A 64-year-old female presented with headaches, progressive short-term memory deficits, and some degree of unbalance without other physical dysfunction. Figure 2 illustrates the pre and postoperative imaging. Brain MRI revealed a single highly contrast-enhanced 2x2x2cm lesion in the pineal region lying over the tectal plate associated with some degree of ventricular enlargement but minimal periventricular lucency. Serum hormonal tests for the pineal tumor diagnosis became negative, and the patient was prepared for direct microsurgical resection.
The patient underwent a right paramedian supracerebellar infratentorial approach in a sitting position with complete tumor removal and opening of the posterior third ventricle. During the immediate postoperative course, the patient did not develop new deficits. The histologic diagnosis was a World Health Organization (WHO) grade IV GBM. The patient received two cycles of temozolomide 150 mg/m2 and fractionated radiation therapy of 36 Gy distributed in daily doses of 3 Gy. However, she developed a Pneumocystis infection and Eaton-Lambert syndrome. Thus, radiochemotherapy was discontinued. The patient´s last evaluation was performed 10 months after surgery. MRI studies did not show any recurrence in the pineal region. However, multiple tumor foci with significant perilesional edema in the right cerebellar parenchyma and the right incisural space resulted in obstructive hydrocephalus. The patient died ten months after surgery due to disease progression.
Patients were retrieved from 43 studies published between 1977 and 2021 (see data in brief-repository data) [1–43]. Moreover, three patients were retrieved from references of one published study  after full-text evaluation and diagnosis confirmation [44–46]. Other three patients were retrieved from other sources after diagnosis confirmation [47–49]. Individual participant data was collected to organize the study characteristics table. The original language of publications included English for all of them.
Patient related variables
We report 72 pineal region GBM patients (42% female) with an average age of 39 ± 19.2 (1 – 74) years. A summary of the findings is provided in Tables 1, 2, and 3. The clinical presentation of the patients did not differ from other pineal region lesions. However, 86% of the pineal GBM patients harbored hydrocephalus.
Among the 72 pineal GBM cases, four were reported as WHO grade IV Gliosarcomas [4, 38, 46]. At initial presentation, 22 (31%) patients did not have tumor imaging reports. 11/50 patients (22%) had intracranial dissemination [1, 3, 11, 12, 22, 24, 43], and only one case reported spinal dissemination . The few cases with CT scan reports did not describe calcifications. MRI studies mostly reported an unspecific lesion with a mean diameter of 29 ± 11 (6-50) mm, irregular borders, and heterogeneous enhancement without cystic components. Tumors with and without hydrocephalus had average diameters of 33.4mm and 26mm, respectively. No statistical difference was found due to the reduced number of cases with reported dimensions. Hydrocephalus management included direct removal of the tumor (10.6%), endoscopic third ventriculostomy (38.3%), and ventriculoperitoneal shunt (51.1%).
Immunogenetic reports were very limited to draw appropriate conclusions since only 30 (42%) patients detailed genetic profile information (Table 4) [1, 2, 6, 11–13, 15, 22, 24, 26, 29, 32, 35, 37, 38, 43, 46, 47]. Among 30 patients, H3K27 gene mutation was studied in 15 patients. IDH1 mutation in 15 patients, and Ki67 in 18 patients. Other common glioma genes mutations such as the PTEN mutation, ATRX lost, 1p/19q co-deletion, MGMTp methylation, EGFRvIII mutation underwent evaluation in 10 or a smaller number of patients. The few cases prevented to find of correlations between the genetic profile and disease progression. The survival analysis of the patients with the genetic profile is described in the section “predictors of overall survival”.
We did not find any association under the univariate and multivariate analysis as well. Thus, the H3K27M mutation did not correlate with any clinical or surgical variable. However, none of the seven H3K27M mutant patients followed GTR while 4/8 (50%) H3K27 wild type followed GTR (Fisher test p:0.077).
Surgical intervention and complementary therapy
Biopsy in 33% of this series, usually associated with radiochemotherapy, represented the most frequent treatment modality. Only 19% of the patients underwent gross total resection. Moreover, 89% and 76% of the patients underwent radiotherapy and chemotherapy, respectively. 75% of the patients underwent both radiochemotherapy and radiotherapy. Radiotherapy protocols included in most cases external boost radiotherapy of 60Gy distributed in daily doses of 2Gy. Chemotherapy protocols included Temozolomide in different schemes. Moreover, vincristine, cisplatin, optune, nivolumab, dichloroacetate, KPT-330, Avastin, bevacizumab, nimustine, valproic acid, thioguanine, lomustine, dibromodulcitol, ifosfamide, and etoposide were also reported.
