Perioperative perfusion assessment to guide and improve decision-making in treatment of peripheral arterial diseases
Bryan Wermelink is a PhD student in the department Multi-Modality Medical Imaging. Promotors are prof.dr. R.H. Geelkerken & prof.dr.ir. W. Steenbergen from the faculty of Science & Technology.
Peripheral arterial disease (PAD) of the lower extremities, as severity advances, will increasingly comprise sufficient blood flow to the microcirculation. This microcirculation supplies individual cells with oxygen and nutrients. When increasingly affected, transitioning from asymptomatic PAD to symptomatic PAD, intermittent claudication (IC) or chronic limb threatening ischemia (CLTI) will be induced. With a high prevalence, increasing with age, PAD is a common and severe disease, with a high burden on the patient’s quality of life and healthcare and societal cost.
For all patients with PAD cardiovascular risk management (CVRM) plays a crucial role in halting further disease evolvement targeting risk factors that can be influenced to improve the health behaviour of the patient. CVRM comprises medication treatment for hypertension and hyperlipidemia, discontinuing tobacco use, treating and controlling diabetes mellitus (DM), anticoagulant therapy, increase of physical activity and weight management. When experiencing IC, the first treatment modality involves supervised exercise therapy (SET). When experiencing CLTI, to relieve rest pain or prevent tissue loss, immediate revascularisation is most appropriate. Current options for revascularisation are open, endovascular or hybrid (combined open and endovascular) interventions. During endovascular and hybrid procedures, X-ray fluoroscopy and angiography imaging, the latter using fluoroscopy combined with a contrast agent to visualise the inside of the vascular tree, are nowadays used to guide the interventionalist in peroperative decision making. Although being the golden standard for many years, both imaging techniques are susceptible to subjective interpretation and inability to assess quantified tissue perfusion. As a result, a peroperative assumed successful revascularisation could therefore be accompanied with bad clinical patient outcomes.
To improve treatment outcomes and provide quantified support for the interventionalist, the aim of this thesis is to gain knowledge on the peroperative implementation of novel imaging techniques to enable tissue perfusion quantification in patients with lower extremity PAD.
To determine which novel imaging techniques were suitable, a systematic review was conducted (chapter 2). Four electronic databases were searched for eligible articles describing a perfusion measurement technique, used in a peri-procedural setting before and within 24 hours after the revascularisation procedure, with the aim of determining the effect of intervention in patients with PAD. This systematic review resulted in an overview of 10 tissue perfusion techniques found in 26 eligible articles. It seemed too early to appoint one of them as a reference standard. An updated version of the review up to December 2023 did not change this conclusion. The scope of future research in this domain should therefore focus on clinical accuracy, reliability, and validation of the presented techniques. Therefore, prospective observational studies, to relate perioperative 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.
In chapter 3 a sidestep was made to the conservative treatment of patients with PAD with IC to determine whether the established guidelines for treatment, with SET, are being complied with and whether deviations from these guidelines could be justified in daily clinical practice. A retrospective single centre cohort study was performed including 420 patients with newly diagnosed IC. For all included patients, the compliance rate with the guidelines for SET was 80.5%. The rate of adequately motivated and defensible practice variation was 15.7%; the rate of unjustified practice variation was 3.8%. Therefore, meaningful care was performed in 96.2% of cases. The important message from this study was that personalised care has a valuable impact on patients, which will apply to the entire thesis. Implementation of perioperative tissue perfusion quantification could be the next big step towards personalised and more meaningful care.
