Radiation therapy of pineal region tumors: 25 new cases and a review of 208 previously reported cases

In, J Rodmrrm Oncologic Bml P/~Ks, Vol. Printed in the U.S.A. All rights resewed.

0360.3016/94 $6.00 t Ml P~QWTKXI Press Ltd.

28. PP. 229-245

Copyright icy1993

@ Brief Communication RADIATION THERAPY OF PINEAL REGION TUMORS: 25 NEW CASES AND A REVIEW OF 208 PREVIOUSLY REPORTED CASES

BRIANG.

FULLER,

Department

M.D.,

DANIEL

S. KAPP,

PH.D.,

M.D. AND RICHARD

Cox, PH.D.

of Radiation Oncology, Stanford University School of Medicine, Stanford, CA 94305

Purpose: Malignant pineal region tumors are rare neoplasms arising in midline structures of the brain. This report analyzes the influence of histology, tumor location, radiation dose, treatment volume, age and cerebrospinal fluid findings on freedom from relapse, freedom from spinal relapse and survival. Methods and Materials: Patient and treatment parameters of 25 cases of pineal region tumors managed at Stanford University are presented, and an additional 208 published cases were reviewed. Univariate and multivariate analysis were performed to delineate parameters predictive of freedom from relapse, freedom from spinal relapse, and survival for all 233 patients. Results: The 5- and lo-year freedom from relapse for Stanford patients was 63% and 46% respectively. The 5and lo-year survival for Stanford patients was 67% and 61% respectively. The S- and lo-year freedom from relapse for the total 233 cases was 66% and 612, respectively. The 5- and lo-year survival for all patients was 74% and 68%, respectively. For the entire group, biopsy confirmed germinoma and non-biopsied tumors had superior freedom from relapse compared to non-germinoma germ cell tumors (p = 0.03, p = 0.005, respectively). Non-biopsied patients had improved survival compared to non-germinoma germ cell tumors (p = 0.004). Pineal parenchymal tumors had worse freedom from relapse compared to non-biopsied patients (p = 0.04). For patients with suprasellar tumors, germinomas were associated with improved freedom from relapse compared to non-germinoma germ cell tumors (p = 0.02). Simultaneous pineal and suprasellar tumors had superior survival compared to solitary tumors of pineal (p = 0.04), suprasellar (p = 0.03), or third ventricle location (p = 0.03). Twenty-two patients (9.4%) developed isolated spinal relapse. Five- and lo-year spinal relapse rates for all patients were 11% and 13%. Survival after spinal relapse was 19%. Pineal parenchymal tumors had lower freedom from spinal relapse compared to nonbiopsied patients (p = 0.001). For tumors located in the pineal gland, germinomas and pineal parenchymal tumors had lower freedom from spinal relapse than did non-biopsied patients (p = 0.006, p = 0.004, respectively). Pineal germinomas had lower freedom from spinal relapse than germinomas with suprasellar location (p = 0.04). Univariate and multivariate analysis identified tumor histology as the most significant predictor of freedom from relapse, freedom from spinal relapse and survival. Conclusion: Histologic type had the greatest impact on outcome. Treatment recommendations should be based on assessment of histologic type and extent of disease. Pineal region tumors, Suprasellar tumors, Prognostic factors, Intracranial germ cell tumors, Spinal metastases. INTRODUCTION

mation; usefulness of cerebrospinal fluid (CSF) analysis; prognostic significance of different histologic subtypes; risk of CSF dissemination and/or spinal metastases; optimal radiation dose and treatment volume; and the routine use of craniospinal radiation. Twenty-five patients with pineal region tumors treated at our institution over a 40-year period were analyzed for the possible prognostic importance of age, tumor location, histologic confirmation, dose, treatment volume, and CSF dissemination in predicting relapse and survival. We have also surveyed 125 publications to gather case material for which adequate diagnostic, therapeutic and follow-up data were provided for each case. Two-hundred and eight pineal region tumors were abstracted from the literature.

In the United States, pineal region tumors comprise 0.4% to 1.O%of all brain tumors, and 3% to 11% of brain tumors in children (8, 38, 43, 64, 69). In Japan and in Taiwan, pineal region tumors have a higher incidence, comprising 2.2%-8% of all brain tumors, and approximately 12.5% of pediatric brain tumors (with germinoma comprising 7.5% of all brain tumors in children) (8,2 1, 37). Radiation has been the standard therapy for more than 50 years ( 11, 40, 66). Because of their low incidence and the evolution of surgical, pathologic, radiologic and radiotherapeutic techniques, there exists today no consensus on key management issues including: the value of histologic confirReprint requests to: Brian G. Fuller, M.D., Department of Radiation Oncology, Stanford University Medical Center, Stanford, CA 94305.

Accepted for publication

229

29 April 1993.

230

1. J. Radiation Oncology 0 Biology 0 Physics

These plus the 25 cases from our institution are analyzed in this report with the goal of establishing evaluation and treatment guidelines. METHODS

AND MATERIALS

Stanford University cases Twenty-five patients with pineal region tumors, including pineal gland, suprasellar, and third ventricle tumors were treated at Stanford University between 1948 and 1988 (Table 1). There were 2 1 males and four females. The median age was 20 years (range 8 to 46 years). Primary tumors were located in the pineal gland in 13 cases; the third ventricle in 6 cases; suprasellar region in 4 cases; and in both the pineal and suprasellar areas in 2 cases. Symptoms. All patients were symptomatic at presentation. Documented symptoms included: visual disturbance, 15 cases; headache, 14 cases; nausea and or emesis, 11 cases; papilledema, 9 cases; endocrine dysfunction in 5 cases (diabetes insipidus, four cases; hypothalamic-pituitary dysfunction, 1 case); anorexia, 4 cases; diplopia, 4 cases; hemibody numbness, 3 cases; memory disturbance, wasting, coma, 2 cases each; depression, seizure, vertigo, and Parinauds syndrome, 1 case each. No patient presented with precocious puberty. Radiographic diagnosis. Skull radiographs, ventriculography, pneumoencephalography and arteriography were used to diagnose tumor in 13 cases and computed tomography (CT) or magnetic resonance imaging (MR) in 12 cases. Surgical procedures and histology. Ventriculoperitoneal shunts were performed prior to radiation in 15 cases. Subtotal excision was performed in 9 cases: prior to radiation in 8 patients; and at the time of recurrence in 1 patient. Three patients had biopsy only. Pretreatment CSF analysis was performed in 6 cases. Tissue diagnosis was based on histological or cytological material in 14 cases: 12 prior to radiotherapy; 1 at the time of recurrence (No. 7); and 1 at autopsy (No. 19). Seven patients had confirmed germinoma: 3 from subtotal excision prior to treatment (NO. 11, 20, 23); 1 from biopsy prior to treatment (No. 18); 1 patient had subtotal excision of a spinal recurrence (NO. 7); and 2 patients were diagnosed with germinomas based upon cytologic analysis of CSF (No. 1 and 15). Pineoblastoma was diagnosed following subtotal excision in 3 cases (No. 3, 6, 12). Malignant teratoma was diagnosed in 3 cases: following biopsy (No. 17) and subtotal excision (No. 5) in 2 patients; and at autopsy in 1 patient who received radiation without biopsy (No. 19). In 1 patient subtotal excision revealed mixed germ cell tumor composed of germinoma and elements of mature teratoma (No. 25). An additional 11 non-biopsied patients were treated without histological confirmation. Treatment andfollow-up. All patients were treated with 4 MV or 6 MV photon irradiation. A variety of field configurations were used. Sixteen patients received partial brain irradiation using opposed lateral, three field, four field, or arc rotation techniques. Two patients received

Volume 28, Number 1, 1994

whole brain irradiation, and seven patients received craniospinal radiation. The total tumor doses ranged from 16.4 to 63.2 Gy; the median total tumor dose was 55 Gy. Whole brain doses ranged from 20 to 40 Gy (median, 38.5 Gy). Spinal doses ranged from 30 to 41.6 Gy, (median, 36.2 Gy). Fraction size ranged from 50 to 250 cGy (median, 180 cGy). Follow-up ranged from 0.6 to 41.8 years (median, 4.3 years). Patient, treatment and followup data are summarized in Table 1. Four of the six patients with pretreatment histopathologically confirmed germinomas were treated to craniospinal fields, one received whole brain with a boost with the tumor, and one received partial brain irradiation. Two of the three patients with pineoblastomas received craniospinal radiation, while the third was treated with partial brain fields. Both patients with malignant teratomas received partial brain treatment, while one patient with a mixed germ cell tumor was treated with craniospinal fields. All but one of the patients treated without a histopathologically confirmed diagnosis received partial brain treatment; one patient received craniospinal treatment. Literature cases One hundred and twenty-five publications on pineal region tumors were reviewed and an additional 208 cases were identified for which the following information was published for each patient: age, sex, biopsy status (whether histology was obtained, and if obtained, histologic description of the tumor), tumor location, radiation dose, treatment volume, follow-up duration, follow-up status, recurrence site, and date of recurrence. Cases in which the above information was lacking were excluded except for two cases which lacked information concerning both histology or whether biopsy was performed, and 36 cases which lacked adequate description of treatment volume. These cases were excluded in analysis of histology and treatment volume, respectively. They were also excluded from univariate and multivariate logistic regression analysis. Patients treated without histological confirmation were included and analyzed as a group. Results of CSF studies were analyzed if available, however, this was not a requirement for inclusion in the study. Cases treated with brachytherapy, radiosurgery, neutron or charged particle beams, or patients who died during treatment were excluded. Histologic diagnoses of pinealoma not otherwise defined, lymphoma, craniopharyngioma or metastasis were excluded. Cases were excluded which received no treatment after diagnosis. Several cases had histologic confirmation at the time of recurrence or at autopsy. These cases were included in the non-biopsied group. Patient material was abstracted from a total of 11 published series (5, 17, 19,20, 32, 35, 55, 59,67, 68, 73). Statistical analysis. Actuarial survival, relapse free survival, and freedom from spinal relapse were calculated from date of referral to time of death (or last follow-up), recurrence or spinal relapse, respectively, using the Kaplan-Meier method (45). The significance of differences

