For the first time, we prospectively evaluated the use of IHC as a potential screening method for germline MMRD in pedHGG. In our cohort of 127 central-review confirmed pedHGG patients with evaluable MMR-IHC, staining of MSH2, MSH6, MLH1, and PMS2 successfully identified all present cases with LS and CMMRD. In addition to cases with germline affection, IHC also detected cases with exclusive somatic MMR gene alterations. Of note, these patients may also qualify for treatment adaptation toward checkpoint inhibition instead of cytotoxic therapy, as it is frequently beneficial for pedHGG patients with germline MMRD [4, 9, 36]. Indeed, 2/3 cases with exclusive somatic MMR alterations showed a molecular signature characteristic for MMRD with elevated TMB, MSI, and ALT phenotype [39] indicating the use of checkpoint inhibition [4, 9, 10, 36, 42]. In our series, IHC detecting protein expression independently of the underlying pathomechanism appeared even superior to identify MMRD compared to extensive molecular panel diagnostics.

Previously, in a study by Chung et al., the diagnostic ability to detect MMRD was compared between Low-pass Genomic Instability Characterization (LOGIC) assay, TMB analysis, traditional MSI panel testing, and IHC. This comparison was performed in 56 mixed childhood cancer cases including 27 CNS tumors [6]. In this mixed cancer cohort, the LOGIC assay reached the highest sensitivity to detect MMRD (100%). However, the sensitivity of IHC (86%) was superior to traditional MSI panel testing (14%) and TMB analysis (80%). With focus on CNS tumors, IHC reached a sensitivity of 93%; 25/27 MMRD cases were successfully identified by IHC. Interestingly, IHC could be successfully performed in eight cases with insufficient sequencing coverage during low-pass whole-genome sequencing. Since in our study, MMR-IHC detected MMRD with 100% sensitivity in 127 investigated cases, the performance of the different assays may vary depending on the respective type of cancer analyzed.

Loss-of-function alterations within the MMR genes can lead to the expression of a defective protein. These cases appear to be rare. Hechtman et al. examined this phenomenon in a cohort predominantly comprising cases of colorectal carcinoma and uterine endometrioid carcinoma. Their study reported that approximately 6% of MSI-high (MSI-H) cases exhibited retained MMR protein expression [18]. Similarly, Chen et al. observed that 7% of MMR-deficient colorectal carcinoma cases (6 out of 82) showed intact MMR immunohistochemical staining [7]. In our study, in all 71 cases with retained protein expression, DNA sequencing analysis never identified any alterations in the MMR genes. However, as MMRD is predominantly studied in tumors other than pedHGG [7, 18], evidence regarding the frequency of this phenomenon in pedHGG remains limited. To mitigate the risk of missing rare MMRD cases with retained MMR protein expression, we recommend incorporating the evaluation of MMRD-associated morphological characteristics [20] into the algorithm of pedHGG diagnostics (Fig. 5).

Adapted diagnostic workflow for routine MMRD screening in pedHGG. We would like to recommend performing MMR-IHC in all cases of pedHGG. In cases with unavailable MMR-IHC, or MMRD-associated tumor cell morphology (severe pleomorphism and/or giant cell features), molecular testing should be performed—if feasible—to identify any underlying MMR gene alteration with the need for subsequent MMR gene germline testing. Based on the present findings, MMR gene germline testing may be promptly initiated in cases of MMR-IHC loss; however, prior molecular analysis of tumor tissue for MMR gene alterations can be performed at the investigator’s discretion. Ideally, a combination of molecular assays covering all potential types of alterations including hypermethylation is implemented. Germline testing is recommended for all MMRD patients. MMR(D) mismatch repair (deficiency), IHC immunohistochemistry

The WHO 2021 recommends immunohistochemical MMRD testing in pedHGG cases with severe pleomorphism or giant cell features [20]. Based on our findings, we suggest an adapted diagnostic workflow for MMRD screening in pedHGG (Fig. 5). Instead of performing MMR-IHC only in pedHGG cases with MMRD-associated morphological characteristics, we suggest performing MMR-IHC in all pedHGG cases at first-line, as in our study, 2/10 MMRD cases lacked MMRD-associated morphological features. In cases with retained MMR protein expression, the mentioned morphological characteristics can then be used to identify rare cases with loss-of-function alterations by further molecular investigations.

The LOGIC assay using low-pass whole-genome sequencing to detect microsatellite instability with much greater sensitivity than traditional MSI panels in childhood chancers seems to represent a robust tool to investigate pedHGG cases with unavailable or difficult to interpret IHC [6]. Traditional MSI panels were originally developed for colorectal cancer [43], and their gene selection and cut-off values may not be directly applicable to other cancer types, particularly pediatric malignancies. A study by Hause et al. highlights the presence of instability signatures and cancer-specific properties of MSI [17]. Furthermore, elevated MSI levels may occur via independent mechanisms [32]. In cases with immunohistochemical MMR protein loss, either WES or TSO 500 determined the percentage of unstable MS sites retrospectively. While in colorectal, esophageal, and lung cancer, MSI-H is defined as > 10% unstable microsatellite sites [33], the present study provides the first evidence that the cut-off for MSI-H might require to be set significantly lower (> 5%) for pedHGG.

We are the first to describe MSH2 hypermethylation as an alternative epigenetic silencing of MSH2 and MSH6 expression in pedHGG. This alternative mechanism had not been determined by routine molecular diagnostics [40]. However, since MSH2 hypermethylation is reported to occur in 24% of MSH2-deficient colorectal carcinoma [29], testing for MSH2 gene promotor hypermethylation should always be performed when a loss of MSH2 and MSH6 protein expression is found by IHC, but no corresponding underlying genetic variant is detected by DNA panel sequencing.

