5-Fluorouracil radiation sensitization-A brief review
John E.Byfield
Medical Director,Radiation Therapy Associates Medical Group, Kern Regional Cancer Center,3550 Q Street, Bakersfield, CA 93301, USA
Key words:5-fluorouracil,radiation,radiosensitization, pharmacology,squamous cancers
Summary
5-Fluorouracil (FUra) has emerged as the most promising clinical radiosensitizer now available. FUra’s capacity to render human cells more sensitive to x-rays was established soon after its synthesis.However, the recognition that the drug’s unusual pharmacology dictated explicit scheduling requirements in man was not realized until recently when work in the author’s laboratory identified the extra-cellular drug concentra-tion X time factors necessary to create the intra-cellular radiosensitive state. Subsequent clinico-pharma-cologic investigations led to the realization that only prolonged continuous infusions combined with ap-propriately fractionated,cyclical radiation therapy would maximize the clinical utility of this approach.In-fused FUra radiation-sensitization therapy reaches its maximum efficacy against the squamous-transitional cancers.This group of human malignancies comprises about 15% of al human cancers.Preliminary data also suggests substantial promise in the local/regional control of rectal and breast cancers. Infused FUra used as a radiosensitizer has the potential to eliminate the need for about 90% of all radical cancer surgery.
Introduction
5-Fluorouracil(FUra)was one of the first rationally synthesized anti-cancer agents,having been created as an analogue of uracil with a view to interfering with RNA synthesis [1].Soon after its introduction it was found to affect the radiation survival curve of cultured mammalian cells [2]. Work by Vietti et al. identified striking FUra radiosensitization in vivo in a mouse leukemia system [3]. These kinds of observations led to a large number of clinical trials which combined bolus FUra and radiation. Many of the early trials showed benefit although the effect was invariably modest [4,5]. Variants of these trials continue to be reported, almost always with the addition of a second chemotherapeutic agent-usually an alkylating agent, and usually in the treatment of gastro-intestinal adenocarcino-mas where bolus FUra remains the (somewhat tar-nished) chemotherapeutic “gold standard” [6,7].

In the following brief review,the data from var-ious pre-clinical and clinical studies done by the author’s group will be summarized and related to studies from other investigators in order to develop a coherent current picture of FUra used as a radio-sensitizer. All of this evidence indicates that prior investigations of FUra + x-ray effects using bolus drug employed a schedule that actually precluded radiosensitization by FUra. Remarkably, the max-imum benefit obtained using FUra as a radiosen-sitizer occurs in its use against tumors of the squa-mous/transitional cell group, cancers that by-and-large share a common embryonic origin. Based on this survey future directions in the application of FUra radiosensitization are suggested.
Pre-clinical studies
In 1977 we first reported quantitative tissue culture 
studies employing human cancer cells treated with FUra and radiation with a view to identifying ex-plicitly the prerequisites of radiosensitization [8,9]. We determined the effects of various treatment combinations and sequences using different drug concentrations,exposure times, and x-ray doses in a comprehensive matrix, changing each variable while holding the others constant [9]. Quantitative cloning permitted a direct analysis of the number of cells capable of surviving each combination. The data confirmed FUra was a potent radiosen-sitizer but equally clearly showed that the sensitiza-tion phenomenon occurred only under strictly de-fined circumstances:
(a) FUra had to be present for at least 24 hours af-ter each radiation exposure in order to estab-lish the radiosensitive state. Prior exposure to the drug (with its removal after x-ray exposure) had no effect on radiation survival.
(b) In order for FUra to render cells sensitive to radiation, a demonstrable degree of cell killing by FUra had to occur. In other words, some ef-fectiveness of drug alone must be seen (equiva-lent quantitatively to killing slightly short of a clinical partial response).
(c) The radiosensitizing effect was found to be independent of the cell line studied provided sufficient FUra was present to establish signifi-cant toxicity. Thus FUra-insensitive human cancers probably cannot be radiosensitized by FUra.
