Likelihood of Diagnosing Neuroblastoma in Isolated Horner Syndrome

Background: The need for an extensive evaluation for neuroblastoma in children with Horner syndrome is controversial. Methods: A retrospective study design was used. The cohort included 47 children with anisocoria who were diagnosed with Horner syndrome and 135 children with neuroblastoma evaluated at a pediatric medical center between 2007 and 2015. To detect neuroblastoma, patients with Horner syndrome underwent brain and cervical MRI, abdominal ultrasound, and/or measurement of urinary vanillylmandelic acid (VMA). The neuroblastoma group was evaluated for signs/symptoms of Horner syndrome at the time of diagnosis. Results: Seven patients with Horner syndrome were lost to follow-up, and the findings of the remaining 40 were categorized according to the age of the patient. Horner syndrome most frequently was idiopathic (58%), and in only 1 patient did the discovery of neuroblastoma precede the appearance of Horner syndrome. In the 21 patients aged 1-18 years, Horner syndrome was acquired in 15 patients and congenital in 6. The most common etiology was trauma (62%). Imaging was performed in 14 patients and VMA testing in 13. Neuroblastoma was diagnosed in 5 patients; in none was it related to Horner syndrome. In the 135 patients with neuroblastoma, most of the tumors were diagnosed at Stage 4 (60%) or Stage 3 (30%) with 53% originating in the abdomen. In one patient (0.74%) with signs/symptoms of Horner syndrome at diagnosis of neuroblastoma, the tumor had been identified prenatally and the diagnosis confirmed by imaging postnatally. Conclusions: The absence of occult neuroblastoma in children with Horner syndrome and of signs/symptoms of Horner syndrome in the children diagnosed with neuroblastoma suggests that Horner syndrome might not be as frequent a cause of neuroblastoma as previously thought. We recommend that full investigation for neuroblastoma be reserved for suspicious cases associated with additional systemic signs or symptoms.

Clinical Research: Epidemiology Meets Neuro-Ophthalmology Section Editors: Heather E. Moss, MD, PhD Stacy L. Pineles, MD Likelihood of Diagnosing Neuroblastoma in Isolated Horner Syndrome Avi Ben Shabat, MD, Shifra Ash, MD, Judith Luckman, MD, Helen Toledano, MD, Nitza Goldenberg-Cohen, MD Background: The need for an extensive evaluation for neuroblastoma in children with Horner syndrome is controversial. Methods: A retrospective study design was used. The cohort included 47 children with anisocoria who were diagnosed with Horner syndrome and 135 children with neuroblastoma evaluated at a pediatric medical center between 2007 and 2015. To detect neuroblastoma, patients with Horner syndrome underwent brain and cervical MRI, abdominal ultrasound, and/or measurement of urinary vanillylmandelic acid (VMA). The neuroblastoma group was evaluated for signs/symptoms of Horner syndrome at the time of diagnosis. Results: Seven patients with Horner syndrome were lost to follow-up, and the findings of the remaining 40 were categorized according to the age of the patient. Horner syndrome most frequently was idiopathic (58%), and in only 1 patient did the discovery of neuroblastoma precede the appearance of Horner syndrome. In the 21 patients aged 1- 18 years, Horner syndrome was acquired in 15 patients and congenital in 6. The most common etiology was trauma (62%). Imaging was performed in 14 patients and VMA testing in 13. Neuroblastoma was diagnosed in 5 patients; in none was it related to Horner syndrome. In the 135 patients with neuroblastoma, most of the tumors were diagnosed at Stage 4 (60%) or Stage 3 (30%) with 53% Department of Ophthalmology (NG-C), Bnai Zion Medical Center, Haifa, Israel; Krieger Eye Research Laboratory (NG-C), Felsenstein Medical Research Center, Beilinson Hospital, Petach