Several molecular methods are available to investigate the underlying MMR gene alterations, including NGS, WES/WGS, copy-number variation (CNV) analysis, and pyrosequencing. However, no single assay can comprehensively detect all types of alterations. CNV analysis is suited for detecting amplifications and deletions, while NGS and WES/WGS are more adept at identifying point mutations, small insertions, and deletions. Pyrosequencing, on the other hand, is highly effective for detecting gene promoter hypermethylation. Therefore, to achieve comprehensive coverage of genetic mechanisms, a combination of molecular assays is necessary. For instance, NGS coupled with pyrosequencing could provide a robust approach. This multi-faceted approach underscores the importance of elaborating diagnostic workflows to maximize sensitivity and to ensure that no underlying genetic mechanism is overlooked (Fig. 5).

Despite the high PPV of the here evaluated screening method, in routine care, MMRD detection in the tumor does not necessarily result in germline testing and counseling of the family [32]. In case of pediatric oncological patients, the barriers for further germline testing might be clearly lower. All patients with MMRD detected in MMR-IHC in this present study underwent germline testing and counseling. This is especially important, since in four of seven patients with germline involvement, MMR-IHC results and subsequent germline testing led to the initial diagnosis of LS in the affected families.

Germline testing of the patients with loss of MMR-IHC identified a LS or CMMRD in 70% of MMRD cases. In these patients’ families, the diagnosis of a CPS enables predictive testing of all relatives at risk. Subsequently, intensive, life-saving surveillance programs can be offered to all variant carriers in the families to prevent cancer or to identify early stages with good prognosis [11, 16, 42].

Our sequencing revealed MSH2 as the most commonly altered gene which is in accordance with the reported findings by the IRRDC [12]. The present observation of the frequently co-altered genes ARTX, NF1, and TP53 is also consistent with the previous findings in CNS tumors with underlying CMMRD [12, 42].

Other than any molecular analysis, MMR-IHC can be easily, rapidly, and cost-effectively performed in neuropathology institutes world-wide [2, 3, 42]. In the present pedHGG series, MMR-IHC screening reduced the number of the cases in which testing for germline MMRD was required from 127 to 10 patients, i.e., to approximately 8%. Considering the 28 patients with non-performable/-evaluable MMR-IHC, the rate of recommended germline testing would still be only 25% of pedHGG patients.

However, technical limitations of routine IHC screening for MMRD in pedHGG patients must be taken into account. If the sample is composed predominantly of resident MMR protein-positive non-neoplastic glial cells or dense immune cell infiltration, loss of MMR protein expression in small pedHGG samples may be missed. Thus, personal experience and caution is required to select cases for germline testing. However, this also applies to the interpretation of results of molecular analyses, where quality may also be limited by low tumor cell count. Consequently, in cases with only very small tissue samples and/or low tumor cell density, germline testing should be recommended at a low threshold. This applies especially when they are diagnosed with IDH-mutant astrocytoma [41] or diffuse pediatric-type HGG, H3-wildtype and IDH-wildtype [40].

Difficult-to-reach tumors located in midline are frequently associated with very small biopsy samples. As none of the screened DMG in this cohort showed a loss of MMR-IHC, it may be assumed that there is no association between DMG and MMRD. This observation needs to be verified in larger patient cohorts but previous findings reporting predominant MMRD association with diffuse pediatric-type HGG, H3-wildtype and IDH-wildtype and IDH-mutant astrocytoma strongly support the absence of MMRD in DMG patients [40]. Furthermore, individual cases of H3 G34-mutant pedHGG with germline MMRD as found in the present series (ID 29) had been reported previously [26].

We additionally assessed potential survival differences between MMRD-associated pedHGG and pedHGG controls lacking MMRD. Despite small patient numbers, EFS and OS of IDH-mutant astrocytoma with loss of MMR-IHC were inferior compared to IDH-mutant astrocytoma controls without underlying MMRD confirming previously reported results [41].

Combined immunotherapy with nivolumab and ipilimumab provides a promising treatment option for patients with MMRD/MSI-H metastatic colorectal cancer [31]. There is evidence for a benefit from immune-directed/synergistic salvage therapies in Replication-Repair-Deficient HGG [9]. The present data may support a potential benefit of checkpoint inhibition in pedHGG MMRD patients although the patient numbers are very small, and patients received checkpoint inhibition only as second-line treatment approach. IDH inhibitors like ivosidenib and the particularly blood–brain barrier permeable vorasidenib may be a suitable combination partner for checkpoint inhibition to improve survival of MMRD patients with IDH-mutant astrocytoma [27].

As MMR-IHC demonstrated a sensitivity of 100% in detecting MMRD, germline genetic testing can be promptly initiated for pedHGG patients with loss of MMR protein expression (Fig. 5), as well as for their relatives at-risk, within 1–2 days. This rapid turnaround facilitates early identification of germline cancer predisposition syndromes and may influence clinical management. Our study may indicate that pedHGG MMRD patients may indeed benefit from checkpoint inhibitor therapy. However, further clinical studies are required to evaluate the efficacy of checkpoint inhibitor therapy as a first-line treatment in pedHGG patients. These future studies could potentially lead to the recommendation of checkpoint inhibitor therapy as a standard first-line option for all pedHGG patients with confirmed MMRD at diagnosis.

In conclusion, IHC represents an easy to perform, cost-effective and fast method to screen for germline MMRD in pedHGG with global applicability and both high sensitivity and specificity. We therefore recommend incorporating MMR-IHC into routine diagnostics of pedHGG. Positive results necessitate patient information about potential underlining CPS and, if desired, individual genetic testing and family counseling to offer life-saving surveillance measures to all germline variant carriers in the families. Furthermore, identified patients may benefit from checkpoint inhibition therapy even at first line.