hyperthermia treatment planning - aapm: the … · hyperthermia treatment planning published for...
TRANSCRIPT
HYPERTHERMIA TREATMENTPLANNING
Published for theAmerican Association of Physicists in Medicine
by the American Institute of Physics
AAPM REPORT NO. 27
HYPERTHERMIA TREATMENTPLANNING†
REPORT OFTASK GROUP NO. 2
HYPERTHERMIA COMMITTEE*
AAPM
Members
P. B. Dunscombe (Task Group Chairman)
Gilbert H. Nussbaum
John W. StrohbehnFrank M. Waterman
*T. V. Samulski, Hyperthermia Committee ChairmanB. R. Paliwal, Past Committee Chairman
August 1989
Published for theAmerican Association of Physicists in Medicine
by the American Institute of Physics
†AAPM Report No. 26 covers performance evaluation of hyperther-mia equipment.
DISCLAIMER: This publication is based on sources andinformation believed to be reliable, but the AAPM and theeditors disclaim any warranty or liability based on or relat-ing to the contents of this publication.
The AAPM does not endorse any products, manufac-turers, or suppliers. Nothing in this publication should beinterpreted as implying such endorsement.
Further copies of this report may be obtained from
American Association of Physicists in Medicine335 E. 45 Street
New York, NY 10017
Library of Congress Catalog Card Number: 89-08 1240International Standard Book Number: 0-88318-643-8International Standard Serial Number: 0271-7344
Copyright © 1989 by the American Associationof Physicists in Medicine
All rights reserved. No part of this publication may be re-produced, stored in a retrieval system, or transmitted inany form or by any means (electronic, mechanical, photo-copying, recording, or otherwise) without the prior writ-ten permission of the publisher.
Published by the American Institute of Physics, Inc.335 East 45 Street, New York, NY 10017
Printed in the United States of America
Acknowledgement
T h e s k i l l f u l t y p i n g o f t h i s m a n u s c r i p t b yMiss Thelma Davidson is acknowledged with thanks.
Contents
1. Introduction
2. Specific Absorption Rate and its Distribution
2.1 The source of energy
2.2 The relative SAR distribution
2.3 Absolute Specific Absorption Rate
2.4 Coupling and Matching Devices
3. Temperature and its Distribution
3.1 Thermometer description
3.2 Intrinsic characteristics of the thermometer
3.3 Sources of error in temperature measurement
3.4 Thermometer positioning
3.5 Control circuits, display and recording
3.6 Surface Cooling
4. Treatment Planning
4.1 Identification and description of target
volume and adjacent tissues.
4.2 General considerations of heating modalities.
(Comparative thermal dosimetry).
4.3 Treatment Planning (Prospective thermal
dosimetry).
4.4 Specification of in-vivo thermometry technique
(Concurrent thermal dosimetry).
4.5 Reconstruction of the treatment in time and
space (Retrospective thermal dosimetry).
5. References
4
1. Introduction
Although hyperthermia was first employed as atreatment for malignant disease in the last century, it isonly relatively recently that its mode of action andclinical application have been subjected to seriousscientific scrutiny. From the early 1970’s a wealth ofbiological data has been accumulated from manyinstitutions with the intention of elucidating themechanisms of hyperthermic cell killing and of identifyingoptimum combinations of this rediscovered modality withthe conventional approaches primarily of radiotherapy andchemotherapy. Since the latter part of the last decade,as the potential clinical benefit of hyperthermia becamemore apparent, considerable effort has been devoted to thedevelopment of techniques which could be used for theproduction and monitoring of elevated temperatures withinthe cancer patient.
At the present time, many clinical institutions haveembarked upon hyperthermia programs largely on the basisof the encouraging results which have accrued so far.It is slowly becoming clearer which anatomical sites canbe effectively heated, the physiological and biochemicalconditions which make tumours amenable to this form oftherapy, and the combinations with radiotherapy andchemotherapy which are likely to meet with the greatestsuccess.
In spite of the high current level of interest inhyperthermia, it remains the case that our knowledge ofits modes of action, either alone or in combination, andour experience in its application are rudimentary.Although there is strong evidence of a therapeutic benefitof hyperthermia under some circumstances at least, acursory study of the literature indicates minimaluniformity in its clinical application.
When, as is typical, hyperthermia is applied incombination with radiotherapy or chemotherapy, variablessuch as dose (from radiation or drugs), sequencing, timedelays and fractionation are all likely to significantlyinfluence outcome. These factors have been dealt withelsewhere2,3 and are beyond the scope of this report.
However, even in the approach to the physical problems ofthe therapeutic elevation of temperature and itsmeasurement, little consistency is to be found in theliterature. In view of the variety of heating modalitiescurrently in use and of the relative infancy ofhyperthermia as a clinical tool this lack of consistencyis perhaps not too surprising. In spite of thisshortcoming the development of hyperthermia as a clinicalmodality has, to date, not been seriously impeded.
The situation in the future, however, is likely to bequite different. Smaller centres with fewer resources ofboth equipment and skilled personnel are entering thefield in large numbers. Unless they are to adopt theinefficient approach of accumulating their own experience,published clinical results need to include sufficientinformation to permit duplication of general techniqueseven with equipment from different suppliers.
If hyperthermia is to be applied with maximumefficiency, criteria for the selection of the optimumheating and thermometric techniques need to be available.If the influence of variables such as radiation dose andfractionation is to be understood and optimized to theadvantage of the patient, it is essential that theparameters describing the hyperthermia aspect of atreatment strategy be defined and controlled as tightly aspossible.
This report is concerned with those parameters whichdescribe the heating modality and its application to thepatient. By analogy with conventional radiotherapy wewish to accumulate the relevant information to permit thedecision making process known as treatment planning totake place. Due to the comlexity of the interactionbetween the non-ionizing radiation beams frequently usedfor heating and the inhomogeneous human body, and due tothe unpredictable physiological response to temperatureelevation, many significant decisions have to be madeduring the treatment. This contrasts with conventionalradiotherapy in which the major technical decisions, inalmost all cases, are made in the absence of the patient.
6
Conventional radiotherapy has now progressed to thestage where there is little room for ambiguity or error ina treatment plan. Definitions of the relevant treatmentparameters are universally understood and accepted;standard treatment techniques have been established formany sites ; calculational techniques are sufficientlyaccurate under most, even complex conditions” ; concensushas been reached on which dosimetric parameters need to bemeasured and to what accuracy.
It should be the aim of the hyperthermia community toattain a similar degree of rigour and precision in theselection and specification of a hyperthermia treatment.Unfortunately, a major and fundamental difficulty isencountered at the outset of this endeavour. Whereas inradiotherapy the basic parameter used to specify thetreatment - the dose - has a clear physical definition,this is not the case in hyperthermia. Of course, dose inradiotherapy cannot alone fully describe a course oftreatment, neither can it in general be used to predictoutcome. However, the identification and definition ofthis basic parameter has probably contributed more to thesuccess of radiotherapy than any other development in itshistory with the possible exception of supervoltagemachines. The specification of a comparable unit for usein hyperthermia has been the subject of study8,9. At thepresent, however, there is no consensus as to a unit ofthermal dose of general applicability. This being thecase it is necessary to proceed on the basis of theexisting information which indicates clearly thatbiological damage is a function of both time of exposureand temperature. As the functional relationship betweenthese two parameters and dose remains unclear, treatmentdescription at this stage requires specification of thetime course of the tissue temperature distribution.
Two further areas of distinction between hyperthermiaand radiotherapy warrant brief discussion. Firstly, withthe exception of brachytherapy techniques, it is the aimof conventional radiotherapy to achieve a uniform dosethroughout the target volume. In the majority of casesthe use of multi-portal supervoltage x-ray beams resultsin this objective being approximately met. In contrast,present hyperthermia technology, excluding whole bodyhyperthermia, yields very non uniform patterns of energydeposition. Furthermore, even if energy deposition were
7
uniform, physiological effects (blood flow) anddifferences in thermal properties such as thermalconductivity, among different tissues would result ingenerally very non uniform temperature distributions.In designing and implementing clinical thermometry it isimportant to appreciate the high degree of temperaturenon-uniformity which is present.
Secondly, prospective treatment planning which playsan essential role in conventional radiotherapy is oflimited use in hyperthermia. For the majority ofconventional radiation treatments it is consideredunnecessary to confirm by in vivo measurement duringtreatment that the correct dose has been delivered.
Exceptions occur, of course, when vital radiosensitiveorgans are in close proximity to the radiation field.However, in general, specification of the radiation beamsand calculational techniques are sufficiently advanced topermit the distribution of energy deposition, under givenanatomical conditions, to be computed with high accuracy.
In contrast, the effect of hyperthermia depends notsimply on energy absorbed per unit mass but in some, asyet undefined, way on the resultant temperature rise andits time course. The bioheat equation describes thedependence of temperature both on the energy absorptionrate and on the rate of dissipation of energy. The formerquantity is, in any event, more difficult to determinethan in the case of ionizing radiation due to, forexample, the complexity of the source field andreflections which can occur during radiofrequency,microwave and ultrasound heating. The major difficulty,however, is encountered in quantitating the heat lossmechanisms at play in the human body. Under conditions ofsteep temperature gradients, conduction is often thedominant mode of heat loss. The other contribution is,of course, blood flow and its significance forhyperthermia has been the subject of much study1 2. Thepresence of this latter unpredictable and time dependentheat transport mechanism has, as its consequence, one ofthe major practical differences between treatment inhyperthermia and that in conventional radiotherapy. Invivo measurement, whether by invasive or non-invasivemeans, will remain essential for each and everyhyperthermia treatment.
8
The subject of this report, treatment planning forhypertyhermia, may be resolved into three components whichare introduced below.
Specific Absorption Rate and its Distribution is thedetermination of the rate of energy absorption per unitmass, wkg-1, in three dimensions under specifiedconditions in standard phantoms or in a patient from agiven treatment machine.
The data accumulated during this procedurecharacterize the heating technique under the specificconditions employed and permit general identification ofthose sites and target volumes which are likely to proveamenable to treatment with a particular technique. Inaddition, the distribution of the Specific Absorption Rate[SAR] forms one group of input parameters which isessential for the calculation of the temperaturedistribution in vivo.
Temperature and its Distribution is the determinationof the distribution of temperature in three dimensions ina patient throughout the course of a clinical treatment.
The temperature data accumulated during the clinicaltreatment form the most valuable record of that treatmentand are the ultimate source of data characterizing theheating session. In practice, and with currentlyavailable invasive thermometry, complete temperaturedistributions in vivo cannot be determined. The state ofthe art is presently limited to recommending minimumprocedures which could probably indicate when anappropriate hyperthermic treatment had been delivered”.It remains a topic of considerable research interest todevise methods of deducing complete temperaturedistributions from the limited measurements possible inthe clinic.
Treatment Planning is the selection of the optimaltreatment technique and the geometrical configuration ofthat technique based on a knowledge of the performance ofavailable treatment machines and on the computation of theexpected three dimensional temperature distribution in apatient.
9
This definition describes closely the decision makingprocess known as treatment planning in radiotherapy. Thedifference in practice between the two is that, in thecase of hyperthermia, the computational aspects are lessadvanced and are subject to far greater uncertainty.
