Clinicopathologic analysis of pineal parenchymal tumors of intermediate differentiation: a multi-institutional cohort study by the Kyushu Neuro-Oncology Study Group

Pineal parenchymal tumors of intermediate differentiation (PPTIDs), which were recognized in the 2007 World Health Organization (WHO) classification, are rare, accounting for less than 1% of all central nervous system tumors. This rarity and novelty complicate the diagnosis and treatments of PPTID. We therefore aimed to evaluate the clinicopathological significance of this tumor. At 11 institutions participating in the Kyushu Neuro-Oncology Study Group, data for patients diagnosed with PPTID were collected. Central pathology review and KBTBD4 mutation analysis were applied to attain the diagnostically accurate cohort. PPTID was officially diagnosed in 28 patients: 11 (39%) with WHO grade 2 and 17 (61%) with WHO grade 3 tumors. Median age was 49 years, and the male:female ratio was 1:2.1. Surgery was attempted in all 28 patients, and gross total resection (GTR) was achieved in 46% (13/28). Adjuvant radiotherapy and chemotherapy were administered to, respectively, 82% (23/28) and 46% (13/28). The 5-year progression-free survival (PFS) and overall survival rates were 64.9% and 70.4% respectively. Female sex (p = 0.018) and GTR (p < 0.01) were found to be independent prognostic factors for PFS and female sex (p = 0.019) was that for OS. Initial and second recurrences were most often leptomeningeal (67% and 100% respectively). 80% (20/25) of patients harbored a KBTBD4 mutation. Female sex and GTR were independent prognostic factors in our patients with PPTID. Leptomeningeal recurrence was observed to be particularly characteristic of this tumor. The rate of KBTBD4 mutation observed in our cohort was acceptable and this could prove the accuracy of our PPTID cohort.


Introduction
Pineal parenchymal tumors are rare, accounting for less than 1% of all primary central nervous system tumors [1][2][3]. They range from World Health Organization (WHO) grade 1 pineocytomas, to grades 2-3 pineal parenchymal tumors of intermediate differentiation (PPTIDs), and grade 4 pineoblastoma. Of those entities, PPTID was originally recognized in the 2007 WHO classification as a tumor that bridges the spectrum between a slow-growing well-differentiated pineocytoma and a malignant pineoblastoma, and that accounts for approximately 45% of all pineal parenchymal tumors [4]. Owing to this tumor's novelty and rarity, diagnosis is complicated, and no standard treatment has yet been established. Maximal safe resection is accepted as the aim of any surgery, but optimal regimens of adjuvant irradiation and chemotherapy remain unclear [3,5,6]. In 1 3 clinical practice, patients with PPTID are therefore treated using strategies derived from a small number of institutional experiences, case reports, and small case series. To establish optimal treatment, additional information about the characteristics of patients with PPTID and their patterns of care and survival outcomes are needed.
With respect to diagnosis, distinguishing PPTID not only from pineocytoma or pineoblastoma, but also from central neurocytoma or oligodendroglioma, has been reported to be difficult because of the morphologic similarities of a sheetlike architecture and monomorphic round cells [7][8][9]. Moreover, histologic criteria for WHO grading remains undefined. The histologic diagnosis and grading of PPTID in previous reports, including systemic reviews, might therefore contain some errors. Recently, next-generation sequencing revealed that a KBTBD4 mutation is present in approximately 75% of patients with PPTID [10]. Although that genetic mutation could help to distinguish PPTID from pineoblastoma, previous systemic reviews have never used KBTBD4 mutation in their diagnoses [11][12][13][14][15]. Thus, no patient cohort has been diagnosed sufficiently accurately to determine the optimal treatment in PPTID.
For the present study, data were collected for 32 patients locally diagnosed with PPTID at 11 institutions participating in the Kyushu Neuro-Oncology Study Group. To exclude interinstitutional diagnostic error, a central pathology review based on the diagnostic criteria in the 2016 WHO classification of tumors was conducted. To further improve the accuracy of the pathology diagnosis, Sanger sequencing was used to examine mutational hotspots in KBTBD4. Based on a more accurately diagnosed cohort, we aimed to identify basic patient characteristics, treatment details, survival outcomes, and prognostic factors of PPTID.

