- Email: [email protected]
Delayed Diagnosis in Children with Intracranial Germ Cell Tumors
Delayed Diagnosis in Children with Intracranial Germ Cell Tumors Roshan V. Sethi, BS1, Rose Marino, MD2, Andrzej Niemierko, PhD3, Nancy J. Tarbell, MD3, Torunn I. Yock, MD3, and Shannon M. MacDonald, MD3 Objective To review symptoms and provider history in a large cohort of patients with germ cell tumors (GCTs) to highlight the variety of manifestations and assess the effect of delayed diagnosis on outcomes.
Study design Patients treated for intracranial pure germinoma and nongerminomatous GCTs at Massachusetts General Hospital between 1998 and 2012 were included (n = 70). The primary outcome was time from onset of symptoms to diagnostic imaging. Delay was defined as an interval of $6 months. Results The median duration of symptoms before diagnostic magnetic resonance imaging was 6 months (range, 2 days to 72 months). Thirty-eight of the 70 patients (54%) had a delayed diagnosis. Patients with suprasellar tumors presented with symptoms related to endocrinopathies, and patients with pineal region tumors presented with symptoms related to hydrocephalus. Most of the patients were evaluated by a general pediatrician (49%) and/or pediatric subspecialists (66%) before diagnosis. Patients with delayed diagnosis saw a greater number of physicians before diagnosis (P = .006). The majority of patients (63%) with delayed diagnosis were seen by 2 or more physicians, and many (40%) were seen by 2 or more subspecialists. Progression-free survival was similar in the patients with delayed diagnosis and those without delayed diagnosis (P = .90), but the former were more likely to present with disseminated disease at diagnosis (34% vs 6%; P = .007). Conclusion A significant proportion of patients with GCT experience a delay in time to diagnosis, in some cases despite evaluation by general pediatricians and specialists. This delay increases the risk of disseminated disease. (J Pediatr 2013;163:1448-53).
G
erm cell tumors (GCTs) are thought to develop from embryonic remnants that fail to migrate during development.1 They occur in midline locations—most commonly the gonads, followed by the mediastinum and the central nervous system (CNS)—with a peak incidence during puberty.2 In the brain, primary GCTs most often are detected in the suprasellar and pineal regions, both of which are midline structures. These tumors are grouped into 2 main categories: pure germinomas (PGs) and nongerminomatous GCTs (NGGCTs), which contain nongerminomatous elements, including teratomas, embryonal carcinomas, endodermal sinus, and choriocarcinomas.1,3-9 These nongerminomatous elements, which confer a worse prognosis, often secrete alpha-fetoprotein (AFP) and human chorionic gonadotropin (HCG). The incidence of CNS GCT in pediatric patients is 0.20 per 100 000, and GCTs represent <5% of all pediatric CNS tumors diagnosed in the US.10 At our institution, a referral center for children with brain tumors,11 we noticed that children with GCT tend to have a longer time to diagnosis compared with children with other pediatric CNS malignancies, in part because of GCTs’ varied and unusual clinical manifestations. This symptomatic period before diagnosis may be prolonged despite evaluation by multiple general and specialist pediatricians. Few previous studies have characterized this delay in diagnosis or examined its consequences.1,12 To raise awareness of the variety of symptom complexes related to specific CNS GCTs, and to assess the impact of delayed diagnosis on outcomes, we reviewed the history of symptoms and outcomes in a relatively large cohort of patients treated at our institution.
Methods All patients with an intracranial GCT treated with proton radiotherapy at the Massachusetts General Hospital (MGH) between 1998 and 2012 were included in this study. A total of 70 patients were identified. Disseminated disease was defined as the presence
AFP CNS CSF DI GCT HCG MGH MRI NGGCT PFS PG
Alpha-fetoprotein Central nervous system Cerebrospinal fluid Diabetes insipidus Germ cell tumor Human chorionic gonadotropin Massachusetts General Hospital Magnetic resonance imaging Nongerminomatous germ cell tumor Progression-free survival Pure germinoma
From the 1Harvard Medical School and Departments of 2 Pediatric Endocrinology and 3Radiation Oncology, Massachusetts General Hospital, Boston, MA R.S. is supported by the Doris Duke Clinical Research Fellowship. N.T. was on the medical advisory board of ProCure until 2008, and N.T.’s spouse continues to serve on medical advisory board of ProCure. The other authors declare no conflicts of interest. Portions of this study were presented during an oral presentation at the International Central Nervous System Germ Cell Tumor Symposium, April 17-19, 2013 in Cambridge, United Kingdom. 0022-3476/$ - see front matter. Copyright ª 2013 Mosby Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2013.06.024
1448
Vol. 163, No. 5 November 2013 of noncontiguous disease in more than 1 anatomic location or tumor cells identified in the cerebrospinal fluid (CSF); patients with bifocal suprasellar or pineal disease were not included in this category. Tumors were classified as PG or NGGCT on the basis of a histology review at MGH, along with AFP and betaHCG markers in the serum and CSF. The majority of patients underwent surgical biopsy or limited resection, followed by chemotherapy and radiotherapy. Assessment of endocrinopathies was based on abnormal laboratory values (using institutional normal ranges) and diagnosis by an attending endocrinologist. In a subset of patients, diagnosis of an endocrinopathy was inferred from prescribed replacement medication. Ethical approval for this study was granted by the MGH Institutional Review Board. We reviewed initial consultation notes written by all providers at MGH and the referring institutions and recorded the presence, duration, and type of symptoms. We noted the type and number of providers consulted by each patient before diagnosis, as well as the circumstances of the patient’s referral for a diagnostic scan. If provider records differed in their estimate of the duration of a symptom, we chose the shortest duration to prevent overestimation of the delay in time to diagnosis. We defined the time to diagnosis as the interval from the onset of symptoms to the date of diagnostic magnetic resonance imaging (MRI). Delayed diagnosis was defined as an interval of $6 months.12 The association between clinical and tumor characteristics and delayed diagnosis was evaluated using the 2-sample t test or Fisher exact test of equality of medians for continuous variables, the c2 test for categorical variables, and multivariate logistic regression. Kaplan-Meier, Cox regression, and log-rank tests were used to evaluate the risk of recurrence. Statistical analyses were performed using Stata 12 (StataCorp, College Station, Texas).
