British Journal of Radiology 74 (2001),262-265 © 2001 The British Institute of Radiology
Variation in the probability of cardiac complications with radiation technique in early breast cancer
P A Canney, MD, FRCR
1
R Sanderson, MSc
2
C Deehan, PhD
2 and
T Wheldon, PhD
3
1 Departments of Radiation Oncology
2 Clinical Physics, Beatson Oncology Centre, Western Infirmary, Dumbarton Road, Glasgow G11 6NT
3 Department of Radiation Oncology, Glasgow University, CRC Beatson Laboratories, Glasgow G61 1BD, UK
 |
Abstract
|
|---|
Cardiac damage is recognized to be a potentially serious side effect of breast cancer radiotherapy, the risk of which may be reduced by the choice of appropriate radiotherapy technique. We have previously described variation in physical dose to the heart dependent upon radiotherapy technique. In this paper we report the calculated improvement in normal tissue complication probability (NTCP) (for cardiac damage) achievable by these methods. Cardiac doses were calculated from dose–volume histograms (DVHs) using a "Helax" planning system for 11 patients with left-sided tumours and 5 patients with right-sided tumours. The DVH reduction algorithm of Lyman and Wolbarst [1989] was applied to each DVH to produce a value for the NTCP. For left-sided tumours, mean NTCP with the standard technique was 7.4±5.6% (range 0.6–17%) and for the optimum technique mean NTCP was 0.3±0.6% (range 0–2%) (p<0.003 for the difference between the two techniques): a predicted reduction in late cardiac complications of 23-fold, which is not clearly evident from viewing the DVH raw data.
|
Introduction
|
|---|
Recent studies of radiation toxicity in the treatment of breast cancer have shown that the effects on normal tissues can constitute a significant clinical problem, particularly increased cardiac mortality, and this may offset any potential survival benefit of treatment [1–5]. As a result of these findings, it has been stressed that radiotherapy techniques used in breast cancer should be designed to minimize cardiac dose [4, 5]. A method of reducing the physical cardiac dose by altering patient positioning during post-operative radiotherapy for breast cancer has previously been described [6]. However, the reduction in physical cardiac dose does not give any indication of the scale of reduction in induced radiation cardiac damage and therefore of the potential clinical benefit that might be expected. The dose–volume histogram (DVH) reduction algorithm of Lyman and Wolbarst [7, 8] was applied to these cardiac radiation dose data to produce a value for the normal tissue complication probability (NTCP) for two different radiotherapy methods, our standard technique and the technique using our new patient positioning.
|
Methods
|
|---|
Reducing the physical radiation dose to the heart
Our standard tangential breast technique has been described previously, and has been compared with an alternative patient treatment position that was expected to reduce the cardiac dose [6]. The standard breast technique uses an isocentric treatment incorporating medial and lateral tangential glancing breast fields. Patients are treated in the supine position, with the arm on the affected side flexed at the elbow and abducted at 90° from the body. A collimator rotation is used rather than an inclined board. A hand grasp is used so that the arm position is stable and reproducible. The clinical target volume (CTV) for tangential fields is determined clinically by the oncologist on the simulator by palpation of the breast tissue. Treatment planning is carried out on the single central slice contour taken at simulation.
The new breast technique uses the same field arrangement, but the patient lies supine on a breast board (Smithers Medical Productions Inc., Acron, OH), with the arm on the treated side placed over the head, gripping the contralateral arm of a "T" grip. The other arm is left by the side. The "T" piece is adjusted so that the patient lies flat and is not rotated during treatment. The CTV was defined by skin marks placed on the patient using exactly the same method as the standard breast technique.
CT was performed with radio-opaque markers placed on the medial field border, on the lateral field border and on the skin over the isocentre. CT scans were transferred to the planning computer where treatment beams were added. The dose to the heart was assessed by means of DVHs. Two DVHs were generated for each patient, one with the standard treatment method and one with the new method.
Application of the DVH reduction algorithm to the calculation of NTCP
There are little clinical data on heart tolerance doses. Moreover, dose–response curves for radiation-induced cardiac mortality have not been published. Emami et al [9] selected pericarditis as a surrogate end-point for radiation-induced heart disease and these data were used to estimate the constants required by the Lyman and Wolbarst model.