In the sections that follow each of the three areasdefined above is discussed in some detail. After a briefreview of the relevant recent literature on these topicssuggestions are offered as to how the current state ofknowledge may best be applied in the clinical environmenttaking into account the wide variation in the equipmentavailable and its degree of sophistication. From theensuing discussion, areas requiring further study can beidentified.
2. Specific Absorption Rate and its Distribution
The aim of performing Specific Absorption Rate [SAR]measurements is to characterize the energy deliveryequipment by determining the pattern of energy deposition.Although a variety of heating techniques is now availablethe only ones for which SAR has little or no meaning arewhole body heating via conduction through the skin and/orthe inhalation of hot gases14 and regional methodsemploying extracorporeal heating of blood1 5.
For some systems calculation may be a very appropriateapproach to applicator design and characterization16,17 .At the present time it is necessary in all cases, however,to confirm by judiciously chosen point SAR measurementsunder simple geometrical conditions in a phantom that theexpected distribution is being produced. A furtheressential reason for performing some physical measurementsof SAR is safety. The accuracy required here must providesensitivity sufficient to identify the presence ofunexpected hot spots which occur under some conditions’*.
2.l The Source of Energy
If the SAR measurements to be made are to beassessed as being reasonable for the techniqueemployed it is necessary to obtain a basicunderstanding of the operation of the applicator.
10
For external radiative electromagnetic andultrasound techniques an important determinant of thedepth of penetration in a specified semi-infinitemedium (defined as that depth at which the intensityof a plane wave has dropped to e -2 of its initialvalue) of the unmodified therapeutic beam is thefrequency. The frequency of operation of a system iseasily established from the supplier’s literature.
The presence of harmonic distortion is usually nota problem with radiative techniques although harmonicswill decrease the penetration of both electromagneticand ultrasound beams. The main restriction on thepresence of harmonics will, in most cases, begovernment regulations covering the emission ofelectromagnetic energy. Depth of penetration whichstrictly applies only to plane waves, has no meaningfor the other methods of induction of hyperthermia andhas little revelance to interstitial microwavetechniques in which the radial intensity decrease andcoherence effects largely determine the SAR pattern.However being the fundamental operating parameter of asystem the frequency, and its range where applicable,should be known.
An additional descriptor of radiativeelectromagnetic methods is the mode of antennaexcitation. The mode of excitation contains muchinformation on the three dimensional fielddistribution” and can be used to suggest the mostinformative series of phantom measurements.
The physical size of an applicator, in addition toinfluencing the penetration of the heating beam19,20
places constraints on the largest area perpendicularto the axis of propagation which can be heated.External microwave applicators will effectively heatareas which are significantly smaller than the area ofthe radiating antenna. An ultrasound beam will notspread to cover an area larger than that of thetransducer aperture before the intensity drops belowthe therapeutic level. In contrast capacitive andinductive approaches may heat areas which extendbeyond the boundaries of the applicators. However,applicator size remains a constraint on the largestarea which can be heated.
11
Experience with regional techniques in which a thicktransaxial section of the patient is exposed tosignificant electromagnetic fields [e.g. B.S.D. AnnularPhased Array System23 and the Magnetrode22 ] is limited.Calculation supported by comprehensive in vitro and invivo measurement is essential23 if the true extent of theheating field is to be discovered.
Whilst a knowledge of the physical dimensions of anapplicator and its mode of excitation can be useful in thedesign of phantoms for studying SAR distributions it isimportant to base treatment planning decisions on moredetailed experimental studies possibly supplemented bycalculation. Generally tissue to be irradiated is in thenear field of a microwave applicator and thus theelectromagnetic field, which may have non-radiativecomponents, does not behave like a plane wave. under suchcircumstances the depth of penetration corresponding tothe operating frequency can be used as a rough guide only.Ultrasound transducers are frequently focussed to improvethe effective penetration of the beam, but even so thetarget tissue is usually in the near field2 4.
However, in spite of these complications familiaritywith the heating technique is the first step in thetreatment planning process. Construction of appropriatephantoms, selection of measurement sites and preliminarydecisions on the anatomical regions which may besuccessfully heated can all be based on knowledge andunderstanding of the heating technique.
2.2 The relative SAR distribution
The relative SAR distribution is necessary for theinput to the calculation of the temperature fieldsproduced during a clinical treatment and thus is anessential element of the treatment planning process.Under some conditions calculation may be the-source of themost comprehensive information on SAP distribution16,25
but in ail cases experimental determination of the overallpattern should be obtained. If the density and specificheat of heated tissue are known it is possible to deducelimited SAR values from in vivo measurements. However, ingeneral, static phantoms are preferred as their propertiesare known, geometry can be standardized and a greaternumber of measurement sites is possible.
12
In recent years phantom materials have beendeveloped which correctlv mimic the electrical andacoustic properties of tissues such as muscle and fat overa wide range o f f requencies2 6 , 2 7 , 2 8 . As the intended useof these materials is to study the distribution of energyabsorption, their thermal properties are not necessarilythe same as those of tissue and can be considerablydifferent from living tissue in which blood flow is oftenthe most significant heat loss mechanism. However, theirthermal properties are in general known to sufficientaccuracy to permit extraction of SAR values frommeasurements of rates of rise of temperature.
To avoid convective effects electromagnetic heatingmodalities are usually studied in solid or semi-solidphantoms. More complete distributions, particularly fromultrasound applicators, may be obtained from liquidphantoms in which a transducer is scanned and the relativeintensity pattern obtained”.
The information obtained from phantom studies iseasier to evaluate or to compare with that expected fromthe applicator employed when the measurements are made ina homogenous material with a flat surface. As theintention is to obtain SAR distributions and not toattempt to simulate the final temperature distribution invivo the correct thermal properties of the phantommaterial are not essential. Blood flow can to some extentbe simulated through the design of dynamic phantoms. Therole of such devices, however, remains unclear at themoment: certainly, they are not relevant to SAR studies.
The presence of electrical or acousticinhomogeneities can perturb the SAR distribution to asignificant degree and must therefore receiveconsideration. Clearly it is not possible to simulate allanatomical configurations that might be encountered.However one or two representative cases can providevaluable additional information on the characteristics ofa particular heating technique1 8. The majority ofinvestigations aimed at characterizing applicators willrely on measurements of rates of rise of temperature insolid or semi-solid phantoms at the commencement ofheating before conduction becomes significant. Thebehaviour of the thermometry system used to make thesemeasurements requires some discussion.
13
There are two approaches in widespread use.Thermography has proved extremely valuable in assessingquickly the temperature distribution through a plane of aphantom. With a temperature resolution of approximately0.2°C and a spatial resolution of 3mm thermography withsplit phantoms is capable, with proper care, of quicklyproducing two dimensional temperature distributions of anaccuracy that is adequate for many purposes. Whenthermography is employed it is essential that the phantombe heated rapidly so that the presence of conduction doesnot significantly influence the temperature distribution.In addition, thermographic examination of the splitphantom must promptly follow heating if distortion of thetemperature pattern, due to heat loss from the phantomsurface is to be avoided. As has been pointed outpreviously 3 1 it is likely that the heating system willhave to be operated at or near maximum output to obtainthe greatest accuracy. A major advantage of thermographyis the ease with which unexpected hot or cold spots may beidentified in phantom studies. These may be due to theproperties of the applicator-phantom combination or thepresence of perturbing thermometers.
When thermographic equipment is neither appropriatenor available resort must be made to point measurementsusing invasive thermometers. It is common practice whenemploying this approach to incorporate into the phantom anarray of tubes through which thermometers can beinserted3 3. Clearly the number of sites at whichmeasurements can be made is restricted. However,interpolation between measurement points can yield theappropriate degree of accuracy in those situations inwhich the SAR is changing only slowly with position.
Point SAR measurements using invasive thermometryare clearly more prone to error when fields of high SARgradient are involved. Examples of these would be the SARdistributions from interstitial antennae and thoseproduced by ultrasonic transducers. In the former casethe most accurate method may be calculation with the grossfeatures of the distribution confirmed by thermographicexamination 33. In the latter, intensity plotting in aliquid phantom using an ultrasonic transducer yieldsinformation of an adequate spatial accuracy” althoughmeasurements in a tissue equivalent U.S. phantom may bepreferred.
14
In those cases in which point temperaturemeasurements are made in solid or semi-solid phantoms caremust be taken to ensure that measurement artifacts do notcompromise the accuracy of the determination”. Thermalinsulation of the sensor can lead to error due to the timedelay before the rate of rise of sensor temperature equalsthat in the phantom3 7. In the presence of hightemperature gradients, e.g., with interstitial techniques,conduction along the probe can lead to significante r r o r3 8’ 3 9. In the case of ultrasound irradiation thepresence of plastics surrounding the thermometer has beenshown to yield artefactual readings4 0
In view of the above discussion the followingguidelines can be offered for the characterization ofheating modalities through relative SAR distributions.Where calculations are considered to be a reliable guide,at least in homogenous tissue, these should be available.Where calculations are not considered helpful experimentalmeasurements should be undertaken and these in any caseshould be used to confirm the gross features of acalculated distribution. Phantoms for such studies ofapplicators for local heating may be constructedaccording to established recipes and procedures2 7 and, toavoid edge effects, should be at least three depths ofpenetration deep and possess a surface area three timesthat of the applicator. Impedance matching procedures,where applicable, should be employed although activesurface cooling should not. The accuracy and sensitivityof the temperature monitoring procedure should besufficient to determine SAR values to within 10% and witha spatial accuracy of 3mm at least in the regions in whichthe SAR exceeds 50% of its maximum value. In general, withelectromagnetic heating it will be reasonably straightforward to attain the required degree of accuracy atsufficient depth in homogenous material. Close to thesurface of the phantom the discontinuity in both thermaland electrical properties could lead to substantial errorand, for this reason, it is suggested that normalizationbe made to the point at 1 cm depth on the central axis ofthe beam. Errors at points proximal to this may wellexceed the 10% value given above.
15
In the strong gradients associated with interstitialtechniques it is unlikely that point measurement can yieldthe required degree of accuracy. However, calculationsfor these situations have received considerable attentionand these may form the main source of information onrelative SAR distributions”‘.
Intensity measurements on ultrasonic applicatorsshould be capable of reaching the degree of accuracyspecified above with appropriate techniques in anappropriate liquid phantom.
Regional techniques in which a transaxial section ofa patient is heated require a somewhat different approach.To maintain simplicity the phantom for these studiesshould be of circular-or elliptical section and; of alength sufficient that at its ends the SAR has dropped to10% relative to its maximum value. In selecting a phantomfor this application the significance of possible resonanteffects should be recognized.
Electromagnetic and ultrasonic coupling materialsshould be used if they are to be employed clinically. Theyshould be applied using a standard, and reproducible andspecified method.
Consideration should be given to extracting therequired amount of information in the minimum time andhere knowledge of the electromagnetic or acoustic fieldpatterns is valuable. Many techniques such as capacitiveand inductive radiofrequency heating and stationaryultrasonic insonation produce fields in homogeneous andsymmetrical media that have been shown to have cylindricalsymmetry. In such cases an SAR distribution in one planecontaining the axis of symmetry of the applicator may besufficient. More complex electromagnetic field patternswill, of course, require more comprehensive investigation.
