[CANCER RESEARCH 39, 3248-3253, August 1979]0008-5472/79/0039-0000$02.00
Time-Temperature Relationship in Hyperthermic Treatment of Malignantand Normal Tissue in Vivo1
Jens Overgaard2 and Herman D. Suit3
Edwin L. Steele Laboratory of Radiation Biology, Department of RadiationMassachusetts 02114
ABSTRACT
The effect of hyperthermia on normal and tumor tissue wasstudied following water bath heating of a methylcholanthreneinduced fibrosarcoma (FSaI) isotransplanted into the feet ofC3H mice. The time-temperature relation for the 50% tumorcontrol dose over the temperature range of 41 .5—45.5°showeda log linear relationship which followed a biphasically modifiedAmrheniusplot. At temperatures above 43°,there was a 50%reduction in heating time to obtain the 50% tumor control dosefor each 1°increase in temperature, corresponding to anactivation energy of 140 kcal/mol. At temperatures below 43°,the curve was steeper, with a tendency to double the treatmenttime for each 0.5°reduction in temperature (activation energy,approximately 230 kcal/mol).
Normal tissue damage in the tumor-bearing foot was estimated at two levels with a 50% response dose assay. Severenormal tissue damage showed a time-temperature relationshipsimilar to the tumor response, thus indicating no variation intherapeutic ratio at different temperatures. However, for slighttissue damage, the therapeutic ratio increased with decreasingtemperatures, yielding a better therapeutic ratio at lower temperatures.
The time-temperature relationship obtained in the FSal fibrosarcoma is supported by other studies and points to a generaltime-temperature relationship for hyperthermic tumor destruction.
INTRODUCTION
The ability of moderate heat treatment to eradicate expenimental tumors in vivo has been established for numerousanimal tumor systems (3, 4, 6, 7, 14, 15, 19, 20, 23—25,29,30).
There is an intense interest in applying this modality to thetreatment of human cancer, either alone or in combination withother modalities, e.g. , ionizing radiation (1, 2, 9, 10, 13, 19,28).
In planning preclinical investigations, we have been concenned with several questions, 2 of which are under currentstudy: (a) the time-temperature relationship required to obtainthe same tumor response and the time-temperature relationshiprequired to obtain specified levels of tissue damage; (b) todetermine if there is a temperature at which the therapeutic
1 Supported in part by Department of Health, Education and Welfare Grant
CAl 331 1 and the Danish Cancer Society.2 Present address: The Institute of Cancer Research, Radiumstationen, OK
8000 Aarhus C, Denmark. To whom requests for reprints should be addressed.3 Andres Soriano Director of Cancer Management, Massachusetts General
Hospital.
Received December 28, 1978; accepted April 26, 1979.
Medicine, Massachusetts General Hospital, Harvard Medical School, Boston,
effect is greater (e.g. , a high tumor effect and low damage tonormal tissue).
A few previous studies have partly focused on these problems. Westemmank(31) and, more recently, Ovengaard andOvergaard (2i , 23) studied the time-temperature relationshiprequired to obtain control in solid animal tumors. Overgaamdand Overgaard found that, at 41 .5—43°,a doubling of treatmenttime for each 0.5°reduction in treatment temperature is mequired for equivalent responses. In the Westermark study, thetreatment time could be reduced to one-half for each 1°increase in temperature between 44 and 48°.Both investigationsused low-frequency electromagnetic diathermy, a techniquewhich may give a somewhat heterogeneous heat distribution,especially in the tumor periphery, and the method used istherefore not optimal in order to obtain a quantitative animaltumor response (20). Furthermore, the technique by whichWestermark measured the tumor temperature is questionable(23, 31).
Cnile(3, 4), on the other hand, studied the effect of Sarcoma180 tumors inoculated into the feet of mice and treated by a
water bath. This technique achieved a uniform heating of thesmall tumor volumes. Crile observed a parallel response in thetumor and severe normal tissue damage, which was especiallyestablished at temperatures higher than 43°.He showed, likeWestermank, that at 43°and above a doubling of the treatmenttime was necessary for each 1°reduction in the temperature.At lower temperatures, the observations by Cnile were somewhat divergent. Because the number of animals in his studywas small, there is some uncertainty as to the quantitativeevaluation of the data. Furthermore, the Sarcoma 180 tumor ishighly antigenic, and spontaneous tumor regressions werefrequently observed.
