1. Introduction
Skin defects in the head, face and neck resulting from chronic wounds, traumatic injuries, or oncological conditions require effective closure. A tailored approach must be developed for each defect based on the patient’s requirements and the extent of the affected tissues [
1]. The reconstructive ladder is a tool used in plastic surgery to evaluate the extent of a defect and select the most appropriate method for treatment. This ladder begins with primary closure, followed by free skin grafts, local flaps, and pedicled flaps, and ends with free flaps, depending on the location and severity of the defect [
2,
3]. When reconstructing defects in the face, head, and neck, surgeons often use regional or distant pedicled flaps. Regional flaps are a suitable option for treating areas where the damage is not too severe and where it is crucial to maintain the natural look of the face. An understanding of the biomechanics of soft tissue and the vascular supply to skin flaps is elementary for assessing skin-/and flap perfusion [
4].
Despite significant advancements in medical, surgical and technical fields, no technique has been found that allows for a consistent, simple, reliable, and cost-effective assessment of perfusion in cutaneous flaps. Therefore, direct clinical evaluation remains the gold standard in the monitoring of skin flaps [
5,
6]. Most head and neck surgeons evaluate flap perfusion based on parameters such as color, reperfusion, scratch testing (quality of the blood), and flap surface temperature [
6,
7].
Thermal imaging (TI), smartphone-based TI (SBTI), and dynamic infrared thermography (DIRT) have now shown promise as an innovative tool in several medical specialties. Using the detection of infrared radiation emitted by the body, thermography produces accurate visual representations of surface temperatures, allowing for areas of increased or decreased blood flow to be highlighted [
8].
As an advanced, non-invasive, and low-cost, mostly smartphone-based (Teledyne FLIR LLC, Wilsonville, OR, USA) [
9,
10,
11,
12,
13] imaging technique, SBTI offers valuable insights into the graduation of burn depths [
9,
10], as well as thermal physiology [
14].
This study aimed to deliver an implementable algorithm for the use of SBTI in the perioperative setting and a prior insight into the accuracy of this TI technique as a reliable assessment for cutaneous perfusion in pedicled flaps in the face, head and neck.
2. Materials and Methods
2.1. Ethics Statement
This study involving human participants adheres to ethical standards. Approval was obtained from the Ethics Committee of the University of Heidelberg (Votum: S-630/2023) on 6 November 2023. The research conforms to the Declaration of Helsinki.
The publication of images and individual participant information is contingent on the authors securing free prior informed consent. The authors confirmed consent during the author’s license to publish without submitting the actual form. A standard patient consent form from Thieme© was obtained by all patients.
2.2. Thermal Imaging Technique
TI and dynamic infrared thermography (DIRT), have now shown promise as an innovative tool in several medical specialties [
14]. Thermal imaging technology utilizes electromagnetic radiation in the near-infrared (NIR) range, specifically wavelengths ranging from 780 to 1400 nm. Unlike visible light, which is perceived by the human eye, NIR radiation is invisible and requires technology-based recording, analysis, and interpretation. By employing specialized cameras sensitive to NIR wavelengths, such as infrared cameras (e.g., the FLIR One System), it becomes possible to capture and visualize NIR radiation [
15]. Using the detection of infrared radiation emitted by the body, thermography produces accurate visual representations of surface temperatures, allowing for areas of increased or decreased blood flow to be highlighted. Over these color-coded heat emission differences, indirect perfusion patterns can be drawn.
2.3. Implementation of SBTI in the Operating Room
Before the evaluation of the SBTI technique in cutaneous flaps, we established an imaging protocol, given in
Figure 1;
Figure 2, where we defined the definite points of interest (T1–T4). The ROI (region of interest) was defined by the center of the flap with an imaging distance taken 30 cm away from the central point of the flap. The room temperature in our OR is controlled by a thermoregulation system, which holds it constantly around 23 °C. The humidity of the room is constantly around 30%.
2.4. Study Course
Our study aimed to integrate thermal imaging into the assessment of patients undergoing reconstructive surgery for defects in the face and neck following tumor resections of varying extents. This investigation was conducted at SLK Kliniken Heilbronn, Germany, within the Department for Otorhinolaryngology/Head and Neck, Plastic Surgery, over four months, from November 2023 to March 2024. We integrated various data such as age, gender, diagnosis, disease stage, diabetes mellitus status, smoking history (pack years), hypertension, anticoagulants, history of previous surgeries in the operative area, details of the surgical procedure (including technique and date), location and size of the defect, intra-operative clinical assessment and Doppler signal, imaging details, postoperative complications, exact timing of post-operative assessments (T4) and the need for revision surgery followed up for two months.
Drawing from findings in prior research, we found that incorporating the FLIR One camera into our surgical protocol was straightforward, albeit requiring a learning curve for both its application and the interpretation of images about the indirect visualization of blood flow based on temperature differentials. Hence, we opted for the FLIR One System (Teledyne FLIR LLC, Wilsonville, OR, USA) for this pilot study. This multispectral camera features a near-infrared camera unit with a resolution of 160 × 120 pixels, covering a temperature range of 0 °C to 400 °C, with a thermal resolution of 70 mK.