A straightforward and useful approach to summarizingSAR data so obtained has been suggested. The usefulthermal field size has been defined as the area enclosedby the iso SAR curve corresponding to 50% of the maximumSAR at the appropriate depth. For external radiative,capacitive and pancake single coil inductive devicesspecification of the useful thermal field size could bemade at a depth of 1cm. Therapeutic ultrasound fields
16
often exhibit significant intensity fluctuations bothaxially and laterally and hence their characteristicsrequire more complete measurement than is generally neededfor electromagnetic fields. An appropriate approach mightbe to specify, at a depth of 1 cm, the size of the fieldas defined by the locations at which the lateral edges ofthe outer peaks have fallen to 50% of their maximumvalues.
Whilst the concept of useful thermal field size isconvenient for applicator characterization and qualitycontrol it is understood that it is not necessarily or ingeneral coincident with the therapeutically relevant fieldsize.
The value of such a simple descriptor for techniquessuch as the Annular Phased Array System or the Magnetrodeis more questionable. In these cases it is necessary,under some standard conditions to identify the location ofthe maximum SAR and define, in the plane containing thecoil or antennae, the area enclosed by the 50% SAR Ifthe position of maximum is on the surface it isrecommended that normalization takes place with respect toa point 1cm deep to avoid the difficulties of measurementat the surface.
For external applicators the thermotherapeuticp e n e t r a t i o n d e p t h43 is a further quantity which aids thespecification of the heating system. This is defined asthe depth at which the central axis SAR falls to one thirdof its value at the surface. This figure gives anindication of the depth at which therapeutic heating ispossible when irradiating plane, homogenous,muscle-like tissue with some surface temperature control.
Finally, representative measurements should be madewith inhomogeneities present.fat27,44
Plane slabs simulatingof thicknesses 5 and 10mm will suffice in most
cases to identify possible hot and cold spots.
The information collected by the above means is theminimum necessary as the input to the treatment planningprocess. In addition, the data so collected will bevaluable for the comparative assessment of techniqueswithin and between centers.
17
2.3 Absolute Specific Absorption Rate
Absolute Specific Absorption Rates measured underexperimental conditions are far less important inhyperthermia than the analogous dose rate is inconventional radiation therapy. Temperature, one of thequantities which is known to strongly influence biologicaleffect, is dependent not only on the rate of energyabsorption and specific heat but also on the largelyunpredictable local energy transfer mechanisms.
However, the relationship between absolute SpecificAbsorption Rate under some standard conditions andmanufacturer supplied power indicators is an essentialcharacteristic of the equipment and should be determined.Assisted by calculation and the relative SAR distributionthis quantity will help to indicate the maximum blood flowfor which therapeutic temperature rises can be achieved.It can also be used to indicate the amount of surfacecooling which must be employed to limit the temperaturerise at the skin.
The most important potential use of absolute SAR isfor input into calculations of temperature distributions.Provided applicator - patient geometry is consistent fromtreatment to treatment, i.e., electrical or acousticimpedance matching is reproducible, knowledge of theabsolute SAR at a given point removes one variable in thecalculation of the temperature distribution.
Absolute Specific Absorption Rate values should bemeasured at a point 10mm deep [to avoid surface effects]in a standard specified phantom material and at the radiallocation of highest SAR. This will usually be on thecentral axis for a microwave or ultrasound device.Dividing the SAR by the net power applied yields theapplicator heating efficiency1 3. For a capacitive heatingsystem both the relative SAR distribution and its absolutevalue at any point depend on the separation of theelectrodes and on the materials between them. Theseparameters should be specified when reporting results.
18
For interstitial microwave techniques, in which thereduction in SAR with distance is rapid, the thermometerused should be positioned 5mm from the central axis of theconductor and at the level of the most proximal break in4 5
the outer conductor. Localized current field techniquesrequire a standardized spacing of the electrodes ifreproducibility is to be achieved. An appropriate spacingis 10mm with the absolute SAR being measured mid-waybetween two electrodes.
Regional approachcs such is the APAS and theMagnetrode are somewhat more difficult to handle. Whilstthe Magnetrode produces zero SAR on its axis this is notthe case with the APAS. For these devices absolute SARshould be determined at a location close to its maximumvalue in a standard phantom and under standard conditionsof geometry and machine settings, such as phase in thecase of APAS.
Perhaps the greatest value of the applicator heatingefficiency is that it facilitates inter-institutioncomparisons of heating techniques. At the present with awide variety of commercial and custom-built equipment inuse there exists considerable difficulty in identifyingcomparable treatments in different centers. Availabilityof a quantity which permits comparison even under rathersimple geometrical conditions will be of value whencollating and examining data collected from differentinstitutions or different modalities in the sameinstitution.
2.4 Coupling and Matching Devices
Coupling and matching devices are frequently employedwith external heating techniques. Frequently such devicesalso serve to limit the temperature rise at the patient’sskin. If a hyperthermia technique is to be adequatelycharacterized it is important that this feature of theequipment be understood.
At radiofrequencies with both the capacitive andinductive approaches, the impedance seen by the generatorcan be altered to maximize the power transfer to thetissue of interest. It is appropriate to identify therange of conditions under which efficient power transfercan take place. The applicator heating efficiency
19
[defined in Section 2.3] should be measured forrepresentative separations of the capacitor plate or coilfrom the muscle equivalent phantom. Of particularimportance is the effect of fat between the applicator andthe muscle like tissue which is to be heated. Theinfluence on the applicator heating efficiency of a l0mmthick fat layer can be assessed by comparing the heatingefficiencies with and without simulated fat in positionunder the configuration that optimizes power transfer inthe absence of fat.
Similar studies are required for external microwaveapplicators where the presence of fat can influence thecoupling efficiency.
It should also be noted that the area of the usefulthermal field (Section 2.2) can be influenced and in somecases determined by the coupling arrangement. Salinefilled bags used with capacitive techniques significantlyinfluence, through their size and conductivity, theh e a t i n g f i e l d . Similarly it is known that bags filledwith non-conducting de-ionized water, often used to couplewaveguides to patients, can influence both the applicatorheating efficiency, which they are intended to maximize,and the SAR distribution”. In designing experiments toinvestigate the influence of fat like tissue on theapplicator heating efficiency it should be recognized thatthe presence of fat will itself alter the SARdistribution. For consistency and to avoid the steep SARgradients at the fat-muscle interface it is recommendedthat the reference point for such measurements be 10 mmdeep in the muscle like tissue.
If such devices are to be employed clinically it isimportant that they be assessed under conditions thatrepresent those likely to be encountered during patienttreatment. The variability in coupling and its influenceon both absolute and relative SAR will form one limit onthe accuracy with which calculated temperaturedistributions can be obtained.
20
3. Temperature and its Distribution
The potential success of a planned hyperthermiatreatment can be gauged from a knowledge of the propertiesof the heating modality employed and the anatomyand physiology of the patient in the region of the target.The actual success of a hyperthermia treatment, in thesense of achieving and maintaining a desired temperaturedistribution, cannot be assured solely on the basis ofprior measurement whether in the laboratory or clinic. I tis certainly the case now and likely to remain so for theforeseeable future that in vivo temperature measurementswill be required during each and every treatment.
Two approaches to thermometry may be identified. Thefirst is invasive thermometry in which temperaturemeasurements are made at specific and predetermined pointswithin the heated region. The number of measurement sitesfor this approach is clearly limited by practicalconsiderations. The alternative of non-invasive methods,which in principle are capable of yielding threedimensional temperature maps, have considerable appeal forhyperthermia applications. Although many such techniqueshave been investigated none has yet been identifiedwhich combines the necessary spatial and temperatureresolution over clinically relevant volumes.
Invasive thermometers are the only devices so farwhich have demonstrated clinical utility and the ensuingdiscussion is therefore devoted to these.
3.1 Thermometer Description
In assessing the suitability of a particulartechnique for a particular heating modality it isnecessary that the underlying physical principles ofoperation be understood. Only with such informationis it possible to evaluate the extent of compatabilitywith the heating technique and the anatomical regionof interest. Thorough reviews of thermometers forhyperthermia have been given elsewhere36,50 and only afew of their major characteristics will be re-iteratedhere.
21
Thermocouples remain popular devices formeasuring temperature in the clinical environmentwith over 60% of reported studies having relied ontheir use for temperature measurement. They offerthe advantages of cost, ease of fabrication, size,adequate accuracy, stability and multisensorcapability. The major drawback of thermocouples haslong been recognized as their interaction with thestrong electromagnetic fields employed frequently inclinical hyperthermia5 2 , 5 3. In addition they sufferfrom from thermal smearing due to heat conduction in theirmetallic leads and, depending on their coating andpackaging are known to interact with ultrasonicf i e l d s 4 0. Sources of error including interactioneffects must be understood and appreciated and aredicussed in Section 3.3 below.
Thermistors are extremely sensitive devices and,when connected to the measurement electronics throughhigh resistance leads, will neither perturb nor beperturbed by electromagnetic fields. Currently,multisensor probes of a size acceptable for clinicaluse have not been fabricated and thus to extract themaximum information from an insertion resort must bemade to tracking techniques55,56. The remaining twoclinically applicable techniques both involvetemperature dependent optical properties of crystalswhich are remotely sensed through non-conducting opticf i b r e s5 7 , 5 8 , 5 9 . Interaction between such thermometersand electromagnetic fields is not measurable and thesedevices are usually described as minimally perturbing.Manufacturer’s literature can be consulted to identifythe smallest catheter into which the sensor will fit.Susceptibility to humidity and the method ofsterilization to be used are general properties ofthermometer systems which should be known beforeembarking on clinical treatments. As measurementduring treatment is vital in clinical hyperthermia,the gross features of the thermometry system must beidentified and appreciated at the treatment planningstage. Armed with such knowledge wiser choices ofthermometer location and measurement technique will bemade.
22
3.2 Intrinsic Characteristics of the Thermometer
Both during the stage of planning a hyperthermiatreatment and reporting the results of a clinicalstudy, awareness of the intrinsic characteristics ofthe thermometry system is valuable. It is well knownthat the biological response of tissue to hyperthermiais a strong function of temperature and from thisperspective high accuracy of the thermometry system isrequired.
It is recommended that thermometry systems forclinical use during local or regional hyperthermia becalibrated to an accuracy of 0.2°C with a precision of0.l°C. Precision is defined in its usual way toreflect the reproducibility of measurement by onethemometer or the agreement among differentthermometers belonging to one thermometry system.These criteria are easily met by most availablethermometry systems in the absence of artefacts(Section 3.3). In the special case of whole bodyhyperthermia highly accurate temperature contol isvital if serious side effects of the treatment are tobe avoided. In this situation, which is accompaniedby a fairly homogeneous temperature distribution, anaccuracy of 0.05° C is recommended.
It should be noted that a temperature accuracy of0.2°C implies effective thermal contact with thetissue being monitored. There are two conditionsunder which this can be difficult. Under somecircumstances, which should be clearly identifiedduring the treatment planning process, precise controlof the surface temperature may be necessary. Tomerely ensure that the surface does not rise abovesome predetermined level may be inappropriate. Insuch cases it will be necessary to ensure that thetemperature sensor accurately reflects the surfacetemperature and not that of any cooling techniqueapplied.
