PII S0360-3016(02)02744-X

CLINICAL INVESTIGATION Brain

HYPOTHYROIDISM IN CHILDREN WITH MEDULLOBLASTOMA:A COMPARISON OF 3600 AND 2340 cGY CRANIOSPINAL RADIOTHERAPY

ARNOLD C. PAULINO, M.D.

Departments of Radiation Oncology and Pediatrics, The University of Iowa College of Medicine, University of Iowa Health Care andChildren’s Hospital of Iowa, Iowa City, IA

Purpose: To determine if low-dose craniospinal irradiation (2340 cGy) with chemotherapy is associated with alower incidence of hypothyroidism compared to standard dose (3600 cGy) with or without chemotherapy inchildren with medulloblastoma.Patients and Methods: Between 1980 and 1999, 32 patients <20 years old survived after craniospinal irradiationwith or without chemotherapy. Twenty patients received 3600 cGy craniospinal irradiation (CSI), whereas 12had 2340 cGy CSI; all patients received a posterior fossa boost to a total dose 5040–5580 cGy. The median agesat the time of CSI for those receiving 2340 cGy and 3600 cGy were 7.2 and 10.2 years, respectively. Chemo-therapy (CT) was employed in 22 children. All children who received 2340 cGy had CT consisting of vincristine,CCNU, and either cisplatin or cyclophosphamide. Ten of 20 (50%) patients receiving 3600 cGy had CT; the mostcommon regimen was vincristine, CCNU, and prednisone. Serum-free thyroxine and thyroid-stimulating hor-mone concentrations were measured in all children at variable times after radiotherapy. Thyroid-stimulatinghormone responses to i.v. thyrotrophin-releasing hormone were assessed in those suspected of having centralhypothyroidism. Median follow-up for children receiving 2340 cGy was 5 years (range: 2–11.2 years), whereasfor those receiving 3600 cGy, follow-up was 12.5 years (range: 2.4–20 years).Results: Eighteen patients (56%) developed hypothyroidism at a median time after radiotherapy of 41 months(range: 10 months to 18 years). Primary hypothyroidism was more common than central hypothyroidism (38%and 19%). All 7 children <5 years developed hypothyroidism, whereas 9 of 15 (60%) ages 5–10 and 2 of 10 (20%)age >10 years had hypothyroidism (p < 0.001). Hypothyroidism was documented in 10 of 12 (83%) who had 2340cGy � CT, 6 of 10 (60%) who had 3600 cGy � CT, and 2 of 10 (20%) who had 3600 cGy without CT (p < 0.025).Conclusions: Current treatment regimens consisting of chemotherapy and 2340 cGy craniospinal irradiationfollowed by a posterior fossa boost for medulloblastoma do not show a reduction of hypothyroidism. Young ageand use of chemotherapy were associated with a higher incidence of hypothyroidism. © 2002 Elsevier ScienceInc.

Hypothyroidism, Craniospinal irradiation, Medulloblastoma, Children, Late effects.

INTRODUCTION

Children with medulloblastoma are at risk for endocrinedysfunction after craniospinal irradiation (CSI). Althoughgrowth hormone deficiency is the most common complica-tion, thyroid dysfunction can occur from either direct irra-diation of the thyroid gland (primary hypothyroidism) orirradiation of the hypothalamic-pituitary gland complex(central hypothyroidism). Radiotherapy (RT) dose is animportant determinant of primary hypothyroidism in chil-dren with Hodgkin’s disease (1). Likewise, RT dose hasbeen implicated in the development of hypothalamic-pitu-itary dysfunction in children receiving prophylactic cranialirradiation for acute lymphoblastic leukemia and definitiveor postoperative treatment for brain tumors (2, 3).

Most reports documenting hypothyroidism after CSI formedulloblastoma have employed doses�3000 cGy (4–6).

Currently, in standard-risk disease, a craniospinal dose of2340 cGy in addition to chemotherapy (CT) is used and hasbeen shown to be without survival decrement (7). The lowerdose of CSI has resulted in better neuropsychological out-come in both younger and older children (8). It is not knownwhether a craniospinal dose of 2340 cGy with CT is asso-ciated with a lower incidence of hypothyroidism comparedto the traditional 3600 cGy with or without CT. In an effortto clarify this issue, we retrospectively reviewed our insti-tutional experience to determine if thyroid dysfunction isaltered by a lower dose of CSI.

