Pineal parenchymal tumor of intermediate differentiation: a single-institution experience

Abstract

Background

Pineal parenchymal tumors are exceedingly rare brain tumors responsible for less than 1% of all adult primary intracranial malignancies in the United States. In this study, we describe the clinicopathologic features, management, and outcomes of patients with pineal parenchymal tumor of intermediate differentiation (PPTID).

Methods

We describe a single-center, multidisciplinary team experience in managing PPTID patients over a 15-year period (January 2000 to January 2015) at The University of Texas MD Anderson Cancer Center (MDACC). Pathology was reviewed by the pathology collaborators (A.G. and G.N.F.) and retrospective chart review was performed for treatment and clinical outcomes.

Results

We identified 17 patients (9 male) with diagnosis of PPTID. Median age at diagnosis of PPTID was 37 years (range, 15-57 years). Follow-up ranged from 0.1 to 162.8 months with 6 reported deaths. Most patients presented with headaches and diplopia. Three patients had neuroaxial dissemination at initial diagnosis, and recurrence of tumor was common (7/16) despite treatment.

Conclusions

No clear prognostic factors were identified in this series. Extension of resection showed a trend toward improved survival. PPTID with neuroaxial dissemination benefits from aggressive initial treatment including craniospinal irradiation and adjuvant chemotherapy, whereas localized disease may be treated traditionally with maximum debulking followed by adjuvant radiotherapy alone. Long-term monitoring is recommended for neurotoxicity and/or late recurrence.

Pineal parenchymal tumors (PPTs) are exceedingly rare brain tumors responsible for approximately 0.8% of all adult primary intracranial malignancies in the United States.¹ PPTs include well-differentiated pineocytoma, poorly differentiated pineoblastoma (PB), and pineal parenchymal tumors of intermediate differentiation (PPTIDs). PPTIDs account for approximately 21% to 54% of all pineal parenchymal tumors in published series, and the variability in reported incidence is thought to be due to difficulty in diagnosis in the spectrum of PPTs.^2–5

Owing to the rarity of PPTID, data regarding the optimal treatment, clinical course, and prognosis is limited. In this study we describe the clinical course, pathology findings, management strategies, and outcomes of 17 patients with PPTID treated at The University of Texas MD Anderson Cancer Center (MDACC) in an effort to better understand this rare tumor entity.

The pineal gland is a specialized circumventricular organ composed of pinealocytes, which are specialized neurosecretory cells that produce melatonin and help regulate circadian rhythm.^6–8 Only about 27% of pineal region tumors originate from the pineal gland itself, and these are collectively referred to as PPTs.⁹ PPTs are composed of a spectrum of pinealocyte-derived neoplasms ranging from well-differentiated World Health Organization (WHO) grade I pineocytoma, intermediate-grade PPTID, to poorly differentiated WHO grade IV PB.⁵ PPTID represents the middle step in the PPT morphologic spectrum. The term PPTID, initially introduced in 1993 by Schild and colleagues,¹⁰ refers to a relatively new entity that was formally codified in 2000, and although it is thought to have biological behavior between pineocytoma and PB, there is no clear pathologic correlation to biological behavior, which complicates prognostication and treatment.¹¹ Definite histologic grading criteria are not yet defined.⁵

The mean age at diagnosis of PPTID is 41 years, although all ages can be affected.⁵ Patients usually present with signs of increased intracranial pressure, ataxia, Parinaud syndrome, or diplopia.^10,11 Associated obstructive hydrocephalus is also common.¹² Imaging studies typically show an invasive, heterogeneously enhancing pineal mass with hyperintense signaling on T2-weighted images (Figure 1). Heterogeneous management and outcome is due in part to our limited knowledge of the biology and molecular genetics. Recently, genome sequencing and global DNA methylation profiling demonstrated that PPTID harbors a KBTBD4 mutation, whereas a DICER1 mutation, DROSHA homozygous deletion, and/or PDE4DIP microduplication are seen only in PB.^13,14 A proposed classification, based on these discoveries, delineates 5 molecular subgroups of PBs. Interestingly, in group 3 there were PBs with lower Ki-67/MIB-1 indices and some others with KBTBD4 mutation more consistent with PPTIDs.¹⁵ However the global DNA methylation identified a sole tumor entity with homogenous and favorable outcome despite being previously diagnosed as PB, providing new insights into the pathogenesis of PPTs¹⁵ and suggesting overlap between PPTID and PB when diagnosing by histology features.