The DVH reduction algorithm of Lyman and Wolbarst [7, 8] was applied to both of the DVHs produced for each patient. The Lyman formula for NTCP assumes that the NTCP is a function of both the fractional volume irradiated and the absorbed dose received by this volume. It was assumed that the volume dependence of the complication probability could be represented by a power-law relationship:
where TD(V) is the tolerance dose for a given partial volume V, TD(1) is the tolerance dose for the full volume and n is a fitted parameter.
The NTCPs thus derived for the two techniques could then be compared for each patient individually to evaluate the benefit of the new technique.
|
Results
|
|---|
16 patients were studied in total. 11 had the left breast irradiated, 5 following mastectomy and 6 following conservative surgery; and 5 had the right breast irradiated, 2 following mastectomy and 3 following conservative surgery.
Reduction in physical dose
These results have been reported in detail elsewhere [6]. Briefly, the reduction in mean cardiac dose on the left side obtained by changing from the standard position to the new position was 60% (p<0.001) and the reduction in maximum dose was 32% (p<0.001). The percentage of cardiac volume included within the 50% isodose contour was 12.3% (SD±7.3) for the standard technique and 2.8% (SD±3.8) for the new technique. There was considerable interpatient variation in the absolute mean, maximum and median cardiac doses with both treatment methods, but the new treatment method resulted in a reduction in all cases (Figure 1
).
View larger version (5K):
[in this window]
[in a new window]
|
Figure 1. Sample dose–volume histograms for a patient with (a) the standard technique and (b) the alternative technique with the new positioning.
|
|
Changes in NTCP
Results for each patient are shown in Table 1. For the left-sided breast treatments the mean NTCP with the standard technique was 7.4±5.6% (range 0.6–17%); for the new technique the mean NTCP was 0.3±0.6% (range 0–2%). A Wilcoxon matched-pairs signed-ranks test for difference in NTCP was performed and the difference between the two techniques was significant (p<0.003).
View this table:
[in this window]
[in a new window]
|
Table 1. Differences in normal tissue complication probability (NTCP) for each patient for left-sided treatments only
|
|
For all patients having the right breast treated, the NTCP with the standard technique was <0.1%; for the new technique the NTCP was not significantly different from zero.
|
Discussion
|
|---|
There are several deficiencies in the methods used to calculate the probability of late cardiac damage. First, the Lyman model itself does not adequately reflect known radiobiological theory, as the radiation response is highly dependent on the radiation protocol employed [10], and the exponents TD(V) and n, once derived from experimental data, apply to that protocol only and should not be extrapolated to other radiation schedules. Second, a surrogate end-point of pericarditis [9] rather than late mortality was used to derive the exponents. Third, the method assumes a functionally homogeneous organ; the level of dose to sensitive structures within the myocardium could be more significant than the overall volume irradiated [2]. Fourth, the model does not discriminate between tissue damage at contiguous sites and geographically separate sites within the treated volume. Several other models have been proposed [2, 11], but all have some of the problems outlined above or require other assumptions to be made. The Lyman method was chosen as it is the simplest method and at present there is no compelling evidence to prefer any one particular method [11]. Rather than absolute values of predicted cardiac damage, we were looking for evidence of changes in NTCP between two clinical techniques employing identical dose and fractionation schedules but with differing cardiac DVHs.
The average predicted NTCP of 7.4% is consistent with the known excess of cardiac deaths reported by Rutqvist et al [5], who showed that at 15 years follow-up there was an increase in risk of cardiac death, from 3% following surgery alone to 10% if the left breast or chest wall had been irradiated. The apparent accuracy of the NTCP may be fortuitous; however, the relative benefit of the radiotherapy techniques shown by the differences in NCTP are likely to reflect a realistic estimate of the scale of improvements that may be achieved in practice. Thus, the new technique is predicted to produce a significant reduction in cardiac normal tissue complications of 23-fold, from 7.4% to 0.3%. The significance of any measurement of physical dose to the heart by comparing DVH statistics would normally be assessed by the clinician based upon past experience. This assessment is at present wholly subjective. Although lower cardiac doses would seem to be better, simple viewing of the raw DVH data cannot indicate the potential magnitude of cardiac damage and thus the clinical significance of any variation between radiotherapy techniques.
