3.3.2. Tissue Oxygenation (StO2)

A normal distribution of all time points analyzed within this study was detected. A statistically significant correlation between the different time points could be observed (each *p* < 0.05). Mean ranks differed significantly between the related samples time points I and II (*p* < 0.001) as well as between I and III (*p* = 0.048), whereas the values between time point II and III did not differ significantly (*p* = 0.076). This corresponded to a strong effect within the compared groups I and II (r = 0.6) and a moderately strong effect for group comparison I and III and II and III (r > 0.3).

#### 3.3.3. Near Infrared Perfusion Index (NIR)

A normal distribution for the measured values of all time points was seen. A statistically significant correlation between the different time points could be shown by Pearson and Spearman test with *p* < 0.01. There were significant differences in mean ranks between the related samples time points I and II as well as I and III (*p* = 0.002), whereas the values between time point II and III did not differ significantly (*p* = 0.213). This corresponded to a strong effect between the compred groups that showed significantly different mean ranks (r each > 0.5), whereas for group II and III only a weak effect could be shown (r = 0.235).

#### 3.3.4. Tissue Hemoglobin Index (THI)

The hypothesis of a normal distribution could not be rejected for time points I and III. Correlation analysis by Pearson and Spearman test could demonstrate a statistically significant correlation between the different time points with *p* < 0.001. Interestingly, statistically significant differences between mean ranks of time points I and II as well as II and III were shown, whereas time points I and III did not show such significant differences (*p* = 0.133). This corresponded to a strong effect between the compared groups that showed significantly different mean ranks (r each > 0.5), whereas for group I and III, a weak effect could be shown only (r = 0.284).

#### 3.3.5. Tissue Water Index (TWI)

For time points II and III, normally distributed values were observed. Both, Spearman and Pearson correlation test were able to show a statistically significant correlation between

all time points. Mean ranks differed significantly between all groups, corresponding to a strong effect between time point I and II (r = 0.626) and a moderate effect between the other time points with r > 0.4. man and Pearson correlation test were able to show a statistically significant correlation between all time points. Mean ranks differed significantly between all groups, corresponding to a strong effect between time point I and II (r = 0.626) and a moderate effect between the other time points with r > 0.4.

The hypothesis of a normal distribution could not be rejected for time points I and III. Correlation analysis by Pearson and Spearman test could demonstrate a statistically significant correlation between the different time points with *p* < 0.001. Interestingly, statistically significant differences between mean ranks of time points I and II as well as II and III were shown, whereas time points I and III did not show such significant differences (*p* = 0.133). This corresponded to a strong effect between the compared groups that showed significantly different mean ranks (r each > 0.5), whereas for group I and III, a

For time points II and III, normally distributed values were observed. Both, Spear-

#### 3.3.6. Return-to-Perfusion Measurement 3.3.6. Return-to-Perfusion Measurement

*J. Pers. Med.* **2021**, *11*, x FOR PEER REVIEW 9 of 17

3.3.4. Tissue Hemoglobin Index (THI)

weak effect could be shown only (r = 0.284).

Shapiro–Wilk and Kolmogorov–Smirnov tests could not reject the hypothesis of normal distribution for StO2, NIR, THI and TWI. There was a statistically significant correlation between the RTP-value and the MAT-measurements for StO<sup>2</sup> (*p* < 0.001), THI (*p* = 0.004) and TWI (*p* = 0.011), whereas no significant correlation could be shown for NIR-index (*p* = 0.179) (Figure 6). Shapiro–Wilk and Kolmogorov–Smirnov tests could not reject the hypothesis of normal distribution for StO2, NIR, THI and TWI. There was a statistically significant correlation between the RTP-value and the MAT-measurements for StO2 (*p* < 0.001), THI (*p* = 0.004) and TWI (p=0.011), whereas no significant correlation could be shown for NIRindex (*p* = 0.179) (Figure 6).