Methods

Pathologic Evaluation

This study was performed under the approval of the institutional review board (protocol PA15-0081). A search of the MDACC institutional database between January 2000 and January 2015 resulted in 17 adult patients (age ≥ 18 years) and 23 specimens with histologically confirmed PPTID diagnoses.

Pathology slides were reviewed and a diagnosis of PPTID or PPT was confirmed by a board-certified neuropathologist (G.N.F.) according to 2016 WHO classification. Synaptophysin and/or neurofilament were used as neuronal markers; positive staining was recorded for cases showing cytoplasmic labeling, and reactivity was further characterized as either focal or diffuse. Mitoses were typically assessed by light microscopy, but several cases also used immunohistochemistry with a more sensitive marker of mitotic activity: phospohistone-H3. Ki-67/MIB-1 immunostaining used an IgG1 kappa, antihuman monoclonal mouse antibody (Dako) clone for cases processed in the MDACC laboratory. The Ki-67 proliferation index was evaluated by manual estimation or quantified based on automated imaging analysis using Aperio ImageScope v.12.1.0.5029 by Leica Biosystems for computer-assisted quantitation. PPTIDs that were assigned grade II had well-differentiated tumor cells resembling pineocytes, but lacked the typical pineocytomatous rosettes commonly found in pineocytomas. Grade II demonstrated expression of neuronal immunomarkers, such as synaptophysin or neurofilament, and showed fewer than 6 mitotic figures per 10 high-power field (hpf). PPTIDs assigned grade III had less-differentiated tumor cells, but lacked the sheets of primitive small blue cells with nuclear molding that are typical of PBs. Tumors designated as grade III had very weak/negative neuronal immunostaining with fewer than 6 mitoses per 10 hpf or demonstrated positive neuronal immunostaining with 6 or more mitotic figures per 10 hpf.

Statistical Analysis and Clinical Outcomes Assessment

Demographic and clinical follow-up data for each patient were extracted from the electronic medical record. The length of follow-up was defined as the duration of time from each patient’s definitive surgery to the time of death or date of his or her most recent clinical or radiographic evaluation. Death was confirmed by review of medical records, tumor registry, death certificate, and/or social security index. Survival analyses were performed using the Kaplan-Meier and Cox proportional hazards method and log-rank test. When the date of death was not known, the record was censored in the analysis as of the date of last follow-up.

Data collected were summarized using standard descriptive statistics and frequency tabulation. Time to event end points including overall survival (OS) and progression-free survival (PFS) was estimated using the Kaplan-Meier method and the comparisons between or among characteristics groups were conducted using the log-rank test. Univariate Cox proportional hazard models were applied to assess the effect of covariates of interest on time to event end points. All computations were carried out in SAS 9.2 and S-plus 8.0 or GraphPad Prism 6.

Results

Patient Characteristics

A search of the MDACC institutional database between January 2000 and January 2015 identified 17 adult patients (age ≥ 18 years) with histologically confirmed PPTID diagnosis. The male to female ratio was 1:1. The median age at initial diagnosis was 37 years (range, 15-57 years). Nearly all patients (16/17) presented with headaches, and some also presented with focal neurological signs, with diplopia being the most common symptom (5/17), followed by ataxia/vertigo (2/17). Three patients had disseminated disease in neuroaxis at diagnosis (Figure 1). The mean duration of follow up was 62.6 months (range, 0.1-162.8 months). At the time of the analysis, there was a limited number of events, with only 6 deaths (1 death in the immediate postoperative period) and 7 progressions. Most of the patients with disease progression (6/7) had homogeneous contrast enhancement on initial MRI (Figure 1), whereas those without recurrent disease (6/9) had both no/mild restriction of diffusion-weighted imaging (DWI) signal and patchy/mild contrast enhancement of the tumor. Additional data regarding demographics, radiological features, tumor location, treatment, and outcome are detailed in Table 1.