231

Pineal region tumors 0 B.G. FULLER etal.

between patient groups was tested using the statistical method of Gehan (34). Pretreatment and treatment factors in 195 patients were tested for prognostic significance for survival, freedom from relapse and freedom from spinal relapse in a univariate and multivariate regression analysis using the method of Cox (22). RESULTS Stanford University cases The 5- and lo-year actuarial freedom from relapse for Stanford University patients was 63% and 45%, respectively. Actuarial survival at five and ten years was 67%

and 6 l%, respectively (Fig. 1). Freedom from relapse for histologically proven germinoma was 7 1% at five years and 24% at ten years. However only one patient was available for IO-year follow-up. One patient (No. 1 I) received partial brain irradiation and failed in the frontal lobe at 6.3 years. Two patients who received craniospinal radiation relapsed; one (No. 15) failed in the CSF at 3 months; the other (No. 23) relapsed in the lungs and abdomen at 5 months. Patient No. 7 received partial brain irradiation initially without biopsy and failed in the cervital spine with a biopsy proven germinoma at 6.6 years. In contrast, the freedom from relapse for non-biopsied patients was 8 1% at five years and 69% at ten years. Of

Table 1. Patient and treatment data Case

Age/Sex

Location

Histology

CSF

Treated volume

Dose GY 50 T 40 B 36 S 55 T 54 T 30.6 BS 55 T 55 T 55 T 39 B/S 54.7 T

c-t

cs

P P

GERM Dx via CSF cytology NBX PINB

N/A

PB cs

40/F 8/F 10/F

S’S P

NBX MT PINB

N/A N/A

N/A

PB PB cs

7

21/M

P

NBX

cs

PB

8 9 10 11 12 13 14 15

8/M 20/M 20/M 13/M 22/M 14/M 46/M 14/M

P/TV P TV SS P/TV P P TV

N/A N/A N/A N/A N/A N/A N/A

c+

PB PB PB PB PB PB PB cs

16.9 T 55 T 55 T 56.2 T 55.8 T 50.2 T 60 T 55 T 40 B/S

16

27/M

P

NBX NBX NBX GERM PINB NBX NBX GERM Dx via CSF cytology NBX

C-k

cs

17 18

14/F 13/M

ss ss

MT GERM

AFP+ N/A

PB WB

19 20

20/M 15/M

P P f ss

NBX GERM

NEG N/A

PB cs

21

32/M

P

NBX

N/A

PB

56 T 40 B/S 55 T 44.5 T 20 B 63.2 T 50 T 38 B 30 s 48.5 T

22 23

12/M 35/M

TV P

NBX GERM

N/A N/A

PB cs

24 25

25/M 20/M

TV P f ss

NBX MIXED GERM CELL

N/A N/A

PB cs

1

17/M

P

2 3

46/M 25/M

4 5 6

N/A

54.8 T 54.8 T 44 B 37 s 55.4 T 52 T 40 B 41 s

Time to relapse

Site of relapse

Status

Follow-up

-

A/NED

4.3 Yrs

-

D/ICD A/WD

14.6 Yrs 1.2 Yrs

D/TP D/TR A/NED

0.6 Yrs 2.4 Yrs 4.2 Yrs

A/NED

18.2 Yrs

A/NED A/NED A/NED D/TR A/WD A/NED D/TR D/TR

40.9 Yrs 10 Yrs 14.2 Yrs 6.4 Yrs 1.6 Yrs 8.5 Yrs 1.1 Yrs 1.9 Yrs

-

A/NED

0.9 Yrs

-

10 MO PDS 26 MO 6.6 Yr 6.3 Yr 1 Yr 11 MO 3 MO

LS/Sacr P ss C Spine Germ FR lobe TV B. Stem/Thai CSF

PDS -

ss

-

D/TP A/NED

0.75 Yrs 4.25 Yrs

PDS -

P MT at autopsy -

D/TP A/NED

1.3 Yrs 9.6 Yrs

6.75 Yr 5 MO

P

A/WD

6.75 Yrs

Chest/Abd

A/NED A/NED

5.6 Yrs 7 Yrs

A/NED A/NED

9.3 Yrs 2 Yrs

-

-

A = Alive; ABD = Abdomen; AFP = Alpha-fetoprotein; B = Brain; C = Cervical; C+ = Cytologically positive; CS = Craniospinal; D = Dead; Dx = Diagnosed; DS = Disease; F = Female; FR = Frontal; GERM = Germinoma; ICD = Intercurrent disease; LS = Lumbar spine; M = Male; MT = Malignant teratoma; MO = Months; N/A = Not available; NBX = Non-biopsied; NEG = Negative; NED = No evidence of disease; P = Pineal; P + SS = Pineal and suprasellar; PINB = Pineoblastoma; PB = Partial brain; PDS = Progressive disease; S = Spine; SS = Suprasellar; SACR = Sacrum; T = Tumor; THAL = ThaIamus; TV = Third ventricle; TP = Tumor progression; TR = Tumor recurrence; WB = Whole brain; WD = With disease; YR = Year.

232

I. J. Radiation Oncology 0 Biology 0 Physics

a. STANFORD

Volume 28, Number I. 1994

PA

100

h

60 40

c. FFR

,

xhec!,(20;

4. 20

d. SURVIVAL

Stanford

I

(25)

Stanford

(25)

-

p=

p=

0.10

I 10

I 20

30

40

0.56

(yea:)

I 20

30

40

TIME Fig. 1. Freedom from relapse (FFR) and survival for Stanford University patients with pineal region tumors and

for 208 published cases. (a) Freedom from relapse and survival for Stanford University patients; (b) Freedom from relapse and survival for all patients reviewed. (c) comparison of freedom from relapse for patients treated at Stanford University and 208 published cases; (d) comparison of survival for patients treated at Stanford University and 208 published cases.

the 11 patients treated without biopsy for whom histopathologic diagnosis remains unknown, seven remain without evidence of recurrence; one (No. 2) died of intercurrent disease at 14.6 years; one (No. 4) had progressive local disease; one (No. 14) failed in the brain stem and thalamus at 11 months; and one (No. 21) failed in the pineal region at 6.75 years. The 2-year freedom from relapse in non-biopsied patients (8 1%) was superior to the freedom from relapse for patients with biopsy proven nongerminoma germ cell histology (43%), p = 0.045. Similarly, when the 5-year freedom from relapse of non-biopsied patients (8 1%) was compared to freedom from relapse for all biopsy-proven patients combined (48%), there was a difference of borderline significance (p = 0.055) favoring the non-biopsied group. There was no significant difference in survival for non-biopsied versus biopsy proven cases (p = 0.4), or non-biopsied versus histologically proven germinomas (p = 0.5). Non-germinoma histologies. There were four cases of non-germinoma germ cell tumors, three were malignant teratomas. Two patients presented with malignant teratomas located in the suprasellar region. Both patients (No. 5 and 17) relapsed locally and died of disease (at 9 and 24 months). One patient who presented with a mass in the pineal region was treated without biopsy (No. 19), and died of locally progressive disease. Autopsy revealed

malignant teratoma. One patient (No. 25) presented with simultaneous pineal and suprasellar masses. Subtotal resection of the pineal mass revealed germinoma with elements of mature teratoma. He is without evidence of recurrence 2 years following craniospinal radiation. He has persistent anorexia, fatigue, short-term memory deficit and hypersomnolence, and he continues to require pituitary hormone replacement. Three patients had histologically proven pineal parenchymal tumors: one patient (No. 6) is without evidence of disease following combined chemotherapy and radiation. The remaining two patients are alive with disease after failing locally (No. 12) or in the lumbar spine (No. 3). Local and regional control. Seven patients (7125, 28%) failed locally at the site of primary tumor. Three patients had persistent disease following radiotherapy and died of tumor progression (No. 4, 17, 19); three recurred locally one to 6.8 years following radiotherapy; one patient (No. 15) developed positive CSF cytology seven months prior to recurrence in the pineal gland, third ventricle, fourth ventricle, and pituitary fossa. The overall crude local control rate was 18/25 (72%). If the one patient who received craniospinal radiation (No. 13) and failed in the lumbar spine 10 months following treatment is considered as an in-field recurrence, the in-field recurrence rate becomes 8125 (32%).

233

Pineal region tumors 0 B. G. FULLER elai.

Four patients failed outside of the treatment volume: three patients developed CNS recurrence in the cervical spine (one patient), frontal lobe (one patient), thalamus and brain stem (one patient); and one patient (No. 23) with germinoma developed shunt metastases five months following craniospinal radiation. Seven patients treated with craniospinal radiation were compared to 16 patients treated with partial brain irradiation. There was no significant difference in freedom from relapse (p = 0.6) or survival (p = 0.8). CSF dissemination. Cerebral spinal fluid results were available for 6/25 (24%) of patients. Cerebral spinal fluid was positive in five patients: cytologically positive in four cases (No. 7 and 16 were nonbiopsied; No. 1 and No. 15 were diagnosed with germinomas based on CSF cytology at initial presentation); and alphafetoprotein (AFP) was elevated in one case with negative cytology (No. 17). The remaining patient (No. 19) had negative cytology and negative AFP and HCG markers. Three of five patients with positive CSF (No. 1, 15, 16) were treated with craniospinal radiation, and two were treated with partial brain fields. Both patients treated with partial brain fields relapsed: one (No. 7) failed in the spine 6.6 years following treatment; and one (No. 17) who presented with biopsy proven malignant teratoma and elevated AFP in CSF, died of progressive local disease. One (No. 15) of three patients who received craniospinal radiation after presenting with positive CSF cytology relapsed. This patient redeveloped malignant cells in the CSF at 3 months following treatment. Seven months later he developed radiographic evidence of extensive local failure and eventually died of recurrent disease. In total, 3/5 (60%) of patients with positive CSF findings developed local or distant failure, compared to 9/20 (45%) of patients for whom CSF analysis was negative, not performed or not documented. Two of 25 patients (8%) developed epidural metastases in the spine. Case No. 3 with pineoblastoma relapsed in the lumbosacral spine ten months following craniospinal radiation. The second spinal failure (No. 7) occurred in a patient who presented with malignant cells in the CSF and received partial brain radiation as initial treatment. He developed an epidural mass at Cb, 6.6 years following treatment. Laminectomy and subtotal excision revealed germinoma. He then was retreated with craniospinal radiation and is now without evidence of disease more than 10 years following retreatment. Table 2. Primary

Pineal gland Third ventricle Suprasellar/Hypothalamus Pineal and suprasellar Thalamus Total

location

Two patients had documented CSF dissemination at recurrence but did not develop spinal metastases: No. 15 relapsed in the CSF prior to extensive local recurrence (as stated above); and one patient (No. 23) presented with abdominal and pulmonary masses at recurrence as a result of ventriculoperitoneal shunt metastases. Both patients had germinomas. The patient with shunt metastases was successfully retreated with combination chemotherapy.

