Evolution of Surgical Management of Pineal Region Tumors in the Pediatric Population: A 17-Year Experience at a Single Institution

Abstract Introduction: Pineal region tumors have historically been challenging to treat. Advances in surgical techniques have led to significant changes in care and outcomes for these patients, and this is well demonstrated by our single institution’s experience over a 17-year-period in which the evolution of diagnosis, treatment, and outcomes of pineal tumors in pediatric patients will be outlined. Methods: We retrospectively collected data on all pediatric patients with pineal region lesions treated with surgery at Children’s National Hospital (CNH) from 2005 to 2021. Variables analyzed included presenting symptoms, presence of hydrocephalus, diagnostic and surgical approach, pathology, and adverse events, among others. IRB approval was obtained (IRB: STUDY00000009), and consent was waived due to minimal risk to patients included. Results: A total of 43 pediatric patients with pineal region tumors were treated during a 17-year period. Most tumors in our series were germinomas (n = 13, 29.5%) followed by pineoblastomas (n = 10, 22.7%). Twenty seven of the 43 patients (62.8%) in our series received a biopsy to establish diagnosis, and 44.4% went on to have surgery for resection. The most common open approach was posterior interhemispheric (PIH, transcallosal) – used for 59.3% of the patients. Gross total resection was achieved in 50%; recurrence occurred in 20.9% and mortality in 11% over a median follow-up of 47 months. Endoscopic third ventriculostomy (ETV) was employed to treat hydrocephalus in 26 of the 38 patients (68.4%) and was significantly more likely to be performed from 2011 to 2021. Most (73%) of the patients who received an ETV also underwent a concurrent endoscopic biopsy. No difference was found in recurrence rate or mortality in patients who underwent resection compared to those who did not, but complications were more frequent with resection. There was disagreement between frozen and final pathology in 18.4% of biopsies. Conclusion: This series describes the evolution of surgical approaches and outcomes over a 17-year-period at a single institution. Complication rates were higher with open resection, reinforcing the safety of pursuing endoscopic biopsy as an initial approach. The most significant changes occurred in the preferential use of ETVs over ventriculoperitoneal shunts. Though there has been a significant evolution in our understanding of and treatment for these tumors, in our series, the outcomes for these patients have not significantly changed over that time.


Introduction
Pineal region tumors in pediatric patients while uncommon remain of significant importance due to their diagnostic and surgical challenges.The pathology that can occur in the pineal region is varied and includes tumors of the pineal parenchyma such as pineocytomas or papillary tumors of the pineal region, germ cell tumors such as germinomas, teratomas, choriocarcinomas, endodermal sinus tumors (also termed yolk sac tumors), as well as embryonal tumors (atypical teratoid rhabdoid tumors, primitive neuroectodermal tumors including pineoblastomas, and medulloblastomas) or other glioneuronal tumors that may occur in or encroach upon the pineal region (Fig. 1).Patients often come to clinical attention for symptoms of hydrocephalus such as headaches and vomiting, as well as eye movement or other visual abnormalities, in addition to precocious puberty [1].
Large database reviews of pediatric brain tumors suggest that pineal tumors make up 0.4-1% of brain tumors [2][3][4].However, these reports are often collated using diagnostic coding that may fail to group germ cell tumors or other tumors that occur in the pineal region together with histologically defined pineal parenchymal tumors, likely leading to an underestimate of the overall incidence of pineal region tumors.Studies that categorize tumors by anatomical location suggest that pineal region tumors make up 2.6-13% of pediatric brain tumors [3,5,6], which is far greater than the 0.4-0.6% of brain tumors that occur in the pineal region in adults [7,8].True pineal parenchymal tumors make up less than 30% of tumors occurring in the pineal region [9].Germ cell tumors are an important pineal region tumor subgroup and make up 3.7-10.6% of all pediatric brain tumors [2][3][4]; 42.4-50% of these occur in the pineal region [10,11].Two-thirds of germ cells and up to 50% of pineal region tumors are germinomas [1,11].

