Complexity. Structural genomic variants or copy number variations (CNV) are known genetic risk factors for various epilepsy syndromes. In fact, CNVs might represent the single most studied type of genetic alterations across a very broad range of epilepsy syndromes. There is, however, a group of patients that is usually not investigated in genetic studies: patients with presumable lesional epilepsies or questionable findings on Magnetic Resonance Imaging (MRI). Many of these epilepsies are usually thought to be secondary to the identified lesion, and genetic risk factors are not considered. In a recent study in the European Journal of Human Genetics last week, we investigated the role of CNVs in a cohort of patients with complex epilepsy phenotypes that were not easily classified into existing categories. Many of patients included had definite or questionable findings on MRI. The results of our study made us wonder whether the boundary between lesional and genetic epilepsies needs to redrawn.
Pattern recognition. We usually don’t consider a genetic cause or a genetic contribution if we have a clear lesional cause for a patient’s epilepsy. A focal cortical dysplasia or a scar after a perinatal injury is usually considered sufficient as the underlying cause. While this distinction appears to be clear-cut at first glance, it sometimes blurs in clinical practice. In a patient with a relatively small lesion on MRI, but epilepsy with severe intellectual disability, do we still consider the lesion as causative? What if the MRI findings were ambiguous, showing gliosis or global atrophy? We would probably consider these findings unrelated, but would still feel hesitant to classify the patient’s epilepsy as non-lesional, idiopathic, or genetic. Initiating genetic diagnostic studies might not necessarily be the next logical step and accordingly, this subgroup of patients has not been included in genetic studies so far. Accordingly, these epilepsies represent “forgotten phenotypes” in genetic studies that we know very little about. We aimed to take a first step towards understanding the role of genetic risk factors by assessing the role of CNVs in these phenotypes.
No difference. To summarize our findings in a single sentence: there is no difference in the frequency of CNVs with regards to the MRI findings. We did not find significant differences in the frequency of CNVs in patients with our without findings on MRI when we assessed 196 patients with available MRI data. 82/196 patients had positive findings on MRI, including acquired (n=26), developmental (n=38) and unclassified (n=18) findings. The frequency of CNVs was not different, independent of whether we look at all rare CNVs, pathogenic CNVs, likely pathogenic CNVs, or CNVs of unknown significance. We had classified the CNVs according to established diagnostic criteria. While the study was not sufficiently powered to find small differences between subgroups, this was not the primary aim of our study. We simply wanted to assess whether it might make sense to screen patients with epilepsy and MRI findings for CNVs. We did not find a convincing argument to exclude these patients from CNV screening, which is further corroborated by the following examples.
Lesional turned genetic. A patient (NL4) with focal epilepsy and a right temporal cyst was found to have a de novo 17p11.2 deletion, compatible with Smith-Magenis-Syndrome. While cysts are not considered epileptogenic per se, they might have focal cortical dysplasia in the surrounding cortex that might be thought to be causative, particularly if the epileptic focus is located in the same region. In the above patient, the lesion was considered causative prior to CNV analysis. Likewise, a patient (D10) with hydrocephalus of unknown origin, assumed to be acquired, turned out to have a terminal deletion of chromosome 6. These findings are in line with prior studies on the role of CNVs in Temporal Lobe Epilepsy, which indicated that the presence or absence of a CNV is not predictive of outcome after temporal lobectomy. How would we handle this in clinical practice? One rule of thumb is to repeatedly ask the question whether a certain MRI finding fully explains the patient’s phenotype. If not, pursuing a genetic diagnosis might be worthwhile.
From EEGO to EJGH. The current study was a long-term project that was underway for almost five years. It initiated as a subproject of a rejected and long-forgotten EU grant on epileptic encephalopathies of genetic origin (EEGO), which resulted in the collaboration between the Dianalund/Copenhagen group and our group in Kiel. In the process of developing this project, we then joined forces with the Utrecht group who had a similar project at the time and eventually contributed the majority of the patients in this project. None of the data in this project was particularly “hot” at any point. Therefore, I think that this publication is a nice demonstration that genetic findings sometimes only make sense if the data is pooled. This way “data can speak”, and a more powerful study may emerge with relevant clinical implications. I would like to thank all my collaborators on this long-term project, which went through many different shapes and formats prior to reaching the final version that we submitted.