Pathologic Features

Twenty-three specimens were gathered from the 17 patients. Ten of the 23 specimens were surgically obtained and processed at an institution other than MDACC. About half the specimens (11/23) were assigned grade II, about half (11/23) were assigned grade III, and one specimen (patient 3) was assigned grade II to III. Grade III tumors had a greater number of mitoses on average, with 8.3 vs 2.4 mitoses/10 hpf or 17.5% vs 5.0% Ki-67/MIB-1. Representative samples are shown in Figure 2. Only grade III tumors showed the presence of necrosis. All 5 grade II tumors that were stained with neurofilament showed positive staining in a predominantly diffuse pattern, compared to only 5 of 9 grade III tumors (3 diffuse, 2 focal). Synaptophysin was positive in 7 of 7 grade II tumors, 9 of 9 grade III tumors, as well as the 1 grade II to III tumor.

Initial Treatment

Surgery

Three patients with dissemination within the neuroaxis at diagnosis underwent biopsy only. The majority of patients with localized intracranial disease underwent gross total resection (GTR) (7/14). Six patients underwent subtotal resection (STR) and one patient had an unknown extent of resection. One patient died in the immediate postoperative period, and was removed from further survival calculation and discussion.

Radiation

All but one patient (case 17) had adjuvant radiotherapy following the surgery. Of the 15 patients who received radiotherapy, 12 received craniospinal irradiation (CSI) (80%). This was further subdivided into 5 patients who received proton beam radiation and 7 patients who received intensity-modulated radiation therapy (IMRT). Two of the patients (cases 1 and 16) received CSI and focal boost. Of the 12 patients who received CSI, 3 patients had cerebrospinal fluid (CSF) dissemination, 3 underwent STR, and 3 had a grade III tumor. Only 3 patients (20%) underwent focal radiation alone; 1 patient received fractionated proton radiotherapy, and the other 2 received stereotactic radiosurgery as initial therapy.

Chemotherapy

Only 4 patients received chemotherapy at the time of initial diagnosis and treatment. This group largely consisted of 3 patients with dissemination along the neuroaxis at diagnosis. The fourth patient (case 9) had localized disease, but had STR and received craniospinal IMRT and adjuvant chemotherapy. The chemotherapy regimen typically used a platinum-based multidrug regimen. Detailed upfront chemotherapy regimens are listed in Table 2.

Treatment at Recurrence

Seven patients had tumor recurrence, including 2 patients (cases 1 and 4) who had disseminated disease at diagnosis; their recurrences occurred at 25.6 months and 63.2 months, respectively. Case 1 had distant failure in the thoracic spinal cord, and case 4 had local failure. Most tumor recurrences from localized disease occurred within 24 months of initial diagnosis, and could be either local or distant. Median time to recurrence was 20.9 months (range, 12.7-159.5 months).

One patient (case 9) had a late recurrence at 160 months, after treatment with adjuvant radiation and chemotherapy. Interestingly, one patient (case 17) who was treated with only surgery at diagnosis underwent stereotactic radiosurgery only at the time of recurrence and had a durable response. Management of patient recurrences varied depending on the prior treatment regimen and is summarized in Table 3.

Discussion

Although PPTIDs are presumed to have intermediate aggressiveness and response to treatment compared to well-differentiated pineocytomas and poorly differentiated PBs, limited clinical data are available regarding the prognosis and optimal treatment of this recently defined tumor entity. Owing to the extreme scarcity of cases available for study, little progress has been made in identifying clinically relevant parameters for determining prognosis and optimal treatment strategies.

Our patient population was fairly balanced in terms of sex, with a male to female ratio of 1:1, in contrast to the reported female predominance seen in some published series.^10,11,16 The mean age at diagnosis was comparable to historical data at 37 years, with a range of 15 to 57 years.^17,18 Three patients (18%) had dissemination in the neuroaxis at the time of initial diagnosis, and there were no documented extraneural metastases. Despite the early dissemination and multiple recurrences, OS was more than 5 years with aggressive treatment for this subgroup.

In terms of prognostic factors, younger age was not confirmed as a poor prognostic factor within this PPTID series, and previous reports may be driven instead by the increased prevalence of PBs in children when considering the whole spectrum of PPTs.^10,16,18 There are mixed reports regarding the extent of resection as a prognostic factor in PPTs, and our series did not find statistically significant differences in OS or PFS for PPTID based on extent of resection.^10,16,19,20 The trend toward improved OS and PFS with resection is likely due to a selection bias, because biopsy was performed only on patients with disseminated disease at diagnosis. Preoperative radiological features with high DWI signal and homogeneous contrast enhancement of the tumor may be associated with high cellular PPTID or PB,²¹ suggesting a higher chance for progression compared to those with only a mild DWI signal and patchy contrast enhancement.

