Giant Supratentorial Brain Tumors in Children: Functional Outcome and Progression-Free Survival Analysis

Central nervous system (CNS) tumors are the most common solid tumors in children [1]. Giant supratentorial brain tumors (GSBTs) are a rare type of brain tumor with lesions larger than 50 mm [2]. These tumors have high morbidity and mortality rates regardless of their histological grade [3], resection size, or adjuvant treatment [4].

Two previous articles have discussed the management of GSBT in pediatric patients; however, both articles based their results on the surgical resection’s extent [2, 3]. These tumors are so complex that measuring surgical success based exclusively on the extent of the resection seems an oversimplification.

In light of the above, this study aimed to identify factors affecting progression-free survival (PFS) in pediatric patients with GSBT treated with surgical excision. The secondary aim was to analyze how these factors affected the functional outcome in the long term.

We performed a retrospective, analytical, single-center cohort study. We reviewed all pediatric patients with a magnetic resonance imaging (MRI)-based diagnosis of GSBT at Pediatric Hospital Prof. Dr. Juan P. Garrahan, Buenos Aires, Argentina, between January 2014 and June 2018. Our Institutional Review Board and the Local Ethics Committee approved this study.

We included all patients younger than 18 years with GSBT, defined according to the medical literature as a tumor larger than 50 mm, who underwent surgical resection as the primary treatment and had complete follow-up at our hospital. We excluded patients with vascular or granulomatous lesions, with incomplete data or less than 24 months of follow-up, and GSBT where surgery was not the primary treatment, like midbrain gliomas extending to the pineal region, optic pathway gliomas, cystic craniopharyngiomas, germinomas, and pineal region lesions.

The aim of the surgery was extensive resection without causing significant damage to the surrounding brain. The resection area usually encases the brain region containing the tumor along with a rim of non-eloquent cortex and white matter, if possible. A large craniotomy was done in all cases to avoid brain ischemia because of postoperative (post-op) edema and raised intracranial pressure. Neuronavigation and intraoperative (IOP) ultrasound were used to locate the limits of the lesion and the main vessels. A microscope was used in all cases for at least part of the surgery. Suction with an ultrasonic aspirator was used whenever possible to maximize the degree of resection and avoid injuring surrounding vessels. The decision to use adjuvant chemotherapy and/or radiotherapy was made according to pediatric protocols for treating CNS tumors in infants.

Complete full blood count and clotting were done before surgery, along with a group and safe and two or more units of blood for the surgery. All patients were admitted on the day of the surgery. Proper hair trimming, and standard prep and drape were done. Antibiotics were given 1 h before the skin incision and repeated every 4 h.

Patients were extubated when possible after surgery and transferred to the PICU, where they remained for a minimum of 5 days. A total blood count was taken 24 h after the surgery, and an MRI was done in the first 72 h after the surgery. Following PICU, the patient was transferred to the ward, where they remained for a minimum of 5 days before discharge.

Data were collected on gender, age, clinical presentation, and duration of symptoms before the diagnosis. Preoperative (pre-op) brain MRIs were analyzed by tumor volume and the lesion’s primary location and were clustered into six areas: frontal, temporal, occipital, or parietal lobe, III ventricle, or lateral ventricles.

Each patient’s functional status score (FSS) [5] was recorded before the surgery, 6 months after, and yearly after that. The FSS is a functional status outcome scale validated for children. Domains of functioning include mental status, sensory, communication, motor, feeding, and respiratory, classified from average (1) to very severe dysfunction (5). The final score goes from 6 to 30, categorized as good outcome ≤ 7, moderate outcome of 8–15, and poor outcome ≥ 16.

We recorded the number of surgical procedures per patient, the extent of resection (partial resection: <79%, debulking: 80–94%, and gross total resection [GTR]: >95%), and the adjuvant treatment. The histopathological diagnosis was recorded and later subcategorized into high grade (grade III/IV) and low grade (grade I/II) based on the 2016 World Health Organization Classification of Tumors of the Central Nervous System [6].

Post-op complications within 30 days after surgery included CNS infections (Is), hydrocephalus (H), cerebral abscesses, subdural collections (SCs), intracranial bleeding requiring surgical intervention, and seizures (SZs). CNS I was established in the event of a positive cerebrospinal fluid culture and clinical symptoms suggesting I (fever, increased white blood cell count, erythema, and purulent drainage from the wound). H was recorded when ventricular dilatation on a brain image was accompanied by clinical symptoms of raised intracranial pressure (including headache, vomiting, nausea, sensory impairment, bulging fontanelle, and VI cranial nerve palsy). Rim-like lesions near the surgical site showing gadolinium enhancement and restriction in diffusion were recorded as brain abscesses. SCs were defined as increased subdural space greater than 0.5 cm on MRI or computer tomography. We also recorded clinician-evidenced SZs lasting more than 30 s.