Results Thirty patients (43%) were diagnosed with NGGCT, and 40 (57%) were diagnosed with PG. The majority of primary
tumors were located in the suprasellar region (28% of NGGCT and 40% of PG), followed by the pineal region (23% of NGGCT and 33% of PG), and multiple midline regions (18% of NGGCT and 26% of PG). The remaining patient presented with a posterior fossa mass. All isolated pineal region tumors occurred in male patients. A similar pattern was noted in patients with multiple midline tumors (13 of 18 males; 72%), but not in those with suprasellar tumors (17 of 28 females; 61%). Overall location distribution differed significantly by sex (P < .0001). Symptom Type and Duration The most common symptom reported by patients with an intracranial GCT was headache (69%). Nausea and vomiting were also common (50%). Correspondingly, many patients were found to have hydrocephalus at diagnosis (50%). Other common symptoms included polyuria and/or polydipsia (59%), double vision (34%), changes in visual acuity or visual field cuts (27%), fatigue (33%), weight loss or poor growth (17%), and premature puberty (14%) (Table I). The symptom profile differed significantly by tumor location. Pineal tumors manifested with symptoms of hydrocephalus, whereas suprasellar tumors caused symptoms related to endocrinopathies (Table I). Ophthalmic symptoms were common in all patients, but the nature of these symptoms differed by location. Pineal region tumors caused photosensitivity or diplopia, and many patients (14 of 23; 61%) had a component of Parinaud syndrome, the triad of impaired upward gaze, convergence nystagmus, and impaired papillary response detected on physical examination. Comparatively fewer patients with suprasellar tumors (8 of 28; 29%) developed declining visual acuity and/or visual field limitations owing to tumor encroachment on the optic nerve and optic chiasm. Patients with multiple midline tumors exhibited various combinations of these findings. Disturbances to the pubertal axis were relatively common. Ten patients, all males, came to attention with premature puberty. Eight of these patients had an HCG-secreting tumor (serum range, 21-15 522 IU/mL) located in the pineal (n = 3),
Table I. Symptoms by tumor histology and location Symptom type
All, n (%)
NGGCT, n (%)
PG, n (%)
Suprasellar, n (%)
Pineal, n (%)
Bifocal, n (%)
Delayed, n (%)
Nondelayed, n (%)
Total Premature puberty Polyuria and/or polydipsia Weight loss Poor growth Nausea/vomiting Fatigue Headache Motor weakness Seizure Ataxia Psychiatric Diplopia Visual acuity/field changes Parinaud syndrome (at diagnosis) Papilledema (at diagnosis) Hydrocephalus (at diagnosis)
70 (100) 10 (14) 41 (59) 14 (20) 12 (17) 35 (50) 20 (29) 48 (69) 2 (3) 3 (4) 6 (9) 5 (7) 23 (33) 19 (27) 13 (19) 14 (20) 35 (70)
30 (43) 7 (23) 16 (53) 4 (13) 4 (13) 16 (53) 7 (23) 23 (77) 0 (0) 2 (7) 2 (7) 1 (3) 12 (40) 7 (23) 7 (23) 8 (27) 17 (57)
40 (57) 3 (8) 25 (63) 10 (25) 8 (20) 19 (48) 13 (33) 25 (63) 2 (5) 1 (3) 4 (10) 4 (10) 11 (28) 12 (30) 6 (15) 6 (15) 18 (45)
28 (40) 3 (11) 26 (93) 8 (29) 8 (29) 10 (36) 11 (39) 16 (57) 1 (4) 1 (4) 3 (11) 2 (7) 5 (18) 8 (29) 0 (0) 2 (7) 3 (11)
23 (33) 3 (13) 1 (4) 0 (0) 0 (0) 16 (70) 2 (9) 22 (96) 0 (0) 2 (9) 2 (9) 3 (13) 13 (57) 5 (22) 14 (61) 10 (43) 22 (96)
18 (26) 4 (22) 14 (78) 6 (33) 4 (22) 9 (50) 7 (39) 9 (50) 1 (6) 0 (0) 1 (6) 0 (0) 5 (28) 5 (28) 4 (22) 2 (11) 10 (56)
38 (54) 6 (16) 31 (82) 11 (29) 10 (26) 14 (37) 13 (34) 20 (53) 2 (5) 2 (5) 3 (8) 2 (5) 10 (26) 11 (29) 2 (5) 6 (16) 12 (32)
32 (46) 4 (12) 10 (31) 3 (9) 2 (6) 21 (66) 7 (22) 28 (88) 0 (0) 1 (3) 3 (9) 3 (9) 13 (41) 8 (25) 11 (34) 8 (25) 23 (72) 1449
THE JOURNAL OF PEDIATRICS
www.jpeds.com
suprasellar (n = 3), or bifocal (n = 2) region. The remaining 2 patients with precocious puberty had bifocal PG tumors associated with normal serum and CSF markers. In addition, 3 of the 7 female patients aged >12 years presented with amenorrhea. Poor growth, with an average duration of 27 months, was the symptom associated with the longest delays in diagnosis, and diplopia and nausea/vomiting, with an average duration of 2 weeks, were present for the shortest delays in diagnosis (Figure 1). The median overall time to diagnosis was 6 months (range, 2 days to 72 months). The median time to diagnosis differed significantly by histology (7.0 months for PG vs 3.4 months for NGGCT; P = .05) and tumor location (16.8 months for suprasellar vs 1 month for pineal vs 6.2 months for bifocal; P < .001). Medical Evaluation Thirty-four patients (49%) were evaluated by a general pediatrician before diagnosis. The majority of patients (66%) were evaluated by a pediatric specialist before diagnosis: 12 patients (17%) were seen by a neurologist, 19 (27%) by an ophthalmologist, 24 (34%) by an endocrinologist, 5 (7%) by an urologist, 5 (7%) by a gastroenterologist, and 2 (3%) by an orthopedist. Thirty-two patients (46%) were evaluated by more than 1 doctor. Thirty-nine patients (56%) were referred for diagnostic MRI from the outpatient office, and 29 (41%) underwent diagnostic MRI in the Emergency Department. The diagnostic MRI was ordered by a specialist in 34 patients (49%) and by a general pediatrician or Emergency Department physician in the remaining 36 patients (51%). An endocrinologist was the most common provider referring patients for diagnostic MRI. Endocrinopathies were common. Forty-two of the 46 patients with suprasellar or bifocal disease were evaluated by an endocrinologist before biopsy or resection. Before the diagnosis of brain tumor, 21 patients (50%) had been diagnosed with diabetes insipidus (DI), 11 (26%) with growth hormone deficiency, 8 (19%) with central hypothyroidism,
Vol. 163, No. 5 and 8 (19%) with central adrenal insufficiency. After diagnosis of a CNS GCT and before any therapeutic intervention, an additional 16 patients (38%) were diagnosed with DI, 8 (19%) with growth hormone deficiency, 13 (31%) with central hypothyroidism, and 12 (29%) with central adrenal insufficiency. All 42 patients with suprasellar or bifocal tumor with endocrine evaluation were diagnosed with at least 1 endocrinopathy. Only 4 patients did not have a diagnosis with DI; 2 of these patients developed the condition after surgery. The most common alternative diagnosis considered before diagnosis of GCT was diabetes mellitus (n = 7); other diagnoses included migraine (n = 2) and sinusitis (n = 3). One patient was thought to be pregnant based on amenorrhea and elevated urine and serum beta-HCG levels (1000 IU/ mL). Several patients received a psychiatric diagnosis, including eating disorder (n = 1), depression (n = 1), obsessive-compulsive disorder (n = 2), anxiety (n = 2), attention deficit hyperactivity disorder (n = 1), and panic disorder (n = 1). Delay to Diagnosis Thirty-eight patients (54%) had a time to diagnosis of $6 months. The demographic and clinical characteristics of patients with and without delayed diagnosis are summarized in Table II. Delay was more likely in patients with a suprasellar tumor (82% for suprasellar vs 22% for pineal vs 55% for bifocal; P < .001), those with a PG histology (68% vs 37%; P = .02), and females (73% vs 46%; P = .04), although the role of sex was confounded by the male predominance of pineal region tumors. In multivariate logistic regression analysis including sex, tumor location, and histology, sex no longer was significant (OR, 0.95; 95% CI, 0.2-4.0; P = .90); suprasellar location (OR vs other, 9.6; 95% CI, 2.3-40.5; P = .002) and PG histology (OR, 4.1; 95% CI, 1.3-13.3; P = .02) remained significant. Patients with delayed diagnosis had on average a similar number of symptoms as patients without delay (P = .20). Table II. Characteristics of patients with and without a delay in diagnosis
Age at diagnosis Median, months (range) $10 years, n <10 years, n Sex, n Male Female Classification, n NGGCT PG Location, n Suprasellar Pineal Bifocal Other
Figure 1. Median duration of symptoms. 1450
Delayed (‡6 months) (n = 38)
Nondelayed (<6 months) (n = 32)
11.7 (6.2-21.7) 29 9
11.4 (6.7-19.2) 24 8
P value .80*
.04† 22 16
26 6
11 27
19 13
23 5 10 0
5 18 8 1
.02† <.001†
*Two-sample t test with equal variances. †Fisher exact test.
