A comparison of conformal and intensity-modulated techniques for oesophageal radiotherapy

Radiotherapy and Oncology 61 (2001) 157–163 www.elsevier.com/locate/radonline

A comparison of conformal and intensity-modulated techniques for oesophageal radiotherapy Christopher M. Nutting a, James L. Bedford b,*, Vivian P. Cosgrove b,1, Diana M. Tait a, David P. Dearnaley a, Steve Webb b a

Academic Department of Radiotherapy, The Institute of Cancer Research and The Royal Marsden NHS Trust, Downs Road, Sutton, Surrey SM2 5PT, UK b Joint Department of Physics, The Institute of Cancer Research and The Royal Marsden NHS Trust, Downs Road, Sutton, Surrey SM2 5PT, UK Received 23 January 2001; received in revised form 13 June 2001; accepted 14 August 2001

Abstract Background and purpose: To investigate the potential of intensity-modulated radiotherapy (IMRT) to reduce lung irradiation in the treatment of oesophageal carcinoma with radical radiotherapy. Materials and methods: A treatment planning study was performed to compare two-phase conformal radiotherapy (CFRT) with IMRT in five patients. The CFRT plans consisted of anterior, posterior and bilateral posterior oblique fields, while the IMRT plans consisted of either nine equispaced fields (9F), or four fields (4F) with orientations equal to the CFRT plans. IMRT plans with seven, five or three equispaced fields were also investigated in one patient. Treatment plans were compared using dose–volume histograms and normal tissue complication probabilities. Results: The 9F IMRT plan was unable to improve on the homogeneity of dose to the planning target volume (PTV), compared with the CFRT plan (dose range, 16.9 ^ 4.5 (1 SD) vs. 12.4 ^ 3.9%; P ¼ 0:06). Similarly, the 9F IMRT plan was unable to reduce the mean lung dose (11.7 ^ 3.2 vs. 11.0 ^ 2.9 Gy; P ¼ 0:2). Similar results were obtained for seven, five and three equispaced fields in the single patient studied. The 4F IMRT plan provided comparable PTV dose homogeneity with the CFRT plan (11.8 ^ 3.3 vs. 12.4 ^ 3.9%; P ¼ 0:6), with reduced mean lung dose (9.5 ^ 2.3 vs 11.0 ^ 2.9 Gy; P ¼ 0:001). Conclusions: IMRT using nine equispaced fields provided no improvement over CFRT. This was because the larger number of fields in the IMRT plan distributed a low dose over the entire lung. In contrast, IMRT using four fields equal to the CFRT fields offered an improvement in lung sparing. Thus, IMRT with a few carefully chosen field directions may lead to a modest reduction in pneumonitis, or allow tumour dose escalation within the currently accepted lung toxicity. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Oesophageal carcinoma; Conformal radiotherapy; Intensity-modulated radiotherapy; Radiation pneumonitis; Dose escalation

1. Introduction The median survival after radiotherapy for carcinoma of the oesophagus is approximately 9 months. The 2-year survival rate is around 10% and the 5-year survival rate is less than 5% [24], although selected series report a 5-year survival rate of 21%, demonstrating the potential of radiotherapy as a curative modality [27]. In recent years, concomitant chemo-radiation schedules have produced encouraging results in randomized trials, with up to 25–30% of patients surviving for 5 years or more, and this is now considered as standard treatment [1,4,12,29]. Even using this more intensive approach, local disease recurrence still occurs in

* Corresponding author. 1 Present address: Department of Medical Physics, Belvoir Park Hospital, Hospital Road, Belfast BT8 8JR, UK.

approximately 40% of cases [1,4,12], and may act as a focus for subsequent metastatic disease. Radiotherapy for carcinoma of the oesophagus presents a particularly difficult treatment planning problem. The planning target volume (PTV) is central, close to the spinal cord, and is almost completely surrounded by lung, a radiosensitive organ with a relatively low radiation tolerance. It has been shown previously that conformal radiotherapy (CFRT) can successfully reduce the irradiation of lung parenchyma [10]. This reduction in lung dose is expected to facilitate tumour dose escalation, which might improve local control [2]. The therapeutic ratio for oesophageal radiotherapy can only be maintained if higher doses can be delivered without an increase in late normal tissue damage, the lung parenchyma and spinal cord being of particular concern. After chemo-radiation, a significant number of patients may survive for 5 years or more, and so the clinical consequences