Patients were followed for a minimum of 24 months for the PFS and overall survival (OS) analysis. PFS was defined as a lack of disease progression, metastasis, or death from any cause and was calculated from the date of the last surgery to the date of progression, metastasis, death, or last contact. OS was recorded from the last surgery date to the last contact or death event. Overall clinical follow-up was defined as the period from diagnosis to last clinical assessment or death.

We also categorized the patients into disease-free, regrowth or recurrence, minimal residual stable lesion, or death. Finally, we used the last post-op FSS recorded in the charts to analyze the clinical, functional outcome. The secondary outcome was the association between the post-op FSS (continuous variable) and the independent variables: age, pre-op FSS, number of surgeries, tumor volume, type of resection, tumor location, histopathological diagnosis, IOP bleeding, and different complications.

Since the sample was considered non-normally distributed due to the number of patients, categorical variables were presented as absolute frequency, and continuous variables were presented as the median and interquartile range (IQR). The primary outcome variable used for the analysis was PFS at 24 months. A Kaplan-Meier method was used to estimate the PFS curves. For the bivariate analysis, we used the log-rank test. The FSS scale and tumor volume variables were categorized for the survival analysis to perform the log-rank test; for the multivariate analysis, we used a Cox proportional hazards method, in which we built a model which included all the variables that, in the bivariate analysis, presented a value of p < 0.2. In addition, hazard ratios (HRs) with their respective confidence intervals (CIs) were presented, and p values <0.05 were considered statistically significant.

After checking for the assumptions, a multivariate analysis with linear regression was used for the secondary outcome. The stepwise forward selection was used, adding terms with p < 0.1 and removing those with p ≥ 0.2. Coefficients with their respective CIs were presented, and p values <0.05 were considered statistically significant. We used STATA/IC 15.1 (4905 Lakeway Dr. College Station, TX 77845, USA) for statistical analysis.

We initially identified 32 patients diagnosed with a GSBT in our database between January 2014 and June 2018, but we analyzed 27. We excluded five patients with non-neoplastic lesions: two corresponded to granulomatous processes and three to vascular lesions – cavernomas and a giant dural arteriovenous fistula.

The median age was six (range 2–12), and 17/27 were male. The most frequent clinical presentation was raised intracranial pressure in thirteen cases, followed by a new onset of paralysis/paresis in six cases and SZs in four. Other less frequent symptoms were behavioral changes, palpable skull mass, ptosis, and sinusitis, each present in one case. The median time from symptom onset to hospital arrival was 4 weeks (range 1–8). Table 1 summarizes the data for the 27 patients.

The most frequent tumor location was the frontal lobe in 11 cases, followed by the parietal lobe in eight cases, the temporal lobe in two cases, the occipital lobe in one case, lateral ventricles in 3 cases, and the third ventricle in 2 cases. The median tumor volume was 117 cm³ (IQR 87–237 cm³), with a range of 57–465 cm³.

The median pre-op FFS value was 7 (6–8). Seventeen patients had an FFS of ≤ 7, and 10 patients had an FSS between 8 and 15. None of the patients presented an FSS score >16 in the pre-op evaluation.

GTR was achieved in 20/27 patients, subtotal in 6/27, and partial resection in 1/27. One surgery was needed in 18 cases, two surgeries in six, and three surgeries in 3 patients.

Profuse IOP bleeding occurred during tumor excision in 7/27 patients. Each case was managed with IOP fluids and blood transfusions without interrupting the surgery.

The histopathological results were highly variable. The most frequent type of tumor found was anaplastic ependymoma (grade III) on 5/27 and atypical teratoid rhabdoid tumor (grade IV) on 5/27. Atypical choroid plexus papilloma (CPP) (grade II), pilocytic astrocytoma (grade I), and glioblastoma IDH mutant (grade IV) were found in 3/27 cases each. An embryonal tumor (C19MC-altered) (grade IV) was found in 2/27 cases. To conclude, choroid plexus carcinoma (grade III), CPP (grade I), fibrous meningioma (grade I), ependymoma (RELA-fusion positive) (grade III), diffuse midline glioma (H3 K27M-mutant) (grade IV), anaplastic ganglioglioma (grade III) were present in 1/27 patients each. Nineteen cases were high-grade tumors (grade III: 8/19, grade IV: 11/19), whereas eight were low grade (grade I: 5/8, grade II: 3/8). Fifteen patients required adjuvant chemotherapy, and thirteen required radiotherapy.

The most frequent post-op complication was H in 9 patients – all treated with a ventriculoperitoneal shunt – followed by CNS I and SCs, each occurring in 6 patients. Two patients presented with SZs, brain abscesses, and intracranial bleeding.