Evolution of Surgical Management of Pineal Tumors
Pineal tumor types show varying proclivities for age and ethnicity groups.Pineal parenchymal tumors comprise around 1% of brain tumors for all pediatric age groups while germ cell tumors peak in infants and teenagers [2,3].Pineal parenchymal tumors as well as germ cell tumors show a male predominance [3], and germ cell tumors occur more often in Asian children [2,11] while pineal parenchymal tumors may occur more in non-Hispanic than in Hispanic children [3].
Tumors of the pineal region vary widely in aggressiveness and optimal management.The overall 5-year survival rate of anatomically defined pineal region tumors is 76.2%, but this is heavily influenced by tumor pathology.Pineal parenchymal tumors are associated with a 88.5% 5-year survival rate and germ cell tumors yield a 72.6-91% 5-year survival rate [3,4,10].Pineoblastomas, which are aggressive embryonal tumors and can disseminate throughout the neuraxis, are associated with only a 46.7% 3-year survival [12].This contrasts with germinomas, which respond well to radiation and have a 90% 5-year survival rate with radiation treatment alone [11]; non-germinonatous germ cell tumors carry a worse prognosis [13,14].
Tumors of the pineal region are variably addressed with combinations of biopsy, resection, radiation therapy, and/or chemotherapy, which is tailored to tumor histology [1].Stereotactic biopsy has been a mainstay for diagnosis of pineal region lesions with 1.3% mortality and 7% morbidity but a high diagnostic yield (approximately 95.8%) [4,19,20].Over time, with the advent of endoscopic techniques and especially the increasing use of ETV to treat the often-associated hydrocephalus in these patients, endoscopic biopsies became progressively more accepted [3,24,25].When it becomes necessary for more extensive surgical resection, such as pineoblastomas, there have been innovations in surgical approaches and techniques to resect these tumors safely and efficiently [1,17].Recently, it has been demonstrated that endoscopicassisted surgery can be utilized in these patients with significantly higher rates of gross total resection and no difference in complication rates as compared to microsurgery [22].
The presence of β-human chorionic gonadotropin (β-HCG) or alpha-fetoprotein (AFP) markers in blood or CSF can deliver a diagnosis of secreting germ cell tumor, which may be treated conservatively with neoadjuvant radiation or chemotherapy before consideration of resection of residual tumor [6,15].However, a subset of germ cell tumors (nonresponsive to radiation/chemotherapy) need resection even after initial chemoradiation, and pineal region tumors without positive germ cell markers require biopsy and/or resection [14].Thus, most of the patients presenting with pineal region lesions will undergo surgery of some form though there is some debate on the best initial surgical optionbiopsy (open, endoscopic, or stereotactic) versus resection (endoscopic or open).The pineal gland is a deep-seated structure and attempts to resect tumors have been historically morbid [16].Over time open microsurgical access strategies to the pineal region have been described and refined, with improvements in patient safety.The most commonly adopted open microsurgical approaches include supracerebellar infratentorial (SCIT), occipital transtentorial approaches, posterior transcortical, and posterior interhemispheric (PIH) transcallosal approaches (Table 1) [17].The armament of surgical options of pineal region tumors was further expanded by the adoption of endoscopic ventricular access, which allows for simultaneous third ventriculocisternostomy to address hydrocephalus [18].Several recent studies have examined the use and effect of specific surgical techniques or algorithms.Abecassis et al. studied their own experience with 50 pineal region tumors over 1 decade and found no significant differences in extent of resection or outcomes among different surgical approaches employed [33].Schulz et al. [20] recently evaluated 25 patients with pineal region tumors without histological diagnosis; only eight of whom needed surgical resection after definitive diagnosis via an initial biopsy and recommended primary biopsy for these lesions prior to surgical resection.However, no study has examined changes that may have occurred in surgical management of these patients over the past 2 decades and how those changes have affected outcomes.Here, we describe our institutional experience and examine the trends with surgical treatment of pediatric pineal tumors in 43 patients over 17 years at Children's National Hospital.

