Proton Beams to Replace Photon Beams in Radical Dose Treatments

ACTA ONCOLOGICA LECTURE

æ

æ

Proton Beams to Replace Photon Beams in Radical Dose Treatments Herman Suit, Saveli Goldberg, Andrzej Niemierko, Alexei Trofimov, Judith Adams, Harald Paganetti, George T. Y. Chen, Thomas Bortfeld, Stanley Rosenthal, Jay Loeffler and Thomas Delaney From the Department of Radiation Oncology, Massachusetts General Hospital, Harvard Medical School, Boston, USA Correspondence to: Herman Suit, Department of Radiation Oncology, Massachusetts General Hospital, 55 Fruit Street, Boston, MA 02114, USA. Fax: /1 617 726 4805. E-mail: [email protected]

Acta Oncologica Vol. 42, No. 8, pp. 800 /808, 2003 With proton beam radiation therapy a smaller volume of normal tissues is irradiated at high dose levels for most anatomic sites than is feasible with any photon technique. This is due to the Laws of Physics, which determine the absorption of energy from photons and protons. In other words, the dose from a photon beam decreases exponentially with depth in the irradiated material. In contrast, protons have a finite range and that range is energy dependent. Accordingly, by appropriate distribution of proton energies, the dose can be uniform across the target and essentially zero deep to the target and the atomic composition of the irradiated material. The dose proximal to the target is lower compared with that in photon techniques, for all except superficial targets. This resultant closer approximation of the planning treatment volume (PTV) to the CTV/GTV (grossly evident tumor volume/subclinical tumor extensions) constitutes a clinical gain by definition; i.e. a smaller treatment volume that covers the target three dimensionally for the entirety of each treatment session provides a clinical advantage. Several illustrative clinical dose distributions are presented and the clinical outcome results are reviewed briefly. An important technical advance will be the use of intensity modulated proton radiation therapy, which achieves contouring of the proximal edge of the SOBP (spread out Bragg peak) as well as the distal edge. This technique uses pencil beam scanning. To permit further progressive reductions of the PTV, 4-D treatment planning and delivery is required. The fourth dimension is time, as the position and contours of the tumor and the adjacent critical normal tissues are not constant. A potentially valuable new method for assessing the clinical merits of each of a large number of treatment plans is the evaluation of multidimensional plots of the complication probabilities for each of ‘n’ critical normal tissues/ structures for a specified tumor control probability. The cost of proton therapy compared with that of very high technology photon therapy is estimated and evaluated. The differential is estimated to be :/1.5 provided there were to be no charge for the original facility and that there were sufficient patients for operating on an extended schedule (6 /7 days of 14 /16 h) with ]/ two gantries and one fixed horizontal beam. Received 12 June 2003 Accepted 18 August 2003

One strategy for improving the efficacy of radiation therapy that has yielded progressively higher success rates has been that of reducing the planning treatment volume (PTV). This is designed to include the grossly evident tumor volume (GTV) and the larger volume which allows for subclinical tumor extensions (CTV) and the PTV with a sufficient margin to accommodate uncertainties in patient (target) setup errors. The following truisms are the basis for these efforts. First, there is no advantage to any patient for any uninvolved tissue to receive any radiation dose. Second, primary radiation injury never develops in unirradiated tissues. Accordingly, a technique that allows a reduced treatment volume is a superior treatment method. That is, there is no basis for research into whether a smaller treatment volume is an advantage. The gain may be very small, but it is virtually impossible to conceive of a negative outcome. There is, however, legitimate research to assess the # Taylor & Francis 2003. ISSN 0284-186X DOI: 10.1080/02841860310017676

magnitude of the gain or the gain versus the cost of the new method, but not whether there is a gain. A major advance in this path to reduction in treatment volume is the use of proton beams. Because of the substantially smaller PTVs required by proton therapy techniques, it is predicted that photon beams will be substantially replaced by proton beams for the definitive radiation treatment of cancer patients within one to two decades (1). This optimism for potential gains from proton beam radiation therapy is based on the Laws of Physics, which determine absorption of energy from protons. These make the achieving of a superior dose distribution comparatively straightforward for most anatomic sites compared with any photon technique, including intensitymodulated x-ray therapy (IMXT). A superior dose distribution is a smaller treatment volume (PTV) that covers the defined target three dimensionally throughout each Acta Oncologica

