PATHOLOGY
Melanotic Neuroectodermal Tumor of Infancy: A Systematic Review Saleh Rachidi,* Amit J. Sood,y Krishna G. Patel,z Shaun A. Nguyen,x Heidi Hamilton,k Brad W. Neville,{ and Terry A. Day, MD# Melanotic neuroectodermal tumor of infancy (MNTI) is a rare tumor, usually diagnosed within the first year of age, with a predilection for the maxilla. Although the tumor is usually benign, its rapidly growing nature and ability to cause major deformities in surrounding structures necessitate early diagnosis and intervention. It is important that medical and dental specialists are prepared to make the diagnosis and proceed with appropriate intervention. The authors performed a systematic review of the 472 reported cases from 1918 through 2013 and provided a comprehensive update on this rare entity that can have devastating effects on young patients. This investigation uncovered age at diagnosis as an important prognostic indicator, because younger age correlated with a higher recurrence rate. The authors also present a case report of a 5-month-old girl diagnosed with MNTI and review her clinical presentation and imaging and histopathologic findings. Ó 2015 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg -:1-11, 2015 Melanotic neuroectodermal tumor of infancy (MNTI) is a rare neoplasm arising mainly in infants, although it also has been reported in older children and adults. It was first described by Krompecher1 in 1918, and other names have been used to describe this lesion, such as melanotic progonoma, pigmented epulis, congenital melanocarcinoma, pigmented adamantinoma, retinal anlage tumor, and pigmented tumor of the jaw of infants. This variable nomenclature reflects the uncertainty about the tumor’s origin, which prevailed for half a century, until Borello and Gorlin2 proposed the neural crest origin of this tumor in 1966. This was based on the fact that a subset of these patients excreted large amounts of vanillylmandelic acid (VMA), which is associated with other neuroectodermal tumors, such as neuroblastoma, ganglioneur-
oblastoma, and pheochromocytoma.3 Neural crest cells are multipotent embryonic cells that ultimately differentiate into various structures, including the odontogenic ectomesenchyme, melanocytes, and neural ganglia. These cells display mesodermal and ectodermal morphologic features at different stages of their ontogeny, explaining the difficulty in deciphering the embryologic origin of these tumors and possibly explaining the biphasic cellular phenotype such tumors display.3 Differential diagnosis includes tumors such as rhabdomyosarcoma, neuroblastoma, Ewing sarcoma, and lymphoma. Clinical context and imaging studies can narrow the differential. Ultimate diagnosis is pathologic and positive staining for markers such as synaptophysin, cytokeratin, and HMB-45 can further assist in
Received from the Medical University of South Carolina, Charleston, SC.
Division of Head and Neck Oncologic Surgery; Professor and Director, Division of Head and Neck Oncology, Head and Neck
*Medical Student, Department of Microbiology and Immunology.
Tumor Center, Department of Otolaryngology-Head and Neck
yResident Physician, Head and Neck Tumor Center, Department
Surgery.
of Otolaryngology-Head and Neck Surgery.
This work is funded by the Department of Otolaryngology-Head
zAssistant Professor, Head and Neck Tumor Center, Department
and Neck Surgery, Medical University of South Carolina.
of Otolaryngology-Head and Neck Surgery.
Address correspondence and reprint requests to Dr Day: 135 Rut-
xAssociate Professor, Head and Neck Tumor Center, Department
ledge Avenue, Charleston, SC, 29425; e-mail: [email protected]
of Otolaryngology-Head and Neck Surgery. kAssistant Professor, Department of Pathology and Laboratory
Received December 17 2014 Accepted March 28 2015
Medicine.
Ó 2015 American Association of Oral and Maxillofacial Surgeons
{Professor,
Division
of
Oral
Pathology,
Department
of
0278-2391/15/00354-7
Stomatology.
http://dx.doi.org/10.1016/j.joms.2015.03.061
#Wendy and Keith Wellin Endowed Chair in Head and Neck Surgery; Vice Chair, Clinical Affairs; Vice Chair and Director,
1
2 identifying the lesion. No biological markers have been established to differentiate benign from malignant lesions. The standard of care for MNTI is surgical intervention, which is mostly curative; however, up to 15% of cases develop recurrence.4 Owing to the extremely low incidence of MNTI, no prospective studies have been conducted to provide novel treatments or prognostic markers that in turn would guide clinical management. In this report, the authors analyze all MNTI cases reported in the literature (472 cases from 1918 through 2013) and show that age at diagnosis is a strong prognostic marker predictive of disease recurrence.
