2
Radiobiologic Concepts for Brachytherapy
Alexandra J. Stewart, Robert A. Cormack, and Kathryn D. Held
Radiobiologic principles are vitally important in the daily use of brachytherapy. Brachytherapy regimes were initially developed empirically with doses determined predominantly by clinical effect and maximum tolerated toxicity. Radiobiologic modeling has allowed prediction of the biological effect of varying dose prescriptions on the tumor and surrounding normal tissue and thus more accurate estimations of probabilities of cure or tissue toxicity.
Brachytherapy has been described as the first form of conformal radiation therapy (1). Radioactive sources are placed within or very close to the cancer, allowing a high cancer-to-normal tissue dose ratio. It must be remembered that accurate source placement remains the single most important factor in brachytherapy such that, in an implant with poor geometry, changing radiobiologic parameters will not improve the clinical outcome (2). The importance of radiobiology and its use within brachytherapy was emphasized by the move from low dose rate (LDR) treatment to fractionated high dose rate (HDR) treatment and is further emphasized by the rise in popularity of pulsed brachytherapy (PB; also known as pulsed dose rate [PDR] brachytherapy).
To introduce the reader to radiobiology and its relationship to brachytherapy, a number of common clinical questions are posed and the answers illustrated using evidence from laboratory work and clinical trials. At the end of the chapter, several worked clinical examples illustrate how radiobiology can be practically applied.
HOW IS BRACHYTHERAPY DIFFERENT FROM EXTERNAL BEAM RADIOTHERAPY?
In external beam radiotherapy (EBRT), a large volume is treated with a relatively homogeneous distribution of dose such that deviations of dose within the volume typically range from 95% to 107% of the prescribed dose (3). In contrast, brachytherapy treats a smaller volume with an extremely heterogeneous dose distribution. The average dose within the target volume is usually far higher than the prescribed dose at the reference isodose on the periphery of the implant. This is tolerated due to the volume–effect relationship: very small normal tissue volumes (eg 1–2 cm3) can tolerate very high doses that larger volumes would not tolerate. There are a few exceptions to this such as spinal cord, though in paraspinal very low dose rate (vLDR) brachytherapy point doses of up to 167.3 Gy to the cord have been described with no subsequent myelitis (4,5).
IS DOSE HETEROGENEITY GOOD OR BAD IN BRACHYTHERAPY?
This is a very interesting question because it depends on the treatment site. In brachytherapy there is a rapid falloff in dose as distance from the source increases, illustrating the inverse square law. With many sources or dwell positions in an implant, the dose throughout an implant may vary widely from the prescription isodose. Therefore, the concept of equivalent uniform dose (EUD) was introduced (6). This involves calculation of the equivalent average dose enclosed within the target volume. In the absence of a spacing device, the farther from the source the prescription isodose is, the higher the EUD will be. The EUD is also higher for single-line sources and lower numbers of dwell positions (7).
Another brachytherapy concept, which is also utilized in EBRT planning, is the dose homogeneity index (DHI). This is calculated as follows:
DHI = V100 − V150/V100
where V100 and V150 are the tissue volumes receiving 100% and 150%, respectively, of the prescribed dose. The higher the DHI, the more uniform the dose distribution within an implant.
Heterogeneity is favored in some situations; in contrast, homogeneity is favored in others, for example, heterogeneity is extremely important in a tandem and ovoid cervix implant. Planning studies using intensity modulated EBRT (IMRT) have shown that delivering a homogeneous dose across the target volume will not result in the ultra-high doses delivered by a tandem and ovoid implant; therefore, an equivalent biological dose cannot be delivered and control will be compromised (8). The doses to organs at risk were also higher for the IMRT plans, which would be predicted to result in increased long-term toxicity. In addition it was very difficult and time consuming when using IMRT to attempt to recreate the dose heterogeneity automatically achieved with brachytherapy.