In a subsequent series of in vitro studies [10] we examined the sensitivity of a variety of human tumor cell lines to FUra alone.Epithelial tumor lines from both adeno- and squamous carcinomas were used.The sensitivity of these lines differed by approximately a factor of two.The time-dose fac-tors for FUra cytotoxicity were also studied be-cause of their relevance to clinical exposures.Un-like other anti-metabolites there was no apparent limit to FUra cell killing.
The salient features of FUra’s interaction with radiation are illustrated diagrammatically in Fig. 1 which shows three radiation survival curves.Curve 1 illustrates the survival of typical human epithelial carcinoma cells when exposed to varying radiation doses (no drug). It can be seen that there is an in-

Rads
Fig. 1. Diagram of 5-Fluorouracil radiosensitization. Curve 1: Radiation survival curve for cells not treated with 5-FU.Curve 2: Radiation survival curve for cells treated post-radiation with sufficient 5FU to kill 50% of the cells without radiation (partial response equivalent). Curve 3: Radiation survival curve with 5FU killing to 10%(typical of cell system very sensitive to 5FU).
itial shoulder region followed by dose-dependent logarithmic decrease in survival.
Curve 2 shows the effect of FUra when added immediately after radiation and left present forat least 24 hours. In this case 50% of the ceIls have been killed by the FUra (50% survival at the zero radiation dose). Under these circumstances the slope of the radiation survival curve is now steeper. This indicates that this particular combination of FUra(i.e.in terms of concentration x time) has enhanced tumor cell killing and the cells have been “radiosensitized”.
Curve 3 shows what happens when even more cells have been killed by FUra (shown by fewer sur-vivors at the zero radiation dose). The survival curve is now even steeper such that every radiation dose kills more cells than would be expected from “additive” killing(where the surviving cell slope would be identical to the slope of the radiation-alone curve). It is important to note that the tumor 
cell killing in curve 3 would be dramatically differ-ent at the clinical level since the FUra effect is itself much greater and radiosensitization has achieved an even greater synergistic effect.The more sensi-tive a tumor to FUra, the more rapidly it will dis-appear under the pressure of a radiosensitizing treatment regimen.
These studies were then compared to the in vivo pharmacologic data generated simultaneously in our clinic in order to re-create at the clinical level the radiosensitization seen in tissue culture.
Pharmacology of 5-Fluorouracil in man
Pharmacological studies of FUra date from the work of Clarkson and colleagues who established its primary features [11], most especially its short half-life (10-15 minutes). Typical bolus doses of FUra disappear from the blood stream rapidly be-cause of hepatic degradation of drug. Since radio-sensitization requires constant drug exposure for at least 24 hours after each x-ray exposure, bolus drug dosing cannot achieve radiosensitization.
Subsequent studies revealed the second pharma-cological factor involved in using FUra as a radio-sensitizer, viz. its non-linear pharmacokinetics [12,13]. Non-linear pharmacokinetics means that there is not a linear relationship between the dose of drug given and the maximum or mean plasma drug level achieved.This may occur for several rea-sons. In the case of FUra it appears to occur be-cause there are at least two mechanisms competing for drug removal under dosing conditions used in man.We believe these two mechanisms have quite different origins and capacities.
The mechanisms which lead to FUra’s non-linear pharmacokinetics are not agreed upon. It is the author’s belief that thetwo clearance mechanisms relevant at infusion dose levels are: (a) removal of drug by proliferating body tissues, largely through incorporation into RNA,and (b) hepatic degrada-tion [13]. These two mechanisms must be an-tagonistic since hepatic degradation “wastes”drug while incorporation of the drug into RNA is a known mechanism for one of FUra’s cytotoxic ef-fects [14].

At the clinical level these two phenomena inter-act leading to the puzzling observation that essen-tially no FUra is demonstrable at infusion levels be-low about 15mg/kg/24 hours (constant infusion). Under such circumstances FUra “clearance”equals the cardiac output (i.e. the drug is totally cleared during a single passage (subject to minor differences in tissue bed uptake). At dose rates higher than about 15 mg/kg/24 hours there is a linear relationship between infusion dose and mean plasma level (from 15 to 65 mg/kg/24 hours).