Clinical Presentation MNTI mainly affects the craniofacial region, with most cases arising in the maxilla,5 although other cases have been reported in areas such as the testes6 and peripheral bones.7 Infants typically present with a rapidly growing, painless, nonulcerative mass affecting the maxilla or other craniofacial structures,4 destroying adjacent bone and causing esthetic deformities and feeding disruption. The mass is typically firm, with smooth swelling and bluish-black coloration.4 In a recent report of a cohort of 18 patients, 12 were male and 6 were female, although no meaningful gender disparity has been noted previously in the literature. Six patients in this series displayed urinary VMA and all were male. All these patients had a negative family history.4
Histopathology Microscopically, 2 cell types are appreciated. The first has a cuboidal, epithelioid appearance with a clustering or tubular pattern. These cells possess melanosomes and are positive for cytokeratin and HMB-45. The other cell type is smaller and round and has a dark nucleus and a limited cytoplasm. These cells resemble neuroblasts and show dense vesicles under electron microscopy.8,9 The infiltrative nature of the tumor and lack of a capsule, in addition to the presence of neuroblast-like cells, might falsely lead to the impression of a malignant nature of the mass if not taken in the context of the patient’s medical history. Histologic markers predicting patient prognosis are not established, with only weak evidence regarding the ability of markers such as Ki-67 and CD99 to predict tumor aggressiveness.10 Another marker, neuron-specific enolase, which has been shown to predict a poorer prognosis in neuroblastoma, showed no prognostic correlation in a cohort of 18 patients with MNTI.4
MELANOTIC NEUROECTODERMAL TUMOR OF INFANCY
Clinical Behavior and Management Given its normally benign nature and age at onset, adequate surgical resection of MNTI is mostly curative, but local recurrence has been observed in 10 to 15% of cases.4 Although usually benign, this tumor can grow rapidly and infiltrate adjacent bone, soft tissue, and the orbit. In addition, a few cases of malignant MNTI have been reported in the literature,8,11-16 and the rate of metastasis is estimated to be approximately 3%.17 Surgery has been recommended for treatment, although the necessary amount of margin clearance has not been clearly delineated.
Report of Case PRESENTATION AND DIAGNOSIS
A 5-month-old girl, well developed and otherwise healthy, presented with a mass in the mouth and left cheek that had been growing for 4 weeks. The lesion was painful with pressure and eating, with no other associated signs or symptoms. The mother reported no change in eating habits, weight, or other problems. The patient was evaluated in the head and neck multidisciplinary clinic and clinical diagnosis was made with a consideration of MNTI (Fig 1). Physical examination showed a submucosal mass of the left maxillary alveolar ridge, extending from the right maxillary primary central incisor to the left maxillary primary canine. Obvious external facial deformity was evident, with a firm, bony protuberance extending near the orbital floor. Computed tomography with contrast showed an expansile mass (4.5 3.1 3.4 cm) involving the left maxillary bone, alveolar ridge, and orbital floor, extending into and crossing the midline. Sclerotic expansion of the bone toward the nasal cavity and multiple cystic-like expansile masses with internal enhancement also were observed. At least 4 teeth also were seen floating in the edges of the mass (Fig 2). For definitive diagnosis, the patient underwent a biopsy in the operating room. Transoral curettage and biopsy of the mass disclosed charcoal-colored gelatinous contents (Fig 3) and the diagnosis of MNTI was made by histologic examination (Fig 4). Staining for the proliferation marker Ki-67 showed 20% positive cells (Fig 4). Consultations were provided by pediatric dentistry, oral and maxillofacial pathology, maxillofacial prosthodontics, and facial plastic and reconstructive surgery. The case was discussed at the head and neck multidisciplinary tumor board and surgical resection was recommended. TREATMENT
After general anesthesia was administered, comprehensive examination was performed. The tumor was protruding into the upper lip, nose, face, and left
3
RACHIDI ET AL
FIGURE 1. Preoperative pictures of patient. A, B, Pictures of the patient with a left maxillary mass at presentation to the clinic. C, Intraoral view of tumor extruding from the hard palate and alveolar process. Rachidi et al. Melanotic Neuroectodermal Tumor of Infancy. J Oral Maxillofac Surg 2015.
cheek, extending to the orbital rim and the left hard palate. Incisions were made in the mucosa, taking margins around the obvious tumor, including the tooth buds on the left side and extending from the midline to the left posterior alveolar ridge. Approximately 30% of the hard palate was involved and required
resection. The tumor was circumferentially exposed and the soft tissue component was elevated from the maxilla and hard palate, allowing for adequate bony margins to obtain good visualization. The mass was visible through the bone, with some dark discoloration. It appeared to be expanding tooth roots on
4
MELANOTIC NEUROECTODERMAL TUMOR OF INFANCY
FIGURE 2. Computed tomogram with contrast and 3-dimensional reconstruction. A, Computed tomogram with contrast showing a large maxillary mass with cystic-like and sclerotic components. B, Computed tomogram-based 3-dimensional reconstruction of the cranium shows loss of maxillary bone at the alveolar ridge, resulting from tumor invasion. C, Normal side contralateral to the tumor shown for comparison. Rachidi et al. Melanotic Neuroectodermal Tumor of Infancy. J Oral Maxillofac Surg 2015.
