The goal of treatment of the cancer patient is to achieve a tumour-free and morbidity-free status. This objective requires irradiation of all local, regional, and distant tumour, by means which produce no or minimal morbidity.

Radiation is an important treatment for the cancer patient in this regard. Correct use of ionizing radiation destroys many early tumours and preserves anatomy and function. Good functional and cosmetic results can be obtained from the radiation treatment of early stage tumours affecting the vocal cords, many sites in the mouth and pharynx, prostate, rectum, vagina, and skin, and also in patients with Hodgkin's disease, uveal melanoma and retinoblastoma. A combination of radiotherapy and conservative surgery can sometimes be used in place of more radical surgery, providing better functional and cosmetic results; this is currently applicable to some patients with carcinoma of the breast or soft tissue sarcomas of the extremities. Radiation is also combined with chemotherapy, with or without surgery, in the treatment of certain patients with carcinoma of the urinary bladder or colorectum, Ewing's sarcoma, rhabdomyosarcoma, or primary lymphoma of bone. Evidence is accumulating to suggest that radiotherapy plus chemotherapy has an advantage in the treatment of non-small cell carcinoma of the lung and locally advanced carcinoma of the uterine cervix. However, the destruction of normal tissue by the malignant process precludes an excellent result being obtained in patients with locally advanced disease at many anatomic sites.

The basic strategy in radiation oncology is to define the target volume correctly in three dimensions; to design an array of radiation beams that will encompass a minimal target volume (the minimum treatment volume); to position and immobilize the patients in order to target the beam with a high degree of accuracy; and to employ a dose and fractionation protocol that has the highest probability of eradicating the tumour, while producing tolerable changes in the normal tissues unavoidably included in the treatment volume. Radiation therapy has an advantage over surgery in that there is little risk of producing symptomatic change in major vessels, nerves, bones, and tendons included in the treatment volume.

The choice of total dose and dose fractionation protocol is a complex process. Factors that need to be taken into consideration include tumour size, histological type, anatomical site, condition of the normal tissues to be included in the treatment volume, presence of disease processes which may be associated with increased radiation sensitivity (diabetes, ataxia telangiectasia, autoimmune disease), and whether the radiation is to be administered alone or in combination with surgery or chemotherapy.

Treatment of epithelial and mesenchymal tumours by radiation alone requires a dose of 65 to 75 Gy (Gray—see later) given at 1.8 to 2 Gy five times a week. This is reduced to 50 to 65 Gy when radiotherapy is combined with surgery which removes all gross disease. The dose also needs to be reduced when chemotherapy is being administered: for most drug regimens the required reduction is slight, but combination with agents such as Adriamycin and actinomycin D requires a reduction in radiation dose of about 10 per cent. Lymphomas are usually treated with doses of about 40 to 50 Gy, while many germinal tumours require doses of about 30 Gy. Stage Mo epithelial tumours are usually treated radically: the dose used is the maximum that is associated with major morbidity in 1 to 2 per cent of treated patients. When palliation of symptoms is the major goal of treatment, lower total doses are given over a short period.

Radiation treatment of early stage disease often produces local control and disease-free survival rates of over 90 per cent. Results are, however, less satisfactory in patients with advanced disease and in those whose tumours develop in anatomical sites immediately adjacent to radiation-sensitive structures, such as the brain-stem or eye, or in sensitive structures such as the bowel, kidney, or liver. Some tumours are apparently unusually resistant to radiotherapy. Conventional radiotherapy of glioblastoma multiforme with a radiation dose of 65 to 70 Gy results in long-term disease-free survival of less than 2 per cent of patients. At the other extreme, high survival rates can be obtained in patients with radiosensitive tumours such as germinoma and seminoma after a radiation dose of only 30 Gy over 3 to 4 weeks.

Four general strategies can be used to improve the efficacy of radiation therapy. These are: improving the dose distribution to concentrate the radiation dose in the target tissues; increasing the differential response between the tumour and normal tissues; developing means for predicting the response of tumour and normal tissue in individual patients; and improving our knowledge of the natural history and patterns of spread of the various types of tumour.

