Unilateral hippocampal wasting after combined-modality therapy for glioblastoma

Introduction

Glioblastoma (GBM), the most common primary brain tumor in adults, is standardly treated with a multimodal approach, including surgery, followed by concurrent radiotherapy (RT) and chemotherapy [1]. Adjuvant therapies can extend survival but are associated with brain toxicity [2]. Recent studies have also documented hippocampal damage associated with both systemic therapy and radiotherapy [3,4]. We previously measured hippocampal volume after chemoradiation and found a pattern of radiation dose-dependent atrophy [5]. Here, we present a case of extraordinarily severe, unilateral hippocampal volume loss after treatment for GBM.

Case Report

A 24-year-old Caucasian man with no history of cancer, seizures, or neurologic disorder presented with new-onset weakness of the left upper extremity and was found to have a right parietal brain mass (Figure 1). He underwent gross total resection, and pathologic examination diagnosed glioblastoma with methylated MGMT. He was started on routine post-operative seizure prophylaxis (levetiracetam), which was discontinued prior to the start of radiation, and discharged home after unremarkable surgical recovery. His Karnofsky Performance Status was 90. He also had headaches and a pseudomeningocele at the surgery site.

Figure 1.

Pre-surgery T1-weighted MRI with IV contrast. A heterogeneously enhancing mass is visible in the right parietal lobe but does not appear to involve the hippocampus.

Six weeks after surgery he began chemoradiation, 60 Gy in 30 fractions with daily temozolomide 75 mg/m2. A pre-RT MRI did not show hippocampal abnormality. After two weeks of chemoradiation he had his first-ever seizure, partial with secondary generalization. He started lacosamide and continued chemoradiation. Three weeks later he presented with slurred speech and hydrocephalus associated with re-accumulation of his pseudomeningocele. An external ventricular drain was placed, followed by a ventriculo-peritoneal shunt. This resulted in a one-week break from chemoradiation, the only interruption.

Adjuvant temozolomide was given in five 28-day cycles, with daily doses of 200 mg/m² for cycles 2–3, and 150 mg/m² for the rest. Notable events during this period include one generalized, tonic-clonic seizure one week after chemoradiation, in the context of possible ventriculo-peritoneal shunt malfunction and noncompliance with lacosamide.

He was then seizure-free for four months, followed by two partial seizures associated with aura (bad taste/smell and abnormal sensation in left upper extremity), prompting addition of topiramate. He was again seizure-free for several months before suffering an episode of focal status epilepticus; topiramate was increased and phenytoin started. This regimen was efficacious at controlling seizures over the next 9 months, though he did continue to have intermittent aura and also had one seizure after decreasing his topiramate dose. In total, during the 20 months after his initial diagnosis, he had three generalized seizures and one episode of focal status epilepticus. Two of the three generalized seizures occurred during non-compliance with prescribed anti-epileptics. Other medications during this period included sertraline for mood disorder, medical marijuana, and cannabis oil.

The patient underwent numerous brain MRI studies. Following treatment, surgical and radiation changes are noted, including an area of enhancement consistent in appearance and clinical evolution with radiation necrosis. There was no apparent recurrence of tumor until approximately 18 months after initial resection, when a 6 mm enhancing focus was detected in the splenium of the corpus callosum. The new lesion continued to grow, and he subsequently underwent salvage therapy.

Also observed on MRI was development of remarkable hippocampal asymmetry. While the immediate pre-treatment images show symmetric, normal-appearing hippocampi, by 10 months post-surgery (6 months after chemoradiation) the right hippocampus was severely atrophied (Figure 2). This atrophy remained stable through the time of tumor recurrence at 18 months post-surgery.

Figure 2.

NeuroQuant segmentation of MRI with hippocampi in yellow(10). An Ommaya reservoir caused the artifact at 19.2 months. Right and left hippocampal volume percent change (compared to pre-chemoradiation) and mean radiation dose also shown.

Discussion

There are no reports in the literature of such severe hippocampal atrophy after standard treatment for GBM or brain metastases. A recent study of 52 patients with primary brain tumors, most of whom underwent surgery and chemoradiation, reported a mean hippocampal volume loss of 6% after RT doses >40 Gy, and the worst atrophy seen in the cohort was 20% [5]. Even in Alzheimer’s disease, the yearly atrophy rate is approximately 4%, compared to <1% in normal aging [6]. The current patient’s right hippocampus shrunk by 45% in less than one year.

This is also the first presentation of a detailed time course of hippocampal volume changes after treatment for GBM (Figure 3). Volume loss in the right hippocampus began shortly after completing chemoradiation and stabilized within approximately 6 months. Conversely, the contralateral hippocampal volume remained relatively stable at all time points.

Figure 3.

Hippocampal volume over time. Measurements made using NeuroQuant (CorTechs, Inc., La Jolla, CA)(10).

Medical records indicate the patient complained of memory loss during this period. Unfortunately, no formal cognitive testing was performed. A link has not been formally established between hippocampal atrophy and cognitive decline in brain tumor patients, though it has in other neurologic conditions [7,8].

There are numerous potential contributors to this patient’s hippocampal atrophy, and it is unclear to what degree each etiology led to the outcome. The surgical resection, while not violating the hippocampus itself, may theoretically have involved damage to afferent and/or efferent hippocampal pathways, leading to Wallerian degeneration or denervation injury. There may have been some microvascular ischemia during surgery, though there was no evidence of infarct on imaging. The patient began having seizures during chemoradiation, possibly themselves resulting from tumor or treatment effects. The episodes were not frequent or unusually severe, but epilepsy patients frequently develop unilateral hippocampal atrophy [9], and there is some evidence that minor structural changes can occur following even a single seizure [10]. The patient was also taking multiple anti-epileptic medications throughout treatment, which may possibly potentiate hippocampal changes. Additionally, he may have had a predisposition to hippocampal sclerosis at baseline, unmasked by GBM treatment. Finally, chemoradiation has been associated with radiation dose-dependent hippocampal atrophy, and his mean hippocampal dose on the affected side was over 50 Gy. A radiation dose-volume histogram for the hippocampi is shown in Figure 4. The severity of the outcome may be evidence of synergistic interactions between some of these possible etiologies.

Figure 4.

(A) Radiation dose overlaid on MRI at time of RT planning. (B) Radiation dose-volume histogram for hippocampi.

The prevalence and precise clinical implications of severe hippocampal atrophy in brain tumor patients (GBM, as in this case, or brain metastases) has not been described. This case report serves as the first formal documentation of this striking outcome and to increase awareness among clinicians of severe hippocampal volume changes possible after combined-modality brain tumor treatment.