RACHIDI ET AL
5
FIGURE 3. Intraoperative tumor images and long-term postoperative picture. A, Preoperative view of tumor protruding from the left maxilla. B, Wide and C, close-up intraoperative views of the tumor showing pigmented and fibrous tissues in the tumor. D, Resected tumor showing pigmentation and fibrous tissue after first surgery. E, Excised tumor after second surgery. F, The patient 1 year after tumor resection. Rachidi et al. Melanotic Neuroectodermal Tumor of Infancy. J Oral Maxillofac Surg 2015.
6
MELANOTIC NEUROECTODERMAL TUMOR OF INFANCY
FIGURE 4. Hematoxylin and eosin and Ki-67 stains. Hematoxylin and eosin staining of A, original biopsy specimen and B, resected specimen. The tumor consists of a biphasic population of cells growing in nests and cords within a dense fibrotic connective tissue stroma. One cell type is the ‘‘small round blue cell’’ component, which has finely stippled nuclear chromatin, inconspicuous nucleoli, and scant cytoplasm. The second cell type is made up of slightly larger, more ‘‘epithelioid’’ cells with cytoplasm that contains abundant, uniform, finely granular melanin pigment. C, Tumor invading trabecular bone in resection specimen. Nests and cords of tumor cells invade the trabecular bone of the maxilla. D, Ki-67 (MIB-1) immunohistochemical staining, which is a cellular proliferation index marker, had a variable staining pattern. The stain highlighted approximately 20% of tumor cells in the original biopsy specimen. Rachidi et al. Melanotic Neuroectodermal Tumor of Infancy. J Oral Maxillofac Surg 2015.
the left maxillary ridge. The left maxilla was resected, extending from the lateral nasal wall to the midline, to the orbital rim, to the zygoma, and to the left posterior alveolar ridge. After bone resection, darkened areas in the thickened bone near the orbit were biopsied for frozen sections. Soft tissue, alveolar ridge, and lateral maxillary margins were resected until deemed clear. However, a final margin was taken superiorly at the orbital floor and found to have positive tumor. Owing to the unknown nature of the remainder of all the margins, it was determined to close the wound and await the final pathology, in case further orbital resection and other areas of resection were required, because excision of these margins would leave a very large facial and orbital deformity. Wound edges were advanced to provide coverage for the open wound
and the rest of the bone was left to granulate secondarily. A follow-up computed tomogram was obtained 3 weeks after surgical resection and depicted a small focus of residual disease at the orbital floor. The patient underwent transoral orbital floor resection followed by obturator placement. Nineteen months after the second resection, the patient was doing well and developing normally with no evidence of disease.
Methods Owing to the retrospective nature of this study, it was granted an exemption in writing by the Medical
7
RACHIDI ET AL
University of South Carolina institutional review board (Charleston, SC). SYSTEMATIC REVIEW
A systematic literature review was conducted by searching the PubMed and National Center for Biotechnology Information database using the keyword search term melanotic neuroectodermal tumor of infancy and the Medical Subject Heading term neuroectodermal tumor, melanotic. All articles that included MNTI cases published from January 2005 through December 2013 were included. Excluded were 1) reports published in a language other than English and without an English-language abstract and 2) cases that were published in another article. This yielded a total of 69 publications, which included the 116 cases (Supplementary Table 1).4,7,12-16,18-80 In addition, data were collected on 356 additional cases that were reported in the literature from 1918 through 2004 from 3 major systematic reviews.18,81,82 These cases were combined with the 116 cases published from January 2005 through December 2013 (Supplementary Table 1) to constitute a total of 472 cases for analysis. This analysis included gender, age at diagnosis, tumor site, and disease recurrence. Articles that did not include at least 1 of these variables were excluded from analysis for that particular variable. STATISTICAL ANALYSIS
All data analyses were performed with SPSS 22.0 (IBM Corporation, Armonk, NY), SigmaPlot 12.5 (Systat Software, Inc, San Jose, CA), and MedCal 12.0.2.0 (MedCalc Software, Belgium). From the pooled data, patient information and demographic variables, such as age, gender, and medical conditions, were described with summary statistics. In general, summary statistics consisted of categorical variables (number of patients in the statistical analysis set and percentage) and continuous variables (number of non-missing observations, mean, median, standard deviation, and interquartile range [IQR]). All continuous variables were assessed for normality by the ShapiroWilk test. Comparisons of outcomes (nominal variables) were performed using the c2 or Fisher exact test. For comparison of continuous variables between 2 groups, an independent t test or a Mann-Whitney test was used for the comparison of age at diagnosis between male and female patients. For comparison of continuous variables for at least 3 groups, 1-way analysis of variance or the Kruskal-Wallis test was performed. When analysis of variance was found to indicate a statistical difference, the Duncan multiple comparison test was used for post hoc analysis. In addition, a regression model was used to show predic-
tor variables, such as age, affecting disease behavior. A relative risk ratio was calculated for tumor recurrence. A P value of .05 was considered statistically significant for all statistical tests.