Ionizing radiation are short wavelength (<10&supminus;&sup8; cm) electromagnetic radiation or charged particles such as electrons, protons, and helium ions). Fast neutrons are neutral particles; they produce secondary charged particles (low energy protons) upon interaction with atoms in the irradiated material. These low energy protons produce a high density of ionization along their path. Their biological effect is different from that of X-rays: clinical testing of fast neutron beams has demonstrated an apparent superiority in treating carcinoma of the salivary glands.

Radiation is almost entirely based on photons and electrons. Photons are designated X-rays when produced by the acceleration of electrons into a high Z target such as tungsten. The X-ray beam is defined in terms of the energy of the electrons incident on the target: this defines the peak energy of the resultant X-rays. Linear accelerators in clinical use produce 4 to 50 MeV beams. The electrons lose their energy on interacting with the outer orbital electrons of the target (e.g. tungsten atoms). This energy is converted into photons (X-rays) or heat. As the energy of the incident electrons increases, the proportion of the energy appearing as photons increases. Thus cooling of the target is relatively easy for linear accelerators, in contrast with the situation for high output diagnostic X-ray units (low energy beams).

Photons are designated gamma rays when they are produced by the transition of a radioactive atom from one excited state to a less excited or energetic state. Clinical gamma ray units are almost exclusively fitted with &sup6;&sup0;Co or ¹³&sup7;Cs sources, which emit photons of about 1.2 and 0.6 MeV, respectively. Gamma rays from a variety of isotopes are used widely in brachytherapy (the insertion of radioactive materials into tissues, usually for a few days). Those most commonly used are ¹&sup9;²Ir, ¹²&sup5;I, ²²&sup6;Ra, and ¹³&sup7;Cs. Those sources are sometimes left in place permanently, for total decay; this particularly applies to ¹²&sup5;I.

Figure 1 2781 illustrates the decrease in delivered dose with depth of tissue for 10 MeV X-rays, 12 MeV electrons, and a modulated energy 160 MeV proton beam. For the photon beam, dose decreases with depth in a complex exponential manner, reaching 50 per cent, 25 per cent, and 12.5 per cent at depths of 17 cm, 32 cm, and 45 cm, respectively. These high energy photon beams have an advantage in the fact that the maximum dose is delivered to tissue at a depth of about 2.5 cm; the dose at the surface of the body is relatively low. In clinical terms, this results in a markedly reduced radiation dose to the skin and subcutaneous tissue, in contrast to the older 250 KeV X-ray units, which delivered 100 per cent of the dose to the skin.

The more powerful accelerators allow greater depth of penetration: the tissue depths of the 50 per cent isodose line are 7 cm, 10 cm, 17 cm, and 24 cm for 250 KvP, &sup6;&sup0;Co, 10 MeV and 34 MeV X-rays, respectively. This is important in the irradiation of deep-seated lesions of the torso. Electron beams also show different characteristics of absorption by tissues. For a 12 MeV electron beam there is a relatively sharp decrease in dose beyond 4 cm; at higher energies this fall-off is less steep, and by 35 MeV, electrons have little clinical advantage over X-rays.

The third curve shown in Fig. 1 2781 is that for a 160 MeV proton beam. There is a relatively flat dose across the range of interest, and a markedly steeper fall-off at the end of the range than is seen for electrons. The steep dose gradient is seen for the entire range of energies pertinent to clinical use. Proton therapy has produced important advances in the treatment of uveal melanomas and sarcomas of the skull base due to the ability to administer higher doses than are feasible with photon beams. Proton therapy for tumours of other sites is being evaluated.

Radiation for clinical use is generated by linear accelerators which produce a wide range of X-ray energies, high dose rates, and isocentric gantries that allow the machine head to rotate 360° around the target in the patient. Almost all patients can be treated in the supine position, and require only short periods of immobilization, and defined target areas of tissue can be treated with smaller margins of normal tissue. These can generate photon and electron beams for clinical use.

Figure 2 2782 shows the dose distributions for two commonly used field arrangements. Figure 2(a) 2782 shows a four field ‘box’ technique: AP/PA and right/left parallel opposed pairs of 10 MeV beams to treat a central pelvic lesion. Figure 2(b) 2782 demonstrates the use of wedge filters to modify the beam profile such that the isodose lines are at 45° instead of being perpendicular to the path of the beam. This provides an elegant solution to the problem of confining the dose to one section of the anatomical part; this is often valuable for part of the treatment of many lesions of the head and neck region.