The plasma FUra levels seen during clinical infu-sions between 15 and 65 mg/kg/day span the con-centrations needed to induce significant cytotoxic-ity in human tumor cells in tissue culture and to in-duce radiosensitization (500 to 2,000 ng/ml for 72-120 hour infusions,cf.ref.[15]. This observa-tion then established a quantitative link between tissue culture data and the drug levels needed for in vivo cytotoxic radiosensization of human tumors.
Scheduling and FUra radiosensitization
The former (radiobiological) studies indicated the “scheduling” requirements for obtaining FUra radiosensitization. The drug must be present for at least 24 hours after each (and every) radiation frac-tion in order that maximum radiosensitization be achieved.Any other combination would fail to radiosensitize.This factor places rigid structures on any clinical program.
The latter (pharmacologic) studies showed that FUra’s intrinsic pharmacokinetics do not immedi-ately lend themselves to reproducing the condi-tions needed for radiosensitization. Indeed, the short half-life of the drug (from hepatic removal) preclude anything other than additive effects when bolus drug is added to any variety of radiation frac-tionation scheme.
Together the two sets of requirements meant that a continuous infusion in which drug is made present for at least 24 hours after each radiation fraction would be optimal.
The next step in the development of a FUra radio-sensitizing regimen was therefore to determine the optimum duration of the infusion. In theory 
the briefest infusion would be 24 hours in duration since this is the minimum radiosensitizing expo-sure.Moertel and colleagues had studied infusions up to 24 hours in duration and not noted any sig-nificant difference in toxicity versus bolus therapy [16].However,Seifert et al.clearly showed that 5 day FUra infusions showed a marked shift in limiting toxicity [17]. We examined 72 hour infu-sions and found an intermediate type of toxicity complicated by CNS side-effects not seen with with shorter infusions where marrow toxicity dominates [13]. Lokich et al. have studied “protracted”‘ infu-sions in which the drug is essentially infused all the time (or at least until toxicity, tumor progression, or a mechanical event causes cessation of therapy).
From this data and from the known requisites of FUra radiosensitization it is reasonable to propose that infusions between 96 and infinite hours can be used for radiosensitizing regimens provided the drug is infused to limiting toxicity. In almost all pa-tients this toxicity will involve some component of each patient’s squamous cell renewal system.
Integration of FUra and radiation
From the results of the above studies it is possible to propose some rules concerning the development of “integrated” drug and x-ray regimens using FUra as a putative radiosensitizer.
(a) The infusion should be at least 96 hours long and should be given to toxicity.The latter stems from the observation that the degree of radio-sensitization achieved is a function of primary FUra cell killing (Fig. 1).While radiosensitiza-tion can be achieved with infusions shorter than 96 hours,CNS side-effects become a limit-ing toxicity and can be lethal [13]. Therefore 96 hours is the “shortest” infusion that is both safe and useful.
(b) The radiation fractionation scheme may be varied at will. The author has successfully em-ployed conventional fractionation (180-200 rad/day),hyperfractionation,and hypofrac-tionation schemes.There is no convincing evi-dence that infused FUra affects the late effects of radiation which are a function,primarily,of

the daily treatment fraction size.The capacity to combine radiosensitizing infused FUra with hyperfractionated radiation is important, par-ticularly in re-treatment patients where toler-ance is limited and can be increased by hyper-fractionation.
(c)The author feels strongly that cyclicity is vital in the application of FUra as a radiosensitizer. The use of cyclical treatment is fundamental to cancer chemotherapy but alien to classical radi-ation therapy (where it is termed “split-course”therapy).
All of the studies and conclusions noted above apply equally to the normal tissues that are respon-sible for “early” or acute radiation side effects. Clearly FUra can radiosensitize normal cells as well.However,this normal tissue radiosensitiza-tion occurs only in the irradiated field and for all in-tents and purposesis relevant only in head and neck and lower pelvic cancers. In both areas suitable frac-tionation (i.e.cyclical therapy) will permit rapid normal tissue recovery [19]. The striking results of infused FUra and radiation in head and neck cancer clearly show that cyclical treatment employing FUra does not suffer from the limitations apparent in split-course radiation treatments [19].