In contrast, in an interstitial breast implant, an improved cosmetic outcome relies on dose homogeneity and heterogeneity is discouraged. A DHI of more than 0.75 is recommended, with a value more than 0.85 being ideal. The toxicity of multicatheter interstitial implants is significantly lower when the DHI within the target volume is higher, that is, the dose is more homogeneous (9,10). The volume of the individual high-dose regions also appears to be important, with large volumes receiving more than 150% and 200% of the prescribed dose being associated with increased rates of fat necrosis and poorer cosmesis (10).
It is important to consider radiobiologic factors when converting from one treatment modality to another. The linear quadratic equation (see later) was used to formulate a dose for multicatheter interstitial HDR for partial breast brachytherapy. A dose of 34 Gy in 10 fractions was calculated as equivalent to 50 Gy in 25 fractions EBRT. For interstitial brachytherapy with a high DHI, the EUD will be very close to the prescribed dose. This dose/fractionation scheme was then used for the single-line source balloon MammoSite catheter (Hologic, Bedford, MA, USA). An examination of the EUD of the MammoSite catheter demonstrates that the EUD is higher than the dose prescribed at the reference isodose and that this effect increases as balloon diameter decreases and with a decreasing number of dwell positions (7). In a small series, the increased EUD was seen to correlate with increased skin hyperpigmentation and telangiectasia. This underlines the importance of considering all radiobiologic factors when implementing different brachytherapy techniques.
WHAT IS DOSE RATE?
Dose rate is a measure of the speed at which a patient is exposed to radiation; it is measured in units of dose per unit time. Three categories of brachytherapy dose rate were defined in the International Commission on Radiation Units and Measurements (ICRU) 38 report (11):
• LDR—Range: 0.4 to 2 Gy/hr. In clinical practice, the usual range is 0.4 to 1 Gy/hr.
• Medium dose rate (MDR)—Range: 2 to 12 Gy/hr. Very rarely used in modern practice.
• HDR—More than 12 Gy/hr.
Permanent seed implants are often termed vLDR as they deliver a high total dose at a very low dose rate, often less than 0.4 Gy/hr. LDR brachytherapy is becoming much less common with the rise of PB. The term PDR is incorrect because the dose rate does not pulse and is in fact quite high (often more than 12 Gy/hr), but due to the short pulse times, the dose delivery overall mimics LDR. PB was developed in an effort to simulate the radiobiologic properties of LDR, but with the advantages of staff radioprotection from remote afterloading and dose optimization from use of a stepping source. A large number of small fractions (pulses) are administered in the same overall time taken for an LDR implant; it could also be termed ultra-fractionated HDR treatment. Hourly pulses are closest in radiobiologic effect to LDR with total dose corrections required as the interval between pulses increase.
DOES DOSE RATE MATTER?
Dose rate is important in LDR and, thus, in the conversion of LDR doses to HDR and PB regimens (12). A randomized study of LDR implants for cervical carcinoma with dose rates of 0.4 versus 0.8 Gy/hr showed a significant increase in late complications in the higher dose rate group (45% vs 30%) with no difference in overall survival or local control (13,14). Similar findings were seen in head and neck implants with an increase in necrosis from 12% at 0.3 to 0.6 Gy/hr to 29% at 0.6 to 1 Gy/hr and no significant change in the tumor control rate at 70 Gy (15). When the dose was decreased to 60 Gy, there was a significant decrease in tumor control at the lower dose rate (66% vs 91%) (15). In contrast, Pierquin found no difference in necrosis or control in interstitial implants with a total dose of 70 Gy and a dose rate ranging from 0.3 to 1 Gy/hr (16).
Therefore, it is felt that when using LDR (and thus probably PB), the dose rate (or equivalent dose rate) should be in the range of 0.3 to 1 Gy/hr, more due to the effects on late complications than on local control. If the dose rate exceeds 1 Gy/hr, a reduction in the total dose should be considered and can be calculated using the biological effective dose (BED) concept as follows. When converting LDR doses to HDR the dose rate effect must be taken into account, but other variables such as treatment time and interval between fractions must also be considered. It was previously suggested that a dose reduction factor should be used. This oversimplifies the dose conversion and it is preferable to carefully consider all radiobiologic factors before converting the dose.
