Optical method using fluence or radiance measurements to monitor ...

REVIEW OF SCIENTIFIC INSTRUMENTS VOLUME 74, NUMBER 1 JANUARY 2003Optical method using fluence or radiance measurements to monitorthermal therapyLee Chin a) and Mihaela PopOntario Cancer Institute, Princess Margaret Hospital, University Health Network, Toronto,Ontario M5G 2M9, Canada and Department of Medical Biophysics, University of Toronto,610 University Avenue, Toronto, Ontario M5G 2M9, CanadaWilliam WhelanDepartment of Medical Biophysics, University of Toronto, 610 University Avenue, Toronto,Ontario M5G 2M9, Canada and Department of Mathematics, Physics, and Computer Science,Ryerson Polytechnic University, Toronto, Ontario M5B 2K3, CanadaMichael Sherar and Alex VitkinOntario Cancer Institute, Princess Margaret Hospital, University Health Network, Toronto,Ontario M5G 2M9, Canada and Department of Medical Biophysics and Department of Radiation Oncology,University of Toronto, 610 University Avenue, Toronto, Ontario M5G 2M9, CanadaPresented on June 24 2002We present data from a two-region coagulated and native tissue Monte Carlo simulation inspherical geometry to assess the feasibility of using optical fluence and radiance information fordetermining the boundary of tissue coagulation during thermal therapy. Our results demonstrate thatradiance offers directional sensitivity that might be useful in monitoring asymmetric coagulationgrowth. Preliminary experimental data from fluence monitoring of radio-frequency kidney thermaltherapy further indicates that fluence information might be insufficient for determining thecoagulation boundary. Therefore, in addition to optical fluence, we are also exploring the potentialof radiance monitoring as a method for establishing the position of the coagulation boundary duringradio-frequency thermal therapy. © 2003 American Institute of Physics.DOI: 10.1063/1.1519681I. INTRODUCTIONInterstitial thermal therapy ITT is currently under investigationfor the treatment of solid tumors. Thin needleapplicators are inserted into the tumor site and energizedwith lasers, microwaves, radio frequency rf, or ultrasound.The energy is absorbed heating the tumor site resulting incoagulative necrosis. Due to the complex and dynamic natureof tissue properties of the tumor tissue during ITT, thedevelopment of real-time monitoring systems that can accuratelydetect tissue damage during the procedure and preservecritical normal structures is vital. Currently, temperaturesensors are commonly used to monitor and control ITTtreatments. However, due to slow thermal conduction,patient-specific variability in thermal damage, and the inabilityto directly sense tissue coagulation, temperature measurementsalone, might be insufficient for assessing treatmentefficacy. Optical point monitoring, where light intensity informationis measured using interstitial optical sensors, providesnear instantaneous response to structural changes thatoccur during treatment. 1,2 This is due to the fact that thescattering coefficient of tissue increases significantly between2 and 7 upon coagulation. The sharp contrast betweencoagulated and native tissue led Iizuka 3 to suggest thatITT might be amenable to optical monitoring.Continuous-wave optical information can be measuredRadiance measurements offer the advantage of directionalinformation but due to the lack of readily available radianceprobes, we and others have primarily employed fluenceprobes for ITT monitoring. However, recently Dickey et al. 4reported that by attaching a small right-angle prism to acleaved fiber, radiance information can be retrieved by performinga 180° rotation of the fiber. These authors employedradiance measurements for photodynamic therapy dosimetry.The potential of utilizing radiance for monitoring of ITT isyet to be explored. Optical monitoring is well suited to laserinterstitial thermal therapies LITT because the laser sourcemay act to simultaneously treat and provide optical energythat can be detected for optical sensing.We previously tested the potential of Monte Carlo MCmodels to relate fluence changes to coagulation radius duringLITT with a single spherical source. 5 Since the coagulationgrowth around the fiber is spherically symmetric, the determinationof coagulation volume can be modeled oneaElectronic mail: chinl@oci.utoronto.caeither in the form of direction-dependent intensity radiance,L(r,ˆ ) where r is the radius from the source and is thesolid angle relative to the radius normal, or directionindependentintensity fluence (r). The fluence Wm 2 can be obtained from the radiance Wm 2 sr 1 by integratingover all solid angles:r4Lr,ˆ d.0034-6748/2003/74(1)/393/3/$20.00 393© 2003 American Institute of PhysicsDownloaded 10 Feb 2003 to 142.150.192.30. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp

394 Rev. Sci. Instrum., Vol. 74, No. 1, January 2003 Chin et al.FIG. 1. Simulation geometry inset and Monte Carlo results showing predictedchanges in fluence, , forward radiance, L f (10°), and backwardradiance, L a (170°), at 2 mm away from the source during LITT employinga spherically emitting source fiber. The fluence behavior has been verifiedexperimentally see Ref. 5. Symbols are Monte Carlo results while the linesare fitted.dimensionally spherically. This geometry allows for the basicphysics of the optical monitoring problem to beinvestigated. However, for practical application the methodneeds to be extended to other geometries such as cylindricalsources source which are more commonly used in ITT.This article extends our previous MC simulation to includeradiance information. We also present preliminary experimentaldata demonstrating fluence monitoring of rf thermaltherapy of ex vivo porcine kidney. Based on our MCresults, we discuss why radiance might prove advantageousover fluence data for this complex geometry.L f . In contrast, the fluence and L a decrease less noticeablyindicating that at 2 mm, the majority of interstitial light intensityis traveling forward away from the source fiber. Asthe coagulation radius passes 2 mm, a light trapping effectcauses an increase in the fluence 150%, and a more significantchange in the L a value 450%. Here, L f remainsalmost constant as opposed to L a in the previous case. Here,the implication is that the increase in fluence is due to photonstraveling back towards the source fiber.The differences between the two radiances are a result ofthe highly forward scattering nature of light in tissue. Forsmall coagulation radii, the increased scattering properties ofcoagulated tissue will have minimal effect on the bulk radiancedistribution since photons scatter forward away fromthe source fiber and readily escape the coagulation volume.Still, there is sufficient scattering to cause collimated photonsto diverge from their original straight-line path. Hence, as thecoagulation front approaches 2 mm, it is primarily L f that isaffected. After passing 2 mm, the increase in fluence iscaused by photons that have scattered numerous times,enough to reverse their initial forward direction. Consequently,L a increases significantly while L f remains relativelyconstant following passing of the coagulation front.The MC results show that compared to fluence, L f and L aoffer increased sensitivity to an approaching and passing coagulationboundary, respectively.The fluence results of the MC model have been validatedexperimentally in well-characterized tissue mimicking albumenphantoms 6 and demonstrated excellent qualitative andquantitative agreement 20%, 5 experiments are underwayto validate the new radiance results.II. PRINCIPLES OF OPTICAL MONITORINGFor a spherical source of optical energy positionedwithin homogeneous tissue, our MC algorithm models thermaldamage as concentric spheres of coagulated and nativeoptical properties centered at the source inset of Fig. 1. Itisassumed that the index of refraction of coagulated tissue isthe same as native tissue. To mimic the ‘‘growth’’ of thedamaged region during heating, separate MC simulations arerun for increasing radii of the inner ‘‘coagulated’’ sphere.Next, changes in light intensity at a given detector positionare plotted versus the inner coagulation radius. Since theextent of damage increases as a function of heating time, thex axis damage radius is effectively a measure of heatingtime with the horizontal scaling of the graph expanding orcontracting depending on the rate of heating for a giventreatment. Additional details of the MC simulation can befound in Ref. 5.Figure 1 shows MC simulation results showing relativechanges at 2 mm away from a spherical source in fluence,radiance facing the source, L f L(10°), and radiance facingaway from the source, L a L(170°). Native ( a0.5 cm 1 , s 2.67 cm 1 ) and coagulated ( a0.7 cm 1 , s 13.1 cm 1 ) optical properties were chosento simulate albumen phantoms, which have similar propertiesas soft tissue. 6 Prior to the coagulation boundary reachingthe 2 mm point, there is a significant decrease 70% inIII. MONITORING RADIOFREQUENCY THERMALTHERAPYRf ablation has shown promise for the minimally invasivetreatment of small solid renal masses. 7,8 We are currentlyinvestigating a system produced by Radio TherapeuticsCorp. Sunnyvale, CA, which delivers rf energy at 460kHz via a LeVeen electrode Radio Therapeutics Corp. consistingof an array of evenly spaced wire electrodes enclosedby a metal cannula. When deployed, the array appears similarin shape to an umbrella Fig. 2 and produces a sphericalthermal lesion in tissue. A recent study in patients that underwentradical nephrectomy following rf ablation showedthat areas of viable tumor were present at the margins of thetreated volume. 8 The untreated areas were attributed to heatsink effects, abnormal tumor vasculature, and uneven rf energydistribution. 8 We envision that optical methods mightplay a role in rf thermal therapy monitoring to ensure completecoagulation of the tumor volume.We performed fluence monitoring of the coagulationfront in fresh ex vivo porcine kidney during rf thermaltherapy using a previously developed protocol that deliveredcontinuous rf energy until a coagulation-induced impedanceof 400 was reached, which caused a cessation of powerdelivery. 7 Figure 2 shows a schematic of our experimentalgeometry. The optical fibers and rf applicator were fixed inan acrylic jig for accurate positioning. A 2 mm diam diag-Downloaded 10 Feb 2003 to 142.150.192.30. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp

