A Systematic Review and Critical Appraisal of Peri-Procedural Tissue Perfusion Techniques and their Clinical Value in Patients with Peripheral Arterial Disease

protocol, letter, congress (n = 32) Not peri-operative (n = 22)


INTRODUCTION
In peripheral arterial disease (PAD), macrovascular stenoses or occlusions cause an inadequate blood supply to the lower limbs. 1 Patients with PAD may therefore suffer from intermittent claudication (IC), rest pain, or non-healing wounds, which all lead to an impaired quality of life. 2 IC is the most common presenting symptom of PAD, which in 5% of patients progresses to chronic limb threatening ischaemia (CLTI). 3 To prevent major tissue loss in patients with CLTI, revascularisation is most appropriate. 2 Clinical outcomes after a revascularisation procedure remain unpredictable when current imaging techniques are used. 4 These techniques mainly focus on assessment of the macrovasculature and do not include the assessment of the microvasculature, which is pivotal in patients with CLTI. Therefore, satisfactory results might be accompanied by poor clinical outcomes and early amputations. Ideally, microcirculation changes should be determined during a revascularisation procedure to guide the vascular surgeon or interventionalist on how extensive the procedure must be to improve local tissue perfusion. 5,6 Many invasive and non-invasive techniques have been introduced in recent years that claim to enable the visualisation and quantification of the microvasculature and tissue perfusion. Unfortunately, none of these techniques is currently widely used peri-procedurally. Non-invasive techniques include laser speckle contrast imaging (LSCI), micro-lightguide spectrophotometry (O 2 C), magnetic resonance imaging (MRI) perfusion (MRIp), near infrared spectroscopy (NIRS), skin perfusion pressure (SPP), and plantar thermography (PT). Invasive techniques include two dimensional perfusion angiography (2D-PA), contrast enhanced ultrasound (CEUS), computed tomography (CT) perfusion imaging, and indocyanine green angiography (ICGA).
The aim of this systematic review was to provide an up to date overview of the peri-procedural applicability of the aforementioned techniques, a brief description of the techniques, their diagnostic accuracy in assessing tissue perfusion, and their relationship to clinical outcomes.

METHODS
This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic review and Meta-Analysis (PRISMA) guidelines. 7 Eligible articles were included if they described a technique to determine tissue perfusion, in patients with PAD, in a peri-procedural setting. Articles had to have focused on perfusion imaging before and within 24 hours of a revascularisation procedure, to determine the effect of the intervention. Imaging techniques were compared with well known conventional techniques like ankle brachial pressure index (ABPI), toe brachial index (TBI), and clinical outcomes such as wound status, improvement in walking distance, or Fontaine classification. Included articles were full text articles published between 1 January 2010 and 31 December 2020. Exclusion criteria were articles that involved experimental treatment with stem cell therapy, that were not performed periprocedurally, or that were animal studies. Furthermore, studies with fewer than 10 patients, commentaries, guidelines and letters to the editor were excluded.

Literature search
Four electronic databases were searched for eligible articles: MEDLINE, Embase, CINAHL, and the Cochrane Central Register of Controlled Trials. The database search was performed using medical subject headings (MeSH) terms for "peripheral arterial disease", "peripheral vascular diseases", "diagnostic imaging", "diagnostic techniques, cardiovascular", "photoacoustic techniques", "microcirculation", "perfusion", "vascular surgical procedures", and "operating rooms" complemented with the keywords "endovascular technique", "revascularisation", and "PTA". Free text words were used to avoid missing recently published manuscripts without a MeSH label. The complete search strategy is available in Supplementary Appendix 1. The titles and abstracts of the studies were independently screened by two authors (B.W. and K.F.M.), who were blinded to the study authors and journal titles. Disagreements were discussed by the two authors. Articles considered for inclusion were independently reviewed by the same two authors. Disagreements were solved by discussion or by consensus after consulting a third author (R.H.G.).

Data collection
The details of eligible articles were collected by two authors (B.W. and K.F.M.) per study and organised using a predetermined data collection form. Extracted data were grouped per technique and structured regarding characteristics, research goal, comparison with conventional techniques, and clinical outcomes. Technical properties, advantages, disadvantages, and clinical applications of the respective techniques were described. Outcomes of interest comprised clinical applicability of the technique, diagnostic accuracy in assessing tissue perfusion, and their relationship with clinical outcomes. The QUADAS-2 tool was used by two independent observers (B.W. and K.F.M.) to assess the risk of bias and applicability of the studies. 8 This tool was used to assess the risk of bias in patient selection, blinded assessment of the index test from the reference standard, and the flow and timing of the study and its measurements. Patient selection, the index test, and reference standard were assessed for concerns regarding applicability. If there was no mention of a reference standard, the risk of bias of index test and reference standard were scored as unknown. 8 For techniques reported in an eligible article, the technical background was described and study outcomes are presented in Table 2.