In chapter 4 the implementation of laser speckle contrast imaging (LSCI) was performed in a cohort study including 100 patients, using the measurement protocol of a prior pilot study. In the pilot study, it appeared that peroperative perfusion measurements were greatly influenced by medication administered during general anesthesia (GA). This pilot study showed that analysis of LSCI data is highly reproducible, next to the already proven stability and reproducibility of the measurements itself. In the cohort study we aimed to determine the added value of perioperative LSCI and identify perioperative LSCI parameters that correlate with midterm clinical patient outcome parameters, including the influence of peroperative administered general anesthesia medication. Thereafter, we aimed to obtain a predictive multivariate regression model to assess the success of revascularisation. General anesthesia significantly increased perfusion. After revascularisation, perfusion in most predetermined regions of interest (ROIs) improved significantly as well. Perioperative tissue perfusion, especially in the toes, combined with Rutherford classification predicted good or bad clinical outcomes with an area under the curve (AUC) of the receiver operator characteristic (ROC) curve of 0.84 at the end of the revascularisation procedure and with an AUC of 0.9 when the difference between pre-and postoperative measurements were used. The addition of a postoperative measurement was clinically relevant and could give the practitioners an opportunity to personalise follow-up within 24 hours after the revascularisation. Unfortunately, no clinically relevant relation between patient-oriented clinical outcome parameters at six months and LSCI parameter improvement during revascularisation was found in the complete cohort, including several stages of PAD. The effect of GA medication overshadowed the more subtle influence of the revascularisation on LSCI parameters.
In chapter 5 a cohort study in patients with PAD was performed, aimed to determine the correlation between standardised acquired Two-dimensional Perfusion Angiography (2D-PA) parameters and midterm relevant clinical patient outcome parameters. From thereon, it was aimed to determine a predictive multivariate regression model to assess the success of revascularisation peroperatively. After revascularisation the arrival time (AT), time to peak (TTP), wash-in rate (WiR) and slope were significantly improved. The AT, TTP and WiR improved significantly more after revascularisation in patients with a good clinical outcome compared to patients with a bad clinical outcome. Preoperative patient characteristics, Rutherford classification and DM, combined with 2D-PA parameter AT in the forefoot in patients with PAD predict a good and bad clinical outcome with an AUC of 0.83. This value represents an excellent diagnostic accuracy. It was concluded that there is a strong relationship between relevant patient-oriented clinical outcome parameters at 6 months and 2D-PA parameters obtained during revascularisation.
In chapter 6 we aimed to develop a (semi)automated algorithm for comprehensive and convenient analysis of fluorescence angiography (FA) parameters. Automated detection of parameters T0 (arrival time of contrast) and Tmax (time it takes to reach maximum intensity) was successful in all subjects. Output of the algorithm had an excellent agreement with the median of the human observations with a intraclass correlation coefficient of 0.95 (95% confidence interval: 0.86 – 0.96). The presented method provided convenient data analysis in the search for effective FA quantification. Future research should expand the data to find adequate threshold values for peroperatively identifying insufficient perfusion and investigate the influence of physiological conditions.
In chapter 7 this analysis method, using T0 and Tmax to determine TTP, was used on a database of 32 FA recordings in patients without impaired bowel perfusion to provide a reference of ‘healthy’ normal values and 9 patients with suspected impaired bowel perfusion. Analysis of the group with unimpaired bowel perfusion provided a median (interquartile range (i.q.r.)) TTP of 4.8 (2.5) s, with 95 per cent of measurements under 7.0 s. The group with suspected impaired perfusion showed a median (i.q.r.) TTP of 9.6 (3.2) s, significantly different compared with the unimpaired group (P=0.002). Measurements after vascularisation was restored demonstrated a median (i.q.r.) TTP of 4.2 (1.6) s, which differed significantly from the median before intervention (P=0.033). This initial study suggests that this method for onsite FA quantification yields inflow-based parameters that enable discrimination between healthy and impaired perfusion and reflect corresponding symptoms.
In conclusion, this thesis shows for the first time a relation between relevant mid-term clinical outcomes and perioperative acquired perfusion parameters from LSCI, 2D-PA and FA. The cohort studies using both LSCI and 2D-PA showed an excellent diagnostic accuracy when a multivariate regression model was used including both patient characteristics and perfusion related parameters. Therefore an important step forward was made in the search to achieve quantified perioperative perfusion based support for patient centred clinical decision making. Next steps should be taken to confirm and improve our findings in cohort studies, with the focus on patients with Rutherford classification 5, which was found to be most impactful in the prediction of clinical outcomes. Finally, an iterative process should follow to automate image and data analysis to obtain a clinically useful, multi-modality real-time perfusion quantification tool.