0167-8140/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0167-814 0(01)00438-8

158

C.M. Nutting et al. / Radiotherapy and Oncology 61 (2001) 157–163

of normal tissue damage must be taken into account when designing new treatment techniques. Intensity-modulated radiotherapy (IMRT) is capable of producing complex dose distributions, and can therefore reduce the dose to radiosensitive organs close to the tumour [20]. In the treatment of carcinoma of the oesophagus, it is hoped that IMRT might offer the potential to improve the uniformity of tumour irradiation, and reduce the dose delivered to lung parenchyma. This study therefore investigates the impact of IMRT on treatment planning for carcinoma of the oesophagus, by comparing a standard CFRT technique to a 9F IMRT technique. The use of fewer beams is then investigated, to determine the extent to which this factor influences the dose distribution. Dose–volume histograms (DVHs) for PTV, lungs, and spinal cord are used as endpoints, as well as the normal tissue complication probability (NTCP) for lung. 2. Materials and methods 2.1. Patients Five patients with oesophageal carcinoma, who had been recently treated with chemo-radiation at the Royal Marsden NHS Trust, were studied. In accord with the local protocol [10], patients received a prolonged venous infusion of 5fluorouracil (5-FU) at 300 mg/m 2 per day for 12 weeks and then 200 mg/m 2 per day for the following 6 weeks. Three-dimensional CFRT, consisting of 55 Gy in 30 daily fractions over 6 weeks, commenced at week 12. Five courses of Cisplatin at 60 mg/m 2 were administered at 3weekly intervals from the commencement of 5-FU, so that the final dose of Cisplatin was given during the first week of radiotherapy. Patients were CT-scanned in a supine position, with images taken at 10–20 mm intervals throughout the entire thorax, including the whole lung volume to allow for calculation of lung DVHs. The clinical target volume (CTV), spinal cord, and lung parenchyma were outlined on each image. The CTV included the oesophageal tumour, with a margin for microscopic tumour extension, and the adjacent lymph nodes. A three-dimensional margin of 15 mm was added to the CTV to account for movement and uncertainty in target definition, creating the PTV. In three patients, the PTV included the spinal cord, so the PTV was edited around the vertebral bodies to exclude the cord, but this did not interfere with the anticipated sub-clinical spread of the existing disease. 2.2. Treatment objectives The aims of planning were to deliver 55 Gy (^10%) to the PTV, to maintain the maximum spinal cord dose at less than or equal to 45 Gy, and minimize the dose delivered to the lungs. The maximum dose of 45 Gy to the spinal cord was considered to be an appropriate clinical tolerance level