Materials and Methods
Information on patients with pineal region lesions treated at Children's National Hospital (CNH) between 2005 and 2021 was retrospectively reviewed from a prospectively collected registry.Exclusion criteria were children who did not receive any surgical intervention at CNH. IRB approval was obtained (IRB: STUDY00000009), and consent was waived due to minimal risk to patients included.
Electronic patient charts were then reviewed for each of these patients and the following information was collected: date of surgery, age at surgery, gender, presenting symptoms, serum/CSF β-HCG and AFP markers, pathology (both permanent and frozen), presence of hydrocephalus at presentation, procedure(s) performed, extent of resection if resection was attempted as reported by surgeon (biopsy, subtotal, gross total), performance of ETV, placement of a CSF shunt, re-operation, placement of an external ventricular drain at surgery, pediatric intensive care unit length of stay, any additional surgeries performed, complications (infection, hemorrhage, new neurologic deficit, or need for reoperation), disposition (home, rehabilitation facility, or other facility), adjuvant chemotherapy or radiation, recurrence (defined as growth of primary disease or occurrence or growth of metastatic disease), length of follow-up, and mortality.Unpaired Student's t and one-way ANOVA tests were used to analyze parametric and non-parametric continuous data, respectively.Categorical data were compared using Pearson χ 2 tests.All statistical analyses were done using PRISM (version 9.2.0,GraphPad Software, San Diego, CA, USA).A p value of less than 0.05 was used to determine significance.

Patient Demographics and Presentation
A total of 43 pediatric patients with pineal region lesions were surgically treated at CNH between January 2005 and December 2021 (Table 2).The average age at treatment was 10 years (range: 5 months to 22 years).There were 22 males (51.3%) and 21 female (48.7%) patients.Thirty-four (79%) patients presented without having undergone any prior surgical intervention.Prior CSF diversion procedures were completed at outside hospitals for 8 patients, including 3 endoscopic third • Great exposure for infratentorial extending tumors • Small posterior fossa in children may limit surgical exposure • Modest surgical window for anterior extending tumors, deep venous structures may limit anterior extent • Midline trajectory • Less effective if tumor above venous system or it extends into lateral ventricles • Operative angle may be steep depending on angle of the tentorium • Often good visualization of deep venous system • Best visualization if gravity used for natural cerebellar retraction (i.e., semi-sitting, Concorde position) • Bridging veins between cerebellum and tentorium need to be sacrificed

Evolution of Surgical Management of Pineal Tumors
ventriculostomies (ETV), 3 shunts, and 2 patients who underwent both ETV and shunt placement before establishing care at CNH. Prior biopsies had been completed for 5 patients.More than half of the patients (n = 24, 55.8%) had headaches as one of the primary symptoms upon presentation.Visual or ocular disturbances (including diplopia and papilledema) were noted in 18 of the patients (41.9%), while 18 also presented with nausea or vomiting.Gait instability (n = 6, 14%) and precocious puberty (n = 1, 2.3%) were also noted.Interestingly, 2 of the patients were asymptomatic from their pineal lesions but received CT scans after sustaining trauma to the head and the lesions were discovered incidentally.Hydrocephalus was present in 38 (88.4%) of patients (irrespective of prior CSF diversion).There were no serum or CSF markers tested for 8 of the patients, and seven (16.2%) showed positive serum CSF AFP or β-HCG.