Type of beam, energy, and treatment technique are chosen to allow the highest feasible concentration of radiation in the target. Efforts to improve photon techniques are being directed at three-dimensional treatment planning, on-line portal imaging, multi-leaf collimator systems, and greater use of enhanced immobilization systems. Research is also being directed into the use of intraoperative electron and photon beam therapy and intraoperative brachytherapy, charged particle beams (e.g. protons), etc.

Radiation is absorbed in discrete events by the orbital electrons and nuclei of the atoms of the irradiated material, producing ionizations and excitations. Ionizations are the result of either direct or indirect action of the radiation. The former involves interaction between a photon and the atoms of the target molecule itself. Indirect effects occur when the photons react with a water molecule, releasing reactive radicals such as OH⪾, which in turn causes damage to the target molecule (DNA).

The observed biological effect of radiation is principally due to the ionizations produced in critical molecules, particularly DNA. The latter is essential for cell survival; incomplete or faulty repair may be incompatible with cell survival. The impact of radiation-induced changes in DNA depends on the particular gene affected and the degree to which the damage can be repaired.

The unit of radiation is the Gray (Gy): 1 Gy is defined as the absorption of 10&sup4; ergs/g of the irradiated material. One Gray is the equivalent of 100 rads, the older nomenclature. The average energy required to produce an ionization is about 35 electron volts (ev); this is equivalent to 1.7 × 10¹&sup4; ionizations/g. For the average mammalian cell, a dose of 1 Gy produces an average of 200 000 ionizations per nucleus of 500 pg, and some 500 ionizations in the 10 pg of DNA. The net result is that 1 Gy produces about 1000 single-strand breaks and 50 double-strand breaks. Single-strand breaks are not judged to be major factors in the lethal effects of radiation: they are repaired rapidly, with a T&sub1;&subsol;&sub2; of only a few minutes. Double-strand breaks are a more serious threat to the cell as their repair is less effective. The lethality of a given dose of radiation shows good correlation with the number of double-strand breaks.

There is good experimental evidence to support the concept that the central target for cell killing by radiation is the DNA. Irradiation to the cytoplasm of mammalian cells using high doses of narrowly focused beams of heavy particles has no observable effect, while passage of only a few particles through the nucleus is fatal. The incorporation of [³H]thymidine, but not [³H]uridine into mammalian cells is highly toxic. The former is incorporated into DNA, which it then irradiates, while uridine is incorporated predominantly into the cytoplasmic RNA. Cells become highly sensitive to the effects of radiation when they have incorporated analogues of thymidine, such as bromodeoxyuridine and iododeoxyuridine, into their DNA. [¹²&sup5;I] is much more toxic when it is incorporated into the DNA as [¹²&sup5;] iododeoxyuridine, than when it is attached to the cell membrane in the form of [¹²&sup5;I] concanavalin A.

Certain oncogenes are associated with increased resistance to radiation. Rat embryo fibroblast lines transformed by the ras or ras and myc oncogenes often exhibit increased resistance to both high and low doses of radiation. Similar studies have not yet been conducted on established cell lines.

Resistance to radiation is rarely induced, even by irradiation over many cell cycles. In fact, cells surviving high single radiation doses may exhibit significant lower resistance. This is in marked contrast to the resistance which often develops following exposure to chemotherapeutic agents.

Successful radiation therapy of a cancer requires that all the tumour cells are killed, while enough normal cells remain to proliferate and reconstitute the irradiated tissue to normal or near-normal status.

Radiation-induced damage can be lethal or non-lethal: the former is of most interest to the clinician since all tumour cells must be killed in order to effect a cure. In this context, a lethal response is the loss of the ability to produce a continuously expanding progeny. Non-viable cells may retain metabolic integrity, and may even divide several times, before their eventual lysis. Although such cells may persist in the tissues for prolonged periods, they are of little consequence for the patient.

The factors determining the observed response of a tissue or organ to radiation include the number and proportion of cells that lose their viability, tissue cellularity/stroma, time course of pyknosis and lysis of the killed cells, time course of removal of cell debris, and proliferative activity and functional integrity of surviving cells. The time of appearance of gross or symptomatic change is principally a function of the cellularity and proliferation kinetics of the cellular constituents. For example, changes in bone marrow are prompt since the marrow is comprised of high numbers of actively dividing cells. In contrast, changes in mature muscle take time to develop, since there are few cells with extremely slow turnover. Radiation effects may thus become apparent in different tissues at different times, even though identical proportions of cells were killed.