for this ‘in-series’ organ in the setting of chemo-radiation. Traditionally, the lung has been considered as an ‘in-parallel’ critical structure, in which the functional sub-units are independent, but there is some controversy as to which model best predicts pneumonitis. At the Royal Marsden NHS Trust, the volume of lung irradiated to greater than 18 Gy (designated V18) has been used for some time to predict patients at high risk. For this study, we also recorded the mean lung dose as there is increasing evidence that this correlates most closely with clinical reports of pneumonitis [15,17], and an NTCP calculation for lung was performed. The heart was close to the PTV, but was not included as an organ at risk. Cardiac toxicity from radiotherapy for oesophageal carcinoma was not a major clinical concern because of the small numbers of long-term survivors. 2.3. Treatment plan parameters The CT images were transferred to VIRTUOS [3], a virtual simulator forming part of the VOXELPLAN treatment planning system (DKFZ-Heidelberg). Three-dimensional conformal plans were created for each patient, using a two-phase technique, with the same target volume for each phase. The first phase consisted of parallelopposed, antero-posterior and postero-anterior (AP–PA) fields, and the second phase an anterior and two posterior oblique fields at gantry angles of approximately 110 and 2508 (Fig. 1). The beam angles were manually adjusted to avoid the spinal cord, and the beam weights were selected in order to maximize PTV dose homogeneity. Wedges were used on the posterior oblique fields when indicated. The fields were conformally shaped using the beam’s eye view of the PTV. A margin of 6 mm was allowed between the PTV and the field edge to allow for the beam penumbra. This margin was chosen after a careful examination of the penumbra produced by conformal blocks and multileaf collimators [11]. Both treatment phases were planned separately, and the proportion of dose delivered by each phase was calculated such that the maximum spinal cord dose remained below 45 Gy, and the V18 was minimized. The two phases of the plan were then combined into a composite plan for DVH calculation [2]. This composite plan consisted of four fields (4F). The CT images and outlines were also transferred by inhouse software to CORVUS (v3.0), an inverse treatment planning system (Nomos Corporation, Pittsburgh, PA) [30]. For each patient, a dose distribution for 9F IMRT was produced. The following inverse planning constraints were used: PTV: goal dose, 55 Gy (^5%) in 30 fractions; lungs: 18 Gy to less than 5–10% of the lung volume; spinal cord: maximum dose, 45 Gy. The lung constraint was chosen to be lower than the values reported by Bedford et al. [2] for CFRT, so as to guide the solution towards lower lung doses. In principle, the aim was to minimize the mean lung dose or lung NTCP, subject to satisfactory PTV dose and sufficiently low spinal cord dose. However, CORVUS

C.M. Nutting et al. / Radiotherapy and Oncology 61 (2001) 157–163

did not offer the option to carry out biological optimization. Instead, the user was required to enter dose–volume constraints, from which an ideal cumulative DVH was constructed. Its optimization engine then carried out simulated annealing using an objective function based upon the difference between the actual and ideal cumulative DVHs. Hence, in this study, the approach was to use experience of previous studies [2] to specify dose–volume constraints which guided the planning system to a satisfactory solution. Further dose statistics and biological indices were then used in the treatment plan evaluation. This type of approach was

159

used by Stein et al. [26]. The specification of treatment objectives in terms of dose–volume points is widespread in commercial treatment planning systems, and this type of strategy is therefore often necessary. The 9F IMRT plan represented the standard geometry for intensity modulation. No attempt was made to optimize the beam orientations since the inverse planning algorithm was expected to reduce the intensity of any beams that were not useful. It was shown by So¨ derstro¨ m and Brahme [25] that IMRT techniques using as few as three fields were able to achieve almost as high complication-free control rates (P 1 ) as 9F techniques. Thus, seven-field (7F), five-field (5F) and three-field (3F) IMRT plans with equispaced beams were also investigated in one patient. For all patients, a 4F plan was also produced with the same geometry as the CFRT plan. This was used to represent an IMRT plan with selected beam orientations. Choice of beam orientation for the IMRT plans was an important factor. So¨ derstro¨ m and Brahme [25] demonstrated that choice of beam orientation influenced IMRT techniques using three fields more than it influenced techniques using five or more fields. This was supported by Stein et al. [26], who showed that notable improvements in DVHs and biological indices could be achieved using orientation optimization for techniques with fewer than five fields. However, beam orientation optimization was found to be extremely difficult to achieve in the case of oesophageal cancer, due to the presence of lung inhomogeneity. The traditional approach has been to neglect the details of the interaction of the radiation fields with the patient in order to speed up the optimization sufficiently. For example, Stein et al. neglected lateral scatter in their optimization scheme, and only introduced a full scatter calculation for the best ten beam orientations [26]. Similarly, Rowbottom et al. [21] partially neglected scatter and inhomogeneity within the patient. However, when the method of Rowbottom et al. [21] was applied to beam orientation optimization in the present study, it was unable to provide any improvement over an IMRT plan with equispaced beam angles. The long-term solution will be to fully incorporate scatter and tissue inhomogeneity into the optimization process. However, such a method was not available for this study, so the beam angles for the CFRT plans were used for the 4F IMRT plans, as these angles were carefully chosen, and were known to succeed for CFRT. 2.4. Treatment plan evaluation

Fig. 1. The field arrangements used for: (a), CFRT; (b), 9F IMRT; and (c), 4F IMRT techniques. Gantry angles are indicated next to the fields. The gantry angles for patient 1 (PT 1) and patient 5 (PT 5) are given in brackets as they are slightly different to those for the remaining patients.