The present experiments have been undertaken to study therelationship between temperature and time using 2 end pointsof tissue response: (a) permanent regression of a tumor growing in the foot pad; and (b) late damage to the tissue of thetumor-bearing foot. The temperatures studied covered therange 41.5—45.5°.
MATERIALS AND METHODS
Animal Tumor SystemMale and female C3Hf/Sed mice, 10 to 12 weeks old, from
our defined-flora and specific-pathogen-free colony were used(26). Cell suspension was prepared by a nonenzymatic technique from fourth- or fifth-generation isotmansplantsof FSaI, amethylcholanthrene-mnducedfibrosancoma (26). For transplant,2 to 5 x i o@viable tumor cells (trypan blue dye exclusion test)in a 2.5-j@lvolume were injected into the foot pad. The tumortransplant take rate was >95%.
0.5 Doubtful change from normal hair on foot. Less than 1- x 1-mm scarat tumor site.
1.0 Slight reddening, no hair on foot.1.5 Red foot and/or slight edema and/or atrophic hairless scar.2.0 Edema of whole foot and/or free tendons at site of scar.2.5 Fused toes and/or edema (more than 2.0).3.0 Severe edema and/or moist reaction in 1 spot.3.5 Moist reaction on entire foot.4.0 Loss of 1 toe.4.5 Loss of 2 or more toes.5.0 Loss of foot.6.0 Loss of foot through ankle.
Time-Temperature Relationship in Heat Treatment
Hyperthermic TreatmentTreatments were performed on tumors with a volume of
about 200 cu mm as determined from the 3 diameters usingthe formula Di x D2 x D3 x (ir/6). This treatment size wasnormally obtained about 10 days after challenge. Unanesthetized mice were fixed in a special jig by tape in such a way thatthe tumor-bearing leg could be immersed in a circulating waterbath with the tumor about 1 cm below the water surface.Special care was taken to avoid impairment of blood flow inthe limb. The water bath temperature was measured with athermometer calibrated against a standard thermometer certified by the National Bureau of Standards. Temperature washeld to within ±0.05°of the desired level by a thermostatcontrol system. In the initial experiments, the intratumoral ternperature was measured by a 25-gauge needle probe (YSIModel 46 Tele Thermometer; Yellow Spring Instrument Cornpany, Yellow Spring, Ohio). Intratumoral temperature stabilizedat 0.2°below water bath temperature with 2 mm of immersion.All temperatures mentioned in this paper refer to water bathtemperatures. Animals could be kept in the treatment positionfor up to 12 hr without major problems.
Evaluation of Results
Tumor Evaluation. The micewere followedtwice weekly forthe first 30 days and then with weekly intervals up to I 20 daysafter treatment. Animals with tumors that were not controlledby the treatment were sacrificed when their tumor reached2000 Cu mm. At completion of an experiment, local controlresults were tabulated and the TCD504values were calculated(27). Excluded from the analysis were animals which diedtumor free during the 120-day period. Only tumor-free survivinganimals were counted as cured. Failure meant local negrowthof tumors, plus a few mice without tumor in the foot, but withnegnowthoutside the treated area (e.g. , in the hip). Mice whichhad a partial on total leg amputation following high-dose treatment were scored as local controls.
Normal Tissue Evaluation. The status of the normal tissuein the tumor-bearing foot was scored together with the tumorresponse. Because the acute damage could be modified by thepresence of tumor, late reactions at 120 days were recordedand calculated. The scoring protocol is presented in Table 1.The late reactions of tumor-bearing feet are described in termsof RD50values (computed in the same manner as TCD50)for 2levels of damage. The therapeutic ratio is defined in this reportas RDso:TCDso.
RESULTS
Effect on Tumor Tissue
Tumor Reaction. Generally, the initialgross responsewasindependent of the treatment time or temperature used in thisstudy.
Immediately after treatment, there was slight cyanosis andprominent edema. By 24 to 48 hr, the overlying skin in mosttumors was dark blue or black. By that time, the edema haddecreased, and the skin overlying the tumors had turned into
4 The abbreviations used are: TCD@, time at hyperthermia that achieves
control of one-half of the treated tumors; RD@,time at hyperthermia that elicitsa specific level of normal tissue reaction in one-half of the treated feet.