PATIENTS AND METHODS

Between 1980 and 1999, 44 children�20 years oldreceived craniospinal RT followed by a posterior fossa

Reprint requests to: Arnold C. Paulino, M.D., Department ofRadiation Oncology, Emory Clinic, 1365 Clifton Road NE, RoomA1300, Atlanta, GA 30322. Tel: (404) 778-4126; Fax: (404)

778-4139; E-mail: arnold@radonc.emory.orgReceived Sep 18, 2001, and in revised form Jan 8, 2002.

Accepted for publication Jan 14, 2002.

Int. J. Radiation Oncology Biol. Phys., Vol. 53, No. 3, pp. 543–547, 2002Copyright © 2002 Elsevier Science Inc.Printed in the USA. All rights reserved

0360-3016/02/$–see front matter

543

boost for medulloblastoma at our institution. Twelve pa-tients died of their malignancy. The remaining 32 childrenwho survived after treatment form the basis of this report.There were 22 boys and 10 girls with a median age at timeof CSI of 8.7 years (range: 21 months to 20 years). Themedian age for children receiving 2340 cGy was 7.2 years(range: 21 months to 13 years), whereas for children treatedwith 3600 cGy, it was 10.2 years (range: 21 months to 20years).

CSI was performed in the prone position using parallel-opposed lateral cranial fields that abutted a posterior spinalfield (9). In older children, upper and lower spinal fieldsseparated by a skin gap were used. The initial craniospinaljunction was set low, just above the shoulders to avoid exitof radiation to the thyroid gland from the posterior spinalfield (10). Junctions were changed every five treatments.Usually, the cranial field length was decreased, whereas thespinal field length was increased with an upward shift of thespinal isocenter. If there were two spinal fields, the upperfield length stayed the same, whereas the lower field lengthincreased; both spinal isocenters were shifted superiorlywith the corresponding skin gap. It is therefore possible forthe thyroid gland to get exit dose from the spine field witha new junction change, even if the original spinal field didnot exit or diverge into the thyroid gland.

Beam energy was 4 mV photons in 17 children, 1.25 mVphotons in 10 children, and 6 mV photons in 5 children.Twenty children received a dose of 3600 cGy, whereas 12had 2340 cGy in 180-cGy fractions to the entire neuraxis.The posterior fossa received a boost to a total dose 5040–5580 cGy using parallel-opposed lateral fields and standardlandmarks (11). The patient, tumor, and treatment charac-teristics of children receiving 2340 cGy � CT, 3600 cGy �CT, and 3600 cGy without CT are presented in Table 1.

Chemotherapy was employed in 22 children. All childrenwho received 2340 cGy had CT consisting of vincristine,

CCNU, and either cisplatin or cyclophosphamide. For the10 who received 3600 cGy CSI, the most common regimenwas 8 cycles of vincristine, CCNU, and prednisone in 7patients.

Serum-free thyroxine (T4) and thyroid-stimulating hor-mone (TSH) concentrations were measured in all children atvariable intervals after RT. Normal ranges for the free T4and TSH at our institution were 0.71–1.85 ng/dL and 0.40–3.50 uIU/mL, respectively. For most patients, thyroid func-tion tests were performed secondary to short stature, coldintolerance, fatigue, and/or constipation. TSH responses toi.v. thyrotrophin-releasing hormone were documented inchildren with growth impairment or suspicion of hypotha-lamic/pituitary axis dysfunction. Primary hypothyroidismwas defined as a low free T4 and an elevated TSH. Com-pensated hypothyroidism was defined as a high TSH levelwith a normal free T4 value. Central hypothyroidism wasdiagnosed when accompanied by a low serum T4 with anonelevated TSH and confirmed by an impaired or delayedTSH response to thyrotrophin-releasing hormone.

Median follow-up for all patients was 8.2 years (range:2–20 years). For children receiving 2340 cGy CSI, medianfollow-up time was 5 years (range: 2–11.2 years), whereasfor those receiving 3600 cGy, follow-up time was 12.5years (range: 2.4 –20 years). The chi-square test was usedto test for statistical significance between the differentpatient and treatment parameters on the development ofhypothyroidism.