The second situation in which poor thermalcontact can impair temperature accuracy arises whentemperatures are changing rapidly37. Pull treatmentplanning requires consideration of intended rates of
23
rise of tissue temperature and particularly for slowlyincreasing temperatures the ‘thermal dose’accummulated during this phase of the treatment can besignificant. However, for rapidly rising temperaturesthe time constant of the probe, which is determinedlargely by the properties of its thermal insulation,could lead to temperature measurement errors andpossibly even a safety hazard due to excessive hotspots not being detected rapidly. An appropriatesampling and display rate will also, of course, benecessary if a short probe rime constant is to beexploited.
The sensitivity of themometry systems in routineclinical use is typically 0.l°C, which is adequate formost purposes. The only situation in which greatersensitivity may be required is during attempts toquantitate blood flow [or effective thermalconductivity] using thermal washout following theinterruption of heating1 1 , 6 0.
The final intrinsic property of a thermometrysystem which should be considered is stabilitv. Thefrequency of calibration is chosen such that, with theknown stability of a system, the precision andaccuracyrecommendations outlined above are not exceeded.Systems whose stability,when measured under the mechanical stressesexperienced during clinical treatment, is so poor thatthese levels cannot be maintained during a singleheating session are not appropriate for clinical use.
3.3 Sources of Error in Temperature Measurement
With the recognition both that in vivotemperature measurement is a vital component ofclinical hyperthermia and that relatively highaccuracy is required due to the rapid dependence ofbiological effects on temperature, considerable efforthas been invested in studies of possible sources oferror. As these errors impact upon clinicalprocedures and hence should be considered during theplanning of a treatment they are discussed in thisdocument. Interference with the sensitive electronics
24
which conditions and records the signals from thetemperature sensor has been a long recognizeddifficulty of making measurements in the presence ofstrong electromagnetic heating fields. Techniques forminimizing the consequences of electromagneticinterference have been discussed” and manythermometry systems have been designed with shieldingand filtering configurations aimed at the eliminationof this problem6 2. The presence of interference can beidentified by laboratory experimentation” or in theclinic by the very rapid change in indicatedtemperature which accompanies changes in appliedpower5 1. If, such an effect is present it must beregarded as a systematic error and steps taken toeliminate it altogether6 4. In those cases in whichelimination is not possible this effect can be handledby making measurements during brief [<1 sec]interruptions of power. At the planning stage thefrequency of measurement and hence of powerinterruption must be decided. To avoid overshootingthe target temperature by more than 0.5°C,measurements will need to be made at times whichcorrespond to this temperature change during theinitial phases of heating. Once the plateautemperature has been attained clinical experience willdictate an appropriate frequency of measurementalthough 1 per minute would seem to be suitable andeasily achievable in most cases.
In addition to interference effects metallicthermometers, as are commonly used, are afflicted byproblems associated with their self heating. Themagnitude of the effect under one particular-set ofexperimental conditions has been quantified”’ and itssignificance for the equilibrium temperaturedistribution assessed6 5. A study of techniques fordealing with self heating errors has concluded thatthe effect on the sensor itself may be effectivelyeliminated by making measurements during powerinterruption 6 6. However, the period after cessationof heating before a reliable measurement can be madeis considerably longer than that required to eliminateinterference effects being several probe timeconstants which could be as long as 5-10 sec. Thissame study66 failed to identify a method for
25
effectively dealing with the perturbation of thetemperature distribution due to an artefactual heatsource within the tissue. Should a thermometer underclinical conditions exhibit a degree of self heatingwhich is likely to significantly perturb thetemperature distribution it is inappropriate for ahyperthermia application.
If a small to moderate degree of self heating isknown to occur [which is insufficient to perturb thetissue temperature by more than 0.1°C] it can be dealtwith by a power interruption of adequate duration.The decision on whether or not to employ such atechnique and how to employ it must be made at thetreatment planning stage and based on the known orexpected characteristics of the thermometer. Othereffects havethermocouples67
been identified during the use of bothand diodes6 8 and the significance of
such artefacts for clinical hyperthermia shouldclearly be assessed.
A further effect which can be significant when thetemperature sensor assembly is thermally conducting issmearing or distortion of the temperature distributiondue to conduction along the sensor. Such an effect isclearly more significant in high temperature gradientssuch as are found frequently in interstitialapplications; close to large blood vessels; in thepenumbra of the heating field and near the surface.At the planning stage some estimate of the expectedmagnitude of the effect should be made3 6 , 3 8 , 6 9. If theerror exceeds or is likely to exceed that stipulatedin Section 3.2 an appropriate correction will berequired.
Thermal lag due to the insulating properties ofthe catheter within which the temperature sensor islocated can, under certain conditions, influenceclinical procedures. If a tracking technique isemployed to maximize the temperature informationobtained per insertion55,56 sufficient time must beallowed at each location for thermal equilibrium to beattained. This time will depend on the time constantof the sensor within its encapsulation3 7. Similarly
26
if it is intended to determine the SAR at certainlocations in vivo by observing the rate of change oftemperature, accommodation must be made for timeconstant effects”.
An additional problem has been identified whenperforming temperature measurements in ultrasonicf ie lds . The interaction of the field with plasticcatheters has been observed3 6 , 4 0 ’ 7 0’ 71 and clearly theselection of a thermometry system for use under theseconditions must acknowledge the existence of thisef fect .
Familiarity with the behaviour of the particularthermometer available is essential if temperatureinformation obtained during clinical hyperthermia isto be a valid and meaningful record of the treatment.
3.4 Thermometer positioning
The aim of thermometry in hyperthermia is, bymeasurement or calculation or by a combination of thetwo, to determine the distribution of temperature [andits time course] in three dimensions. Particularly inview of the marked lack of homogeneity of temperatureachievable with almost all heating modalities,position accuracy is of vital importance. In Section4 of this report recommendations concerning thelocations of thermometers during clinical hyperthermiaare given. In this section general considerationsconcerning thermometer positioning are discussed,
With the exception of whole body hyperthermia,temperature gradients of between 1 and 10°C cm -1 arecharacteristic of current clinical hyperthermiaheating technology. The lower value is achievablewith regional heating techniques while higher valuesare typical of gradients encountered during theapplication of interstitial techniques, near thesurface of externally heated sites and, in general,near tumour/normal tissue boundaries60. Thereproducibility of set up from treatment to treatmentshould ideally be 2mm at least where position withrespect to both patient anatomy and applicator must be
27
considered. It is recognized however that this levelof precision may, under clinical conditions, bedifficult to reach.
In the absence of other constraints it is clearthat thermometry will be more reproducible if thelocations of measurement are in regions wherelarge temperature gradients are not expected. Unlessthere are good reasons, which will be consideredduring the treatment planning process, measurementsites close to electrical or thermal discontinuitiesshould not be used to control the treatment. Largeblood vessels which could constitute significant heatsinks6 9 may require special consideration if they liewithin the intended therapeutic volume of the heatingf ie ld .
The use of linear arrays of sensors is aconvenient technique for increasing the temperatureinformation which can be obtained from the insertionof one catheter. An alternative method which employsonly one thermometer to reach the same objective is alinear tracking technique5 5. In view of therecommended precision of positioning it is unlikely tobe appropriate to routinely make temperaturemeasurements at locations closer together than about5mm. Exceptions to this may be unavoidable ifcritical regions of the patient coincide with steeptemperature gradients and in such cases precautionswill be necessary3 8.
In addition to considering the precision oflocation of temperature measurement, accuracy shouldalso be of concern. Characteristics of thethermometry technique which can lead to systematicerrors in the identification of the location ofmeasurement must be recognized and appropriateallowances made. Several investigations of sources ofthis type of error have been made38,67 . It appearsthat with thermally conducting probes, conductionalong the probe can have a significant influence whichincreases as the temperature gradient increases.
28
Finally, a technique for identifying the locations oftemperature measurement relative to the patient’s anatomyand the applicator position must be selected beforeclinical studies commence. The specification of atemperature achieved during clinical hyperthermia withouta clear and accurate description of the location of themeasurement is of limited value. Unfortunately, manypublished reports reflect this shortcoming. Severaltechniques are available for identifying the location of atemperature measurement. Computed tomography in thepresence o f radio--opaque skin markers to indicateapplicator position is universally applicable whenavailable. Stereo or orthogonal x-ray films taken withthe probe catheter filled with a radio opaque substancecan be employed to identify the position in threedimensions of the catheter. Techniques and computerprograms designed for use with radiation brachytherapy arewell known and available. Tomographic techniques such asultrasound and MRI can also be used to help identifycatheter location.
Bearing in mind that catheters used to containclinical thermometers have diameters which usually are notless than 1mm, a realistic spatial accuracy for clinicalhyperthermia is 2mm. The techniques mentioned above aregenerally capable of identifying catheter location withrespect to patient anatomy to this degree of accuracy.
3.5 Control circuits, display and recording
With the advance of technology, increasinglysophisticated methods are available for the real timemanipulation, display and recording of the large amount oftemperature data accummulated during a clinicalhyperthermia session. It is clearly an important aspectof the planning of a treatment to decide:
a) which and how much temperature data will berecorded?
b) at what time intervals?
c) which signals will be used to determine the courseof treatment? and
d) will automatic feedback control be employed?
29
The use of microprocessor controlled thermometry whichprobably contribute to the majority of current clinicaltreatments allows considerable flexibility in theapplication of real time thermometry. High capacitystorage devices offer the added advantage of thepossibility of detailed retrospective analyses of aheating session and the value of this facility will bediscussed in the next section. Here the features ofcontemporary thermometry systems which are likely to be ofgreatest value to hyperthermia dosimetry are considered ingeneral terms.
Multichannel thermometry is now regarded as essentialdue to the inhomogeneous nature of the temperature fieldproduced during clinical hyperthermia. The large amountof data yielded by such an approach brings with it theproblem of how to present information to the operator inan easily digestible form. One difficulty commonlyencountered is correlating the various temperaturechannels, which are usually identified by numbers, withtheir positions in a patient and hence their relativeimportance in determining the course of treatment. I t i smandatory that a diagram of relevant patient anatomy andthe probe catheters be available at each treatment tofacilitate interpretation of the data presented.
Digital displays of individual temperature channelsare, in the absence of other information. difficult tointerpret as a treatment is being assessed in the clinic.A bar graph presentation is of somewhat more value as someidea of the temperature distribution may be gleaned moreeasily. Possibly the most informative display for theoperator controlling the treatment is a graph showing thetime course of each temperature channel. It is notuncommon for patient movement or some other factor tonecessitate readjustment of the applicator position duringtreatment. A visual representation of the treatmentbefore and after such an adjustment greatly aids therapid optimization of set up.
Two additional operating characteristics ofthermometry systems warrant brief discussion. The firstis the frequency with which each temperature channel isread and the presented data updated. With typical
30
temporal temperature gradients of 1°C min-1 a readfrequency of 10 min-1 for each channel is consistent withthe recommended measurement accuracy. The second is thefrequency of storage of temperature data on disk or tapefor later analysis. Under most conditions storage of dataat 1 min intervals will provide sufficient information tocharacterize the behaviour of temperature at eachmeasurement site. Circumstances under which a morefrequent storage of temperature information will benecessary are special investigations such as thequantitation of blood flow through thermal clearancetechniques or the determination of SAR in vivo byexamining the initial rate of rise of tissue temperature.