Acta Oncologica 42 (2003)

treatment session. In addition, a superior treatment also means a smaller dose to the volume of normal tissues in the low and intermediate dose volumes outside of the beam paths. For each of these different volumes, proton techniques can deliver a lower dose. Normal tissues are those that are not known or suspected to be involved by tumor on a gross or microscopic basis at a clinically important probability. The basis for such an unqualified positive view for gains from technical advances is that a decrease in the PTV increases the tolerance of the patient to radiation and hence permits a higher dose to the target. This will inevitably yield some increment in tumor control probability (TCP). Concurrently, owing to the smaller treatment volume, there would be a reduction in the severity and frequency of treatment-related morbidity. The maximum gain in TCP as a result of reduced treatment volume will be for tumors for which the standard treatment method yields a TCP in the range of :/0.2 /0.75, that is the range in TCP for which the dose response curve is maximally steep. The slope of the dose response curve for tumor control probability of a moderately well-stratified population of human tumors has been accepted as characterized by a g50 of :/2. Thus, if the TCP were 0.4 for the standard treatment and the dose to the target increased by 10% by using a new technology, then the TCP would be increased by :/20 percentage points; i.e. from 40% to :/60%. There may well be a concurrent reduction in normal tissue complication probability (NTCP). For the reference treatment, the NTCP for Grade III/IV injuries would in nearly all circumstances be 5/5% and, accordingly, be in the relatively flat portion of the dose /response curve. Despite the high g50 of :/4 for NTCP for class III/IV injuries, the maximum that this could be expected to improve by would be from, say, about 5% down to perhaps 1%.

MATERIAL AND METHODS In this article the clinical need for further improvements in the efficacy of radiation therapy is discussed. This is followed by consideration of the first proposal for the use of proton beams in radiation therapy and the comparative depth dose curves for a photon and a modulated energy proton beam. The start of proton radiation therapy and its subsequent growth is described. Several illustrative proton treatment plans are then presented followed by a brief and selective review of outcome results of proton radiation therapy. Two of the coming technical developments in radiation oncology are considered, which will be in high technology photon and proton therapy. In our final section we consider the relative cost of very high technology photon therapy versus proton radiation therapy.

Protons to replace photons for radical dose therapy

801

RESULTS Need for improved local therapy Despite the many and impressive improvements in the efficacy of radiation oncology, alone or combined with other modalities (surgery, chemical and biological agents or genetic manipulations) during recent decades, there are, as of 2003, important problems concerning local failure, normal tissue complications, reduced tolerance to chemotherapy and or surgery. With the use of substantially advanced radiation technologies, that is smaller treatment volumes, there is almost a certainty that the magnitude of each of these categories of problems can be decreased. The extent of the decrease will be site specific. Initial proposal for proton radiation therapy In 1946 Robert Wilson of Harvard University Physics Department (2) published the initial proposal for the use of protons in radiation therapy. Wilson had played a central role in the development in 1945 of the atomic bombs used in the attacks on Hiroshima and Nagasaki, Japan. Upon his return to academic life, one of Wilson’s initial efforts was to devise a peaceful application of nuclear physics to human health needs; hence, his investigation of protons as a therapeutic beam. This resulted in his, now famous, 1946 paper, ‘Radiological Use of Fast Protons’. Comparative depth dose curves for high-energy photon and modulated-energy proton beams The depth dose curves for a 15 MeV linear accelerator x-ray beam and a modulated energy 150 MeV proton beam from our new proton therapy facility are presented in Fig. 1a. The advantages of the proton beam are immediately evident. Consider that the proton energy was selected such that the end of range would be just deep to the distal margin of the defined target. Accordingly, there would be essentially zero dose deep to the target and a lesser dose proximal to the target for the proton beam, excepting for superficially sited lesions. Two features of the depth dose distribution should be noted in planning proton radiation therapy. First, the range of protons of a specified energy is dependent on the density heterogeneity of the tissue in the beam path. The effective density along the path of each proton ‘beamlet’ can readily be determined by analysis of the CT scan data and the optimal energy distribution accordingly selected. Secondly, the penumbra width of proton beams to photon beams varies with depth in tissue. For the 95% /80% width, the penumbra is narrower for a proton beam than for a 15-MeV photon beam up to a depth of 25 cm and for all depths investigated for the 18 MeV xray beam, as illustrated in Fig. 1b. However, the relationship for the 80% /20% is more complex. Specifically, the penumbra width of a proton beam is narrower than that of the 15 MeV x-ray beam to a depth of 17 cm and 22 cm