RESULTS
The database searches resulted in 3 910 identified records, of which 569 were duplicates. After title and abstract screening, 3 230 articles were excluded according to the exclusion criteria. Extensive full article review resulted in the exclusion of another 85 articles. Finally, 26 articles, describing 10 techniques, were found to be eligible for inclusion. The study flow diagram is shown in Fig. 1.
Details of the included studies are presented in Tables 1 e 3. Six non-invasive techniques were described in 11 articles including in total 523 patients (Fontaine II e IV). Four invasive techniques were described in 16 articles including in total 653 patients (Fontaine II e IV). Table 1 presents the characteristics of the included studies; Table 2 provides the research goals and clinical outcomes data. Table 3 summarises the advantages and disadvantages of the respective techniques. In two articles, 9,10 two different tissue perfusion techniques were described. The risk of bias and applicability concerns of the included studies according to QUADAS-2 tool are shown in Table 4 and Fig. 2.

Non-invasive techniques
Laser speckle contrast imaging. LSCI uses a coherent laser light to illuminate tissue. This coherent laser light scatters on the surface of tissue, creating an interference pattern called a speckle pattern. 10 The motion of red blood cells (RBCs) in the microcirculation changes the speckle pattern over time, resulting in blurring of the image. Blurring is increased by a higher velocity or number of RBCs and displayed in real time blood flow maps. 10,11 Magnetic resonance imaging perfusion. MRIp uses arterial spin labelling (ASL) to measure absolute tissue perfusion. ASL uses the inversion of water molecules in blood as an endogenous contrast agent. 12 By subtracting tagged images from a control image, a perfusion signal is obtained. 12 Regions of interest (ROIs) are drawn in muscle groups, from where a perfusion time course is extracted. Thereafter, parameters of interest can be extracted. 12 Other promising MRI techniques do measure tissue perfusion but did not fulfil the inclusion criteria of this review. These techniques are described in more detail in several other studies. 13e15 Micro-lightguide spectrophotometry. O 2 C, or "oxygen to see", uses a combination of laser Doppler flowmetry and spectroscopy. O 2 C is determined using white light and laser light, with a penetration depth up to 2 mm reaching the dermis. 16,17 RBC movement causes a laser Doppler shift, which is detected as blood velocity. Spectroscopy is used to determine the amount of haemoglobin in a skin volume. Overall flow can be extracted into the following parameters: oxygen saturation (sO 2 ); relative haemoglobin (rHB) amount; relative blood flow; and blood flow velocity. 16 Near-infrared spectroscopy. NIRS measures single tissue oxygen saturation, continuously, with a maximum imaging depth of approximately 15 e 20 mm, including muscle tissue. 18,19 Measurements are performed using an optode, housing the light source (red and near infrared spectrum). Light is emitted through the sampled tissue and is partly absorbed and reflected, which is recorded by photodetectors. 19,20 Oxygenated and deoxygenated haemoglobin have  different absorption spectra for red and near infrared light, making it possible to determine the proportion of oxygenated haemoglobin using NIRS. The single tissue oxygen saturation value in the measured tissue therefore reflects the ratio (%) between concentrations of oxygenated and deoxygenated haemoglobin. 19

Study outcomes
Two studies, including 14 and 30 patients with Fontaine III e IV PAD, respectively, were found. The study by Boezeman et al. showed no significant improvement of single tissue oxygen saturation directly after revascularisation. 19 ABPI or TBI showed a significant improvement after four weeks; however, no correlation with single tissue oxygen saturation directly after revascularisation was determined. 19 The study by Kundra et al. showed a significant improvement in single tissue oxygen saturation after surgery, which correlated with Doppler signals. 21 Plantar thermography. Thermography detects infrared radiation, typically emitted from skin, which presents regional temperature as a heat zone image. Both the plantar and dorsal foot are measured using a digital infrared thermal imaging system, with a standardised temperature range of 17 C e 34 C. Software converts the temperature into a colour coded image. 22 Skin perfusion pressure. SPP can be measured on the dorsal and plantar surface of the foot. A laser Doppler probe placed beneath a blood pressure cuff determines the systolic blood pressure needed to restore the blood circulation in the microcirculation. 9,10,23e27