RESULTS

The patient, treatment, and thyroid toxicity characteristicsfor all children are presented in Table 2. Eighteen patients(56%) developed hypothyroidism at a median time of 41months (range: 10 months to 18 years) after CSI for me-dulloblastoma. Primary and central hypothyroidism were

Table 1. Patient, tumor, and treatment characteristics according to treatment regimen

2340 cGy CSI � CT(n � 12)

3600 cGy CSI � CT(n � 10)

3600 cGy CSI(n � 10)

Age (years)Median 7.2 7.1 13Range 1.7–13 1.7–12 5.1–20

Gender, M:F 10:2 4:6 8:2Stage*

T1–T3a 8 7 10T3b–T4 3 3 0M0 11 5 7M� 1 5 3

Posterior fossa doseMedian 5400 cGy 5400 cGy 5400 cGyRange 5040–5580 cGy 5400–5580 cGy 5000–5580 cGy

RT treatment duration (days)Median 43 50 46Range 41–89 42–85 41–96

Abbreviations: CSI � craniospinal irradiation; CT � chemotherapy.* Chang stage.

544 I. J. Radiation Oncology ● Biology ● Physics Volume 53, Number 3, 2002

seen in 12 (38%) and 6 (19%) patients, respectively. Of the12 children who developed primary thyroid injury, 2 hadcompensated hypothyroidism. The median time to develop-

ment of primary hypothyroidism was 33 months (15 monthsto 18 years), whereas for central hypothyroidism, it was 43months (range: 10 months to 5 years). Central hypothyroid-ism was found in 1 of 8 (13%) and 5 of 10 (50%) receiving3600 and 2340 cGy, respectively (p � 0.25, chi-square).

Age was found to be a statistically significant parameterin the development of hypothyroidism (Fig. 1). Hypothy-roidism was seen in all 7 children �5 years old, 9 of 15(60%) ages 5–10, and 2 of 10 (20%) �10 years of age (p �0.001, chi-square test). Hypothyroidism was more commonin children who received CT and CSI (Fig. 2). It was seenin 10 of 12 (83%) receiving 2340 cGy and CT, 6 of 10(60%) receiving 3600 cGy and CT, and 2 of 10 (20%)receiving 3600 cGy without CT (p � 0.025, chi-square test).For patients receiving 2340 cGy � CT, median time tohypothyroidism was 25 months (range: 10 months to 9years), whereas for those receiving 3600 cGy � CT, mediantime was 49 months (range: 15 months to 12.7 years). Forthe 2 patients who developed hypothyroidism after CSIalone, time to development of thyroid dysfunction was 10

Fig. 1. Influence of age on development of hypothyroidism inchildren treated with craniospinal irradiation for medulloblastoma(p � 0.001, chi-square test).

Table 2. Patient treatment and thyroid toxicity characteristics

Patient GenderAge at RT

(years) Thyroid statusTime to development ofhypothyroidism (months)

Follow-up(months)

2340 cGy � CT1 F 13.0 Primary hypothyroidism 17 1202 M 1.8 Central hypothyroidism 60 1083 M 9.8 Euthyroid – 1344 M 10.0 Primary hypothyroidism 20 255 M 3.0 Central hypothyroidism 38 726 M 4.7 Central hypothyroidism 56 917 M 6.0 Primary hypothyroidism 22 308 M 9.0 Central hypothyroidism 20 689 M 5.0 Euthyroid – 39

10 F 7.3 Primary hypothyroidism 28 5311 M 7.0 Central hypothyroidism 10 2412 M 3.5 Primary hypothyroidism 108 111

3600 cGy � CT13 M 5.9 Euthyroid – 15614 F 8.3 Euthyroid – 4715 F 11.9 Euthyroid – 14816 F 12.0 Euthyroid – 22017 F 9.0 Primary hypothyroidism 105 13218 M 4.2 Primary hypothyroidism 54 23419 M 6.9 Primary hypothyroidism 15 7020 F 3.4 Primary hypothyroidism 22 17521 M 1.8 Primary hypothyroidism 152 19522 F 10.5 Central hypothyroidism 45 66

