J.F. Fowler

A review of αβ ratios for experimental tumors: Implications for clinical studies of altered fractionation

Clinical interest in the use of more and smaller dose fractions in radical radiotherapy has been stimulated by recent reviews of experimental results with normal tissues. It has been found that if the dose per fraction is reduced (i.e., in hyperfractionation) there is sparing of late responding normal tissues relative to those which respond early. This phenomenon can be understood in terms of the shapes of the underlying dose effect relationships, which can be described using the linear quadratic equation. The ratio (alpha/beta) of the linear (alpha) and quadratic (beta) terms is a useful measure of the curviness of such dose effect curves. Low alpha/beta values (1.5 to 5 Gy) have been observed for late responding normal tissues and indicate that radiation damage should be greatly spared by the use of dose fractions smaller than the 2 Gy used in conventional radiotherapy. By contrast the high alpha/beta values (6-14 Gy) observed for acutely responding normal tissues indicate that the response is relatively linear over the dose range of clinical interest. Hence less extra sparing effect is to be expected if lower doses per fraction are administered. If tumors respond in the same way as acutely responding normal tissues then hyperfractionation might confer a therapeutic gain relative to late responding normal tissues. We have reviewed published results for experimental tumors irradiated in situ and either assayed in situ or after excision. The alpha/beta ratios were usually at least as high as those for acutely responding normal tissues, and 36/48 tumors gave values greater than 8 Gy. Low values of less than 5 Gy were obtained for only 4/48 tumors. There are considerable technical problems in interpreting these experiments, but the results do suggest that hyperfractionation might confer therapeutic gain relative to late responding normal tissues on the basis of differences in repair capability. In clinical practice more efficient reoxygenation, cell cycle redistribution and decreased overall treatment time might also confer therapeutic gain.

Potential for increasing the differential response between tumors and normal tissues: Can proliferation rate be used?☆

Rapid proliferation of malignant cells has not previously been emphasized as a major source of failure to control tumors. Evidence is presented that the effective doubling times of clonogenic cells in human tumors during multifraction radiotherapy are in the range of a few days, that is, similar to the pre-treatment Potential Doubling Times and much shorter than Volume Doubling Times. Evidence from animal tumor studies leads to the same conclusion. Accelerated fractionation should be considered for individual human tumors whose LI is measured (e.g., by flow cytometry and the BUdR antibody) and found to be too high.

Review: Total Doses in Fractionated Radiotherapy—Implications of New Radiobiological Data

The total dose in radiotherapy has been adjusted in the past for different fractionation schedules by the use of empirical formulae such as NSD, TDF and CRE. It is now appropriate to consider fractionation factors which include more biological insight in their formulation than was possible earlier. It has become clear, from both clinical and experimental animal data, that the total dose in multi-fraction irradiations depends more critically on size of dose-per-fraction for late than for early damage to normal tissues. This difference has been interpreted as due to different shapes of the underlying dose-response curves. The late reactions respond with more curvature in the dose-response curve, i.e. with more repair capability at very low doses per fraction, than the early tissue reactions. A linear-quadratic relationship for the dose-response curves has been found to fit experimental data well, with few exceptions. This paper reviews this interpretation and explores some of its implications for radiotherapy and for radiobiology applied to therapy. Of many repair factors that have been suggested, the ratio alpha/beta (of the linear to the quadratic coefficients) is one that should be independent of the level of damage assayed. Values of alpha/beta of about 10 Gy have been reported for a number of early tissue responses but a range of values from about 1 to 5 or 6 Gy for late responses. It is a current challenge to radiobiology to explain why this difference occurs. Once such values are known for different tissues--and the dangers of premature assumptions are emphasized--calculations are possible which might be useful in radiotherapy as an alternative to NSD, TDF, CRE etc. Some data are presented on the magnitude of differences from these previously used empirical formulae, with a discussion about how easily detected the discrepancies might be in clinical practice. Applications to hypofractionation, hyperfractionation and accelerated fractionation are illustrated.

Skin reactions in mice after multifraction x-irradiation.