Surgical Approach
A chart of the sequential surgical procedures in the case review is depicted in Figure 2. A biopsy was the first surgical procedure for most patients (n = 27; 62.8%), while 37.2% of patients underwent surgical resection as a first procedure.Twenty patients underwent endoscopic biopsy, and 19 of these patients underwent concurrent ETV, while 1 received a concurrent shunt.Stereotactic biopsy was completed in 2 patients.The remaining 5 initial biopsies were conducted via alternate approaches (lumbar laminectomy, SCIT approach via suboccipital craniotomy, PIH approach, and a subfrontal craniotomy; often pursuing a more easily accessible site of disseminated disease.Diagnostic yield for initial biopsy procedures was 92.6%.One stereotactic biopsy was nondiagnostic and a PIH approach was completed later for biopsy in that patient.One endoscopic biopsy attempt failed due to poor visualization, and this patient later underwent a subtotal resection via PIH approach. Overall, patients who underwent biopsy as their first surgical procedure were less likely to have undergone a prior surgical procedure at another hospital than patients whose first procedure was not a biopsy (2 of 27 patients or 7.4% in the biopsy first group who underwent prior procedures vs. 6 of 16 patients or 37.5% in the open resection or alternate procedure group who underwent prior procedures, χ 2 = 0.014).Patients who underwent biopsy first were also less likely to ultimately experience a complication during their treatment course compared to patients who underwent a resection or alternative procedure (2 of 27 patients or 7.4% vs. 7 of 16 patients or 43.8%, χ 2 = 0.0046).There was no difference between patients who underwent biopsy at first surgical procedure compared to another procedure in patient age (10.3 years vs. 8.8 years, p = 0.38), patient mortality (11.1% vs. 12.5%, χ 2 = 0.89), or serum/CSF marker positivity (18.5% vs. 12.5%, χ 2 = 0.61).
Of the patients whose first procedure was a biopsy, 12 of the 27 (44.4%)went on to undergo a surgery for resection (4 pineoblastomas, 3 pineal parenchymal tumors, 2 germinomas, 2 non-germinomatous germ cell tumors, and 1 glial tumor).For 10 of these patients, resection occurred within 3 months of the initial biopsy; 1 patient with germinoma underwent resection 18 months later, and 1 patient with disseminated astrocytoma underwent resection 49 months later.The pathology for the 15 patients who underwent initial biopsy but did not undergo further resection consisted of 8 germinomas, 2 glial tumors, 2 pineoblastomas, 2 pineal parenchymal tumors, and 1 ganglioglioma.The series included 32 patients who underwent an open approach at some point; 26 of these 6% no resection, χ 2 = 0.41).However, patients who underwent resections were significantly more likely to experience a complication (9 of 26 or 34.6% resection vs. 0 of 17 or 0.0% no resection, χ 2 = 0.0089).

Treatment of Hydrocephalus
ETV was employed to treat hydrocephalus in 26 of the 38 patients (68.4%), irrespective of whether the procedure was performed at CNH or another insti-tution, while 21 of the 38 (52.6%) patients eventually obtained a ventriculoperitoneal shunt (VPS).10 of the 21 VPSs that were placed were done as initial treatment for the hydrocephalus and no ETV was attempted in these patients.The use of VPS as the initial treatment for hydrocephalus declined over the study period.Out of the 10 shunts placed as first-line hydrocephalus therapyfour were placed in the 10 patients (40%) treated between 2005 and 2009 while the other six were placed in the 28 patients treated between 2011 and 2021 (21.4%).When excluding patients who either did not have hydrocephalus at presentation but developed it after surgery for the lesion or those who received a shunt at an outside institution prior to presenting to CNH, this decrease in use of VPS as initial treatment for hydrocephalus becomes even more evidentthree out of 10 (30%) between 2005 and 2009 versus two out of 28 (7.1%, p = 0.04) between 2011 and 2021.Eleven of the 26 (42.3%)ETVs performed for treatment of the hydrocephalus were eventually converted to a VPS.Prior to 2010, all 5 ETVs performed had to be eventually converted to a VPS, while only 6 of the 21 ETVs (28.6%) performed between 2011 and 2021 had to be converted to a VPS (p = 0.004).5 of the 38 patients (13.2%) with hydrocephalus did not need any treatment and the hydrocephalus resolved after treatment of lesion.

Pathology
Most of the tumors in our series were germinomas (germinomas (n = 13, 30.2%), followed by pineoblastomas (n = 10, 23.2%), glial tumors (n = 6, 14%), and mature teratomas (n = 4, 9.3%).There were also two atypical teratoid rhabdoid tumors and one each of mixed malignant germ cell tumor, papillary tumor, ganglioglioma, benign reactive pineal gland cyst, diffuse midline glioma, dermoid cyst, and a high-grade neuroepithelial tumor (Fig. 3).Intraoperative frozen pathology was sent in 38 patients.Of these 38, 7 (18.4%)patients' final pathology was in disagreement with the frozen pathology.This occurred in 4 cases where the frozen pathology suggested germinoma but final pathology yielded grade II/III pineal parenchymal tumor, papillary tumor of the pineal region, astrocytic neoplasm, and low grade pineocytoma versus neurocytoma.Other disagreements between frozen and final pathology included a frozen result of mixed germ cell tumor that later resulted as pineal parenchymal tumor of indeterminate malignancy, a frozen result of ependymoma tumor that later resulted as high-grade neuroepithelial tumor, and a frozen result of pineocytoma tumor that later resulted as ganglioglioma.The decision to resect versus conduct a biopsy alone varied somewhat with tumor pathology.Patients who had germinomas were less likely to undergo a resection attempt (4/13 or 30.8% of patients with germinomas underwent resection while 21/35 or 60% of patients with non-germinoma tumors underwent attempted resection, χ 2 = 0.044).However, there was no significant difference in likelihood of undergoing a resection surgery for germ cell tumors overall, pineal parenchymal tumors, pineoblastomas, or glial tumors (χ 2 = of 0.36, 0.43, 0.71, and 0.66, respectively).