These “rules” suggest that optimal treatment will involve tumors that develop from transformed cells derived from normal tissues sensitive to in-fused FUra.These include all squamous tissues(in-cluding bladder transitional cell tumors) and breast carcinoma.Stomach carcinomas react very much like squamous cell cancers of the distal esophagus(which are very sensitive to FUra radio-sensitization) despite their disparate histology. Less sensitive are the colo-rectal carcinomas for which additional measures are required. Tumors insensitive to FUra cannot be radiosensitized using this approach.
Practically speaking, a 5 day infusion at 25-30 mg/kg/day will radiosensitize virtually all FUra-sensitive cancers as described in references 20 and 21 and listed in Table 1. The author currently uses 180 rad per day (which has less adverse late effects) but protracts therapy. This dose is given 5 days a week in conjunction with a 5 day (120 hour) FUra infusion.After this treatment cycle the patient is 
Table 1. Cancers sensitive to 5FU radiosensitization
Cancer Incidence Surgical procedure
Esophagus 9,300 Esophagectomy
Anus 2,500 Miles procedure
Larynx 11,700 Laryngectomy
Vulva 4,400 Vulvectomy
Penis 1,300 Phallectomy
Bladder 40,500 Cystectomy
Rectum* 42,000 Miles procedure
Breast 123,900 Mastectomy
The total “radical” cancer surgical procedures performed in the U.S.each year is about 242,500 of which about 97% are sus-ceptible to 5FU radiosensitization.*Rectal carcinoma is only slightly sensitive to 5FU.
rested 9 days.The duration of this inter-cycle rest period should be kept at a minimum (i.e. 9 days) unless toxicity remains significant then further recovery,usually 1 week is advised.The regimen can be conducted totally as an out-patient with any reasonably compliant,intelligent patient.Tumors in which this regimen has proven useful are listed in Table 1. The reader will note that the list in-cludes the majority of human cancers otherwise re-quiring radical surgery. The reader is referred to re-cent reviews by the author [20,21] for a more detailed discussion of current clinical results with FUra radiosensitization.
Future directions
The available clinical data supporting the conten-tion that FUra used as a clinical radiosensitizer has great potential has been summarized elsewhere in this group of papers and by the author [20,21]. While the “optimal” regimen has yet to be estab-lished,the biweekly 5 day schedule with FUra taken to limiting toxicity currently appears close to ideal. The addition of a second chemotherapeutic agent is the most obvious next step and many current regimens employ such an agent (usually cisplatin or Mitomycin C). As yet the benefit of these addi-tional compounds has not been established when compared to infused FUra, used properly,a single agent. Frequently, the FUra dose is reduced to permit the inclusion of these second agents and

such reduction is probably unwise. Therefore it is important that their inclusion be justified by care-fully designed studies where their contribution can be distinguished from the contribution of radiosen-sitizing FUra which is considerable.
The other major area for current research is the study of this approach against tumors whose defini-tive therapy involves a radical surgical resection carrying major morbidity.Since the infused FUra-radiation program is effective against many such cancers (Table 1) it seems plausible that it may substitute for such radical surgery in achieving loco-regional tumor control.The elimination of the need for such procedures is a worthy goal while the medical community awaits effective systemic therapy against most of the tumors responsive to this regimen.