802

H. Suit et al.


is no postulated or known radiation biological advantage for proton therapy. The relative biological effectiveness (RBE) for protons is accepted as 1.10 in the SOBP and is a generic value (3). That is, the one value is applied to all tissues, dose/fraction, total doses, total time and treatment volumes. (There is a slightly higher RBE at the end of range of the SOBP. However, the principle effect of that small region of higher RBE is to extend the range by a few millimeters and not to create a significant local ‘hot spot’.) Early proton therapy

Fig. 1. (a) Central axis depth dose curve for a 15 MV linear accelerator x-ray beam and a 150 MV proton beam with the Spread Out Bragg Peak. (b) Penumbra width versus depth for the 95% to 80% dose level for a 15 MeV and an 18 MeV x-ray beam and a proton beam. (c) Penumbra width versus depth for the 80% to 20% dose level for a 15 MeV and an 18 MeV x-ray beam and a proton beam.

for the 18 MeV x-ray beam, as illustrated in Fig. 1c. Thus, the penumbra for proton beams is narrower than that for photon beams for at least a depth of 17 cm. Interest in proton beam therapy is based exclusively on the lesser dose to normal tissues throughout the body. There

The first therapeutic use of protons was at the University of California at Berkeley, 1955. However, they shortly changed to helium ion beam therapy and later extended their study of particle therapy to carbon and neon ion beams. The second program in proton treatment of patients was by the team at Uppsala in 1957 (4 /6). The Swedish program was followed by programs conducted by the Massachusetts General Hospital (MGH) Neurosurgical Group working with the Harvard Cyclotron Laboratory (HCL), the Joint Institute for Nuclear Research (JINR), Dubna Russia (1964) and the Institute of Experimental and Theoretical Physics (ITEP), Moscow (1968). These programs were based on physics research facilities and beam time for patients was extremely limited. This meant that there was no opportunity to investigate standard fractionated treatment protocols. The treatments were based principally on single dose, or in a few instances 2 to 4 dose fractions, and the targets were predominately benign intracranial lesions. In the wake of those initial efforts, there has been a substantial expansion in the number of hospital-based proton therapy programs with almost full-time beam availability. For a current tabulation of the proton therapy centers in the world, examine Table 1 (data for this Table was obtained from Janet Sisterson, 2003). The Table presents the numbers of patients treated at each of the centers that treat deeply sited lesions and then those centers using proton energies of :/65 MeV against ocular lesions. Additional new proton therapy programs are to be implemented at MD Anderson; the University of Pennsylvania; Trentino, Italy; and the University of Florida Heidelberg (proton and 12Cion); Seoul, Korea; Wanjie, Zibo, China; Beijing, China; Shizoka, Japan; and Munich. Use of standard fractionated dose protocols in proton beam therapy By 1970, the Harvard Cyclotron and its 160 MeV proton beam was of minimal interest to basic physics and was therefore available for clinical work on the basis of 4 day/ week proton and one day/week photon treatment to investigate standard fractionated proton radiation therapy of the cancer patient. This program was initiated in 1974 based on :/2 CGE/fraction (CGE /Gy /RBE of 1.1). Our strategy was to concentrate resources on a fairly small



803

Table 1 World proton-treated patients / totals Institution Deep lesions Berkeley 184 Uppsala

Site

CA. USA Sweden 1989 Boston, MA

First Rx

1954 1957 313 MGH-MEEI-HCL 1961 NPTC, MGH 2001 Dubna Russia 1967 /1974 1983 154 Moscow Russia 1969 St. Petersburg Russia 1975 PMRC, Tsukuba Japan 1983 /2000 Loma Linda CA. USA 1990 Louvain-la-Neuve Belgium 1991 /1993 Orsay France 1991 iThemba LABS South Africa 1993 MPRI IN USA 1993 PSI (200 MeV) Switzerland 1996 NCC, Kashiwa Japan 1998 HIBMC, Hyogo Japan 2001 PMRC (2), Tsukuba Japan 2001 NPTC, MGH MA USA 2001 Wakasa Bay Japan 2002 Orbital lesions only Chiba Japan 1979 PSI (72 Me1V) Switzerland 1984 Clatterbridge England 1989 Nice France 1991 UCSF / CNL CA USA 1994 TRIUMF Canada 1995 H. M. I, Berlin Germany 1998 INFN-LNS, Catania Italy 2002 Combined Total

