Comparison of High Dose Rate, Low DoseRate, and High Dose Rate FractionatedRadiation for Optimizing Differences in

Radiosensitivities In Vitro

Ruth C. Wilkins, Ph.D.,1* C.E. Ng, Ph.D.,2 and G.P. Raaphorst, Ph.D.21Health Canada, Radiation Protection Bureau, Ottawa, Ontario, Canada

2Ottawa Regional Cancer Centre, Ottawa, Ontario, Canada

SUMMARY Radiotherapy is administered with the assumption that all patients respondsimilarly to radiation although radiosensitivity does vary from patient to patient, resultingin different degrees of early and late effects. Because the dose given to a patient is limitedby the response of normal tissue in the treatment field, it would be beneficial to determinethe sensitivity of this normal tissue prior to therapy. Previous studies to predict radiosen-sitivity have used surviving fractions after a single dose given in vitro, however, differencesin cell survival at this low level of kill are not easy to resolve. In this study, we set out toevaluate the use of alternative dose regimens which may better resolve differences inradiosensitivity. We have examined several radiation protocols for predictive value, in-cluding survival after high doses (6 Gy) at both high (112 cGy/min) and low (.882 cGy/min)dose rates and after fractionated doses of 2 Gy (6 fractions). A sensitive human fibroblastline (S11358) cultured from a patient showing severe effects after therapy is comparedwith a cell line (OMB1) cultured from an apparently normal subject. Differences betweenthese cell lines have been compared with those between two human melanoma cell lines(SKMEL3 and HT144) which have shown resistant and sensitive response to radiation invitro respectively. In both fibroblast and melanoma cell lines, the difference in the survivalof normal and sensitive cells increased with increasing dose regardless of whether irra-diation was delivered as low dose rate, high dose rate, or as fractionated doses. We proposethat radiation doses which more closely mimic clinical treatment are more suitable thansurviving fraction after 2 Gy (SF2) for in vitro evaluation of relative radiosensitivities of cellpopulations. Radiat. Oncol. Invest. 6:209–215, 1998. © 1998 Wiley-Liss, Inc.

Key words: predictive assay; fibroblasts; fractionated radiation; intrinsicradiosensitivity

INTRODUCTIONIn radiotherapy, the dose-limiting factor is oftendamage to the normal tissue in the treatment field.Because there is considerable variation in normaltissue response between patients, the maximumdose is usually calculated from the tolerance of the

small percentage of sensitive patients. Therefore, ifthese sensitive patients could be identified prior totreatment, the dose given to the remaining patientscould be increased. This increase in dose could po-tentially improve local control and cure to the non-sensitive patients by up to 20% [1].

Work Performed at Ottawa Regional Cancer Centre, Ottawa, Ontario

*Correspondence to: Ruth Wilkins, Ph.D., Health Canada, Radiation Protection Bureau, 775 Brookfield Rd., PostalLocator 6303B, Ottawa, Ontario, K1A 1C1 Canada. Phone: (613) 941-7263; Fax: (613) 941-1734; E-mail: RuthWilkins@hc-sc.gc.ca

Received 2 February 1998; Revised 8 April 1998; Accepted 26 June 1998

© 1998 Wiley-Liss, Inc.

Radiation Oncology Investigations 6:209–215 (1998)

There appears to be a genetic component toradiosensitivity, and in certain inherited syn-dromes, this genetic component dominates theoverall response to radiotherapy [2], as shown inpatients with ataxia telangiectasia (AT), which ischaracterized by hypersensitivity to ionizing radia-tion [3,4]. Skin fibroblasts and other cell types fromthese patients have also shown this sensitive re-sponse in vitro [5].

In radiotherapy patients with no known ge-netic syndromes, there is clinical evidence of largeindividual differences in the radiosensitivity of nor-mal tissue [6]. If this variation also has a geneticcomponent, normal human fibroblasts should pro-vide a system to detect those individual differences.

Several studies claim that differences in radio-sensitivity can be measured in vitro [7]. It has beenshown that the range of radiosensitivities in severalfibroblast and lymphocyte cultures is larger thancan be explained by sampling or methodology er-rors [8]. The level of radiosensitivity has also beencorrelated to the degree of the patients’ reactionafter treatment [9–11]. In order to measure radio-sensitivity, several endpoints have been used, in-cluding SF2 and a [10], SF3.5 [12], D0 and D10

[11], and D1 [9], which have been measured afterboth high dose rate (HDR) and low dose rate(LDR) irradiation. However, there has been muchdebate whether the SF2 model, or models based onlow levels of survival, are relevant predictors ofradiosensitivity [13]. The wide distribution of SF2

values for cells grown in culture from normal tissuehas been correlated with late reactions to radio-therapy [9,14,15], although a correlation to earlyreactions is still controversial [10,11,16].