For both CFRT and IMRT plans, the dose was calculated using a pencil-beam model with equivalent path length inhomogeneity correction. Both VIRTUOS and CORVUS were supplied with beam data from a 6-MV linear accelerator (Elekta Oncology Systems, Crawley, UK). The dosimetric accuracy of both systems had been previously established, and previous studies had shown the two systems to produce mutually compatible results [19]. For the PTV,

160

C.M. Nutting et al. / Radiotherapy and Oncology 61 (2001) 157–163

The NTCP for lung was calculated using BIOPLAN [22]. The calculations incorporated the Kutcher–Burman histogram reduction scheme [14] in conjunction with the Lyman model [16]. Before histogram reduction, the dose at each point within the lung was converted to the biologically effective dose (BED) using an a /b ratio of 3.0 Gy. The parameters proposed by Kwa et al. [15], based on treatment plans with inhomogeneity correction for 540 patients, were then used (TD50 ¼ 30:5 Gy, m ¼ 0:3, and n ¼ 1:0). The NTCP was calculated for each lung volume separately, and the mean lung NTCP was then calculated for all ten lung volumes. Statistical significance of each comparison was assessed using a two-tailed Student’s t-test. This required the data for the population of ten lung volumes to be normally distributed, which was established by creating normal plots (quantile–quantile plots) for the results of the CFRT treatment.

3. Results Typical dose distributions for the CFRT, 9F IMRT and 4F

Fig. 2. Central axis dose distributions for: (a), CFRT; (b), 9F IMRT; and (c), 4F IMRT. Isodoses of 5.5 (10%), 18, 45 and 50 Gy (90%) are shown. Note the larger volume of lung within the 18 Gy isodose on the 9F IMRT plan, compared with the CFRT plan. Note also that the 90% isodose conforms more closely to the PTV with both of the IMRT plans than with the CFRT plan, due to inherent penumbra optimization.

the mean dose and dose range were calculated. The minimum dose to the PTV was defined as the dose received by 99% of the volume, and the maximum dose was defined as the dose received by 1%. This convention, used widely in IMRT planning [5,9,13,19], was to prevent a few outlying voxels from spuriously influencing the results. The maximum dose to the spinal cord, the mean lung dose, and lung V18 for each plan were recorded.

Fig. 3. DVHs for patient 1. (a) 9F IMRT (solid lines) vs. CFRT (dotted lines); (b), 4F IMRT (solid lines) vs. CFRT (dotted lines).

C.M. Nutting et al. / Radiotherapy and Oncology 61 (2001) 157–163

161

Table 1 Mean a results for CFRT and 9F IMRT for five oesophageal cancer patients

Table 3 Mean a results for CFRT and 4F IMRT for five oesophageal cancer patients b

Structure

P value

Structure

– 0.09 0.1 0.06 0.2 0.2 0.03 –

Mean dose (Gy) 55.7 ^ 1.0 55.7 ^ 1.0 Minimum dose (%) 94.2 ^ 2.8 94.0 ^ 1.9 Maximum dose (%) 106.6 ^ 3.9 105.8 ^ 1.8 Dose range (%) 12.4 ^ 3.9 11.8 ^ 3.3 Spinal cord Maximum dose (Gy) 44.5 ^ 0.5 44.5 ^ 0.6 Lungs Mean dose (Gy) 11.0 ^ 2.9 9.5 ^ 2.3 18.8 ^ 11.9 14.1 ^ 10.1 V18 (%) NTCP (%) 1.0 ^ 0.7 0.6 ^ 0.4

Parameter

CFRT

9F IMRT

Mean dose (Gy) 55.7 ^ 1.0 55.7 ^ 1.0 Minimum dose (%) 94.2 ^ 2.8 91.3 ^ 3.4 Maximum dose (%) 106.6 ^ 3.9 108.2 ^ 3.0 Dose range (%) 12.4 ^ 3.9 16.9 ^ 4.5 Spinal cord Maximum dose (Gy) 44.5 ^ 0.5 42.5 ^ 3.0 Lung Mean dose (Gy) 11.0 ^ 2.9 11.7 ^ 3.2 18.8 ^ 11.9 22.2 ^ 12.8 V18 (%) NTCP (%) 1.0 ^ 0.7 1.0 ^ 0.8 PTV

a

Parameter

CFRT

4F IMRT

PTV

P value – 0.9 0.6 0.6 0.9 0.001 0.001 0.008

a

Mean ^ 1 SD.