Rev. Sci. Instrum., Vol. 74, No. 1, January 2003Photoacoustic and photothermal phenomena395nostic source fiber Resonance Optics, Dennis, MA deliveringenergy from a diode laser Diomed, England at 805 nmand 0.5 W was positioned 5 mm from a 800 m diam fluencesensor Resonance Optics positioned at 2 mm from the tipof the rf electrode. The light source was positioned 7 mmfrom the electrode tip. The fluence sensor was attached to aPDA55 photodiode Thorlabs, Newton, NJ that convertedthe light signal to a photovoltage reading. The average of 30photovoltage signals was read each second and displayed inreal time on a PC. Following the treatment, the kidney wascut open in the plane containing the optical fibers.Figure 2 shows the changes in fluence measured duringrf thermal therapy of the kidney. The curve can be dividedinto three regions: I 0t50 s, fluence remains constant;II 50t180 s, fluence begins to rise slowly increasing by12%; and III 180t240 s, a sudden and sharp fluenceincrease with a relative change of 225% was observed. Itis likely that the constant light intensity of region I indicatesthat coagulation has yet to occur. The gradual fluenceincrease observed in region II suggests the onset of coagulationaround the source tines. The sudden increase in lightintensity in region III possibly signifies that the coagulationfront has passed or is close to the fluence sensor location.Visual examination of the thermal lesion after the treatmentshowed that the resulting coagulation boundary passed thefluence probe.Unlike the LITT case, the complex evolution of the coagulatedregion during rf ablation relative to the diagnosticoptical source makes prediction of the coagulation locationdifficult. During LITT, a significant increase in fluence occurswhen the coagulation radius passes the sensor. 1 However,for our chosen rf monitoring geometry, the coagulationradius grows towards the diagnostic source as opposed toaround it, which could potentially cause an increase in fluenceeven before the front passes the optical sensor. Hence,for optical monitoring of rf thermal therapy, fluence alonemight be insufficient to predict the extent of coagulation. Insuch asymmetric problems, measurements of directionallight intensity might prove more useful.FIG. 2. Experimental geometry inset and measured changes in opticalfluence during rf thermal therapy of ex vivo kidney.IV. APPLICATION OF RADIANCE TO MONITORING OFRF THERMAL THERAPYBased on the MC results of Sec. II, we hypothesize thatemploying a radiance probe rotating between 0° and 180°might provide the additional information necessary to determinethe location of the coagulation radius during rf thermaltherapy. We envision that prior to the coagulation boundarypassing the radiance sensor, L a would increase steadily asthe coagulation front approached the sensor, while L f wouldremain relatively constant. Using this approach, one wouldstrategically place the radiance probe at a desired location ofthe final treatment boundary, and wait for a change in theforward flux caused by coagulated tissue between the sensorand diagnostic source. We have recently constructed radianceprobes and are in the process of designing an automatedrotation system for the application of radiance monitoring ofrf kidney thermal therapy.ACKNOWLEDGMENTSFinancial support for this work was provided by the NationalCancer Institute of Canada with funds from the CanadianCancer Society. The authors would like to thank Dr.Brian Wilson for suggesting the possibility of radiance monitoringof ITT. Also, thanks to Victor Yang and DwayneDickey for assistance in building the radiance probes.1 L. C. L. Chin, W. M. Whelan, M. D. Sherar, and I. A. Vitkin, Phys. Med.Biol. 46, 2407 2001.2 L. O. Svaasand, T. Spott, J. B. Fishkin, T. Pham, B. J. Tromberg, and M.W. Berns, Phys. Med. Biol. 44, 801 1999.3 M. N. Iizuka, M.Sc. thesis, University of Toronto 1999.4 D. J. Dickey, R. B. Moore, D. C. Rayner, and J. Tulip, Phys. Med. Biol.46, 2359 2001.5 L. C. L. Chin, W. M. Whelan, and I. A. Vitkin, Phys. Med. Biol. submitted.6 M. N. Iizuka, M. D. Sherar, and I. A. Vitkin, Lasers Surg. Med. 25, 1591999.7 R. A. Rendon, M. Gertner, M. Sherar, M. R. Asch, K. R. Kachura, J.Sweet, and M. A. S. Jewett, J. Urol. 166, 2922001.8 R. A. Rendon, J. R. Kachura, J. R. Sweet, Jr., M. R. Gertner, M. D. Sherar,M. Robinette, J. Tsihlias, J. Trachtenberg, H. Sampson, and M. A. S.Jewett, J. Urol. in press.Downloaded 10 Feb 2003 to 142.150.192.30. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/rsio/rsicr.jsp