Genomic Diagnoses for Ectopic Intracerebral Calcifications

Discussion

We present the results of comprehensive genomic testing in patients with EICs regardless of family history. Our cohort consisted of 55% female and 45% male patients, which is consistent with previous studies reporting a higher prevalence of basal ganglia calcifications in female patients.¹⁴ The overall diagnostic yield of genomic testing for EICs was 34% in our study. 46% (7 of 15) of patients who received a diagnosis developed symptoms in adulthood. This mirrors the overall cohort where 47% of patients presented after age 18 years and suggests that age at symptom onset does not predict pretest probability of having an underlying genetic etiology for EICs.

Our cohort represents individuals who have features consistent with a rare disease and sought extensive medical evaluation and thus is not reflective of the general population of patients with EICs. Because a significant portion of patients with EICs are asymptomatic and therefore may not ever seek medical attention, our results likely overestimate the diagnostic yield of genomic testing on the general EIC population after secondary EIC has been ruled out. In addition, those who are referred have typically undergone extensive prior testing for both acquired and genetic etiologies, which implies an above average level of access to care as well as medical literacy. There are several reasons patients with known EIC genes may not have been identified in our cohort before inclusion in this study. First, genetic testing is currently not standard of care in all cases of EICs. Second, the clinical significance of EICs is not well understood; the phenotypes associated with this finding range from asymptomatic to more prototypical neurologic syndromes such as neuropsychiatric syndromes or movement disorders. As such, brain imaging may not have been indicated in all cases before research evaluation or EICs may have been seen as an incidental finding in some of our participants rather than an anchor for diagnosis. Finally, clinical genomic analysis pipelines are commonly phenotype rather than brain imaging-driven and EIC-causing variants may not have been considered or reported by previous testing laboratories because of the lack of clear clinical correlation between EICs and the patient's presenting symptoms.

Of the 9 patients with an affected first-degree relative, 3 (33%) had one of the 4 canonical familial calcification genes, which is lower than the 55% yield in previous published reports.¹⁰ This may reflect ascertainment bias in our cohort because patients with EICs are more likely to be tested for these 4 genes, and positive cases would be less likely to be referred to the UDN. Our most common pathologic finding in this cohort was SLC20A2, which was previously reported to account for around 40% of all primary familial brain calcification cases⁹ and was present in 4 of our familial cases in this cohort. Of the remaining 2, one received a diagnosis of XMEN (MIM 300853) and the other received a diagnosis of Aicardi-Goutieres syndrome type 7 (MIM 610181). While sequencing of additional family members, especially parents, has previously been shown to increase diagnostic yield of genomic testing, it did not substantially change the diagnostic yield in this cohort in the absence of a positive family history. This likely reflects a large contribution of familial rather than de novo variations toward disease prevalence.

Nine of 33 (27%) in sporadic cases of EIC in our cohort received a diagnosis. Three cases, 2 with SLC20A2 and 1 with PDGFB, had classic primary familial brain calcification genes. The remaining patients received diagnosis of Labrune syndrome (MIM 614561), Aicardi-Goutieres syndrome type 7 (MIM 615846), PKAN (MIM 234200), or leukoencephalopathy with spheroids 2 (MIM 221820). Although some of these patients had nonspecific findings unrelated to calcifications such as growth retardation in one patient with Aicardi-Goutieres syndrome and leukoencephalopathy in one patient with Labrune syndrome, many of these cases did not exhibit classic signs associated with their disease such as cysts, spheroids, or clear iron deposition which likely contributed to their referral to the UDN. These cases were unlikely to have been diagnosed without genomic testing, highlighting the heterogeneity of disease presentations and potential pitfalls of relying on clinical findings to perform targeted genetic testing. Notably, given that SNORD118 is a noncoding RNA, this may be missed on exome sequencing but would be detected on genome sequencing.

Among the 29 cases with EIC who did not receive a genomic diagnosis, 4 were found to be heterozygous for a possibly causative variant in a canonical gene associated with familial brain calcifications. Further research is necessary to determine whether these variants are causative. In addition, 3 more individuals were found to have pathogenic variants that explained their clinical presentation but not their EICs. One patient with bilateral basal ganglia calcifications and childhood-onset dystonia was found to be heterozygous for the known pathogenic deletion in TOR1A (MIM 605204) causative of DYT1. By definition, DYT1 primary dystonia is not expected to have structural brain abnormalities, perhaps suggesting this patient might have a blended phenotype with a yet uncharacterized cause of EICs. Two others had complex clinical presentations, one with pathogenic variants in GBA (MIM 606463) and one with potentially causative variants in COLQ (MIM 603033) which is undergoing further evaluation. These diagnoses would explain some, but not all, of the respective probands' phenotypes. For these cases, ambiguity remains if EICs are a rare feature of these disorders or if these individuals have blended phenotypes from more than one condition and/or gene-gene interactions which might result in EICs.

Of the 39 individuals in our cohort who had bilateral calcifications in the basal ganglia consistent with the classic description “idiopathic basal ganglia calcification” (MIM PS213600), 12 received a diagnosis on genomic testing of which 6 had a pathogenic change in a gene associated with this phenotype. Of the remaining 5 cases that did not have basal ganglia involvement, 3 received a genomic diagnosis. None of the 3 were found to have one of the canonical genes which supports the inclusion of basal ganglia involvement as a criterion for this entity. There were no clear correlations between our genetic findings with clinical symptoms or age at presentation; however, because of the nature of the UDN, atypical presentations are likely overrepresented in our cohort, and asymptomatic forms of IBG are excluded by the nature of the UDP program.

The causative genes we identified are involved in a variety of physiologic and cellular processes including ribosomal RNA processing, prostaglandin transport, magnesium transport, platelet and vascular endothelial growth, coenzyme A biosynthesis, DNA replication, and innate immunity. The wide range of functions of pathogenic changes that can lead to EICs points to this phenotype being at the convergence of disturbances in multiple pathways. More research is needed to understand how these processes interact and contribute to abnormal calcium deposition.

In conclusion, we present genomic, clinical, and imaging findings in 44 patients who had EICs and did not have a diagnosis after extensive routine workup. Extrapyramidal movement disorders were the most common clinical finding, and the basal ganglia was the most common site of calcification on imaging. We identified a genetic diagnosis through genomic testing in 25% of sporadic cases and 58% of cases with a similarly affected first-degree relative (dominant EIC). Because our cohort is not representative of all patients with EICs, more research needs to be done to determine the clinical utility of genetic testing in EICs and the generalizability of our findings. Our findings suggest that genetic testing should be considered for patients with neurologic or nonspecific neuropsychiatric symptoms along with EICs. Most of our solved cases were not due to one of the 4 genes associated with PFBC, even when the classically described basal ganglia calcifications were present. These findings support the use of genomic testing rather than targeted or panel testing. The presence of Labrune syndrome, caused by a small nonprotein coding element, in our cohort supports the use of genome rather than exome sequencing.