Postoperative Course
The majority of patients (n = 38, 88.4%) were discharged home post-operatively; the remainder were discharged to a rehabilitation or other facility.The median length of postoperative follow-up was 47 months.Chemotherapy was administered to 33 (76.7%) of patients, and radiation was administered to 31 (72.1%).Recurrence during follow-up occurred in 9 patients (20.9%).The mortality rate across the series was 11.6% (Fig. 4).All but 1 of the 5 patients who died did so after recurrence.Recurrence occurred in 3 patients with pineoblastoma, 3 glial tumors, 1 high grade neuroepithelial tumor, 1 germinoma, and 1 papillary tumor of the pineal region.Mortality occurred in 3 pineoblastoma patients, 1 patient with a germinoma, and 1 patient with an H3K27 mutant diffuse midline glioma.All patients who experienced recurrence or died underwent chemotherapy, radiation, or both.

Complications
Surgical complications, defined as infection, neurological deficit, or pseudomeningocele occurred in 9 of the 43 patients (20.9%).All complications occurred in patients who underwent attempted resection.The majority of these ( 6) were new neurological deficits; 2 patients had postoperative infections and 1 patient had a postoperative pseudomeningocele.New neurological deficits consisted of extraocular movement abnormalities in 3 patients, a visual field cut in 1 patient, a venous infarct causing leftsided weakness in 1 patient, and a fixed pupil with loss of cough/gag in 1 patient.Four patients had complications related to compromise of the internal cerebral veintwo were intentionally sacrificed, and two were inadvertent injuries.The 2 patients who underwent intentional internal cerebral vein sacrifice for operative exposure experienced no clinical complications, although 1 showed T2 signal change suggestive of edema in the corresponding thalamus.One child who underwent PIH also had inadvertent injury to the internal cerebral vein which led to a subtotal resection; this patient had no new postoperative deficits and therefore is not included as a complication.The last patient with an inadvertent internal cerebral vein compromise experienced a venous infarct causing left-sided weakness and is included in the complications above.