Sethi et al
ORIGINAL ARTICLES
November 2013 Patients with delay saw a greater number of physicians (P = .006) before diagnosis. Most (63%) patients with delay were seen by 2 or more physicians, and many (40%) were seen by 2 or more specialists. Patients with delayed diagnosis were more likely to have disseminated disease at diagnosis (34% vs 6%; P = .007). Of the 15 patients with disseminated disease, 9 (60%) had a PG histology and 6 (40%) had an NGGCT histology. After adjusting for histology, patients with delayed diagnosis were more likely to receive craniospinal irradiation rather than whole-ventricle or involved-field irradiation (OR, 10.4; 95% CI, 1.2-86.3; P = .03). At a median follow-up of 34 months from the date of diagnosis, progression-free survival (PFS) at 3 years was similar in the 2 groups of patients (93% delayed vs 83% nondelayed; P = .20) (Figure 2). There was no difference in 3-year PFS between patients with an NGGCT and those with a PG (88% vs 89%; P = .90). In a subgroup analysis of patients with PG histology, 3-year PFS was similar in patients with delayed diagnosis and those without delay (90% vs 90%; P = .60). One patient with a nondisseminated NGGCT whose diagnosis was not delayed died of disease after local recurrence at 15.3 months after diagnosis. Reports of poor growth and changes in weight could not be verified by available records; thus, we repeated the major analysis without the contribution of these symptoms. All comparisons produced similar results and statistical significance. Eleven patients (16%) were followed with serial MRI for a median of 9 months (range, 1.5-54.5 months) before the eventual diagnosis of brain tumor. Nearly all (10 of 11; 91%) came to attention with central DI; the remaining patient exhibited precocious puberty. On the first MRI study, all patients demonstrated infundibular (pituitary stalk) thickening, but this was not considered conclusive for GCT. Serum and CSF markers were initially normal in all patients. Progressive changes in the size of the infundibulum
eventually led to biopsy in all 11 patients. In 3 patients, the first biopsy analysis revealed lymphocytic infiltrate consistent with lymphocytic hypophysitis. In 2 of these patients, after progressive growth of the infundibulum, HCG markers in the CSF eventually rose to abnormal levels (8 IU/mL and 33.2 IU/mL) despite previously negative markers, and both patients were diagnosed with PG. The third patient continued to exhibit negative markers, but infundibular growth led to a second biopsy analysis consistent with PG. In 7 patients, the first biopsy analysis was diagnostic of either PG (n = 6) or NGGCT (n = 1). One patient with PG histology developed elevated serum AFP (491 ng/mL) and CSF AFP (663 ng/mL) shortly before chemotherapy and was reclassified as NGGCT. The remaining patient underwent biopsy analysis after chemotherapy; fibrosis and reactive changes were noted, but markers before chemotherapy were suggestive of NGGCT. Overall, all 3 patients with eventual diagnosis of NGGCT had negative CSF and serum AFP and HCG markers at presentation, but eventually developed positive markers at the time of biopsy (n = 1), and shortly before chemotherapy (n = 2). Most patients (10 of 11; 91%) followed with serial MRI experienced a delay of >6 months, with a median time to diagnosis of 23 months (range, 4.5-72 months). Two patients exhibited dissemination at the time of diagnosis. In 3 patients, the first brain MRI revealed no abnormalities. All 3 patients came to medical attention with a history of polyuria and polydipsia and were followed with serial MRI studies. None underwent initial serum and CSF evaluations. The infundibulum appeared thickened on follow-up scan after 3 months in 1 patient, 16 months in 1 patient, and 36 months in 1 patient. Two patients underwent biopsy analysis, which was suggestive of PG in both cases. The third patient was diagnosed on the basis of elevated serum HCG (12 IU/mL) and CSF HCG (53 IU/mL) detected at the time of the last MRI. One of the patients with PG histology had elevated HCG (292 IU/mL) at the time of diagnosis and was considered to have an NGGCT. We performed a subgroup analysis on the risk of dissemination that excluded patients with a normal first MRI (n = 3) or nondiagnostic first biopsy analysis (n = 3). We found that patients with delayed diagnosis continued to exhibit an increased risk of dissemination (36% for delayed vs 9% for nondelayed; P = .02).
Discussion
Figure 2. PFS for patients with delayed (6 months) and nondelayed (<6 months) diagnosis.
In our study cohort, patients with symptoms of $6 months duration were more likely to have disseminated disease. In addition, most patients, particularly those with delayed diagnosis, had undergone multiple specialist evaluations before diagnosis, likely owing to the subtle and insidious nature of their symptoms. The median time to diagnosis of GCT in our cohort was 6 months. Other retrospective reviews of GCTs have reported an estimated duration of symptoms ranging from 39 days (median) to 8.4 months (mean).1,12-14 Jennings et al,1 in a
Delayed Diagnosis in Children with Intracranial Germ Cell Tumors
1451
THE JOURNAL OF PEDIATRICS
www.jpeds.com
meta-analysis of 215 patients with GCT, found that 35% of the patients with a PG were symptomatic for >6 months, and 50% of the patients with delayed diagnosis were symptomatic for >2 years. In 2007, Crawford et al12 analyzed time from symptom onset to diagnosis in 30 patients with CNS GCTs. A smaller proportion of patients (9 of 30) had a time to diagnosis of >6 months. Similar to our findings, however, the authors identified endocrinopathies secondary to suprasellar tumors as the major factor in delayed diagnosis; as in our series, growth delay was associated with the longest median duration of symptoms.12 Pediatric CNS tumors other than CNS GCTs are associated with significantly shorter symptom duration. For medulloblastoma, ependymoma, and pilocytic astrocytoma, most estimates fall between 24 days (median) and 7.7 months (average); the majority of studies report a median duration of symptoms of <4 months.13,15-20 In a meta-analysis of 98 studies and 22 619 patients, Brasme et al21 identified a median time to diagnosis for all pediatric CNS tumors of 7 weeks. Our findings underscore the difficulties faced in diagnosing this rare type of tumor. Few other pediatric malignancies can produce symptoms persisting longer than 2 years. Even fewer CNS malignancies can produce symptoms that result in a differential diagnosis so broad as to include diabetes mellitus and pregnancy. The extent of the difficulty in diagnosis depends on the region of brain involved. GCTs located in the pineal region manifest quickly with familiar brain tumor symptoms (headache, nausea/vomiting, and papilledema) related to hydrocephalus from blocked CSF flow. In addition, at the time of diagnosis, most patients with pineal tumors exhibited 1 or more of the features of Parinaud syndrome (impaired upward gaze, convergence nystagmus, and impaired papillary responses). In contrast, patients with a GCT located in the suprasellar region showed slowly progressive symptoms related to endocrinopathies. Nearly all (26 of 28; 93%) came to medical attention with polyuria and polydipsia, with premature puberty, poor growth, and amenorrhea less common. Importantly, our data indicate that endocrinopathies are not always accompanied by radiographic changes in the neurohypophysis; a small “occult” tumor is capable of disturbing hormone production. We defined the date of diagnosis as the date of the diagnostic brain MRI. In a subset of patients whose first brain MRI was not conclusive, nondiagnostic infundibular thickening or no abnormality was present, often in the setting of central DI. The ideal management of these patients is unclear, particularly given that initial biopsy analysis was suggestive of a different diagnosis (lymphocytic hypophysitis) in 3 patients. In 2000, Maghnie et al22 reviewed the outcomes of 79 children with central DI. Twenty-nine children (37%) had equivocal infundibular thickening on the first MRI. In these children, the eventual diagnosis was idiopathic DI in 18, Langerhans cell histiocytosis in 5, GCT in 5, and autoimmune polyendocrinopathy in 1. Because GCTs were rarer than idiopathic DI, even in the setting of infundibular thickening, the authors recommended pursuing the 1452