Patient no. Date of total

30 73 Jan.-02 9 116 229 84 3 539 1 029 700 7176 21 2 157 433 34 99 161 30 145 229 2 145 3 712 1 201 1 951 448 77 317 24 33 398

1957 1976 Dec.-02 Dec.-02 Dec.-02 Dec.-02 June-98 July-00 May-02 Jan.-02 Dec.-02 Dec.-99 Dec.-01 Dec.-02 Jan.-02 Dec.-02 Dec.-02 June-02 Apr.-02 Dec.-02 Dec.-02 June-02 July-02 Dec.-02 Dec.-02 Dec.-02

Source : J Sisterson / January 2003.

number of anatomic sites for which local control was low by conventional photon therapy techniques and our analysis of comparative treatment plans indicated that by using proton beam techniques a substantially higher dose could be delivered than by photon techniques. Our first patient (in January 1974) was a young pediatric patient who had a posterior pelvic rhabdomyosarcoma managed by proton beam therapy and chemotherapy. The patient achieved a complete primary response but later succumbed to distant metastatic disease. The initial targets selected for study were skull base and cervical spine sarcomas, sarcomas at several sites and uveal melanomas. The sites studied were later expanded to include the prostate, several of the pediatric tumors, paranasal sinus tumors, meningiomas and glioblastoma. Additionally, the proton beam is used in stereotactic proton radiosurgery and stereotactic radiation therapy for AVMs, acoustic neuromas, selected CNS tumors, and so on. Proton beam therapy has been and is under intensive investigation at many centers worldwide. There have been numerous successful efforts in advancing proton treatment planning and delivery and in clinical evaluations of the

efficacy of proton radiation therapy. Selected results are mentioned below. Illustrative proton dose distributions Pelvic lymph nodes. The dose distributions at the mid-pelvis for IMXT and a single posterior pelvic proton field are presented in Fig. 2. As shown, for IMXT, the dose was 19 and 35 Gy to the central pelvic and anterolateral intestinal tissues. At those sites the proton dose was, of course, essentially zero. Medial distal thigh. Proton radiation treatment of a soft tissue sarcoma of the medial distal thigh delivers near negligible doses to the femoral shaft (see Fig. 3a). The functional status of the elbow joint following proton radiation therapy for an ante-cubital fossa sarcoma by means of a single anterior proton beam designed in such a way that the distal beam edge in three dimensions extended only to the periosteum is shown in Fig. 3b and c. This patient enjoyed near normal function for 18 years, after which time he succumbed to pancreatic cancer.

804

H. Suit et al.


Fig. 2. Dose distribution for radiation treatment of the pelvic lymph nodes by 6 MV IMXT and a single posterior proton beam with the dose levels shown at two points in the intestinal structures.

Medulloblastoma . The standard radiation treatment technique for these children is a single posterior 4 MeV x-ray beam which delivers an important exit dose to portions of the heart, esophagus, lung, gastrointestinal (GI) tract and ovaries in female patients. In stark contrast, the dose to those structures in proton radiation therapy is virtually zero (Fig. 4). Carcinoma of the lung. For stage I non-small cell lung cancer (NSCLC) in medically or technically operable patients, high-dose, small-field radiation therapy has been investigated. One of the technical challenges is the correct allowance for target motion. At present, this is achieved by controlling beam ‘On’ for selected phases of the respiratory cycle, i.e. gated treatment. The dose distribution in the IMXT and IMPT for treatment of a stage I NSCLC is presented in Fig. 5. The target dose is set to be the same for the two treatment techniques. There is a dramatic decrease in the volumes of lung tissue that receive low to intermediate dose levels using the IMPT technique. For the IMXT plan, :/one-fifth of patients receiving 15 CGE delivered in 30 equal doses. This dose is expected to result in radiation-induced cancer at a low but not zero probability (see Hall & Wuu (7)). Outcome results of proton treatment of cancer patients Chordomas and chondrosarcomas of the skull base/cervical spine. The two most common neoplasms of these anatomic sites are chordomas and chondrosarcomas. Surgical and photon radiation treatments of these lesions have usually failed because of the anatomical constraints of the extremely close proximity of the tumor to critical and radiationsensitive CNS structures. Simple evaluation of comparative treatment plans made clear that substantially higher target doses were readily feasible by proton therapy than the 55 Gy feasible dose by conventional photon therapy. That is, dose levels of :/70 CGE were practical with accepted dose constraints to the CNS structures of concern. The 10-year local control results are 95% for chondrosarcomas and 45% for chordomas ((8), and N. J. Liebsch & J. Munzenrider,