Literature cases The primary location and histologies of the total 233 combined Stanford University and literature cases are listed in Table 2. Actuarial freedom from relapse and survival for the entire group is shown in Figure 1. Freedom from relapse at 5- and 10 years was 66% and 62%, respectively. Survival at 5- and 10 years was 74% and 68%, respectively. Comparison of the patients treated at Stanford University to the patients abstracted from the literature reveals no significant difference in freedom from relapse or survival (p = 0.1 and p = 0.56, respectively). Histology. Histologies were grouped as follows: nonbiopsied; germinomas; pineal parenchymal tumors (including pineocytoma and pineoblastoma); and non-germinoma germ cell tumors (including malignant teratoma, mixed germ cell tumor, embryonal carcinoma, endoderma1 sinus carcinoma, mixed teratoma). Freedom from relapse and survival by histologic type are shown in Figure 2. There was a significant difference in freedom from relapse favoring germinomas over non-germinoma germ cell tumors. At five and ten years the freedom from relapse for germinomas were 66% and 54%, respectively, while that of non-germinoma germ cell tumors were 32% and 32%, (p = 0.03). Freedom from relapse at five and ten years for the non-biopsied group was 73% and 70%, respectively. This was significantly better than the freedom from relapse for the non-germinoma germ cell tumor group (p = 0.005). For the pineal parenchymal tumor group, freedom from relapse at 5 years was 50% (no patient in this group had greater than 5 years follow-up). This was significantly worse than that for the non-biopsied group (p = 0.04). There was no significant difference in freedom from relapse when comparing germinomas to pineal parenchymal tumors (a = 0.1 l), germinomas to non-biopsied patients (p = 0.3) or pineal parenchymal tumors to non-germinoma germ cell tumors (p = 0.9). For tumors located in the pineal gland there was a significant difference in freedom from relapse only when

and histology

of 233 pineal region tumors

Germ

PPT

NGE

NBX

24 1 38 3 3 69

11 1 12

13 4 17

95 10 11 16 1 133

GERM = Germinoma; PPT = Pineal parenchymal tumors; NGE = Non-germinoma * Includes one oligodendroglioma and one low grade astrocytoma.

germ cell tumors:

Glioma*

Total

2 2 NBX = Non-biopsied.

145 12 53 19 4 233

234

I.

J. Radiation Oncology 0 Biology 0 Physics

Volume 28, Number 1, 1994

SURVIVAL

FREEDOM FROM RELAPSE I II

1

b. ALL PATIENTS 1

a. ALL PATIENTS

L

I

1

Gehan P-dues nbxvs. nge 0.004 ge vs. nge O&67

-

Gehan P-values nbx vs. nge 0.009

-

I

c. PINEAL

100

60

4o

II

nge (13)

I

we

(13)

Gehan P-values u nbx vs. nge O&339

e. SUPRASELLAR 80 60 -

nbx (11) 40-k 20

ge (37)

L we

(4)

Gehan P-values nbx vs. ge 0.44 nbx vs. nge 0.97

Gehan P-values ge vs.nge 0.024

TIME

t

(years)

Fig. 2. Tumor histology and outcome. (a) and (b) Freedom from relapse and survival for all patients based on histologic group; (c) and (d) Freedom from relapse and survival for patients with tumors located in the pineal gland. (e) and (f) Freedom from relapse and survival for patients with tumors located in the suprasellar region. Abbreviations: germinoma, ge; non-biopsied, nbx; malignant non-germinoma germ cell tumors, nge; pineal parenchymal tumors, ppt (two patients with glial tumors were excluded).

comparing patients with non-germinoma

with non-biopsied

tumors

to those

germ cell tumors (p = 0.04, Fig. 2~). For tumors located in the pineal gland, the 5- and lo-year freedom from relapse for non-biopsied patients was 73% and 69%, while that for patients with non-germinoma germ cell tumors was 35% and 35%, respectively. There were no significant differences in freedom from relapse for the other histologic groups with pineal gland location. For the suprasellar location, freedom from relapse for germinomas and non-biopsied cases were com-

pared (Fig. 2e). There was no significant difference (p = 0.2). However, for tumors with suprasellar location there was a significant difference in freedom from relapse favoring germinomas over those with non-germinoma germ cell tumors (a = 0.02). Survival at 5- and 10 years for the non-biopsied group was 8 1% and 78%, respectively (Fig. 2b). This was significantly superior to the 5- and 1O-year survival for the nongerminoma germ cell tumor group (36% and 36%, respectively) (p = 0.004). There was a trend for improved

Pineal region tumors 0 B. G. FULLER et al.

235

b. NON-BIOPSIED

a. ALL PATIENTS 80 60 40

Gehan P-values p+ss vs. tv 0.080 p+ss vs. ss 0.019

20

80 60

Gehan P-value csrt vs. pb 0.059 10

2;

30

Gehan P-value p+ss vs. ss 0.080 10

40

TIME

20

30

40

(years)

Fig. 3. Tumor location and outcome. (a) Comparison of survival based on tumor location; (b) and (d) survival and freedom from relapse for non-biopsied patients based on tumor location. (c) Freedom from relapse for suprasellar tumors based on treatment volume. Abbreviations: craniospinal radiation. csrt; hypothalamic, ht; pineal, p; partial brain, pb: pineal and suprasellar/multiple midline tumors, p+ss: third ventricle, tv; suprasellar, ss; whole brain, wb.

survival for patients with germinomas compared to nongerminoma germ cell tumor patients (p = 0.07). Fiveand 1O-year survival for germinoma patients was 76% and 63%. There were no other significant differences in survival detected for other comparisons of histologic type. For patients with primary tumors located in the pineal gland, a difference in survival was seen favoring the nonbiopsied group over the non-germinoma germ cell tumor group (p = 0.009). There was no significant difference in survival for non-biopsied patients compared to patients with germinomas located in the suprasellar region (p = 0.4) (Fig. 2f). Location ofprimary tumors. Freedom from relapse and survival by primary tumor location is shown in Figure 3. Tumor location was grouped as: pineal gland; suprasellar; pineal and suprasellar region; third ventricle; and other, that is, hypothalamus, thalamus. There were no significant differences in freedom from relapse when analysis was limited to groups defined by primary tumor location alone. For non-biopsied patients, improved freedom from relapse was associated with pineal plus suprasellar location (83%) compared to suprasellar location (52%), however this difference was not significant (p = 0.08) (Fig. 3d). Comparison of survival by location of primary tumor (Fig. 3a) revealed a significant difference favoring pineal plus suprasellar location over pineal gland location (p = 0.04),

suprasellar location (p = 0.03) and over third ventricle primaries (p = 0.03). When survival for non-biopsied patients was analyzed by location of primary tumor, a significant difference emerged favoring pineal plus suprasellar (100%) over suprasellar (63%) location (p = 0.02, Fig. 3b). Likewise, there was a trend of improved survival for non-biopsied patients with primaries located in pineal plus suprasellar locations compared to third ventricle location, although the difference was not significant (p = 0.08). When analysis was limited to groups defined by histologically proven germinomas or non-germinoma germ cell tumors, there were no differences in survival associated with location. Radiation dose. Freedom from relapse and survival were compared for all patients based on radiation dose to the tumor. Patients were grouped according to tumor doses less than or greater than 40, 45, 50 and 55 Gy. No correlation of dose and freedom from relapse or survival was found. When subgrouped according to histologic type or treatment volume, again no dose response relationship was seen for freedom from relapse or survival. Analysis of actuarial freedom from spinal relapse as a function of dose was not performed. Treatment volume. Patients were grouped by radiation fields used: that is, partial brain, whole brain and craniospinal radiation. Thirty-six patients did not have a de-

1. J. Radiation Oncology 0 Biology0 Physics

a.