Vol. 163, No. 5 diagnosis of GCT only in those patients with progressive infundibular enlargement. Patients with delayed diagnosis were more likely to have disseminated disease at diagnosis, but were not more likely to experience recurrence after treatment. Nevertheless, extensive growth across several critical regions of the brain may increase morbidity. In addition, patients with disseminated disease receive craniospinal irradiation regardless of histology. There are many implications of receiving radiotherapy to the entire neuroaxis rather than targeted to the ventricular system or involved field; the larger field significantly increases the risk of potential late effects, including additional endocrinopathy, auditory damage, induction of second malignancy, growth impairment of the vertebral body, and cognitive loss.23-27 In our analysis, delayed diagnosis was associated with tumor histology and location. In particular, PG tumors in the suprasellar region were more likely to manifest with subtler, although not fewer, symptoms. In addition, patients with delayed diagnosis underwent examination by more subspecialists than those without delay. Similarly, in a study of 330 patients with pediatric CNS tumors, Shay et al19 found that examination by neurologists and ophthalmologists was associated with a delay in diagnosis, whereas examination by an emergency department physician was associated with a shorter time to diagnosis. The authors acknowledged a contribution of selection bias to their results; this also plays a significant role in our findings. The present study has several limitations. The estimate of duration of symptoms is subject to recall bias. We were not able to separate patient-associated delay from physicianassociated delay; some patients did not come to the attention of a physician until shortly before diagnosis. We also did not have detailed records of weight and height before diagnosis, and thus could not confirm the findings of poor growth and weight loss reported by the parents and providers. Finally, our data were collected retrospectively. Patients with GCT often are evaluated by a broad spectrum of pediatric providers. The diverse and unusual symptoms of GCTs make diagnosis difficult. A broad review of systems may aid recognition of this unusual entity. Early diagnosis may reduce the risk of dissemination at presentation, as well as the morbidity of treatment. n Submitted for publication Mar 19, 2013; last revision received May 6, 2013; accepted Jun 13, 2013. Reprint requests: Shannon M. MacDonald, MD, Department of Radiation Oncology, Massachusetts General Hospital, 55 Fruit St, Yawkey 112, Boston, MA 02114. E-mail: [email protected]
References 1. Jennings MT, Gelman R, Hochberg F. Intracranial germ-cell tumors: natural history and pathogenesis. J Neurosurg 1985;63:155-67. 2. Villano JL, Propp JM, Porter KR, Stewart AK, Valyi-Nagy T, Li X, et al. Malignant pineal germ-cell tumors: an analysis of cases from three tumor registries. Neuro Oncol 2008;10:121-30. 3. Dearnaley DP, A’Hern RP, Whittaker S, Bloom HJ. Pineal and CNS germ cell tumors: Royal Marsden Hospital experience 1962-1987. Int J Radiat Oncol Biol Phys 1990;18:773-81.
Sethi et al
ORIGINAL ARTICLES
November 2013 4. Matsutani M, Sano K, Takakura K, Fujimaki T, Nakamura O, Funata N, et al. Primary intracranial germ cell tumors: a clinical analysis of 153 histologically verified cases. J Neurosurg 1997;86:446-55. 5. Matsutani M, Takakura K, Sano K. Primary intracranial germ cell tumors: pathology and treatment. Prog Exp Tumor Res 1987;30:307-12. 6. Matsutani M, Ushio Y, Abe H, Yamashita J, Shibui S, Fujimaki T, et al. Combined chemotherapy and radiation therapy for central nervous system germ cell tumors: preliminary results of a Phase II study of the Japanese Pediatric Brain Tumor Study Group. Neurosurg Focus 1998;5:e7. 7. Modak S, Gardner S, Dunkel IJ, Balmaceda C, Rosenblum MK, Miller DC, et al. Thiotepa-based high-dose chemotherapy with autologous stem-cell rescue in patients with recurrent or progressive CNS germ cell tumors. J Clin Oncol 2004;22:1934-43. 8. Robertson PL, DaRosso RC, Allen JC. Improved prognosis of intracranial non-germinoma germ cell tumors with multimodality therapy. J Neurooncol 1997;32:71-80. 9. Smith AA, Weng E, Handler M, Foreman NK. Intracranial germ cell tumors: a single institution experience and review of the literature. J Neurooncol 2004;68:153-9. 10. Dolecek TA, Propp JM, Stroup NE, Kruchko C. CBTRUS statistical report: primary brain and central nervous system tumors diagnosed in the United States in 2005-2009. Neuro Oncol 2012;14(Suppl 5):v1-49. 11. MacDonald SM, Trofimov A, Safai S, Adams J, Fullerton B, Ebb D, et al. Proton radiotherapy for pediatric central nervous system germ cell tumors: early clinical outcomes. Int J Radiat Oncol Biol Phys 2011;79:121-9. 12. Crawford JR, Santi MR, Vezina G, Myseros JS, Keating RF, LaFond DA, et al. CNS germ cell tumor (CNSGCT) of childhood: presentation and delayed diagnosis. Neurology 2007;68:1668-73. 13. Kukal K, Dobrovoljac M, Boltshauser E, Ammann RA, Grotzer MA. Does diagnostic delay result in decreased survival in paediatric brain tumours? Eur J Pediatr 2009;168:303-10. 14. Hardenbergh PH, Golden J, Billet A, Scott RM, Shrieve DC, Silver B, et al. Intracranial germinoma: the case for lower-dose radiation therapy. Int J Radiat Oncol Biol Phys 1997;39:419-26. 15. Dobrovoljac M, Hengartner H, Boltshauser E, Grotzer MA. Delay in the diagnosis of paediatric brain tumours. Eur J Pediatr 2002;161:663-7.
16. Mehta V, Chapman A, McNeely PD, Walling S, Howes WJ. Latency between symptom onset and diagnosis of pediatric brain tumors: an Eastern Canadian geographic study. Neurosurgery 2002;51:365-72. 17. Reulecke BC, Erker CG, Fiedler BJ, Niederstadt TU, Kurlemann G. Brain tumors in children: initial symptoms and their influence on the time span between symptom onset and diagnosis. J Child Neurol 2008;23:178-83. 18. Saha V, Love S, Eden T, Micallef-Eynaud P, MacKinlay G. Determinants of symptom interval in childhood cancer. Arch Dis Child 1993;68:771-4. 19. Shay V, Fattal-Valevski A, Beni-Adani L, Constantini S. Diagnostic delay of pediatric brain tumors in Israel: a retrospective risk factor analysis. Childs Nerv Syst 2012;28:93-100. 20. Wilne SH, Ferris RC, Nathwani A, Kennedy CR. The presenting features of brain tumours: a review of 200 cases. Arch Dis Child 2006;91:502-6. 21. Brasme JF, Morfouace M, Grill J, Martinot A, Amalberti R, BonsLetouzey C, et al. Delays in diagnosis of paediatric cancers: a systematic review and comparison with expert testimony in lawsuits. Lancet Oncol 2012;13:e445-59. 22. Maghnie M, Cosi G, Genovese E, Manca-Bitti ML, Cohen A, Zecca S, et al. Central diabetes insipidus in children and young adults. N Engl J Med 2000;343:998-1007. 23. Merchant TE, Rose SR, Bosley C, Wu S, Xiong X, Lustig RH. Growth hormone secretion after conformal radiation therapy in pediatric patients with localized brain tumors. J Clin Oncol 2011;29:4776-80. 24. Mulhern RK, Fairclough D, Ochs J. A prospective comparison of neuropsychologic performance of children surviving leukemia who received 18-Gy, 24-Gy, or no cranial irradiation. J Clin Oncol 1991;9:1348-56. 25. Neglia JP, Friedman DL, Yasui Y, Mertens AC, Hammond S, Stovall M, et al. Second malignant neoplasms in five-year survivors of childhood cancer: childhood cancer survivor study. J Natl Cancer Inst 2001;93: 618-29. 26. Merchant TE, Gould CJ, Xiong X, Robbins N, Zhu J, Pritchard DL, et al. Early neuro-otologic effects of three-dimensional irradiation in children with primary brain tumors. Int J Radiat Oncol Biol Phys 2004;58:1194-207. 27. Hartley KA, Li C, Laningham FH, Krasin MJ, Xiong X, Merchant TE. Vertebral body growth after craniospinal irradiation. Int J Radiat Oncol Biol Phys 2008;70:1343-9.