Discussion
This report details the presentation and surgical management of a series of 43 children presenting to Children's National Hospital with pineal region tumors between 2005 and 2021.Only a small number of modern pediatric case series of pineal tumors have been published, and several of those published focus on germ cell tumors, making results difficult to compare [10,19].However, analysis of the patients reported here offered several relevant observations regarding presentation, management, and outcomes of these patients, some of which we contextualize below in comparison with other series of pediatric pineal region tumors.
When compared to other series, our cohort reported a lower incidence of headache (55.8% in our cohort vs. Evolution of Surgical Management of Pineal Tumors 76.0%), nausea/vomiting (41.9% vs. 72.0%),and motor deficits (5.0% vs. 20.0%),but a similar incidence of visual abnormalities (41.9% vs. 48.0%)[20].This series was also notable for the relatively equal representation of male and female patients, as compared to series that often show male predominance [21].The 88.4% of patients presenting with hydrocephalus in our series is comparable to the 81-82% observed in other pediatric series [20,22].Interestingly, our series had a relatively high rate of cases in which frozen pathology did not agree with final pathology; this suggests that surgeons should be circumspect before making clinical decisions based on frozen pathology.The majority of patients in our series underwent an initial biopsy, but a relatively large number of patients (37.2%) underwent surgical resection as first procedure, which is higher than reported in some series [20].This may in part due to the number of patients presenting from other centers with established diagnoses.The diagnostic yield of 92.6% reported here is comparable to other series occurring during the same time period, which report diagnostic yields of 95.8% [20].As expected, patients with germinomas were less likely to be brought to the OR for resection [17].The 44% of initial biopsy patients in our series that went on to undergo resection is comparable to the 54.2% post-biopsy resection rate reported in another series [20].Our biopsy/endoscopic complication rate was 0% which is less than or comparable to the 4% permanent complication rate reported in another series of endoscopic biopsies [20].The complications noted were all in open biopsies or resections.One case, in which the internal cerebral vein was inadvertently sacrificed, yielded the most severe new neurologic deficit.A small number of similar internal cerebral venous infarcts have been reported [23], accentuating the need for care in the decision to sacrifice venous structures, especially the internal cerebral vein.
Our series also demonstrates the significant changes that have occurred in the management and outcomes of these patients over the years.Prior to the advent of microneurosurgery in the 1970s and widespread use of CT imaging in the 1970s, pineal region tumors were treated conservatively with placement of VPS and irradiation [6].However, as CT imaging and neurosurgical techniques advanced, attempts were made to classify and distinguish these tumors in terms of histology and response to treatment [26].Over time, open surgery became the gold standard for these patients to obtain a definitive histological diagnosis [26].These open surgical procedures were associated with a high rate of morbidityas high as 10% in certain reports [7].In light of this, neurosurgeons have been pursuing novel approaches to this region.
Fukushima et al. were the first to describe endoscopic treatment of pineal region lesions in 1973 [32].As the technique became refined and better hemostatic techniques were developed, several case reports and short series were published describing the use of endoscopes for simultaneous ventriculostomies and biopsies of tumors [6,8,11,12,16,18].However, it was not until the mid to late 2000s that the technique truly began to gain ground with several series being published examining the effects of combined ETV and endoscopic biopsies in patients with pineal region tumors [3,23].In 2004, Yamini et al. presented 54 patients as part of a clinical series and literature review who underwent simultaneous ETV and endoscopic biopsy.Their results were encouraging demonstrating high rates of accurate diagnosis and low rates of VPS placement in their combined cohorts [34].Our series illustrates that these series clearly had a direct impact on the therapeutic options for these patients leading to a significantly higher rate of ETVs performed for management of hydrocephalus after 2010.Over the timeline of our study, ETV became the preferred technique.26 of the 38 patients with hydrocephalus received an ETV and only 2 patients between 2011 and 2021 had a VPS placed as the primary treatment for hydrocephalus.The rate of VPS placement due to failure of ETV also decreased significantly in the last 10 years from 100% conversion rate prior to 2010 to 28.6% conversion rate in 2011-2021.Rates of biopsies and resection, average length of stay, adjuvant treatment, and other outcomes remained stable throughout the 17 years.ETVs for management of hydrocephalus are not without complication.The ETV Success Score (ETVSS) developed by Kulkarni et al. predicts close to an 80% success rate at 6 months for ETVs performed in children in our cohort (average age of 6 years, tectal tumor, with no previous shunt) [35].More recent studies have shown that this score correlates with long-term success and independence from VPS as well [13].Though initially (2005-2009) our cohort's conversion rate from ETV to VPS was much higher at 100%, it did decrease significantly and was closer to the expected rate of 28.6% in the last 10 years.The decrease in conversion rate was most likely due to the surgeon's experience with the procedure as well as improvement in techniques over the decade.Nevertheless, our ETV failure rate of 28.6% was higher than would be expected and this may be due to the nature of these tumors in addition to the added debris and blood in the ventricular system after biopsy and resection.It is important to note that three of the six ETVs that had to be converted also underwent tumor resections as opposed to simple biopsies.Overall management strategies for these patients are based on either presumptive diagnosis based on imaging characteristics and tumor markers or histopathologic diagnosis if biopsy is performed.Treatment often includes some combination of surgical resection and chemoradiotherapy depending on type of tumor.Over the last few years, all aspects of care for these patients have changed significantly.As surgical techniques continue to improve, it has been clearly demonstrated that for certain tumor types, maximal safe resections should be the goal.For pineoblastomas, specifically, Parikh et al. [27] have shown greater overall survival when patients underwent gross total resection or neartotal resection compared to those who had subtotal resections or biopsies alone.Similar studies are needed for other tumor types occurring in this region.The three major surgical approaches used for these tumors are PIH, infratentorial supracerebellar, and occipital transtentorial.Patient-specific anatomy, specific tumor position and extension and surgeon's familiarity and comfort allow for choosing one over another (Table 1).Though advances in neuronavigation and microsurgery have made these approaches safer, use of endoscopes to perform these could potentially further improve outcomes [31].
To properly assign patients to appropriate treatment tracks of chemoradiotherapy, primary surgical resection, or combination thereof, precise and accurate diagnosis is imperative.The 2021 WHO classification of CNS tumors now includes subtypes of pineoblastomas and pineal parenchymal tumors of intermediate differentiation based on specific recent advances in molecular understanding of these tumors [28].These subtypes aid in precise prognostication for these patients.It also allows for risk stratification of these patients based on molecular aberrations and targeted therapy [29].For example, patients with pineoblastomas and MYC amplification and FOXR2 overexpression, potential treatment with BETbromodomain inhibition is being considered, while in those with pineoblastomas and RB1 alteration, treatment with nortriptyline is under investigation [29].When considering chemoradiotherapy, germinomas continue to remain highly sensitive to radiation.Localized germinomas may be treated with reduced dose radiation localized to region of interest, while disseminated germinomas continue to be treated with craniospinal irradiation.However, it is important to note that the recommended dose of radiotherapy has been reduced [30].For the other tumor types, com-bination of neoadjuvant chemotherapy, a "secondlook" surgery for residual after neoadjuvant therapy, and radiation continues to be the standard of care [29].Studies are underway for different combinations of neoadjuvant therapy and potential reduced dose of radiation based on prognostication based on molecular pathology [29].
It is important to note that despite the advances noted in some studies described above, the rates of biopsy, resection, adjuvant therapy, and outcomes did not change through the years in our cohort.This may be due to a number of factors, including the small number of patients included here.Additionally, all of the patients included in our series received surgical intervention for biopsy or resection for definitive diagnosis of these tumors.This may be higher than seen in literature or other clinical centers as we only included patients who presented for neurosurgical attention.Patients who had positive markers and had no need for neurosurgical intervention were excluded and so our cohort may not capture patients who did not require definitive tissue diagnosis.Though chemoradiotherapy is becoming increasingly more effective, surgical resection is still recommended in cases of pineal parenchymal tumors, teratomas, and most glial tumors [17].As tests for markers become more sensitive and adjuvant therapies are enhanced, we may expect the need for biopsies and microsurgical resections of these tumors to decline in the future.
Limitations to our study include the retrospective nature of the study and the small number of patients.There has also been an increased tendency to diagnose some of these lesions based on serum and CSF markers without requiring a tissue diagnosis.If these patients did not require a biopsy and did not have symptomatic hydrocephalus, they may not have presented to our neurosurgical service and thus not captured in this analysis.