Ethics
This study was carried out in accordance with the principles of the Helsinki Declaration. The ethics guidelines for medical and biological research involving human subjects in Japan were followed [16]. Approval was obtained from the Institutional Review Board of Miyazaki University (no. G-0121) and all collaborating institutes.

Patient samples and clinical information
The 32 patients with PPTID were diagnosed between 2007 and 2021. Samples from that cohort consisting of formalinfixed paraffin-embedded sections and frozen tissues were collected for pathology evaluation and DNA extraction. Clinical data obtained from medical records included age at the time of surgery, sex, symptoms at presentation, Karnofsky performance status (KPS), imaging, type of surgery, administration of chemotherapy and radiation, recurrence, and prognosis [progression-free survival (PFS), overall survival (OS)].

Pathology evaluation
Central review of all pathology slides was conducted by a senior board-certified neuropathologist (T.I.), who used mitotic figures, expression of neuronal immunomarkers, and the Ki-67 proliferation index to grade the tumors based on the 2016 WHO classification. Synaptophysin or neurofilament, or both, were used as neuronal markers.

DNA extraction
Genomic DNA was extracted from frozen tissue and formalin-fixed paraffin-embedded sections. The QIAamp DNA Mini Kit (Qiagen, Valencia, CA, USA) was used for tissue and the WaxFree DNA Extraction Kit (TrimGen, Sparks, MD, USA) was used for sections according to the manufacturer's protocol. The resulting DNA concentrations were measured with a Qubit R Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA).

Analysis of KBTBD4 hotspot mutations
Mutational hotspots in KBTBD4 were examined using direct DNA sequencing. A spanning 135-bp fragment was identified with the sense primer 5′-TTG TGC CAC CAG ATC ACT GCGG-3′ and the antisense primer 5′-GTC CCG AGG CAA AGG AGC ACAC-3′ [11]. The 25-μL reaction mixture contained 50-ng tumor genomic DNA, 1-μL each of sense and antisense primer (10 μM), and 12.5-μL GoTaq Hot Start Green Master Mix (Promega, Fitchburg, WI, USA). Genomic DNA then underwent polymerase chain reaction amplification, with initial denaturation at 95° for 2 min and 35 cycles of amplification consisting of denaturation at 95° for 30 s, annealing at 58° for 30 s, and extension at 72° for 20 s. Final extension was performed at 72° for 5 min. The sequence reactions used a Big Dye Terminator v3.1 Cycle Sequencing Kit (Thermo Fisher Scientific) with a 20-μL reaction mixture containing 1-μL polymerase chain reaction product, 4-μL polymerase chain reaction buffer, 2-μL sense primer (5 μM), and 1 μL Big Dye Terminator Ready Reaction Mix. Initial denaturation was performed at 96° for 1 min, followed by 25 cycles of amplification (denaturation at 96° for 30 s, annealing at 50° for 5 s, and extension at 60° for 4 min). Sequencing products were immediately submitted to direct sequencing on an ABI PRISM 3130xl Genetic Analyzer (Thermo Fisher Scientific).

Statistical analyses
Ki-67 proliferative indexes were compared using the t-test. PFS and OS were evaluated by the Kaplan-Meier method and compared using the log-rank test. Multivariate analyses of prognostic factors were performed in Cox proportional hazards models. Differences of p < 0.05 were considered statistically significant. Statistical analyses were performed in the Stat Mate III software application (ATMS, Tokyo, Japan).

Recurrence pattern and treatments at recurrence
Tumor recurrence was observed in 32% of the patients (9/28) ( Table 2). Of 9 patients with recurrence, 3 patients (33%, 3/9) were WHO grade 2 and the others were grade 3. And all recurrent patients harbored KBTBD4 mutations. The recurrence was leptomeningeal in six patients (67%, 6/9) and local in two patients (22%, 2/9). One patient of the nine with recurrence underwent biopsy. Six received radiotherapy, with half receiving stereotactic radiotherapy (3/6), and half, whole-spine irradiation (3/6). Three received chemotherapy (the ICE regimen in all cases). Secondary recurrences were observed in five patients, and all were leptomeningeal. The predominance of leptomeningeal recurrences (both initial and secondary) suggests that leptomeningeal dissemination is the specific recurrence pattern in PPTID.