3600 cGy23 F 11.7 Euthyroid – 9624 M 20.0 Euthyroid – 2925 M 14.2 Euthyroid – 24026 M 17.0 Euthyroid – 5727 M 9.9 Euthyroid – 22928 M 6.9 Euthyroid – 13329 F 18.0 Euthyroid – 20430 M 20.0 Euthyroid – 11431 M 12.1 Primary hypothyroidism 216 23832 M 5.2 Primary hypothyroidism 120 129

Abbreviation: CT � chemotherapy.

545Hypothyroidism in medulloblastoma ● A. C. PAULINO

and 18 years. Gender was not found to have an impact onthyroid dysfunction.

DISCUSSION

Hypothyroidism is a well-known complication of radia-tion therapy for childhood malignancies. The majority ofour current information is derived from children treated forHodgkin’s disease and head-and-neck sarcomas where thethyroid gland is in the direct path of the radiation beam.Limited information is available regarding hypothyroidismin children after craniospinal irradiation. Although hypothy-roidism can occur as a result of exit dose to the thyroidgland from a spinal field, it can also result from directirradiation of the hypothalamic-pituitary complex duringcranial irradiation.

Primary hypothyroidism is more prevalent than centralhypothyroidism. In our study, the cause of thyroid dysfunc-tion was direct injury to the thyroid gland in 38% of casesand injury to the hypothalamic-pituitary complex in 19%.This is consistent with a previous report from Livesey andBrook, who documented primary hypothyroidism in 28%and central hypothyroidism in 3% (6). Chin et al. docu-mented a 62% rate of primary hypothyroidism in childrentreated with conventional RT to a mean dose of 3500 cGy(5). Most reports, however, state that up to one-third ofpatients develop primary hypothyroidism after CSI (12–14).

In our analysis, the most important factors for develop-ment of thyroid dysfunction were the age of child at time ofRT and the use of CT. Data regarding the influence of ageon hypothyroidism are conflicting. In patients withHodgkin’s disease, the incidence of hypothyroidism rose

from 15% for children �5 years to 39% for those 15–20years of age; however, most of the younger children weretreated with a lower dose of RT (15). Our study showed ahigher incidence of hypothyroidism in children. Unlike thestudy by Hancock et al. (15), a significant proportion of ourpatients were �5 years of age and treated with CSI doses of2340 or 3600 cGy. Furthermore, central hypothyroidismdoes not result from RT in Hodgkin’s disease, because thepituitary-hypothalamic complex is not irradiated. In anotherstudy of children treated with cranial RT or CSI, age wasnot found to influence thyroid dysfunction (13). Our studyconfirms two previous reports regarding the higher inci-dence of thyroid dysfunction with the use of CT and RT (6,13). In one of these studies, CSI and CT were associatedwith a 69% incidence of primary hypothyroidism. Limitedinformation is available regarding the influence of chemo-therapy alone in the development of hypothyroidism. In onestudy of 54 adults treated with chemotherapy alone forlocally advanced Hodgkin’s disease, 24 (44%) developedhypothyroidism. The agents used in the above study in-cluded CCNU, vincristine, vinblastine, procarbazine, andprednisolone (16). Most of the patients in our study whoreceived chemotherapy had CCNU and vincristine, whichmay be responsible for the higher incidence of hypothyroid-ism in those receiving CSI and chemotherapy.

The influence on thyroid dysfunction of lowering the CSIdose from 3600 cGy to 2340 cGy did not seem to be asimportant as the omission of CT in the treatment of medul-loblastoma. Whatever benefit there may have been in low-ering the CSI dose may have been negated by the use ofchemotherapy. There does seem to be a trend toward morecentral hypothyroidism compared to primary hypothyroid-ism with a lower CSI dose. Five of the 10 patients (50%)irradiated at 2340 cGy developed hypothyroidism second-ary to hypothalamic injury, whereas 7 of 8 (88%) patientsirradiated at 3600 cGy had primary thyroid gland injury.