**Figure 6.** Boxplot diagram of the Return-to-Perfusion (RTP) measurements. The red box highlights the area between the 1st and 3rd quartile. The stars indicate the different subjects' measurements that had a MAT >8 s. As can be observed, some values lie between the 1st and the 3rd quartile, whereas others are located within the 4 inter quartiles range (IQR). Others lie due to their extreme values beyond these areas and are therefore considered outliers. With few exceptions, the measurements defined as outliers belong to the same patients. As can be seen, RTP-values for StO2, THI and TWI of patients with a clear pathological MAT typically lie beyond the 4 IQR (Nr. 30, 35, 60, 108). Others, with a MAT of 9–15 s–depending on the definition–rated as pathologic or non-pathologic, typically fall within the 4 IQR. The allocation of the values on the X-axis has no meaning. **Figure 6.** Boxplot diagram of the Return-to-Perfusion (RTP) measurements. The red box highlights the area between the 1st and 3rd quartile. The stars indicate the different subjects' measurements that had a MAT >8 s. As can be observed, some values lie between the 1st and the 3rd quartile, whereas others are located within the 4 inter quartiles range (IQR). Others lie due to their extreme values beyond these areas and are therefore considered outliers. With few exceptions, the measurements defined as outliers belong to the same patients. As can be seen, RTP-values for StO<sup>2</sup> , THI and TWI of patients with a clear pathological MAT typically lie beyond the 4 IQR (Nr. 30, 35, 60, 108). Others, with a MAT of 9–15 s–depending on the definition–rated as pathologic or non-pathologic, typically fall within the 4 IQR. The allocation of the values on the X-axis has no meaning.

#### *3.4. Patient Case*

A 63-year-old patient presented with an oral mucosal lesion of the right floor of the mouth that had been present for a year. Clinically, there was an ulcerating lesion of about 2 cm. During staging, no further suspicions lesions could be detected; radiologically, neither lymphatic, nor osseus metastases were found. The interdisciplinary head and neck tumor board recommended resection of the tumor and bilateral neck dissection. To cover the defect, we planned to harvest a RFFF. However, the MAT showed a very poor perfusion of the right arm with a reperfusion time of over 20 s; the results could be confirmed by HSI. When the MAT was performed of the left arm, a time to reperfusion of 11 s was measured. However, the HSI measurement showed adequate perfusion with a satisfactory RTP-value for StO2, NIR, THI, and TWI (Figure 7). Considering the measurable reperfusion, the decision was made to harvest a RFFF from the left arm. With constant monitoring of the oxygen saturation by means of pulse oximetry, the RFFF could be harvested without complications. In the postoperative follow-up, the graft was adequately perfused and healed well (Figure 8).

*3.4. Patient Case* 

and healed well (Figure 8).

A 63-year-old patient presented with an oral mucosal lesion of the right floor of the mouth that had been present for a year. Clinically, there was an ulcerating lesion of about 2 cm. During staging, no further suspicions lesions could be detected; radiologically, neither lymphatic, nor osseus metastases were found. The interdisciplinary head and neck tumor board recommended resection of the tumor and bilateral neck dissection. To cover the defect, we planned to harvest a RFFF. However, the MAT showed a very poor perfusion of the right arm with a reperfusion time of over 20 s; the results could be confirmed by HSI. When the MAT was performed of the left arm, a time to reperfusion of 11 s was measured. However, the HSI measurement showed adequate perfusion with a satisfactory RTP-value for StO2, NIR, THI, and TWI (Figure 7). Considering the measurable reperfusion, the decision was made to harvest a RFFF from the left arm. With constant monitoring of the oxygen saturation by means of pulse oximetry, the RFFF could be harvested

**Figure 7.** In (**A**) the measurements at the three different time points are shown. A baseline measurement was taken (time point I, both arteries open), then both, the radial and the ulnar artery were occluded (time point II) and at time point III, the ulnar artery was released. At time point I, tissue oxygenation (StO<sup>2</sup> ) was 45%. After occluding the arteries, the values decreased to 35.67% and increased again after release of the ulnar artery (51%) as can be seen in (**B**). This observation is consistent with the measurements of the NIR perfusion index (NIR) showing values of 51.33 at time point I, 40.67 at time point II and 55.67 at time point III (see B). The image shows very impressively the trend of the hemoglobin index (THI) during the course of the experiment. With both arteries open, the THI was 32.67; after occlusion of both arteries, the values decreased to 13 and increased again after release of the ulnar artery (43). As can be seen in the images in A, tissue water index (TWI) increased during artery occlusion showing values of 45.33 while both, the baseline measurement and the measurement after release of the ulnar artery showed values of 35. The Return-to-perfusion (RTP) value indicates the difference in percent between the measurements at time point III and time point I. Here, RTP-values for StO<sup>2</sup> were −13.33%, for NIR −8.46%, for THI −31.62% and for TWI −1.91%. Those values corresponded to a strong return to perfusion after release of the ulnar artery showing a save perfusion of the hand by the ulnar artery alone.

artery showing a save perfusion of the hand by the ulnar artery alone.

