144 Radiation Oncology ??Biology ??Physics November 1986,Volume 12, Sup. 1

days over 7 weeks. To date 45 patients (aged 2-57 years) have been entered on study with a median followup of 51 weeks. Only 5 of 45 patients (11%) have progressed to date. Toxicity has been minimal with 2 patients demonstrating local tumor necrosis and neurologic deterioration managed with steroids and subsequent clinical and scan improvement. These results will be compared with our previous studies, but are currently significantly better than any previously reported treatment using conventional radiation therapy with or without chemotherapy.

103

RADIOTHERAPEUTIC CONSIDERATIONS IN THE TREATMENT OF HEMANGIOBLASTOMAS OF THE CENTRAL NERVOUS SYSTEM

Stephen R. Smalley, M.D., Paula J. Schomberg, M.D., John D. Earle, M.D., Edward R. Laws, Jr., M.D.

Mayo Clinic and Mayo Foundation, Rochester, Minnesota 55905

Ninety-six patients (pts) with hemangioblastoma of the central nervous system (CNS) were seen at our institution from January 1963 to August 1983. Twenty-seven of these (15 males, 12 females) received radiation therapy and are the subject of this report. Six pts had von Hippel-Lindau Syndrome, and 4 patients presented with polycythemia secondary to the hemangioblastoma. The median age at diagnosis was 48 years (range 20-68). Two clinical groups were apparent; those that received post-operative radiation therapy for clinically suspect, or microscopically positive margins (6 pts) and those who underwent radiation therapy (XRT) for gross residual disease (21 pts). Median follow-up was 99 months following radiation therapy. Those with gross residual disease were particularly unfavorable in that 12/21 patients in this group had manifested a-total of 17 local recurrences prior to XRT. Six of the 21 pts with oross residual underwent subtotal resection (STR) prior to XRT. One patient initially explored and found unr&ectable was rendered resectable after planned pre-operative XRT (40 Gy, 2 Gylfx). This patient remains alive and free of disease 108+ months following his radiation therapy. Because the combined modality approach did not allow assessment of local control, he was excluded from the gross residual cohort in terms of time-dose relationship analysis.

The cohort with gross residual disease appeared to a demonstrate time-dose relationship. We divided the XRT cohort into 3 groups on the basis of TDF administered; 1) high dose group - TDF greater than 75, 2) moderate dose group - TDF of 65-75 and 3) low dose group - TDF less than 65. The high dose group achieved local control in 78% (7/9) with a median local disease free interval (DFI) of 164+ months. The moderate dose group achieved local control in only 22% (2/9) with a median DFI of 67 months. One patient received a TDF of less than 65 and recurred locally in 4 months. Thirty-three percent of both the high dose and low dose group received STK prior to XRT. The six patients who underwent STR prior to XRT achieved a 50% local control rate and a 98 mo median DFI (mean DFI 141+ mo) while those without STR achieved local control in 38% with a median DFI of 66 mo (mean DFI 69+ mo). The 6 pts with microscopic residual or clinically suspect margins seemed to require less aggressive irradiation. Two of 2 pts in the moderate dose group maintained local control (DFI 82+, 226+ mos) as did 2/3 high dose pts (DFI 92+, 183+ mos). One patient treated with a low TDF (34) recurred with a DFI of 101 months.

Disease relapse was observed in 13 patients. One failed in an area out of the previous XRT field (second primary) and one marginal failure was recorded. Local recurrence occurred after long disease free intervals (median DFI - 67 months). Local recurrence, though delayed in its time course, was an adverse finding with 8 of 12 patients dying secondary to either disase progression (7 pts) or post-operative mortality from attempted surgical salvage (1 pt.). Three local failures were salvaged with surgery and are alive and disease free at 37.5+, 109+ and 247+ mos following relapse.

We conclude that hemangioblastomas of the CNS are amenable to XRT when optimal surgical resection is not possible. Aggressive therapy with TDFs of greater than 75 is necessary if gross residual disease is present. Subtotal resection, if technically feasible, may be beneficial. Local failure can occur after long disease free intervals, and carries an unfavorable prognosis, although surgical salvage is possible.

104 LONG TERM RESULTS OF A PILOT STUDY OF LOW DOSE CRANIAL-SPINAL RADIATION FOR CEREBELLAR MEDULLOBLASTOMA

Brand, William N., Schneider, Philip A. and Tokars, Roger P.

