J.R. Castro

EXPERIENCE IN CHARGED PARTICLE IRRADIATION OF TUMORS OF THE SKULL BASE: 1977-1992

PURPOSEnTo review the experience at University of California Lawrence Berkeley Laboratory in using charged particles to irradiate primary neoplasms of the skull base and those extending to the skull base from the nasopharynx and paranasal sinuses.nnnMETHODS AND MATERIALSnDuring the period from 1977 to 1992, 223 patients were irradiated with charged particles at the Lawrence Berkeley Laboratory for tumors either arising in or extending to the skull base, of whom 48 (22%) had recurrent lesions, either post previous surgery or radiotherapy. One hundred twenty-six patients had lesions arising in the cranial base, mostly chordoma (53), chondrosarcoma (27), paraclival meningioma (27) with 19 patients having other histologies such as osteosarcoma or neurofibrosarcoma. There were also 31 patients with primary or recurrent squamous carcinoma of the nasopharynx extending to the skull base, 44 patients with major or minor salivary gland tumors, mostly adenocarcinoma, and 22 patients with squamous carcinoma of the paranasal sinuses, all with cranial base extension.nnnRESULTSnLocal control and survival appeared improved in tumors arising in the skull base, following the ability with charged particles to deliver high doses (mean of 65 Gy-equivalent) with relative sparing of the adjacent normal tissues. The Kaplan-Meier 5-year local control was 85% for meningioma, 78% for chondrosarcoma, 63% for chordoma and 58% for other sarcoma. Follow-up ranged from 4-191 months with a median of 51 months.nnnCONCLUSIONnCharged particle radiotherapy is highly effective in controlling cranial base lesions which have have been partially resected. Better tumor localization with CT and MRI, improved 3-D treatment planning and beam delivery techniques have continued to reduce the level of serious complications and increase local control and survival.

Dose volume histogram analysis of liver radiation tolerance.

Eleven patients with carcinoma of the pancreas or biliary system received heavy charged particle radiation treatments and whole liver heavy charged particle radiation at Lawrence Berkeley Laboratory. Doses to the whole liver ranged from 10 to 24 Gray-equivalent (the biological equivalent of 10 to 24 Gray of low-LET photon radiation), whereas the dose to the primary lesion ranged from 53.5 to 70 Gray-equivalent (GyE). The fraction size was 2 to 3 GyE. The liver received partial as well as whole organ irradiation. Integral dose volume histograms for the liver were obtained in all 11 patients. An integral dose volume histogram displays on the ordinate the percentage of liver that was irradiated in excess of the dose specified on the abcissa. In this study, the clinical liver radiation tolerance of these patients is correlated with the information contained in an integral dose volume histogram. One patient developed radiation hepatitis. The integral dose volume histogram of this patient differed from the dose volume histograms of the other 10 patients. This difference was greatest in the range of doses between 30 and 40 GyE. Our results suggest that liver doses in excess of 30 to 35 GyE should be limited to 30% of the liver or less when 18 GyE of whole liver radiation is delivered at 2 GyE per fraction in addition to primary radiation of the pancreas or biliary system.

Definitive postoperative irradiation of bile duct carcinoma with charged particles and/or photons☆

PURPOSEnTo determine the rates of survival and local control in patients with bile duct adenocarcinomas treated with post-operative photons and/or charged particles.nnnMETHODS AND MATERIALSnA retrospective study was performed analyzing all patients with bile duct adenocarcinomas who received radiotherapy through the University of California San Francisco and at Lawrence Berkeley Laboratory between 1977 and 1987, a total of 62 patients. University of California San Francisco patients received photon therapy (median dose 5400 cGy), and Lawrence Berkeley Laboratory patients were treated with the charged particles helium and/or neon (median dose 6000 cGyE). Forty-eight patients were treated post-operatively with curative intent, 30 with photons and 18 with particles. Thirty-six patients in the study had gross residual disease; none had microscopically negative margins.nnnRESULTSnThe overall two-year actuarial survival was 28%: 44% for particle-treated patients and 18% for patients treated with photons (p = .048). Median actuarial survival was 23 months in particle patients and 12 months in photon patients. Local control was also improved, though less significantly, in patients treated with particles (median disease-free survival 20 months vs. 4.5 months, p = .054). A univariate and multivariate analysis was performed and revealed that only extent of residual disease predicted local failure and overall survival; no other prognostic factors were identified.nnnCONCLUSIONnCompared to conventional photon radiotherapy, treatment with post-operative charged particle irradiation at Lawrence Berkeley Laboratory appeared to offer a survival advantage in this non-randomized series. Additional investigation into protection of surrounding normal tissue with better dose localization through the use of charged particles is planned.

Design of beam‐modulating devices for charged‐particle therapy

The computer modeling program used to design beam-modulating devices for charged-particle therapy at Lawrence Berkeley Laboratory has been improved to allow a more realistic description of the beam. The original program used a single calculated Bragg peak to design the spread Bragg peak. The range of this curve was shifted so that Bragg curves of varying ranges could be superimposed. The new version of the program allows several measured Bragg curves with different ranges to be used as input, and interpolates between them to obtain the required data for the superposition calculation. The experimental configuration for measuring these input curves simulated therapy conditions. Seven beam-modulating propellers with spread Bragg-peak widths ranging from 2.2 to 14.4 cm were designed and constructed for a 215-MeV/u helium beam using this new design program. Depth-dose distributions produced by these new propellers were in good agreement with predicted distributions, and these propellers are currently being used clinically.

Helium-ion radiation therapy at the Lawrence Berkeley Laboratory: recent results of a Northern California Oncology Group Clinical Trial.