Fig. 3. (a) Dose distribution for radiation treatment of the medial thigh by a single medial proton beam. Notice that the distal edge of the beam is contoured to avoid the femoral shaft. (b) and (c) This patient had proton radiation treatment for an ante-cubital fossa sarcoma with the distal edge of the beam extending to the anterior aspect of the joint. The patient survived for 18 years with excellent function of the joint and use of the upper extremity.

pers. comm. 2003). Analyses of chordoma outcome results demonstrated significantly higher local control rates for male patients than for female patients. Terahara & Niemierko (9) reported TCD50 values of 63.1 CGE and 75.4



805

actuarial local result at 15 years for the first 2 069 patients treated for uveal melanoma by the MGH, HCL and Massachusetts Eye and Ear Infirmary team is 95% (11). The death rate due to distant metastasis varied between 5% and 65% depending on the clinical stage of the melanoma. The eye was retained in 84% of patients. Visual acuity ]/20/ 200 was retained by 80% to 0% at 10 years depending on lesion site and size. Egger et al. (12) reported on a series of 2 435 patients treated with protons to 60 CGE in 4 equal doses at the Paul Scherrer Institute between 1984 and 1998. The local control result was 94.8% at 10 years. Of the 73 local failures, 51 were described as at the immediate margin of the treated volume. Control rates have improved substantially in the patients treated since 1993.

Fig. 4. Treatment of a young patient with medulloblastoma by the conventional posterior 4 MV x-ray beam and resultant intermediate high dose to the anterior tissues/organs. In sharp contrast, treatment by a posterior proton beam delivers essentially zero dose to all tissues anterior to the anterior surface of the vertebral bodies.

Fig. 5. Demonstrating a substantial dose to large portions of the uninvolved lung by the IMXT plan as compared with the very tightly confined dose using the proton beam technique.

CGE for male and female subjects, respectively. They also estimated the g50 to be 3.7 for both genders. Accordingly, for female patients, the dose to the GTV has been increased to 83 CGE. The dose constraints for the brainstem surface and mid-portion are currently set at 67 CGE and 53 CGE, respectively. Clearly, there is a need for more effective treatment for chordomas. A pertinent fact regarding the results obtained with chondrosarcoma is that 74 patients who were referred to our institution with the diagnosis of chordoma had the diagnosis changed to chondrosarcoma (10). Uveal melanoma. Uveal melanomas have been treated traditionally by enucleation and for the past few decades by application of a radioactive plaque over the base of the melanoma. The latter has been effective against small- to intermediate-sized lesions not close to the fovea or disc. Local control at 5 years has been reported to be 87%. The

