Our analysis suggests that our institutional policy in the management of pediatric CM results in improved long-term outcome in these patients. Our data indicates that: i) seizure control requires extended resection of the cavernomas beyond pure lesionectomy, ii) it is safe to manage asymptomatic lobar and small (< 1cm) deep CMs conservatively as no patient in our study experienced a severe event, and iii) less aggressive surgery in deep CM may require a repeat of surgery due to re-hemorrhage and symptom recurrence, but an excellent long-term neurological outcome can be expected. Based on this experience we propose a treatment algorithm to guide clinical decision-making in the treatment of children with CM.
Even though there are some reports that are comparable to our cohort in terms of age, gender distribution, clinical presentation, and number of lesions, our data on the conservatively treated children is unique [1, 8, 24, 31]. To the best of our knowledge, no study to date has included a conservative group as a comparison for the outcome analysis. Such an analysis, however, is essential to understand the natural history of these lesions and enables adequate treatment allocation.
Since baseline characteristics differ significantly between groups, the goal of comparing surgically and non-surgically treated groups is not to determine whether either of these strategies leads to better results, but to compare the postoperative clinical progress to the natural course of the disease. According to our analysis, mRS scores significantly improved from time of diagnosis to time of last follow-up in both treatment groups. Furthermore, the difference in median mRS at the time of diagnosis between groups dissolved during the observation period. This finding not only indicates that surgery in patients assigned to an intervention is beneficial but also confirms the adequate selection for conservative treatment at our institution.
Lobar lesions
CMs with lobar locality predominately present with seizures and surgery aims to achieve seizure freedom in case of pharmaco-resistance. Only severe hemorrhage due to a cavernoma necessitates urgent surgical removal [21]. In our cohort, 72% of patients had seizures at the time of diagnosis and seizure freedom was achieved in 86% of these patients. However, 2 patients required a second surgery due to seizure recurrence after initial postoperative seizure freedom for 1 and 3.5 years, respectively. Both underwent second epilepsy surgery and were seizure-free thereafter. In both cases, the resected CM was located adjacent to an eloquent region (motor cortex and visual cortex) and aggressive resection was avoided initially. Since no permanent deficits were encountered after the second more extended surgery, an initially more aggressive approach may have been justified in these cases, which may have avoided seizure recurrence and the need for a second surgery. Therefore, a more restrained approach may not be successful if patients with lobar CM and epilepsy are allocated to a surgical treatment. Furthermore, a conservative treatment may lead to a good seizure control at a first glance, but recurrence and new onset seizures may be encountered during the observation period. In our study population, one patient encountered seizure recurrence one year after cessation of AEDs and one additional patient experienced new-onset seizures during the observation period. In both cases AEDs were (re)established.
The postoperative seizure rates in pediatric patients with CMs reported in previous studies are diverse. For instance, a multi-center study conducted by Hugelshofer et al. that included 41 pediatric patients with CM and seizures found postoperative rates of 72% Engel I and 11% Engel II [12]. Sawarkar et al. reported seizure outcomes of 94.1% Engel I and 5.9% Engel II in a cohort including 19 patients with seizures [24]. Whereas Hugelshofer et al. reported gross total resection including the hemosiderin rim in all patients with seizures, Sawakar et al. reported resection of the hemosiderin rim only in 41% of the cases. Other studies reported 100% postoperative seizure-free rates after CM resection in pediatric patients. However, those studies by Xia et al. and Di Rocco et al. did not use a formal seizure classification, nor did they define the term “seizure-free” [31, 8]. Despite the limited number of studies that assessed seizure outcome after CM resection in pediatric patients, numerous studies focusing on adult CM patients reported findings similar to our results, suggesting postoperative seizure-free rates of approximately 75% [23]. Thus, differences in reported rates of seizure-free patients may depend on other factors than extent of resection of hemosiderin and may be explained by differences in time of follow-up or different policies in the use of AEDs.
This excellent epilepsy outcome in surgically treated patients with lobar CM can be discussed in the context of the risk for hemorrhage and operative complication rate. In our study population, only one patient in the conservative group suffered from symptomatic hemorrhage. Thus, the observed annual hemorrhage rate in the subgroup of patients with lobar lesions was 0.4%. Annual hemorrhage rates of pediatric lobar CMs reported in the literature range between 0.5% and 2.1% [10]. Regarding the operative morbidity, we encountered no new permanent deficit with an overall complication rate of 8%. These findings are in line with previous reports and further support the utilization of our approach.[8, 12] The case of a patient who suffered from seizures and underwent resection of a lobar CM with an extensive hemosiderin deposit is described below.