We report on the first decade of the helium-ion radiotherapy clinical trial being carried out at the Lawrence Berkeley Laboratory. Over 500 patients have now been treated. We have had very good results to date in treating patients with small tumors critically located near a radiation-sensitive organ which would preclude delivering a curative dose with conventional radiotherapy. On the other hand, patients with larger tumors where the tumor dose cannot be increased more than 10% over conventional radiotherapy have not responded well to helium ion radiotherapy. This is illustrated by discussing selected patient groups in detail, namely those with uveal melanoma, small, low-grade tumors near the central nervous system, carcinoma of the pancreas, and carcinoma of the esophagus.

Preliminary results in heavy charged particle irradiation of bone sarcoma

Between 1979 and 1989, 17 patients with unfavorable bone sarcoma were treated wholly or in part with heavy charged particle irradiation (helium and/or neon ions) at the University of California Lawrence Berkeley Laboratory. The majority of tumors were located near critical structures such as the spinal cord or brain. Gross tumor was present in all but two patients at the time of irradiation. Six patients were treated for recurrent disease. Histologies included osteosarcoma, Ewings sarcoma, and recurrent osteoblastoma. Four of the osteosarcomata were believed to have been induced by previous therapeutic irradiation for various tumors. Follow-up time since initiation of radiation ranged from 7 to 118 months (median 40 months). The 5-year Kaplan-Maier local control rate was 48%; the corresponding survival rate was 41%. Over half the patients succumbed to distant metastases despite the majority of patients receiving chemotherapy. In this preliminary study, we have shown that heavy charged particle irradiation can be effectively used for control of bone sarcoma. A Phase II trial is warranted to determine optimal treatment for unresectable or gross residual disease.

Sensitivity of helium beam-modulator design to uncertainties in biological data.

The goal in designing beam-modulating devices for heavy charged-particle therapy is to achieve uniform biological effects across the spread-peak region of the beam. To accomplish this, the linear-quadratic model for cell survival has been used to describe the biological response of the target cells to charged-particle radiation. In this paper, the sensitivity of the beam-modulator design in the high-dose region to the values of the linear-quadratic variables alpha and beta has been investigated for a 215-MeV/u helium beam, and implications for higher LET beams are discussed. The major conclusions of this work are that, for helium over the LET range of 2 to 16 keV/mu, uncertainties in measuring alpha and beta for a given cell type which are of the order of 20% or less have a negligible effect on the beam-modulator design (i.e., on the slope of the spread Bragg peak); uncertainties less than or equal to 10% in the dose-averaged LET at each depth are unimportant; and, if the linear-quadratic variables for the tumor differ from those used in the beam-modulator design by a constant factor between about 0.5 and 3, then the resultant nonuniformity in the photon-equivalent dose delivered to the tumor is within +/- 25%. It is also shown that for any ion, if the nominal values of alpha or beta used by the beam-modulator design program differ from their actual values by a constant factor, then the maximum errors possible in the beam-modulator design may be characterized by two limiting depth-dose curves such that the ratio of the dose at the proximal end of the spread Bragg curve to the dose at the distal end of the spread peak is given by alpha distal/alpha prox for the steepest curve, and square root of beta distal/beta prox for the flattest curve.

The effect of patient motion on dose uncertainty in charged particle irradiation for lesions encircling the brain stem or spinal cord.

A specialized charged-particle radiotherapy technique developed at Lawrence Berkeley Laboratory (LBL) is applied to patients with lesions abutting or surrounding the spinal cord or brain stem. This technique divides the target into two parts, one partially surrounding the critical structure (brain stem or spinal cord) and a second excluding the critical structure and abutting the first portion of the target. Compensators are used to conform the dose distribution to the distal surface of the target. This technique represents a novel approach in treating unresectable or residual tumors surrounding the spinal cord or brain stem. Since the placement of the patient with respect to beam-shaping devices is critical for divided-target treatments, a method for calculating dose distributions reflecting random patient motion is proposed, and the effects of random patient motion are studied for two divided-target patient examples. Dose-volume histograms and a normal-tissue complication probability model are used in this analysis. For the patients considered in this study, the normal-tissue-complication probability model predicts that random patient motion less than or equal to 0.2 cm is tolerable in terms of spinal cord complications.

Evaluation of fixed- versus variable-modulation treatment modes for charged-particle irradiation of the gastrointestinal tract

The clinical usefulness of variable-modulation dose delivery of neon ion and proton beams over fixed-modulation beams is evaluated for several patients with tumors in the gastrointestinal tract by comparing dose distributions, dose volume histograms, and predictions of normal tissue complication probabilities calculated with the two methods. Both techniques provide excellent coverage of the target volume with neon ion and proton beams. The advantage of variable modulation is that less dose is delivered proximal to the target volume. For tumors in the gastrointestinal tract, this implies that less dose is given to the liver, gut, kidneys, and lungs. For the ten patients considered in this study, variable-modulation reduced the total integral dose by an average of 17% for neon ion beams and by 18% for protons as compared to fixed-modulation. If the tumor volume is excluded, the reduction in the integral dose to normal tissues ranged from 15% to 32% for neon ions and from 18% to 34% for proton beams. These gains are larger than those anticipated on the basis of an analytic study by Goitein and Chen [Med. Phys. 10, 831-840 (1983)], which predicted integral dose reductions of the order of 10% for protons and 14% for neon ions. They are also larger than those reported in a similar study by Urie and Goitein [Med. Phys. 16, 593-601 (1989)] for proton irradiation of skull-base tumors. This is probably because the tumors in the GI tract considered in this study were more irregularly shaped than Goitein and Chens analytic model assumes. The results of this study also suggest that due to increased sparing of normal tissues, the number of different portal directions required to achieve a satisfactory treatment plan will be reduced for variable-modulation beam delivery systems. This implies that variable-modulation treatment plans will be easier to execute than current fixed-modulation plans.