Three-dimensional fluorescence lifetime tomography Anuradha Godavarty a 327 Votey Building, Department of Computer Science, University of Vermont, Burlington, Vermont 05405 Eva M. Sevick-Muraca 1011 Richardson Building, Photon Migration Laboratories, Texas A&M University, College Station, Texas 77843 Margaret J. Eppstein 327 Votey Building, Department of Computer Science, University of Vermont, Burlington, Vermont 05405 Received 20 August 2004; revised 19 December 2004; accepted for publication 23 December 2004; published 18 March 2005 Near-infrared fluorescence tomography using molecularly targeted lifetime-sensitive, fluorescent contrast agents have applications for early-stage cancer diagnostics. Yet, although the measurement of fluorescent lifetime imaging microscopy FLIM is extensively used in microscopy and spec- troscopy applications, demonstration of fluorescence lifetime tomography for medical imaging is limited to two-dimensional studies. Herein, the feasibility of three-dimensional fluorescence- lifetime tomography on clinically relevant phantom volumes is established, using i a gain- modulated intensified charge coupled device CCD and modulated laser diode imaging system, ii two fluorescent contrast agents, e.g., Indocyanine green and 3-3’-Diethylthiatricarbocyanine iodide differing in their fluorescence lifetime by 0.62 ns, and iii a two stage approximate extended Kalman filter reconstruction algorithm. Fluorescence measurements of phase and amplitude were acquired on the phantom surface under different target to background fluorescence absorption 70:1, 100:1 and fluorescence lifetime 1:1, 2.1:1 contrasts at target depths of 1.4–2 cm. The Bayesian tomography algorithm was employed to obtain three-dimensional images of lifetime and absorption owing to the fluorophores. © 2005 American Association of Physicists in Medicine. DOI: 10.1118/1.1861160 I. INTRODUCTION Optical-based molecular imaging and tomography using near-infrared NIR fluorescent contrast agents represents an emerging technology, which may advance cancer diagnostic imaging. In optical imaging, NIR light between 700–900 nm wavelengths propagates deeply into tissues and can differen- tiate diseased from normal tissues based on the differences in the endogenous absorption owing to oxy- and deoxy- hemoglobin, as demonstrated from breast imaging studies. 1–7 In order to increase the optical detection of small tissue le- sions not contrasted by neovascularization, molecular target- ing and reporting fluorescent contrast agents have been de- veloped. These agents can potentially improve the target:background T:B optical contrast ratio between nor- mal and diseased tissues with greater specificity and sensi- tivity over optical imaging performed using endogenous ab- sorption contrast alone. 8–10 In recent years, three-dimensional 3D fluorescence- enhanced optical tomography has been developed for time- dependent and time-independent measurements of light propagation for in vivo small animal studies 11,12 and large phantom studies. 13–19 All developments have been based upon T:B contrast due to fluorophore absorption cross sec- tion. Yet fluorescence offers the added capability of deter- mining lifetime or radiative decay kinetics as a method for assessing the local environmental conditions independent of fluorophore concentration. 20,21 Hence, by employing lifetime-sensitive fluorescing contrast agents, tomographic analysis of the fluorescence measurements may differentiate diseased tissues from normal tissues. Pioneering work to de- velop “reporting” fluorophore-protein conjugate contrast agents that demonstrate changes in fluorescence decay kinet- ics in the presence of specific proteases has been reported by Weissleder and colleagues 22–24 and is the subject of develop- ment in other laboratories. 25 Tomographic imaging of fluorescence lifetime in 2D 26–28 and 3D 29,30 has been demonstrated from time-dependent simulated measurements. While 3D tomographic reconstruc- tion of lifetime-sensitive fluorophore absorption cross sec- tion has been experimentally reported from time-independent measurements on small animals, 31 fluorescence lifetime to- mography has been limited to 2D experimental studies using time-dependent measurement schemes. 32 To date, 3D fluo- rescence lifetime tomography has not been demonstrated us- ing experimental measurements, but has been limited to simulated studies. 29,30 In this contribution, the feasibility of 3D fluorescence life- time tomography is demonstrated on clinically relevant tis- sue phantoms using fluorescent contrast agents differing in their fluorescence optical properties and time-dependent frequency-domain fluorescence measurements. Clinical translation of 3D fluorescence lifetime tomography will re- quire both: i the development of lifetime sensitive fluores- cent contrast agents, which “report” environmental condi- tions through a change in decay kinetics and ii the development and demonstration of a tomography algorithm that can reconstruct the 3D distribution of spatially-varying 992 992 Med. Phys. 32 „4…, April 2005 0094-2405/2005/32„4…/992/9/$22.50 © 2005 Am. Assoc. Phys. Med.