The 24-month PFS was 51.85%, and the OS was 74.04%. A PFS-ending event or treatment failure occurred in 13/27 patients. Relapse or regrowth was reported in 6/13, while death was reported in 7/13 patients. The median time to relapse or regrowth was 3.5 months (IQR 2–10), while the median time to death was 8 months (IQR 4–19), with 2 patients surviving less than 4 months after surgery.

Of the remaining 14 patients who were progression-free after 24 months, 11 were disease-free after a median of 27 months (IQR 25–48), and 3 patients had a minimal residual stable tumor after a median of 27 months (26–39). Table 2 shows the Cox proportional hazards analysis for the variables that on the bivariate analysis presented a value of p < 0.2. Patients with post-op FFS >16 had a worse PFS compared to patients with a post-op FSS <15 (HR 4.96; 95% CI: 1.30–18.94; p = 0.01) (shown in Fig. 1a). Patients with more than three surgeries had worse PFS than patients with one or two surgeries (HR 11.06; 95% CI: 2.23–60.01; p = 0 0.003) (shown in Fig. 1b). Regarding the histopathological results, grade IV tumors were associated with worse PFS than grade I, II, and III tumors (HR 3.24; 95% CI: 1.03–10.16; p = 0.04) (shown in Fig. 1c). Finally, patients with CNS Is had worse PFS than patients without an I (HR 4.08; 95% CI: 1.29–12.85; p = 0.01) (shown in Fig. 1d). There were no significant differences between the size of the resection (gross total, debulking, or partial) and PFS at 24 months (HR 1.09; 95% CI: 0.35–3.33; p = 0.87). Figure 1 shows the 24-month PFS graphs for the various subgroup analyses. Neither of the other variables showed any association with the PFS at 24 months.

The median post-op FSS value was 7 (IQR 6–9). 16 patients had an FSS <7, 8 patients had an FSS between 8 and 15, and three patients had an FSS >16. We found an association between some of the variables analyzed in the model and the post-op FSS. This model explained 82% of the variance (R2 = 0.82, F (11,15) = 6.38, p < 0.0007).

After adjusting for the different variables, four factors were found to increase the post-op FSS. Being a male increased the post-op FSS scale by 3.7 points (Coef: 3.79; 95% CI: 1.09–6.47; p = 0.009), high-grade lesions by 4.34 points (Coef: 4.34; 95% CI: 2.30–6.38; p ≤ 0.001), H by 5.79 points (Coef: 5.79; 95% CI: 3.58–7.99; p ≤ 0.001), and intracranial bleeding after surgery by 4.5 points (Coef: 4.52; 95% CI: 0.67–8-46; p = 0.02). By contrast, radiotherapy (Coef: −4.19; 95% CI: −6.22 to −2.16; p = 0.001) and chemotherapy (Coef: −1.89; 95% CI: −3.07 to −0.09; p ≤ 0.05) reduced the post-op FSS by 4.19 and 1.89 points, respectively.

Over the last few years, concerns have been raised about how long-term results are assessed in pediatric neuro-oncology patients [7]. It is common to find publications in which the outcome is measured primarily by surgical resection; however, this approach overlooks the patient’s functional outcome and quality of life. An appropriate evaluation of the success of the treatment should therefore include a complete analysis of surgical and functional outcomes with long-term follow-up [8].

GSBT in pediatric patients has been reported by Guo et al. [2], who analyzed 14 cases, and Oliveira et al. [9], who analyzed 23 cases. However, both publications measured their outcomes mainly on the extent of the surgical resection without an objective functional assessment.

In this article, the authors presented a cohort series of 27 GSBT with a minimum follow-up of 24 months, where the functional outcome was measured by an objective functional scale – the FSS – and the oncological outcome was measured by PFS and OS. Different from what Guo et al. reported, no association was found between the magnitude of the excision and the PFS at 24 months, even though 74% of the patient had a GTR – Figure 2 shows a patient diagnosed with anaplastic ependymoma, treated with GTR and adjuvant therapy, who presented regrowth of the lesion less than 2 months after surgery.

As expected [3, 10, 11], histological classification played a crucial role in PFS. The histological variability of GSBT made it challenging to assess each tumor independently, but high-grade tumors showed a significant reduction in the PFS at 24 months. Less expected was the association between clinical variables, like post-op complications or functional outcomes, and PFS. For example, patients with an FSS >16 after surgery significantly decreased the PFS, as well as patients with more than three surgeries or patients with a post-op I (shown in Fig. 1).