Our aim is to investigate whether differencesin survival levels measured in human melanomacell lines and normal fibroblasts would be moreresolvable after fractionated radiation comparedwith either HDR or LDR irradiation. This protocolmimics the external beam therapy, which can beresponsible for both early and late reactions.

METHODS

Cell Culture

Melanoma Cell LinesSKMEL3 is an adherent, fibroblast-like, pig-mented, radioresistant melanoma cell line estab-lished from a 41-year-old female (ATCC). HT144is a nonpigmented, radiosensitive melanoma cellline established from a 29-year-old male (ATCC).Cells were cultured in a mixture of DMEM/F-12(Gibco) containing 7.5% fetal bovine serum (FBS)and 7.5% newborn calf serum (NCS) (Gibco),

0.1 mM nonessential amino acids (Gibco), 20 mMHepes (Boehringer Mannheim), and 10 mMNaHCO3 (Sigma) in a humidified atmosphere of98% air, 2% CO2 at 37°C. Cell lines were testedregularly for mycoplasma contamination.

All experiments were carried out with conflu-ent cultures. For the HDR and fractionated irradia-tion, confluence was obtained by inoculating 5 ×105 cells into 4 mL of medium in a 25 cm2 flask onday 0, feeding the cells with fresh medium on day5, and using the cells for experiments on day 8. Thecells were in stable plateau-phase on the day of theexperiment as shown by their growth curves [17].For the LDR irradiations, 5 × 104 cells were platedin 1.3 cm2 glass vials on day 0, fed on day 5, andused for experiments on day 8.

Primary CultureThe OMB1 cell line was grown from a 4 mmpunchbiopsy obtained from the forearm of a 46-year-oldmale volunteer who gave informed consent afterthe nature of the procedure was fully explained.The tissue was transported in saline and imme-diately soaked for two 10 min periods in me-dium containing high concentrations of antibiotics(1000 IU/mL penicillin G sodium, 1000mg/mLstreptomycin sulphate, 2.5mg/mL amphotericin B)but without serum. The tissue was transferred to apetri dish and minced into small pieces. The pieceswere pressed to the bottoms of four 25 cm2 flasks(4–5 pieces per flask) containing 2 mL of mediumwith 20% FBS and 1 × antibiotics (100 IU/mLpenicillin G sodium, 100mg/mL streptomycin sul-phate, 0.25mg/mL amphotericin B). The cultureswere observed daily until cell growth could be seenat which time the volume of medium was increasedto 4 mL. After the cultures were established, theserum levels were changed to 7.5% FBS and 7.5%NCS with no subsequent change in the growthcharacteristics. All experiments were done on cellswith passage numbers less than 10. The cultureswere fed every 5 days until there was enough cellgrowth to trypsinize and culture. All experimentswere carried out with confluent culture. For HDRirradiations, confluence was obtained by inoculat-ing 1 × 105 cells into a 25 cm2 flask in 4 mL ofmedium on day 0, feeding the cells with fresh me-dium on day 5, and using the cells for experimentson day 10. The cells were in stable plateau-phaseon the day of the experiment as shown by theirgrowth curves in Figure 1. For LDR irradiations, 5× 103 cells were plated on day 0, fed on day 5, andused for experiment on day 10.

The S11358 cell line was grown from a biopsyfrom a 33-year-old female patient with cervical

210 Wilkins et al.: Optimizing Differences in Radiosensitivity

cancer who had severe chronic normal tissue injuryfollowing conventional irradiation [18]. The cells,already in culture, were obtained from the AtomicEnergy of Canada Limited, Chalk River. The cellswere maintained and experimented upon under thesame conditions as the OMB1 cell line.