The relative seriality model has been applied to DVH data by Gagliardi et al [2] for a similar group of patients. They estimated values for predicted cardiac complication rates for differing radiotherapy techniques from mean DVHs derived from detailed studies of a small number of patients and then extrapolated to the whole population. Our study estimated the individual risk for each patient and then calculated a mean value for the individual NTCPs. We decided not to differentiate between the myocardium alone and the whole heart including the blood volume as it was not possible from our CT scans to ascertain whether the heart was in systole or diastole at the time of the planning CT. In contrast, Gagliardi et al [2] attempted to differentiate between myocardial volume and whole heart volume inclusive of the blood volume therein, which received irradiation, but gave values of NTCP for both volumes. They predicted cardiac mortality of 4.7% for the total heart volume and 6.8% for the myocardium alone. Despite the differences in methodology, the results of our studies and those of Gagliardi et al [2] are very similar and both predict a NTCP close to the historically observed level of excess cardiac deaths.
An alternative view is that as the increase in late cardiac mortality is for left-sided treatments rather than right-sided treatments [5, 12], a realistic aim would be to reduce the physical heart dose delivered by left-sided radiotherapy treatments to that received during right-sided radiotherapy treatments. We did not accomplish this, but the differences in physical dose when comparing our new technique employed for left breast treatments and the standard technique for right-sided treatments were small. Comparing the NTCP of 0.3% for left-sided treatments using the new technique with the NTCP of 0.006% for right-sided treatments using the standard technique, there was a borderline statistically significant difference between the two (p<0.047; Kruskal–Wallis test), indicating that the aim of reducing the complications for left-sided treatments to those observed for right-sided treatments is close to being achieved.
Received for publication May 17, 2000.
Revision received September 20, 2000.
Accepted for publication December 21, 2000.
|
References
|
|---|
-
Cuzick J, Stewart H, Rutqvist L, Houghton J, Edwards R, Redmond C, et al. Cause-specific mortality in long-term survivors of breast cancer who participated in trials of radiotherapy. J Clin Oncol 1994;12:447–53.[Abstract]
-
Gagliardi G, Lax I, Ottolennghi A, Rutqvist LE. Long-term cardiac mortality after radiotherapy of breast cancer—application of the relative seriality model. Br J Radiol 1996;69:839–46.[Abstract/Free Full Text]
-
Host H, Brennhovd IO, Loeb M. Postoperative radiotherapy in breast cancer—long-term results from the Oslo study. Int J Radiat Oncol Biol Phys 1986;12:727–32.[Medline]
-
Jones JM, Ribeiro GG. Mortality pattern over 34 years of breast cancer patients in a clinical trial of post-operative radiotherapy. Clin Radiol 1989;40:204[Medline]
-
Rutqvist LE, Lax I, Fornander T, Johansson H. Cardiovascular mortality in a randomised trial of adjuvant radiation therapy versus surgery alone in primary breast cancer. Int J Radiat Oncol Biol Phys 1992;22:887–96.[Medline]
-
Canney PA, Deehan C, Glegg M, Dickson J. Reducing cardiac doses in post-operative irradiation of breast cancer patients: the relative importance of patient positioning and CT scan planning. Br J Radiol 1999;72:986–93.[Abstract]
-
Lyman JT, Wolbarst AB. Optimization of radiotherapy III. A dose–volume histogram reduction algorithm. Int J Radiat Oncol Biol Phys 1987;13:103–9.[Medline]
-
Lyman JT, Wolbarst AB. Optimization of radiotherapy IV. A method of assassing complication probabilities from dose–volume histograms. Int J Radiat Oncol Biol Phys 1989;17:433–6.[Medline]
-
Emami B, Lyman J, Brown A, Coia L, Goitein M, Munzenride JE, et al. Tolerance of normal tissue to therapeutic radiation. Int J Radiat Oncol Biol Phys 1991;21:109–22.[Medline]
-
Millar WT, Canney PA. The derivation and application of the equations for fractionated continuous irradiation, including the effects of multiple incomplete repair processes: part 1—derivation of the equations. Int J Radiat Biol 1993;64:275–91.[Medline]
-
Hamilton CS, Chan LY, McElwain DLS, Denham JW. A practical evaluation of five dose–volume histogram reduction algorithms. Radiother Oncol 1992;24:251–60.[Medline]
-
Rutqvist LE, Johansson H. Mortality by laterality of primary tumour among 55,000 breast cancer patients from the Swedish Cancer Registry. Br J Cancer 1990;61:866–8.[Medline]
Top
Abstract
Introduction