SummaryFractionated doses of 240 kV X-rays were given to one hind leg of mice, and skin reactions were recorded up to 5 weeks after the last dose. The following schedules of irradiation were used: 15 equal fractions in 18 days (15F/18d), 9F/18d, 9F/10d, 5F/9d, 5F/4d, 3F/4d, 2F/2d and single doses. The dose required to produce a given average skin reaction over the period 10 to 32 days, or a corresponding period after multifraction doses, increased steadily with time and fraction number; except for the 9F/10d schedule which produced reactions significantly less than would be expected from any simple sequence. The results are discussed in comparison with other experiments on skin. After allowing for concurrent proliferation, a dose–response curve was derived which had a Dq of 550 rads and a straight portion for doses above 1200 rads. The ratio of initial to final slopes was approximately 0·4 and this indicated the ratio of irreparable to reparable X-ray injury in the skin.

Repair in mouse lung for up to 20 fractions of X rays or neutrons

Local irradiation of the mouse thorax followed by the measurement of lung damage up to 17 months after irradiation has been carried out with up to 20 fractions of 3 MeV neutrons or of 240 kV X rays. Doses per fraction down to 0.28 and 1.5 Gy respectively were used. Repair capacity and RBE values were assessed by measuring breathing rate and lethality at monthly intervals up to 17 months. Only a small sparing of neutron damage was found. Sparing with X rays continued to increase as the size of each fraction was decreased, and was the main influence on the RBE values. The single-dose RBE was approximately 1.8, increasing to approximately 5 at the lowest dose per fraction measured. Dose-response curves derived for each fraction were well fitted by the formula alpha d + beta d2 where the repair parameter alpha/beta has values of 2-4 Gy after X irradiation. A slight fall of alpha/beta with time after X irradiation was observed, from about 4 Gy for pneumonitis to about 2 Gy for late fibrosis. This was significant for lethality but not for the increase of breathing rate. With neutrons the value of alpha was much higher than with X rays and a trend of increasing value of alpha at later times after irradiation was seen. Use of the linear quadratic dose-response formula predicts a continuing increase in the sparing of X-ray damage in lung as doses per fraction are decreased below those used here, and a limiting low-dose RBE of about 7.

Repair in mouse lung between multiple small doses of X rays

Multiple fraction experiments have been carried out to determine the response of mouse lung to repeated small doses of 240 kV X rays down to 150 rad/fraction using breathing rate and lethality to assess damage. Two experimental approaches were used to measure the effect of small doses in vivo: (1) multiple equal doses and (2) multiple priming doses followed by a large test dose. Analysis was performed using the multitarget two-component model and the linear quadratic model of cell survival. The amount of repair was calculated as a function of either dose per fraction (FR) or total dose (Frec). Both FR and Frec increased with decreasing dose per fraction but the change in FR was small. The advantage of Frec was that it varied more rapidly with dose per fraction than FR, so that possible differences between tissue repair capabilities are more visible on plots of repair as a function of dose per fraction. FR and Frec both decreased with the level of single-dose isoeffect injury; thus neither parameter is acceptable for comparing repair capability of different normal tissues with widely differing single-dose end point levels. Beta/alpha values were calculated and found to be a more acceptable index of repair capability than either FR or Frec because unlike those two parameters, beta/alpha varied little with level of damage. Beta/alpha values of 1.7 to 4.2 krad-1 were obtained for both lung death and increased breathing rate and are clearly intermediate between the lower beta/alpha ratios for acute reactions, i.e., skin and intestine, and the higher values for late reactions in kidney and spinal cord.

Fractionation with X rays and neutrons in mice: response of skin and C3H mammary tumours

Abstract The effectiveness of fast neutron therapy was tested where it was unlikely to show any advantage relative to conventional X rays, i.e. on a mouse mammary tumour which contains radioresistant hypoxic cells, but which reoxygenates extensively after a large X-ray dose. The therapeutic effects were estimated by comparing the degree of early skin reaction associated with a standard frequency of local tumour control assessed at 150 days. Five and nine fractions of X rays or fast neutrons were given at the expected optimum spacing for re-oxygenation after X rays. Single doses were also used. Five fractions of X rays given in nine days were indeed found to be as effective as either of the fractionated neutron treatments, which were closely similar to each other, but nine fractions of X rays in 18 days were considerably less effective. These results suggest that X rays can be made as effective as fast neutrons, by a sharply optimal choice of fractionation dose and interval. However, to make this choice re...