Study outcomes
All six SPP studies, including 315 patients, had a cohort ranging from 16 to 147 patients with Fontaine IIb e IV PAD, showed a significantly improved SPP on the dorsal or plantar side of the foot after intervention. 9,10,23e25, 27 Ikeoka et al. also showed a significant improvement in ABPI and ankle pressure. 25

Invasive techniques
Two dimensional perfusion angiography. 2D-PA determines tissue perfusion based on digital subtraction angiography (DSA) images, acquired during endovascular treatment. 9,28e33 ROIs are drawn to determine region specific time attenuation curves (TACs), also called time density curves. A TAC shows a graph comparing contrast intensification against time. 31 From this TAC, a wide range of parameters such as arrival time, time to peak (TTP) and wash in rate can be extracted. 9,28e34 Furthermore, ratios for outflow and inflow can be determined (i.e., TTP outflow /TTP inflow ), to overcome potential limitations of standardised pump injection. 28 These parameters are used to convert DSA images into colour coded images.

Study outcomes
All eight 2D-PA studies, which included 257 patients, ranging from 16 to 68 patients with Fontaine IIb e IV PAD, showed that an increase in blood flow after revascularisation could be instantly measured and quantified using multiple parameters. 9,28e34 Furthermore, correlations were found between the ABPI, TBI or SPP and 2D-PA parameters. 9   To evaluate the use of parametric colour coding and analysis of TACs as a real time quantitative measure of perfusion after EVT Washout phase parameters showed a significant reduction in time required for contrast to decay to a specified percentage after peak (T 90% , T 80% , T 70%, T 60% , and T 50%). Percentages of contrast decay at specified time intervals after peak (I 1s , I 2s , I 3s , I 4s , and I 5s ) were increased significantly To assess whether the regional evaluation of foot blood volume may guide direct revascularisation and if it will lead to better perfusion improvement than indirect revascularisation

Median
Contrast enhanced ultrasound. CEUS uses an intravenous injection of microbubbles combined with ultrasound, which allows for analysis of the intravascular distribution of this hyperechogenic contrast agent over time. 35 CEUS can be performed with a penetration depth of 4 -15 cm. 35,36 Quantitative measurements are performed by determining the TTP contrast intensity. 36 Computed tomography perfusion imaging. CT perfusion imaging uses a flat panel detector angiographic system to capture real time parenchymal blood volume, using an automated intra-arterial injection of an iodine based contrast agent. 37 Imaging of the entire foot and ankle can be performed. Post-processing software is used for three dimensional reconstruction and conversion to colour coded perfusion maps.
Indocyanine green angiography. ICGa uses a laser light source, in the near infrared light spectrum (650 e 900 nm), combined with intravenous injection of ICG. 13,38 A charged coupled device camera captures a real time image. 39,40 ICGa provides approximately 3 e 7 mm of tissue penetration, 41e43 and could therefore be used as an indicator of superficial tissue perfusion. 41 Dedicated software can be used to analyse quantitatively ICG intensity in the entire image or in a chosen ROI. 39,40,42e45 Study outcomes Six studies including 335 patients, ranging from 30 to 101 patients with Fontaine IIb e IV PAD, were found. A significant improvement in perfusion directly after revascularisation, by means of multiple ICGa perfusion parameters, was found. 39,40,42e45 All studies measured ABPI and showed significant improvement after revascularisation. Colvard et al., 39 Patel et al., 40 and Rother et al. 42 showed a correlation of different ICGa perfusion parameters with an overall change in ABPI. Mironov et al. showed that none of the perfusion variables was a significant predictor of wound healing. 44 However, the studies Impairing motion artefacts and foot deformation between prerevascularisation and post-revascularisation images 37 The use of ionising radiation 37 High cost 37 ICGA ICG is a water soluble, non-radioactive, nonionising, and non-toxic contrast agent 40 Low penetration depth 13 Expensive imaging systems 13 Measurements affected by temperature and medication (vasoactive substances) 13 The need for intravenous contrast administration 13 LSCI ¼ laser speckle contrast imaging; MRIp ¼ magnetic resonance imaging perfusion; O 2 C ¼ micro-lightguide spectrophotometry; NIRS ¼ near infrared spectroscopy; PT ¼ plantar thermography; SPP ¼ skin perfusion pressure; DSA ¼ digital subtraction angiography; 2D-PA ¼ two dimensional perfusion angiography; CEUS ¼ contrast enhanced ultrasound; CT-PI ¼ computed tomography perfusion imaging; ICGA ¼ indocyanine green (ICG) angiography.
were difficult to compare owing to heterogeneity in imaging devices and protocols.