Results Since the first report of MNTI in 1918, 3 major systematic reviews have summarized the cases of MNTI,18,81,82 the most recent of which included cases reported before 2004. In the present report, a total of 69 articles with cases were identified in the English-language literature from 2005 through 2013, and these articles included 116 cases of MNTI (Table 1, Supplementary Table 1). Analysis of cases reported from 2005 through 2013 showed that 59% of cases were male and 41% were female (Table 1). The tumors were mostly located in the maxilla, followed by the mandible and skull (Table 1). In terms of tumor size, a total of 70 cases reported the size of the mass, with the greatest dimension ranging from 0.5 to 20.5 cm (average, 3.85 2.7 cm; median, Table 1. PATIENT CHARACTERISTICS—CASES FROM 2005 THROUGH 2013
Age at diagnosis (mo) Range Average (SD) Median (IQR) Boys/girls, % Site, n (%) Maxilla Skull Mandible Other Greatest dimension (cm) Range Average (SD) Median (IQR) Treatment, n (%) Surg Curettage or debulking Chemo Surg + Chemo Surg + RT Chemo + RT Surg + Chemo + RT Follow-up (mo) Range Average (SD) Median (IQR)
0-156 10.4 (23.1) 5 (4) 59/41 70 (60.3) 21 (18.1) 12 (10.3) 13 (11.2) 0.5-20.5 3.85 (2.7) 3.5 (2.68) 73 (64.6) 28 (24.7) 2 (1.8) 6 (5.3) 2 (1.8) 1 (0.9) 1 (0.9) 0.067-312 60.7 (67) 30 (84.5)
Abbreviations: Chemo, chemotherapy; IQR, interquartile range; RT, radiotherapy; SD, standard deviation; Surg, surgical resection. Rachidi et al. Melanotic Neuroectodermal Tumor of Infancy. J Oral Maxillofac Surg 2015.
8
MELANOTIC NEUROECTODERMAL TUMOR OF INFANCY
3.5 cm; IQR, 2.68 cm; Table 1). Male and female patients showed comparable sizes as measured by tumor length (data not shown). In addition to the subgroup analysis performed on recent cases (2005 through 2013), a comprehensive analysis was performed of all 472 cases in the literature from 1918 through 2013 (Table 2). Of patients with reported genders (n = 423), 56% were male and 44% were female (Table 2, Fig 5A). The age range was 0 (at birth) to 804 months (median, 4.5 months; IQR, 3 months; Table 2). Male and female patients were diagnosed at comparable ages, with a median of 4.25 months in male patients (IQR, 3 months) and 4 months in female patients (IQR, 3 months). However, a larger proportion of female patients were diagnosed after 3 years of age (6.6%) compared with male patients (2.7%), raising the average age of diagnosis in to 22.2 85.5 months in female patients compared with 9.7 29.6 months in male patients, without being statistically different (P = .91; Fig 5B). Partly consistent with previous reports, the most common site of the tumor was the maxilla, followed by the skull and the mandible (Table 2, Fig 5C). Treatment modality was available for 249 cases. The vast majority involved tumor resection (n = 208)
Table 2. PATIENT CHARACTERISTICS—COMPREHENSIVE ANALYSIS (CASES FROM 1918 THROUGH 2013)
Age at diagnosis (mo) Range Average (SD) Median (IQR) Boys/girls, % Site, n (%) Maxilla Skull Mandible Other Treatment, n (%) Surg Curettage or debulking Chemo Surg + Chemo Surg + RT Chemo + RT Surg + Chemo + RT Follow-up (mo) Range Average (SD) Median (IQR)
0-804 15 (59.7) 4.5 (3) 56/44 279 (62.2) 70 (15.6) 35 (7.8) 64 (14.3) 188 (75.5) 37 (14.9) 3 (1.2) 15 (6.0) 3 (1.2) 1 (0.4) 2 (0.8) 0.067-408 53.4 (69.4) 24 (63.5)
Abbreviations: Chemo, chemotherapy; IQR, interquartile range; RT, radiotherapy; SD, standard deviation; Surg, surgical resection. Rachidi et al. Melanotic Neuroectodermal Tumor of Infancy. J Oral Maxillofac Surg 2015.