The final consideration to be reviewed at the time ofplanning a hyperthermia treatment is automatic control ofthe power output of the heating device based on incomingtemperature information. Technically this approach isstraightforward but it requires a high level of confidencein the selected clinical technique and control temperaturepoint(s) if it is to be implemented. Hyperthermia isstill at the stage where operator and attending physiciancompetence and experience are vital ingredients for asuccessful hyperthermia program and, with very fewexceptions, it seems unlikely that control of a clinicaltreatment session will be delegated to a machine in theforeseeable future.
This section of the report has addressed in generalterms the accummulation of temperature data during aclinical treatment. With present invasive thermometry theinformation obtained in this way will clearly be spatiallyincomplete. Methods by which this shortcoming may beremedied will be among the topics to be discussed underthe heading of Treatment Planning.
3.6 Surface Cooling
Although a surface cooling facility need not ingeneral influence the SAR distribution it frequentlyserves the additional function of an electric or acousticimpedance matching device (Section 2.4). In any event itscapability will need to be appreciated if effectivetreatment planning is to undertaken.
31
Forced air systems are relatively easy to implementwhen external microwave or radiofrequency techniques areemployed. With the flow of air parallel to the patientsurface cooling may be uneven if the surface anatomy isirregular and this effect should be recognized. Air flowthrough the applicator and therefore normal to the surfacemay be more appropriate than parallel flow and should bedesigned such that the air flow is greatest where thepower density is greatest.
Water, whether de-ionized or containing sodiumchloride, flowing through a flexible “bolus” between theapplicator and the patient is probably the most commonlyused means of limiting the rise of skin and near surfacetemperature.
With such devices it is necessary to appreciate thatthe efficiency of heat removal from the skin depends onthe thermal contact berween the surface and the coolingl i q u i d4 8. Under conditions of uneven surface, contact maybe poor in places resulting in relatively high andunpredictable temperatures.
4. Treatment Planning
Treatment planning, whilst clearly an essential aspecto f clinical hyperthermia, is fraught with difficulty.Certainly the confidence placed in radiotherapy treatmentplans can not be justified in the case of hyperthermia.However, recent years have seen considerable activity inthis area resulting in the development of philosophies andtechniques. In what follows different paths by whichtreatment planning for hyperthermia may be approached aresuggested. Various degrees of sophistication can beemployed along these paths and in general the limitationon sophistication will be imposed by limitations ofresources. The four approaches described below follow theoriginal suggestion by Cetas and Roemer72 with theirclassification given in parentheses for Sections 4.2 to4.5.
4.1 Identification and description of the target volumeand adjacent tissues
32
The obvious first step in any attack on localizeddisease is the identification of the target. It is widelyrecognized that the precision with which this can becarried out in the case of malignant disease is in generalpoor. Generous margins around identifiable disease areemployed whenever possible to account for microscopicspread. It is also noted that the limitation on targetvolume for conventional radiotherapy is frequently imposedby the proximity of adjacent normal and radiosensitivetissues. In spite of these difficulties it is essentialto employ the most precise and sensitive techniquesavailable to achieve this first step in the treatmentplanning process.
The most appropriate and convenient currentlyavailable technique for localization of the tumour anddisplay of adjacent anatomy is X-ray transmission computedtomography [CT]. Multiple consecutive slices containprobably the most complete and spatially preciseinformation available for any anatomical location. Forcertain lesions ultrasonic scanning can be a satisfactorytechnique. Visual inspection and palpation can serve toidentify the circumference of many superficial tumors suchas chest wall recurrences although precise informationconcerning the depth of invasion will be lacking. Planefilm X-radiography is routinely employed for localizationfor conventional radiotherapy and, although it can fulfilla similar role for hyperthermia, it is inferior to CT inboth applications. The contribution of Magnetic ResonanceImaging to target localization for hyperthermia is, atthis time, unclear.
As in conventional radiotherapy, treatment planningfor hyperthermia requires as its first step a series ofpatient contours with the target volume indicated. CT isthe most appropriate source of this information andideally it should be available in parallel contiguousplanes.
An important point to note at the treatment planningstage is the need to record accurately the locations ofthermometers used during the clinical application ofhyperthermia. In many studies to date only the vaguestinformation is available concerning thermometer positionand this shortcoming certainly hampers attempts to compare
33
data not only from different centers but also from thesame center. The representation of patient anatomy andtarget data in the form of contours should allow for theentry of thermometer position at a later time.
In addition to the spatial resolution available withX-ray computed tomography a further reason for employingthis imaging modality is its ability to discriminateclearly between different tissues and hence identifyelectric or acoustic inhomogeneities. Very few centerscan currently handle such information accurately with theaim of deducing Specific Absorption Rate distributions.However the behaviour of electromagnetic and ultrasonicradiation in different tissues is qualitativelyunderstood 49,73,74 and therefore the presencerparticularly of adipose tissue, bone and air cavitiesshould be shown on the patient contour.
Thermal inhomogeneities are also obviously ofconsiderable significance for clinical hyperthermia andthese must be indicated on the patient contour. Ofparticular significance in this category is the presenceof large blood vessels which have been shown tosignificantly influence the temperature distribution intheir neighbourhoods6 9. Ideally the blood flowdistribution in and out of the plane of the plan should beknown and not just the sites of gross differences. At themoment it is possible to do little more than infer fromanatomical data the likely degree of non-uniformity offlow within the heating field, however, the development ofthermal washout techniques”” hold promise for the future.
Currently the majority of hyperthermia treatmentsare performed in conjunction with radiotherapy. The dosedistribution from interstitial or external sources ofionising radiation can be calculated with high accuracy inalmost all cases and both treatment planning andretrospective evaluation of clinical studies clearlyrequire details of the radiation dose distribution. Avital component of treatment documentation is the recordof adjuvant therapy and in the case of radiotherapy thisrecord must include details of the dose distribution withrespect to the actual patient anatomy.
34
In summary, treatment planning for hyperthermiarequires, as it does for radiotherapy, an identifiedtarget volume superimposed upon the actual patient’sanatomy. Such data in a quantity [ie. number of planes]which are sufficient to describe the geometry of theheated region to acceptable accuracy are essential as thefirst step of the planning process. Few centers arecapable of even estimating SAR distributions from maps ofelectrical or acoustic tissue properties; however, suchinformation should be indicated where possible on thepatient contour. Evidence of non-uniform blood perfusionshould likewise be indicated as significant temperaturenon-uniformity may well be present as a result.
4.2 General considerations of heating modalities(Comparative thermal dosimetry)
Comparative thermal dosimetry for use in treatmentplanning describes the process of selecting the mostappropriate heating technique and applicator for aparticular application. This process does not depend upondetailed accurate predicted temperature distributions butrather on matching available heating techniques to thetarget through a knowledge of the behaviour of the heatingmodality.
Targets may be grossly classified for this purposeinto superficial [distal surface within 4 cm of the skin]and deep [all other sites]. Most of the clinical dataaccumulated to the present time has been gained from thetreatment of superficial sites and the techniques forproducing a therapeutic temperature rise under suchconditions have been extensively studied. The mostpopular approach remains microwave heating usingfrequences which range up to 2450 MHz. Additional“external” techniqescapacitive heating76
include ultrasonic irradiation7 5,and inductive heating77 although the
latter, due to its toroidal heating pattern, is notstraightforward to implement for most superficial sites.
The heating of deep seated tumours remainsproblematical. Several techniques have been developedwhich, theoretically at least, could be expected toproduce therapeutic elevation of temperature deep in thebody. These include inductive heating’““’ ultrasonicirradiation” and the use of constructive interference of
35
electromagnetic waves81. It may be concluded that, atthis stage, such techniques of heating deep seated tumoursare still very much experimental and should not beundertaken in those centres which lack the physics,engineering and medical support for their safeimplementation.
An approach to clinical hyperthermia which has foundmany followers in recent years and which may be applied toboth superficial and deep seated tumours is interstitialhyperthermia. Clearly surgical intervention of acomplexity depending on site is necessary if this approachis to be adopted. However, as with interstitialradiotherapy techniques the treated volume is localizedand, with effective treatment planning, neighbouringnormal tissues may be spared damage.
The first stage in comparative treatment planning isto ensure that the target in three dimensions isadequately covered by the proposed heating technique.Such a comparison will initially be undertaken assuminghomogeneity of electric [or acoustic] and thermophysicalproperties.
The next stage is, through a knowledge of thephysical behaviour of the heating beam in heterogeneousstructures, to consider the likely impact of local anatomyon the SAR distribution. An extreme example of thisprocess is the virtual impossiblility of employingultrasound to heat regions with overlying air or bone.More commonly encountered complications of this type,however, concern the effect of subcutaneous fat layers andunderlying bone for example when treating chest-wallrecurrences. In the case of heating through a layer offat it should be appreciated that excessive heating of thefat layer will occur when using capacitive techniqueswhilst adipose tissue is relatively transparent tomicrowave radiation49 unless there is a component of theelectric field perpendicular to the fat muscleinterface 8 2. The presence of ribs relatively close to thesurface will lead to the reflection of both microwave andultrasonic beams with the possibility of the production oflocal hot spots. In comparing available heatingtechniques such effects must be recognized at the planningstage, and taken into account in the selection of the mostappropriate modality.
36
A physical property of the heating technique and ofthe applicator employed which must also be considered isthe penumbra of the SAR distribution. It is clear thatthe SAR penumbra is not identical to the temperaturepenumbra due to smearing through conduction andconvection. However, the sharper the SAR penumbra thesharper will be the temperature penumbra. If the targetvolume lies in close proximity to potentially thermallysensitive or low heat capacity normal tissues it may benecessary to select a heating modality such as ultrasoundin which a steep fall off of SAR can be achieved.
A related consideration is field shaping. Theflexibility to shape the heating field of a microwaveapplicator is very limited as the matching between thepatient, applicator and generator can be impaired by thepresence of reflectors.antennae83
Recently developed microstripdo, however, permit some flexibility in the
selection of the size and shape of microwave heatingf ie lds . Scanning techniques of the type employed withultrasonic irradiation can achieve field shaping with theuse of complex mechanical and control systems8 0 , 8 4 c o n t r o lof multielement transducers has been used to the samee n d3 0. With capacitive heating custom designed salinebags can be used to influence the heating field4 6.
Field shaping is probably most easily accomplishedwith interstitial techniques. Although the hyperthermiaequivalent of Patterson-Parker or Quimby rules are not yetavailable, and in fact the development of such rules maynot be possible because of living tissue dynamics, aknowledge of the heating pattern around an interstitialapplicator may be used to identify with reasonableaccuracy the volume likely to be raised to therapeutictemperature levels for any given configuration ofapplicators. The actual temperature rise achieved and itsdistribution will, as with any other technique, depend onheat transfer mechanisms and, in particular, on bloodf l o w4 1.