The relationship between radiation dose and probability of cell kill is basic information required by the radiation oncologist. Figure 3 2783 shows a plot of survival fraction of HeLa cells versus dose of radiation. The curve is characterized by a shoulder region followed by a straight line. Similar experiments performed on cells irradiated in vivo showed survival curves were similar to those obtained in vitro.

Cell survival curves have now been generated for a large number of mammalian cell lines: the shape is generally close to that shown in Fig. 3 2783.

The multitarget single-hit model is often used to describe cell survival curves. This can be expressed as: Equation 33

where SF is survival fraction, and D is the radiation dose.

The curve is a straight line at survival fractions of less than 0.1. The slope of this straight line is the reciprocal of D&sub0;, which is the dose that reduces the survival fraction to 0.37 (i.e. kills 0.63 of the cells). The range of D&sub0; values for aerobic cells is 1.2 to 2.2 Gy for epithelial and mesenchymal cells and 0.7 to 1.1 Gy for bone marrow and lymphoid cells; n is 2 to 20 and 1 to 2 for these two classes of cell. n is the back extrapolation of the straight portion of the curve to the ordinate.

The dose range relevant to clinical radiation oncology is 1 to 3 Gy/fraction. Cell survival in this dose range is better described by the linear quadratic model: Equation 34

where &agr; and &bgr; are constants. &agr; represents single-hit events and is proportional to dose; this component dominates the response at low radiation doses. The &bgr; component represents interacting events, and is the major process in cell activation at high doses; the effect is proportional to dose&sub2;. For radiation administered in F equal dose fractions, the equation becomes Equation 35

The parameter &agr;/&bgr; has been widely accepted for describing the effects of dose fractionation and the responses of cells to low doses of radiation. It has a value in the range 1 to 5 for late-responding tissues, and 6 to 20 in early responding tissues and most tumours. Thus, when the effects on late responding tissue are critical to patient tolerance, it is beneficial to use small doses (1 - 2 Gy) per fraction. Gains have been demonstrated with the use of 1.2 Gy fractions rather than 2 Gy fractions in the treatment of carcinomas of the head and neck. Radiation treatment for curable patients is now based upon small doses per fraction in most centres.

Radiation sensitivity is a function of the position of the cell in the replication cycle. Although this shows some variation between cell lines, cells in the M stage are most sensitive, while those in S stage are least sensitive (Fig. 4) 2784. Cells in G1 and early S show intermediate sensitivity, although cells at the G1/S interphase are usually sensitive. The effectiveness of a radiation dose therefore depends on the distribution of cell ages and the variation of sensitivity with age. Radiation delays the progression of cells through the cell cycle: this effect is dependent on cell line, dose, and cell age at irradiation. The delay is maximal for cells in G2; the result is a rapid decline in the mitotic index, as cells accumulate in G2 but cannot enter M. After a period of time which varies with cell line and radiation dose, surviving cells begin to progress through the cell cycle, and the mitotic index returns to normal.

Design of dose fractionation regimens needs to take into account the proliferation of cells during the inter-treatment interval: this serves to increase the dose required to inactivate the tumour. This proliferation has prompted clinical trials of accelerated treatment protocols, in which the patient receives two or three fractions of 1.0 to 1.8 Gy in one day. Tumours with evidence of rapid cell proliferation are candidates for study of accelerated treatment protocols.

The first quantitative description of repair of radiation-induced damage was published in 1959: prolonging the interval between radiation treatments of cell lines from 0 h to 2 h increased the survival fraction, due to repair. Further prolongation resulted in a decrease in survival fraction, reflecting the greater sensitivity of cells that survived the first dose as they progressed out of the more resistant phases of the cell cycle (Fig. 5) 2785. Some cells show an increase in radiosensitivity as more fractions are administered. This suggests that surviving cells retain some degree of radiation damage which reduces their ability to withstand subsequent exposures.