Irradiation

Cells were irradiated at a HDR using x-rays (150kVp, 1 mm Al filter) at a dose rate of 112 cGy/min.HDR irradiations were carried out at 4°C in orderto minimize the amount of repair occurring duringthe irradiation procedure [17]. For LDR irradia-tions, cells were irradiated using226Ra sources at adose rate of 0.88 cGy/min. Cells were kept at 37°Cthroughout the irradiation. Fractionated irradiationswere performed using x-rays, as above, with 2 Gygiven three times daily with a minimum of 6 hrbetween irradiations. Six fractions were adminis-tered for a total of 12 Gy. Cells were held at 4°Cduring the irradiation and incubated at 37°C be-tween irradiations.

Clonogenic Cell Survival

Cell survival was measured by the colony formingassay. Confluent cells were trypsinized and platedat the appropriate number in 60 mm or 100 mmtissue culture dishes to achieve approximately 50colonies per dish. Dishes were incubated at 37°Cfor 11–14 days, stained, and colonies containingmore than 50 cells were counted. Each experimentwas repeated three times with the error bars repre-senting the standard error of the mean (S.E.M.).

Method of Fitting Survival Curves

The survival curves were fitted using the linearquadratic model, S4 exp(-aD-bD2) by taking thenatural logarithm of the equation and fitting a sec-ond order polynomial to the equation lnS-4 -aD-bD2. The fits were performed using a graphics pro-gram Grapher (Golden Software).

Flow Cytometry

The cell cycle distribution was examined duringfractionated radiation using propidium iodide stain-ing on a Coulter Epics XL flow cytometer. The cellcycle distribution was analyzed using the Multi-cycle program (Phoenix Flow Systems).

RESULTS

The two melanoma cell lines, SKMEL3 andHT144, have been shown to be radioresistant andradiosensitive respectively [17,19]. Figures 2 and 3show their survival curves after HDR, LDR, andfractionated irradiation. The large difference intheir radiation responses is evident, even after lowdoses. These differences have been quantified inTable 1, which shows the surviving fractions of theresistant and sensitive cells after different doses ofeach radiation protocol. These values were takenfrom the fitted curves. TheP value from the stu-dentst-test indicates the significance of the differ-ences between the cell lines. Table 1 shows thedifferences in survival of the SKMEL3 and HT144

Fig. 2. Survival curves of melanoma cells lines after highdose rate (HDR) and low dose rate (LDR) irradiation. Thedata were fitted to the linear quadratic model. The error barsrepresent the S.E.M. of three independent experiments.

Fig. 1. Growth curve of OMB1 and S11358 cells fed onday 5. The initial concentration of cells was 4 × 103 cells/cm2.

Wilkins et al.: Optimizing Differences in Radiosensitivity 211

cells lines to be highly significant (P < .001) at alllevels for all three radiation protocols. The in-creased survival due to fractionated radiation, how-ever, allows survival to be measured after higherdoses (8 Gy) of radiation in which the differencebetween the two cell lines becomes even more sig-nificant (P < .0001).

This method was then used to evaluate thedifferences in radiosensitivity between the two hu-man fibroblast lines cultured from skin biopsies.These two cell lines, OMB1 and S11358 came froma normal subject and radiosensitive patient respec-tively. Their radiation survival curves are shown inFigures 4 and 5. The differences in their radiosen-sitivities in vitro are much smaller than with themelanoma cell lines; however, the same trend ap-pears. TheP value indicates that the difference be-tween the survival of the two cell lines is not sig-nificant for any radiation protocol up to 6 Gy. Frac-tionation, in this case, allowed survival to bemeasured after 12 Gy, at which time theP valuedecreased below 0.05, indicating a significant dif-ference between survival.

Figures 6 and 7 show that the cell cycle dis-tributions of the melanoma and fibroblast cellsthroughout the fractionation experiment changedvery little.

DISCUSSION

Much work has been done on tumor tissue grown invitro as a predictor of radiosensitivity and tumor T

able

1.Su

rviv

ing

Frac

tion

sA

fter

Var

ious

Dos

esof

Rad

iati

on

Cel

llin

e

LDR

HD

R2

Gy/

frac

tion

2G

y6

Gy

2G

y6

Gy

2G

y6

Gy

maxa

SK

ME

L30.

71(0

.03)b

0.35

(0.0

4)0.

93(0

.16)

0.31

(0.0

2)0.

93(0

.07)

0.59

(0.0

4)0.

41(0

.01)

HT

144

0.11

7(0

.007

)8.

6×

10−4

(4×

10−

5)

0.03

4(0

.000

4)3.

