QIRT 10 10 th International Conference on Quantitative InfraRed Thermography July 27-30, 2010, Québec (Canada) Time-Frequency Analysis of Skin Temperature in a Patient with a Surface Tumor Monitored with Infrared Imaging By C.G. Scully 1,2 , W. Liu 2 , J. Meyer 2 , A. Dementyev 2 , K.H. Chon 1 , P. Innominato 3,4,5 , F. Lévi 3,4,5 , and A.M. Gorbach 2 1 Worcester Polytechnic Institute, Worcester, MA, USA 2 National Institutes of Health, Bethesda, MD, USA, gorbacha@mail.nih.gov 3 INSERM, U776 «Biological rhythms and cancers», Paul Brousse Hospital, Villejuif, France, francis.levi@inserm.fr 4 Univ Paris Sud 11, SO776, Orsay, France 5 AP-HP, Department of Medical Oncology, Chronotherapy Unit, Paul Brousse Hospital, Villejuif, France Abstract Infrared thermography was used to assess skin temperature in a patient with a surface tumor across a range of frequencies relating to blood flow fluctuations. Using a high acquisition rate of image capture and wavelet analysis we were able to study temperature dynamics from low frequencies related to vasomotion (~0.01 Hz) up to the cardiac rate (~1 Hz). We show that IR imaging can be used to monitor skin temperature in spatial, time, and time-frequency domains allowing the study of coupling between different microvascular territories. 1. Introduction Blood flow in the microcirculation can often be assessed by use of optical imaging techniques such as Laser Doppler and Laser Speckle. These techniques allow for frequency analysis of blood flow fluctuations across a spectrum of frequencies that are related to NO-independent endothelial activity (Range V: 0.005 – 0.0095 Hz), NO-dependent endothelial activity (Range VI: 0.0095 – 0.021 Hz), neurogenic (sympathetic) activity (Range IV: 0.021 – 0.052 Hz), myogenic activity (Range III: 0.052 – 0.145 Hz), respiration (Range II: 0.145 – 0.6 Hz), and heart rate (Range 1: 0.6 – 2 Hz) [1]. Use of frequency analysis techniques allows the study of the contribution of each physiological component to the overall local blood flow regulation. Previously, skin temperature has also been used to study these components, but the low-pass filter characteristics of the skin attenuate higher frequency temperature changes. Bandrivskyy et al. measured skin temperature with thermistors along with simultaneous laser Doppler measurements and used Wavelet Phase Coherence (WPC) to show the effects of blood flow on skin temperature up to the 0.1 Hz range, at which point skin temperature oscillations contributed to blood flow are masked by noise [2]. Infrared (IR) imaging reflects skin temperature through radiation following conduction and convection of heat to the surface of the skin, but does not rely on inertia such as a thermistor. Therefore, with IR measurement it is possible to assess the temperature of living tissue at more depth than the surface. Because of this there is an expectation that IR imaging would allow the analysis of lower and higher frequencies through the skin. Gorbach et al. used IR imaging to study blood flow oscillations in the kidney using wavelet analysis [4], and during intraoperative monitoring of human brain with an IR camera heart rate in individual vessel segments on the cortex was monitored [5]. Garbey et al. measured the cardiac pulse using IR through the skin, with the consequence of filtering out low frequency oscillations relating to blood flow [3]. Tumor microvasculature is known to be atypical in structure as well as function. The exact nature of vascular abnormalities that may be present in a specific tumor at a specific stage is not well known, but may include increased angiogenesis, non-specific flow directions, and stagnant flow [6]. These differences in structure and function can diminish blood flow regulation in tumors and lead to decreased effectiveness of medication [6]. The oscillatory nature of the microvasculature in tumors compared to normal tissue has not been well characterized. The objective of this work is to assess the ability of an IR camera to monitor skin temperature fluctuations as related to blood flow and compare differences in temperature patterns between skin regions in a patient presenting with a surface tumor. Time-frequency analysis between regions of interest within and outside of the tumor boundaries was performed. A thermally conductive region, described in [7], around the bridge of the nose has previously been associated with brain temperature stemming from the hypothalamic thermoregulatory center. A secondary objective of this work is to assess stability and differences in brain tunneling temperature (BTT) between the right and left side of the nose monitored with IR imaging. 2. Methods 2.1. Image Acquisition IR images were acquired with an FLIR SC7700 camera (3-5 micron wavelength, 640x512 pixels/image, 14 bits) with a temperature resolution of 0.017°C at an acquisition rate of 10 Hz. CNUC TM is an FLIR calibration process that adjusts camera properties during imaging to correct for room temperature. This function was enabled during the patient imaging procedure. The patient gave written informed consent for this non-invasive procedure. The patient was seated in a chair and asked to look on a color target above the camera lenses and stay stationary for 25 minutes during imaging. A total of 15,000 images were collected. Pixel size was determined to be ~0.62 mm/pixel by measuring known distances in the image. http://dx.doi.org/10.21611/qirt.2010.165