Delayed Diagnosis in Children with Intracranial Germ Cell Tumors
1453
Study design Patients treated for intracranial pure germinoma and nongerminomatous GCTs at Massachusetts General Hospital between 1998 and 2012 were included (n = 70). The primary outcome was time from onset of symptoms to diagnostic imaging. Delay was defined as an interval of $6 months. Results The median duration of symptoms before diagnostic magnetic resonance imaging was 6 months (range, 2 days to 72 months). Thirty-eight of the 70 patients (54%) had a delayed diagnosis. Patients with suprasellar tumors presented with symptoms related to endocrinopathies, and patients with pineal region tumors presented with symptoms related to hydrocephalus. Most of the patients were evaluated by a general pediatrician (49%) and/or pediatric subspecialists (66%) before diagnosis. Patients with delayed diagnosis saw a greater number of physicians before diagnosis (P = .006). The majority of patients (63%) with delayed diagnosis were seen by 2 or more physicians, and many (40%) were seen by 2 or more subspecialists. Progression-free survival was similar in the patients with delayed diagnosis and those without delayed diagnosis (P = .90), but the former were more likely to present with disseminated disease at diagnosis (34% vs 6%; P = .007). Conclusion A significant proportion of patients with GCT experience a delay in time to diagnosis, in some cases despite evaluation by general pediatricians and specialists. This delay increases the risk of disseminated disease. (J Pediatr 2013;163:1448-53).
G
erm cell tumors (GCTs) are thought to develop from embryonic remnants that fail to migrate during development.1 They occur in midline locations—most commonly the gonads, followed by the mediastinum and the central nervous system (CNS)—with a peak incidence during puberty.2 In the brain, primary GCTs most often are detected in the suprasellar and pineal regions, both of which are midline structures. These tumors are grouped into 2 main categories: pure germinomas (PGs) and nongerminomatous GCTs (NGGCTs), which contain nongerminomatous elements, including teratomas, embryonal carcinomas, endodermal sinus, and choriocarcinomas.1,3-9 These nongerminomatous elements, which confer a worse prognosis, often secrete alpha-fetoprotein (AFP) and human chorionic gonadotropin (HCG). The incidence of CNS GCT in pediatric patients is 0.20 per 100 000, and GCTs represent <5% of all pediatric CNS tumors diagnosed in the US.10 At our institution, a referral center for children with brain tumors,11 we noticed that children with GCT tend to have a longer time to diagnosis compared with children with other pediatric CNS malignancies, in part because of GCTs’ varied and unusual clinical manifestations. This symptomatic period before diagnosis may be prolonged despite evaluation by multiple general and specialist pediatricians. Few previous studies have characterized this delay in diagnosis or examined its consequences.1,12 To raise awareness of the variety of symptom complexes related to specific CNS GCTs, and to assess the impact of delayed diagnosis on outcomes, we reviewed the history of symptoms and outcomes in a relatively large cohort of patients treated at our institution.
Methods All patients with an intracranial GCT treated with proton radiotherapy at the Massachusetts General Hospital (MGH) between 1998 and 2012 were included in this study. A total of 70 patients were identified. Disseminated disease was defined as the presence
AFP CNS CSF DI GCT HCG MGH MRI NGGCT PFS PG
Alpha-fetoprotein Central nervous system Cerebrospinal fluid Diabetes insipidus Germ cell tumor Human chorionic gonadotropin Massachusetts General Hospital Magnetic resonance imaging Nongerminomatous germ cell tumor Progression-free survival Pure germinoma
From the 1Harvard Medical School and Departments of 2 Pediatric Endocrinology and 3Radiation Oncology, Massachusetts General Hospital, Boston, MA R.S. is supported by the Doris Duke Clinical Research Fellowship. N.T. was on the medical advisory board of ProCure until 2008, and N.T.’s spouse continues to serve on medical advisory board of ProCure. The other authors declare no conflicts of interest. Portions of this study were presented during an oral presentation at the International Central Nervous System Germ Cell Tumor Symposium, April 17-19, 2013 in Cambridge, United Kingdom. 0022-3476/$ - see front matter. Copyright ª 2013 Mosby Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2013.06.024
1448
Vol. 163, No. 5 November 2013 of noncontiguous disease in more than 1 anatomic location or tumor cells identified in the cerebrospinal fluid (CSF); patients with bifocal suprasellar or pineal disease were not included in this category. Tumors were classified as PG or NGGCT on the basis of a histology review at MGH, along with AFP and betaHCG markers in the serum and CSF. The majority of patients underwent surgical biopsy or limited resection, followed by chemotherapy and radiotherapy. Assessment of endocrinopathies was based on abnormal laboratory values (using institutional normal ranges) and diagnosis by an attending endocrinologist. In a subset of patients, diagnosis of an endocrinopathy was inferred from prescribed replacement medication. Ethical approval for this study was granted by the MGH Institutional Review Board. We reviewed initial consultation notes written by all providers at MGH and the referring institutions and recorded the presence, duration, and type of symptoms. We noted the type and number of providers consulted by each patient before diagnosis, as well as the circumstances of the patient’s referral for a diagnostic scan. If provider records differed in their estimate of the duration of a symptom, we chose the shortest duration to prevent overestimation of the delay in time to diagnosis. We defined the time to diagnosis as the interval from the onset of symptoms to the date of diagnostic magnetic resonance imaging (MRI). Delayed diagnosis was defined as an interval of $6 months.12 The association between clinical and tumor characteristics and delayed diagnosis was evaluated using the 2-sample t test or Fisher exact test of equality of medians for continuous variables, the c2 test for categorical variables, and multivariate logistic regression. Kaplan-Meier, Cox regression, and log-rank tests were used to evaluate the risk of recurrence. Statistical analyses were performed using Stata 12 (StataCorp, College Station, Texas).
