Radiotherapy and Oncology, 17 (1990) 29-35 Elsevier

29

RADION 00606

Considerations in fractionated proton radiation therapy: clinical potential and results M. Austin-Seymour, M. Urie, J. Munzenrider, C. Willett, M. Goitein, L. Verhey, R. Gentry, P. McNulty, A. Koehler and H. Suit Department of Radiation Medicine, Massachusetts GeneralHospital Boston, Massachusetts, U.S.A. and Harvard Cyclotron Laboratory, Cambridge,Massachusetts, U.S.A. (Received 5 October 1988; revision received 6 February 1989; accepted 17 February 1989)

Key words: Proton radiation therapy; Comparative treatment planning

Summary Protons have a finite range in tissue and can provide a better concentration of radiation dose in the tumor than conventional X-rays in certain situations. The development of optimized treatment plans for X-rays and protons followed by a comparative evaluation is one method of selecting tumor sites best suited for proton treatment. The preliminary results of comparative treatment planning for base of skull tumors and carcinoma of the prostate are discussed. These comparisons suggest a clinical gain for proton treatment of tumors in these locations. The clinical experience with fractionated proton treatment of several tumor sites is also discussed. The results of high dose proton treatment ofchordomas and low grade chondrosarcomas of the base of skull is particularly promising: an actuarial 5-year local control of 7 8 ~ has been obtained in 50 patients followed for a minimum of 22 months.

Introduction Protons have a finite range in tissue. This quality distinguishes protons from X-rays and allows the use of unique beam arrangements. Modulated protons deliver a homogeneous dose to a defined target and can provide a better concentration of radiation dose in the tumor than conventional Addressfor correspondence: Dr. Herman D. Suit, Department of Radiation Medicine, Cox 3, Massachusetts General Hospital, Boston, MA 02114, U.S.A.

X-rays when the tumor is close to a radiosensitive normal structure. The physical characteristics of protons combined with meticulous treatment planning and accurate execution of the plan contribute to the clinical usefulness of protons. At the outset certain tumor sites, such as paraspinal tumors, appear to be particularly attractive for proton treatment and indeed paraspinal tumors and tumors of the base of the skull have been among the first categories of clinical investigation. On the other hand, it is not obvious that protons would be of benefit at every tumor

0167-8140/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

30 site and finite resources limit the number of sites which can be studied. One method of selecting tumors for proton treatment is the use of comparative treatment planning. Optimized treatment plans for X-rays and protons are developed using the multidimensional planning system described below and are then evaluated. This is an area of intense effort currently at Massachusetts General Hospital (MGH), and some of the initial comparisons are described in this manuscript.

Materials and methods

It is necessary to understand the multidimensional treatment planning system prior to evaluating treatment plans [6,7]. This system is currently in routine use for proton planning and particularly challenging photon cases at MGH. The proton patient is initially immobilized in either a supine or seated position with an alpha cradle and thermoplastic face mask. This immobilization technique is used because most sites requiring planning at this time are located in cervical or cranial regions. A C T scan is then done with the immobilized patient in the treatment position. The scan is performed with a special CT scanner which has been modified to scan patients in either a supine or seated position. The fixed horizontal proton beam necessitates treatment in the seated position in some cases. The slice thickness is 3 mm at 3 mm increments through the region of the tumor. The tumor volume and the relevant normal structures are then outlined on each slice of the treatment planning scan using an interactive computerized treatment planning system [6]. All available information, including the radiographic studies, clinical findings, and operative findings are used in this task. Treatment fields for the tumor are designed which minimize the dose to designated normal structures. The beam's eye view is a powerful tool for this purpose [7]. Brass apertures are used to shape the field edges and Lucite compensators are designed to shape the

distal end of the beam range to the tumor [4,24]. The aperture and the compensator are both produced by a computer controlled milling machine [5]. Distributions of dose using all treatment fields are computed on each CT slice. Before each treatment, the position of the tumor relative to the beam aperture is verified radiographically using an iterative process. The final step in this process is the careful comparison of a portal film and an alignment film generated from the CT data [7]. The field placement accuracy is within 2 mm. The mean movements for an immobilized patient are less than 1 mm [25]. On the average, each dally set-up and treatment takes 45-60 min. The process of comparative treatment planning involves formulating optimal treatment plans for X-rays and for protons using this multidimensional treatment planning. It is assumed that these treatment plans will be delivered using the iterative radiographic set-up procedures described above. An additional assumption is the availability of protons with sufficient range for the plans.