WHAT ARE α/β RATIOS AND WHERE DO THEY COME FROM?
The most commonly used radiobiologic model to relate biological effect, often taken to be cell surviving fraction (SF), to dose is the linear quadratic equation (17,18):
SF = exp – (αD + βD2)
where the α component represents a single ionizing radiation event that produces damage that is not repairable and increases in a linear pattern with dose; thus, it is influenced by overall dose rather than fractionation. The β component represents damage caused by two sublethal ionizing events that can combine to form a lethal event. This damage is potentially repairable and increases in a quadratic pattern. It is influenced by fractionation and dose rate as well as by overall dose. The α/β ratio is a measure of how a tissue will respond to a change in total dose, fractionation, or dose rate; it is also termed fraction sensitivity. For early-reacting normal tissues such as bowel mucosa, which express damage from radiotherapy in days to weeks after irradiation, the α/β ratio is high, for example, 10 to 20 Gy. For late reacting normal tissues, such as spinal cord, which express damage from radiotherapy in the months to years following irradiation, the α/β ratio is low, for example, 1 to 6 Gy (19).
α/β ratios can be determined using outcome data from clinical trials. This is demonstrated from EBRT data in breast (20,21) and prostate cancer (22). There are fewer data from brachytherapy studies, possibly because there are fewer of them, and also because many involve a combination of EBRT and brachytherapy.
WHAT IS BED AND HOW DOES IT MATTER?
Although some older papers used total nominal dose when reporting radiotherapy technique, this method is not optimal, as it does not account for the effects of fractionation in brachytherapy treatment. The BED is a method of calculation of the isoeffective consequence of different fractionation schedules using the linear quadratic principle and as such is a measure of the probable efficacy of a course of radiation.
The BED for a fractionated HDR treatment involving N fractions, each of dose d, is given as:
It employs the individual α/β ratio of a tissue or tumor and therefore the radiobiologic effects of a treatment course can be calculated for each different tissue type. The α/β ratio used to calculate a particular BED equation is placed in subscript after the dose units, for example, using an α/β ratio of 3, results are reported in Gy3. The BED equation can also be used in combination with damage threshold effects to predict the risk of damage in tissues, for example, in carcinoma of the cervix it was seen that severe rectal complications occurred after a threshold BED of 125 Gy3 was delivered to a rectal reference point and that complications occurred incrementally at an approximate rate of 1% per additional Gy3 (23). The standard BED equation can be modified to take into account other treatment factors such as overall time, incomplete repair and repair rates in tissues or tumors, dose rates, and so on.
The equivalent dose in 2 Gy fractions (EQD2), given by the EQD2 equation is used to compare cumulative BED results.
DOES FRACTIONATION MATTER?
Stitt et al (25) showed that the probability of late damage increases as the number of fractions of HDR decreases (ie, the dose per fraction increases). This is also related to the percentage of dose that the normal tissue receives. In intracavitary brachytherapy for cervix cancer if the normal tissue were to receive 100% of the dose, 30 low dose fractions would be needed for LDR late complication equivalence. If normal tissue received 90%, 12 to 16 medium dose fractions would be needed and if it received 80%, four to six higher dose fractions could be used. Hama et al (26) showed that there were increased late complications if four fractions or less were used. However, Patel et al (27) used 18 Gy in two fractions and still showed a very low rate of late complications. With more accurate dose estimation using image-guided brachytherapy it may be possible to decrease the number of fractions or increase the treatment dose per fraction if it can be ascertained that normal tissues are receiving clinically tolerated proportions of the treatment dose.
WHY DO WE USE DIFFERENT RADIONUCLIDES IN DIFFERENT SITUATIONS?
HDR is generally administered using the radionuclide iridium-192, which has been chosen for its physical properties, particularly half-life and specific activity. However, for vLDR brachytherapy, a variety of radionuclides are available and their radiobiologic properties can be used to determine which may be most appropriate in different clinical situations.