Mean ^ 1 SD. The beam angles for the IMRT plans are equal to those of the CFRT plans. b

IMRT plans are shown in Fig. 2. DVHs comparing CFRT with 9F and 4F IMRT are shown in Fig. 3 for patient 1. For the PTV, the 9F IMRT technique produced similar minimum and maximum doses and dose inhomogeneity to the CFRT technique (Table 1). The spinal cord maximum dose was respected with both techniques. Nine-field IMRT reduced the volume of lung irradiated to high doses (above 20–30 Gy), but at the expense of larger volumes of lung receiving lower doses. At the clinically relevant threshold of 18 Gy, 9F IMRT increased the irradiated volume of lung (P ¼ 0:03). This can be seen by comparing the volume of lung within the 18 Gy isodose in Fig. 2a,b, and by examining the dose statistics in Table 1. No improvement in mean lung dose was achieved with 9F IMRT compared with CFRT (P ¼ 0:2). Experimentation with different inverse planning constraints did not improve on these results. The effect of reducing the number of equispaced fields is shown in Table 2 for patient 1. IMRT techniques using seven or five equispaced fields failed to reduce the lung V18, or to reduce the mean lung dose, compared with CFRT. A 3F IMRT plan reduced the mean dose, but raised the V18. This was not improved by changing the inverse planning constraints. More promising results were obtained for the 4F IMRT plan using beam orientations chosen from the beam’s eye view to avoid the spinal cord and lung. The mean dose to the PTV for the 4F IMRT plans was comparable with the CFRT and 9F IMRT techniques (Table 3). The mean range of the PTV dose for the 4F IMRT plans was also comparable with the CFRT plans. The maximum spinal cord dose was equivalent to CFRT. The 4F IMRT plans significantly

reduced the lung V18 from 18.8 to 14.1% (P ¼ 0:001). The mean lung dose was reduced from 11.0 Gy with the CFRT plans to 9.5 Gy with 4F IMRT (P ¼ 0:001). The dose to the lungs was consequently much lower than that with the 9F plans. From the DVH in Fig. 3b, it can be seen that the 4F IMRT technique reduced the volume of lung receiving both high and lower doses of radiation. NTCP calculations showed a significant reduction in the risk of grade 2 radiation pneumonitis with IMRT using this customized beam arrangement, compared with CFRT (P ¼ 0:008). 4. Discussion IMRT offers the greatest benefit when the tumour is concave [13,18,19]. For oesophageal carcinoma, where the PTV is approximately cylindrical, the benefit is therefore expected to be small. The results of this study are in accord with this expectation, with a small benefit in terms of lung sparing being demonstrated, depending upon the IMRT technique used. The choice of IMRT technique is clearly important. Although a 9F plan with equispaced beams is often proposed as the ultimate in practical IMRT, it has been shown to be inappropriate in the present context. The 9F IMRT plans produce acceptable dose distributions for the PTV, but fail to improve the dose to the lungs, compared with CFRT. This is a consequence of the effect which has been referred to as “a little to a lot or a lot to a little” [31]. Equispaced beam arrangements with seven, five and three

Table 2 Optimisation of IMRT techniques using equispaced fields for patient 1 Structure

Parameter

CFRT

9F IMRT

7F IMRT

5F IMRT

3F IMRT

PTV

Mean dose (Gy) Minimum dose (%) Maximum dose (%) Dose range (%) Maximum dose (Gy) Mean dose (Gy) V18 (%)