**Figure 7.** In (**A**) the measurements at the three different time points are shown. A baseline measurement was taken (time point I, both arteries open), then both, the radial and the ulnar artery were occluded (time point II) and at time point III, the ulnar artery was released. At time point I, tissue oxygenation (StO2) was 45%. After occluding the arteries, the values decreased to 35.67% and increased again after release of the ulnar artery (51%) as can be seen in (**B**). This observation is consistent with the measurements of the NIR perfusion index (NIR) showing values of 51.33 at time point I, 40.67 at time point II and 55.67 at time point III (see B). The image shows very impressively the trend of the hemoglobin index (THI) during the course of the experiment. With both arteries open, the THI was 32.67; after occlusion of both arteries, the values decreased to 13 and increased again after release of the ulnar artery (43). As can be seen in the images in A, tissue water index (TWI) increased during artery occlusion showing values of 45.33 while both, the baseline measurement and the measurement after release of the ulnar artery showed values of 35. The Return-toperfusion (RTP) value indicates the difference in percent between the measurements at time point

for TWI −1.91%. Those values corresponded to a strong return to perfusion after release of the ulnar

**Figure 8.** A TIVITA scan immediately after transplantation of the RFFF into the floor of the mouth is shown on the left. The circular mark (white dotted line) indicates the graft inside the patient's mouth. The measurement shows a (typical) initially very high oxygenation of the graft with correspondingly high NIR values and adequately low THI values. In the course of time, up to postoperative day 3 shown on the right, a reduction of the StO2 and NIR values and a homogenization of the measured values over the area of the graft are visible. Based on previous studies [21], it could be determined that this process is typical for the proper healing and perfusion of the graft. **Figure 8.** A TIVITA scan immediately after transplantation of the RFFF into the floor of the mouth is shown on the left. The circular mark (white dotted line) indicates the graft inside the patient's mouth. The measurement shows a (typical) initially very high oxygenation of the graft with correspondingly high NIR values and adequately low THI values. In the course of time, up to postoperative day 3 shown on the right, a reduction of the StO<sup>2</sup> and NIR values and a homogenization of the measured values over the area of the graft are visible. Based on previous studies [21], it could be determined that this process is typical for the proper healing and perfusion of the graft.

#### **4. Discussion 4. Discussion**

The radial artery is located in the lateral intermuscular septum between the brachioradialis and flexor carpi radialis muscles [1]. Entering the hand, the radial artery gives rise to the princeps pollicis artery and radial indices artery. The deep palmar arch is formed by the dorsal radial artery and the deep branch of the ulnar artery. Four palmar metacarpal arteries arise from the deep palmar arch and converge with the common palmar digital arteries. The superficial palmar arch is formed by the ulnar artery and the superficial branch of the radial artery with four common palmar digital arteries arising from the arch. The common digital arteries then divide into two proper palmar digital arteries [22,27,28] (Figure 9). The radial artery is located in the lateral intermuscular septum between the brachioradialis and flexor carpi radialis muscles [1]. Entering the hand, the radial artery gives rise to the princeps pollicis artery and radial indices artery. The deep palmar arch is formed by the dorsal radial artery and the deep branch of the ulnar artery. Four palmar metacarpal arteries arise from the deep palmar arch and converge with the common palmar digital arteries. The superficial palmar arch is formed by the ulnar artery and the superficial branch of the radial artery with four common palmar digital arteries arising from the arch. The common digital arteries then divide into two proper palmar digital arteries [22,27,28] (Figure 9). *J. Pers. Med.* **2021**, *11*, x FOR PEER REVIEW 12 of 17

**Figure 9.** Palmar arch of the right hand. Entering the hand, the radial artery gives rise to the princeps pollicis artery and radial indices artery. The deep palmar arch is formed by the dorsal radial artery and the deep branch of the ulnar artery. The superficial palmar arch is formed by the ulnar artery and the superficial branch of the radial artery. **Figure 9.** Palmar arch of the right hand. Entering the hand, the radial artery gives rise to the princeps pollicis artery and radial indices artery. The deep palmar arch is formed by the dorsal radial artery and the deep branch of the ulnar artery. The superficial palmar arch is formed by the ulnar artery and the superficial branch of the radial artery.