Hepatocellular carcinoma. A series of 236 inoperable (medically or technically) patients with hepatocellular carcinoma at Tsukuba Cancer Center have been treated with large dose/fraction proton therapy (K Tokuuye, pers. comm. 2003). Specifically, the mean dose was 5 CGE /16 or 80 CGE in 16 fractions. As evidence of the unfavorable status of the patients, 77 had two tumors. Treatment was administered using a respiration, gated technique. The CTV was designed as the GTV/5 /10 mm and the PTV was the CTV/10 mm. Local control at 10 years was estimated to be 85% and 70% for primary and secondary treatment, respectively. However, essentially all patients developed additional tumors in the remaining hepatic tissues. Stage I non-small cell lung cancer (NSCLC). For the small peripheral NSCLCs, fairly small-field, high-dose/fraction proton treatment has yielded very high local control rates. Bush et al. from Loma Linda University (LLU) used 5.1 CGE /10 in 40 patients (13). They have estimated the 3year local control, overall survival and disease-free survival results to be 86%, 34% and 52%, respectively. Carcinoma of the prostate. Shipley et al. (14) reported MGH results of a Phase III trial of stage III/IV carcinoma of the prostate in which patients were randomized to receive 50.4 Gy photon irradiation with a boost dose to 67.2 Gy by photons versus 77.2 CGE by protons. The 8-year local control rates for the 189 patients who completed the planned treatment were 60% vs. 77%, p/0.089, respectively. The 8-year results for Gleason 8 /10 patients were 19% and 84%, p/0.001, favoring the proton arm. There were more treatment-related morbidities among the proton arm, viz urethral stricture rates were 8% vs.19% and rectal bleeding rates of 12% vs. 32%, respectively. Slater et al. of LLU (J. Slater, pers. comm. 2003) analyzed the results of a study of 1 255 T1 /T2 patients who underwent proton treatment during the period 1990 /1997 and for whom there had been no previous surgery and no hormone therapy. Radiation dose levels were 74 /75 CGE at 1.8 /2 CGE/

806

H. Suit et al.

fraction. The 10-year disease-free survival rates were 92%, 81%, 64% and 53% for patients whose pretreatment prostate-specific antigen (PSA) values were 5/4.0, 4.1 / 10.0, 10.1 /20.0 and /20.0, respectively. These results were entirely comparable with those of the surgical series from Johns Hopkins University. Glioblastoma. For this tumor, our attempt at the MGH to have a major impact on disease-free survival failed. Twentythree patients were given 1.8 CGE on a b.i.d. schedule ( /6 h between doses) to 90 CGE in 5 weeks. Of these patients, there was one in-field failure and one patient was still alive at 5 years with good neurological functional status (15). Pediatric tumor patients. Our experience in this most important patient group has been consistently positive regarding the achieved dose distribution. As of this date, our experience with tumors of a specific type and site is too limited in number to make a serious comment. Future technical developments A pertinent fact is that as new technological developments are being applied in the clinic, there is no evidence that marginal failures have become an increasing problem. In fact, the opposite seems to be true. This is primarily due to the markedly better diagnostic imaging and the ability to ensure that we have the target on the beam. As we move towards further reductions in the PTV, the requirements for the stringent controls in maintaining the target within the beam become progressively more difficult to realize. This means that radiation oncology is moving towards imagedirected therapy, so there will be on-line, continuous correction of the position of the target vis-a`-vis the beam. This means the implementation of true 4-D treatment planning and delivery, the fourth dimension being time. Intensity modulated proton therapy (IMPT) IMPT is a pencil beam scanning technique in which a small circular beam is scanned many times across the defined treatment field, with the energy and intensity varying so that the dose in each voxel can be optimized (16). This technique is being studied at the Paul Scherrer Institute, Villigen, Switzerland, using a beam of full width at halfmaximum (FWHM) of 9 mm. However, plans are being developed at several centers for IMPT employing smaller beams (down to 5 mm FWHM). This will provide an even further reduction in the PTV. Pencil beam scanning has two major advantages compared with passive energy modulation as used in all other centers at this writing: 1) The capacity to contour the proximal edge of the spread out Bragg peak and 2) a lower scattered dose. The advantage of the IMPT technique over passive beam energy modulation in terms of dose gradient from the PTV into the nearby normal tissues is judged to be substantially less than that for


