Application—Infrared Thermal Imaging Camera to Detect Blood Flow Double

Infrared Thermal Imaging Camera to Detect Blood Flow Double

   Popular clinical markers of local or peripheral perfusion such as center-to-toe temperature difference, skin mottling and capillary refill time are subjective indicators that lack the required sensitivity and specificity to identify patients with compromised peripheral perfusion. In the last few decades, non-invasive contact-based optical techniques have increasingly been adopted in clinical settings for quantitative and qualitative assessment of blood perfusion in peripheral tissue. Specifically, Near-infrared Spectroscopy (NIRS) has been used to measure tissue oxygenation and hemoglobin concentration, and a pulse oximeter (PulseOx) is routinely used to measure arterial oxygen saturation (SpO2). Laser Doppler Flowmetry (LDF) is another popular modality used to measure micro-circulatory blood flow in tissue, especially to assess wound healing and for skin disease research.

   However, all of these contact-based optical modalities can only measure blood perfusion at a specific location on the skin surface, i.e., the point-of-contact, and their measurements are sensitive to the exact placement of the probe. The high spatial variability in blood perfusion across tissue such as the skin limits the clinical utility of such single point contact-based modalities.

Figure 1. Incremental vascular occlusion experiment.

      In this paper, we develop a novel camera-based, multi-sensor, motion-robust blood perfusion imaging modality, that can reliably measure spatial maps and temporal trends of peripheral blood perfusion over the skin surface or internal tissue. It can be implemented both as a bedside patient monitoring system, e.g., in an ICU or the operating room, as well as a hand-held imaging tool to visualize blood perfusion at surgical sites, wounds, and ulcers. Further, in this work, we propose to use a novel brightness invariant optical flow algorithm to reliably track and compensate the motion of the skin surface during perfusion imaging. In this paper, we show, for the first time, that our novel approach of using a brightness invariant optical flow algorithm reduces the error in blood perfusion estimate below 10% in diverse motion scenarios when using existing optical flow algorithms that assume brightness constancy as is usually done.

   In the lab-based study, we conducted several controlled blood flow occlusion experiments on healthy participants of varying skin tones. We limited the evaluation of it and its comparisons with competing methods to only the palm region because of the ease of access and experiment simplicity. However, it methodology can easily be extended to obtain blood perfusion maps from other body parts such as the face, foot, and the abdomen. Through these experiments, we demonstrate that it can easily detect arterial and venous blood flow occlusion and identify different levels of partial blood flow occlusion. 

   Additionally, it can also potentially be used to obtain blood perfusion maps from internal tissues using existing camera attached to an endoscope or a laparoscope with minimal modification. Hence, it could be employed in minimally invasive surgeries to visualize blood perfusion, e.g., for identifying anastomotic failures after surgical interventions such as bowel resection and intestinal surgeries, and also for localizing cancerous tissue that usually shows distinctly different blood perfusion signatures.

Reference

Mayank Kumar, James W. Suliburk, Ashok Veeraraghavan, et al. It: a camera-based, motion-robust and highly sensitive blood perfusion imaging modality. Scientific Reports. 10(4825):1-17, 2020.