Results

In Figure 1, proton and 10 MV X-ray treatment plans for a base of skull tumor are represented. This tumor is immediately adjacent to the brainstem and the left temporal lobe. The transverse CT slice also shows the anatomic relationship between the tumor and both globes and optic nerves. The dose to the tumor in both plans is 70 Gy. In both plans a posterior oblique beam has been used to spare the brainstem; right and left lateral fields have also been used. The proton dose distribution is more closely contoured to the tumor volume than the X-ray plan. With protons, the dose to the right globe, optic nerve, and temporal lobe is considerably less than with X-rays. Figure 1 shows that the dose to the right optic nerve is 62 Gy in the X-ray plan and only 31 Gy in the proton plan. This difference is due to the finite range in tissue of protons; the

31

Fig. I(A) A proton treatment plan for base of skull tumor; (B) a 10 MV X-ray plan. The tumor dose is 70 Gy in both plans. Dose is represented in 10 tones of gray with each change corresponding to a 6 Gy decrement in dose. In addition, dose is specified numerically. The arrow points to the right optic nerve. The dose at the tip of the arrow is 31 Gy in the proton plan and 62 Gy in the X-ray plan.

likelihood of radiation optic neuropathy on the fight should be significantly reduced with protons. With X-rays, the dose to the right retina is as high as 30 Gy, while with protons the retina receives essentially no dose. In both plans there is a gradient of dose in the right temporal lobe. In the proton plan this gradient ranges from 16 to 28 Gy, which contrasts with the gradient of 34-50 Gy in the X-ray plan. Figure 2 shows two plans for the treatment of carcinoma of the prostate. The proton plan has four proton fields (AP, PA, right lateral and left lateral) for the comprehensive treatment of the prostate and adjacent nodal area; a perineal field is used for the prostate tumor boost. The X-ray plan uses an unevenly weighted 4-field box technique for treatment of the prostate and nodes and right and left lateral fields for the prostate tumor boost. The particular CT slice shows the portion of a locally advanced prostate cancer which invades the seminal vesicles. A circular probe is present in the rectum in order to distend the rectum posteriorly. In both plans a dose of 70 Gy is delivered to the prostate tumor. With protons the dose to much of the rectum, the bladder, and both hip joints is considerably less than with X-rays. Because the prostate cancer is immediately adjacent to the anterior wall of the rectum and the posterior wall of the bladder, the dose to the portion of these structures is 70 Gy in both plans. However, with protons a smaller portion of these structures receives 70 Gy than with X-rays. In both plans there is a gradient of dose across the posterior recto-sigmoid, the bladder, and the hip joints. For example, in the X-ray plan, the dose ranges from 45 to 55 Gy in the posterior recto-sigmoid while with protons the dose is less than 45 Gy. In addition to tumors of the base of skull and prostate tumors, other tumor sites currently being evaluated using comparative treatment planning include lung, esophagus, pancreas, nasopharynx, thyroid, and cervix. Tools such as dose volume histograms are being used to aid in the comparison process [ 1].

32 Discussion

Fig. 2(A) A proton plan for a locally advanced carcinoma of the prostate invading seminal vesicles; (B) a 10 MV X-ray plan for the same tumor. The tumor dose is 70 Gy in both plans with each change of gray tone corresponding to a 5 Gy decrement in dose.