the difference between standard photon and IMXT methods. Four dimensional image guided radiation therapy The treatment planning and delivery of dose to moving targets requires accurate definition of the position and the three dimensional contours throughout each treatment session. During irradiation, the beam must encompass the target, either through radiation beam tracking of the target or compensatory motion of the patient, i.e. to keep the target in the beam as planned. Although such techniques are not currently available for clinical use, they are being rapidly developed, and are being designed to provide on-line diagnostic quality imaging of the target and essentially concurrent adjustment of the beam or patient position to reduce to a technical minimum the margin between the PTV and CTV/GTV. Insertion of radio-opaque fiducial markers is likely to be required as part of the planning in order to achieve correct guidance of the radiation beam. During our investigation of four-dimensional effects on dose distribution, one member of our group (HP) developed fourdimensional Monte Carlo dose calculations with modification of the underlying geometry during the Monte Carlo run. Correction for distortion of imaged contour Treatment planning CT scans are commonly performed under conditions of quiet respiration, following the tenet of ‘scanning the way you treat’. Typically, motion artefacts are seen as irregularities in the beam’s eye view projection of the GTV; enveloping the target with a smoothed aperture is generally considered adequate. In recent studies by one of us (GTYC), it is shown that significant shape and location distortions can arise during the scanning of moving objects. Spherical objects undergoing periodic sinusoidal motion simulating respiration (period of 4 s, peak-to-peak motion of 2 cm) can result in elongation or shortening of the imaged object by /2 cm, with centroid displacements of /1 cm on a high performance multidetector CT scanner. In Fig. 6 we present a selection of resultant images for CT scanning of a 6-cm diameter spherical object, which was moved by sinusoidal motion of 1 cm amplitude with a periodicity of 4 s (comparable with a normal respiratory cycle). These distorted images of the contour/shape of the spherical object may be reduced markedly by performing a 4-D CT scan or by diminishing motion through gating. Pareto optimization Future treatment planning systems are projected to generate a large number of plans upon which an assessment of the best plan in terms of the balance of NTCP estimates for each normal tissue 1, 2, 3 . . .. . . for a specified TCP. This evaluation of many plans will be difficult because of the problem that as the dose distribution is adjusted to lower



Fig. 6. A series of CT images of a 6-cm spherical object moving at the speed of a lung tumor under normal respiration. Image A was taken with no motion. There is a substantial distortion of the contours of the object when imaged during motion.

the NTCP for tissue 1, there would likely be an increase in NTCP for tissues 2 and 3. One of us (TB) proposed that Pareto graphical display of NTCP for one organ at risk (OAR) against the NTCP values for a second OAR (multidimensional displays where there is concern about complications arising in multiple organs) for a defined TCP. The Pareto optimization concept originated in economics. The plan is relatively straightforward in that there is a multidimensional plot of the NTCP1 vs. NTCP2 vs. NTCP3 . . . NTCPn. From this, an ‘optimal’ dose distribution can be selected with relative ease. This is illustrated in Fig. 7. Here, equivalent uniform dose (EUD) has been used as a surrogate for NTCP. The clinical applicability is now under evaluation but is predicted to become a useful tool in the clinical assessment of the relative merits of several treatment plans. Costs of proton therapy compared with costs of high technology photon therapy Proton therapy is more costly than photon therapy. For a thorough and detailed analysis of the relative costs, see the excellent paper by Goitein & Jermann (17). We estimate that the cost increment for proton therapy varies from

Fig. 7. This illustrates a Pareto plot of the EUD (equivalent uniform dose) (normal tissue complication probability (NTCP) could be plotted for the same information) to the spinal cord and the parotid salivary gland for a radiation treatment yielding a specified tumor control probability. Pareto optimal /m /; dominated j.

:/1.5 to 3 when comparing proton versus photon systems with fully equivalent patient support assemblies, on line portal imaging, optimized treatment planning systems, including biological effect distribution, etc. This broad range of 1.5 /3 in higher costs is a function of several factors, e.g. number of patients treated/year, number of treatment rooms, number of gantries, number of hours of operation/day and operation days/week, functional lifetime of the cyclotron/synchrotron, maintenance of the proton and photon machines and gantries, the need to allow for costs of financing the building/equipment of the facility, etc. To reduce the cost differential, the proton medical facility would have to be sited such that large numbers of appropriate patients would be referred and there would be sufficient staff for operation on an extended basis. The cost increase factor of 1.5 is estimated to be appropriate were there to be ]/ three gantries with one horizontal beam and a sufficiently high number of patients to have a full schedule and operation of :/14 h/day and no financing charge. A relevant fact when considering the cost of a new therapeutic modality is that, to date, each technical development throughout the history of radiation oncology that

Table 2 Technical developments adopted in 1950 /2003 Portal Films Gantries CT, MRI, MRS, PET Imaging Stereotactic radiation surgery/therapy IMXT Proton Therapy

807

Simulators 60 Co and linear accelerator beams Computers for treatment planning IORT 3-D conformal XRT Image-guided brachytherapy

