuracy measurements were further normalized to fluorescence‐positive (TP and FP) or fluorescence‐negative regions (TN and FN) within each wound, fluorescence imaging demonstrated a sensitivity of 74.5%, specificity of 84.5%, PPV of 78.2%, NPV of 81.7% and an accuracy of 80.3%. This region‐specific calculation provides a more balanced assessment of the accuracy within fluorescence regions while removing bias from differences in wound size and fluorescence distribution. These findings show that within positive or negative regions, fluorescence imaging is reliable for both predicting DRT failure and confirming integration, respectively. Overall, the success rate when applying DRTs over fluorescence‐positive regions per cm² was 14.4% while the success rate of DRTs placed over fluorescence‐negative regions was 85.6%. This highlights the strong negative predictive value of fluorescence imaging and demonstrates that the absence of fluorescence was strongly associated with successful DRT integration, while positive fluorescence signalled a high risk of failure.

4. Discussion

While DRT placement prior to skin grafting can improve reconstructive outcomes in wounds with exposed bone/tendon, their success hinges on meticulous wound bed preparation. Infection control can be difficult in these wounds, particularly with quiescent infections like osteomyelitis [23], where clinical signs of infection can subtle or absent [24]. Traditional wound swabs are also limited in this setting and may not reliably reflect bacterial burden [4, 25, 26]. This presents an urgent need for innovative tools that can help clinicians more accurately identify bacterial contamination and infection in wounds requiring DRTs to avoid complications, improve patient outcomes and reduce costs. Here, we demonstrate that real‐time, intraoperative fluorescence imaging (MolecuLight i:X) can accurately predict DRT outcomes based on the presence of fluorescence signals on the wound. Importantly, because the surgical team was blinded to fluorescence findings, this study was designed to assess predictive performance rather than to evaluate the impact of fluorescence‐guided clinical decision‐making on outcomes. Our study highlights the potential value of integrating fluorescence imaging into the intraoperative clinical workflow for enhanced infection control.

Our study cohort consisted of patients with wounds secondary to crush or laceration injuries, with a mean of 32.3% wound surface area exhibiting exposed bone and/or tendon. We observed an average DRT engraftment rate of 84%. Despite all patients being clinically and microbiologically cleared of infection, we found a 63% overall complication rate following DRT placement, including seroma in 34% and infection in 28% of cases. While infection and seroma formation are a common occurrence with DRT placement [27], our complication rate is higher than those reported in the literature. Alet et al. reported 13.3% seroma and 6.7% infection complication rates following DRT placement among 15 patients with severe traumatic wounds involving exposed deep functional structures [28]. In a study involving previously infected diabetic foot ulcers with exposed bone and tendon, Integra DRT placement resulted in 13% of patients developing complications (e.g., infection or unsuccessful coverage of bone/tendon) [29]. A systematic review including 1254 reconstructed wound sites found Integra DRTs were associated with a 16.9% infection complication rate [5]. This discrepancy between complication rates in our patient cohort compared with previous literature may be due to various factors, including patient and wound characteristics, wound bed preparation protocol, swabbing technique and post‐operative care, among others. Despite this, missed bacterial burden remains a significant issue that can be mitigated by improved wound bed preparation and more accurate, guided wound cultures.

Point‐of‐care fluorescence imaging addresses this unmet clinical need by providing highly accurate, objective information to aid in the detection of clinically significant bacterial loads (> 10⁴ CFU/g) during each stage of surgical wound reconstruction [10, 30]. Numerous studies have supported its effectiveness in identifying moderate‐to‐high bacterial presence, guiding more precise debridement and enhancing clinical decision‐making in various wound care settings [12]. For instance, a recent RCT study found that approximately 30% of the total bacterial load remained following initial debridement as per standard of care, even when performed by experienced surgeons [17]. Intraoperative fluorescence‐guided debridement significantly reduced this residual bacterial load, enabling more complete wound bed preparation. Sandy‐Hodgetts et al. found that 76% of surgical wounds assessed had elevated bacterial loads exceeding 10⁴ CFU/g, confirmed by wound biopsies, which were missed due to the poor accuracy and sensitivity of clinical assessment [30]. Expert‐interpreted fluorescence imaging further enhanced bacterial detection by 11.3‐fold compared to clinical assessment alone. Li et al. employed fluorescence imaging to enhance real‐time intraoperative debridement of infected wounds requiring flap reconstruction, which resulted in improved clinical outcomes including a significant reduction in re‐infection rates [18]. These data demonstrate that, even when best practices for preoperative wound bed preparation are followed, opportunities remain to enhance bacterial detection and management, particularly involving wound bed preparation ahead of skin substitutes or grafting.

We have previously demonstrated that fluorescence imaging is a highly accurate diagnostic tool for assessing bacterial bioburden and predicting outcomes split‐thickness skin graft sites in burn patients [19]. Building on this evidence, the current study evaluated the capability of intraoperative fluorescence imaging to predict the outcomes of skin substitutes (DRT) in a subset of 13 patients. We detected fluorescence‐positive regions in 9 (69%) of patients' wounds, which is consistent with fluorescence‐positivity rates reported in some [17, 18, 30], but not all studies [31, 32] employing fluorescence imaging in a surgical setting. In our prior work, 12 of 38 wounds (32%) displayed positive intraoperative fluorescence signals [19]. This discrepancy may be attributed to differences in wound aetiology and characteristics. For example, clinical signs and symptoms of infection can be more difficult to assess in wounds with exposed bone or tendon [26]. Notably, all fluorescence‐positive wounds experienced partial graft loss despite the absence of overt clinical indicators of infection [19]. Pathogenic bacteria were identified in only one of nine fluorescence‐positive wounds (11%) using traditional cultures, compared to six of nine wounds (67%) using fluorescence‐guided sampling. While the reasons for the suboptimal performance of traditional wound cultures to detect superficial bacteria are unclear, these data are in line with several other studies demonstrating poor accuracy of traditional cultures collected from wounds with exposed bone [26, 33]. These findings further support fluorescence imaging as a more reliable method for bacterial detection and highlight its diagnostic value in guiding targeted, site‐specific sampling.