Table 1
Score for heat-induced normal tissue damage to the tumor-bearingfoot
a dry crust, which usually persisted for about 1 week. Thecrust would normally fall off spontaneously, leaving a palehairless scar approximately 2 mm in diameter providing thatcomplete regression was achieved. In some cases, there wasa defect in the skin, and the underlying tendons were exposed.In most cases, the foot healed with a normal skin covering,which included negrowth of hair.
In almost all tumors which ultimately were not controlled,there was grossly evident tumor at the time of detachment ofthe crust. This recurrence grew with approximately the samedoubling time as that of the untreated tumor. In tumors treatedwith very low heat doses, only a 2- to 3-day inhibition of growthwas noted followed by normal growth.
Typically, the control or failure could be estimated within afew weeks. On only very few occasions (<2%) was a tumorfound to recur locally after total disappearance. The longesttime from treatment to definitive disclosure of a recurrent tumorwas about 5 weeks (Chart 1).
Dose Response. Table 2 shows the treatment time acquiredat a given temperature to achieve 50% local control in a FSaItumor. The dose-response curves for tumor control at 41.5,42.5, 43.5, 44.5, and 45.5°are presented in Chart 2. Thesecurves have almost parallel slopes on the logit-log dose gridused in Chart 2, suggesting that there is no temperaturedependent difference in the dose response in the range between 41 .5 and 45.5°.
Time-Temperature Relation. The relationshipbetween timeand temperature required to obtain equivalent tumor control(TCD50)is shown in Chart 3, which demonstrates a modifiedArrhenius plot of the TCD50obtained at different temperaturesbetween 41 .5 and 45.5°.The time-temperature relationshipfollowed a variable pattern. Above 43°,a 1°reduction in thetreatment temperature will result in a doubling of treatment timerequired to obtain an equivalent therapeutic effect. However,below 43°,the curve is steeper, with a tendency to maximallydouble the treatment time required for an equivalent effect foreach 0.5°reduction in treatment temperature.
Effect on Normal Tissue
Foot Reaction. The acute normal tissue reaction in the footincreased with temperature or treatment time. In treatments atlow temperatures (4i .5°),themewas no noticeable change inthe non-tumor-bearing part of the foot immediately after treatment. At higher temperatures, the acute changes were mainlyin the form of edema which was in some instances severe andlater developed into necrosis of the peripheral part of the foot.
This occasionally could result in amputation of toes or in partialor complete loss of the foot. The progression of the normaltissue damage reached a maximum within 2 to 3 weeks aftertreatment; when amputation occurred, it was during that time.
100 %
90
80
70
60
50
40
30
20
10
0
DAYS AFTER TREATMENT
Chart 1. Time to clearance of tumor or to local failure following single hyperthermic treatment at 41 .5—45.5°.A definitive estimate of tumor response wasobtained within 40 days after treatment. No additional changes occurred in theremainder of the 120-day observation period.
After the acute damage, there was normally a moderate repairor reconstruction before more chronic conditions were established about 3 to 4 weeks after treatment. In general, a correlation between acute and chronic damage was noted eventhough the acute reaction was often difficult to evaluate because the tumor response itself interfered with observations ofthe damage to normal tissues of the foot. No significant changein the normal tissue reaction occurred beyond i month following treatment (Chart 4).
Dose Response. The RD50at a given temperaturebetween42°and 45.5°is shown in Table 2. Level 1 response was thelowest degree of definitive tissue change. This was mainlycosmetic and did not produce functional defects. On the otherhand, the damage at Level 4 was severe, was irreversible, andinvolved pronounced impairment of the function of the limb andloss of at least one toe. Dose-response curves for Levels I and4 were of a common slope for all temperatures tested. Thisslope was approximately the same as found for the tumorresponse curve (Chart 2). However, a RD50value for the Level4 foot damage could not be obtained at 4i .5 and 42°at themaximum exposure times which were feasible in our expenimental set-up.
Time-Temperature Relation. Chart 5 shows modified Amrhenius plots of the time-temperature relationship of heat treat
4 0 ment causing damage to the mouse foot at 2 specific levels.
