Infrared thermography for the assessment of lumbar sympathetic blocks in patients with complex regional pain syndrome

Resumen

Complex regional pain syndrome (CRPS) is a debilitating chronic pain condition that usually affects one limb, and it is characterized by its misunderstood underlying pathophysiology, resulting in both challenging diagnosis and treatment. To avoid the patients' impairment quality of life, the achievement of both an early diagnosis and treatment marks a turning point. Among the different treatment approaches, lumbar sympathetic blocks (LSBs) are addressed to alleviate the pain and reduce some sympathetic signs of the condition. This interventional procedure is performed by injecting local anaesthetic around the sympathetic ganglia and, until now, it has been performed under different imaging techniques, including the ultrasound or the fluoroscopy approaches. Since infrared thermography (IRT) has proven to be a powerful tool to evaluate skin temperatures and taking into account the vasodilatory effects of the local anaesthetics injected in the LSB, the use of IRT has been considered for the LSBs assessment. Therefore, the purpose of this thesis is to evaluate the capability of IRT as a complementary assessment technique for the LSBs procedures performance. To fulfil this aim, three studies have been conducted implementing the IRT in patients diagnosed with lower limbs CRPS undergoing LSBs. The first study focuses on the feasibility of IRT as a complementary assessment tool for LSBs performance, that is, for the confirmation of the proper needle position. When LSBs are performed, the correct needle placement is critical to carry out the procedure technically correct and, consequently, to achieve the desired clinical outcomes. To verify the needle placement position, imaging techniques have traditionally been used, however, LSBs under radioscopic guidance do not always ensure an exact performance. For this reason, the thermal alterations induced by the local anaesthetics, have been exploited and assessed by means of IRT. Thus, the LSB procedure was considered successfully performed when thermal changes within the affected plantar foot were observed in the infrared images after the lidocaine injection. The second study deals with the quantitative analysis of the thermal data collected in the clinical setting through the evaluation of different temperature-based parameters extracted from both feet. According to the results, the proper LSB success prediction could be achieved in the first four minutes after the block through the evaluation of the feet skin temperatures. Therefore, the implementation of IRT in the clinical setting might be of great help in assessing the LSBs performance by evaluating the plantar feet temperatures in real time. Finally, the third study addresses the quantitative analysis by implementing machine learning (ML) tools to assess their capability to automatically classify LSBs. In this study, a set of thermal features retrieved from the infrared images have been used to evaluate four ML algorithms for three different moments after the baseline time (lidocaine injection). The results indicate that all four models evaluated present good performance metrics to automatically classify LSBs into successful and failed. Therefore, combining infrared features with ML classification models shows to be effective for the LSBs procedures automatic classification. In conclusion, the use of IRT as a complementary technique in daily clinical practice for LSBs assessment has been evidenced entirely effective. Since IRT is an objective method and it is not very demanding to perform, it is of great help for pain physicians to identify failed procedures, and consequently, it allow them to reverse this situation.