The vascular abnormality leading to impaired perfusion after raising the RFFF is a combined condition of an incomplete ulnar arterial supply to the hand and a missing com-The vascular abnormality leading to impaired perfusion after raising the RFFF is a combined condition of an incomplete ulnar arterial supply to the hand and a missing

munication between deep and superficial palmar arch. Coleman and Anson reported 12% of specimens to show a combination of the two abnormalities [29]. Strauch et al. reported

tively. Other studies found the collateral circulation to be absent in 2% up to 20% of cases

In this regard, a higher incidence of pathological results in Allen's test could be expected. In a large study with 1000 patients undergoing cardiac catheterization, 49% had a normal Allen's test (cut-off time <5 s). The authors classified 24% as borderline (5–9 s) and 27% as abnormal (>10 s) [22]. In 1990, Hosokawa et al. showed 5.8% of 1470 patients examined within the hospital to have an abnormal Allen's test (time until recover of color 5 s). Unilateral abnormality was observed in 4.4%, bilateral abnormality in 1.4% of cases [2]. The incidence of an abnormal test increased with age (incidence of abnormality >80 years 6.9%). Since the average age of developing oral cancer is 62 years [33], a more accurate and safer test method is needed to ensure adequate blood supply by the ulnar artery when

Nuckols et al. reported a sensitivity of 65% and a specificity of 76% (positive predictive value (PPV) = 93%, negative predictive value (NPV) = 35%) for Allen's test with a cutoff time of 5 s [28]. This corresponds to the results of Husum et al. who indicated a NPV of 0.992, which yields a false-positive rate of to a 0.8% (1/100 hands with normal Allen's test and an inadequate collateral circulation) [28]. According to this study, a normal test result would incorrectly indicate inadequacy in about 50% of cases [28]. A systematic review by Romeu-Bordas et al. evaluated the reliability and validity of Allen's test in patients prior to radial artery puncture. They concluded Allen's test to show inadequate diagnostic validity for screening deficits in the collateral circulation. Because of this, Allen's test is termed to be no adequate predictor of hand ischemia: "Therefore, Allen's test should not be systematized prior to performing an arterial puncture as an isolated screening test for collateral arterial circulation deficits of the hand and should not be considered

[28]. Complete superficial arches occur in 84% up to 90% of cases [4,27,30–32].

harvesting RFFF.

communication between deep and superficial palmar arch. Coleman and Anson reported 12% of specimens to show a combination of the two abnormalities [29]. Strauch et al. reported the superficial arch and the deep arch to be incomplete in 21% and in 3% of cases, respectively. Other studies found the collateral circulation to be absent in 2% up to 20% of cases [28]. Complete superficial arches occur in 84% up to 90% of cases [4,27,30–32].

In this regard, a higher incidence of pathological results in Allen's test could be expected. In a large study with 1000 patients undergoing cardiac catheterization, 49% had a normal Allen's test (cut-off time <5 s). The authors classified 24% as borderline (5–9 s) and 27% as abnormal (>10 s) [22]. In 1990, Hosokawa et al. showed 5.8% of 1470 patients examined within the hospital to have an abnormal Allen's test (time until recover of color 5 s). Unilateral abnormality was observed in 4.4%, bilateral abnormality in 1.4% of cases [2]. The incidence of an abnormal test increased with age (incidence of abnormality >80 years 6.9%). Since the average age of developing oral cancer is 62 years [33], a more accurate and safer test method is needed to ensure adequate blood supply by the ulnar artery when harvesting RFFF.

Nuckols et al. reported a sensitivity of 65% and a specificity of 76% (positive predictive value (PPV) = 93%, negative predictive value (NPV) = 35%) for Allen's test with a cut-off time of 5 s [28]. This corresponds to the results of Husum et al. who indicated a NPV of 0.992, which yields a false-positive rate of to a 0.8% (1/100 hands with normal Allen's test and an inadequate collateral circulation) [28]. According to this study, a normal test result would incorrectly indicate inadequacy in about 50% of cases [28]. A systematic review by Romeu-Bordas et al. evaluated the reliability and validity of Allen's test in patients prior to radial artery puncture. They concluded Allen's test to show inadequate diagnostic validity for screening deficits in the collateral circulation. Because of this, Allen's test is termed to be no adequate predictor of hand ischemia: "Therefore, Allen's test should not be systematized prior to performing an arterial puncture as an isolated screening test for collateral arterial circulation deficits of the hand and should not be considered an absolute contraindication for performing a transradial puncture presenting an abnormal result in the Allen's test" [8]. False-negative Allen's test could result in hand ischemia and necrosis, whereas false-positive test results cause a change of primary treatment plan, possibly resulting in a suboptimal therapy situation [4]. Initially, ischemic hand complications (IHC) appear in the form of pallor and progressive darkening of the skin. Chronic complications include pain, cold intolerance, ulceration, tissue necrosis and gangrene of the digits [4].