56.2 93.0 110.0 17.0 43.9 11.5 15.5

56.2 92.3 108.9 16.6 43.5 11.8 24.2

56.2 92.2 110.2 18.0 43.9 11.9 27.1

56.2 91.1 109.6 18.5 44.9 11.7 26.1

56.2 89.8 109.8 20.0 33.0 10.9 23.5

Spinal cord Lungs

162

C.M. Nutting et al. / Radiotherapy and Oncology 61 (2001) 157–163

fields gradually change the situation to one in which a high dose is given to a relatively small portion of lung. However, multi-field equispaced beam arrangements do not improve the estimated lung toxicity. So¨ derstro¨ m and Brahme [25] and Stein et al. [26] have previously shown that for prostate tumours, the use of three or four beams can produce acceptable dose distributions. This work supports this conclusion and reaffirms their view of the importance of beam direction optimization. In this study, beam directions chosen on the basis of the corresponding CFRT directions provide very acceptable dose distributions. The resulting IMRT plans yield good PTV dose homogeneity, and a reduction in the mean lung dose for equivalent maximum radiation dose to the spinal cord. There is also a reduction in the V18 and NTCP for lung. However, further development of a beam direction optimization algorithm for the thorax is required. In particular, scatter and tissue inhomogeneity effects need to be incorporated into such an algorithm without excessively increasing the calculation time. The success of the 4F IMRT plan is probably due to its improved field shaping compared with the CFRT plan. Although with CFRT, the penumbra margin has been carefully optimized for the width of the measured penumbra [11], the actual 90% isodose inevitably varies in relation to the PTV, depending upon the precise overlap of the fields. This could be improved by a field-shape optimization algorithm [28], but this is beyond the normal practice in CFRT. With IMRT, the penumbra margin is selected dosimetrically, rather than by applying a fixed geometric rule. Moreover, with IMRT, the fluence within the PTV can be adjusted so that the penumbra is narrower than that with CFRT [23]. This has also been used to reduce the field length in radiotherapy for lung cancer [8]. This effect is probably present in this study, although the CORVUS planning system produces highly modulated fields, with intensity varying substantially across the width of each field, so that the exact effect is difficult to visualize. In general, the approximate mean weights of the intensity-modulated fields are comparable with the beam weights of the CFRT fields, as would be expected. With the 4F IMRT plan, the reduction in mean lung dose is 1.5 Gy and the mean reduction in lung V18 is 4.7%. These improvements are small, but are of similar magnitude to those reported by other authors comparing CFRT and IMRT for stage III non-small cell lung carcinoma [7]. The latter have been used as a basis for dose escalation in this tumour site [6]. Interestingly, the reduction in lung dose is also of a similar magnitude to that seen when conventional and CFRT techniques are compared for oesophageal carcinoma [2]. It can be seen from the dose distributions in Fig. 2 that the use of predominately anterior and posterior portals increases the proportion of the mediastinum in the high dose region. The heart and great vessels are predicted to receive doses of 45–55 Gy. The clinical consequences of this could be accel-

erated ischaemic heart disease in long-term survivors, as seen after the irradiation of the heart in breast cancer patients. The magnitude of this effect is difficult to predict, but may have a detrimental effect on the therapeutic ratio. However, this is not of immediate concern since the survival time is not yet long enough. For this reason, we have not included the heart and the great vessels in the optimization scheme. This is likely to remain the case even with dose escalation. The lung NTCP values are in the range of 0.2–2.7%, significantly lower than those seen clinically in this patient group. This difference is likely to be due to the use of concomitant chemotherapy, which has not been accurately accounted for in the NTCP parameters [15]. Moreover, the NTCP calculation assumes that the dose per fraction delivered to the lung is equal throughout treatment. This assumption is correct for the IMRT plans, but incorrect for the composite conformal plans. In the latter case, some of the lung is treated only in the second phase, at a higher dose per fraction, so that the calculated NTCP may be an underestimate. However, although the NTCP is recognized to be only an estimate of lung damage, the calculated probabilities are consistent with the dosimetric statistics. The data presented here suggest that IMRT may be used either to reduce the risk of pneumonitis after irradiation, or to escalate the dose delivered to oesophageal tumours. If tumour dose escalation within currently acceptable mean lung dose were attempted, then it may be possible to deliver a dose in excess of 60 Gy as long as the optimization algorithm could maintain the spinal cord dose within tolerance. This hypothesis could be explored in a clinical study with normal tissue complications or tumour control as endpoints.