Tikva, Israel; The Ruth and Bruce Rappaport Faculty of Medicine (NG-C), Technion, Haifa, Israel; Sackler Faculty of Medicine (ABS, HT, SA), Tel Aviv University, Tel Aviv, Israel; Department of Radiology (JL), Rabin Medical Center, Beilinson Hospital, Petach Tikva, Israel; Department of Pediatric Oncology (HT, SA), Schneider Children's Medical Center of Israel, Petach Tikva; Israel. Supported in part by the Zanvyl and Isabelle Krieger Fund, Baltimore, MD (NG-C). The funding organization had no role in the design or conduct of this research. Partially presented at the 2016 Israel Ophthalmology Society (ILOS) meeting, Tel Aviv, Israel, and the 2017 NANOS annual meeting, Washington DC. The authors report no conflicts of interest. A. Ben Shabat and S. Ash have contributed equally to this work. Address correspondence to Nitza Goldenberg-Cohen, MD, Department of Ophthalmology, Bnai Zion Medical Center, Haifa 3339419, Israel; E-mail: [e-mail] 308 originating in the abdomen. In one patient (0.74%) with signs/symptoms of Horner syndrome at diagnosis of neuroblastoma, the tumor had been identified prenatally and the diagnosis confirmed by imaging postnatally. Conclusions: The absence of occult neuroblastoma in children with Horner syndrome and of signs/symptoms of Horner syndrome in the children diagnosed with neuroblastoma suggests that Horner syndrome might not be as frequent a cause of neuroblastoma as previously thought. We recommend that full investigation for neuroblastoma be reserved for suspicious cases associated with additional systemic signs or symptoms. Journal of Neuro-Ophthalmology 2019;39:308-312 doi: 10.1097/WNO.0000000000000764 © 2019 by North American Neuro-Ophthalmology Society P ediatric Horner syndrome may be congenital or acquired. Its incidence is estimated at 1.42 per 100,000 children (1). Most cases of pediatric Horner syndrome are either idiopathic or caused by trauma, but neuroblastoma must be considered in the differential diagnosis (2-4). Neuroblastoma is a neuroendocrine tumor of the sympathetic nervous system that almost exclusively occurs in children. It is the most common extracranial solid tumor of infancy and the third most common childhood cancer after brain tumors and leukemia (5). The incidence of neuroblastoma is 1.54 per 100,000 children per year, and the prevalence is estimated at 1 per 7,000 births. Ninety percent of cases are diagnosed before the age of 5 years, including 37% in infants, usually during the first month of life (6). Although the tumor may arise at any location along the sympathetic nervous system, the most common site is the adrenal gland (40%), followed by the abdomen (25%), thoracic cavity (15%), cervical region (5%), and pelvic sympathetic ganglia (5%) (7). The need for extensive evaluation for neuroblastoma in children with Horner syndrome is controversial. George et al (8) suggested that complete imaging evaluation be performed only in children with heterochromia or positive Ben Shabat et al: J Neuro-Ophthalmol 2019; 39: 308-312 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Clinical Research: Epidemiology Meets Neuro-Ophthalmology findings on the urinary vanillylmandelic acid (VMA) test. Mahoney et al (3) recommended MRI for all cases of pediatric Horner syndrome. Smith et al (1) in a study of 20 patients (,19 years) with Horner syndrome found no cases of an underlying mass or neuroblastoma and suggested that current recommendations need to be reconsidered. In 2015, on the basis of a review of the clinical signs and symptoms of Horner syndrome, Miller and Kanagalingham (4) proposed that, in the absence of a history of trauma, all patients with congenital or new-onset Horner syndrome should undergo thorough evaluation including careful anamnesis, complete physical examination, urinary catecholamine test, and MRI of the brain, neck, chest, and