Pathologically, about half the specimens met criteria for grade II and the other half met criteria for grade III based on previous WHO guideline. There was no statistically significant difference in PFS or OS based on WHO grade, which further reinforces the need for prognostically significant biomarkers in PPTIDs. A recent report by Raleigh and colleagues described 2 morphologic subtypes (small-cell and large-cell) of PPTID with distinct clinical outcomes, and further research regarding these subtypes is warranted.²⁰ In addition, survival difference between PPTID and PB may depend of molecular features (KBTBD4 mutant vs KBTBD4 wild type) and their global DNA methylation status.^13–15

Published data regarding optimal treatment modalities, such as postsurgical radiation with or without adjuvant chemotherapy, are limited. Some studies have shown better tumor control using CSI in neoplasms with potential for dissemination, including PPTID,¹⁰ and the observed pattern of practice in our series certainly shows a preference for CSI even for localized disease, because most patients (12/16) received CSI as an adjuvant therapy. This preference was based on the presence of CSF dissemination, residual tumor, grade III tumor, or young adult age. There were no statistically significant differences in PFS or OS between proton and IMRT CSI. Despite the benefit shown by radiotherapy, the long- and short-term toxicity associated with CSI has spurred interest in exploring localized adjuvant radiotherapy to minimize radiation exposure.²² Our series included only a small number of patients who received localized radiotherapy alone, but with one remarkable case (case 17) showing no evidence of disease 98 months after local radiotherapy alone.

There is less consensus regarding chemotherapy in PPTID. Platinum-based adjuvant chemotherapy was used in all patients with disseminated disease at diagnosis, but only rarely used in patients with localized disease at diagnosis. Case 9 had the most aggressive adjuvant therapy for a patient with localized disease, with adjuvant CSI and chemotherapy, and had a long duration of response (160 months), whereas the other recurrent localized PPTIDs typically progressed within 2 years after administration of adjuvant radiotherapy. However, the follow-up and sample size are not robust enough to see whether chemotherapy at the time of recurrence produces similarly durable response. There was no preferred chemotherapy regimen at the time of recurrence.

Owing to the rarity of PPTIDs, our study was limited by a small sample size, lack of genomic profiling, as well as the inherent biases associated with the retrospective nature of the review and the selection bias from a referral-based tertiary academic center. Undersampling bias may be also considered in those patients with disseminated disease who underwent biopsy only. Furthermore, many patients included in this study received their initial treatment at different facilities, which contributed to the heterogeneity of initial treatments. Ten specimens were obtained and processed at outside institutions, which may have contributed to variation in the quality of tissue processing and immunostaining intensity. Additionally, the pathology specimens sent to MDACC did not always include the standardized or necessary immunohistochemical stains, and paraffin-embedded blocks were often unavailable for additional tissue-based studies, including molecular studies.

Conclusions

This series contributes to the scarce data available regarding the prognosis, clinical course, and treatment of PPTIDs. PPTIDs have potential for neuraxial dissemination, and it is crucial to fully stage the patient at the time of the diagnosis with whole-spine MRI with and without contrast as well as CSF cytology. In disseminated disease, aggressive initial treatment is appropriate with CSI and adjuvant chemotherapy. In localized disease, patients have been traditionally treated with maximum debulking followed by adjuvant radiotherapy, and rarely adjuvant chemotherapy.

Future directions for research should address identifying and validating prognostic factors in a large-scale registry to appropriately stratify patients and further refine the role of adjuvant radiation and/or adjuvant chemotherapy and its impact on OS, especially in the setting of GTR. Given the variety of radiation delivery modalities available, studies comparing CSI vs localized radiation therapy for OS and long-term toxicity would be helpful in balancing the survival benefit and treatment-related morbidity. We report one case of delayed and localized radiation that conferred durable response, and larger studies are needed to determine whether delayed radiation has the same clinical efficacy as early radiation. Though most of the observed recurrences occurred early (within 24 months) in our series, several patients had late recurrences, indicating that long-term follow-up is necessary in this patient population.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Conflict of interest statement. None declared.

References