The comparative treatment planning process described above predicts an advantage for proton treatment of base of skull tumors and prostate carcinoma. Clinical experience with proton treatment of these sites and other tumor sites has been gained through a collaborative program between the Radiation Medicine Service at Massachusetts General Hospital, the Retina Service at Massachusetts Eye and Ear Infirmary (MEEI) and the Harvard Cyclotron Laboratory (HCL) [13,14,18-20]. This program was initiated in 1974. At HCL, protons are accelerated to 160 MeV, which have a 15.9 cm range in soft tissue. The dose decreases very rapidly at the distal end of the beam range: the dose drops from 80 to 20~o in 4 mm. The lateral penumbra is also sharp with the distance between the 90 and 20~o varying between 3.0 and 8.5 mm, depending on beam conditions and depth. The relative biological effectiveness (RBE) of modulated protons relative to cobalt-60 has been studied using several assays [20-23]. An RBE value of 1.1 has been used for clinical proton radiation therapy. Dose is expressed in the units of Cobalt Gray Equivalent (CGE) and is obtained by multiplying the physical dose in Gy by 1.1. As of December 31, 1986, a total of 1462 patients had received fractionated proton treatments (Table I). Informed consent was obtained from all patients prior to treatment. Over 1000 of these patients had uveal melanomas, and the success of proton treatment is well established [8-10]. Other major categories of patients include 109 with chordomas or chondrosarcomas, 96 with prostate cancers, and 63 with sarcomas of soft tissue. The range limitation of 15.9 cm has played an important role in selection of patients: treatment of tumor sites in the chest and abdomen has rarely been possible. Most of the patients with chordomas or chondrosarcomas have had tumors in the base of the skull; a total of 88 such patients have received proton treatment. These are rare tumors which

33 TABLE I Patients receiving fraetionated proton treatments (MGH/ MEEI/HCL- January 1974-December 1986). Category

Number

Uveal melanoma Chordoma/chondros arcoma Sarcoma Soft tissue Bone Prostate Non-randomized Randomized Other CNS Head and neck Miscellaneous Ano-rectal carcinoma Non-uveal melanomas

1006 109 63 16 64 53 43 37 40 18 13 TotN

1462

are locally aggressive [ 11 ]. Metastases occur infrequently. These tumors occur in close proximity to critical structures of the central nervous system, including the brainstem, spinal cord, the optic chiasm, and the brain itself. They almost always regrow following surgical resection alone since radical resection is rarely possible [12]. When radiation treatments are given post-operatively using conventional treatment modalities, most patients experience local recurrence of their tumor [2,16]. A dose of 55 Gy is commonly given using megavoltage techniques. Based on a review of the literature, an estimate of the local control rate at 5 years is 3 5 ~ [2,15-17]. Treatment techniques similar to those shown in Fig. 1 have been used for the proton treatments of these patients. Fifty patients have a minimum follow-up time of 22 months. There were 27 chordomas and 23 grade I-II chondrosarcomas. The median tumor dose was 69 CGE with a range from 56.9 to 74.4 CGE. This is 25 ~ higher than the median dose for cases reported in the literature using conventional treatment modalities. The treatments were given over approximately 8 weeks with a daily dose of 1.8-2.0 CGE. Followup has been by clinical examinations and CT scans. The median follow-up time is 34 months.

The 5-year actuarial local control rate is 78~o. Six patients have developed new neurologic symptoms with an increase in tumor size on CT at 8-33 months. In addition to these local failures, one patient developed recurrent disease anterior to the original lesion and another developed metastatic disease in a supraclavicular lymph node. The actuarial disease-free survival rate at 5 years is 7 6 ~ . Five patients have developed complications producing functional impairment. One patient developed unilateral blindness and seizures associated with an enhancing area in the temporal lobe on CT. Another patient with diabetes mellitus developed bilateral visual loss. An additional patient developed the sudden onset of unilateral blindness. Two other patients have chronic and severe headaches. Proton treatment appears to be an effective treatment modality for these tumors. The 5-year actuarial local control of 78 ~o compares favorably with the results of conventional radiation techniques. Certainly, further follow-up is necessary to judge the long-term efficacy of these treatments. Sixty-three patients with soft tissue sarcomas have received proton treatment. Most patients received this as a component of their treatment, but some patients were treated entirely with protons. These patients were selected for treatment because certain critical normal tissues could be excluded from the high dose treatment volume by the use of protons. These critical tissues include the spinal cord, the bowel, the ovaries, the globes, and certain joints. Overall, 54 out of 63 patients have local control of their tumor with a mean follow-up time of 36 months. Patients with paraspinai sarcomas have especially benefited by the use of protons with 14 out of 15 patients achieving local control. The 15 patients with paraspinal sarcomas have had proton treatment using techniques similar to those illustrated in Fig. 1. These patients have been followed for a mean of 35 months with a range from 6 to 96 months. All patients had surgical removal of as much tumor as feasible. The