40 - 45yrs

(13)

20 1

P=

0.026

P=

0.013

t

80 60 -

20 1

t

Fig. 4. Comparison of freedom from relapse (a) and survival (b) for patients less than 5 years of age and those 5 years and older. scription of the field size and were therefore excluded from this portion of the analysis. Of the remaining 197 patients, 109 were treated with partial brain, 27 with whole brain, and 61 with craniospinal radiation fields. There was no

significant difference in freedom from relapse or survival when comparisons were made based on treatment volume alone. Similarly, there was no difference in freedom from relapse or survival associated with treatment volume for the subgroup of non-biopsied patients, for patients with germinomas or for patients with pineal gland location of primary tumor. Patients with suprasellar primary tumors who received craniospinal radiation had a superior 5-year (86% vs. 53%) and lo-year (86% vs. 41%) freedom from relapse compared to similar patients treated with partial brain fields. However this difference was of borderline significance, p = 0.059 (Fig. 3~). There was no significant difference in survival. Age. Patients less than 5 years old had significantly worse freedom from relapse (p = 0.03) and survival (p = 0.01) in comparison to patients older than five years (Fig. 4). Similarly, patients less than 10 years old had worse 5and lo-year survival (66% and 65%, respectively) than patients older than 10 years (75% and 70%, respectively). However, this difference was of borderline significance (p = 0.06), and actuarial survival curves merged after 16 years, There were no other significant associations of age and survival. Distant metastases. Nine patients (9/233, 3.8%) developed metastases outside of the neuraxis. Four patients

Volume 28, Number 1, 1994

had abdominal, two abdominal and chest, and three had pulmonary metastases associated with CNS recurrence. Five of these patients were initially treated without biopsy, one patient had histologically proven germinoma, two patients initially presented with embryonal carcinoma and one patient presented with malignant teratoma. All patients had ventriculoperitoneal shunts. Seven patients were believed to have developed the abdominal or abdominal and pulmonary metastases as a result of shunt metastases. Development of abdominal (peritoneal) metastases was not associated with a reduction in survival (p = 0.59, data not shown). Three patients developed chest metastases presumably through hematogenous spread in association with pineal (one case, pineoblastoma), brain and spine (one case, choriocarcinoma) and spinal (one case, malignant teratoma) recurrences. Spinal relapse. Patients were considered to have spinal metastases if so reported in the literature. It was assumed that these cases represented epidural or intramedullary metastases. Three patients relapsed in the brain and spine. Twenty-two patients developed isolated spinal recurrence. One patient with isolated spinal recurrence had received spinal irradiation only as initial treatment and is excluded from further analysis of isolated spinal relapse. For all patients reviewed, the crude isolated spinal relapse rate was 9.4%. The actuarial 5- and lo-year spinal relapse rates were 11% and 13%, respectively. Of the 2 1 patients analyzed with isolated spinal relapse, two had received whole brain, 15 received partial brain, and four received craniospinal radiation. Nine patients had histologically confirmed germinomas, two had pineoblastomas, two had non-germinoma germ cell tumors, seven patients were non-biopsied and one patient had an oligodendroglioma. Four patients were reported alive after spinal relapse: two with disease and two without evidence of disease after additional therapy. The 5-year actuarial survival following spinal relapse was 19% (Fig. 5). This was significantly worse than the 5-year survival for all patients without spinal recurrence (78%) p = < 0.000 1.

P4

0.0001

no spinal ret (212)

20 -

,

spinal ret (21) 10

20

TIME

30

40

(years)

Fig. 5. Survival following spinal recurrence (ret).

231

Pineal region tumors 0 B. G. FULLER efal. I

I

I

b. PINEAL

a. ALL PATIENTS , we
nbx (96)

we (13)

) PPt (12)

Gehan nbx vs. nbx vs. ppt vs.

P-valuss ge 0.061 ppt 0.001 ge 0.088

HPPt (11) Gehan P-values nbx vs. ge 0.006 nbx vs. ppt 0.004

h gs (24)

I

I

I

I

d. POSITIVE CSF

80 60

P (24) pborwb Gehan P-value: 10

20

TIME

30

(7)

0.038 40

(years)

Gehan P-value: 10

20

TIME

30

0.075 40

(years)

Fig. 6. Analysis of isolated spinal relapse. (a) Freedom from spinal relapse by histologic type; (b) freedom from spinal relapse by histology for primary tumors located in the pineal gland; (c) comparison of spinal relapse for

patients with germinomas as a function of tumor location; (d) spinal relapse in patients with maker or cytologically positive CSF. Abbreviations: craniospinal radiation, csrt; germinoma, ge; non-germinoma germ cell tumor, nge; non-biopsied, nbx; p, pineal; partial brain irradiation, pb; pineal parenchymal tumor, ppt: suprasellar, ss; whole brain radiotherapy,

wb.

Pineal parenchymal tumors had a worse freedom from spinal relapse rate (66%) at 5 years compared to confirmed germinomas (84%), however the difference was not significant (p = 0.09; Fig. 6a). Similarly, patients with germinoma had worse j-year freedom from spinal relapse (84%) than did non-biopsied patients (94%), but the difference was of borderline significance (p = 0.06). Fiveyear freedom from spinal relapse for pineal parenchymal tumors was significantly worse than for the non-biopsied group (p = 0.00 1). There was no significant difference in freedom from spinal relapse when analysis was limited to groups defined by tumor location alone. However, pineal parenchymal tumors and germinomas located in the pineal region had significantly worse freedom from spinal relapse rates in comparison to non-biopsied patients with pineal location (p = 0.004) and (p = 0.006), respectively (Fig. 6b). For pineal location the 5-year freedom from spinal relapse for the non-biopsied group was 93% and

that for germinomas and pineal parenchymal tumor groups with pineal location were 70% and 64%, respectively. For patients with germinoma, pineal location was associated with worse freedom from spinal relapse in comparison to suprasellar location, p = 0.04. The actuarial freedom from spinal relapse at 5- and 10 years for germinomas with suprasellar location was 92% and 92%, respectively; while that of germinomas with pineal location was 70% and 60%, respectively (Fig. 6~). There was no significant difference in freedom from spinal relapse among the different histologic groups arising in the suprasellar region. There was no significant difference in freedom from spinal relapse when comparison was based solely on the use of craniospinal radiation in comparison to whole brain (p = 0.7) or partial brain (p = 1.0). For germinomas, the use of craniospinal radiation had no significant impact on freedom from spinal relapse in comparison to whole brain (p = 0.8) or partial brain (p = 0.4)

238

1. J. Radiation Oncology 0 Biology 0 Physics

radiation. Similarly, in the non-biopsied group there was no difference in freedom from spinal relapse associated with the use of partial brain, whole brain or craniospinal radiation. CSF analysis. Results of CSF analysis were available for 39 patients. Patients with positive CSF (positive cytology and/or elevated markers) were not at increased risk of spinal relapse in comparison with those who had negative CSF (p = 0.6). However there was a significantly less favorable freedom from spinal relapse associated with positive CSF compared to CSF not available, p = 0.003 (data not shown). The actuarial freedom from spinal relapse for patients with positive CSF vs. the larger group of patients for which CSF information was not available was 68% vs. 92%, respectively, at 10 years. Seventeen patients with positive CSF who received craniospinal radiation were compared to those who received whole brain (one patient) or partial brain (six patients) radiation (Fig. 6d). At 5 years the actuarial freedom from spinal relapse rate was 87% for the craniospinal radiation group and 57% for the group not treated with craniospinal radiation. Although suggestive, this difference was not significant (p = 0.08). Perhaps this is due to the small number of patients with CSF results available for comparison. Regression analysis. Cox regression analyses were performed on 195 patients for whom all pertinent data was available for evaluation. Age, sex, histology, lack of histologic diagnosis, location of primary tumor, tumor dose and treatment volume (partial brain, whole brain, craniospinal radiation), were analyzed for associations with freedom from relapse, freedom from spinal relapse and survival. Univariate analysis identified histological type as the most significant predictor of freedom from relapse (p = 0.0 18), freedom from spinal relapse (p = 0.0 12) and survival (p = 0.034). Similarly, a multivariate model also identified histologic type as the most significant predictor of freedom from relapse (p = 0.0 13), freedom from spinal relapse (p = O.Ol), and survival (p = 0.035). The addition of other parameters did not result in a better model. Patients with non-biopsied tumors had higher freedom from relapse and freedom from spinal relapse than any histologic group. Germinomas and non-biopsied tumors were associated with the highest survival rates.

DISCUSSION

Tumors of the pineal region pose an interesting challenge to the clinician due to their rarity, and due to the diverse histopathologic group of neoplasms typically encountered. Differing opinions exist regarding key issues in the diagnosis and management. For the most part, prospective randomized studies are lacking. Most published studies are small, retrospective analyses, which include cases lacking histologic confirmation, and which span several decades. We undertook this study to determine if there were trends in evaluation, treatment and outcome

Volume 28, Number 1. 1994

which would support rational decisions in case management. Stanford University cases

Our 5-year freedom from relapse and survival rates (63% and 67%, respectively) are similar to those reported by other investigators (2, 13). The significantly superior freedom from relapse for non-biopsied patients (82% vs. 50% for biopsy proven cases), is particularly noteworthy. While it has been argued that biopsy may increase the risk of dissemination and therefore have a negative impact on prognosis (42, 80, 85), the alternative explanation, that a significant number of non-biopsied patients with benign lesions have been unnecessarily radiated, may be more plausible. Contrary to the notion that biopsy or resection may increase the risk of metastases, several investigators have found no increased risk of metastases associated with biopsy or resection (3 1, 44, 5 1, 59, 6 1). The significant differences in outcome for non-biopsied versus biopsy confirmed patients suggests that caution should be used when interpreting studies in which results for non-biopsied patients are not reported as a separate group. Furthermore, the inclusion of non-biopsied patients within a retrospective series may result in selecting only malignant tumors within biopsy confirmed subgroups. Estimates of the percentage of benign lesions found upon biopsy or resection of pineal or suprasellar region masses range from 10% to 36% (3 1, 44, 59, 78, 84). These lesions include benign cysts (dermoid, glial, epidermoid, arachnoid, colloid), harmatomas (lipomatous, arteriovenous malformations, cavernous angiomas), aneurysms (vein of Galen), benign teratomas, pineocytomas with neuronal differentiation, cystic astrocytomas, meningiomas, low-grade oligodendrogliomas, and granulomas. The inclusion of such benign conditions in nonbiopsied groups might explain the more favorable outcome associated with non-biopsied patients in some series. In our 25 patients, treatment volume had no impact on freedom from relapse or survival. The three patients who recurred after craniospinal radiation, did so with manifestations of CSF metastases at initial relapse (i.e., recurrence of positive CSF in one case; lumbar spine recurrence in one case; and shunt metastases to the abdomen and lung in one case). The majority of patients failing partial brain irradiation did so locally (6 out of 9, 66%). Three patients who received partial brain fields developed out-of-field recurrence in the cervical spine, frontal lobe and brainstem/thalamus. Interestingly, the ratio of in-field to out-of-field recurrence was 2: 1 following either craniospinal radiation or partial brain irradiation. A propensity for late recurrence in the spine and elsewhere has been reported (2, 5, 58, 71, 80, 82). Three of the 25 (12%) patients in the Stanford University group relapsed more than five years following treatment. Two of three had biopsy proven germinomas at recurrence. It is, therefore, premature to consider patients free of significant risk of recurrence at five years following treatment.