Having considered applicator performance and thelikely consequences for the SAR distribution of electricalor mechanical inhomogeneities it is next necessary toassess the significance of thermal and physiologicalnon-uniformities. For superficial sites the most common
37
and easily handled thermal inhomogeneity occurs at thesurface. At the planning stage a decision has to be madeon the required surface temperature rise and this decisionwill depend upon whether or not the surface forms part ofthe target volume. It is frequently the case that it isrequired to provide skin sparing and this, fortunately, isrelatively straightforward. Both water cooling devices85
which can form part of the impedance matching network andair cooling have been effectively employed. It isimportant to recognize, however, that very steeptemperature gradients can occur at cooled surfaces andthat accurate monitoring of surface temperature isdi f f icult . It should also be recognized that whilstsurface cooling obviously cannot alter the SARdistribution produced by a particular applicator it canhave a significant influence on the temperaturedistribution even to a depth of a few cm in the tissue8 6.In particular the position of maximum temperature will bemoved deeper into the tissue and be reduced in magnitude.The consequences of surface cooling will be to increasethe depth at which a therapeutic temperature rise can beachieved without causing excessive temperatures moresuperficially and an increased requirement for absorbedpower to produce the required temperatures.
Thermal and physiological inhomogeneities apart fromthose encountered at the surface are much more difficultto handle. In preparing the anatomical data for use inhyperthermia treatment planning it was noted that thepresence and location of such inhomogeneities should berecorded. Should significant inhomogeneities exist withinthe treated field it must be recognized at this stage thatsevere temperature non-uniformities will ensue. Anassessment of the likely impact of these on both thesafety and likely efficacy of a hyperthermia treatmentshould be made early on in the treatment planning process.
The discussion to this point has reviewed thefactors which will influence the choice of applicators andtechniques for a particular treatment. At the currentstate of knowlege and computational ability a substantialamount of intuition is used to make the decisions requiredfor hyperthermia treatment planning. Certainly theobjectivity employed in assessing competing radiation
38
treatment plans cannot at this time be used forhyperthermia. Limited progress is however being made inthis direction with some published data comparing heatingsystems through predicted temperature distributions inmodels which incorporate typical blood f lows for dif ferento r g a n s 8 7. As, ideally, it is the estimated temperaturedistributions upon which a decision should be made suchefforts will undoubtedly aid comparative thermaldosimetry.
In summary, the selection of the most appropriateheating system for a particular patient and tumour site isprobably the most important decision to be made during thetreatment planning process. Although the information uponwhich this decision can currently be based is bothincomplete and almost certainly inaccurate anunderstanding of the physical and physiological processeswhich determine the distribution of temperature elevationin vivo is essential to maximize the probability ofeffective treatment being delivered.
4.3 Treatment Planning (Prospective thermal dosimetry)
Ideally, one wishes to carry out a prospectiveprocedure for hyperthermia which is analagous to thatroutinely employed in conventional radiotherapy and with acomparable degree of accuracy. Unfortunately, in the caseof hyperthermia a dynamic variable, temperature, and itsdistribution are ultimately required whereas in ionizingradiation therapy, the key quantity, dose is passive andmerely describes the energy deposit ion. Th is d i f f e rencebetween radiotherapy and cl inical hyperthermia is chieflyresponsible for placing accurate prospective temperaturedosimetry out of reach at the present.
In spite of this dif f iculty, however, some progressin this area has been made. The f irst step in computingthe temperature distribution in vivo is to establish thedistribution of SAR. This quantity possesses the units ofradiation dose rate [W/kg] and is clearly not influencedby physiological processes. Several heating modalities inroutine clinical use employ coherent beams of microwavesor ultrasound and, particularly in the case of the former,
39
phase relationships must be taken into account. Thisaspect of the computation of SAR results in greatercomplexity than that encountered with incoherent x-raybeams. However, progress has already been made incalculating SAR distributions under realistic anatomicalc o n d i t i o n s8 7 , 8 8 it is to be expected that with increasesin computational speed increased accuracy will beachievable routinely. Such computations should beundertaken whenever possible as a guide to applicatorselection.
As mentioned earlier the major difficultyencountered in true treatment planning for hyperthermia istranslating the rate of energy absorption into temperatureelevation.- Required for this-procedure is completeknowledge of the distribution of the thermal properties ofthe tissue being heated. The most significant of theseproperties is frequently convective heat loss throughblood flow and the manner in which it responds to localtemperature elevation.
At the moment and for the foreseeable future thereis no way of establishing reliable input blood flow dataupon which to base calculations of temperature. The bestthat can be achieved at the moment in this regard is toperform calculations based on easiest-to-heat andhardest-to-heat conditions89 which are selected to bracketthe most likely values of the actual thermal properties ofthe anatomical region of interest.
In summary, treatment planning for hyperthermia asunderstood in analogy to radiation therapy is notcurrently possible and indeed may never be possible.Calculations of SAR distributions, however, are beingperformed and with reasonable accuracy under certainconditions. whilst the information provided by suchcalculations is limited in its value for clinicalhyperthermia, it does have several uses. Such SARdistributions permit comparative thermal dosimetry to beundertaken with more confidence and hence can lead to amore enlightened selection of the optimum heating system.Also, in the absence of gross thermal inhomogeneities thegross features of the temperature distribution will bestrongly influenced by the SAR distribution. Thus, under
40
many conditions the locations of hot and cold spots may bepredicted on the basis of SAR data to an accuracy which issufficient to provide guidance in the clinical set up.Strong temperature gradients can be expected to accompanystrong SAR gradients and the knowledge of the locations ofsuch regions may well influence the placement ofthermometers. It is recognized that slight positionerrors could result in significant temperature errors whengradients are strong.
4.4 Specification of in-vivo thermometry technique(Concurrent thermal dosimetry)
Thermometry during a clinical treatment is the mostreliable approach at the moment to safe and efficacioustherapy. It is also the most widely practised. It isclear from the previous section that accurate prospectivedosimetry is not a realistic option at this time. Whilethe optimum technique can and should be based at least onunderstanding, and preferably calculations of, theexpected heating field, reliance cannot be placed oncalculated temperature distributions carried out beforethe thermophysical properties of the tissues have beenexperimentally established.
Furthermore, although attempts are being made todevelop non-invasive techniques of temperature mapping49
these are still, at best, a long way from possessingeither the accuracy or clinical utility for routine use.Thus, invasive monitoring at the time of treatment remainsthe only acceptable approach to clinical thermometry.
A clear limitation in determining the threedimensional temperature distribution in and around thetumour, and this, of course, is the ultimate goal, is thelimited number of points at which temperature can bemeasured. As, for reasons of sterility, temperatureprobes are most commonly inserted into catheters it may bemore appropriate to consider the measurements to belimited to a set of lines. In section 3.2 of this reportit was pointed out that tracking techniques or the use ofthermometer arrays maximizes the information obtainablefrom each catheter. Such techniques are to be encouraged.It is also desirable to implant as many catheters as
41
possible although this will be l imited by the number oftemperature channels which can be monitored and by patientcomfort .
If the number of paths, along which temperature canbe monitored is l imited how are these to be chosen? Acomprehensive set of animal data has strongly indicatedthat the most important predictor of tumour response[other factors being equal] is minimum tumourtemperature 1 3. On the basis of this conclusion it wouldappear reasonable to position at least half the availablecatheters along paths which intersect the periphery of thetarget volume.
Using the contour data produced as described inSection 4.1 together with calculated or estimated SARdis t r ibut ions i t i s poss ib le t o ident i fy l ike ly co ldregions of the target. Efforts should be made to ensurethat catheters pass close to such points. Of equalimportance are expected hot spots occurring in adjacentnormal tissue. In the absence of reliable calculated SARdistributions estimates need to be made of locations ofpossible hot spots based on the behaviour of theapplicator employed and the anatomy of the heated region.Such locations should, whenever possible, be monitored astheir temperatures may pose the limit on the intratumourtemperature achievable.
As a general rule it should be appreciated thattemperature measurements in regions of strong temperaturegradients are dif f icult to interpret due to the problem oflocating catheters precisely at predetermined positions.Such considerations are particularly relevant when thedecision is being made whether or not to monitor surfacetemperature. In add i t i on i t i s d i f f i cu l t t o ensurethermal contact between the thermometer and the skinwithout impairing surface cooling and hence influencingthe parameter being measured. It is easy to question theaccuracy of such surface temperature determinations. Oncethe eff icacy of a particular surface cooling technique hasbeen clinically established, l imited thermometry channelsmay be more usefully employed in monitoring subcutaneoustemperatures.
42
The placement of catheters described above has beenrecommended for those cases, which are the majority, wherea safe and effective treatment is the sole objective. Inother words, it is not the intention to attempt toreconstruct the three dimensional temperature field.
Clearly if the facilities are being developed for thederivation of the full description of the temperaturefield then alternative catheter arrangements may bedesirable. To date some progress has been made in thisrespect and concerted efforts by the hyperthermiacommunity to this end continue9 0. However, development ofthis important area is still limited and norecommendations can be given as to catheter placement forsuch purposes.
A vital aspect of in-vivo thermometry is the recordand specification of thermometer location. It is afeature of the majority of published clinical reports thatsuch information is notable by its absence. Such ashort-coming undoubtedly impairs intercomparison ofclinical studies and impedes the development of optimumclinical protocols. It is essential that the locations ofthe catheter tracks be identified as completely aspossible on the patient contour data. Further it isessential that the measurement points along those tracksbe reproducible and known.
Clearly an accurate technique of locating thecatheter track needs to be selected and routinelyemployed. X-ray computed tomography is the ideal tool forthis purpose as catheter location is overlaid directly onpatient anatomy. If such equipment is unavailable analternative choice is to use ultrasonic scanning which,provided the surface contour is not significantlydistorted through the action of scanning, is often capableof locating the catheter position. The easiest andpossibly most accessible technique in a radiotherapydepartment may however be the use of orthogonal X-rayfilms taken on a simulator. It is relativelystraightforward to reconstruct the relative positions ofradioopaque markers representing thermometer position36
and, provided these can be related to patient anatomy, allthe required information is obtained.
43
It cannot be stressed enough that recordingtemperatures however accurately without specifyingprecisely where these temperatures were measured withrespect to the target volume is almost meaningless.
Concurrent thermal dosimetry has as its aim not onlyto confirm that an acceptable temperature distribution isachieved in and around the target but also to provide theinformation that permits control of the treatment. Mostcommonly a single applicator is employed for superficialsites and the only control available over the temperaturedistribution is the total power delivered - the relativetemperature distribution cannot be deliberately altered.In such cases the coldest part of the target should beidentified during the temperature elevation phase and thisused to decide when the predetermined target temperaturehas been reached. Depending on the site it may benecessary to declare maximum temperatures at otherlocations and clearly, in such cases, several conflictinglimits may be placed on the power which can be applied tothe applicator. In the absence of upper temperaturelimits being encountered, control of the treatment isbased solely on one temperature channel representing thecoldest part of the target. It is important to verifyregularly during the treatment that, due to, for example,temperature induced blood flow changes, the grosstemperature distribution has not altered and that thecontrol channel is indeed at the coldest part of thetarget. Similarly, due precautions must be taken toensure that predetermined normal tissue temperatures arenot exceeded.
With multiple applicators, eg., interstitial antennae,or with applicator scanning techniques control becomesmore complex as both the magnitude of the temperatureelevation and its distribution is under operator control.Here a highly interactive process will be advantageous asconsiderable control can be exercised over the temperaturedistribution produced even under conditions of changingblood flow9 1.