Case presentation
In a male 1.5-year-old patient who was exhibiting seizures, a large right occipital CM was resected adjacent to the area of the visual cortex (Fig. 6A). As shown in Fig. 6B, resection was restricted to the CM itself without radical removal of the hemosiderin rim. After surgery, the patient was seizure-free with seizure recurrence observed 3.5 years after the surgical intervention. Follow-up MRI showed persisting hemosiderin deposits in the cortex around the resection cavity (Fig. 6C). Video EEG monitoring confirmed localization of the epileptogenic zone in the field of resection. After extension of the initial resection, which included all hemosiderin remnants, the patient achieved seizure-freedom (Fig. 6D).
Lesions of the brainstem and basal ganglia
The current study assessed 18 children with CMs of the brainstem or basal ganglia with the majority (78%) receiving surgery. The decision for a conservative treatment in the remaining 4 children was based on the low symptom burden, small lesion size, and difficult surgical accessibility of the lesion. No hemorrhage was seen in any of the conservatively managed patients (median follow-up time: 59.5 months). Although deep location is reported to be a risk factor for hemorrhage, lesion size is likely to be crucial [10, 21]. According to a study by Li et al., which included 85 pediatric patients with untreated brainstem CMs, a lesion diameter of > 2 cm is associated with higher hemorrhage risk [16]. An association between size and hemorrhage risk has also been reported by Kupersmith et al. in a series of 37 adult brainstem CMs with a cut-off value of 10 mm [14]. In our cohort, conservatively managed CMs had a median diameter of 0.8 cm (0.6–0.9 cm) compared with a median diameter of 2.4 cm (1.0-4.2 cm) in surgically treated patients. Consequently, the observed benign course of the conservative group may be attributable to the significantly smaller lesions size.
The surgically treated group of patients was characterized by neurological impairment at the time of presentation. Hemiparesis was the leading neurological deficit followed by other symptoms such as reduced state of consciousness and cranial nerve palsies. Significant improvement of the median mRS score from 4 to 1 in the surgery group confirms the substantial benefit from surgery in this group. Gross-total resection after last surgery was achieved in 64% of patients with no newly encountered permanent neurological deficit. This is in marked contrast to other reports. Abla et al, for instance, reported a rate of 25% of permanent new deficits after surgery in 40 pediatric patients with brainstem CMs [1]. These findings are in line with a study by Li et al, who reported 21.2% permanent new deficits in a cohort of 52 children with brainstem CMs [17]. The discrepancy between our and previous findings is likely due to the fundamentally different treatment approach. While it has been our policy to be less aggressive, leaving residual disease in place before putting the patients at risk for permanent neurological compromise, and to re-operate in case of recurrence, others share the opinion that total removal is necessary irrespective of the potential morbidity during surgery [5, 17]. These two treatment concepts can be discussed under various aspects; the fact, however, that no child with deep CM in our study population experienced neurological compromise due to re-hemorrhage in residual disease or multiple surgeries strongly favors the less aggressive approach utilized in this study. This difference in conceptualizing surgery for deep CM is confirmed when rates of total resection are compared. Li et al. who employed a more aggressive approach reported a gross-total resection rate of 94.2% compared with 69% gross-total resection rate after last intervention in our study [17]. Although Abla et al. did not provide information on gross-total resection rates, the difference in surgical approaches may be extrapolated by comparing the numbers of procedures with our cohort. Only 5% of patients underwent re-operation in the cohort assessed by Abla and colleagues, whereas almost half of the patients in our study population received at least two surgical procedures.[1] The main argument for an aggressive approach has been advocated to be the potential for re-hemorrhage. According to the study by Li et al., the annual hemorrhage rate declined from 12.3–0.5% after surgery [17]. However, Abla et al. reported an annual postoperative hemorrhage rate of 5.25% versus 6.2% revealed during follow-up in the current analysis [1]. Applying these findings to a corresponding clinical situation in which patients, parents, and surgeons are diversely involved with regard to their understanding of the medical issue, the decision about the intended extent of resection must be an individual and very personalized one. The data in our study suggest that a cautious approach, not primarily aiming at gross total resection, results in a clear benefit for patients with a relatively small complication rate and no permanent new neurological deficit. Our policy is further supported by the exemplary case described below.
Case presentation
A 3-year-old female patient with symptomatic hemorrhage of a large CM located in the pons (Fig. 7A) presented with VI nerve palsy and mild hemiparesis. After hematoma evacuation and partial resection of the CM via a retrosigmoidal approach (Fig. 7B), the patient recovered and developed well. One year later, asymptomatic progression was seen during a routine follow-up MRI (Fig. 7C). For this reason, an elective surgery was performed, using an approach via the telovelar route. Given that no safe corridor could be identified based on the findings from neuromonitoring without putting the geniculum of the facial nerve or the medial lemniscus pathway at risk, no resection was performed. Although the lesion was still evident during a follow-up MRI almost four years later (Fig. 7D), the patient experienced no further neurological symptoms since the last surgical intervention.