To assess the predictive capability of fluorescence imaging for DRT failure, we used area‐based diagnostic metrics to correlate fluorescence signals with DRT integration or failure. After accounting for differences in wound size and fluorescence distribution, we found fluorescence imaging to be a reliably accurate predictor of DRT success with a sensitivity of 74.5% and specificity of 84.5%, a PPV of 78.2%, NPV of 81.7% and an accuracy of 80.3%. These findings suggest that the presence of bacterial fluorescence is a strong negative prognostic indicator for site‐specific DRT integration. We observed a high success rate over fluorescence‐negative areas (85.6%), suggesting that the absence of fluorescence signals may reliably predict successful integration. However, even with moderate sensitivity, fluorescence imaging plays an important role by providing bacteria‐specific information that enhances intraoperative assessment and supports surgical judgement during DRT placement. Conversely, the success rate for fluorescence‐positive regions was just 14.4%. This spatial analysis highlights the clinical utility of fluorescence imaging in guiding more effective wound bed preparation to improve DRT integration rates when interpreted in conjunction with existing clinical and microbiological assessment strategies.

The current data validates our previous work demonstrating that fluorescence imaging can predict skin graft outcomes in burn patients with a sensitivity and specificity of 92.3% and 96%, respectively, compared to Levine cultures (sensitivity, 30%; specificity, 76%) [19]. Further, we previously demonstrated a 99.2% chance of success if a skin graft was placed over a fluorescence‐negative region, whereas the graft‐take rate diminished to 27.9% when placed over fluorescence‐positive regions. Together, the area‐based diagnostic metrics from these two studies highlight the predictive performance of fluorescence imaging in guiding optimal skin graft and skin substitute placement. These data further support the integration of fluorescence imaging into pre‐ and intraoperative clinical workflows, particularly during wound bed preparation, to mitigate infectious complications and possibly improve patient outcomes.

Skin substitutes are inherently costly. Integra DRTs, the most widely commercially available and commonly used DRT, used in the present study, costs approximately US$18 per cm [28]. Shakir et al. reported the successful integration of Integra DRTs and skin grafts incurred 2.4 times lower total costs compared to those that failed ($24,510.75 vs. $59,680.11; p < 0.01), with preoperative wound colonization more frequently observed in the failed cases [34]. Frame et al. demonstrated that following DRT placement, wounds that developed infections were significantly associated with less than 100% DRT take, suboptimal autograft take and increased infectious complications following skin grafting [35]. Considering our findings demonstrating an increased likelihood of DRT failure in fluorescence‐positive regions, indicating elevated bacterial presence, these data suggest that fluorescence imaging may have the potential to reduce failure‐related costs by enabling earlier detection of elevated bacterial burden; however, a formal cost‐effectiveness analysis was not performed in this study. Future studies incorporating prospective economic evaluations are needed to quantify the cost impact of fluorescence imaging–guided wound bed preparation.

This study is not without limitations. First, the sample size of patients who underwent fluorescence imaging was small (n = 13), which may limit the generalizability of the findings. Accordingly, these findings are hypothesis‐generating. Future prospective studies with larger, multicenter cohorts are needed to validate these findings and assess their strength across diverse trauma populations and care settings. Second, wound sizes varied considerably, ranging from small to extensive injuries. While we normalized our diagnostic measurements to account for differences in wound size and fluorescence distribution, wound heterogeneity with respect to anatomic location, injury mechanism and degree of bone or tendon exposure may independently influence bacterial burden and DRT integration, and residual confounding related these factors cannot be fully excluded. Future studies with larger cohorts should incorporate stratified analyses based on wound characteristics to more clearly define the impact of these variables on predictive performance. Third, fluorescence image acquisition and interpretation were performed by a single trained, blinded investigator, which may limit evaluation of reproducibility. Although standardized training and manufacturer‐recommended interpretation criteria were used, future studies should incorporate multiple independent raters and formal inter‐rater reliability analyses to further evaluate consistency and generalizability of fluorescence signal interpretation. Fourth, microbiological culture data were limited to the fluorescence‐imaged subset and obtained at a single intraoperative time point. While intraoperative cultures were selected to align with the timing of fluorescence imaging assessment, deeper tissue sampling and serial cultures may further strengthen microbiological correlations and should be considered in future studies.

Taken together, our findings demonstrate that the absence of clinical signs and symptoms and negative wound culture results do not reliably exclude the presence of clinically significant bacterial loads that may impair DRT integration. Conversely, fluorescence‐guided cultures more accurately reflected bacterial burden compared with traditional cultures obtained using the Levine technique. Fluorescence imaging also demonstrated high accuracy in predicting DRT outcomes, whereby DRTs placed over fluorescence‐positive regions had markedly lower integration rates compared to those placed over fluorescence‐negative regions. These results support the value of intraoperative fluorescence imaging as an adjunctive tool to improve wound bed preparation ahead of DRT placement for optimal patient outcomes. Future interventional studies will be important to determine whether integrating fluorescence‐guided strategies will further optimise DRT integration and postoperative outcomes.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest. While the authors received sponsorship from MolecuLight Inc. to present this work at the 2025 American Burn Association conference, this funding did not involve, nor did it influence, the design, data collection, analysis or interpretation of the study findings.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.