followed by curettage only (n = 37; Table 2). No difference in recurrence rate was observed between curettage and resection (data not shown). For tumor recurrence by gender, male and female patients showed equal recurrence rates at 5 years after diagnosis (Fig 5D). Importantly, the comprehensive systematic review of all 472 cases showed that age at diagnosis is a robust prognostic indicator in MNTI. Infants who were diagnosed within the first 2 months of birth had a shorter disease-free survival time, and all recurrences occurred within 6 months from treatment (Fig 5E). In contrast, those diagnosed after 4.5 months of age had a minimal risk of recurrence, and diagnosis from 2.5 to 4 months conferred an intermediate risk (Fig 5E). Therefore, age at diagnosis is an important predictor of disease recurrence (P for trend
Melanotic Neuroectodermal Tumor of Infancy: A Systematic Review Saleh Rachidi,* Amit J. Sood,y Krishna G. Patel,z Shaun A. Nguyen,x Heidi Hamilton,k Brad W. Neville,{ and Terry A. Day, MD# Melanotic neuroectodermal tumor of infancy (MNTI) is a rare tumor, usually diagnosed within the first year of age, with a predilection for the maxilla. Although the tumor is usually benign, its rapidly growing nature and ability to cause major deformities in surrounding structures necessitate early diagnosis and intervention. It is important that medical and dental specialists are prepared to make the diagnosis and proceed with appropriate intervention. The authors performed a systematic review of the 472 reported cases from 1918 through 2013 and provided a comprehensive update on this rare entity that can have devastating effects on young patients. This investigation uncovered age at diagnosis as an important prognostic indicator, because younger age correlated with a higher recurrence rate. The authors also present a case report of a 5-month-old girl diagnosed with MNTI and review her clinical presentation and imaging and histopathologic findings. Ó 2015 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg -:1-11, 2015 Melanotic neuroectodermal tumor of infancy (MNTI) is a rare neoplasm arising mainly in infants, although it also has been reported in older children and adults. It was first described by Krompecher1 in 1918, and other names have been used to describe this lesion, such as melanotic progonoma, pigmented epulis, congenital melanocarcinoma, pigmented adamantinoma, retinal anlage tumor, and pigmented tumor of the jaw of infants. This variable nomenclature reflects the uncertainty about the tumor’s origin, which prevailed for half a century, until Borello and Gorlin2 proposed the neural crest origin of this tumor in 1966. This was based on the fact that a subset of these patients excreted large amounts of vanillylmandelic acid (VMA), which is associated with other neuroectodermal tumors, such as neuroblastoma, ganglioneur-
oblastoma, and pheochromocytoma.3 Neural crest cells are multipotent embryonic cells that ultimately differentiate into various structures, including the odontogenic ectomesenchyme, melanocytes, and neural ganglia. These cells display mesodermal and ectodermal morphologic features at different stages of their ontogeny, explaining the difficulty in deciphering the embryologic origin of these tumors and possibly explaining the biphasic cellular phenotype such tumors display.3 Differential diagnosis includes tumors such as rhabdomyosarcoma, neuroblastoma, Ewing sarcoma, and lymphoma. Clinical context and imaging studies can narrow the differential. Ultimate diagnosis is pathologic and positive staining for markers such as synaptophysin, cytokeratin, and HMB-45 can further assist in
Received from the Medical University of South Carolina, Charleston, SC.
Division of Head and Neck Oncologic Surgery; Professor and Director, Division of Head and Neck Oncology, Head and Neck
*Medical Student, Department of Microbiology and Immunology.
Tumor Center, Department of Otolaryngology-Head and Neck
yResident Physician, Head and Neck Tumor Center, Department
Surgery.
of Otolaryngology-Head and Neck Surgery.
This work is funded by the Department of Otolaryngology-Head
zAssistant Professor, Head and Neck Tumor Center, Department
and Neck Surgery, Medical University of South Carolina.
of Otolaryngology-Head and Neck Surgery.
Address correspondence and reprint requests to Dr Day: 135 Rut-
xAssociate Professor, Head and Neck Tumor Center, Department
ledge Avenue, Charleston, SC, 29425; e-mail: [email protected]
of Otolaryngology-Head and Neck Surgery. kAssistant Professor, Department of Pathology and Laboratory
Received December 17 2014 Accepted March 28 2015
Medicine.
Ó 2015 American Association of Oral and Maxillofacial Surgeons
{Professor,
Division
of
Oral
Pathology,
Department
of
0278-2391/15/00354-7
Stomatology.
http://dx.doi.org/10.1016/j.joms.2015.03.061
#Wendy and Keith Wellin Endowed Chair in Head and Neck Surgery; Vice Chair, Clinical Affairs; Vice Chair and Director,
1
2 identifying the lesion. No biological markers have been established to differentiate benign from malignant lesions. The standard of care for MNTI is surgical intervention, which is mostly curative; however, up to 15% of cases develop recurrence.4 Owing to the extremely low incidence of MNTI, no prospective studies have been conducted to provide novel treatments or prognostic markers that in turn would guide clinical management. In this report, the authors analyze all MNTI cases reported in the literature (472 cases from 1918 through 2013) and show that age at diagnosis is a strong prognostic marker predictive of disease recurrence.