Pineal region tumors 0 B. G. FULLER d a/.

Careful continued follow-up, prompt detection and successful treatment can result in long-term survival in approximately 19% of cases with spinal relapse. In our series case #7 was treated with resection and irradiation after relapsing at C6 with germinoma greater than 6 years following partial brain irradiation. He remains without evidence of disease now greater than ten years since additional therapy was completed. All cases Histology. Analysis of the pooled Stanford University and literature patients underscores the important influence of tumor histology on outcome. As reported by others (31, 43, 61, 68, 83, 85) we found that histologic type influences survival, freedom from relapse and spinal recurrence (Fig. 2 and Fig. 6). When possible, routine histologic confirmation prior to therapy would be essential for planning appropriate management. Since the presence of various histopathological components (i.e., malignant non-germinoma germ cell tumor elements; neuronal differentiation) can profoundly impact prognosis (69) complete or subtotal resection is preferable to needle biopsy in that a larger amount of representative tissue is made available for histologic characterization. Proponents of treatment without biopsy who advocate a “radiation test” to determine the histologic nature of lesions (15, 35, 53, 67) might observe that the results in non-biopsied patients are excellent and therefore there is little to be gained from histologic confirmation. However, it is unlikely that all non-biopsied tumors are malignant or should receive radiation. The “radiation test”, which was developed in an era when biopsy was associated with a 20-50% mortality (1 l), has several problems. First, all patients are committed to a course of radiation. This is unjustified given that lo-36% of pineal region tumors are benign (31, 44, 59, 78, 84) and would not benefit from radiation therapy. Such pineal lesions include dermoid, glial, epidermoid, arachnoid, and colloid cysts, harmatomas, benign teratomas, meningiomas, low grade cystic astrocytomas or oligodendrogliomas, and granulomas. In pediatric patients, radiation for such benign lesions may not only be unnecessary, but may be induce long-term morbidity even at relatively low doses (i.e., 2000 cGy “test” doses). Cognitive impairment (30, 64); impaired learning and cerebral calcifications (1), impaired cerebral blood flow particularly in patients less than age 15 when treated (70) memory deficits (12) and the late development of second tumors (33,47) have been linked to brain irradiation for pineal region tumors in children. Use of craniospinal radiation may impair bone growth as well (29). Therefore, initiating a course of radiation without attempting to obtain a tissue diagnosis should be avoided. Patients who present in communities which lack necessary neurosurgical resources for safe biopsy should be referred to larger centers where such expertise is available. Second, the rate of radiographic response does not predict histology or curability (20) and therefore cannot replace histologic

239

confirmation. Tumors other than germinomas such as embryonal carcinoma (63), yolk sac tumors (6, 57), pineocytomas (23, 28) and pineoblastomas (14, 52) may show rapid responses to radiation. Third, as discussed by Edwards et al. (3 1), the “radiation test” strategy obviously fails in the situation of a partial response. The resulting situation is one in which the time to diagnosis has been prolonged and the probability of treatment related morbidity in patients who harbor benign disease may have been increased (59). Finally, decisions regarding adjuvant therapies (i.e., spinal irradiation, chemotherapy), cannot be made rationally without knowledge of tumor histology and natural history. Tumor location. Analysis of tumor location suggests that patients with simultaneous pineal plus suprasellar tumors had superior survival when compared to those with pineal only or suprasellar only locations (p = 0.04, p = 0.03, respectively; Fig. 3). Also, for the subset of nonbiopsied patients, there was a trend towards improved freedom from relapse for pineal plus suprasellar location compared to pineal or suprasellar location alone. Differences in outcome associated with tumor location have been reported by others (54,8 1). Sano et al. (74) reported that suprasellar germinomas had an improved IO-year survival (92%) when compared to pineal germinomas (61%). Onoyama et al. (61) found better survival associated with suprasellar location when compared to pineal location, however this difference disappeared when glioma and teratoma cases were excluded. Excellent outcome for “multiple midline tumors” has been previously described (68, 79). There are several possible explanations for the observed superior survival for pineal plus suprasellar cases, the most obvious being that patients with simultaneous pineal plus suprasellar masses harbor germinomas, whereas patients with solitary pineal or suprasellar masses, may harbor a wider range of histologies, some less favorable (i.e., malignant teratoma, pineoblastoma, glioma, chordoma, metastases). Although the radiographic findings of simultaneous pineal/third ventricle and suprasellar/parasellar masses may be considered highly suspicious for germinomas, other malignant histologies may produce multiple midline masses (7, 56). We feel that the rarity of these cases justifies histological analysis whenever possible to confirm the diagnosis. A second possible explanation is that most patients with simultaneous pineal plus suprasellar masses received more extensive radiation fields (68). While this may be true, our analysis of treatment volume does not suggest that more extensive treatment alone accounts for improved outcome (see discussion of treatment volume). Rich et al. (68) reported that multiple midline tumors (none which were biopsied) had excellent freedom from relapse and survival when treated with craniospinal radiation. It is plausible that the favorable outcome associated with simultaneous pineal plus suprasellar tumors was most likely due to the predominance of germinomas in those cases. Multiple benign masses simultaneously

240

1. J. Radiation Oncology 0 Biology 0 Physics

occupying the pineal plus suprasellar or parasellar regions is unlikely, and the favorable clinical outcome observed would have been unlikely for other malignant histologies. The observations of Rich et al. (68) were strengthened in our analysis which included cases from that series and an additional 11 cases for a total of 19 pineal plus suprasellar or multiple midline tumor cases. Dose. The histologic similarity of pineal germinoma and testicular seminoma is well documented (69). While complete responses to as low as 1600 cGy have been reported for germinomas (lo), it is generally held that intracranial germinomas require higher radiation doses for tumor control than do seminomas (2,24,46,6 1,73,80). Kersh et al. (46) found that the 5-year survival for patients with intracranial germ cell tumors receiving greater than 50 Gy was 78% versus 58% for those receiving less than 50 Gy. Onoyama et al. reported improved survival for doses of 1500 to 1699 rets (5-year survival, 90%) vs. 1499 rets and below (5-year survival, 47%). However, the later study included a group of patients in which therapy was interrupted because of poor medical condition (6 1). Abay et al. (2) also reported a dose response relationship. Local control was achieved in 87.5% of patients receiving greater than or equal to 50 Gy, while only 33% of patients were locally controlled who received less than 30 Gy. Sung et al. reported a 100% recurrence rate for patients receiving 38 to 40 Gy, 47% recurrence rate for patients receiving 38 to 45 Gy, and a 10% recurrence rate for patients receiving 50 to 55 Gy (80). Dattoli and Newall (24) reported local failure in one patient with biopsy proven germinoma treated to 3960 cGy, while nine patients with germinomas who received 5000-5500 cGy remained free of disease. In all of these studies, no detailed statistical analysis of dose and outcome was reported (2, 24, 46, 6 1, 80). In 1979, Salazar et al. (73) published a study similar to our report. In that review, a total of 67 patients were analyzed to determine the impact of dose on survival. There was a statistically significant improvement in median survival associated with doses greater than 1500 ret (9.25 years) versus doses less than 1500 ret (4.5 years). However, the significant difference in survival disappeared after five years. Our own analysis found no significant difference in freedom from relapse or survival associated with dose. Perhaps differences in tumor size, extent of surgical debulking, how radiation dose was delivered or reported, or other factors (i.e., inclusion of benign tumors among non-biopsied patients; inconsistencies in accuracy of radiographic localization and field size used with similar doses reported by different investigators during different periods) obscured any relationship between dose and outcome. Treatment volume. In addition to the indications for biopsy of pineal region tumors, the choice of appropriate treatment volume has been a topic of debate. The most experience has been gained treating non-biopsied patients with partial brain fields (11, 17, 19,20, 73, 74,80,83, 85) during a period which spans the pre and post CT era.

Volume 28, Number 1, 1994

There have been several retrospective comparisons of outcome based on treatment volume. Bradfield et al. (17) reported improved survival associated with larger treatment volumes, but no statistical analysis was provided. Rao et al. (67) found no statistically significant difference in patients based on whole brain or partial brain treatment. Shibamoto et al. (77) reported no difference in freedom from relapse or survival associated with treatment volume for 66 non-randomized patients, 43% of which were nonbiopsied. Salazar et al. (73) compared disease-free survival as a function of irradiated volume in a total of 67 patients from a combined literature review. They found improved survival for patients with whole brain radiation compared to those receiving only partial brain treatment, however the statistical significance of this observation was not reported and the improvement in disease-free survival was not seen after 4 years. Dattoli and Newall (24) reported excellent survival in patients with germinomas treated with partial brain fields and recommended that treatment volumes be limited to the tumor plus 2 cm margins. Advocates of routine craniospinal radiation (25, 32, 35, 53, 68) cite the excellent freedom from relapse, survival, and freedom from spinal relapse associated with craniospinal radiation, as well as the low likelihood of subsequent cure following spinal recurrence, as support for routine craniospinal radiation. These observations were probably based on series with small patient numbers (2, 15, 17, 25, 32, 46, 50, 68, 80) and series which often include biopsied patients as well as patients selected on the basis of response to the “radiation test” (2, 15, 17,25, 35, 46, 68). The likely inclusion of patients with benign disease within many series may have resulted in higher survival and freedom from spinal relapse rates associated with craniospinal radiation. In addition, the small patient numbers in most series do not allow matched comparisons of craniospinal radiation with more limited treatment ( 17, 80). However, Kersh et al. (46) reported a statistically significant difference in survival associated with craniospinal radiation (87.5%) compared to whole brain or partial brain treatment (56%). Similarly, Rich et al. (68) found a significant improvement in outcome associated with craniospinal radiation in comparison to partial brain treatment. In both of these retrospective series, not only was there a large percentage of patients who were nonbiopsied, but both series had large differences in the number of patients who received craniospinal radiation compared to those who received partial brain treatment. Furthermore, comparable freedom from relapse, freedom from spinal relapse, and survival rates have been reported by others following the use of whole brain or partial brain treatment volumes (24, 51, 54). Since spinal failure has been reported for all malignant histologies encountered in the pineal region, inadequate diagnosis (i.e., “radiation test”) and/or treatment (i.e., low doses of radiation; delays in definitive treatment; poor tumor localization and inadequate treatment volume; inadequate surgical debulking of glioma or malignant teratoma) of the primary site