For Level 4 damage, themeis a log linear time-temperaturerelationship in the temperature interval 42.5—44.5°with anactivation energy of 140 kcal/mol.
The Level 1 damage has a steeper time-temperature curvewith an activation energy of about 200 kcal/mol at 42—44.5°.
It:@ 60
:73k. 5Q14J
@ 40
@ 30
20
to
10 20 30
90 -
80 -
70 -
Chart 2. Dose-response curves to obtain tumor control in the FSalfibrosarcoma in vivo. An approximately parallel response is found at41 .5—45.5°.Bars, 95% confIdence limit.
Chart 3. Time-temperature relationship to obtain a hyperthermic isoeffectaCD@@@)in the FSal fibresarcoma in tnvo. For temperatures higher than 43°,thismodified Arrhenius plot has an activation energy of I 40 kcal/mol. For lowertemperatures, the activation energy increased towards approximately 230 kcal/mel. Bars, 95% confidence limit.
____________________ J42° 43― 440 450
42° 43° 44° 45°
TEMPERATI/REChart 5. Time-temperature relationship to hyperthermic isoeffect (RD@)at 2
different levels of normal tissue damage. The modified Arrhenius curves show alog-linear relationship for Level 1 damage between 42 and 45.5° with anactivation energy of 200 kcal/mei. For the Level 4 damage estimated in thetemperature range 42.5 to 45.5°,the activation energy is 149 kcal/mol.
— level 4
level 1
.— —:::: 425°
00LOInc@c3
1.@
k.:73
Lu
1000
C,,14j
:3
I'..
0U,
43 5°
100 . °-
30 120
DAYS AFTER TREATMENTChart 4. Relationship between early and late damage to the tumor-bearing
feet after local hyperthermia at 42.5, 43.5, or 44.5°.Ne significant variationswere observed between damages recorded 30 or 120 days after treatment.
Therapeutic Ratio
Chart 6 shows the therapeutic ratios (RDso:TCDse)for singletreatments at 42—45.5°.For severe chronic damage (Level 4),the therapeutic ratio is constant at 1.5 for treatment temperatunes @42.5°,but for the Level 1 response the therapeutic ratiodecreased from a value of @1.75 at 42°down to 0.75 at 45.5°.
DISCUSSION
The present findings showed that hyperthermic treatment at
44@5 0
TEMPERATURE10 Chart 6. Therapeutic ratio for 2 specific levels of normal tissue damage. The
therapeutic ratio for the mild skin damage [RD,@(Level 1):TCD@]decreased withincreasing temperatures at 42—45.5°.However, for the severe Level 4 damage,a constant therapeutic ratio of 1.5 is observed at 42.5—45.5°.
41 .5—44.5°was able to locally eradicate 50% of the FSaIfibrosarcomas growing in the C3H mouse foot without causingmajor change in the normal tissue.
The relationship between temperature and time for the TCD50was, fontemperatures @43°,a decrease in treatment time bya factor of 2 for each degree increase in temperature. Thiscorresponds to an activation energy of approximately 140kcal/mol. For temperatures lower than 43°, the activationenergy appears to increase with decreasing temperature up toabout 230 kcal/mol; this corresponds approximately to a doubling of the treatment time for each 0.5°reduction in temperature.
(20, 23)Present studyRobinson et a!.(24)sarcoma44.0—48.0
139Westermark (31)
J. Overgaard and H. D. Suit
Table3
a Calculated by least-square fits of published data.b Based on TCDSO or similar dose-response data.
The time-temperature relationship observed at temperaturesabove 43°is in agreement with other studies of experimentaltumors heated in vivo, both by a water bath (3, 4, 11, 24, 25)and by diathermy (20, 31). Also the tendency to an increasedactivation energy observed at temperatures below 42.5—43.5°has previously been observed in other studies (3, 20, 23, 24)(Table 3). These data point towards a biphasic time-temperatune relationship for tumor control in vivo.
Since a similar biphasic pattern with approximately the sameactivation energies has also been obtained from in vitro survivalcurves from different mammalian cell lines (2, 5, 13), a generaltime-temperature relationship to hyperthermic destruction issuggested by these studies.