Molecular oxygen is a powerful sensitizer of mammalian cells to radiation. The dose of radiation required to produce a defined biological effect under hypoxic conditions (<1 mmHg) is approximately three times that required to achieve the same effect under aerobic conditions (≥20 mmHg). Oxygen must be present in the cells at the time of irradiation for this effect to be realized.

The clinical importance of this effect lies in the fact that the maximal oxygen enhancement ratio is obtained at physiological O&sub2; concentrations. Early studies of oxygen diffusion gradients within tissues suggested that the po&sub2; would decrease to near zero at necrotic regions of tumours. This would result in the existence of a population of viable, hypoxic and, therefore, radiation resistant, tumour cells near or at the edge of the necrotic zone. This may be a decisive factor in the failure of radiation treatment of some tumours. Measured po&sub2; values on squamous cell carcinomas of the uterine crevix and of the head and neck region have been shown in several reports to correlate with local control achieved by radiation therapy alone. There is wide variability in stage, namely, only some tumours have important fractions of their cells existing under hypoxic conditions.

The pathophysiology which results in oxygen deficiency also limits access of drugs and other antitumour agents to the tumour.

The radiation sensitivity of hypoxic cells depends on the degree and duration of hypoxia. While acutely hypoxic cells exhibit full resistance, chronically hypoxic cells may be less resistant. During fractionated radiotherapy, tissue po&sub2; often increases to the level seen in the earliest stages of tumour growth. This may be a consequence of tumour regression, with reduction in intercapillary distances, a decrease in the number of metabolically active (and therefore oxygen-consuming) cells, or decreased intratumoural pressure leading to better blood flow. These phenomena have been hypoxic mechanisms and the assessment of the fraction of hypoxic cells a subject for considerable research.

Several electron affinic compounds are also strong sensitizers of hypoxic cells.

The use of two toxic agents may kill more cells than would be expected from the effects of the two agents used singly. The effect may be simple multiplication of the two toxicities; for example, if survival fraction is 0.2 for radiation and also for chemotherapy, the survival fraction would be 0.04 for their combined use. There is for some agents, a radiation sensitization, that is, an effect substantially greater than expected by simple combination of toxicities.

Knowledge of the shape and position of the dose-response curves for tumours and the critical normal tissues is essential to the design of improved treatment strategies. Successful treatment requires that the dose - response curve for tumour control be to the left of that for damage to normal tissue. An improved treatment means that the two curves are further separated.

Two end-points commonly used in experimental radiation therapy are tumour growth delay and tumour control (absence of growth for the period of observation): the latter is clearly the goal of clinicians.

Let us consider a model tumour which is composed of M clonogens, and let us assume that these clonogens die only as a results of a direct radiation effect, and that tumour recurrence is inevitable if one or more clonogens survives. The relationship between tumour control probability (TCP), survival fraction (SF) and number of clonogens (M), is: Equation 36

From this equation, lnTCP = SF × M. LnTCP thus represents the average number of surviving cells per tumour: TCP values of 0.1, 0.37, and 0.5 correspond to 2.3, 1.0, and 0.693 surviving cells. The average number of cells surviving per tumour following a dose of radiation equivalent to the TCD&sub5;&sub0; (that which kills 50 per cent of the tumour cells) is 0.693, since in some tumours the number of surviving cells is 2, 3, 4, etc. A mean of 1 to 10 cells survive in tumours at tumour control probabilities above 0.01.

The slope of the dose-response curve is an important determinant of the gain to be expected from a specific increase in dose. Although assessment of dose-response is difficult, analyses of the limited data available suggest that the slopes obtained in clinical studies are flatter than those determined in animal systems. This response is largely due to the greater heterogeneity seen in human tumours.

The goal of dose fractionation is to employ doses per fraction and inter-treatment interval that give the greatest differential effect between tumour and normal tissue. The repair process takes 4 to 6 h; in normal tissues such as nerves, kidney, and lung the inter-treatment interval needs to be 6 to 12 h for near complete repair. One strategy is to reduce the dose given to 1.0 - 1.2 Gy per fraction such that the effects of repair are minimized. Such hyperfractionation regimens have shown significant gain in the treatment of head and neck tumours and urinary bladder tumours.

There is great interest in accelerating the radiation treatment by administering two fractions per day (6 - 8 h between fractions) so as to reduce sharply the proliferation of surviving clonogens during the inter-treatment intervals. This approach appears to have increased effectiveness of radiation against squamous cell carcinomas, especially of the head and neck region.