9×

10−5

(5×

10−

6)

0.03

4(0

.002

)3.

8×

10−4

(6×

10−

5)

1.1

×10

−5

(5×

10−

6)

Pva

lue

<.0

01<

.001

.005

<.0

01<

.001

<.0

01<

.000

1

OM

B1

0.40

(0.0

2)0.

11(0

.03)

0.37

(0.0

9)0.

025

(0.0

05)

0.55

(0.0

5)0.

22(0

.03)

0.11

(0.0

2)S

1135

80.

422

(0.0

01)

0.10

(0.0

2)0.

37(0

.03)

0.01

3(0

.003

)0.

53(0

.03)

0.15

(0.0

1)0.

026

(0.0

04)

Pva

lue

.4.8

.9.0

7.8

.1.0

192

LDR

4lo

wdo

sera

te.

HD

R4

high

dose

rate

.a M

axim

umdo

se:

8G

yfo

rm

elan

oma

cell

lines

,12

Gy

for

fibro

blas

tce

lllin

es.

bV

alue

sin

pare

nthe

ses

are

the

S.E

.M.

with

each

expe

rimen

tre

peat

eda

min

imum

ofth

ree

times

.

Fig. 3. Survival curves of melanoma cells lines after highdose rate (HDR) and fractionated radiation. The data werefitted to the linear quadratic model. The error bars representthe S.E.M. of three independent experiments.

212 Wilkins et al.: Optimizing Differences in Radiosensitivity

response [14,20–24]. Some trends have been seenbetween tumor sensitivity in vitro and patient re-sponse to radiotherapy. It is, however, the responseof the normal tissue that limits the dose adminis-tered to the patient. The goal is to give the tumorthe highest dose possible without damaging normaltissue in the treatment field, so it is normal tissue

sensitivity that is required for treatment planning.The tumor is also a heterogeneous environmentwith many factors that might affect the tumor re-sponse. This makes it difficult to choose one factoron which to base a predictive assay for intrinsicradiosensitivity of the tumor [25,26].

The first thing to notice from these results isthat, with the fibroblast lines, there is no differencein the surviving fraction after 2 Gy of radiationwith either HDR or LDR. Thus we were not able toseparate these two cell lines based on that singlemeasurement. This is not surprising because no re-pair was allowed to occur before plating the cellsafter a single HDR dose.

The best method for distinguishing betweenthe two melanoma cell lines is to irradiate with highdoses of fractionated irradiation, allowing repair tooccur between doses. Although the difference be-tween the fractionated survival of the two cell linesis no more significant after the same dose of HDRradiation, increased survival allows higher doses offractionated radiation to be used. For the two mela-noma cell lines, it was expected that the differencein SF after LDR irradiation would be greater, be-cause repair would also occur during irradiation.

Fig. 4. Survival curves of fibroblasts after high dose rate(HDR) and low dose rate (LDR) irradiation. The data werefitted to the linear quadratic model. The error bars representthe standard error of the mean (S.E.M.) of three independentexperiments.

Fig. 5. Survival curves of fibroblasts after high dose rate(HDR) and fractionated irradiation. The data were fitted tothe linear quadratic model. The error bars represent theS.E.M. of three independent experiments.

Fig. 6. Cell cycle analysis of melanoma cell lines afterfractionated radiation. The percentage of cells in each of thecell phases G1, S, and G2/M are shown. The data are fromone representative experiment.

Wilkins et al.: Optimizing Differences in Radiosensitivity 213

However, this did not appear to be the case (Table1). In our single-dose experiments, cells wereplated immediately after irradiation, which pre-vented repair from occurring before plating. Oneexplanation for a smaller separation after LDR isthat HDR fractionated irradiations induced a repairmechanism in the SKMEL3 line that was not in-duced with continuous LDR [27]. This observationwill be further investigated. Although fractionatedtherapy is no better at predicting the differences inradiosensitivity after the same doses are adminis-tered, it may become a better predictor of radiosen-sitivity if higher doses can be used.

The same results were seen with the fibroblastcells lines, although at a much lower level of sepa-ration. The fractionated radiation showed the great-est separation, again at higher doses than could bemeasured with HDR.