34 median dose is 66 CGE, which is well above the threshold dose for spinal cord damage. One patient had local regrowth of tumor 6 months after treatment and simultaneously developed distant metastases. Another patient recurred just inferior to the original tumor volume at 6 months and also developed distant metastases. No patient has had treatment related spinal cord damage. Protons have been used to treat locally advanced prostate carcinomas using a perineal field similar to the boost field shown in Fig. 2. Unlike the proton plan in Fig. 2, the prostate tumor and adjacent nodal areas are treated using X-rays prior to the perineal boost. With this treatment technique, the dose to the posterior rectal wall is significantly reduced compared to X-rays alone. A phase II study was done to determine whether or not a 10~o higher dose could be given without an increase in complications. Two groups of patients treated concurrently were compared: 118 patients were treated with conventional megavoltage technique to 68.4Gy, and 64 patients were boosted with protons after X-ray treatment to the pelvis to a total dose 73.5 CGE. Local control was slightly better in the proton boost patients, while a similar incidence of bladder and rectal complications was observed in both groups [3]. A randomized trial was initiated in December 1981. Patients with prostatic carcinoma extending beyond the capsule receive 50.4 Gy to the pelvis in 28 fractions with either 4 or 25 MV X-rays and then are randomized to receive either 12 fractions of 2.1 CGE with a perineal proton field or eight 2.1 Gy fractions with opposed lateral 25 MV X-rays to total doses of 75.6 or 67.2 Gy, respectively. Between December 1981 and December 1986, 117 patients were randomized. This is approximately 6 0 ~ of the number required to demonstrate a 15 ~o difference in the local control between the two groups. In summary, the comparative treatment planning process facilitates an exploration of the theoretical advantages of protons. In certain tumor sites the advantages of protons have been

substantiated by the M H G - H C L clinical experience. The comparative treatment planning method aids in the selection of sites for future clinical investigations.

Acknowledgements This work supported by Grant CA 21239 for the National Cancer Institute, D H H S and by contract Particle Treatment and Planning from the National Cancer Institute, D H H S .

References 1 Austin-Seymour, M., Chen, G.T.Y., Castro, J.R., Saunders, W.M., Pitluck, S., Woodruff K.H. and Kessler, M. Dose volume histogram analysis of liver radiation tolerance. Int. J. Radiat. Oncol. Biol. Phys. 12: 31-35, 1986. 2 Cummings,B.J., Hodson, D. I. and Bush, R.S. Chordoma: the results of megavoltageradiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 9: 633-642, 1983. 3 Duttenhaver, J.R., Shipley, W.U., Perrone, T.L. et al. Protonsor megavoltagex-raysas boosttherapyfor patients irradiated for localizedprostatic carcinoma: an earlyphase I-II comparison.Cancer51: 1599-1604, 1983. 4 Goitein, M. Compensation for inhomogeneities in charged particle radiotherapyusing computed tomography. Int. J. Radiat. Oncol. Biol. Phys. 4: 499-508, 1978. 5 Goitein,M., Abrams, M., Gentry, R., Uric, M., Verhey, L. and Wagner, M. Planning treatment with heavy charged particles. Int. J. Radiat. Oncol. Biol. Phys. 8: 2065-2070, 1982. 6 Goitein, M. and Abrams, M. Multi-dimensionaltreatment planning.I. Delineationof anatomy.Int. J. Radiat. Oncol. Biol. Phys. 9: 777-787, 1983. 7 Goitein, M. and Abrams, M. Multi-dimensionaltreatment planning.II. Beam'seye-view,back projection and projection through CT sections. Int. J. Radiat. Oncol. Biol. Phys. 9: 789-797, 1983. 8 Gragoudas, E. S., Goitein, M., Koehler, A., Constable, I. J., Wagner,M. S., Verhey,L., Tepper,J. E., Suit, H. D., Brockhurst, R.J., Schneider, R.J. and Johnson, K.N. Proton irradiationofchoroidalmelanomas.Arch. Ophthalmol. 96: 1583-1591, 1978. 9 Gragoudas,E. S., Goitein,M., Verhey,L., Munzenrider, J. E., Urie,M., Suit,H. D. and Koehler,A. Proton beam irradiation of uveal melanomas.Arch. Ophthalmol. 100: 928-934, 1982.