The radiation dose required to inactivate a population of cells increases with cell number. There is a low likelihood of radiation therapy being successful when epithelial and mesenchymal tumours are more than 5 cm in diameter, unless the dose is increased to a level associated with significant late atrophic changes.

The radiation dose needed can be reduced by surgical removal of the clinically evident tumour mass. Such surgery often removes those cells most likely to be hypoxic, and thus resistant to radiation. Radiation is then only required to inactivate the relatively small numbers of cells which are infiltrating the grossly normal tissues adjacent to the lesion. The sequelae of radical surgery and of aggressive radiation therapy can thus be avoided, with improved functional and cosmetic results by combining the two modalities.

The efficacy of this combined approach has been demonstrated in a study in mice carrying transplanted fibrosarcomas (Fig. 6) 2786. Clinical experience in management of patients with carcinoma of breast, and head and neck region, and soft tissue sarcoma indicates a clear gain in functional and cosmetic status by this combined modality approach.

The response of normal tissues to irradiation has been studied extensively in experimental animals.

An acute, or early, response to radiation is defined as that appearing within 3 months of the exposure. Late responses occur after this time. The most important difference between early and late responding tissues is in their sensitivity to different dose fractionation regimens. In the clinical situation, late-responding tissues are responsible for most treatment-related morbidity. In most situations, a clear advantage can be gained by the use of small fractions, although the duration of treatment is extended.

Early and late responding tissues also differ in their cell proliferation kinetics: stem cells of early responding tissues are capable of rapid repopulation. This results, for example, in prompt re-epithelialization of a denuded surface, and in a severe, acute reaction to treatment which quickly resolves. Stem cells in some tissues, such as periosteum and liver, only proliferate in response to damage. Some organs, such as lung, have both an early and a late component to their response to radiation.

Radiation-induced malignancy is a severe late effect: in adults the cumulative incidence at 20 years is estimated to be 0.5 per cent. It is more common following radiotherapy during childhood, and particularly following combined radiotherapy and chemotherapy, with an estimated cumulative incidence at 20 years of 10 to 20 per cent in children receiving radiation and intensive chemotherapy for Ewing's sarcoma.

Experiments in mice have shown that bone marrow cells are relatively sensitive to radiation, with a D&sub0; of 1 Gy and n of 1 to 2.5. Cell survival curves have also been established for a series of normal early-responding tissues, including skin, crypt cells of the jejunum, stomach, colon, and testis.

Histopathologically, two major classes of radiation-induced damage to the lung are pneumonitis and fibrosis. Radiation induces changes in tissue levels of hydroxyproline and surfactant, functional changes in breathing rate and gas exchange capacity, and, at very high levels, death. Early studies demonstrated the lethality of irradiation of the thorax in experimental animals due to pulmonary and oesophageal damage. Oesophageal stricture and perforation occurred within 20 days of a single radiation exposure, while death due to lung damage occurred late, within about 160 days. Measurement of functional changes, such as breathing frequency and amplitude, have been used to examine fractionation sensitivity, LET, and the effects of concomitant chemotherapy on lung function. Increasing the radiation dose from 10 to 20 Gy increases respiratory rate and reduces amplitude. Such changes are evident before the histological development of pneumonitis. Biochemical assays, such as measurement of collagen accumulation or hydroxyproline or surfactant level, demonstrates a dose-response within a limited dose range.

Irradiation of the kidneys produces degeneration, atrophy, and necrosis of tubules; the glomeruli exhibit thickening of the basement membrane followed by sclerosis and necrosis. The primacy of tubular and glomerular damage in radiation nephritis is uncertain. Histologically, tubular damage correlates with radiation dose, while glomerular damage is more variable in degree and frequency. Microcolonies of tubule cells are detectable within 6 months of irradiation. The radiation dose required to induce tubular collapse decreases with time, indicating the progressive nature of the tubular damage.

Functional assays that have been used to assess the response of the kidney to irradiation include glomerular filtration rate (indicating loss of nephron function), effective renal plasma flow (indicating loss of tubule function), &sup8;&sup6;Rb uptake, and [&sup5;¹Cr] EDTA excretion (urinary frequency). Radiation nephritis is progressive: LD&sub5;&sub0; values at 6, 12, and 16 months are 23.6, 17.9, and 12.8 Gy, respectively.