5. Conclusions Intensity-modulated treatment plans using conventional beam angles can provide acceptable dose homogeneity within the PTV and reduce lung irradiation in oesophageal radiotherapy. This improvement in the dose distribution may reduce the risk of radiation pneumonitis, or could allow dose escalation. It has not been possible to produce acceptable dose distributions using a 9F IMRT technique or other IMRT techniques using equispaced beam angles, but in agreement with reports in the literature [25,26], a technique using four customized beam directions has provided a more appropriate treatment plan.

Acknowledgements The authors are grateful to the Nomos Corporation for their generous loan of the CORVUS treatment planning system. The authors would like to thank their collaborators at the German Cancer Research Centre, DKFZ-Heidelberg, for providing VOXELPLAN. This work was generously

C.M. Nutting et al. / Radiotherapy and Oncology 61 (2001) 157–163

supported by a programme grant from The Cancer Research Campaign under grant reference SP2312/0201.

References [1] Al-Sarraf M, Martz K, Herskovic L, et al. Progress report of combined chemoradiotherapy versus radiotherapy alone in patients with esophageal cancer: an intergroup study. J Clin Oncol 1997;15:277–284. [2] Bedford JL, Viviers L, Guzel Z, Childs PJ, Webb S, Tait DM. A quantitative treatment planning study evaluating the potential of dose escalation in conformal radiotherapy of the oesophagus. Radiother Oncol 2000;57:183–193. [3] Bendl R, Pross J, Hoess A, Keller MA, Preiser K, Schlegel W. VIRTUOS — a program for virtual radiotherapy simulation and verification. In: Hounsell AR, Wilkinson JM, Williams PC, editors. Proceedings of the 11th International Conference on the Use of Computers in Radiation Therapy, Manchester, March 20–24, 1994, Manchester: North Western Medical Physics Department, 1994. pp. 226–227. [4] Bosset JF, Gignoux M, Triboulet JP, et al. Chemoradiotherapy followed by surgery compared with surgery alone in squamous cell cancer of the oesophagus. N Engl J Med 1997;337:161–167. [5] Cardinale RM, Benedict SH, Wu QW, Zwicker RD, Gaballa HE, Mohan R. A comparison of three stereotactic radiotherapy techniques; arcs vs. noncoplanar fixed fields vs. intensity modulation. Int J Radiat Oncol Biol Phys 1998;42:431–436. [6] Derycke S, Van Duyse B, De Gersem W, De Wagter C, De Neve W. Non-coplanar beam intensity modulation allows large dose escalation in stage III lung cancer. Radiother Oncol 1997;45:253–261. [7] Derycke S, De Gersem WRT, Van Duyse BBR, De Neve WCJ. Conformal radiotherapy of stage III non-small cell lung cancer: a class solution involving non-coplanar intensity-modulated beams. Int J Radiat Oncol Biol Phys 1998;41:771–777. [8] Dirkx MLP, Heijmen BJM. Beam intensity modulation for penumbra enhancement and field length reduction in lung cancer treatments: a dosimetric study. Radiother Oncol 2000;56:181–188. [9] Eisbruch A, Marsh LH, Martel MK, et al. Comprehensive irradiation of head and neck cancer using conformal multisegmental fields: assessment of target coverage and noninvolved tissue sparing. Int J Radiat Oncol Biol Phys 1998;41:559–568. [10] Guzel Z, Bedford JL, Childs PJ, et al. A comparison of conventional and conformal radiotherapy of the oesophagus: work in progress. Br J Radiol 1998;71:1076–1082. [11] Heisig S, Shentall GS, Mirza K, Mayles WPM. Application of the GE TARGET planning computer to multi-leaf collimator treatments. In: Hounsell AR, Wilkinson JM, Williams PC, editors. Proceedings of the XIth International Conference on the Use of Computers in Radiation Therapy, Manchester, March 20–24, 1994, Manchester: North Western Medical Physics Department, 1994. pp. 16–17. [12] Herskovic A, Martz K, Al-Sarraf M, et al. Combined chemotherapy and radiotherapy compared with radiotherapy alone in patients with cancer of the esophagus. N Engl J Med 1992;326:1593–1598. [13] Khoo VS, Oldham M, Adams EJ, Bedford JL, Webb S, Brada M. Comparison of intensity-modulated tomotherapy with stereotactically