Therefore, a secure method for vascular assessment is needed. In daily routine, some supplemental properties are necessary to make the test feasible: the test should be noninvasive, fast as well as easy to perform and the evaluation needs to be objective and reproducible. Furthermore, the method needs to have a good predictive ability with a high sensitivity, specificity and accuracy. In this study, HSI was shown to detect perfusion deficits during MAT. HSI provides both, topographical and spectral information in an objective, reproductive and measurable manner. Combinations of values allows drawing conclusions about tissue perfusion. High THI and low StO<sup>2</sup> indicates venous congestion, whereas low THI and low StO<sup>2</sup> points to an arterial occlusion. A high NIR and a low StO<sup>2</sup> indicated that deep tissue perfusion is given whereas superficial layers are undersupplied, whereas the contrary case points to a critical situation as superficial supply can clinically hide saturation problems in deeper tissue layers [21]. Moreover, if the reliability of Allen's test as a screening tool is shown by a high number of successful and complication-free radial forearm free flap transfers, the remaining percentage of hand ischemia in the presence of non-pathological test results suggests the need for a more secure measurement method to increase patients' safety.

The results of this study indicate that in patients with a non-pathological MAT (bloodrefill time of less than nine seconds) all parameters collected during hyperspectral imaging significantly differ between both, the baseline measurement and the measurements taken during complete artery occlusion as well as between the measurement after release of the ulnar artery and during artery occlusion. Furthermore, THI-RTP values correlated

with the MAT results. Hence, it can be concluded that the calculated ratio between the HSI-measurements at the beginning and the end of the test (RTP value) representing the ability to full reperfusion by only the ulnar artery reflects the time to reperfusion measured during MAT.

On the one hand, this indicates a reliable differentiation between perfusion and occlusion–as already shown by Grambow et al. in a rat in vivo model [26]–as well as the confirmation of non-pathological MAT. Therefore, the use of hyperspectral imaging for additional diagnostics in combination with MAT would be a useful tool to verify the correct performance of the test (differentiation of occlusion and perfusion) as well as to confirm the final diagnosis of a non-pathological MAT. With the aid of the RTP-value, a correlation between hyperspectral imaging and MAT could further be supported. We could not only show a safe differentiation between perfusion and occlusion status during a nonpathological MAT, but the system can also detect a pathological reperfusion based on the different parameters. This can be observed by the no longer significant differences between time points I and II as well as between II and III, but–due to the impaired reperfusion at time point III–significant differences between perfusion (time point I) and occlusion (time point II) as well as between perfusion (time point I) and reperfusion (time point III). Interestingly, we could show a statistically significant correlation between the RTP-value and the MAT results. This is in accordance with the observed outliers when comparing the hyperspectral data of patients with a non-pathological MAT and those with a pathological MAT. Here, it is clearly shown that the hyperspectral measurement values that are far beyond the norm correlate well with certainly pathological MAT values (AT max.).

Due to the overall group size and especially the rather small number of pathological MAT readings in our study, it was not possible to define safe cut-off values, rendering the fitting of neural-networks impossible. Nonetheless, the trends observed in this study indicate a potential use of hyperspectral imaging for reperfusion analysis thus offering an alternative or supplementary method to the gold standard.

Based on current data, an advantage of HSI over clinical assessment alone has already been observed for perfusion monitoring of microvascular anastomosed grafts, particularly by inexperienced personnel. Compared to visual assessment of the hand reperfusion during MAT, HSI offers some clear advantages as it provides an objective, reproducible method also feasible for non-medical personnel that has no interobserver error and–in contrast to the MAT–gives a visual and measurable feedback in case of insufficient artery occlusion or potential test error due to palmar hyperextension.

HSI is used both clinically and experimentally for numerous indications: In anesthesia and ICU, the technique is already used in critically ill patients to monitor micro- and macro-circulatory changes, tissue perfusion and oedema formation to reduce the negative effects of hemodynamic incoherence. Prior to HSI, skin mottling and capillary refill time were used to assess hemodynamic parameter, but, as in MAT, the inter-observer variability demonstrated contradictory findings and a variety of cut-off values were suggested [34]. In vascular surgery, HSI is used to provide objective decision criteria for determining the extent of amputation and to make predictions about the chance of healing of the amputation wound [35]. What is more, HSI has been frequently used in monitoring perfusion in microvascular anastomoses, both in experimental setups and in the clinical practice. Here, a clear advantage over the visual assessment of the grafts could be shown in numerous studies [21,26,36]. Similar results could also be shown in transplant medicine. Here Sucher et al. proved that by means of HSI it is possible to objectively assess whether an organ (the kidney) is suitable for transplantation even before the surgery. In addition, it is possible to check immediately after the transplantation whether the blood vessels have been sutured correctly and the organ is sufficiently supplied with blood. Until now, this was also decided by visual assessment, so the new HSI technique offers a clear advantage [37]. In visceral surgery, HSI is used for blood flow analysis, for example, to determine the extent of resection in the case of mesenteric ischemia, but also to assess anastomoses, as well as tubular gastric blood flow in esophageal resections [38,39]. At the current time,