To illustrate this, interstitial seed placement for permanent thoracic implants is presented as an example; see Table 2.1. Iodine (125I) seeds have traditionally been used, predominantly due to the ease of availability. A dose of 80 to 120 Gy at 0.5 to 1 cm from the implant has been used with success and minimal toxicity (28−31). An alternative with a shorter half-life and thus higher dose rate may be preferable. Use of palladium (103Pd) has been described (32,33) but the lack of availability meant that its use has generally been superseded by cesium (131Cs) (34). As 131Cs is a new source to be used in clinical practice, radiobiologic calculations were undertaken to determine how it should be used. The 131Cs prescribed dose determination was based on the linear-quadratic formulation with assumptions regarding the α/β ratios for late responding lung tissue, tissue repair constant, average tumor doubling time, repopulation rates and so on. If an α/β for lung cancer of less than 5 is assumed, and given the relative effectiveness of 125I in lung cancer cell killing in the literature, then a prescribed dose of 60 to 80 Gy (60−70 Gy used for lesions less than 1 cm) is assumed to be reasonable especially because of the significantly higher dose rate (35).
Table 2.1 The physical properties of radionuclides commonly used in interstitial thoracic brachytherapy
WHAT DIFFERENCE DOES IT MAKE TO PRESCRIBE AT DIFFERENT DEPTHS?
The choice of prescription point is very important in brachytherapy. Due to the sharp reduction of dose with distance, the physical dose depends not only on the prescription point but also on the dose rate. As distance from the source increases, not only does the dose decrease but also the dose rate; see Figure 2.1. Therefore, a change from a prescription point that receives 10 Gy at 1.5 Gy per hour to a point that receives 10 Gy at 0.5 Gy per hour results in a large increase in SF. Therefore, a fall of dose and dose rate causes a larger reduction in cell kill than a reduced dose or dose rate used in isolation. To illustrate this using vaginal cylinder brachytherapy, if 5 Gy per fraction is prescribed at the cylinder surface, to achieve the same biologic effect, a higher dose will have to be prescribed at 0.5 cm from the surface, for example, 6 Gy; see Case 2.8.
This concept can also be used to protect normal tissues outside the target volume. As mentioned earlier, the EUD is also important and choice of a prescription point farther from the source may result in a higher EUD and larger volumes of high dose areas within the target, which could be significant for late toxicity (7). This could be balanced with a higher dose of homogeneity at the prescription point, achieved due to less rapid dose falloff with distance (the Inverse Square Law).
Figure 2.1 Survival curve plot for α = 0.15 Gy–1 and β = 0.05 Gy–2 for a treatment time of 24 hours at various dose rates. Note that for a treatment that gave 10 Gy at 1.5 per hour (*) to a very small volume and if a larger volume received 0.5 Gy per hour (*) to a dose of 5 Gy, there is a striking difference in the survival fraction.
WHAT ARE THE FOUR Rs AND HOW DO WE/SHOULD WE CAPITALIZE ON THEM?
Factors contributing to the response of tissues to radiotherapy have been labeled the four Rs (40). These are repair, repopulation, reoxygenation, and reassortment. A fifth R could be added—radiosensitivity—though this is an integral property of any cell and affects radiotherapy response but cannot be altered. The addition of exogenous radiosensitizers has not been widely studied in brachytherapy.
Repair
Sublethally damaged cells are capable of repair if they are allowed sufficient time. If they are exposed to further irradiation before repair is complete, the sublethal damage may become lethal. The lower the radiation dose rate to which a cell is exposed, the more likely it is that repair will occur in that cell. Late reacting normal tissues seem more capable of repair than many tumor cells so, at a given fractionated therapeutic dose, tumor is preferentially killed over normal tissue (41).
The time course of LDR treatment over several days allows for sublethal damage repair. The short treatment time of HDR treatment prohibits this repair during the actual irradiation. However, if an interval between HDR fractions of more than 6 hours is maintained, full repair of normal tissues can occur (41). In order for HDR to be equivalent to LDR for tumor cell–kill effect, the dose per fraction should be low with multiple fractions given; in essence, this is what PDR is. However, this is not clinically practical for HDR. For example when treating carcinoma of the cervix, most centers use four to six fractions of HDR brachytherapy and have similar survival and complication rates to LDR. This effect may be explained using the repair half-life.