Pineal region tumors 0 B. G.

make analysis of the influence of spinal “prophylaxis” difficult. This is particularly relevant in non-biopsied patients. Those opposed to routine craniospinal radiation cite the low frequency of isolated spinal failure among survivors (average in our literature review was 9.4%) as evidence supporting the use of more limited radiation fields, and believe that the risk of growth abnormalities in children and gonadal dysfunction which can result from spinal radiation is not warranted (5 1,73). While there is no consensus regarding the use of whole brain radiation or the appropriate volume and structures to be included in partial brain fields, in our analysis, craniospinal radiation was not associated with a significant difference in outcome compared to whole brain or partial brain treatment. An improved freedom from relapse for patients with suprasellar primary tumors who received craniospinal radiation was of borderline significance (p = 0.059). We conclude that there is no clear support for prophylactic spinal radiation and that its routine use in non-biopsied patients without evidence of spinal or CSF involvement, or without simultaneous pineal plus suprasellar lesions is not indicated. Craniospinal radiation should be reserved for cases with: (a) spinal metastases at presentation; (b) marker or cytologically positive CSF, (c) simultaneous pineal plus suprasellar involvement; (d) malignant non-germinoma germ cell tumors; (e) documented pineoblastoma; (f) pineocytoma without neuronal, or neuronal and astrocytic differentiation (14, 36, 75). Spinal relapse. We found that histology had the greatest impact on spinal relapse. Pineal parenchymal tumors had the highest 5-year spinal relapse (34%) compared to nonbiopsied cases (5%, p = 0.001) or biopsy confirmed germinomas (20%, p = 0.057). A review of the literature revealed a combined spinal relapse for pineal parenchymal tumors of approximately 4 1% (20/49) (14,20,23,28, 59, 64, 65, 85). In patients with pineoblastoma or pineocytoma lacking benign features, adjuvant spinal radiation should be routinely used, although there is no data clearly showing that craniospinal radiation reduces the risk of spinal relapse. The results of our review for germinomas are similar to two recent reports which addressed the risk of spinal relapse and the indications for craniospinal treatment (16,5 1). We found that non-biopsied patients have the lowest actuarial spinal relapse (3% with craniospinal radiation, 6% with whole brain, and 8% in the partial brain group). The 11% failure rate reported by Lindstat et al. (5 1) for non-biopsied patients treated with craniospinal radiation is somewhat higher than the 3% noted in our study. This is perhaps due to differences in series selected for review. Composite rates of 3% and 8% were reported by Brada and Rajan for radiosensitive tumors or histologically non-verified tumors respectively, treated with craniospinal radiation (16). For biopsy confirmed germinomas, we found no significant difference in spinal relapse for patients receiving craniospinal radiation (SW), compared to whole brain

FULLER et nl.

241

(17%) p = 0.8, or partial brain radiation (26%) p = 0.4. The 13-23% spinal relapse for confirmed germinomas treated with partial brain irradiation and the 5-8% risk of spinal relapse following craniospinal radiation noted by Brada and Rajan ( 16) and by Lindstat et al. (5 1) are similar to our results. The higher rate of spinal relapse seen in the patients who received partial brain irradiation in our series, is most likely due to inclusion of a higher percentage of patients from the pre-CT era. Improved outcome associated with advances in contemporary diagnostic and staging practices has been reported for medulloblastoma (26) and it is likely that many patients treated with partial brain radiation had poorer staging and tumor localization (17, 20, 73, 80) than the majority of patients receiving craniospinal radiation treated in the post-CT era (25,32,68). It seems reasonable, based on a spinal relapse rate of 26% [and greater than 40% in some series (go)], for authors reviewing data from the pre-CT era to advocate routine craniospinal radiation for all biopsy proven germinomas (53, 68, 80). Indeed, review of data presented by Brada and Rajan reveals that the majority of cases with spinal relapse in confirmed germinomas treated with partial brain radiation were reported prior to 1980 (16). Contemporary series of biopsy proven germinomas treated without spinal radiation report spinal relapse rates which are superior to that for the craniospinal radiation group in the present review. Combining patients from series by Dattoli and Newall (24), Edwards et al. (3 I), Fields et ul. (32) and Matsutani el ul. (54) gives a total of one reported spinal relapse out of 40 cases, for a spinal relapse rate of 2.5%. Except for one (32) these latter reports were not included in our pooled review because of insufficient data on each patient. In our opinion, with the careful use of MR or CT, myelography, and CSF cytology and marker analysis, the contemporary risk of spinal relapse for biopsy proven germinomas without evidence of dissemination should be less than 10%. For patients with biopsy proven germinomas, we found an increased risk of spinal relapse for tumors located in the pineal gland in comparison to suprasellar location (p = 0.04). This was also described by Jenkin et al. (42) who found a correlation of spinal metastasis with pineal location (p = 0.003). In the cases reported to have positive CSF, craniospinal radiation was associated with improved freedom from spinal relapse. Although the number of patients studied is small and the observed difference (87% for craniospinal radiation and 57% without craniospinal radiation) was not significant (p = 0.08) the trend is suggestive enough to support recommending adjuvant spinal radiation in all patients with cytologically and/or marker positive CSF. In addition, spinal MR or CT myelography should be used to identify any sites of macroscopic involvement. Chemotherapy. The relatively poor survival of patients with non-germinoma germ cell tumors and the responsiveness of extraneural metastatic non-germinoma germ

242

1. J. Radiation Oncology 0 Biology 0 Physics

cell tumors to chemotherapy, suggest that chemotherapy may play a role in the primary treatment of patients with non-germinoma germ cell tumors. Single agent or combination regimens have been used for neoadjuvant (4, 9, 18, 41, 48, 62, 63) or adjuvant (48, 63, 76) therapy, as well as for treatment of CNS or extraneural recurrence (6, 27, 49, 60, 72, 86). Several authors have proposed replacement of radiotherapy with chemotherapy following resection of intracranial germ cell tumors (4, 18). However, there is no data suggesting that cure rates for pineal region tumors are higher following chemotherapy alone. The true value of neoadjuvant or adjuvant chemotherapy in prolonging survival of pineal region tumors is unknown (4,6), and the possible sequelae of chemotherapy remain to be clarified (3,4, 59, 62). At this time, short follow-up, small patient numbers, and results of non-randomized studies prevent definitive statements concerning the efficacy of chemotherapy for malignant pineal region tumors. The poor prognosis associated with non-germinoma germ cell tumors, as well as the sequelae of CNS radiation in patients less than three years of age, justifies pilot studies testing adjuvant chemotherapy for non-germinoma germ cell tumors, and neoadjuvant chemotherapy for young patients for whom radiotherapy might be delayed until CNS development is more complete.

CONCLUSION Based on review of our results and the pertinent literature, we feel that the following radiation therapy treatment guidelines can be supported. Initial evaluation In addition to physical, neurologic and ophthalmologic examination, CBC, chest x-ray, one should obtain complete endocrine screening for hypothalamic and pituitary dysfunction, and serum chemistries including serum tumor markers (HCG, AFP). Lumbar puncture for cytology, chemistry and tumor maker analysis should be performed routinely in patients with malignant tumors, and repeated more than once if studies are negative. In addition to contrast enhanced CT or MR scans of the brain, CT myelography and/or MR scan of the spine should be routinely performed on patients with malignant histologies. Stereotaxic biopsy or resection, or craniotomy and resection for tissue diagnosis should be attempted in all cases unless there exist specific contraindications. Radiotherapy Based on tissue diagnosis, patients with residual disease following resection of mature teratoma or pineocytoma with neuronal or neuronal and astrocytic differentiation should be evaluated for possible limited field radiotherapy to residual disease if CNS development is complete. However, the true value of radiotherapy in this situation is unknown. Children with partially resected benign le-

Volume 28, Number 1, 1994

sions including low grade astrocytomas, oligodendrogliomas, mature teratomas, pineocytomas with neuronal differentiation, meningiomas, well-differentiated ependymomas, and dermoid or epidermoid cysts should have long-term observation without further treatment unless recurrence develops. Patients with isolated pineal or suprasellar involvement by germinoma with minimal extension to contiguous structures and negative CSF and spinal imaging studies, should receive partial brain radiation encompassing at least the tumor with 2 cm margins. Total dose to the tumor should be 50-55 Gy. Extensive pineal or suprasellar germinoma with involvement of contiguous structures, should be treated with 38-40 Gy to the whole brain, and 15 Gy boost to the tumor bed or site(s) of post-surgical residual disease. Craniospinal radiotherapy is felt indicated for simultaneous pineal and suprasellar involvement, or in cases of CSF dissemination (i.e., cytologic or marker positive CSF, and/or spinal metastases). The neuraxis including the sacrum to at least S3 should receive at least 30 Gy, followed by a boost to the intracranial tumor bed and to radiologically evident spinal metastases. Patients with non-germinoma germ cell tumors and negative CSF studies and negative spinal imaging studies should receive radiation to the neuraxis with a boost to the primary site. Preferably, adjuvant systemic chemotherapy should be given in a protocol setting. Cases with radiologic evidence of intramedullary or epidural spinal involvement should also receive boosts to areas of spinal involvement. Adjuvant chemotherapy with intrathecal infusion should be considered. Patients with pineoblastoma and pineocytoma lacking benign differentiation (neuronal or neuronal and astrocytic differentiation) should receive radiation to the neuraxis and a boost to the tumor bed. Patients with positive spinal imaging studies should receive a boost to the area of spinal involvement. Adjuvant chemotherapy should be considered routinely, preferably in a protocol setting. Consideration of intrathecal chemotherapy is indicated in patients with positive CSF or macroscopic spinal metastases. Children less than three years of age with malignant pineal region tumors should be considered for neoadjuvant chemotherapy in a protocol setting (including intrathecal infusions for CSF or spinal involvement). Chemotherapy should be given until maximum response is reached or until tumor progression ensues, with the goal of delaying radiotherapy until the patient is greater than 3 years of age. Follow-up evaluations for patients with malignant histologies should include physical and neurologic examinations, endocrine evaluation, lumbar puncture for CSF analysis and neuroimaging studies. Pediatric patients should have yearly neuropsychiatric and intellectual function testing. Given the propensity for late relapse, as well as the risk of second cancers, all patients should be followed at regular intervals for the remainder of their lives.