Clinical Presentation MNTI mainly affects the craniofacial region, with most cases arising in the maxilla,5 although other cases have been reported in areas such as the testes6 and peripheral bones.7 Infants typically present with a rapidly growing, painless, nonulcerative mass affecting the maxilla or other craniofacial structures,4 destroying adjacent bone and causing esthetic deformities and feeding disruption. The mass is typically firm, with smooth swelling and bluish-black coloration.4 In a recent report of a cohort of 18 patients, 12 were male and 6 were female, although no meaningful gender disparity has been noted previously in the literature. Six patients in this series displayed urinary VMA and all were male. All these patients had a negative family history.4
Histopathology Microscopically, 2 cell types are appreciated. The first has a cuboidal, epithelioid appearance with a clustering or tubular pattern. These cells possess melanosomes and are positive for cytokeratin and HMB-45. The other cell type is smaller and round and has a dark nucleus and a limited cytoplasm. These cells resemble neuroblasts and show dense vesicles under electron microscopy.8,9 The infiltrative nature of the tumor and lack of a capsule, in addition to the presence of neuroblast-like cells, might falsely lead to the impression of a malignant nature of the mass if not taken in the context of the patient’s medical history. Histologic markers predicting patient prognosis are not established, with only weak evidence regarding the ability of markers such as Ki-67 and CD99 to predict tumor aggressiveness.10 Another marker, neuron-specific enolase, which has been shown to predict a poorer prognosis in neuroblastoma, showed no prognostic correlation in a cohort of 18 patients with MNTI.4
MELANOTIC NEUROECTODERMAL TUMOR OF INFANCY
Clinical Behavior and Management Given its normally benign nature and age at onset, adequate surgical resection of MNTI is mostly curative, but local recurrence has been observed in 10 to 15% of cases.4 Although usually benign, this tumor can grow rapidly and infiltrate adjacent bone, soft tissue, and the orbit. In addition, a few cases of malignant MNTI have been reported in the literature,8,11-16 and the rate of metastasis is estimated to be approximately 3%.17 Surgery has been recommended for treatment, although the necessary amount of margin clearance has not been clearly delineated.
Report of Case PRESENTATION AND DIAGNOSIS
A 5-month-old girl, well developed and otherwise healthy, presented with a mass in the mouth and left cheek that had been growing for 4 weeks. The lesion was painful with pressure and eating, with no other associated signs or symptoms. The mother reported no change in eating habits, weight, or other problems. The patient was evaluated in the head and neck multidisciplinary clinic and clinical diagnosis was made with a consideration of MNTI (Fig 1). Physical examination showed a submucosal mass of the left maxillary alveolar ridge, extending from the right maxillary primary central incisor to the left maxillary primary canine. Obvious external facial deformity was evident, with a firm, bony protuberance extending near the orbital floor. Computed tomography with contrast showed an expansile mass (4.5 3.1 3.4 cm) involving the left maxillary bone, alveolar ridge, and orbital floor, extending into and crossing the midline. Sclerotic expansion of the bone toward the nasal cavity and multiple cystic-like expansile masses with internal enhancement also were observed. At least 4 teeth also were seen floating in the edges of the mass (Fig 2). For definitive diagnosis, the patient underwent a biopsy in the operating room. Transoral curettage and biopsy of the mass disclosed charcoal-colored gelatinous contents (Fig 3) and the diagnosis of MNTI was made by histologic examination (Fig 4). Staining for the proliferation marker Ki-67 showed 20% positive cells (Fig 4). Consultations were provided by pediatric dentistry, oral and maxillofacial pathology, maxillofacial prosthodontics, and facial plastic and reconstructive surgery. The case was discussed at the head and neck multidisciplinary tumor board and surgical resection was recommended. TREATMENT
After general anesthesia was administered, comprehensive examination was performed. The tumor was protruding into the upper lip, nose, face, and left
3
RACHIDI ET AL
FIGURE 1. Preoperative pictures of patient. A, B, Pictures of the patient with a left maxillary mass at presentation to the clinic. C, Intraoral view of tumor extruding from the hard palate and alveolar process. Rachidi et al. Melanotic Neuroectodermal Tumor of Infancy. J Oral Maxillofac Surg 2015.
cheek, extending to the orbital rim and the left hard palate. Incisions were made in the mucosa, taking margins around the obvious tumor, including the tooth buds on the left side and extending from the midline to the left posterior alveolar ridge. Approximately 30% of the hard palate was involved and required
resection. The tumor was circumferentially exposed and the soft tissue component was elevated from the maxilla and hard palate, allowing for adequate bony margins to obtain good visualization. The mass was visible through the bone, with some dark discoloration. It appeared to be expanding tooth roots on
4
MELANOTIC NEUROECTODERMAL TUMOR OF INFANCY
FIGURE 2. Computed tomogram with contrast and 3-dimensional reconstruction. A, Computed tomogram with contrast showing a large maxillary mass with cystic-like and sclerotic components. B, Computed tomogram-based 3-dimensional reconstruction of the cranium shows loss of maxillary bone at the alveolar ridge, resulting from tumor invasion. C, Normal side contralateral to the tumor shown for comparison. Rachidi et al. Melanotic Neuroectodermal Tumor of Infancy. J Oral Maxillofac Surg 2015.