The limitations of our study include the different intervalsat which thyroid function tests were obtained. Certainly wecould have diagnosed thyroid dysfunction earlier in somepatients had there been a “ regular” schedule. Thermolumi-nescent dosimetry or phantom measurements of dose to thethyroid gland were not routinely performed during the studyperiod. Also, we had only 12 children who received 2340cGy, and this may have affected results. Although thefollow-up time for the lower CSI dose is shorter, 83% haddeveloped hypothyroidism at last follow-up. The results ofour current study indicate that the present treatment regimenfor standard-risk medulloblastoma is not associated with alower rate of hypothyroidism.

REFERENCES

1. Constine LS, Donaldson SS, McDougall IR, et al. Thyroiddysfunction after radiotherapy in children with Hodgkin’sdisease. Cancer 1984;53:878–883.

2. Blatt J, Lee P, Suttnes J, et al. Pulsatile growth hormonesecretion in children with acute lymphoblastic leukemia after

1800 cGy cranial irradiation. Int J Radiat Oncol Biol Phys1988;15:1001–1006.

3. Rose SR, Lustig RH, Pitukcheewanont P, et al. Diagnosis ofhidden central hypothyroidism in survivors of childhood can-cer. J Clin Endocrinol Metab 1999;84:4472–4479.

Fig. 2. Influence of type of treatment on development of hypothy-roidism in children with medulloblastoma (p � 0.025, chi-squaretest).

546 I. J. Radiation Oncology ● Biology ● Physics Volume 53, Number 3, 2002

4. Oberfield SE, Allen JC, Pollack J, et al. Long-term endocrinesequelae after treatment of medulloblastoma: Prospectivestudy of growth and thyroid function. J Pediatr 1986;108:219–223.

5. Chin D, Sklar C, Donahue B, et al. Thyroid dysfunction asa late effect in survivors of pediatric medulloblastoma/primitive neuroectodermal tumors: A comparison of hyper-fractionated versus conventional radiotherapy. Cancer1997;80:798 –804.

6. Livesey EA, Brook CGD. Thyroid dysfunction after radiother-apy and chemotherapy of brain tumors. Arch Dis Child 1989;64:593–595.

7. Packer RJ, Goldwein J, Nicholson HS, et al. Treatment of chil-dren with medulloblastomas with reduced-dose craniospinal ra-diation therapy and adjuvant chemotherapy: A Children’s CancerGroup Study. J Clin Oncol 1999;17:2127–2136.

8. Mulhern RK, Kepner JL, Thomas PR, et al. Neuropsychologicfunctioning of survivors of childhood medulloblastoma ran-domized to receive conventional or reduced-dose craniospinalirradiation: A Pediatric Oncology Group Study. J Clin Oncol1998;16:1723–1728.

9. Paulino AC. Radiotherapeutic management of medulloblas-toma. Oncology 1997;11:813–823.

10. Narayana A, Jeswani S, Paulino AC. The cranial-spinal junc-tion in medulloblastoma: Does it matter? Int J Radiat OncolBiol Phys 1999;44:81–84.

11. Tarbell NJ, Smith AR, Adams J, et al. The challenge ofconformal radiotherapy in the curative treatment of medul-loblastoma. Int J Radiat Oncol Biol Phys 2000;46:265–266.

12. Duffner PH, Cohen ME, Anderson SW, et al. Long-termeffects of treatment on endocrine function in children withbrain tumors. Ann Neurol 1983;14:528–532.

13. Ogilvy-Stuart AL, Shalet SM, Gottamaneni HR. Thyroidfunction after treatment of brain tumors in children. J Pediatr1991;119:733–737.

14. Constine LS, Woolf PD, Cann D, et al. Hypothalamic-pitu-itary dysfunction after radiation for brain tumors. N EnglJ Med 1993;328:87–94.

15. Hancock S, Cox R, McDougall I. Thyroid diseases aftertreatment for Hodgkin’s disease. N Engl J Med 1991;325:599–605.

16. Sutcliffe SB, Chapman R, Wrigley PFM. Cyclical combina-tion chemotherapy and thyroid function in patients with ad-vanced Hodgkin’s disease. Med Pediatr Oncol 1981;9:439–448.

547Hypothyroidism in medulloblastoma ● A. C. PAULINO