808

H. Suit et al.

achieved an important reduction in treatment volume has been accepted into clinical practice. Each of these developments has increased the time/effort, space and staff and the total cost of treatment. Consider the technological advances that were adopted into standard practice during the period 1950 /2003, as listed in Table 2. That proton radiation therapy does achieve a superior dose distribution is not contested. However, there is a need to assess the gain versus the cost. This is a value judgment and must be based on the actual cost increment relative to the highest technology photon treatment and how the gains are to be assessed. ACKNOWLEDGEMENTS This programme has been developed in collaboration with the faculty and staff at the Massachusetts General Hospital, the Harvard Cyclotron Laboratory and the Massachusetts Eye and Ear Infirmary. We thank Dr M. Goitein for his critical contribution to this programme. We are also indebted to the important support, grant NCI Ca 21239, given by the National Cancer Institute from 1976 on a continuous basis through to the present time.

REFERENCES 1. Suit H. The Gray Lecture 2001: Coming technical advances in radiation oncology. Int J Radiat Oncol Biol Phys 2002; 53: 798 /809. 2. Wilson RR. Radiological use of fast protons. Radiology 1946; 47: 487 /91. 3. Paganetti H, Niemierko A, Ancukiewicz M, et al. Relative biological effectiveness (RBE) values for proton beam therapy. Int J Radiat Oncol Biol Phys 2002; 53: 407 /21. 4. Falkmer S, Fors B, Larsson B, Lindell A, Naeslund J, Stenson S. Pilot study on proton irradiation of human carcinoma. Acta Radiol 1964; 59: 33 /51. 5. Fors B, Larsson B, Lindell A, Naeslund J, Stenson S. Effect of high energy protons on human genital carcinoma. Acta Radiol 1964; 2: 384 /98.

Acta Oncologica 42 (2003) 6. Graffman S, Jung B, Norhman B, Bergstrom R. Supplementary treatment of nasopharyngeal tumours with high energy protons. Acta Radiol 1967; 6: 361 /7. 7. Hall E, Wuu CS. Radiation-induced second cancers: the impact of 3D-CRT and IMRT. Int J Radiat Oncol Biol Phys 2003; 56: 83 /8. 8. Munzenrider JE, Liebsch NJ. Proton therapy for tumors of the skull base. Strahlenther Onkol 1999; 175(Suppl 2): 57 /63. 9. Terehara ANA, Goitein M, Finkelstein D, et al. Analysis of the relationship between tumor dose inhomogeneity and local control in patients with skull base chordoma. Int J Radiat Oncol Biol Phys 1999; 45: 351 /8. 10. Rosenberg AE, Nielsen GP, Keel SB, et al. Chondrosarcoma of the base of skull: a clinicopathologic study of 200 cases with emphasis on its distinction from chordoma. Am J Surg Pathol 1999; 23: 1370 /8. 11. Gragoudas E, Li W, Goitein M, Lane AM, Munzenrider JE, Egan KM. Evidence-based estimates of outcome in patients irradiated for intraocular melanoma. Arch Ophthalmol 2002; 120: 1665 /71. 12. Egger E, Schalenbourg A, Zografos L, et al. Maximizing local tumor control and survival after proton beam radiotherapy of uveal melanoma. Int J Radiat Oncol Biol Phys 2001; 51: 138 / 47. 13. Bush D, Slater JD, Bonnet R, et al. Proton beam radiotherapy for early-stage lung cancer. Chest 1999; 116: 1313 /9. 14. Shipley WU, Verhey LJ, Munzenrider JE, et al. Advanced prostate cancer: the results of a randomized comparative trial of high dose irradiation boosting with conformal protons compared with conventional dose irradiation using photons alone. Int J Radiat Oncol Biol Phys 1995; 32: 3 /12. 15. Fitzek MM, Thornton AF, Rabinov JD, et al. Accelerated fractionated proton/photon irradiation to 90 cobalt gray equivalent for glioblastoma multiforme: results of a phase II prospective trial. J Neurosurg 1999; 91: 251 /60. 16. Lomax AJ, Cella L, Weber D, Kurtz JM, Miralbell R. Potential role of intensity-modulated photons and protons in the treatment of the breast and regional nodes. Int J Radiat Oncol Biol Phys 2003; 55: 785 /92. 17. Goitein M, Jermann M. The relative costs of proton and x-ray radiation therapy. Clin Oncol 2003; 15: S37 /50.