RACHIDI ET AL
5
FIGURE 3. Intraoperative tumor images and long-term postoperative picture. A, Preoperative view of tumor protruding from the left maxilla. B, Wide and C, close-up intraoperative views of the tumor showing pigmented and fibrous tissues in the tumor. D, Resected tumor showing pigmentation and fibrous tissue after first surgery. E, Excised tumor after second surgery. F, The patient 1 year after tumor resection. Rachidi et al. Melanotic Neuroectodermal Tumor of Infancy. J Oral Maxillofac Surg 2015.
6
MELANOTIC NEUROECTODERMAL TUMOR OF INFANCY
FIGURE 4. Hematoxylin and eosin and Ki-67 stains. Hematoxylin and eosin staining of A, original biopsy specimen and B, resected specimen. The tumor consists of a biphasic population of cells growing in nests and cords within a dense fibrotic connective tissue stroma. One cell type is the ‘‘small round blue cell’’ component, which has finely stippled nuclear chromatin, inconspicuous nucleoli, and scant cytoplasm. The second cell type is made up of slightly larger, more ‘‘epithelioid’’ cells with cytoplasm that contains abundant, uniform, finely granular melanin pigment. C, Tumor invading trabecular bone in resection specimen. Nests and cords of tumor cells invade the trabecular bone of the maxilla. D, Ki-67 (MIB-1) immunohistochemical staining, which is a cellular proliferation index marker, had a variable staining pattern. The stain highlighted approximately 20% of tumor cells in the original biopsy specimen. Rachidi et al. Melanotic Neuroectodermal Tumor of Infancy. J Oral Maxillofac Surg 2015.
the left maxillary ridge. The left maxilla was resected, extending from the lateral nasal wall to the midline, to the orbital rim, to the zygoma, and to the left posterior alveolar ridge. After bone resection, darkened areas in the thickened bone near the orbit were biopsied for frozen sections. Soft tissue, alveolar ridge, and lateral maxillary margins were resected until deemed clear. However, a final margin was taken superiorly at the orbital floor and found to have positive tumor. Owing to the unknown nature of the remainder of all the margins, it was determined to close the wound and await the final pathology, in case further orbital resection and other areas of resection were required, because excision of these margins would leave a very large facial and orbital deformity. Wound edges were advanced to provide coverage for the open wound
and the rest of the bone was left to granulate secondarily. A follow-up computed tomogram was obtained 3 weeks after surgical resection and depicted a small focus of residual disease at the orbital floor. The patient underwent transoral orbital floor resection followed by obturator placement. Nineteen months after the second resection, the patient was doing well and developing normally with no evidence of disease.
Methods Owing to the retrospective nature of this study, it was granted an exemption in writing by the Medical
7
RACHIDI ET AL
University of South Carolina institutional review board (Charleston, SC). SYSTEMATIC REVIEW
A systematic literature review was conducted by searching the PubMed and National Center for Biotechnology Information database using the keyword search term melanotic neuroectodermal tumor of infancy and the Medical Subject Heading term neuroectodermal tumor, melanotic. All articles that included MNTI cases published from January 2005 through December 2013 were included. Excluded were 1) reports published in a language other than English and without an English-language abstract and 2) cases that were published in another article. This yielded a total of 69 publications, which included the 116 cases (Supplementary Table 1).4,7,12-16,18-80 In addition, data were collected on 356 additional cases that were reported in the literature from 1918 through 2004 from 3 major systematic reviews.18,81,82 These cases were combined with the 116 cases published from January 2005 through December 2013 (Supplementary Table 1) to constitute a total of 472 cases for analysis. This analysis included gender, age at diagnosis, tumor site, and disease recurrence. Articles that did not include at least 1 of these variables were excluded from analysis for that particular variable. STATISTICAL ANALYSIS
All data analyses were performed with SPSS 22.0 (IBM Corporation, Armonk, NY), SigmaPlot 12.5 (Systat Software, Inc, San Jose, CA), and MedCal 12.0.2.0 (MedCalc Software, Belgium). From the pooled data, patient information and demographic variables, such as age, gender, and medical conditions, were described with summary statistics. In general, summary statistics consisted of categorical variables (number of patients in the statistical analysis set and percentage) and continuous variables (number of non-missing observations, mean, median, standard deviation, and interquartile range [IQR]). All continuous variables were assessed for normality by the ShapiroWilk test. Comparisons of outcomes (nominal variables) were performed using the c2 or Fisher exact test. For comparison of continuous variables between 2 groups, an independent t test or a Mann-Whitney test was used for the comparison of age at diagnosis between male and female patients. For comparison of continuous variables for at least 3 groups, 1-way analysis of variance or the Kruskal-Wallis test was performed. When analysis of variance was found to indicate a statistical difference, the Duncan multiple comparison test was used for post hoc analysis. In addition, a regression model was used to show predic-
tor variables, such as age, affecting disease behavior. A relative risk ratio was calculated for tumor recurrence. A P value of .05 was considered statistically significant for all statistical tests.