Conclusion
This series describes surgical approaches and outcomes in pediatric patients with pineal region tumors over a 17-year-period at our institution.The most significant changes in treatment have occurred in the preferential use of ETVs over VPS for associated hydrocephalus.Complication rates were higher with open resection, reinforcing the safety of pursuing endoscopic biopsy as an initial approach.Finally, in our limited experience, outcomes of these patients did not significantly change over that time.

Fig. 3 .
Fig. 3. Pie-chart depicting the composition of different tumor pathologies in our cohort.

Fig. 4 .
Fig. 4. Kaplan-Meier survival curve for all patients.Vertical lines indicate patients censored due to end of followup period.

Table 1 .
Four most common surgical approaches to pineal region • Minimally invasive• Limited exposure that requires extension of tumor into 3rd ventricle • May offer opportunity to address hydrocephalus via ETV • Difficult to remove meaningful volume of tumor • Best for biopsy • Bleeding may obscure visualization PIH transcallosal • May be shortest path to anterior 3rd ventricle lesions • Presence of bridging veins, need bilateral sinus exposure • Direct path to tumor between internal cerebral veins • Difficult to reach infratentorial portions of tumor Occipital transtentorial • Wide surgical field of view • Deep venous structures (e.g., vein of Galen, etc.) • Direct visualization of venous structures (e.g., vein of Galen) • Oblique trajectory can be disorienting • Avoid bridging veins • Need to cut tentorium Supracerebellar infratentorial