Histological examination shows that interstitial infiltration by chronic inflammatory cells is followed by interstitial fibrosis. Although no reports suggest the presence of early and late responses, as are seen in the lung, the early reaction progresses to chronic renal damage. However, the microcolony formation by tubular cells is indicative of repopulation, and tubular damage is not progressive beyond 6 months after radiation exposure.

Myelitis is a dose-limiting, severe, late complication of radiation therapy. Its pathogenesis has been studied in a variety of animal models. Irradiation of a few segments of the spinal cord of animals causes loss of reflexes of the lower extremities and the development of dose-dependent paresis-paralysis. Histological changes include demyelination and necrosis of nerve roots, white matter necrosis, and vascular damage, specifically in the arterioles. This suggests three possible targets: oligodendrocytes, vascular endothelial cells, and Schwann cells. After a single radiation dose of 25 Gy, damage becomes apparent after a latent period of about 4 months. The damaging effects of smaller doses become apparent after longer latent periods.

The dose of radiation required to induce paralysis in 50 per cent of treated animals (MD&sub5;&sub0;) depends on the length of spinal cord irradiated. MD&sub5;&sub0; values 6 months after treatment were 59 and 74 Gy for irradiated segments 12 and 6 mm long, respectively. This effect is equivalent to the volume effect which has been observed in virtually all normal tissues.

In general, the smaller the radiation field, the higher the dose required to cause a given degree of change. The model proposed to explain this volume effect proposes that normal tissues can be divided into functional subunits. One functional subunit is defined as the largest structural or operational unit of cells capable of being regenerated from a single surviving clonogen. For the kidney, the functional subunit is the nephron, each of which is a structural entity independent of its neighbours. In tissues such as skin, functional subunits cannot be defined structurally, since clonogens are arranged in sheets. In this type of issue, the definition is operational: a functional subunit is the maximum area or volume that can be restored from a single clonogen. The functional subunits may be arranged in parallel (as in skin) or in series (as in the spinal cord). Elimination of a functional subunit has a greater effect when they are arranged in series than when they are in parallel, resulting in a greater volume effect.

Some tissues show biphasic responses to irradiation. Irradiated skin develops epilation and dry or wet desquamation, usually within 30 days of irradiation. These reactions subside, with the exception of epilation, and late reactions, characterized by fibrosis, necrosis, and/or telangiectasia occur after 3 months or more. The reaction seen in the mouse foot following a single dose of radiation is shown in Fig. 8 2788. The foot reaction becomes visible macroscopically after 10 to 13 days, reaching a maximum at days 20 to 23. Reduction in the reaction follows, and the late reaction starts to appear at about 200 days. The early reaction is due to damage to the skin epithelial cells and hair follicles and to blood vessels. Tissue oedema, a late reaction caused by increased exudate from blood or lymph vessels, is a cause of tissue atrophy and necrosis. Tissue fibrosis is also a consequence of early reactions.

Brahme A. Dosimetric precision requirements in radiation therapy. Acta Radiol Oncol, 1984; 23: 379 - 91.

Brock WA, Baker FL, Geara F, Goepfert H, Peters LJ. In vitro radiosensitivity of tumor cells and local tumor control by radiotherapy (abstract). In: JD Chapman, GF Dewey, GF Whitmore, eds. Radiation Research. A Twentieth Century Perspective. Congress Abstracts, Vol. 1. New York: Academic Press, 1991: 379 - 91.

Brown JM. Evidence for acutely hypoxic cells in mouse tumours and a possible mechanism of reoxygenation. Br J Radiol, 1979; 52: 650 - 6.

Elkind MM, Sutton H. X-ray damage and recovery in mammalian cells inculture. Nature 1959; 184: 1293 - 5.

Hendry JH, Thames HD, eds. Fractionation in Radiotherapy. London: Taylor & Francis, 1987.

Gatenby RA, et al. Oxygen distribution in squamous cell carcinoma metastases and its relationship to outcome of radiation therapy. Int J Radiat Oncol Biol Phys, 1988; 14: 831 - 8.