abdomen. We would add abdominal ultrasonography to that list. However, evaluating children for neuroblastoma is not without risk. Sedation is usually required often with propofol, which can have a neurotoxic effect in young patients (9). In addition, there is the potential of adverse effects of gadolinium administered during MRI (10). Prompted by these reports and our 15-year experience, we sought to determine whether a comprehensive evaluation for neuroblastoma is necessary in children with Horner syndrome. In addition, we wanted to calculate the percentage of children with neuroblastoma in whom Horner syndrome was the presenting sign. METHODS The initial study cohort consisted of 47 patients with anisocoria who were diagnosed with Horner syndrome with positive cocaine eyedrop testing and 135 patients with neuroblastoma. All patients were identified retrospectively by search of the database of a pediatric medical center from 2007 to 2015. The study was approved by the institutional ethics committee. The medical files of the anisocoria group were reviewed for ptosis and the positive cocaine test. Data were collected on the type of Horner syndrome (congenital or acquired), specific cause (in acquired Horner syndrome), and means of evaluation for neuroblastoma (brain and/or cervical MRI, abdominal ultrasound, and VMA test). Horner syndrome was considered to be acquired if it was diagnosed subsequent to a traumatic injury (including surgery or difficult labor), even if the diagnosis was made before the age of 1 year. A diagnosis of congenital Horner syndrome could also be made after the age of 1 year on the basis of previous photographs. The medical records of the neuroblastoma group were reviewed for the presence of signs and symptoms of Horner syndrome at the time of diagnosis of neuroblastoma. Additional data included patient age, disease stage at the time of diagnosis, tumor origin, presence of N-Myc activation, and disease relapse. Ben Shabat et al: J Neuro-Ophthalmol 2019; 39: 308-312 RESULTS The age and sex distribution of the patients with anisocoria due to Horner syndrome and the patients with neuroblastoma and additional signs and symptoms of Horner syndrome at presentation are shown in Table 1. The workup and rates of positive findings are summarized in Table 2. Horner Syndrome Group Of the 47 children with Horner syndrome, 22 were females and 25 were males with a mean age of 4.0 years at diagnosis (median 4.1 years). The patients were divided into 2 groups by age for further analysis: less than 1 year (25 patients, including 14 less than 6 months) and 1-18 years (22 patients, including 7 aged 1-2 years). The mean age of the younger group was 6 months (median 5 months), and of the older group, 8.0 years (median 7.0 years). Horner Syndrome Group ,1 Year All patients younger than 1 year at diagnosis had anisocoria with a positive cocaine test and 11 (44%) had ptosis, and all but one, with a traumatic surgical injury at age 10 months, had congenital Horner syndrome (95%). Four had a traumatic labor and delivery, and 2 had an anatomical abnormality or dysmorphism. Six patients were lost to follow-up and excluded from the rest of the analysis. Complete evaluation was performed in 15 of the remaining 19 patients (79%). The VMA test, performed in 18 patients, was positive in 1 (6%), who was known to have neuroblastoma since birth. Of the 18 patients evaluated with an abdominal ultrasound, one had positive findings. This was the patient with a known neuroblastoma. Four patients (22%) had positive findings on MRI; the other 11 patients who underwent MRI had negative findings. Horner Syndrome Group 1-18 Years Among the 22 patients older than 1 year with Horner syndrome, ptosis was found at presentation in 19 (86%) and/or heterochromia in 5 (23%). One patient was excluded from further analyses for