Results Thirty patients (43%) were diagnosed with NGGCT, and 40 (57%) were diagnosed with PG. The majority of primary
tumors were located in the suprasellar region (28% of NGGCT and 40% of PG), followed by the pineal region (23% of NGGCT and 33% of PG), and multiple midline regions (18% of NGGCT and 26% of PG). The remaining patient presented with a posterior fossa mass. All isolated pineal region tumors occurred in male patients. A similar pattern was noted in patients with multiple midline tumors (13 of 18 males; 72%), but not in those with suprasellar tumors (17 of 28 females; 61%). Overall location distribution differed significantly by sex (P < .0001). Symptom Type and Duration The most common symptom reported by patients with an intracranial GCT was headache (69%). Nausea and vomiting were also common (50%). Correspondingly, many patients were found to have hydrocephalus at diagnosis (50%). Other common symptoms included polyuria and/or polydipsia (59%), double vision (34%), changes in visual acuity or visual field cuts (27%), fatigue (33%), weight loss or poor growth (17%), and premature puberty (14%) (Table I). The symptom profile differed significantly by tumor location. Pineal tumors manifested with symptoms of hydrocephalus, whereas suprasellar tumors caused symptoms related to endocrinopathies (Table I). Ophthalmic symptoms were common in all patients, but the nature of these symptoms differed by location. Pineal region tumors caused photosensitivity or diplopia, and many patients (14 of 23; 61%) had a component of Parinaud syndrome, the triad of impaired upward gaze, convergence nystagmus, and impaired papillary response detected on physical examination. Comparatively fewer patients with suprasellar tumors (8 of 28; 29%) developed declining visual acuity and/or visual field limitations owing to tumor encroachment on the optic nerve and optic chiasm. Patients with multiple midline tumors exhibited various combinations of these findings. Disturbances to the pubertal axis were relatively common. Ten patients, all males, came to attention with premature puberty. Eight of these patients had an HCG-secreting tumor (serum range, 21-15 522 IU/mL) located in the pineal (n = 3),
Table I. Symptoms by tumor histology and location Symptom type
All, n (%)
NGGCT, n (%)
PG, n (%)
Suprasellar, n (%)
Pineal, n (%)
Bifocal, n (%)
Delayed, n (%)
Nondelayed, n (%)
Total Premature puberty Polyuria and/or polydipsia Weight loss Poor growth Nausea/vomiting Fatigue Headache Motor weakness Seizure Ataxia Psychiatric Diplopia Visual acuity/field changes Parinaud syndrome (at diagnosis) Papilledema (at diagnosis) Hydrocephalus (at diagnosis)
70 (100) 10 (14) 41 (59) 14 (20) 12 (17) 35 (50) 20 (29) 48 (69) 2 (3) 3 (4) 6 (9) 5 (7) 23 (33) 19 (27) 13 (19) 14 (20) 35 (70)
30 (43) 7 (23) 16 (53) 4 (13) 4 (13) 16 (53) 7 (23) 23 (77) 0 (0) 2 (7) 2 (7) 1 (3) 12 (40) 7 (23) 7 (23) 8 (27) 17 (57)
40 (57) 3 (8) 25 (63) 10 (25) 8 (20) 19 (48) 13 (33) 25 (63) 2 (5) 1 (3) 4 (10) 4 (10) 11 (28) 12 (30) 6 (15) 6 (15) 18 (45)
28 (40) 3 (11) 26 (93) 8 (29) 8 (29) 10 (36) 11 (39) 16 (57) 1 (4) 1 (4) 3 (11) 2 (7) 5 (18) 8 (29) 0 (0) 2 (7) 3 (11)
23 (33) 3 (13) 1 (4) 0 (0) 0 (0) 16 (70) 2 (9) 22 (96) 0 (0) 2 (9) 2 (9) 3 (13) 13 (57) 5 (22) 14 (61) 10 (43) 22 (96)
18 (26) 4 (22) 14 (78) 6 (33) 4 (22) 9 (50) 7 (39) 9 (50) 1 (6) 0 (0) 1 (6) 0 (0) 5 (28) 5 (28) 4 (22) 2 (11) 10 (56)
38 (54) 6 (16) 31 (82) 11 (29) 10 (26) 14 (37) 13 (34) 20 (53) 2 (5) 2 (5) 3 (8) 2 (5) 10 (26) 11 (29) 2 (5) 6 (16) 12 (32)
32 (46) 4 (12) 10 (31) 3 (9) 2 (6) 21 (66) 7 (22) 28 (88) 0 (0) 1 (3) 3 (9) 3 (9) 13 (41) 8 (25) 11 (34) 8 (25) 23 (72) 1449
THE JOURNAL OF PEDIATRICS
www.jpeds.com
suprasellar (n = 3), or bifocal (n = 2) region. The remaining 2 patients with precocious puberty had bifocal PG tumors associated with normal serum and CSF markers. In addition, 3 of the 7 female patients aged >12 years presented with amenorrhea. Poor growth, with an average duration of 27 months, was the symptom associated with the longest delays in diagnosis, and diplopia and nausea/vomiting, with an average duration of 2 weeks, were present for the shortest delays in diagnosis (Figure 1). The median overall time to diagnosis was 6 months (range, 2 days to 72 months). The median time to diagnosis differed significantly by histology (7.0 months for PG vs 3.4 months for NGGCT; P = .05) and tumor location (16.8 months for suprasellar vs 1 month for pineal vs 6.2 months for bifocal; P < .001). Medical Evaluation Thirty-four patients (49%) were evaluated by a general pediatrician before diagnosis. The majority of patients (66%) were evaluated by a pediatric specialist before diagnosis: 12 patients (17%) were seen by a neurologist, 19 (27%) by an ophthalmologist, 24 (34%) by an endocrinologist, 5 (7%) by an urologist, 5 (7%) by a gastroenterologist, and 2 (3%) by an orthopedist. Thirty-two patients (46%) were evaluated by more than 1 doctor. Thirty-nine patients (56%) were referred for diagnostic MRI from the outpatient office, and 29 (41%) underwent diagnostic MRI in the Emergency Department. The diagnostic MRI was ordered by a specialist in 34 patients (49%) and by a general pediatrician or Emergency Department physician in the remaining 36 patients (51%). An endocrinologist was the most common provider referring patients for diagnostic MRI. Endocrinopathies were common. Forty-two of the 46 patients with suprasellar or bifocal disease were evaluated by an endocrinologist before biopsy or resection. Before the diagnosis of brain tumor, 21 patients (50%) had been diagnosed with diabetes insipidus (DI), 11 (26%) with growth hormone deficiency, 8 (19%) with central hypothyroidism,
Vol. 163, No. 5 and 8 (19%) with central adrenal insufficiency. After diagnosis of a CNS GCT and before any therapeutic intervention, an additional 16 patients (38%) were diagnosed with DI, 8 (19%) with growth hormone deficiency, 13 (31%) with central hypothyroidism, and 12 (29%) with central adrenal insufficiency. All 42 patients with suprasellar or bifocal tumor with endocrine evaluation were diagnosed with at least 1 endocrinopathy. Only 4 patients did not have a diagnosis with DI; 2 of these patients developed the condition after surgery. The most common alternative diagnosis considered before diagnosis of GCT was diabetes mellitus (n = 7); other diagnoses included migraine (n = 2) and sinusitis (n = 3). One patient was thought to be pregnant based on amenorrhea and elevated urine and serum beta-HCG levels (1000 IU/ mL). Several patients received a psychiatric diagnosis, including eating disorder (n = 1), depression (n = 1), obsessive-compulsive disorder (n = 2), anxiety (n = 2), attention deficit hyperactivity disorder (n = 1), and panic disorder (n = 1). Delay to Diagnosis Thirty-eight patients (54%) had a time to diagnosis of $6 months. The demographic and clinical characteristics of patients with and without delayed diagnosis are summarized in Table II. Delay was more likely in patients with a suprasellar tumor (82% for suprasellar vs 22% for pineal vs 55% for bifocal; P < .001), those with a PG histology (68% vs 37%; P = .02), and females (73% vs 46%; P = .04), although the role of sex was confounded by the male predominance of pineal region tumors. In multivariate logistic regression analysis including sex, tumor location, and histology, sex no longer was significant (OR, 0.95; 95% CI, 0.2-4.0; P = .90); suprasellar location (OR vs other, 9.6; 95% CI, 2.3-40.5; P = .002) and PG histology (OR, 4.1; 95% CI, 1.3-13.3; P = .02) remained significant. Patients with delayed diagnosis had on average a similar number of symptoms as patients without delay (P = .20). Table II. Characteristics of patients with and without a delay in diagnosis
Age at diagnosis Median, months (range) $10 years, n <10 years, n Sex, n Male Female Classification, n NGGCT PG Location, n Suprasellar Pineal Bifocal Other
Figure 1. Median duration of symptoms. 1450
Delayed (‡6 months) (n = 38)
Nondelayed (<6 months) (n = 32)
11.7 (6.2-21.7) 29 9
11.4 (6.7-19.2) 24 8
P value .80*
.04† 22 16
26 6
11 27
19 13
23 5 10 0
5 18 8 1
.02† <.001†
*Two-sample t test with equal variances. †Fisher exact test.