For concurrent thermal dosimetry to be effectivelyimplemented many decisions prior to treatment arerequired. Amongst these the most important are thespecification of the location at which the temperature is
44
to be monitored and the maximum allowed and minimumdesirable temperature to be attained at key locations.
4.5 Reconstruction of the treatment in time and space(Retrospective Thermal Dosimetry)
The general goal here is twofold: first, theacquisition and evaluation of a large data base of theclinical, thermal information--treatment temperatures, SARvalues, and all relevant patient physical, anatomical, andmedical information [ie. the standard, recommendedradiological practice extended to include the hyperthermiatreatment data]; second, a detailed, computer model based,evaluation of the hyperthermia treatment to attempt toextend the amount of information that can be obtained fromthe clinical thermometry. For this second area, sincetemperatures are, at blest, measured along a few tracksduring clinical treatments, thus most of the locations inthe tumour have not actually had their temperaturesmeasured. This, of course, represents a very significantgap in the knowledge of what actually happened during atreatment, and is a fundamental problem for all clinicalhyperthermia. For example, treatments may fail becausethe actual minimum temperature in the tumour issignificantly less than therapeutic, even though theminimum measured tumour temperature may be therapeuticallyacceptable and indeed indicate that a very good treatmenthas occured. Thus, barring the successful technicaldevelopment and clinical implementation of non-invasivethermometry techniques, one must resort to developingimproved computer models of the treatments--models whichuse the clinically measured temperatures at the sensorlocations as a basis of predicting the completetemperature field. This is, of course a very difficultproblem whose successful resolution is to be considered inthe future. However, given the fundamental, importantnature of the expected results--knowledge of the completetemperature field--it is a problem well worth attacking.Presently, as one would expect, all of the efforts tosolve this problem are in the early research stages, andno tools are available for practical, clinicalapplication. However, the early research results lookpromising, with simple one dimensional layered modelshaving been successfully used to predict experimental
45
r e s u l t s 9 2 , 9 3 and some early numerical and experimentalresults showing tha the full three dimensional problem maybe solvable9 0. More progress can be expected in thefuture since this is a deterministic problem which cancertainly be solved up to a certain level -- the limits ofwhich are not presently known. When they are found, onemay be able to answer the ubiquitious question of how manytemperature sensors are needed in a treatment, and whereshould they be placed; questions whose answers are unknownat present.
46
1.
2.
3.
4.
5.
6.
7.
8.
9.
J. Overgaard, "The design of clinical trials inhyperthermic oncology", in Physics and Technologyin Hyperthermia, eds, S.B. Field and C. Franconi,Amsterdam: Martinus Nijhoff, [1987]
J. Overgaard, "Influence of sequence and interval onthe biological response to combined hyperthermiaand radiation", In Cancer therapy by hyperthemia,drugs and radiation, ed. L.A. Dethlefsen, NCIMonograph 61, pp 325-332, [1982]
J.L. Meyers, I. van Kersen and G.M. Hahn, "Tumourresponses following multiple hyperthrmia and X-raytreatments: role of thermotolerance at thecellular level", Cancer Res. 46, 5691-5695, [1986]
H.E. Johns and J.R. Cunningham, "The Physics ofRadiology" 4th Edition, Thomas, Springfield,Illinois [1983]
G.E. Hanks, D.F. Herring and S. Kramer, "Patterns ofcare outcome studies: Results of the nationalpractice in cancer of the cervix", Cancer 51,959-967, [1983]
T.R. Mackie, E. El-Khatib, J. Battista, J. Scringer,J. VanDyk and J.R. Cunningham, "Lung dosecorrections for 6 and 15 MV X-rays", MedicalPhysics 12, 327-332, [1985]
Proc. Symp. on quality Assurance of RadiotherapyEquipment. AAPM Symposium Proceedings No 3, AIP,New York [1983]
S.B. Field, "Studies relevant to a means ofquantifying the effects of hyperthermia", Int. J.Hyperthermia 3, 329-296, [1987]
J. Overgaard, "Some problems related to the clinicaluse of thermal isoeffect doses", Int. J.Hyperthemia 3, 329-336 [1987]
47
10.
11.
12.
13.
14.
15.
16.
17.
18.
H.H. Pennes, "Analysis of tissue and arterial bloodtemperatures in the resting human forearm", J.Appl. Physiol. 1, 93-122, [1984]
F.M. Waterman, L.T. Tupchong, R.E. Nerlinger, J.Matthews, "Determination of blood flow in humantumours during local hyperthrmia", [Abstract] Med.Phys. 14, 492, [1987]
B. Emami and C.W. Song, "Physiological Mechanisms inhyperthermia: a review" Int. J. Radiat. Onc. Biol.Phys. 10, 289-295, [1984]
M.W. Dewhirst, D.A. Sim, S. Sapareto and W.G. Connor,"The importance of minimum tumour temperature indetermining early and long term responses ofspontaneous pet animal tumours to heat andradiation", Cancer Res. 44, 43-50, [1984]
T,S. Herman, C.S. Zukoski and R.N. Anderson, "Reviewof the current status of whole body hyperthermiaadministered by water circulation techniques", inCancer Therapy by Hyperthermia, Drugs andRadiation. ed. L.A. Dethlefsen. NCI Monograph 61[1982]
R. Cavaliere, G. Moricca and A. Caputo, "Regionalhyperthermia by perfusion", in Proc. InternationalSymposium on Cancer Therapy by hyperthermia andRadiation, eds. J.E. Robinson and M.J. Wizenberg,ACR, [1985]
J.W. Strohbehn, "Calculation of absorbed power intissue for various hyperthermia devices", CancerResearch [Suppl.] 44, 4781s-4787s, [1984]
K. Hynynen, D.J. Watmough and J.R. Mallard, "Designof ultrasonic transducers for local hyperthermia",Ultrasound in Med. Biol., 7, 392-402, [1981]
A.W. Guy, "Electromagnetic fields and relativeheating patterns due to a rectangular aperturesource in direct contact with bilayered biologicaltissue", IEEE Trans. Microwave Theory Tech.MYY-19, 214-223, [1971]
48
19.
20.
21.
22.
23.
24.
25.
26.
27.
A.Y. Cheung, W.M. Golding and G.M. Samaras, "Directcontact applicators for microwave hyperthermia",J. Microwave Power, 16, 151-159, [1981]
P.F. Turner and L. Kumar, "Computer solutions forapplicator heating patterns" in Cancer Therapy byHyperthermia, Drugs and Radiation. ed. L.A.Dethlefsen, NCI Monograph 61, [1982]
T.V. Samulski, D.S. Kapp, P. Fessenden and A.Lohrbach, "Heating deep seated eccentricallylocated tumours with an annular phased arraysystem: a comparative clinical study using twoannualr array operating conditions", Int. J.Radiation Oncol. Biol. Phys. 13, 83-94 [1987]
B.R. Paliwal, F.A. Gibbs and A.L. Wiley, "Heatingpatterns induced by a 13.56 MHz radiofrequencygenerator in large phantoms and pig abdomen andthorax", Int. J. Radiation Onc. Biol. Phys. 8,857-864, [1982]
C.K. Chancy, J.L.G. Kern and R.L. Levin, "Energydeposition patterns in an amputated human lowerleg heateed with a miniannular phased array".Ned. Phys. 15, 17-23, [1988]
W. Swindell, "Ultrasonic Hyperthermia" in PhysicalTechniques in Clinical Hyperthermia, eds. J.W.Hand and J.R. James, Research Studies Press,Letchworth, England, 228-332, [1987]
C.T. Tsai, C.H. Durney and D.A. Christensen,"Calculated power absorption patterns forhyperthermia applicators consisting of electricdipole arrays", J. Microwave Power 19, 1-13 [1984]
A.W. Guy, "Analyses of electromagnetic fields inducedin biological tissues by thermographic studies onequivalent phantom models", IEEE Trans. MicrowaveTheory Tech. MIT-19, 205-214 [1971]
C.R. Chou, G.W. Chen, A.W. Guy and K.H. Luk,"Formulas for preparing phantom muscle tissue atvarious radiofrequences", Bioelectromagnetics 5,435-441 [1984]
49
28.
29.
30.
31.
32.
33.
34.
35.
E.L. Madsen, J.A. Zagzebski, R.A. Banjvie and R.E.Jutila, "Tissue mimicking materials for ultrasoundphantoms", Med. Phys. 5, 391-394, [1978]
J.B. Leonard, K.R. Foster and T.W. Athey, "Thermalproperties of tissue equivalent phantommaterials", IEEE Trans. Biomed. EngineeringBME-31, 533-536 [1984]
H.R. Underwood, E.C. Burdette, K.B. Ocheltree andR.L. Magin, "A multielement ultrasonichyperthermia applicator with independent elementcontrol." Int. J. Hyperthermia 3, 257-267 [1987]
C.K. Chou, A.W. Guy, J.A. McDougall, A. Dong and K.H.Luk, "Thermographically determined specificabsorption rate patterns of 434 MHz applicators".Med. Phys. 13, 385-390 [1986]
T.C. Cetas, "Temperature measurement in microwavediathermy fields: principles and probes", in Proc.Int. Symp. on Cancer Therapy by Hyperthermia andRadiation, eds. M.J. Wizenberg and J.E. Robinson,ACR, Maryland, 193-203 [1976]
P.B. Dunscombe, K. Gammampila and N.W. Ramsey, "Anexperimental comparison of conventional diathermyequipment for the production of localizedhyperthermia in man". Proc. 1st Meeting ofEuropean Group of Hyperthermia in RadiationOncology, eds. G. Arcangeli and F. Mauro, Masson,Milan, 9-14 [1980]
T.C. Cetas, J.M. Hevezi, M.R. Manning and E.J.Ozimek, "Dosimetry of interstitialthermoradiotherapy" in Cancer Therapy byHyperthemia, drugs, and Radiation, NCI Monograph61, 505-507 [1982]
P.D. Edmonds, E.R. Lee and P. Fessenden, "Spatialdistribution of heating by ultrasound transducersin clinical use, indicated in a tissue-equivalentphantom." Proc. IEEE Ultrasonics Symposium, SanFrancisco, 908-912 [1985]
50
36.
37.
38.
39.
40.
41.
42.
43.
44.