For our investigation, we took thermal images from a total of n = 11 axial pattern and random pattern pedicled skin flaps of the face and neck (e.g., bilobed flaps, Rieger flaps, Hatchet flaps, island flaps, etc.) using the FLIR One camera, resulting in a total of n = 44 images. These images were captured at various stages: after marking (T1), after flap elevation (T2), upon completion of surgery (T3), and 24 h postoperatively (T4) (see
Figure 2).
2.5. Scoring System
The SBTI study group was graded by 1 corresponding to ΔT = 0 °C, indicating perfect perfusion of the entire flap. Grades of 2, 3, 4, and 5 corresponded to ΔT ≥ 2 °C, ΔT ≥ 4 °C, ΔT ≥ 6 °C, and ΔT ≥ 8 °C, respectively, with grade 5 indicating no perfusion, from the flap center to the healthy periphery. For temperature interpretation, the rainbow temperature scale scheme was used (see
Figure 3).
A control group was included, with clinical assessments used as the gold standard for comparison. The control group was graded on a scale from 1 to 5 based on the following criteria: a grade of 1 indicated good recapillarization (<1 s), no color difference between the flap and surrounding tissue, and peripheral flap microbleeding from its edges. A grade of 2 indicated good recapillarization (<2 s), no or low color difference, and peripheral flap microbleeding. A grade of 3 indicated recapillarization (>2 s), brighter flap color, and peripheral bleeding. A grade of 4 indicated no peripheral bleeding, delayed recapillarization signs, and a white flap. A grade of 5 indicated no peripheral bleeding, a pale flap, and no peripheral bleeding of the flap.
3. Results
3.1. Implementation
The implementation process for the study began with planning and drafting the study protocol inspired by Nischwitz et al. [
1], without a cooling phase between the imaging points. (see
Figure 3) Following this, the necessary SBTI camera was procured, and the implementation commenced. However, several obstacles were encountered during this phase. Initially, images were collected by different individuals (nursing staff) without adhering to the study protocol guidelines consequently. As a result, not all collected images could be used for the study and had to be excluded. Specifically, six patients were excluded due to the following reasons: inadequate distance (n = 2) or incomplete capture of the region of interest (n = 1). Additionally, patients were excluded if all required images from T1 to T4 were not available (n = 3), leading to a loss of follow-up. To address these issues, a team of two surgeons was involved in conducting SBTI in the operating room.
3.2. Patient Specifications
The dataset encompasses a total of 11 (4 m; 7 f) patients of varying ages [64:93;77,37 yrs] who underwent reconstructive surgery for facial, head and neck defects [1 × 1 cm:6 × 6 cm] following diagnoses such as squamous cell carcinoma (SCC, n = 3), basal cell carcinoma (BCC, n = 6), malignant melanoma (MM, n = 1), or actinic keratosis (n = 1). Considering perioperative risk profiles, three patients had diabetes mellitus and six patients had hypertension. A total of two were smokers, with 30 and 40 pack years, respectively. The surgical techniques employed were diverse, including procedures such as Rieger flaps (n = 5), Esser flap (n = 1), bilobed flaps (n = 1, See
Figure 4), two-island flaps, one transposition flap, and one Hatchet flap. In total, we included n = 3 axial pattern and n = 8 random pattern flaps. Surgery dates spanned from August 2023 to February 2024. Intraoperative assessments consistently indicated good perfusion, with a Doppler signal present in n = 1 cases, a Hatchet flap (Supratrochlear Artery, see
Figure 5). One post-operative dehiscence in a cervical advancement flap was recorded which was treated by secondary intention, with revision surgery not occasionally necessary. Furthermore, this patient had an elevated complication risk due to his advanced age of 84. The intraoperative SBTI grading was “3” in T2 and T3 not in line with a grade of “2” at timepoints T2 and T3 in clinical assessment.
3.3. Descriptive Analysis of SBTI
As depicted in the two subsequent tables, the assessment of imaging data by both surgeons demonstrates consistent outcomes, albeit exhibiting a greater variability from time points T2–T4 in SBTI imaging (see
Table 1 and
Table 2).
3.4. Outcome
The pedicled skin flaps examined all showed a preserved, uniform thermal image signature peri- and postoperatively for time points T1–4 in both surgeons’ evaluations. This corresponded to a good postoperative result regarding the vitality of the tissue transfer in the clinical assessments 24 h and one week postoperatively. A single postoperative dehiscence occurred in a cervical advancement flap, which exhibited a diminished thermal signature in the periphery intraoperatively. Both evaluating surgeons consistently graded it as “3”, which means >4 °C difference from T2 to T4 in SBTI evaluations, while clinical assessments yielded a “2,” indicating a more favorable outcome. The dehiscence was not associated with diabetes or smoking.