We reported 14/27 patients free of disease progression at the end of the follow-up (median time of 27 months). Of these, 11 were disease-free (shown in Fig. 3, 4), and three had a stable residual tumor. Not all disease-free patients underwent GTR of the lesion, as 3 cases were partial excisions which, with adjuvant treatment, achieved complete regression of the lesion. Considering that two-thirds of the tumors analyzed were WHO grade III or IV, a PFS of 51.85% and an OS of 74.04% represent a favorable outcome since it was also accompanied by survival with an acceptable post-op FSS (median post-op FSS score of 7 with IQR 6–9).

We also identified risk factors associated with worse functional outcomes – worse FSS results. Some of those, such as WHO high-grade lesions and patient gender, cannot be changed. However, others can be minimized by appropriate and timely intervention, such as H, bleeding, and I. Complications like H should be treated as soon as possible. In some cases, the resection of the lesion is enough to treat the problem, as in the case of CPP; nevertheless, we need to remember that the brain parenchyma of these patients suffers from chronic hypoxia [12, 13], so minimal changes in the hydrodynamics of the cerebral spinal fluid could cause massive consequences. Correct hemostasis after surgery is an essential step in any surgery, but even more when operating lesions of this size. The surgeon must take sufficient time to control any minor bleeding before closure, especially considering the extended new space left by removing the lesion.

The functional outcome appears to impact the patient’s OS significantly. In total, 7 patients died within 2 years of their last surgery but none of them during the surgical procedure; however, 2 patients with atypical teratoid rhabdoid tumors experienced post-op complications and died within 4 months after the surgery. Although post-op complications have not been directly identified as contributing to the death, we can assume they influenced the outcome since these patients had the worst post-op FSS score.

GSBTs in children represent a challenge due to histopathological variability, surgical and anesthetic difficulties, and overall complications, which increase morbidity and mortality [14]. Due to their mass effect, raised intracranial pressure is usually the onset symptom. Thus, surgery plays an essential role in the initial treatment. The prognosis is based mainly on histology, genomics, and resection. However, we cannot apply the same principles we use for “normal-size” brain tumors to GSBT because lesion’s size is an independent variable that modifies the overall outcome. To get a complete resection, we need aggressive surgery and wide-field radiotherapy that cause massive sequelae affecting the patient’s OS. So, the size of the lesion overpowers the histological diagnosis and the oncological prognosis.

With the development of multidisciplinary teams, neuro-oncology care has changed toward shared care between neuro-oncology and neurosurgery teams [15]. Some specialists continue to support aggressive treatment for these lesions, wherein the end seems only to prolong the suffering of the patient and the family [16, 17]. When the clinical course gets complicated by severe sequelae, the OS is poor. We could ask ourselves if some of these children should not receive any treatment at all. From the moment the lesion is diagnosed, it is important to consider offering the patient the best treatment, taking into account not only the type of excision or the possibility of adjuvant treatment but also the functional status since apparently influencing both mortality and disease-free survival.

The main limitation of this study relies on the variability of the pathology, which reduces the sample size for statistical analysis. A decision to group the patients into high- and low-grade tumors based on the WHO classification was done to reduce the heterogeneity of the sample. In addition, there are limitations in the validity of the data due to the retrospective collection. However, the data were extracted from electronic medical records, which were prospectively evaluated in all patients with GSBT. Misclassification bias in image review was limited by controlling the tumor diameter with two independent researchers. Selection bias was eliminated by including all patients with GSBT diagnosed and treated at our institution with a minimum follow-up of 24 months. Patients from a vast referral network with an ethnically diverse background were included, with the functional outcomes and surgical outcome measurement, making the results externally generalizable.

GSBT in pediatric patients is a complex entity that requires multidisciplinary management. Surgical management along with survival quality of life should be considered when choosing the best treatment. Factors influencing long-term PFS were high-grade histopathology, the need for three or more surgeries, post-op FSS >16, and CNS Is.

Ethical approval was waived by the local Research Committee at Pediatric Hospital Prof. Dr. Juan P. Garrahan, given the study's retrospective nature and that all the procedures being performed were part of the routine care. The study was conducted following the Declaration of Helsinki. All the images and information presented in this article are anonymized, and the submission does not include images that may identify the person. Written informed consent was obtained from the participants' parent/legal guardian/next of kin to participate in the study.

The authors have no relevant financial or nonfinancial interests to disclose.

The authors did not receive support from any organization for the submitted work.

Amparo Saenz was in charge of conception and design, data analysis and interpretation, and drafting of the work. Yamila Basilotta was in charge of data acquisition. Emma Dalton was in charge of reviewing the manuscript for clarity. Romina Argañaraz was in charge of conception and design and revising it critically for important intellectual content. Beatriz Mantese was in charge of revising it critically for important intellectual content and final approval of the version to be published.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.