Sethi et al
ORIGINAL ARTICLES
November 2013 Patients with delay saw a greater number of physicians (P = .006) before diagnosis. Most (63%) patients with delay were seen by 2 or more physicians, and many (40%) were seen by 2 or more specialists. Patients with delayed diagnosis were more likely to have disseminated disease at diagnosis (34% vs 6%; P = .007). Of the 15 patients with disseminated disease, 9 (60%) had a PG histology and 6 (40%) had an NGGCT histology. After adjusting for histology, patients with delayed diagnosis were more likely to receive craniospinal irradiation rather than whole-ventricle or involved-field irradiation (OR, 10.4; 95% CI, 1.2-86.3; P = .03). At a median follow-up of 34 months from the date of diagnosis, progression-free survival (PFS) at 3 years was similar in the 2 groups of patients (93% delayed vs 83% nondelayed; P = .20) (Figure 2). There was no difference in 3-year PFS between patients with an NGGCT and those with a PG (88% vs 89%; P = .90). In a subgroup analysis of patients with PG histology, 3-year PFS was similar in patients with delayed diagnosis and those without delay (90% vs 90%; P = .60). One patient with a nondisseminated NGGCT whose diagnosis was not delayed died of disease after local recurrence at 15.3 months after diagnosis. Reports of poor growth and changes in weight could not be verified by available records; thus, we repeated the major analysis without the contribution of these symptoms. All comparisons produced similar results and statistical significance. Eleven patients (16%) were followed with serial MRI for a median of 9 months (range, 1.5-54.5 months) before the eventual diagnosis of brain tumor. Nearly all (10 of 11; 91%) came to attention with central DI; the remaining patient exhibited precocious puberty. On the first MRI study, all patients demonstrated infundibular (pituitary stalk) thickening, but this was not considered conclusive for GCT. Serum and CSF markers were initially normal in all patients. Progressive changes in the size of the infundibulum
eventually led to biopsy in all 11 patients. In 3 patients, the first biopsy analysis revealed lymphocytic infiltrate consistent with lymphocytic hypophysitis. In 2 of these patients, after progressive growth of the infundibulum, HCG markers in the CSF eventually rose to abnormal levels (8 IU/mL and 33.2 IU/mL) despite previously negative markers, and both patients were diagnosed with PG. The third patient continued to exhibit negative markers, but infundibular growth led to a second biopsy analysis consistent with PG. In 7 patients, the first biopsy analysis was diagnostic of either PG (n = 6) or NGGCT (n = 1). One patient with PG histology developed elevated serum AFP (491 ng/mL) and CSF AFP (663 ng/mL) shortly before chemotherapy and was reclassified as NGGCT. The remaining patient underwent biopsy analysis after chemotherapy; fibrosis and reactive changes were noted, but markers before chemotherapy were suggestive of NGGCT. Overall, all 3 patients with eventual diagnosis of NGGCT had negative CSF and serum AFP and HCG markers at presentation, but eventually developed positive markers at the time of biopsy (n = 1), and shortly before chemotherapy (n = 2). Most patients (10 of 11; 91%) followed with serial MRI experienced a delay of >6 months, with a median time to diagnosis of 23 months (range, 4.5-72 months). Two patients exhibited dissemination at the time of diagnosis. In 3 patients, the first brain MRI revealed no abnormalities. All 3 patients came to medical attention with a history of polyuria and polydipsia and were followed with serial MRI studies. None underwent initial serum and CSF evaluations. The infundibulum appeared thickened on follow-up scan after 3 months in 1 patient, 16 months in 1 patient, and 36 months in 1 patient. Two patients underwent biopsy analysis, which was suggestive of PG in both cases. The third patient was diagnosed on the basis of elevated serum HCG (12 IU/mL) and CSF HCG (53 IU/mL) detected at the time of the last MRI. One of the patients with PG histology had elevated HCG (292 IU/mL) at the time of diagnosis and was considered to have an NGGCT. We performed a subgroup analysis on the risk of dissemination that excluded patients with a normal first MRI (n = 3) or nondiagnostic first biopsy analysis (n = 3). We found that patients with delayed diagnosis continued to exhibit an increased risk of dissemination (36% for delayed vs 9% for nondelayed; P = .02).
Discussion
Figure 2. PFS for patients with delayed (6 months) and nondelayed (<6 months) diagnosis.
In our study cohort, patients with symptoms of $6 months duration were more likely to have disseminated disease. In addition, most patients, particularly those with delayed diagnosis, had undergone multiple specialist evaluations before diagnosis, likely owing to the subtle and insidious nature of their symptoms. The median time to diagnosis of GCT in our cohort was 6 months. Other retrospective reviews of GCTs have reported an estimated duration of symptoms ranging from 39 days (median) to 8.4 months (mean).1,12-14 Jennings et al,1 in a
Delayed Diagnosis in Children with Intracranial Germ Cell Tumors
1451
THE JOURNAL OF PEDIATRICS
www.jpeds.com
meta-analysis of 215 patients with GCT, found that 35% of the patients with a PG were symptomatic for >6 months, and 50% of the patients with delayed diagnosis were symptomatic for >2 years. In 2007, Crawford et al12 analyzed time from symptom onset to diagnosis in 30 patients with CNS GCTs. A smaller proportion of patients (9 of 30) had a time to diagnosis of >6 months. Similar to our findings, however, the authors identified endocrinopathies secondary to suprasellar tumors as the major factor in delayed diagnosis; as in our series, growth delay was associated with the longest median duration of symptoms.12 Pediatric CNS tumors other than CNS GCTs are associated with significantly shorter symptom duration. For medulloblastoma, ependymoma, and pilocytic astrocytoma, most estimates fall between 24 days (median) and 7.7 months (average); the majority of studies report a median duration of symptoms of <4 months.13,15-20 In a meta-analysis of 98 studies and 22 619 patients, Brasme et al21 identified a median time to diagnosis for all pediatric CNS tumors of 7 weeks. Our findings underscore the difficulties faced in diagnosing this rare type of tumor. Few other pediatric malignancies can produce symptoms persisting longer than 2 years. Even fewer CNS malignancies can produce symptoms that result in a differential diagnosis so broad as to include diabetes mellitus and pregnancy. The extent of the difficulty in diagnosis depends on the region of brain involved. GCTs located in the pineal region manifest quickly with familiar brain tumor symptoms (headache, nausea/vomiting, and papilledema) related to hydrocephalus from blocked CSF flow. In addition, at the time of diagnosis, most patients with pineal tumors exhibited 1 or more of the features of Parinaud syndrome (impaired upward gaze, convergence nystagmus, and impaired papillary responses). In contrast, patients with a GCT located in the suprasellar region showed slowly progressive symptoms related to endocrinopathies. Nearly all (26 of 28; 93%) came to medical attention with polyuria and polydipsia, with premature puberty, poor growth, and amenorrhea less common. Importantly, our data indicate that endocrinopathies are not always accompanied by radiographic changes in the neurohypophysis; a small “occult” tumor is capable of disturbing hormone production. We defined the date of diagnosis as the date of the diagnostic brain MRI. In a subset of patients whose first brain MRI was not conclusive, nondiagnostic infundibular thickening or no abnormality was present, often in the setting of central DI. The ideal management of these patients is unclear, particularly given that initial biopsy analysis was suggestive of a different diagnosis (lymphocytic hypophysitis) in 3 patients. In 2000, Maghnie et al22 reviewed the outcomes of 79 children with central DI. Twenty-nine children (37%) had equivocal infundibular thickening on the first MRI. In these children, the eventual diagnosis was idiopathic DI in 18, Langerhans cell histiocytosis in 5, GCT in 5, and autoimmune polyendocrinopathy in 1. Because GCTs were rarer than idiopathic DI, even in the setting of infundibular thickening, the authors recommended pursuing the 1452
Vol. 163, No. 5 diagnosis of GCT only in those patients with progressive infundibular enlargement. Patients with delayed diagnosis were more likely to have disseminated disease at diagnosis, but were not more likely to experience recurrence after treatment. Nevertheless, extensive growth across several critical regions of the brain may increase morbidity. In addition, patients with disseminated disease receive craniospinal irradiation regardless of histology. There are many implications of receiving radiotherapy to the entire neuroaxis rather than targeted to the ventricular system or involved field; the larger field significantly increases the risk of potential late effects, including additional endocrinopathy, auditory damage, induction of second malignancy, growth impairment of the vertebral body, and cognitive loss.23-27 In our analysis, delayed diagnosis was associated with tumor histology and location. In particular, PG tumors in the suprasellar region were more likely to manifest with subtler, although not fewer, symptoms. In addition, patients with delayed diagnosis underwent examination by more subspecialists than those without delay. Similarly, in a study of 330 patients with pediatric CNS tumors, Shay et al19 found that examination by neurologists and ophthalmologists was associated with a delay in diagnosis, whereas examination by an emergency department physician was associated with a shorter time to diagnosis. The authors acknowledged a contribution of selection bias to their results; this also plays a significant role in our findings. The present study has several limitations. The estimate of duration of symptoms is subject to recall bias. We were not able to separate patient-associated delay from physicianassociated delay; some patients did not come to the attention of a physician until shortly before diagnosis. We also did not have detailed records of weight and height before diagnosis, and thus could not confirm the findings of poor growth and weight loss reported by the parents and providers. Finally, our data were collected retrospectively. Patients with GCT often are evaluated by a broad spectrum of pediatric providers. The diverse and unusual symptoms of GCTs make diagnosis difficult. A broad review of systems may aid recognition of this unusual entity. Early diagnosis may reduce the risk of dissemination at presentation, as well as the morbidity of treatment. n Submitted for publication Mar 19, 2013; last revision received May 6, 2013; accepted Jun 13, 2013. Reprint requests: Shannon M. MacDonald, MD, Department of Radiation Oncology, Massachusetts General Hospital, 55 Fruit St, Yawkey 112, Boston, MA 02114. E-mail: [email protected]
References 1. Jennings MT, Gelman R, Hochberg F. Intracranial germ-cell tumors: natural history and pathogenesis. J Neurosurg 1985;63:155-67. 2. Villano JL, Propp JM, Porter KR, Stewart AK, Valyi-Nagy T, Li X, et al. Malignant pineal germ-cell tumors: an analysis of cases from three tumor registries. Neuro Oncol 2008;10:121-30. 3. Dearnaley DP, A’Hern RP, Whittaker S, Bloom HJ. Pineal and CNS germ cell tumors: Royal Marsden Hospital experience 1962-1987. Int J Radiat Oncol Biol Phys 1990;18:773-81.