References
1.Heidelberger C,Ansfield FJ: Experimental and clinical use of fluorinated pyrimidines in cancer chemotherapy.Cancer Res 23:1226-1243,1963
2.Bagshaw M:Possible role of potentiators in radiation ther-apy. Amer J Roentgenol 85:822-833, 1961
3.Vietti T,Eggerding F,Valeriote F: Combined effect of x-radiation and 5-fluorouracil on survival of transplanted leukemia cells. J. Natl Cancer Inst 47:865-870,1971
4.Moertel CG,Childs DS, Reitemeier RJ,et al.: Combined 5-fluorouracil and supervoltage radiation therapy of locally unresectable gastrointestinal cancer. Lancet 2:865-867, 1969
5.Lo TCM,Wiley Jr AL,Ansfield F,et al.:Combined radi-ation therapy and 5-fluorouracil for advanced squamous cell carcinoma of the oral cavity and oropharynx: a ran-domized study. Amer J Roentgenol 125:229-235,1976
6.Gastrointestinal Tumour Study Group: A comparison of combination chemotherapy and combined modality ther-apy for locally advanced gastric carcinoma. Cancer 49: 1771-1777,1982
7.Gastrointestinal Tumor Study Group:Prolongation of the disease free interval in surgically resected rectal cancer.New Eng J Med 312:1465-1472,1985
8.Byfield JE,Chan PYM,Seagren SL:Radiosensitization by 5FU:Molecular origins and clinical scheduling implica-tions.(Abstr) Proc Amer Assoc Ca Res 18:74, 1977
9.Byfield,JE,Calabro-Jones PM, Klisak I, et al.:Pharmaco-logic requirements for obtaining sensitization of human tumor cells in vitro to combined 5-fluorouracil or ftorafur and x-rays. Inter J Rad Oncol Biol Phys 8:1923-1933, 1982
10.Calabro-Jones PM,Byfield JE,Ward JF,et al.:Time-dose 
relationships for 5-fluorouracil cytotoxicity against human epithelial cells in vitro. Cancer Res 42:4413-4420,1982
11.Clarkson B, O’Connor A, Winston, La R, et al.:The phys-iologic disposition of 5-fluorouracil and 5-fluoro-2′-deoxy-uridine in man. Clin Pharmacol Ther 5:581-610, 1962
12.Myers CE:The pharmacology of the fluoropyrimidines. Pharm Rev 33:1-15,1981
13.Byfield JE,Sharp TR,Frankel SS,et al.:Phase I and phar-macologc study of 72-hour infused 5-fluorouracil in man. Amer J Clin Oncol 8:429-440, 1985
14.Valeriote F, Santelli G: 5-fluorouracil (FUra) Pharmacol Ther 24:107-132,1984
15.Byfield JE,Frankel SS,Sharp TR, et al.:Phase I and phar-macologic study of 72-hour infused 5-fluorouracil and hyperfractionated radiation. Inter J Rad Oncol Biol Phys 11:791-800,1985
16.Moertel CG,Schutt AJ, Reitemeier RJ,et al.: A compari-son of 5-fluorouracil administered by slow infusion and rapid injection. Cancer Res 32:2717-2719,1972
17.Seifert P,Baker LH, Reed ML, et al.:Comparison of con-tinuously infused 5-fluorouracil with bolus injection in treatment of patients with colo-rectal cancer. Cancer 36: 123-128,1975

18.Lokich J,Bothe A, Fine N, et al.:Phase I study of pro-tracted venous infusion of 5-fluorouracil. Cancer 48:2565-2568,1981
19.Byfield JE,Sharp TR,Tang S,et al.:Phase I and II trial of cyclical 5-day infused 5-fluorouracil and coincident radia-tion in advanced cancer of the head of neck. J Clin Oncol 2: 406-413,1983
20.Byfield JE:Theoretical basis and clinical applications of 5-fluorouracil used as a radiosensitizer. In:Rosenthal CJ,Rotman M (eds) Clinical Applications of Continuous Infusion Chemotherapy and Concomitant Radiation Ther-apy.Plenum Publishing Corp. New York, 1986, pp 113-125
21.Byfield JE:The clinical use of 5-fluorouracil and other halopyrimidines as radiosensitizers in man. In: Lokich J (ed)Cancer Chemotherapy by Infusion.Precept Press, Chicago,1987,pp 479-501
Address for offprints: J.E. Byfield,Kern Regional Cancer Center, 3550 Q Street, Bakersfield, CA 93301, USA