Our study showed that in the range of 42.5—45.5°the timeto obtain severe normal tissue damage (Level 4) was parallel tothe time to obtain tumor control, namely, no variation in therapeutic effect at temperatures 42.5°(Chart 6; Table 2). Atlower temperatures, the treatment time was not sufficiently longto obtain RD50values for Level 4 damage since very few miceshowed severe reaction. Our impression, not an experimentalfact, is that the normal tissue damage was relatively less (i.e.,increasing therapeutic gain) at lower temperatures.
An increasing therapeutic gain at low temperatures is supported by the observation that the RD50values for Level 1damage showed a steeper time-temperature relation than didthe TCD50and the RD50(Level 4) responses. This results in ahigher therapeutic ratio at a low temperature if the ratio isbased on the mild Level 1 damage (Chart 6; Table 2).
An activation energy of about 140 kcal/mol for severe tissuedamage has been observed in a number of different tissuesheated to temperatures above 42°(3, 4, 8, 12, 16-1 8). [Forfurther discussion, see the comprehensive review by Field(8)]. Accordingly, the time-temperature relationship appears tobe similar for tumor control and severe normal tissue damage,at least for temperatures above 42.5—43°.
The higher activation energy observed for the mild skindamage may indicate that the 2 levels of damage are causedby different mechanisms or targets. However, the significanceof these findings must still be understood.
The time-temperature relationship to achieve TCD50valuesfor heat treatment of the FSaI tumor was found to follow a loglinear relationship which at temperatures @43°corresponds toa doubling of the treatment time for each degree of reductionin the temperature. At temperatures below 43°,the relationship
was steeper with a tendency to treatment time doubling for0.5°reduction in treatment temperature.
For mild degrees of skin damage, the time-temperature melationship was steeper, resulting in relatively less tissue damage at lower temperatures.
However, at none of the temperatures where severe normaltissue damage could be produced (42.5—44.5°)was anychange in the therapeutic ratio observed. This observation is inagreement with other reports on the time-temperature relationships.
REFERENCES
1. Cavaliere, R., Ciocatto, E. C., Giovanella, B. C., Heidelberger, C., Johnson,R. 0., Margottini, M. , Mondovi, B. Moricca, G., and Ressi-Fanelli, A. Selective heat sensitivity of cancer cells. Biochemical and clinical studies. Cancer,20: 1351-1381. 1967.
2. Connor, W. G., Garner, E. W., Miller, R. C., and Boone, M. L. M. Prospectsfor hyperthermia in human cancer therapy. Radiology, 123: 497—503,1977.
3. Crile, G. Heat as an adjunct to the treatment of cancer. Cleve. Clin. 0. . 28:75-89, 1961.
4. Crile, G. The effects of heat and radiation on cancers implanted on the feetof mice. Cancer Res., 23: 372—380,1963.
5. Dewey, W. C., Hopwood, L. E., Sapareto, S. A., and Gerweck, L. E. Cellularresponses to combinations of hyperthermia and radiation. Radiology, 123:463-474, 1977.
6. Dickson, J. A., and Muckle, 0. C. Total-body hyperthermia versus primarytumor hyperthermia in the treatment of the rabbit VX-2 carcinoma. CancerRes., 32: 1916—1923,1972.
7. Dickson, J. A., Shah, S. A., Waggott, D. . and Whalley, W. B. Tumoreradication in the rabbit by radiofrequency heating. Cancer Res., 37: 2162—2169, 1977.
8. Field, S. B. The response of normal tissues to hyperthermia alone or incombination with x-rays. In: C. Streffer (ed), Cancer Therapy by Hyperthermia and Radiation. Proceedings of the 2nd International Symposium.Essen 1977. pp. 37—48.Munich: Urban and Schwarzenberg, 1978.
9. Har-Kedar, I., and Bleehen, N. M. Experimental and clinical aspects ofhyperthermia applied to the treatment of cancer with special reference tothe role of ultrasonic and microwave heating. Adv. Radiat. Biol. , 6: 229—266, 1976.
10. Hyperthermia in the treatment of the cancer patient. Cancer (Phila.), 37:2075-2083, 1976.
11. Johnson, H. J. The action of short radio waves on tissues. Ill. A comparisonof the thermal sensitivities of transplantable tumours in vivo and in vitro. Am.J. Cancer, 38: 533-550, 1930.
