Pineal parenchymal tumors of intermediate differentiation (PPTIDs) are a subset of pineal parenchymal tumors (PPTs) with histological and biological features between well-differentiated pineocytoma and poorly differentiated pineoblastoma [10]. These tumors present diagnostic challenges, and recent molecular profiling has resulted in a substantial rate of reclassification [8, 9]. Previous attempts to stratify PPTIDs into distinct groups based on mitotic count and neurofilament expression, although initially implemented into the 2007 WHO classification, did not show significant relevance in subsequent studies [1,2,3]. Hence, there is currently no definite grading criteria for PPTIDs in the latest WHO classification [1]. PPTIDs have been classified morphologically into diffuse, lobular, pleomorphic and transitional subtypes (Fig. 1 and Supplementary Figs. 1–4) [1,2,3]. In addition, they have been epigenetically categorized as PPTID-A and PPTID-B [2]. Small insertions in the KBTBD4 gene have been suggested to be characteristic for PPTIDs, albeit not present in every case [5, 7]. The prognostic significance of KBTBD4 insertion is still unclear [7]. Furthermore, the association between morphological and molecular subtypes of PPTIDs, and the prognostic significance of these subtypes remains largely unknown.

Fig. 1
figure 1

a–h PPTIDwild of transitional subtype depicting areas with diffuse growth pattern and clear-cell morphology (a) and, also, rosette-forming areas with typical histology of pineocytoma (b). Although both areas showed elevated proliferation index (c and d, hotspot Ki67: 8.7%), the neurofilament protein was significantly less expressed in diffuse area (e and f). Hotspot-Ki67 was measured via QuPath in order to reduce the inter-observer variability (g). Molecular analyses revealed a wild-type KBTBD4 gene, loss of chromosome 13q (h) and a DNA methylation class distinct from PPTID-A and -B. i Unsupervised DNA methylation-based t-SNE showed that PPTIDwt are epigenetically more similar to pineocytoma than PPTIDmut. j–l Kaplan–Meier curves by log-rank test. KBTBD4 insertions (j, p value 0.03) and hotspot Ki67 greater than or equal to 8% (k, p value 0.02) were significantly associated with worse progression-free survival (PFS). The 2007 WHO grading system (l, p value 0.8) failed to correlate with the PFS

Full size image

In an effort to address these challenges, we studied a cohort of 34 PPTID patients. The cases have been diagnosed and included into the study following a thorough review of histological slides and molecular findings of all pineal tumors from our internal database for which the diagnostic slides and DNA material were available (n = 110).

All PPTID samples were subjected to comprehensive histological and molecular analyses. The findings were correlated with patients’ survival for 19 patients with available survival data.

The material and methods, as well as the clinical and histological findings, are summarized in the Supplementary data, online resource.

KBTBD4 insertions were detected in 24 out of 34 PPTIDs. PPTIDs with wild type (PPTIDwt) and mutant (PPTIDmut) KBTBD4 gene showed no significant difference in proliferation index (hotspot Ki67 measured via QuPath software), mitotic count, age and sex distribution. PPTIDmut had significantly higher cell density (12,160 vs. 7954 cells/mm2; p value < 0.001), a smaller cell size (78 vs. 100 µm2, p value < 0.001), and more frequently showed profound loss of NFP expression compared to PPTIDwt (70 vs. 20%, p value 0.02). PPTIDmut and PPTIDwt were mostly of diffuse (62%) and transitional subtypes (70%), respectively.

Genome-wide DNA methylation analysis showed that the 10 PPTIDwt clustered separate from the 24 PPTIDmut, but together with pineal gland/cyst and pineocytoma (Fig. 1i). Loss of chromosome 13q was more common in PPTIDwt than PPTIDmut (60 vs. 13%, p value < 0.01), mostly as the single alteration in the copy-number profiles of PPTIDwt. Of note, only one of the 8 pineocytoma in our cohort showed loss of chromosome 13q.

There was one radiological tumor recurrence in PPTIDwt during follow-up (n = 9, mean 44, range 6–169 months), whereas recurrences were documented in 7 out 10 PPTIDmut (mean 29, range 9–87 months). The presence of insertions in KBTBD4, methylation class PPTID, diffuse morphology subtype and hotspot Ki67 greater than or equal to 8% were significantly associated with a worse progression-free survival (PFS) by log-rank test (Fig. 1j–k and Supplementary Fig. 5). The 2007 WHO grading system showed no correlation with the PFS (p value 0.8; Fig. 1l). Insertion in KBTBD4 was the only significant variable (p value 0.04) in the Cox regression model that incorporates KBTBD4 status, hotspot-ki67, morphology subtype, mitotic count, extent of resection (EOR) and adjuvant therapy. Although prior works are controversial regarding the prognostic significance of EOR and adjuvant therapy [4, 6], they failed to reach the level of significance for association with PFS in our study (Supplementary Table 1 and Supplementary Fig. 6).

In summary, genetic and epigenetic profiling identifies two broader subgroups of PPTIDs with distinct histological features and clinical course: (1) the PPTIDs with KBTBD4 insertions which have the methylation profile subtypes PPTID-A or PPTID-B and frequently have a small-cell morphology and an unfavourable clinical course and (2) the PPTIDs without KBTBD4 insertions, herein suggested to be called PPTID-C, which have a methylation profile most similar to pineocytoma and frequently show a large-cell morphology, loss of chromosome 13q, and a favourable clinical course. The incorporation of this subgrouping in grading of PPTIDs may improve risk stratification.