The promise of molecular profiling of choroid plexus tumors for diagnostic and prognostic stratification: where to go from here?

Choroid plexus tumors are rare, making up only about 0.2% of all CNS tumors, and most commonly arise in young children.¹ Based on histology, choroid plexus tumors have been subclassified into choroid plexus papilloma (CPP, World Health Organization [WHO] grade I), atypical choroid plexus papilloma (aCPP, WHO grade II), and choroid plexus carcinoma (CPC, WHO grade III), with aCPP showing increased mitotic activity and CPC displaying common histologic hallmarks of malignancy.² CPPs are typically treated with surgical resection alone, resulting in excellent outcome. In contrast, the outcome in CPC remains poor, especially when gross total resection is not feasible.^3–6 While adjuvant chemotherapy and radiotherapy appear to improve survival in CPC, the use of radiotherapy is limited, especially in infants and very young children due to its toxicity. Whether aCPPs behave (and should be treated) like CPPs or CPCs is controversial, with some data suggesting that this may be age dependent.⁷

In this issue, Thomas et al describe a large series of 92 patients with choroid plexus tumors, in which genome-wide methylation profiling and unsupervised hierarchical cluster analysis led to the identification of 3 distinct subgroups, termed cluster 1, cluster 2, and cluster 3. Cluster 1 was associated with young age and predominantly supratentorial location, cluster 2 with adult age and predominantly infratentorial location, and cluster 3 with young age and predominantly supratentorial location. The authors found that cluster 3 tumors carried a poor prognosis, while patients in clusters 1 and 2 fared significantly better. Strikingly, all of the CPCs (WHO grade III) were found in cluster 3. In contrast, CPPs were dispersed among all 3 clusters, and with the exception of a single patient, no recurrences were observed. Surprisingly, aCPPs (WHO grade II) were also found among all 3 clusters, but recurrences were seen in only those in cluster 3. These findings are in contrast to a recent paper from another group using genome-wide methylation and gene expression profiling of 100 choroid plexus tumors.⁸ Merino et al identified 2 clusters, with one comprising nearly all CPCs and the other containing most CPPs and aCPP, whereas Thomas et al's findings indicate that a subset of aCPPs cluster with CPCs and have a recurrence risk more similar to CPCs. Another interesting finding is that adult and pediatric CPPs clustered separately, pointing to differences in biology, although the outcomes were similarly favorable in both groups.

Tumor protein 53 mutation status and DNA copy number profiles are known to be associated with outcome in choroid plexus tumors.^8,9 In the current study, overexpression of p53 was found in 37% of CPCs but in only 1 aCPP (1/25, 4%), and in none of the CPPs. The single aCPP with p53 overexpression was in cluster 3, together with all of the CPCs. Of note, however, this patient remained recurrence free despite the higher-risk methylation profile and p53 overexpression.

Genome-wide methylation profiling is emerging as a powerful new diagnostic tool for brain tumors,¹⁰ and in some entities, such as gliomas^11,12 and ependymomas,^13,14 methylation profiling appears to be a better predictor of clinical behavior compared with traditional histopathology and WHO grading.

Thomas et al's paper in this issue demonstrates the promise and limitations of methylation profiling for choroid plexus tumors. In their study, methylation profiling identified a subset of CPPs and aCPPs that carried an increased risk of recurrence and clustered with CPCs. Multivariate regression analysis revealed that only WHO grade was an independent predictor of overall survival, and only WHO grade and p53 status were independent predictors of progression-free survival.

Besides maximally safe surgical resection, there is no standard treatment approach for aCPPs. This is reflected in the data collected for the study by Thomas et al, which indicate that aCPP patients received a wide variety of adjuvant therapies, including chemotherapy, radiation therapy, a combination of both, or no adjuvant therapy.

The findings of Thomas et al suggest that genome-wide methylation profiling may provide additional prognostic information, especially in aCPPs. The limitations of the study include the retrospective nature, non-uniform treatment approaches, incomplete or limited clinical information in many patients, and lack of p53 sequencing data.

Where to go from here? The data from Thomas et al and Merino et al clearly highlight the potential of genome-wide methylation profiling, gene expression profiling, DNA copy number analysis, and p53 status (ideally determined by sequencing rather than immunohistochemistry) to provide additional layers of information to improve risk stratification in choroid plexus tumors. To translate these findings into better prognostication and improved outcomes for choroid plexus tumor patients, further studies are needed. First, the findings should be validated in smaller but clinically well-annotated cohorts of patients, with the aim to identify the most promising combination of specific biomarkers. Second, well-designed prospective clinical trials will be necessary to ultimately realize the promise of molecular-based risk stratification and improved risk-adapted treatment for patients with choroid plexus tumors.