The survival curves after fractionated radia-tion were upward bending, which would suggest achange in the radiosensitivity. Flow cytometry ex-periments were conducted to address the possibilityof changes in the cell cycle distribution. These re-sults indicate little or no change in cell cycle dis-tribution, however, this method does not discrimi-nate between live, dying, or dead cells. Therefore,a small amount of cell proliferation may be occur-ring, and there could be some decrease in the per-centage of live sensitive cells. These results simplyindicate that there was no large shift in the cell

cycle distribution and no accumulation in the G2

phase of cell cycle as would be expected in expo-nentially growing cells after irradiation. The up-ward bending could therefore be explained bysmall amounts of repopulation or repair occurringbetween fractions, increased resistance being in-duced by previous doses, or by preferential killingof a subpopulation of sensitive cells resulting in amore resistant subpopulation. Upward bending inthe LDR experiment using OMB1 cells may also bedue to some of the same effects.

Although previous results have shown corre-lations between patient response and survival fol-lowing single acute HDR irradiation [9,12], theseresults show that, when trying to detect small dif-ferences in radiosensitivity, survival after a single2 Gy dose cannot predict any differences. The dif-ferences that will be seen in the patients will notappear until much higher doses have been admin-istered using fractionated irradiations. Using frac-tionated irradiation as a predictor allows higherdoses to be given to the cells and allows repairbetween doses.

In conclusion, to measure differences in radio-sensitivity, two factors seem to be important. Thefirst is that a radiation protocol that allows repair tooccur may be essential because repair may play apart in radiosensitivity. The second is that givingacute doses could induce an increased repairmechanism that does not occur after LDR. In orderto be able to include these two factors in a predic-tive assay, a reasonable choice for radiation proto-col would be a fractionated radiation scheme thatalso has the advantage of mimicking the protocolsused in the clinic.

REFERENCES

1. Nyman J, Burnet NG, Turesson I, Peacock JH: Corre-lation betweenin vitro radiosensitivity and normal tis-sue response. Part I—normal tissue response. Br J Can-cer 66:6–6, 1992.

2. Bentzen SM, Overgaard M, Overgaard J: Clinical cor-relations between late normal tissue endpoints after ra-diotherapy: implications for predictive assays of radio-sensitivity. Eur J Cancer 29A:1373–1376, 1993.

3. Murnane JP, Kapp LN: A critical look at the associationof human genetic syndromes with sensitivity to ioniz-ing radiation. Semin Cancer Biol 4:93–104, 1993.

4. Taylor AMR, Harnden DG, Arlett CF, Harcourt SA,Lehmann AR, Stevens S, Bridges BA: Ataxia telangi-ectasia: a human mutation with abnormal radiation sen-sitivity. Nature 258:427–429, 1975.

5. Cole J, Arlett CF, Green MH, Harcourt SA, Priestley A,Henderson L, Cole H, James SE, Richmond F: Com-parative human cellular radiosensitivity: II. The sur-vival following gamma-irradiation of unstimulated (G0)

Fig. 7. Cell cycle analysis of fibroblasts after fractionatedirradiation. The percentage of cells in each of the cell phasesG1, S, and G2/M are shown. The data are from one repre-sentative experiment.

214 Wilkins et al.: Optimizing Differences in Radiosensitivity

T-lymphocytes, T-lymphocyte lines, lymphoblastoidcell lines and fibroblasts from normal donors, fromataxia-telangiectasia patients and from ataxia-telangiectasia heterozygotes. Int J Radiat Biol 54:929–943, 1988.

6. Turesson I: Individual variation and dose dependencyin the progression rate of skin telangiectasia. Int J Ra-diat Oncol Biol Phys 19:1569–1574, 1990.

7. Little JB, Nove J, Strong LC, Nichols WW: Survival ofhuman diploid skin fibroblasts from normal individualsafter X- irradiation. Int J Radiat Biol 54:899–910, 1988.

8. Geara FB, Peters LJ, Ang KK, Wike JL, Sivon SS,Guttenberger R, Callender DL, Malaise EP, Brock WA:Intrinsic radiosensitivity of normal human fibroblastsand lymphocytes after high- and low-dose-rate irradia-tion. Cancer Res 52:6348–6352, 1992.

9. Geara FB, Peters LJ, Ang KK, Wike JL, Brock WA:Prospective comparison of in vitro normal cell radio-sensitivity and normal tissue reactions in radiotherapypatients. Int J Radiat Oncol Biol Phys 27:1173–1179,1993.