Follow up and outcome
Table 1 summarizes the survival and radiological outcomes. The average follow-up (FU) of pineal region GBM patients was 12 ± 10 (0 – 42) months. 30% of the patients were alive at the last FU. The MOS of this series was 12 months. Information regarding the functional status of the patients at the last clinical evaluation was limited. At the last MRI evaluation of the patients, 23 (32%) patients did not report postoperative imaging. 13/49 (27%) patients did not present disease progression, and four of them were tumor-free. Among the rest of patients, 24/49 (49%) had local progression, 12/49 (25%) had intracranial metastasis, although 6/11 (55%) patients with intracranial metastasis at initial presentation did not report postoperative imaging. 7/49 (14%) patients presented spinal metastasis. All seven spinal metastases presented without pineal tumor progression but with intracranial metastasis in five cases. Similarly, 10/12 (83%) patients with regional intracranial metastasis presented without the primary pineal tumor progression. Thus, 12/14 (86%) patients with regional or spinal metastasis at the last FU did not associate primary pineal tumor progression. 4/49 (8.2%) patients had postoperative complications characterized by postoperative hemorrhage and infarctions.
Table 2 compares the differences among the primary surgical management of the patients. Metastatic disease at initial presentation did not have any association with the type of surgical treatment (Biopsy versus surgical removal) (Fisher exact test, p = 0.7). Regional and spinal metastases at the last MRI evaluation was similar among the different surgical treatment. However, patients undergoing biopsy and complementary therapy had reduced local progression of the disease compared to those following a more invasive surgical removal of the tumor (Fisher exact test, p = 0.021). Under univariate analysis, no other variables had a significant association with the progression of the disease at the last evaluation. The mortality of patients following biopsy was significantly inferior compared with patients undergoing surgical removal. The MOS of patients following biopsy was 19 months, while the MOS of patients following surgical removal was 12 months (Long Rank test, p = 0.018). Table 3 details the radiochemotherapy delivery. The mortality of patients following radiochemotherapy was significantly inferior compared with patients undergoing surgery alone. The MOS of patients following radiochemotherapy was 18 months, while the MOS of patients following surgery alone was six months (Long Rank test, p = 0.002). Moreover, the mortality of patients following biopsy plus radiochemotherapy (MOS of 19 months) was significantly inferior to patients following surgical removal without radiochemotherapy (MOS of 10 months) (Long Rank test, p = 0.010).
Predictors of overall survival
After univariate Cox survival analysis, predictors of better overall survival included biopsy procedures, radiotherapy, chemotherapy, and the association of radiotherapy plus chemotherapy (Table 5, Figure 3). After Cox survival multivariate regression analysis, independent variables associated with better survival rates included biopsy procedures and the association of radiotherapy plus chemotherapy (Table 5, Figure 4). It is worth noting that after multivariate regression analysis, radiotherapy (OR: 2.47, 95% CI: 0.78 – 7.86, p: 0.13) and chemotherapy (OR: 2.87, 95% CI: 0.75 – 10.97, p: 0.12) separately did not show independent association with better survival of the patients.
Due to the limited number of cases, we did not find differences in the Cox survival model evaluation among the 30 patients with different genetic profiles. Thus, the MOS of patients with H3K27M mutant was 23 months, and the MOS of patients with H3K27 wild type was ten months (Long Rank test, p = 0.3). However, it is worth noting that 5/7 (71%) patients with H3K27M mutant kept alive along 12±6 (7-23) months compared with the 88% (7/8) mortality of the H3K27 wild type patients after 13±7 (2-24) months (Fisher p = 0.041). Regarding other genetic markers, the distribution of the genetic profile did not significantly affect the overall survival of the patients.
Table 1 details the characteristic of the study population classified as patients younger and older than 20 years old. High male frequency in the adult population contrasted with the high female frequency in the pediatric population. However, the differences were not statistically significant (p = 0.063). Regarding treatment modalities, while adult patients followed surgical removal in most cases, pediatric patients followed biopsy (p = 0.023). Differences were insignificant regarding the sex, hydrocephalus, tumor dimensions, intracranial or spinal dissemination at initial presentation, the extent of surgical resection, radiotherapy delivery, chemotherapy delivery, postoperative complications, disease progression at the last MRI evaluation, clinical FU, nor survival rate.