lack of complete data. Twenty patients (95%) had second- or third-order neuron syndrome, and 1 (5%), a first-order neuron Horner syndrome. Although older than 1 year, 6 patients (29%) were categorized as having congenital Horner syndrome on the basis of family photographs. These patients were followed up for at least 1 year. Complete evaluation was performed in 11 patients (52%) and 4 were followed up by observation only. The VMA test, performed in 13 patients (62%), was positive in 4, all with a known neuroblastoma in whom the Horner syndrome was trauma-induced. MRI, performed in 12 patients (57%), was positive in 6, including the 4 patients with known (unrelated) neuroblastoma and 2 patients with Horner syndrome due to surgical injury or hypothalamic 309 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Clinical Research: Epidemiology Meets Neuro-Ophthalmology TABLE 1. Characteristics of patients with anisocoria diagnosed with Horner syndrome Sign/Symptoms at Presentation Age at Diagnosis No. pts Whole group 0-18 yr Younger subgroup ,6 mo 6-12 mo 0-2 mo Older subgroup 1-2 yr 2-18 yr 1-18 yr Sex (M:F) Mean Age Anisocoria Ptosis Heterochromia 47 25:22 4.0 ± 5.1 yr 47 (100%) 30 (64%) 8 (17%) 14 11 25 8:6 8:3 16:9 3.0 ± 1.0 mo 10.0 ± 1.0 mo 6.0 ± 4.0 mo 14 (100%) 11 (79%) 25 (100%) 6 (43%) 5 (36%) 11 (44%) 2 (14%) 1 (7%) 3 (12%) 7 15 22 1:6 8:7 9:13 17.0 ± 3.0 mo 11.0 ± 4.1 yr 8.0 ± 6.0 yr 7 (100%) 15 (100%) 22 (100%) 6 (86%) 13 (87%) 19 (86%) 0 5 (33%) 5 (23%) glioma. Of 3 patients (14%) who underwent CT scan, 2 had positive findings: 1 after carotid artery dissection and 1 with an infection at the apex of the lung. The 6 patients with negative findings on all modes of evaluation were eventually diagnosed with idiopathic Horner syndrome. Horner syndrome was not due to neuroblastoma in any of the children aged 1-18 years. Neuroblastoma Group Of the 135 patients in the neuroblastoma group, only 1 (0.74%) had signs/and symptoms of Horner syndrome (anisocoria, miosis, and anhidrosis) at the time of diagnosis of neuroblastoma. This patient was diagnosed at birth with mediastinal neuroblastoma by prenatal ultrasound and postnatal identification of an abdominal mass. Eighty-one patients (60%) had a Stage 4 disease at diagnosis (60%) and 31 (23%) had Stage 3; 51 (38%) had tumor relapse and 38 (28%) were positive for N-Myc on genetic analysis. The tumor originated in the abdomen in 72 patients (53%), adrenal gland in 27 (20%), mediastinum in 11 (8%), cervical region in 3 (2%), and pelvis in 1 (1%). In 21 patients (16%), data on the tumor origin were lacking (Table 3). DISCUSSION In our study, no cases of occult neuroblastoma were found in the children with Horner syndrome. Although Horner syndrome was identified in 5 children in the neuroblastoma group, its occurrence was preceded by the diagnosis of the neuroblastoma in all cases. These findings suggest that Horner syndrome is an unlikely sign of occult neuroblastoma. TABLE 2. Evaluation for neuroblastoma in patients with Horner syndrome Test Performed/Positive Result, n (%) Age at Diagnosis No. pts Test VMA US MRI CT/CTA Whole group 0-18 yr 40 No. tested No. positive 31 (78) 5 (16) 35 (88) 11 (31) 27 (68) 10 (37) 4 (10) 3 (75) Younger subgroup ,6 mo 12 No. No. No. No. No. No. tested positive tested positive tested positive 11 (92) 1 (9) 7 (100) 0 18 (95) 1 (6) 11 1 7 3 18 4 (92) (9) (100) (43) (95) (22) 10 1 5 3 15 4 (83) (10) (71) (60) (79) (27) 1 (8) 1 (100) 0 No. No. No. No. No. No. tested positive tested positive tested positive 6 2 7 2 13 4 6 4 11 3 17 7 (86) (67) (79) (27) (81) (41) 5 4 7 2 12 6 (71) (80) (50) (29) (57) (50) 0 6-12 mo 7 0-12 mo 19 Older subgroup 1-2 yr 7 2-18 yr 14 1-18 yr 21 (86) (33) (50) (29) (62) (31) 1 (5) 1 (100) 3 2 3 2 (21) (67) (14) (67) CTA, computed tomographic angiography; pts, patients; US, ultrasound; VMA, vanillylmandelic