DISCUSSION
In this systematic review, 10 techniques were found that assessed tissue perfusion in patients suffering Fontaine II e IV PAD before and within 24 hours after revascularisation procedures. Twenty-three of the 26 included studies had a small sample size (n < 50) and only investigated the feasibility of determining the change in tissue perfusion with these techniques. No diagnostic accuracy or correlation with treatment outcomes, such as wound infection and healing, amputation rate, need for re-admission, quality of life or mobility, were demonstrated. 46 The results of the 26 eligible studies, which were mostly of poor quality according to the QUADAS-2 tool, were not sufficient to substantiate implementation in daily clinical practice yet. Comparing and pooling of data and results was not  meaningful because of the limited number of studies per technique. Besides, heterogeneity in inclusion criteria, patient selection, measurement protocols, follow up time, measurement of clinical outcomes, and clinical endpoints made pooling impossible. Normally, the QUADAS-2 tool is used to assess studies evaluating diagnostic tests that compare the diagnostic accuracy of the index test vs. a reference standard test. Six of the included studies in this review did not describe a reference standard and the remaining studies showed high heterogeneity within described reference standards: ABPI; TBI; clinical classifications; wound healing; or amputation rate. This is an important limitation of the included studies; however, considering available quality assessment tools, there was no reasonable alternative to the QUADAS-2 tool. Before assessment of the microcirculation can be implemented as standard care, multiple issues have to be resolved. Measurement protocols need to be optimised and standardised; techniques should be validated in large clinical cohorts; reliability assessments should be performed; and cutoff values need to be determined with a high sensitivity and specificity. To do so, a well defined study population of patients with Fontaine III e IV PAD should be included and analysed on major clinical outcomes such as the aforementioned wound infection and healing, amputation rate, need for re-admission, quality of life, and mobility. 46 One of the main reasons to perform endovascular or open revascularisation procedures in patients suffering from PAD is to improve tissue perfusion and skin oxygenation of the lower leg and foot. This is of the utmost importance in patients with ischaemic ulcers, to facilitate healing. Ideally, the increase in perfusion and oxygenation of the diseased tissue can be monitored in real time during an intervention. Endovascular revascularisation is currently considered successful when an arterial stenosis or occlusion is overcome and no haemodynamically significant lesion is left behind. Outcomes are therefore focused on anatomical results and thus the macrovasculature. It may be argued that as long as the tissue oxygenation and perfusion parameters in the lower leg and/or foot do not increase the intervention may be considered as not successful. Real time monitoring of these parameters may guide the vascular surgeon or interventionalist during the procedure to extend it (if possible) and to revascularise more feeding arteries. So far, none of the described techniques seem capable of doing this. Future studies should focus on this, and also try to associate the peri-procedural findings of changes in tissue perfusion and skin oxygenation with clinical outcomes including improvement in walking distance, pain relief, and time to wound healing. To do so, the first step would be to define validated normal values for the tissue perfusion techniques. It would be of great help if the course of tissue perfusion and skin oxygenation levels could be monitored in the early post-intervention period. It may be argued that tissue perfusion takes some time after revascularisation to set a new equilibrium. So far, it is unknown when this new equilibrium is reached. Repeated measurements in the early post-intervention period, even at home, may be helpful in determining a decrease that may be associated with early treatment failures and the need to perform additional Doppler ultrasound, CT angiography, or magnetic resonance angiography, and eventually early reintervention. Finally, the cost of equipment was not studied. Equipment cost was difficult to determine because its place in the clinic has not yet crystallised and the reimbursement systems are country and sometimes even hospital dependent.

CONCLUSION
This systematic review provides an overview of 10 tissue perfusion techniques used before and within 24 hours after revascularisation procedures of the lower extremity to treat PAD. Within the broad inclusion criteria, only 26 articles were found to be eligible for inclusion in this review. Ideally, a tissue perfusion technique should guide the vascular surgeon or interventionalist in real time throughout the entire revascularisation procedure and be related to major clinical outcomes such as improvement in Fontaine classification and time to wound healing. The technique should be non-invasive, non-operator dependent, accurate, cost effective, and fast. At this time, evidence remains low regarding the diagnostic accuracy of these techniques. It is too early to recommend one of the currently available techniques as a decision tool in the treatment of patients with PAD. Prospective observational studies, to relate periinterventional assessments with clinical outcomes after a certain length of follow up, are necessary as a first step in the implementation of one of these techniques into daily vascular practice.

CONFLICTS OF INTEREST
None.

FUNDING
None.