Sethi et al
ORIGINAL ARTICLES
November 2013 4. Matsutani M, Sano K, Takakura K, Fujimaki T, Nakamura O, Funata N, et al. Primary intracranial germ cell tumors: a clinical analysis of 153 histologically verified cases. J Neurosurg 1997;86:446-55. 5. Matsutani M, Takakura K, Sano K. Primary intracranial germ cell tumors: pathology and treatment. Prog Exp Tumor Res 1987;30:307-12. 6. Matsutani M, Ushio Y, Abe H, Yamashita J, Shibui S, Fujimaki T, et al. Combined chemotherapy and radiation therapy for central nervous system germ cell tumors: preliminary results of a Phase II study of the Japanese Pediatric Brain Tumor Study Group. Neurosurg Focus 1998;5:e7. 7. Modak S, Gardner S, Dunkel IJ, Balmaceda C, Rosenblum MK, Miller DC, et al. Thiotepa-based high-dose chemotherapy with autologous stem-cell rescue in patients with recurrent or progressive CNS germ cell tumors. J Clin Oncol 2004;22:1934-43. 8. Robertson PL, DaRosso RC, Allen JC. Improved prognosis of intracranial non-germinoma germ cell tumors with multimodality therapy. J Neurooncol 1997;32:71-80. 9. Smith AA, Weng E, Handler M, Foreman NK. Intracranial germ cell tumors: a single institution experience and review of the literature. J Neurooncol 2004;68:153-9. 10. Dolecek TA, Propp JM, Stroup NE, Kruchko C. CBTRUS statistical report: primary brain and central nervous system tumors diagnosed in the United States in 2005-2009. Neuro Oncol 2012;14(Suppl 5):v1-49. 11. MacDonald SM, Trofimov A, Safai S, Adams J, Fullerton B, Ebb D, et al. Proton radiotherapy for pediatric central nervous system germ cell tumors: early clinical outcomes. Int J Radiat Oncol Biol Phys 2011;79:121-9. 12. Crawford JR, Santi MR, Vezina G, Myseros JS, Keating RF, LaFond DA, et al. CNS germ cell tumor (CNSGCT) of childhood: presentation and delayed diagnosis. Neurology 2007;68:1668-73. 13. Kukal K, Dobrovoljac M, Boltshauser E, Ammann RA, Grotzer MA. Does diagnostic delay result in decreased survival in paediatric brain tumours? Eur J Pediatr 2009;168:303-10. 14. Hardenbergh PH, Golden J, Billet A, Scott RM, Shrieve DC, Silver B, et al. Intracranial germinoma: the case for lower-dose radiation therapy. Int J Radiat Oncol Biol Phys 1997;39:419-26. 15. Dobrovoljac M, Hengartner H, Boltshauser E, Grotzer MA. Delay in the diagnosis of paediatric brain tumours. Eur J Pediatr 2002;161:663-7.
16. Mehta V, Chapman A, McNeely PD, Walling S, Howes WJ. Latency between symptom onset and diagnosis of pediatric brain tumors: an Eastern Canadian geographic study. Neurosurgery 2002;51:365-72. 17. Reulecke BC, Erker CG, Fiedler BJ, Niederstadt TU, Kurlemann G. Brain tumors in children: initial symptoms and their influence on the time span between symptom onset and diagnosis. J Child Neurol 2008;23:178-83. 18. Saha V, Love S, Eden T, Micallef-Eynaud P, MacKinlay G. Determinants of symptom interval in childhood cancer. Arch Dis Child 1993;68:771-4. 19. Shay V, Fattal-Valevski A, Beni-Adani L, Constantini S. Diagnostic delay of pediatric brain tumors in Israel: a retrospective risk factor analysis. Childs Nerv Syst 2012;28:93-100. 20. Wilne SH, Ferris RC, Nathwani A, Kennedy CR. The presenting features of brain tumours: a review of 200 cases. Arch Dis Child 2006;91:502-6. 21. Brasme JF, Morfouace M, Grill J, Martinot A, Amalberti R, BonsLetouzey C, et al. Delays in diagnosis of paediatric cancers: a systematic review and comparison with expert testimony in lawsuits. Lancet Oncol 2012;13:e445-59. 22. Maghnie M, Cosi G, Genovese E, Manca-Bitti ML, Cohen A, Zecca S, et al. Central diabetes insipidus in children and young adults. N Engl J Med 2000;343:998-1007. 23. Merchant TE, Rose SR, Bosley C, Wu S, Xiong X, Lustig RH. Growth hormone secretion after conformal radiation therapy in pediatric patients with localized brain tumors. J Clin Oncol 2011;29:4776-80. 24. Mulhern RK, Fairclough D, Ochs J. A prospective comparison of neuropsychologic performance of children surviving leukemia who received 18-Gy, 24-Gy, or no cranial irradiation. J Clin Oncol 1991;9:1348-56. 25. Neglia JP, Friedman DL, Yasui Y, Mertens AC, Hammond S, Stovall M, et al. Second malignant neoplasms in five-year survivors of childhood cancer: childhood cancer survivor study. J Natl Cancer Inst 2001;93: 618-29. 26. Merchant TE, Gould CJ, Xiong X, Robbins N, Zhu J, Pritchard DL, et al. Early neuro-otologic effects of three-dimensional irradiation in children with primary brain tumors. Int J Radiat Oncol Biol Phys 2004;58:1194-207. 27. Hartley KA, Li C, Laningham FH, Krasin MJ, Xiong X, Merchant TE. Vertebral body growth after craniospinal irradiation. Int J Radiat Oncol Biol Phys 2008;70:1343-9.
Delayed Diagnosis in Children with Intracranial Germ Cell Tumors
1453