10. Burnet NG, Nyman J, Turesson I, Wurm R, YarnoldJR, Peacock JH: Prediction of normal-tissue toleranceto radiotherapy from in vitro cellular radiation sensitiv-ity. Lancet 339:1570–1571, 1992.

11. Loeffler JS, Harris JR, Dahlberg WK, Little JB: In vitroradiosensitivity of human diploid fibroblasts derivedfrom women with unusually sensitive clinical responsesto definitive radiation therapy for breast cancer. RadiatRes 121:227–231, 1990.

12. Johansen J, Bentzen SM, Overgaard J, Overgaard M:Evidence for a positive correlation between in vitroradiosensitivity of normal human skin fibroblasts andthe occurrence of subcutaneous fibrosis after radio-therapy. Int J Radiat Biol 66:407–412, 1994.

13. Raaphorst GP: Prediction of radiotherapy response us-ing SF2: is it methodology or mythology? RadiotherOncol 28:187–188, 1993.

14. Brock WA, Tucker SL, Geara FB, Turesson I, Wike J,Nyman J, Peters LJ: Fibroblast radiosensitivity versusacute and late normal skin responses in patients treatedfor breast cancer. Int J Radiat Oncol Biol Phys 32:1371–1379, 1995.

15. Geara FB, Peters LJ, Ang KK, Wike JL, Brock WA:Radiosensitivity measurement of keratinocytes and fi-broblasts from radiotherapy patients. Int J Radiat OncolBiol Phys 24:287–293, 1992.

16. Begg AC, Russell NS, Knaken H, Lebesque JV: Lackof correlation of human fibroblast radiosensitivity in

vitro with early skin reactions in patients undergoingradiotherapy. Int J Radiat Biol 64:393–405, 1993.

17. Bichay T, Feeley MM, Raaphorst GP: A comparison ofheat sensitivity, radiosensitivity and PLDR in four hu-man melanoma cell lines. Melanoma Research 2:63–69, 1992.

18. Brown RC, Raaphorst GP, Danjoux CE, Drouin P:Multi-organ fatal chronic complications following ra-diation treatment for cancer of the cervix—results offibroblast assay. Gynecol Oncol 61:116–121, 1996.

19. Ng CE, Bussey AM, Raaphorst GP: Inhibition of po-tentially lethal and sublethal damage repair by campto-thecin and etoposide in human melanoma cell lines. IntJ Radiat Biol 66:49–57, 1994.

20. Brock WA, Baker FL, Wike JL, Sivon SL, Peters LJ:Cellular radiosensitivity of primary head and necksquamous cell carcinomas and local tumor control. IntJ Radiat Oncol Biol Phys 18:1283–1286, 1990.

21. Brock WA, Baker FL, Peters LJ: Radiosensitivity ofhuman head and neck squamous cell carcinomas in pri-mary culture and its potential as a predictive assay oftumor radiocurability. Int J Radiat Biol 56:751–760,1989.

22. West CM, Davidson SE, Burt PA, Hunter RD: Theintrinsic radiosensitivity of cervical carcinoma: corre-lations with clinical data. Int J Radiat Oncol Biol Phys31:841–846, 1995.

23. West CM, Davidson SE, Roberts SA, Hunter RD: In-trinsic radiosensitivity and prediction of patient re-sponse to radiotherapy for carcinoma of the cervix. BrJ Cancer 68:819–823, 1993.

24. Weichselbaum RR, Beckett MA, Vijayakumar S, Si-mon MA, Awan AM, Nachman J, Panje WR, GoldmanME, Tybor AG, Moran WJ, et al.: Radiobiologicalcharacterization of head and neck and sarcoma cellsderived from patients prior to radiotherapy. Int J RadiatOncol Biol Phys 19:313–319, 1990.

25. Peters LJ, Brock WA: Cellular radiosensitivity as pre-dictors of treatment outcome: where do we stand? Int JRadiat Oncol Biol Phys 25:147–148; 1993.

26. Peters LJ: The ESTRO Regaud lecture. Inherent radio-sensitivity of tumor and normal tissue cells as a predic-tor of human tumor response. Radiother Oncol 17:177–190, 1990.

27. Lambin P, Fertil B, Malaise EP, Joiner MC: Multipha-sic survival curves for cells of human tumor cell lines:induced repair or hypersensitive subpopulation? RadiatRes 138:S32–S36, 1994.

Wilkins et al.: Optimizing Differences in Radiosensitivity 215