Pineal region tumors 0 B. G.

FULLER

et al.

243

REFERENCES 1. Abaoyami, 0.; Chun, M. S.; Kelly, K. S. Cerebral calcifications and learning disabilities following cranial irradiation for medulloblastoma. J. Nat. Med. Assoc. 82:833-836;1990. 2. Abay, E. 0.; Laws, E. R.; Grade, G. L.; Buckman, J. E.; Forbes, G. S.; Gomez, M. R.; Scott, M. Pineal region tumors in children and adolescents (treatment by shunting and radiotherapy). J. Neurosurg. 55889-895; 198 1. 3. Allen, J. C. Controversies in the management of intracranial germ cell tumors. Neurol. Clinics. 9:44 l-452; 199 1. 4. Allen, J. C.; Kim, J. H.; Packer, R. J. Neoadjuvant chemotherapy for newly diagnosed germ-cell tumors of the central nervous system. J. Neurosurg. 67:65-70;1987. 5. Amendola, B. E.; McClatchey, K.; Amendola, M. A. Pineal region tumors: Analysis of treatment results. Int. J. Radiat. Oncol. Biol. Phys. 10:991-997;1984. 6. Anderson, B. A. Intracranial endodermal sinus tumorrole for chemotherapy. Can. J. Neurol. Sci. 14: 166171;1987. 7. Aoyama, I.; Makita, Y.; Nabeshima, S.; Motomochi, M. Intracranial double teratomas. Surg. Neurol. 17:383387; 1982. 8. Araki, C.; Matsumoto, S. Statistical reevaluation of pinealoma and related tumors in Japan. J. Neurosurg. 30:146149;1969. 9. Asano, N.; Oka, H.; Takse, K.; Honda, H.; Ueda, S.; Matsumoto, K.; Hirose, T. Intracranial and intraspinal dissemination from pineal yolk sac tumor treated by PVB therapyCase report. Neurol. Med. Chir. (Tokyo) 30:483-488; 1990. 10. Aydin, F.; Ghatak, N. R.; Radie-Keane, K.; Kinard, J.; Land, S. D. The short term effect of low dose irradiation on intracranial germinoma: a pathologic study. Cancer 69:23222326;1992. 11. Baggenstoss, A. H.; Love, J. G. Pinealomas. Arch. Neurol. Psychiatry 4 1: 1187- 1206; 1939. 12. Bendersky, M.; Lewis, M.; Mandelbaum, D. E.; Stanger, C. Serial follow-up of a child following craniospinal irradiation. Devel. Med. Child. Neurol. 30:816-820;1988. 13. Bloom, H. J. G. Primary intracranial germ cell tumors. Clin. Oncol. 2:23-46;1983. 14. Borit, A.; Blackwood, W.; Mair, W. G. P. Separation of pinealoma from pineoblastoma. Cancer 45: 14081418;1980. 15. Boyd, M. C.; Steinbok, P. Pineal region tumors. Can. J. Neuro. Sci. 12:35-38;1985. 16. Brada, M.; Rajan, B. Spinal seeding in cranial germinoma. Brit. J. Cancer 61:339-340;1990. 17. Bradfield, J. S.; Perez, C. A. Pineal tumors and ectopic pinealomas: Analysis of treatment and failures. Radiology 103: 399-406; 1972. 18. Casteneda, V. L.; Parmeley, R. T.; Gieser, C. F.; Saldivar, V. A.; Mullins, J. K.; Marlin, A. E. Postoperative chemotherapy for primary intracranial germ cell tumor. Med. Ped. Oncol. 18:299-303; 1990. 19. Chang, C. G.; Kageyama, N.; Kobyashi, T.; Yoshida, J.; Negoro, M. Pineal tumors: Clinical diagnosis, with special emphasis on the significance of pineal calcification. Neurosurgery 8:656-668; 198 1. 20. Chapman, P. H.; Linggood, R. M. The management of pineal area tumors: A recent reappraisal. Cancer 46: 12531257;1980. 2 1. Committee of Brain Tumor Registry of Japan. Neurol. Med. Chir. (Tokyo) 32:391-542;1992. 22. Cox, D. R. Analysis of binary data. London: Chapman and Hall, 1970. 23. D’Andrea, A. D.; Packer, R. J.; Rorke, L. J.; Baliniuk, L. T.; Sutton, L. N.; Bruce, D. A.; Schut, L. Pineocytoma

of childhood: a reappraisal of natural history and response to therapy. Cancer 59:1353-1357;1987. 24. Dattoli, M. J.; Newall, J. Radiation therapy for intracranial germinoma: the case for limited volume treatment. Int. J. Radiat. Oncol. Biol. Phys. 19:429-433; 1990. 25. Deamaley, D. P.; A’Hem, R. P.; Whittaker, S.; Bloom, H. J. G. Pineal and CNS germ cell tumors: Royal Marsden Hospital experience 1962-1987. Int. J. Radiat. Oncol. Biol. Phys. 18:773-78 1; 1990. 26. Deutsh, M. The impact of myelography on the treatment re$ults for medulloblastoma. Int. J. Radiat. Oncol. Biol. Phys. 10:999-1003; 1984. 27. De Tribolet, N.; Barrelet, L. Successful chemotherapy of pinealoma (letter). Lancet 2: 1228- 1229; 1979. 28. Disclafani, A.; Hudgins, R. J.; Edwards, M. S. B.; Wara, W.; Wilson, C. B.; Levin, V. A. Pineocytomas. Cancer 63: 302-304; 1989. 29. Donaldson, S. S. Effects of irradiation on skeletal growth and development. In: Green, D. M., D’Angio, G. J., eds. Late effects of treatment for childhood cancer. New York, NY: Wiley-Liss, Inc.; 1992:63-70. 30. Duffner, P. K.; Cohen, M. E.; Thomas, P. R. M.; Lansky, S. B. The long-term effects of cranial irradiation on the central nervous system. Cancer 56: 1841-1846; 1985. 31. Edwards, M. S.; Hudgins, R. J.; Wilson, C. B.; Levin, V. A.; Wara, W. M. Pineal region tumors in children. J. Neurosurg. 68:689-697; 1988. 32. Fields, J. N.; Fulling, K. H.; Thomas, P. R. M.; Marks, J. E. Suprasellar germinoma: Radiation therapy. Radiology 164:247-249; 1987. 33. Garre, M. L.; Soffietti, R.; Carini, S.; Cama, A.; Giordana, M. T.; Gomaglia, P.; Vigliani, M. T.; Gandus, S.; Becker, E.; Andreussi, L. Delayed neurotoxicity and second malignant brain tumor after treatment for medulloblastoma (Abst.). International Symposium on Pediatric Neuro-Oncology, Seattle, WA, June 1-3, 1989;A38. 34. Gehan, E. A. A generalized Wilcoxon test for comparing arbitrarily singly censored samples. Biometrika 52:203233;1965. 35. Griffin, B. R.; Griffin, T. W.; Tong, D. Y. K.; Russell, A. H.; Kurtz, J.; Laramore, G. E.; Groudine, M. Pineal region tumors: results of radiation therapy and indications for elective spinal irradiation. Int. J. Radiat. Oncol. Biol. Phys. 7:605-608;198 1. 36. Herrick, M. K.; Rubinstein, L. J. The cytological differentiating potential of pineal parenchymal neoplasms (true pinealomas). A clinicopathological study of 28 tumours. Brain 102:289-320;1979. 37. Ho, D. M.; Liu, H. C. Primary intracranial germ cell tumor. Pathologic study of 5 1 patients. Cancer 70: 1577- 1584; 1992. 38. Hoffman, H. J. Pineal region tumors. Prog. Exp. Tumor Res. 30:281-288;1987. 39. Horowitz, M. B.; Hall, W. A. Central nervous system germinomas. Arch. Neurol. 48:652-657;199 1. 40. Horrax, G. Treatment of tumors of the pineal body. Arch. Neurol. Psychiatr. 64:227-242; 1950. 4 1. Itoyama, Y.; Kochi, M.; Yamaoto, H.; Kuratsu, J.; Uemura, S.; Ushio, Y. Clinical study of intracranial nongerminomatous germ cell tumors producing alfafetoprotein. Neurosurg. 27:454-460; 1990. 42. Jenkin, D.; Berry, M.; Chart, H.; Greenburg, M.; Hendrick, B.; Hoffman, H.; Humphreys, R.; Sonley, M.; Weitzman, S. Pineal region germinomas in childhood: treatment considerations. Int. J. Radiat. Oncol. Biol. Phys. 18:541545;1990. 43. Jennings, M. T.; Gelman, R.; Hochberg, F. Intracranial germ

244

44. 45.

46.

47.

48.

49.

50.

5 1.

52. 53.

54.

55. 56.

57.

58.

59.

60.

6 1.

62.

63.