A murine model of lip epidermal/mucosal reactions to X-irradiation

A new radiobiological test system has been developed for lip epidermal/mucosal reactions in mice. This is intended for use in investigations of the effect of non-standard fractionation and of modifying drugs on oral radiation reactions in human cancer patients. An arbitrary scale of scores was devised, with separate scores for oedema of the lips and for erythema or exudation. After single doses of 13-20 Gy, the mouse lip epidermal reactions began at 5 days, reached a peak about 10-13 days, and had fallen to low values, but not to zero, by 21 days. Several different periods for averaging the reaction scores were tested for relative steepness and variability, the most useful being 10-12 days inclusive or the 12th day score alone. The use of longer periods of averaging led to apparent saturation of the scores. It was found that large doses of X-rays repeated at 21-23 day intervals did not lead to escalating waves of reactions unless each dose was greater than 17 Gy. With these larger doses, escalation of reactions occurred even if the intervals were extended.

RECOVERY AND REPOPULATION IN MOUSE SKIN AFTER IRRADIATION WITH CYCLOTRON NEUTRONS AS COMPARED WITH 250-kv X RAYS OR 15-Mev ELECTRONS

In a previous publication, Fowler et al. (1) discussed the short-term recovery of mouse skin after 15-Mev electron irradiation, the recovery being measured in terms of the parameter D2 D1, where D2 is the total dose given in two fractions to produce the same biological effect as a single dose, D1. For mammalian cells in tissue culture and in vivo, the magnitude of the recovery from sublethal damage as described by Elkind and Sutton (2) has been shown to be smaller after high-LET irradiation (3-6). Using an intact mammalian system (the survival of mice for more than 4 days after single or split doses of cyclotron neutrons or X-rays), Hornsey et al. (7) found a small value of D2 D1 after neutrons, but a relatively large value after X-rays. Previous work with pig skin had, however, shown the short-term recovery after fast neutrons to be about two-thirds that after X-rays, allowing for RBE (8). In the present paper the mouse skin system is used to compare the results of fast neutron irradiation with those of 250-kv X-rays and 15-Mev electrons. The work may be divided into two parts. The first consists of a comparison of the time of appearance of skin reactions, their rate of increase, and their subsequent rate of decrease, after fast neutron as compared with X-ray or 15-Mev electron irradiation. The skin reactions observed are presumed to be due to epithelial cell depopulation, so that the rate of decrease of the skin reaction is taken as a measure of the repopulation rate of epithelial cells (19). The second part consists of the investigation of the magnitude (and rate) of the short-term recovery in mouse skin, measured by the dose increment D2 D1,

Radiosensitization of Chinese Hamster Cells by Oxygen and Misonidazole at Low X-ray Doses

The radiosensitization of Chinese hamster V79 cells in vitro by air and misonidazole at low X-ray doses (0.2-6.0 Gy) had been studied. These survival data, together with high-dose data, were fitted to the linear quadratic model ln S = -(alpha D + beta D2), deriving estimates of alpha and beta by six different methods to illustrate the influence of the statistical treatment on the values so derived. This in vitro study clearly demonstrated that the survival parameters alpha and beta are dependent to some degree on the method of analysis of the raw survival data; however, their ratios, the values of oxygen enhancement ratios (OERs) and radiosensitizer enhancement ratios (SERs) derived from the different methods, are similar. All methods of analysis give reduced OERs at low radiation doses for combined low- and high-dose X-ray data. However, the OERs are still appreciably high, ranging from 2.45 to 2.50 for an oxic dose of 2 Gy. All methods of analysis gave reduced SERs at low doses for combined low and high X-ray dose data for hypoxic cells irradiated in 1 mmol dm-3 misonidazole. At survival levels corresponding to doses of 2 Gy in the presence of 1 mmol dm-3 misonidazole and SERs ranged from 1.2 to 1.5.