3.5. Limitations
While the application of smartphone-based thermal imaging (SBTI) in flap monitoring presents several advantages, notable limitations must be considered. Thermal imaging demonstrated limited accuracy in small flaps approximately one centimeter in length, often exceeding the resolution capabilities of current technology and leading to potential misidentification. The practical use of SBTI is further constrained by its sensitivity to surface temperature changes, which can be influenced by external conditions such as room temperature differences, high humidity, and the use of cold wet gauze on the tissue. Additionally, anatomical challenges in regions like the nose and ears, and the impact of comorbidities such as smoking and diabetes mellitus, can skew thermal readings and reduce the reliability of SBTI in detecting true perfusion deficits.
For the depiction of our data, descriptive statistics were used. Inferential statistics are less useful for small sample sizes like n = 11 because the ability to generalize findings to a larger population is severely limited due to high variability and the increased risk of errors. Therefore, using descriptive statistics provides a clear and accurate summary of the data without making unsupported generalizations.
4. Discussion
Despite progressive medical and surgical, as well as technical developments, no technique has been found that permits a uniform, simple valid, and cost-effective postoperative flap assessment. As a result, direct clinical assessment remains the gold standard [
5,
6]. The use of thermal imaging to assess perfusion is not a new concept, in the identification of perforators in free and pedicled perforator flaps [
16,
17] and perfusion monitoring in free and pedicled flaps [
13]. Multiple studies focus on the assessment of DIEP flaps [
1] and lower limb reconstruction with propeller flaps [
16] but, to our knowledge, no study was yet conducted using SBTI in head-/face-/and neck pedicled skin flaps. Further, we missed focus on implementation protocols and their obstacles in other studies.
In our study, we aimed to establish an implementable algorithm for SBTI use in the perioperative setting and evaluate its interpretability and accuracy in assessing cutaneous perfusion in pedicled flaps. The implementation of our SBTI protocol encountered initial challenges, including inconsistent image collection and exclusion of patients due to protocol deviations. Due to a late reaction on this topic, we had to exclude six patients, lowering our potential validity. However, with the involvement of a dedicated surgical team, these challenges were addressed, highlighting the importance of standardized protocols and interdisciplinary collaboration in adopting new technologies.
Our patient cohort consisted of individuals undergoing reconstructive surgery for various diagnoses, with diverse surgical techniques employed. Despite this heterogeneity, intraoperative assessments consistently indicated good perfusion, supporting the efficacy of our SBTI protocol in evaluating flap viability. Statistical analysis revealed consistent outcomes in both clinical and SBTI assessments, with SBTI demonstrating a greater variability at certain time points. Furthermore, the SBTI had a higher sensitivity due to our advancement flap which showed postoperative dehiscence. The dT was higher than 4 °C between the ROI in the center compared with the periphery of the flap, whereas the clinical assessment graded it by 2, not assuming a compromised perfusion state. Additionally, pedicled skin flaps showed preserved thermal signatures peri- and postoperatively, correlating with positive clinical outcomes and no further perfusion related complications.
The defined acquisition (T1–T4) of a thermal image signature as a surrogate marker for preserved tissue perfusion by SBTI proved to be simple, inexpensive, noninvasive, and efficient in the present sentinel study. The most relevant time points in SBTI seemed to be T2 and T3 in SBTI due to prognosis of perfusion complications in the postoperative frame, this should be further addressed in studies. Thermal imaging stands out among the assessment techniques due to its non-invasive nature and affordability, particularly when implemented via smartphone-based devices. It offers the advantage of intraoperative detection of vascular insults and provides valuable postoperative monitoring in pedicled flaps. However, there are some limitations of thermal imaging which can limit the accuracy. First, its utility is constrained by its limitation to surface temperature changes, and its specificity may vary depending on the circumstances. For example, using a cold wet gauze lying on the flap for several minutes, it normally takes 2–3 min until the superficial temperature in the TI resolves to its primary temperature. Depending on body composition, e.g., high subcutaneous fat proportion, which can decelerate rewarming of the tissue. Although cooling the tissue temperature may be of argument within perforator identification [
1], it is not useful for the perfusion assessment in pedicled skin flaps via SBTI. Second, the evaluation of flaps of the nose and the ears showed continuous cold areas; therefore, an assessment with thermal imaging can bring inaccurate results. As a third point, smoking and diabetes mellitus can have a negative impact in the assessment of skin flaps due to perfusion impairment. These three factors of influence in thermal imaging could not be precisely addressed within our study and should be the focus of further studies.
In conclusion, while this pilot study demonstrates promising potential for assessing pedicled skin flap perfusion using smartphone-based thermal imaging (SBTI), several limitations warrant cautious interpretation of the results. The study’s small sample size and limited validation against established clinical methods reduce the generalizability of the findings, highlighting the need for larger, more rigorous studies. The accuracy of SBTI was limited in small flaps and anatomically challenging regions, such as the nose and ears, and was influenced by external factors like surface temperature changes and patient comorbidities, such as smoking and diabetes mellitus. To enhance the reliability and clinical utility of SBTI, future research should focus on increasing sample sizes, strengthening validation protocols, developing advanced imaging techniques, and refining assessment time points. Addressing these limitations will help build on the promising results of this pilot study and improve the efficacy of SBTI in clinical practice.