numerous other applications of the hyperspectral camera have already been developed. In addition to special products for the analysis of wounds and soft tissue, camera systems for the operating room and an endoscopic version have also been developed. For the resection of brain tumors, MRI scans are taken preoperatively in order to locate the tumor and subsequently plan the surgical intervention. Intraoperatively, there is currently a lack of tools to locate the tumor with certainty. Data to date indicate that it will most likely be possible to identify the tumor and resect it safely using this method [40]. This large number of potential application areas and the already established use in everyday clinical life point to a great benefit of hyperspectral technology in the future. In particular, subjective assessment criteria can be replaced by means of this technique and thus the therapy of patients can be improved.

As the ambient light conditions affect the parameter values, cautious interpretation is demanded. To the authors knowledge, this is the first study assessing the feasibility of HSI to collateral circulation prior to RFFF harvest. Regarding the number of cases within this study, the present results need to be classified as of descriptive nature. Further clinical studies must be conducted in order to set cut-off values indicating a save arterial refill during HSI assisted MAT. In general, perfusion markers should return to the level of the baseline measurement after releasing the ulnar artery. If the parameters stay low, an adequate perfusion is not guaranteed. What is more, in addition to the measurements performed during this study, further measurements at later time points could have been performed to assess the long-term which could have provided further insight. However, in this study, further measurements were deliberately omitted because we aimed to (1) to ensure the greatest comparability possible between the gold standard and the new method and (2) develop a new method for everyday clinical use. This method should be as simple, fast, and objective as possible, so further, time-consuming follow-up measurements were not necessary. With this study, we aimed to investigate the new method exactly as it should be applied in clinical practice. Despite the more limited assessability of long-term perfusion due to the reduction in the number of measurements, we were able to demonstrate that this method can clearly distinguish between appropriate, questionable, and non-appropriate donor sites. Nevertheless, especially in patients with impaired perfusion (questionable and non-appropriate), the assessment of tissue perfusion in later recordings would be of interest. A clear distinction could be made between delayed complete reperfusion and lack of reperfusion, thus identifying patients who would be expected to develop long-term tissue damage.

As the assessment of collateral perfusion by the ulnar artery is not just mandatory in the field of flap raising, but also for arterial puncture in anesthesia, coronary artery intervention and bypass, there is a high need for such a technique.

#### **5. Conclusions**

This study was able to show that, using HSI, a safe differentiation between perfusion and occlusion is possible and that this method has a good correlation to the present gold standard, the MAT. Therefore, HSI could serve as an additional method for the assessment of the collateral circulation of the hand prior to invasive interventions involving the damage or harvesting of radial artery. HSI provides some advantages over the single visual assessment, as it provides objective, reproducible results without interobserver error, can also be applied by non-medical personnel, and gives a visual and measurable feedback. Yet, limitations are imposed by the poor data situation and examination-related measurement errors.

**Author Contributions:** Conceptualization, P.W.K., D.H. and D.G.E.T.; methodology, P.W.K., D.H. and D.G.E.T.; software, P.W.K. and D.G.E.T.; validation, P.W.K., D.G.E.T., D.H., P.B., R.K. and S.K.; formal analysis, P.B., D.H. and R.K.; investigation, P.B. and D.H.; resources, P.W.K. and D.G.E.T.; data curation, P.B. and D.H.; writing—original draft preparation, D.H., P.B. and R.K.; writing—review and editing, D.H., P.B., D.G.E.T., R.K., S.K. and P.W.K.; visualization, D.H.; supervision, P.W.K. and D.H.; project administration, P.W.K. and D.H. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of Rhineland-Palatinate (registration number: 2020-15022\_1, date of approval: 06/25/2020).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data supporting the conclusions of the article is included within the article. The raw data analyzed during the current study is available from the corresponding author on reasonable request.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**