Goffinet DR, Choi KY, Brown JM. Combined effects of hyperthermia and ionizing radiation on the adult mouse spinal cord. Radiation Res, 1977; 72: 238 - 45.

Gray LH. The concentration of oxygen dissolved in tissue at the time of radiation as a factor in radiotherapy. Br J Radiol, 1953; 26: 638 - 48.

Hall EJ. Radiobiology for the Radiologist, 3rd edn. Philadelphia: JB Lippincott Co, 1988.

Hewitt HB, Wilson CW. A survival curve for mammalian leukemia cells irradiated in vivo. Br J Cancer, 1959; 13: 69 - 75.

Hockel M, et al. Intratumoral pO&sub2; predicts survival in advanced cancer of the uterine cervix. Radioth Oncol, 1993, xx

Horiot JC, et al. Hyperfractionation versus conventional fractionation in oropharyngeal carcinoma: final analysis of a randomized trial of the EORTC cooperative group of radiotherapy. Radioth Oncol, 1992; 25: 231 - 41.

Horwich A. Combined Radiotherapy and Chemotherapy in Clinical Oncology. London: Edward Arnold, 1992.

Kolstad P. Intercapillary distance, oxygen tension and local recurrence in cervix cancer. Scand J Clin Lab Invest, 1968; 106: 145 - 57.

Little JB. Repair of sub-lethal and potentially lethal radiation damage in plateau phase cultures of human cells. Nature, 1969; 225: 804 - 6.

McKenna WG, et al. The role of the H-ras oncogene in radiation resistance and metastasis. Int J Radiat Oncol Biol Phys, 1990; 18: 849 - 59.

Malaise EP, Guichard M, Siemann DW, eds. Chemical modifiers of cancer treatment. Int J Radiat Oncol Biol Phys, 1989; 16:

Powell S, McMillan TJ. DNA damage and repair following treatment with ionizing radiation. Radiother Oncol, 1990; 19: 95 - 108.

Puck TT, Marcus PI. A rapid method for viable cell titration and clone production with HeLa cells in tissue culture: The use of x-irradiated cells to supply conditioning factors. Proc Natl Acad Sci USA, 1955; 41: 432 - 7.

Puck TT, Marcus PI. Action of x-rays on mammalian cells. J Exp Med 1956: 105: 653 - 66.

Puck TT, et al. Action of x-rays on mammalian cells. II. Survival curves of cells from normal human tissues. J Exp Med, 1957; 106: 485 - 500.

Sinclair WK. X-ray induced heritable damage (small-colony formation) in cultured mammalian cells. Radiat Res, 1964; 21: 584 - 611.

Sinclair WK. Cyclic x-ray responses in mammalian cells in vitro. Radiat Res, 1968; 33: 620 - 43.

Suit HD, Shalek RJ, Wette R. Radiation response of C3H mouse mammary carcinoma evaluated in terms of cellular radiation sensitivity. In Shalek RJ, et al, eds. Cellular Radiation Biology, Baltimore: Williams & Wilkins Co. 1965: 514 - 30.

Thomlinson RH, Gray LH. Histological structure of some human lung cancers and possible implications for radiotherapy. Br J Cancer, 1955; 9: 539 - 49.

Till JE, McCulloch EA. A direct measurement of the radiation sensitivity of normal mouse bone marrow cells. Radiat Res, 1961; 14: 213 - 22.

Todoroki T, Suit HD. Therapeutic advantage in pre-operative single dose radiation combined with conservative and radical surgery in different size murine fibrosarcomas. J Surg Oncol, 1985; 29: 207 - 15.

Urano M, Kenton LA, Kahn J. The effect of hyperthermia on the early and late appearing mouse foot reactions and on the radiation carcinogenesis: Effect on the early and late appearing reaction. Int J Radiat Oncol Biol Phys, 1988; 15: 159 - 166.

West CM, Davidson SE, Hendry HJ, Hunter RD. Prediction of cervical carcinoma response to radiation therapy. Lancet, 1991; 338: 818.

Withers HR. Biologic basis for altered fractionation schemes. Cancer 1985; 55: 2086 - 95.

Withers HR, Taylor JMG, Maciejewski B. Treatment volume and tissue tolerance. Int J Radiat Oncol Biol Phys, 1988; 14: 751 - 9.