Radiation Oncology Center, Northwestern Memorial Hospital, Chicago, Illinois

Between May 1974 and March 1983, 44 children with histologically verified cerebellar medulloblastoma were seen for post-operative cranial-spinal irradiation following attempted total tumor removal. 6 patients were excluded from review because they received all or part of their treatment at another institution (3) or did not complete the planned course of irradiation (3). All of the 38 remaining patients were treated by a pre- viously described technique on a 4 MeV Linear Accelerator with 55 Gy delivered to the primary tumor site. Prior to December 1978, 19 consecutive children (Group A) had spinal prophylactic doses of 30-40 Gy and brain pro- phylactic doses of 40-50 Gy. After that date, 25 Gy was given to the cranial-spinal axis of 19 consecutive children (Group B). This lower dose was arbitrarily selected with the hope of reducing morbidity in treated

Proceedings of the 28th Annual ASTRO Meeting 145

survivors and achieving the same tumor control. Risk factors that define good and poor prognosis were evalu- ated for each group, and there were no differences noted. Myelography and CSF cytology were not routinely performed. Follow-up for the 38 patients ranges from 20 months to 124 months. For the low risk patients, survival (12/15 or 80%) was independent of cranial-spinal radiation dose (Group A 6/8, Group 3 6/7). For the high risk patients survival was poor (9/23 or 39%), not dependent on cranial-spinal radiation dose (Group A 5/11, Group B 4/12), and associated with failure at the primary site (10/14), often with CSF seeding (8/10). The other 4 failures include 2 who had moved outside the United States (details of failure are unknown), 1 with supratentorial, CSF seeding and distant metastases, and 1 with distant metastasis only. There were no isolated spinal failures.

This pilot study shows that the prophylactic radiation dose to the cranial-spinal spinal axis can be decreased to 25 Gy without jeopardizing control rate and survival in patients with medulloblastoma.

105

V.A. Martial-Vega, M.D. *, M.D. Wharam, M.D. *, S. Leibel, M.D. #, A. Clark, M.D. -, R. Zweig, M.D. 0, S.E. Order, M.D., Sc.D. *

*-Radiation Oncoloy Department, Johns Hopkins Hospital, Baltimore, Maryland,#-Radiation Oncology Department, U.C.S.F., San Francisco, CA.,*- Department of Histopathology, Foothills Hospital, Alberta, Canada, Pathology Department, Johns Hopkins Hospital.

TREATMENT OF SUPRATENTORIAL GRADE III AND IV GLIOMAS WITH SHORT COURSE HIGH DOSE RADIATION

Ninety-two patients with malignant supratentorial gliomas diagnosed from 1977 to 1983 received split-dose external beam radiotherapy. The initial course of radiation consisted of 3000 cGy whole brain in ten fractions five days a week. After a two week rest, treatment was continued to a portal restricted to the CT demonstrated abnormality plus a margin for an additional 2100 cGy(tota1 5100 cGy/17fx/36 days). The optic chiasm and hypothalamus were excluded from the high dose region. Three patients with Grade II glioma, three lacking histologic confirmation and two without biopsy were excluded from survival analysis. No patients were lost to follow-up. Surviving patients were followed 68 months(median): reviewed. All remaining-patient;

Actuarial Survival

range 45-95 months. Craniotomy specimens were- histologically were followed until death.

Grade III and IV Grade III

=<40 years old

>40 years old Grade IV

=<40 years old >40 years old

Median 1 year Zyears 5 years Survivors

9.6 mos. 44.8% 18% 10%(10/84) 17.5 mos. 66.5% 38.5% 33%(5/21) 35 mos. 75% 61% 45.4%(4/11) 45,66,70,

95ilOimoi. 8 mos. 40% 12% lO%(l/lO) 62 mos. 7.6 mos. 37.7% 11.5% 2X(1/63) 30 mos. 68% 58% 12%(1/11) 51,84 mos. 6 mos. 30.1% 2% 0%(0/52)

No patient developed optic nerve or chiasm injury. One patient, a 70 months survivor, had biopsy documented radionecrosis and residual hemiparesis. The incidence of necrosis among 62 patients alive 6 months or moretand therefore at risk of necrosis) is l/62(1.6%). The incidence among survivors is l/8. The nominal standard dose for this regimen is 1734 ret. The predictive value of this formula for normal brain tolerance will be compared to the "equivalent dose" and "neuret" formulae.

We conclude: llthat this regimen provides a survival probability equivalent to conventional treatment for Grade III and IV supratentorial gliomas, 2)that it is well tolerated and has an acceptable risk-benefit ratio, and 3)that its advantage for the patient compared to the conventional schedule is reduced time requirement and cost.