P. Fessenden, E.R. Lee and T.V. Samulski, "Directtemperature measurement", Cancer Research (Supple)44, 4799s-4804s [1984]
F.M. Waterman "The response of thermometer probesinserted into catheters", Med. Phys. 12, 368-372[1985]
B.E. Lyons, T.V. Samulski and R.H. Britt,"Temperature measurements in high thermalgradients: I. The effects of conduction", Int. J.Radiat. Onc. Biol. Phys. 11, 951-962, [1985]
T.V. Samulski, B.E. Lyons and R.H. Britt,"Temperature measurements in high thermalgradients: II Analysis of conductive errors.",Int. J. Radiat. Oncol. Biol. Phys. 11, 963-971,[1985]
K. Hynynen, C.J. Martin, D.J. Watmough and J.R.Mallard, "Errors in temperature measurement bythermocouple probes during ultrasound inducedhyperthermia", Brit. J. Radiol. 56, 969-970 [1983]
J.W. Strohbehn, B.S. Trembly and E.B. Douple, "Bloodflow effects on the temperature distributions froman invasive microwave attenna array used in cancertherapy", IEEE Trans. Biomed. Engineering BME-29,649-661, [1982]
S. Allen, G. Kantor, H. Bassen and P. Rugera, "CDRHrf phantom for hyperthermia systems evaluations",Int. J. Hyperthermia. 4, 17-24 [1988]
G.H. Nussbaum, "Quality Assessment and assurance inclinical hyperthermia: requirements andprocedures", Cancer Res.[1984]
[Suppl.] 44, 4811s-4817s
J.J.W. Langendijk and P. Nilsson, "Hyperthermiadough: a fat and bone equivalent phantom to testmicrowave/radiofrequency hyperthermia heatingsystems", Phys. Med. Biol. 30, 709-712 [1985]
51
45. J.D. Doss, "Calculation of electric fields inconductive materials". Ned. Phys. 9, 566-573[1982]
46. N.M.S. Reddy, V. Maithreyan, A. Vasanthan, I.S.Balakrishnan, B.K. Bhaskar, R. Jayaraman, V.Shawnta and S. Krishnamurthi, "Local R.F.Capacitive hyperthermia: thermal profiles andtumour response."[1987]
Int. J. Hyperthermia 3, 379-387
47. D.L. Denman, M.J. Kolasa, H.R. Elson, B.S. Aron andJ.G. Keriakes, "Specific absorption rates insimulated tissue media for a 10 x 10 cm 915 MHzwaveguide applicator",[1987]
Med. Phys. 14, 681-686
48. J.J.W. Lagendijk, "Heat transfer in tissue", in
49. J.
50. T.
51. P.
52. K.
Physics and Technology of Hyperthermia, eds. S.B.Field and C. Franconi, Amsterdam: MartinusNijhoff, [1987]
Conway and A.P. Anderson, "Electromagnetictechniques in hyperthermia", Clin. Phys. PhysiolMeas. 4, 287-318 [1986]
Cetas, "Thermometry", in Physics and Technology inHyperthemia, eds. S.B. Fields and C. Franconi(Amsterdam: Martinus Nijhoff) [1987]
Carnochan, R.J. Dickinson and M.C. Joiner, "Thepractical use of thermocouples for temperaturemeasurment in clinical hyperthermia", Int. J.Hyperthermia 2, l-19 [1986]
Gammapila, P.B. Dunscombe, B.M. Southcott andA.J. Stacey, "Thermocouple thermometry inmicrowave fields", Clin. Phys. Physiol. meas. 2,285-292 [1981]
53. D.P. Chakraborty and I.A. Brezovich, "Error sourcesaffecting thermocouple thermometry in R.F.electromagnetic fields", J. Microwave Power 17,17-28 [1982]
52
54.
55.
56.
57.
58.
59.
60.
61.
62.
R.R. Bowman, "A probe for measuring temperature inradiofrequency heated material", IEEE Trans.Microwave Theory Tech. MTT-24, 43-45 [1976]
F.A. Gibbs, "Thermal mapping in experimental cancertreatment with hyperthermia: description and useof a semi automatic system", Int. J. Radiat. Onc.Biol. Phys. 9, 1057-1063 [1983]
M.J. Engler, M.W. Dewhirst. J.W. Winget and J.R.Oleson, "Automated thermometer scanning forhyperthermia treatment monitoring, Int. J. Rad.Onc. Biol. Phys. 13, 1377-1382 [1987]
D.A. Christensen, "A new non-perturbing temperatureprobe using semiconductor band edge shift", J.Bioengineering 1, 541-545, [1977]
K.A. Wickersheim and R.V. Alvers, "Recent advances inoptical temperature measurements", IndustricalResearch and Development 21, 82-89 [1979]
K.A. Wickersheim, "A new fiberoptic thermometrysystem for use in medical hyperthermia",Proceedings SPIE conference on optical fibers inmedicine. Cambridge, MA. No. 713-34, Sept. 19,1986.
T.V. Samulski, P. Fessenden, R. Valdagni and D.S.Kapp. "Correlation of thermal washout rate, steadystate temperature and tissue type in deep[ seatedrecurrent or metastatic tumours", Int. J. Radiat.Oncol. Biol. Phys. 13, 907-916 [1987]
D.P. Chakraborty and I.A. Brezovich, "Error sourcesaffecting thermocouple thermometry in R.F.electromagnetic fields", J. Microwave Power 17,17-28 [1982]
P. Horseman, D. Law, P. Dunscombe, K. Gammampila andB.M. Southcott, "A computer basedmultithermocouple system for use in microwavef ie lds" , J. Red. Engineering Tech. 7, 96-100[1983]
53
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
P.B. Dunscombe, J. McLellan and K. Malaker, "Heatproduction in microwave irradiated thermocouples",Med. Phys. 13, 457-461 [1986]
K. Gammampila, P.B. Dunscombe, B.M. Southcott andA.J. Stacey, Thermocouple thermometry inmicrowave fields", Clin. Phys. Physiol. Meas. 2,285-292 [1982]
R.T. Constable, P. Dunscombe, A. Tsoukatos and K.Malaker, "Perturbation of the temperaturedistribution in microwave irradiated tissue due tothe presence of metallic thermometers", Med. Phys.14, 385-388 [1987]
P.B. Dunscombe, R.T. Constable and J. McLellan,"Minimizing the self-heating artefacts due to themicrowave irradiation of thermocouples", Int. J.Hyperthermia, 4, 437-445 [1988]
R.J. Dickinson, "thermal conduction errors ofmanganin-constantan thermocoupd arrays", Phys.Med. Biol. 30, 445-453 [1985]
P.W. Barth and J.B. Angell, "Thin linear thermometerarrays for use in localized cancer hyperthemia",IEEE Trans. Electron Devices ED-29, 144-150 [1982]
J.J.W. Lagendijk, "The influence of blood flow inlarge vessels on the temperature distribution inhyperthermia", Phys. Med. Biol. 27, 17-23 [1982]
F.M. Waterman and D.B. Leeper, "Temperature artifactsproduced by thermcouples used in conjunction with1 and 3 MHz", Int. J. Hyperthermia (in press,1989).
F.M. Waterman, R.E. Nerlinger and D.B.Leeper, "Catheter induced temperature artifacts inultrasound hyperthermia", Int. J. Hyperthermia (inpress, 1989).
T.C. Cetas and R.B. Roemer, "Status and futuredevelopments in the physical aspects ofhyperthermia", Cancer Research [Suppl.] 44,4894s-4901s [1984]
54
73.
74.
75.
76.
77.
78.
79.
80.
81.
P. Fessenden, E.R. Lee, T. Anderson, J.W. Strohbehn,J. Meyer, T.V. Samulski and J.B. Marmor,"Experience with a multitransducer ultrasoundsystem for localized hyperthermia of deep tissue",IEEE Trans. Biomed. Engineering BME-31, 126-135[1984]
P.P. Lele, "Physical aspects and clinical studieswith ultrasonic hyperthermia", in Hyperthermia inCancer Therapy, ed. F.C. Storm, G.K. Hall Medical,Boston, 333-367. [1983]
P.M. Corry, B. Barlogie. E.J. Tilchen and E.P.Armour, "Ultrasound induced hyperthermia for thetreatment of human superficial tumours", Int. J.Radiat. Oncol. Biol. Phys. 8, 1225-1229 [1982]
I.A. Brezovich, M.D. Lilly. J.R. Divant and D.B.Richards, "A practical system for clinicalradiofrequency hyperthermia", Int. J. Radiat.Oncol. Biol. Phys. 7, 423-430 [1981]
J.W. Hand, J.L. Ledda and N.T.S. Evans,"Considerations of radiofrequency inductionheating for localized hyperthermia", Phys. Med.Biol. 27, 1-16 [1982]
J.R. Oleson, "Hyperthermia by magnetic induction: 1.Physical characteristics of the technique", Int.J. Radiat. Oncol. Biol. Phys. 8, 1747-1756 [1982]
P.M. Corry, J. Khalid, T.S. Kong, E.P. Armour, F.J.McGraw, T. Leduc, "Evaluation of equipment forhyperthermic treatment of cancer", Int. J.Hyperthermia 4, 53-74 [1988]
K. Hynynen, R. Roemer, D. Anhalt, C. Johnson, Z.X.Xu, W. Swindell and T. Cetas, "A scanned,focussed, multiple transducer ultrasonic systemfor localized hyperthermia treatments", Int. J.Hyperthermia 3, 21-35 [1987]
P.F. Turner, "Hyperthermia and inhomogeneous tissueeffects using an annular phased array", IEEETrans. Microwave Theory Tech. MIT 32, 874-882,[1984]
55
82.
83.
84.
85.
86.
87.
88.
89.
90.
D.A. Christensen and C.H. Durney, Hyperthermiaproduction for cancer therapy: a review offundamentals and methods", J. Microwave Power, 16,89-105 [1981]
J.W. Hand, A.J. Hind and J.L. Cheetham, "Multielementmicrowave array applicators for localizedhyperthermia", Strahlentherapie 161, 535, [1985]
G. Ter Haar and J.W. Hopewell, "The induction ofhyperthermia by ultrasound; its value andassociated problems: II Scanned-plane transducer",Phys. Med. Biol. 30, 1327-1333 [1985]
C.W. Song, J.G. Rhee, C.K.K. Lee, and S.H. Levitt,"Capacitive heating of phantom and human tumourswith an 8 MHz radiofrequency applicator (ThermotronRF-8]", Int. J. Radiat. Oncol. Biol. Phys. 12,365-372 [1986]
K.R. Foster, H.N. Kritikos and H.P. Schwan, "Effectof surface cooling and blood flow on the microwaveheating of tissue", IEEE Trans., Biomed.Engineering BME-25, 313-316 [1978]
K.D. Paulsen, J.W. Stohbehn and D.R. Lynch,"Comparative theorectical performance for twotypes of regional hyperthermia systems", Int. J.Radiat. Oncol. Biol. Phys. 11, 1659-1671 [1985]
D.W. Armitage, H.H. LeVeen and R. Pethig,"Radiofrequency induced hyperthermia - computersimulation of specific absorption ratedistributions using realistic anatomical models",Phys. Med. Biol. 28, 31-42 [1983]
R.B. Roemer, T.C. Cetas, J.R. Oleson, S. Halae andA.Y. Matloubich, "Comparative evaluation ofhyperthermia heating modalities. I. Numericalanalysis of thermal dosimetry bracketing cases",Radiat, Res. 100, 450-472, [1984]
S.T. Clegg and R.B. Roemer, "Estimation of threedimensional temperature fields from noisytemperature measurements during hyperthermia", inpreparation
56
91. P. Fessenden, D.S. Kapp and T.V. Samulski, "Clinicalmicrowave applicator design", ProceedingsAAPM/NAMG School on Physical Aspects ofHyperthermia, Durkam, N.C. [1987]
92. R.B. Roemer, K. Forsyth, J.R. Oleson, S.T. Clegg andD.A. Sim, "The effect of hydralazine dose on blood.perfusion changes during hyperthermia", Int. J.Hyperthermia, to be published.
93. M. Plancot, B. Prevost, M. Chive, J.J. Rabre, R.Ledee and G. Jiaun, "A new method for thermaldosimetry in microwave hyperthermia usingmicrowave radiometry for temperature control" Int.J. Hyperthermia 3, 9-20 [1987]
57