[14]

[15]

[16] [17]

[18]

[19]

[20] [21]

[22]

[23] [24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

163

guided conformal radiotherapy for brain tumors. Int J Radiat Oncol Biol Phys 1999;45:415–425. Kutcher GJ, Burman C. Calculation of complication probability factors for non-uniform normal tissue irradiation: the effective volume method. Int J Radiat Oncol Biol Phys 1989;16:1623–1630. Kwa SLS, Lebesque JV, Theuws JCM, et al. Radiation pneumonitis as a function of mean lung dose: an analysis of pooled data of 540 patients. Int J Radiat Oncol Biol Phys 1998;42:1–9. Lyman JT. Complication probability as assessed from dose–volume histograms. Radiat Res 1985;104:S13–S19. Martel MK, Ten Haken RK, Hazuka MB, Turrisi AT, Fraass BA, Lichter AS. Dose–volume histogram and 3-D treatment planning evaluation of patients with pneumonitis. Int J Radiat Oncol Biol Phys 1994;28:575–581. Nutting C, Rowbottom C, Convery D, et al. Intensity modulated radiotherapy (IMRT) in tumours of the head and neck. Br J Radiol 1999;72:S98–S99. Nutting CM, Convery DJ, Cosgrove VP, et al. Reduction of small and large bowel irradiation using an optimized intensity-modulated pelvic radiotherapy technique in patients with prostate cancer. Int J Radiat Oncol Biol Phys 2000;48:649–656. Nutting C, Dearnaley DP, Webb S. Intensity modulated radiation therapy: a clinical review. Br J Radiol 2000;73:459–469. Rowbottom CG, Nutting CM, Webb S. Beam-orientation optimization of intensity-modulated radiotherapy: clinical application to parotid gland tumours. Radiother Oncol 2001;59:169–177. Sanchez-Nieto B, Nahum AE. BIOPLAN: software for the biological evaluation of radiotherapy treatment plans. Med Dosim 2000;25:71– 76. Sharpe MB, Miller BM, Wong JW. Compensation of X-ray beam penumbra in conformal radiotherapy. Med Phys 2000;27:1739–1745. Smith TJ, Ryan LM, Douglass HO, et al. Combined chemoradiotherapy vs. radiotherapy alone for early stage squamous cell carcinoma of the esophagus: a study of the Eastern Cooperative Oncology Group. Int J Radiat Oncol Biol Phys 1998;42:269–276. So¨ derstro¨ m S, Brahme A. Which is the most suitable number of photon beam portals in coplanar radiation therapy? Int J Radiat Oncol Biol Phys 1995;33:151–159. Stein J, Mohan R, Wang XH, et al. Number and orientations of beams in intensity-modulated radiation treatments. Med Phys 1997;24:149– 160. Sykes AJ, Burt PA, Slevin NJ, Stout R, Marrs JE. Radical radiotherapy for carcinoma of the oesophagus: an effective alternative to surgery. Radiother Oncol 1998;48:15–21. van Dalen S, Keijzer M, Huizenga H, Storchi PRM. Optimization of multileaf collimator settings for radiotherapy treatment planning. Phys Med Biol 2000;45:3615–3625. Walsh TN, Noonan N, Hollywood D, Kelly A, Keeling N, Hennessy TPJ. A comparison of multimodal therapy and surgery for esophageal adenocarcinoma. N Engl J Med 1996;335:462–467. Xing L, Curran B, Hill R, et al. Dosimetric verification of a commercial inverse treatment planning system. Phys Med Biol 1999;44:463– 478. Zaider M, Amols HI. A little to a lot or a lot to a little: is NTCP always minimized in multiport therapy? Int J Radiat Oncol Biol Phys 1998;41:945–950.