Results Since the first report of MNTI in 1918, 3 major systematic reviews have summarized the cases of MNTI,18,81,82 the most recent of which included cases reported before 2004. In the present report, a total of 69 articles with cases were identified in the English-language literature from 2005 through 2013, and these articles included 116 cases of MNTI (Table 1, Supplementary Table 1). Analysis of cases reported from 2005 through 2013 showed that 59% of cases were male and 41% were female (Table 1). The tumors were mostly located in the maxilla, followed by the mandible and skull (Table 1). In terms of tumor size, a total of 70 cases reported the size of the mass, with the greatest dimension ranging from 0.5 to 20.5 cm (average, 3.85 2.7 cm; median, Table 1. PATIENT CHARACTERISTICS—CASES FROM 2005 THROUGH 2013
Age at diagnosis (mo) Range Average (SD) Median (IQR) Boys/girls, % Site, n (%) Maxilla Skull Mandible Other Greatest dimension (cm) Range Average (SD) Median (IQR) Treatment, n (%) Surg Curettage or debulking Chemo Surg + Chemo Surg + RT Chemo + RT Surg + Chemo + RT Follow-up (mo) Range Average (SD) Median (IQR)
0-156 10.4 (23.1) 5 (4) 59/41 70 (60.3) 21 (18.1) 12 (10.3) 13 (11.2) 0.5-20.5 3.85 (2.7) 3.5 (2.68) 73 (64.6) 28 (24.7) 2 (1.8) 6 (5.3) 2 (1.8) 1 (0.9) 1 (0.9) 0.067-312 60.7 (67) 30 (84.5)
Abbreviations: Chemo, chemotherapy; IQR, interquartile range; RT, radiotherapy; SD, standard deviation; Surg, surgical resection. Rachidi et al. Melanotic Neuroectodermal Tumor of Infancy. J Oral Maxillofac Surg 2015.
8
MELANOTIC NEUROECTODERMAL TUMOR OF INFANCY
3.5 cm; IQR, 2.68 cm; Table 1). Male and female patients showed comparable sizes as measured by tumor length (data not shown). In addition to the subgroup analysis performed on recent cases (2005 through 2013), a comprehensive analysis was performed of all 472 cases in the literature from 1918 through 2013 (Table 2). Of patients with reported genders (n = 423), 56% were male and 44% were female (Table 2, Fig 5A). The age range was 0 (at birth) to 804 months (median, 4.5 months; IQR, 3 months; Table 2). Male and female patients were diagnosed at comparable ages, with a median of 4.25 months in male patients (IQR, 3 months) and 4 months in female patients (IQR, 3 months). However, a larger proportion of female patients were diagnosed after 3 years of age (6.6%) compared with male patients (2.7%), raising the average age of diagnosis in to 22.2 85.5 months in female patients compared with 9.7 29.6 months in male patients, without being statistically different (P = .91; Fig 5B). Partly consistent with previous reports, the most common site of the tumor was the maxilla, followed by the skull and the mandible (Table 2, Fig 5C). Treatment modality was available for 249 cases. The vast majority involved tumor resection (n = 208)
Table 2. PATIENT CHARACTERISTICS—COMPREHENSIVE ANALYSIS (CASES FROM 1918 THROUGH 2013)
Age at diagnosis (mo) Range Average (SD) Median (IQR) Boys/girls, % Site, n (%) Maxilla Skull Mandible Other Treatment, n (%) Surg Curettage or debulking Chemo Surg + Chemo Surg + RT Chemo + RT Surg + Chemo + RT Follow-up (mo) Range Average (SD) Median (IQR)
0-804 15 (59.7) 4.5 (3) 56/44 279 (62.2) 70 (15.6) 35 (7.8) 64 (14.3) 188 (75.5) 37 (14.9) 3 (1.2) 15 (6.0) 3 (1.2) 1 (0.4) 2 (0.8) 0.067-408 53.4 (69.4) 24 (63.5)
Abbreviations: Chemo, chemotherapy; IQR, interquartile range; RT, radiotherapy; SD, standard deviation; Surg, surgical resection. Rachidi et al. Melanotic Neuroectodermal Tumor of Infancy. J Oral Maxillofac Surg 2015.
followed by curettage only (n = 37; Table 2). No difference in recurrence rate was observed between curettage and resection (data not shown). For tumor recurrence by gender, male and female patients showed equal recurrence rates at 5 years after diagnosis (Fig 5D). Importantly, the comprehensive systematic review of all 472 cases showed that age at diagnosis is a robust prognostic indicator in MNTI. Infants who were diagnosed within the first 2 months of birth had a shorter disease-free survival time, and all recurrences occurred within 6 months from treatment (Fig 5E). In contrast, those diagnosed after 4.5 months of age had a minimal risk of recurrence, and diagnosis from 2.5 to 4 months conferred an intermediate risk (Fig 5E). Therefore, age at diagnosis is an important predictor of disease recurrence (P for trend
