Natural History and MRI Follow-up of Patients with Cerebral Cavernous Malformations

Background

Cerebral cavernous malformations (CCMs) are vascular malformations of the brain and spinal cord lined by endothelial cells and lacking smooth muscle. Due to the disruption of endothelial cell permeability and tight junctions, patients may present with focal neurological deficits, headaches, or seizures, with or without hemorrhage. In some cases, CCMs may exist as incidental findings. CCMs can be classified as sporadic (usually accompanied by a developmental venous anomaly, or DVA) or familial (without DVA and with multiple lesions).

Over the past 20 years, more than 25 natural history studies have attempted to define the symptomatic hemorrhage rate of CCMs. Early studies used retrospective methods, assuming these lesions had been present since birth, but these conclusions were inaccurate. Subsequently, several natural history studies based on prospective data calculations relied primarily on retrospectively collected data and were limited by the data collection methods available at the time. These early studies had inherent limitations due to incomplete medical records and a lack of imaging data. Furthermore, the lack of long-term follow-up and large patient samples added uncertainty. Three existing meta-analyses of natural history studies are commonly referred to by clinicians, but their conclusions are limited due to heterogeneity among the studies. Therefore, clinicians and patients also want to know if patients with two hemorrhages are more likely to have a third hemorrhage, but existing studies have paid little attention to this.

Source Information

This article was written by Kelly D. Flemming, MD, Robert D. Brown Jr., MD, MPH, Giuseppe Lanzino, MD, and others from the Mayo Clinic, and published in the Journal of Neurosurgery on May 24, 2024.

Study Objectives

This study aimed to report the natural history of a contemporary cohort of CCMs with prospective multi-source data collection, analyze the utility of follow-up MRI, and define the rates of first symptomatic hemorrhage (SH) and severe SH for CCMs through integrated electronic health records and detailed imaging review by investigators. It also aimed to evaluate the risk and timing of prospective secondary SH and identify risk factors for hemorrhage.

Methods

Study Design

The research team established a database in 2015 that included consenting adult patients with radiologically confirmed brain or spinal cord CCMs. The study excluded patients under 18 years old, postoperative patients with an alternative cause, and patients who underwent surgery within 3 months of diagnosis. Only patients with follow-up of more than one year were included, and cases of spinal cord CCMs and radiation-induced CCMs were excluded.

Data Collection and Processing

Initial Clinical Information

Detailed electronic medical records were reviewed, and patients were asked to complete an initial medical history questionnaire. Data collected included demographics, lesion type, comorbidities, and initial presentation. Patients were classified as having familial CCMs (confirmed by genetic testing or the presence of multiple lesions without DVA) or sporadic CCMs, and selective medication use was recorded.

Initial Imaging

Initial MRI images were reviewed by neuroradiologists and the first author (K.D.F.), measuring the maximum cross-sectional diameter on T2-weighted images and the presence of hemosiderin deposition within the lesions. The number of CCMs and the location of the symptomatic or largest lesion were recorded.

Prospective Follow-up Data Collection and Outcome Measures

SH events during follow-up were collected through electronic medical records, annual written questionnaires, and in-person follow-up assessments. Functional outcomes were assessed using the modified Rankin Scale (mRS) score. All follow-up MRI images obtained 3 months or more after consultation were reviewed.

Data Analysis

Descriptive statistics were used to analyze patient clinical and imaging characteristics, and Kaplan-Meier curves were used to analyze the risk of first and secondary prospective SH. Univariate and multivariate Cox proportional hazards models were used to analyze potential risk factors for hemorrhage.

Results

Patient and Lesion Characteristics

The study included 315 patients, 58.7% of whom were female, and 19.7% had familial CCMs. At diagnosis, 37.1% of patients had a ruptured CCM, and 28.9% of CCMs were located in the brainstem. The total follow-up time was 2679.6 patient-years, with a median follow-up duration of 6 years. During the follow-up period, 96 SH events occurred, with an annual SH rate of 4.6%. For ruptured CCMs, the 5-year cumulative SH rate was 41.2%, while for unruptured CCMs, it was 6.1%.

Risk Factor Analysis

Univariate analysis showed that ruptured status at diagnosis and persistent or new T1 hyperintense signal on follow-up MRI at 3 months were predictors of future SH. In multivariate analysis, after adjusting for age, sex, and CCM location, ruptured status at diagnosis and persistent or new hyperintense signal on follow-up MRI remained independent predictors.

Conclusions

In this contemporary CCM cohort, we found that the 5-year cumulative SH rate for ruptured CCMs was 41.2%, while for unruptured CCMs, it was 6.1%. Patients with a ruptured diagnosis and new T1 hyperintense signal on follow-up MRI had a higher risk of re-hemorrhage. Changes in MRI images during follow-up can serve as biomarkers of disease activity and provide important references for clinical management.

Study Significance

Through prospective multi-source data collection, this study demonstrates important details about the natural history of CCMs, providing valuable data for clinicians regarding the management of CCMs, particularly in assessing the risk of re-hemorrhage and formulating follow-up plans. The study also highlights the indispensable value of follow-up MRI examinations in the diagnosis and management of CCMs.