35 10 Gragoudas, E. S., Seddon, J. M., Egan, K., Glynn, R., Munzenrider, J., Austin-Seymour, M., Goitein, M., Verhey, L., Uric, M. and Koehler, A. Long term results of proton beam irradiated uveal melanomas. Ophthalmology, 94: 349-353, 1987. 11 Heffelfinger, M. J., Dahlin, D. C., MacCarty, C. S. and Beabout, J.W. Chordomas and cartilaginous tumors at the skull base. Cancer 32: 410-420, 1973. 12 Higinbotham, N.L., Phillips, R.F., Farr, H.W. and Hustu, H.O. Chordoma. Cancer 20: 1841-1850, 1967. 13 Munzenrider, J. E., Shipley, W. U. and Verhey, L.J. Future prospects of radiation therapy with protons. Semin. Oncol. 8: 110-124, 1981. 14 Munzenrider, J.E., Austin-Seymour, M.M., Blitzer, P. J., Gentry, R., Goitein, M., Gragoudas, E. S., Johnson, K., Koehler, A. M., McNulty, P., Moulton, G., Osborne, E., Seddon, J. M., Suit, H. D., Urie, M., Verhey, L. J. and Wagner, M. Proton therapy at Harvard. Strahlentherapie 161: 756-763, 1985. 15 Perzin, K. H. and Pushparaj, N. Nonepithelial tumors of the nasal cavity, paranasal sinuses, and nasopharynx. A clinicopathologic study. XIV: Chordomas. Cancer 57: 784-796, 1986. 16 Phillips, T. L. and Newman, H. Chordomas. In: Modern Radiotherapy and Oncology: Central Nervous System Tumors, pp. 184-203. Editor: T. H. Deeley, Butterworth, Boston, 1974. 17 Saxon, J. P. Chordoma. Int. J. Radiat. Oncol. Biol. Phys. 7: 913-915, 1981. 18 Suit, H. D., Goitein, M. G., Tepper, J. E. and Verhey, L. Clinical experience and expectation with protons and heavy ions. Int. J. Radiat. Oncol. Biol. Phys. 3:115-125, 1977.

19 Suit, H. D., Goitein, M., Munzenrider, J. E., Verhey, L., Gragoudas, E., Koehler, A.M., Urano, M., Shipley, W.U., Linggood, R.M., Friedberg, C. and Wagner, M. Clinical experience with proton beam radiation therapy. J. Can. Assoc. Radiol. 31: 35-39, 1980. 20 Suit, H. D., Goitein, M., Munzenrider, J. E., Verhey, L., Blitzer, P., Gragoudas, E., Koehler, A.M., Uric, M., Gentry, R., Shipley, W., Urano, M., Duttenhaver, J. and Wagner, M. Evaluation of the clinical applicability of proton beams in definitive fractionated radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 8: 2199-2205, 1982. 21 Tepper, J. E., Verhey, L., Goitein, M., Suit, H. D. and Koehler, A.M. In vivo determinations of RBE in a higher energy modulated proton beam using normal tissue reactions and fractionated dose schedules. Int. J. Radiat. Oncol. Biol. Phys. 2:1115-1122, 1977. 22 Urano, M., Goitein, M., Verhey, L., Mendiondo, O., Suit, H. D. and Koehler, A. Relative biological effectiveness of a high energy modulated proton beam using a spontaneous murine tumor in vivo. Int. J. Radiat. Oncol. Biol. Phys. 6: 1187-1193, 1980. 23 Urano, M., Verhey, L. J., Goitein, M., Tepper, J. E., Suit, H. D., Mendiondo, O., Gragoudas, E. S. and Koehler, A. Relative biological effectiveness of modulated proton beams in various murine tissues. Int. J. Radiat. Oncol. Biol. Phys. 10: 509-514, 1983. 24 Urie, M., Goitein, M. and Wagner, M. Compensating for heterogeneities in proton radiation therapy. Phys. Med. Biol. 29: 553-556, 1983. 25 Verhey, L.J., Goitein, M., McNulty, P., Munzenrider, J. E. and Suit, H.D. Precise positioning of patients for radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 8: 289-294, 1981.