Orton (41) has theorized that the repair half-life of late responding normal tissue in the cervix is longer than the 1 to 1.5 hour estimates proposed by other investigators (42,43). If the repair half-life were 1.5 hours, an HDR dose of 2 to 3 Gy per fraction would be equivalent to LDR at 0.5 Gy/hr. In contrast, if it were 4 hours, HDR doses of 5 to 12 Gy per fraction would be equivalent. The latter matches current practice more closely. The longer repair half-life would reduce the sublethal damage repair estimates of LDR, making HDR superior for preventing late normal tissue complications. Of course, repair may not simply be a function of time and may have fast and slow components (44).
Repopulation
During a course of radiation therapy, repopulation in late responding normal tissues generally does not occur, but in early responding normal tissues, repopulation can start within 2 to 3 weeks and increase the tissue tolerance. Likewise, in many tumors, especially rapidly growing ones, tumor cell repopulation or accelerated repopulation begins within a few weeks of the start of radiation exposure and may necessitate an increase in dose, if the radiation treatment time is prolonged. For example, in carcinoma of the cervix, studies have shown improved tumor control and increased survival when radiotherapy is given in the shortest overall time (45,46). This is because shorter treatment times allow less time for accelerated repopulation to occur. The continuous administration of LDR prevents repopulation during treatment. Studies have shown that the use of HDR at the end of a radiotherapy regime may result in increased overall treatment time. Okkan et al (47) showed that the average time to complete treatment when HDR was used was 70 days, compared to 57 days when using LDR. This may decrease the chances of tumor control.
Chen et al (48) showed that, when treating cervix cancer with HDR brachytherapy, if treatment were prolonged more than 63 days there was a significant decrease in disease-free survival from 83% to 65% (P = .004) and in local control from 93% to 83% (P = .02). However, the limit of 63 days would be felt by many investigators to be too long as most studies have used a maximum of 55 days to assess the effects of prolonged treatment time (45,46). Importantly, no difference in late complications was seen in the less than 63 day treatment group, suggesting that there is no morbidity benefit in extending overall treatment time. The fractionated nature of HDR allows for integration of the brachytherapy within the EBRT schedule, allowing shorter overall treatment times. Adequate EBRT should be given to allow tumor shrinkage, as a retrospective study has shown decreased control when weekly brachytherapy was used in bulky tumors from week 1 of EBRT (49). It should be noted that this study was in the era of point A-based prescribing before image-guided target definition so that control would probably be better now as tumor coverage improves. However, if the tumor is large at the inception of brachytherapy, normal tissue toxicity will be higher as larger volumes of organs-at-risk will inevitably be treated, leading to a recommendation of starting brachytherapy after 20 fractions of EBRT in patients with bulky tumors.
Reoxygenation
In carcinoma of the cervix, the effect of hypoxia on tumor control has been well documented with decreased survival in patients with a low initial hemoglobin level (50,51). Acute hypoxia is secondary to constriction of capillaries within the irradiated field and takes approximately 8 hours to pass. Chronic hypoxia is due to growth of tumor beyond the capacity of the tumor to form new vasculature and corrects as the tumor shrinks and oxygen diffuses in from surrounding vessels; this takes days to weeks to occur. Due to the length of administration of LDR, time may allow for acute hypoxia to correct within the tumor during treatment. With HDR treatment the tumor may shrink between insertions, allowing for areas of chronic hypoxia to be reoxygenated. LDR has a lower oxygen enhancement ratio than HDR (52).
Reassortment
There is a theoretical advantage of an improved effect due to tumor cell cycle reassortment using LDR treatment as cells will pass out of the relatively radio-resistant phases of late S and early G1 into the more radiosensitive phases of G2 and M during the overall treatment time. However, in practice, the effect of reassortment has not been shown to play any demonstrable role in clinical radiotherapy.
Hopefully, these common questions have addressed many of the day-to-day radiobiologic issues that you encounter regularly. We now present a series of case vignettes—clinical scenarios that aim to demonstrate these theories in practice.