I. J. Radiation Oncology 0 Biology 0 Physics

cell tumors: natural history and pathogenesis. J. Neurosurg. 63:155-167;1985. Jooma, R.; Kendall, B. Diagnosis and management of pineal tumors. J. Neurosurg. 58:654-665;1983. Kaplan, E. L.; Meier, P. Nonparametric estimation from incomplete observations. J. Am. Statist. Assoc. 53:457481;1958. Kersh, C. R.; Constable, W. C.; Eisert, D. R.; Spaulding, C. A.; Hahn, S. S.; Jerette, J. M.; Marks, R. D. Primary central nervous system germ cell tumors: Effect of histologic confirmation on radiotherapy. Cancer 6 1:2 148-2 152; 1988. Kitanaka, C.; Noboyuki, S.; Tadayoshi, N.; Nakamura, H.; Genka, S.; Nakagawa, K.; Akanuma, A.; Aoyama, H.; Takakura, K. Postradiation astrocytoma. J. Neurosurg. 70: 469-474; 1989. Kobayashi, T.; Yoshida, J.; Ishayama, J.; Noda, S.; Kito, A.; Kido, Y. Combination chemotherapy with cisplatin and etopiside for malignant intracranial germ cell tumors. An experimental and clinical study. J. Neurosurg. 70:678681;1989. Kumura, E.; Aria, N.; Hayakawa, T.; Ushio, Y.; Nakata, H.; Murasawa, A.; Mogami, H.; Ikezoe, J.; Minami, Y. Successful combination chemotherapy (cisplatinum, vinblastine, and bleomycin) against peritoneal dissemination of intracranial germ cell tumor. Neurosurg. 18:802-804; 1986. Legido, A.; Packer, R. J.; Sutton, L. N.; D’Angio, G.; Rorke, L, B.; Bruce, D. A.; Schut, L. Suprasellar germinomas in childhood: a reappraisal. Cancer 63:340-344; 1989. Linstadt, D.; Wara, W. M.; Edwards, M. S. B.; Hudgins, R. J.; Sheline, G. E. Radiotherapy of intracranial germinomas: the case against routine craniospinal irradiation. Int. J. Radiat. Oncol. Biol. Phys. 15:291-297;1988. Linggood, R. M.; Chapman, P. H. Pineal tumors. J. Neurooncol. 12:85-91;1992. Loeffler, J. S.; Mauch, P. Tumors in the pineal region. In: Textbook of uncommon cancer. NY: John Wiley & Sons Ltd.; 1988:693-705. Matsutani, M.; Takaura, K.; Sano, K. Primary intracranial germ cell tumors: pathology and treatment. Prog. Exp. Tumor Res. 30:307-312;1987. Mincer, F.; Meltzer, J.; Botstein, C. Pinealoma: a report of twelve irradiated cases. Cancer 37:27 13-27 18; 1976. Miyamachi, K.; Abe, H.; Tashiro, T.; Aida, T. A case of multiple teratomas combined with embryonal carcinoma. No Shinkei Geka. Neurol. Surg. 14:1161-l 165;1986. Murovic, J. A.; Ongley, J. P.; Parker, J. C.; Page, L. K. Manifestations and therapeutic considerations in pineal yolk sac tumors. Case report. J. Neurosurg. 55:303-307;1987. Nagasawa, S.; Kikuchi, H.; Yamashita, J.; Mitsuno, K. Intracranial and spinal germinomas occurring four years after spinal cord germinoma. Neurol. Med. Chir. (Tokyo) 3 1: 729-731;1991. Neuwelt, E. A. An update on the surgical treatment of malignant pineal region tumors. Clin. Neurosurg. 30:397428;1985. Okuno, S.; Ishikawa, J.; Nozaki, K.; Yamaoto, K. Recurrent intracranial germinoma refractory to conventional irradiation: effective chemotherapy consisting of cis-platin and etopiside. Neurol. Med. Chir. (Tokyo) 32:35 l-355;1992. Onoyama, Y.; Ono, K.; Nakajima, T.; Hiraoka, M.; Abe, M. Radiation therapy of pineal tumors. Radiology 130:757760; 1979. Packer, R. J.; Grossman, R. I.; Rorke, L. B.; Sutton, L. N.; Siegel, K. R.; Littman, P. Brainstem necrosis after pre-irradiation high dose methotrexate. Childs Nerv. Syst. 1:355358;1985. Packer, R. J.; Sutton, L. N.; Rorke, L. B.; Rosenstock, J. G.; Zimmerman, R. A.; Littman, P. A.; Bilaniuk, L. T.;

Volume 28, Number 1. 1994

64.

65.

66.

67.

68.

69. 70.

7 1.

72.

73.

74.

75. 76.

77.

78. 79.

80.

8 1. 82.

Bruce, D. A.; Schut, L. Intracranial embryonal cell carcinoma. Cancer 54:520-524; 1984. Packer, R. J.; Sutton, L. N.; Rosenstock, J. G.; Rorke, L. B.; Bilaniuk, L. T.; Zimmerman, R. A.; Littman, P. A.; Bruce, D. A.; Schut, L. Pineal region tumors of childhood. Pediatrics 74:97- 102; 1984. Phillips, P. C.; Carson, B. S.; Wharham, M. D.; Eggleston, J. C.; Robb, P. A.; Strauss, L. C.; Killmon, D. T.; Cordon, B. C. Pineal parenchymal tumors in children and adults: Treatment results (Abst.). International Symposium on Pediatric Neuro-Oncology, Seattle, WA, June 1-3, 1989;A29. Poppen, J. L.; Marino, R. Pinealomas and tumors of the posterior portion of the third ventricle. J. Neurosurg. 28: 357-364; 1968. Rao, Y. T. R.; Medini, E.; Haselow, R. E.; Jones, T. K.; Levitt, S. H. Pineal and ectopic pineal tumors: the role of radiation therapy. Cancer 48:708-7 13; 198 1. Rich, T. A.; Cassady, J. R.; Strand, R. D.; Winston, K. R. Radiation therapy for pineal and suprasellar germ cell tumors. Cancer 55:932-940; 1985. Rubenstein, L. J. Cytogenesis and differentiation of pineal neoplasms. Human Path. 12:44 l-448; 198 1. Sakai, N.; Nakatani, K.; Shirakami, S.; Araki, Y.; Iamo, Y.; Nishamura, Y.; Andou, T.; Yamada, H.; Yanagawa, S. Delayed effects of radiation on patients with germ cell tumors with special reference to cerebral blood flow. No. To. Shinkei 42:161-166;1990. Sakai, N.; Yamada, H.; Andoh, T.; Hirata, T.; Shimizu, K.; Shinoda, J. Primary intracranial germ-cell tumors: a retrospective analysis with special reference to long-term results of treatment and the behavior of rare types of tumors. Acta Oncol. 27:43-50; 1988. Sakai, N.; Yamada, H.; Nishimura, Y.; Yoshimura, S.; Andou, T. Suprasellar malignant teratoma with 1 l-year survival. Neurol. Med. Chir. (Tokyo) 32:70 l-705; 1992. Salazar, 0. M.; Castro-Vita, H.; Bakos, R. S.; Feldstein, L. M.; Keller, B.; Rubin, P. Radiation therapy for tumors of the pineal region. Int. J. Radiat. Oncol. Biol. Phys. 5: 491-499;1979. Sano, K.; Matsutani, M. Pinealoma (germinoma) treated by direct surgery and postoperative irradiation: A long term follow-up. Childs Brain 8:8 l-89; 198 1. Scheithauer, B. W. Neuropathology of pineal region tumors. Clin. Neurosurg. 30:351-383;1985. Senan, S.; Rampling, R.; Kaye, S. B. Malignant pineal teratomas: A report on three patients and the case for craniospinal irradiation following chemotherapy. Radiother. Oncol. 22:209-2 13; 199 1. Shibamoto, Y.; Abe, M.; Yamashita, J.; Takahashi, M.; Hiraoka, M.; Ono, K.; Tsutsui, K. Treatment results of intracranial germinoma as a function of irradiated volume. Int. J. Radiat. Oncol. Biol. Phys. 15:284-290;1988. Stein, B. M.; Fetell, M. R. Therapeutic modalities for pineal region tumors. Clin. Neurosurg. 30:445-455; 1985. Sugiyama, K.; Uozumi, T.; Kiya, K.; Mukada, K.; Arita, K.; Kurisu, K.; Hotta, T.; Ogasawara, H.; Sumida, M. Intracranial germ cell tumor with synchronous lesions in the pineal and suprasellar regions: Report of six cases and review of the literature. Surg. Neurol. 38: 114- 120; 1992. Sung, D. II.; Harisiadis, L.; Chang, C. H. Midline pineal tumors and suprasellar germinomas: highly curable by irradiation. Radiology 128:745-75 1; 1978. Takakura, K. Intracranial germ cell tumors. Clin. Neurosurg. 30:429-444;1985. Tokoro, K.; Chiba, Y.; Murase, S.; Yagishita, S.; Kyuma, Y. Subarachnoid dissemination of pineal germinoma 9 years after radiation therapy without local relapse. Neurol. Med. Chir. (Tokyo) 31:725-728;1991.

Pineal region tumors 0 B. G. FULLER etal.

83. Ueki, K.; Tanaka, R. Treatment and prognosis of pineal tumors-Experience of 110 cases. Neurol. Med. Chir. 20: l-26;1980. 84. Vaquero, J.; Ramiro, J.; Martinez, R.; Bravo, G. Neurosurgical experience with tumors of the pineal region at Clinica Puerta de Hiro. Acta Neurochir. (Wein) 116:23-32;1992.

245

85. Wara, W. M.; Jenkin, D. T.; Evans, A.; Ertel, I.; Hittle, R.; Ortega, J.; Wilson, C. B.; Hammond, D. Tumors of the pineal and suprasellar region: Childrens cancer study group treatment results 1960-1975. Cancer 43:698-701;1979. 86. Waterson, J.; Priest, J. R. Control of extraneural metastasis of a primary intracranial nongerminomatous germ cell tumor. J. Neurosurg. 71:601-604;1989.