12. Law, M. P., Ahier, R. G.. and Field, S. B. The response of the mouse ear toheat applied alone or combined with x-rays. Br. J. Radiol., 51: 132—138,1978.
13. Leith, J. T., Miller, R. C., Garner, E. W., and Boone, M. L. M. Hyperthermicpotentiation. Biological aspects and applications to radiation therapy. Cancar (Phila.), 39: 766—779,1977.
14. Marmor, J. B., Hahn, N., and Hahn, G. M. Tumor cure and cell survival afterlocalized radiofrequency heating. Cancer Res., 37: 879—883,1977.
15. Mendecki, J., Friedenthal, E., and Botstein, C. Effects of microwave-inducedlocal hyperthermia on mammaryadenocarcinoma in C3H mice. Cancer Res..36:2113-2114,1976.
16. Meritz, A. R., and Henriques, F. C. Studies of thermal injury. II. The relativeimportance of time and surface temperature in the causation of cutaneousburns. Am. J. Pathol., 23: 695-720, 1947.
17. Morris, C. C.. Myers, R., and Field, S. B. The response of the rat tail tohyperthermia. Br. J. Radiel., 50: 576-580, 1977.
18. Okumura, Y., and Reinhold, H. S. Heat sensitivity of rat skin. Eur. J. Cancer,14:1161—1166,1978.
19. Overgaard, J. The effect of local hyperthermia alone, and in combinationwith radiation, on solid tumors. In: C. Streffer (ad.), Cancer Therapy byHyperthermia and Radiation. Proceedings of the 2nd International Symposium, Essen 1977, pp. 49—61. Munich: Urban and Schwarzenberg, 1978.
20. Overgaard, J. Biological effect of 27. 12 MHz short-wave diathermic heatingin experimental tumors. IEEE Trans. Microwave Theory Tech.. MTT-26:523—529.1978.
22. Overgaard, K., and Okkels, H. Uber den Einfluss der WÃ rmebehandlungaufWoods Sarkem. Strahlentherapie, 68: 587-61 9, 1940.
23. Overgaard, K., and Overgaard, J. Investigations on the possibility of athermic tumour therapy—I. Short-wave treatment of a transplanted isologous mouse mammary carcinoma. Eur. J. Cancer, 8: 65—78,1972.
24. RobInson, J. E., McCready, W. A., and Slawson, R. G. Thermal sensitivities
of mouse mammary tumors. In: C. Streffer (ed), Cancer Therapy by Hyperthermia and Radiation. Proceedings of the 2nd International Symposium,Essen 1977, pp. 242—244.Munich: Urban and Schwarzenberg, 1978.
25. Suit, H. D. Hyperthermic effects on animal tissues. Radiology, 123: 483-487, 1977.
26. Suit, H. D., Sedlacek, R., Wagner, M., Orsi, L., Siiobrclc, V., and Rothman,K. J. Effect of Corynebacterium parvum on the response to irradiation of aC3H fibrosarcoma. Cancer Res., 36: 1305-1314, 1976.
27. Suit, H. D., Shalek, R. J., and Wette, R. Radiation response of C3H mousemammary carcinoma evaluated in terms of cellular radiation sensitivity. In:Cellular Radiation Biology, M. D. Anderson Hospital and Tumor Institute atHouston, pp. 514—530.Baltimore: Williams & Wilkins, 1965.
28. Suit, H. D., and Shwayer, M. Hyperthermia: potential as an anti.tumor agent.Cancer (Phila.), 34: 122—129, 1974.
29. Thrall, D. E., Gerweck, L. E., Gillette, E. L., and Dewey, W. C. Response ofcells in vitro and tissues in vivo to hyperthermia and x-irradlation. Adv.Radlat. Blol., 6: 21 1—227,1976.
30. Thrall, D. E., Gillette, E. L., and Bauman, C. L. Effect of heat en the C3Hmouse mammary adenocarcinoma evaluated in terms of tumor growth. Eur.J. Cancer, 9: 871 -875, 1973.
31 . Westermark, N. The effect of heat upon rat tumors. Scand. Arch. Physiol.52: 257-322, 1927.