acid. 310 Ben Shabat et al: J Neuro-Ophthalmol 2019; 39: 308-312 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Clinical Research: Epidemiology Meets Neuro-Ophthalmology TABLE 3. Age and stage at diagnosis of neuroblastoma Stage Age 0-18 mo 18 mo-18 yr 0-18 yr Mean ± SD 1 2 3 4 4s Total 8.0 ± 6.0 mo 3.1 ± 1.1 yr 2.1 ± 2.0 yr 10 (18%) 2 (3%) 12 (9%) 4 (7%) 1 (1%) 5 (4%) 20 (36%) 11 (14%) 31 (23%) 15 (27%) 66 (83%) 81 (60%) 6 (11%) 0 6 (4%) 55 80 135 It is noteworthy that the workup yielded nonneuroblastoma findings of glioma, carotid artery dissection, and infection at the apex of the lung. However, in these 3 cases, the patient's history was highly suggestive of the correct diagnosis. In the neuroblastoma group, 5 patients had Horner syndrome. In 4, the syndrome developed after surgery to remove the tumor. Only one patient had Horner syndrome secondary to neuroblastoma, although it was detected after the malignant disease was diagnosed. The approach to any case of pediatric Horner syndrome is challenging, and many factors need to be taken into account in decisions regarding evaluation and management (4). The results of this study call into question some of the current guidelines in the literature. First, studies have shown that one-fifth of all neuroblastomas are diagnosed either antenatally or in the first 3 months of life (7,11). The introduction of routine prenatal ultrasound 3 decades ago has significantly increased the rate of diagnosis of fetal neuroblastoma. In Israel, where this study was conducted, women may undergo up to 4 obstetric ultrasonography examinations during pregnancy (12,13). A consistently negative finding provides some reassurance of a low risk of occult neuroblastoma. Second, in young children (up to age 2 years), the clinician must take into account the risks of sedation and gadolinium when considering patient evaluation for neuroblastoma. Third, the introduction of new technologies that allow for the simultaneous measurement of the 2 urinary catecholamine metabolites, homovanillic acid (HVA) and VMA, that are crucial to the diagnosis of neuroblastoma, has potentially reduced the need for imaging. HVA and VMA are expressed as a ratio of creatinine concentration. Studies have shown that single-spot urine tests are equally effective as time-consuming and difficult-to-obtain 24-hour urine collections (14). Fourth, age at presentation is an important factor in distinguishing congenital from acquired Horner syndrome. Although there is general consensus that patients with acquired Horner syndrome must undergo complete and thorough evaluation to rule out any life-threatening illness, this may not be true for congenital Horner syndrome. In our initial cohort, 25 patients were younger than 1 year when diagnosed with Horner syndrome, including 14 who Ben Shabat et al: J Neuro-Ophthalmol 2019; 39: 308-312 were younger than 6 months. None of the 19 included in the analysis were found to have neuroblastoma. Therefore, comprehensive evaluation may not be mandatory in this age group. This assumption might be extended to age 2 years given that 7 of the 22 children in our older anisocoria group were aged between 1 and 2 years. Fifth, evaluation of children with Horner may be restricted to suspicious cases without a history of trauma, surgery, or pneumonia, in the presence of other signs or symptoms of neuroblastoma. Highly suspicious cases are defined by the following criteria: Horner syndrome with ptosis and miosis and a positive cocaine test; no history of traumatic labor or iatrogenic surgery; in toddlers, instability during walking or other neurological signs. Our study was limited by the relatively small number of patients in the Horner syndrome group, which may have been partly due to selecting a well-represented population with careful documentation and follow-up. However, the