Survival outcomes and prognostic factors
Median PFS for all the patients was 132 months. The estimated 5-year PFS was 64.9%. Median OS was not reached. The estimated 5-year OS was 70.5% (Fig. 2). Univariate analysis showed that OS was significantly better in female than in male patients (median: not reached for women vs. 105 months for men, p = 0.024) (Supplementary 3, 4). PFS was significantly better for patients who underwent GTR than for those who underwent subtotal resection, partial resection, or biopsy (median: not reached for GTR vs. 46 months for subtotal resection, partial resection, or biopsy; p < 0.01) (Supplementary 3, 5). Age, WHO grade or radiotherapy that were reported as prognostic factors previously had no significant effect on survival duration in this study. Other factors including KPS score, tumor size, Ki-67 proliferation index, KBTBD4 mutation status, and chemotherapy also showed no association with longer survival (Supplementary 3). Multivariate analyses demonstrated that female sex (p = 0.018) and GTR (p < 0.01) were independent prognostic factors for PFS and female sex (p = 0.019) was also that for OS (Table 3).

Discussion
In the present study, characteristics, treatments, and survival outcomes of patients with PPTID were analyzed based on a patient cohort that was more accurately diagnosed than in the past, including through central pathology review and KBTBD4 mutation status. To evaluate differences between our data and those from earlier reports, we compared our data with the results of a systemic review that included 127 patients with PPTID from 29 studies [12]. In our cohort, median patient age was older (49 vs. 33 years) and the female preponderance was greater (male:female ratio: 1:2.1 vs. 1:1.6). In both cohorts, headache and gait abnormality were the two main symptoms, and the rates of leptomeningeal recurrence were close (67% vs. 62.5%). Those data suggest that female preponderance and leptomeningeal recurrence are major characteristics of PPTID and that the two main symptoms are headache and gait abnormality. With respect to survival, the 5-year PFS was better in our cohort than in the systematic review (64.9% vs. 52.2%), but the 5-year OS was worse (70.5% vs. 84.1%). Those differences in the survival data suggest that our findings might more accurately reflect the progression pattern in PPTID. As mentioned, many patients with PPTID experience hard-to-control leptomeningeal recurrences despite multiple treatments. Thus, theoretically, PFS and OS must be closely, potentially suggesting a more "purely PPTID" cohort in our study than in the systemic review cohort.
Female sex and GTR were independent prognostic factors in our cohort. Maximal safe resection was established as the standard aim of surgery since many reports showed GTR as the better prognostic factor [5,12,15]. With respect to female sex, Mallick et al. also reported that factor as prognostic of a better outcome in PPTID [12]. Interestingly, female patients have also been reported to experience significantly better survival in pineoblastoma [17,18]. Moreover, all pineal parenchymal tumors-including pineocytoma, PPTID, and pineoblastoma-demonstrate a female predominance [10,12,19,20]. Although no specific molecular feature of pineal parenchymal tumors has been reported in female patients, sex might be one of the important factors determining tumor characteristics. Some earlier reports recommended that patients with a resection less than GTR receive adjuvant radiotherapy [12,[21][22][23]. However, our study demonstrated an inverse result: radiotherapy was associated with trend toward worse prognosis. We considered that result to have been influenced by treatment bias. At initial treatment, 12 of 13 patients whose resection was less than a GTR received radiotherapy, and two of those who had leptomeningeal dissemination received craniospinal irradiation. At recurrence, six of nine recurrence patients including two who undergone radiotherapy as initial treatment received irradiation. Thus, patients who had residual, recurrent, or disseminated leptomeningeal disease received irradiation aggressively based on earlier reports in the literature that had demonstrated the effectiveness of radiotherapy in PPTID. Survival in our cohort was not inferior to that in previous reports [7,12,15]. So, our data might therefore suggest the efficacy of the therapeutic irradiation strategy used in our cohort. Although hard to conduct, prospective randomized controlled trials would be required to prove the efficacy of radiotherapy in PPTID.
In the present study, we found no correlation between WHO tumor grade and survival. WHO grade was previously reported to be an independent predictor of survival in PPTID. Fauchon et al. reported a large survival advantage in grade 2 disease compared with grade 3 disease (5-year OS: 74% vs. 39%) [7]. Lu et al. similarly reported a statistically significantly longer OS duration in grade 2 disease than in grade 3 disease (mean: 108 vs. 44 months) [15]. In contrast, like us, Yu et al. failed to demonstrate that WHO grade is prognostic, and Mellick excluded WHO grade as a candidate prognostic factor [6,12]. We considered these conflicting reports to be the result of undefined histologic grading criteria. Two major reports concerning the criteria for WHO grading have been published. Sato et al. reported that necrotic appearance, mitosis rate, and immunohistochemical expression of neurofilament protein are useful for determining the WHO grade [24]. In contrast, Jouvet et al. reported that fewer than six mitoses and strong immunopositivity for neurofilaments indicated grade 2 disease, while more than six mitoses or less than six mitoses without strong immunostaining for neurofilaments indicated grade 3 disease [25]. Neither set of criteria was adopted for the WHO classification. As we enter the era of comprehensive tumor molecular data collection, WHO grade might be unimportant factor for tumor prognosis.
In a recent study, comprehensive analysis using nextgeneration sequencing revealed KBTBD4 mutations in PPTID [10,13,14]. Two identified mutations harbor identical somatic small in-frame insertions (p.R313delinsPRR and p.R313_M314insRPP) in the KBTBD4 gene [11].  Although the mechanism by which these mutations drive tumorigenesis in PPTID is currently uncertain, the presence of these mutations is considered useful for distinguishing PPTID from pineoblastoma [10,13,14]. In our study, KBTBD4 mutations were analyzed to achieve more accurate diagnoses in our cohort. Liu et al. showed that 74% of patients with PPTID harbor a KBTBD4 mutation, and Uchida et al. showed that 66.7% harbor a mutation [10,11]. In line with those reports, we found that 80% of our cohort harbored a mutation, a result that we consider to proving the accuracy of the diagnoses in our cohort. Liu et al. also showed that patients positive for a KBTBD4 mutation were older than those who were mutation-negative (median age: 40 vs. 15 years) [10]. We observed no correlation between age and KBTBD4 mutation status in our cohort (median age: 48 years for patients testing positive vs. 53 years for patients testing negative). But also in our cohort, prognosis tended to be better in patients negative for a KBTBD4 mutation than in patients who were mutation positive (log-rank test: p = 0.089 for PFS, p = 0.181 for OS). However, an evaluation of the 20 KBTBD4 mutation-positive and 5 mutationnegative patients demonstrated treatment bias, in that all mutation-negative patients underwent GTR and received no chemotherapy. Multivariate analysis also dismissed the efficacy of KBTBD4 mutation status as a prognostic factor. Moreover, with respect to other factors including sex, KPS, MRI findings, tumor size, WHO grade, Ki-67 proliferation index and radiotherapy, no statistically significant differences were found between patients with and without KBTBD4 mutation (Supplementary 6). Considering all the foregoing findings, no clinical implications of KBTBD4 mutation status were uncovered in our study. However, Liu et al. hypothesized that KBTBD4 mutation is the disease-defining driver gene for PPTID and a candidate therapeutic target [10]. In the future, KBTBD4 mutation status might not only be accepted as the key factor in diagnosing PPTID, but also as a determiner of clinical status with respect to prognosis and targeted therapy.
Our study has limitations. Its multi-institutional retrospective cohort design could result in bias with respect to choose of treatment strategy. Variation in treatments, including radiotherapy and chemotherapy, at each institution could have affected the survival data. Furthermore, the limited number of patients might explain the absence of the statistical power needed to evaluate the effectiveness of adjuvant therapy in PPTID. The study design also affected that statistical power. Comprehensive molecular analyses were not performed in this study. Although analysis of KBTBD4 mutation status could act as a surrogate in diagnosis, the genetic background of our cohort could have better informed the analysis of the clinical characteristics of PPTID.

Conclusions
Our PPTID cohort demonstrated that median age was 49 years-old and female patients were preponderant. Headache and gait abnormality were the main initial symptoms. Survival analyses showed a 5-year PFS of 64.9% and a 5-year OS of 70.4%, with female sex and GTR as independent prognostic factors for PFS and female sex as that for OS. Recurrences were predominantly leptomeningeal (67% of initial recurrences and 100% of second recurrences), a pattern that is particularly characteristic of PPTID. The finding that 80% of our patients harbored a KBTBD4 mutation supported the diagnostic accuracy of our PPTID cohort. A prognostic role for, and benefit of, adjuvant therapy was not evident in the study. Although hard to conduct, prospective randomized controlled trials with comprehensive molecular analyses will be required to establish standard therapy for PPTID.