follow-up time was relatively short (1 year) in some cases. In conclusion, Horner syndrome might not be as frequent a cause of neuroblastoma as previously thought. In our case series, this association appears to be very unusual. Our results suggest that the presence of Horner syndrome is not frequently associated with neuroblastoma, and full investigations for neuroblastoma in affected patients can be reserved for suspicious cases based on historical details and clinical findings. STATEMENT OF AUTHORSHIP Category 1: a. Conception and design: S. Ash and N. GoldenbergCohen; b. Acquisition of data: A. Ben Shabat, S. Ash, J. Luckman, and N. Goldenberg-Cohen; c. Analysis and interpretation of data: A. Ben Shabat, S. Ash, J. Luckman, H. Toledano, and N. Goldenberg-Cohen. Category 2: Drafting the manuscript: A. Ben Shabat, H. Toledano, and N. Goldenberg-Cohen; Revising it for intellectual content: S. Ash, H. Toledano, and N. Goldenberg-Cohen. Category 3: a. Final approval of the completed manuscript: A. Ben Shabat, S. Ash, H. Toledano, J. Luckman, and N. Goldenberg-Cohen. REFERENCES 1. Smith SJ, Diehl N, Leavitt JA, Mohney BG. Incidence of pediatric Horner syndrome and the risk of neuroblastoma: a population-based study. Arch Ophthalmol. 2010;128:324- 329. 311 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Clinical Research: Epidemiology Meets Neuro-Ophthalmology 2. Jeffery AR, Ellis FJ, Repka MX, Buncic JR. Pediatric Horner syndrome. J AAPOS. 1998;2:159-167. 3. Mahoney NR, Liu GT, Menacker SJ, Wilson MC, Hogarty MD, Maris JM. Pediatric Horner syndrome: etiologies and roles of imaging and urine studies to detect neuroblastoma and other responsible mass lesions. Am J Ophthalmol. 2006;142:651- 659. 4. Miller N, Kanagalingam S. Horner syndrome: clinical perspectives. Eye Brain. 2015;7:35-46. 5. Gurney JG, Ross JA, Wall DA, Bleyer WA, Severson RK, Robison LL. Infant cancer in the U.S.: histology-specific incidence and trends, 1973 to 1992. J Pediatr Hematol Oncol. 1997;19:428-432. 6. London WB, Castleberry RP, Matthay KK, Look AT, Seeger RC, Shimada H, Thorner P, Brodeur G, Maris JM, Reynolds CP, Cohn SL. Evidence for an age cutoff greater than 365 days for neuroblastoma risk group stratification in the Children's Oncology Group. J Clin Oncol. 2005;23:6459-6465. 7. Ries LAG, Smith MA, Gurney JG, Linet M, Tamra T, Young JL, Bunin GR. Cancer Incidence and Survival Among Children and Adolescents: United States SEER Program 1975-1995. Bethesda, MD: National Cancer Institute, 1999. NIH Pub. No. 99-4649. 312 8. George ND, Gonzalez G, Hoyt CS. Does Horner's syndrome in infancy require investigation? Br J Ophthalmol. 1998;82:51- 54. 9. Rappaport BA, Suresh S, Hertz S, Evers AS, Orser BA. Anesthetic neurotoxicity-clinical implications of animal models. N Engl J Med. 2015;372:796-797. 10. Rogosnitzky M, Branch S. Gadolinium-based contrast agent toxicity: a review of known and proposed mechanisms. Biometals. 2016;29:365-376. 11. Nuchtern JG. Perinatal neuroblastoma. Semin Pediatr Surg. 2006;15:10-16. 12. Bronshtein M, Zimmer EZ. Prenatal ultrasound examinations: for whom, by whom, what, when and how many? Ultrasound Obstet Gynecol. 1997;10:1-4. 13. Hendler I, Blackwell SC, Bujold E, Treadwell MC, Mittal P, Sokol RJ, Sorokin Y. Suboptimal second-trimester ultrasonographic visualization of the fetal heart in obese women: should we repeat the examination? J Ultrasound Med. 2005;24:1205-1209; quiz 1210-1211. 14. Cangemi G, Barco S, Reggiardo G, Viscardi E, Di Cataldo A, Garaventa A, Melioli G, Conte M. Interchangeability between 24-hour collection and single spot urines for vanillylmandelic and homovanillic acid levels in the diagnosis of neuroblastoma. Pediatr Blood Cancer. 2013;60:E170-E172. Ben Shabat et al: J Neuro-Ophthalmol 2019; 39: 308-312 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited.