Article Text

Download PDFPDF

Subependymal nodules, giant cell astrocytomas and the tuberous sclerosis complex: a population-based study
  1. F J K O’Callaghan1,2,4,5,6,
  2. C N Martyn3,
  3. S Renowden4,
  4. M Noakes5,
  5. D Presdee5,
  6. J P Osborne2,5
  1. 1
    Bristol Royal Hospital for Children, Bristol, UK
  2. 2
    School for Health, University of Bath, Bath, UK
  3. 3
    MRC Environmental Epidemiology Unit, Southampton, UK
  4. 4
    Frenchay Hospital, Bristol, UK
  5. 5
    Royal United Hospital, Bath, UK
  6. 6
    Faculty of Medicine and Dentistry, University of Bristol, Bristol, UK
  1. Dr F J K O’Callaghan, Department of Paediatric Neurology, the Bristol Royal Hospital for Children, Institute of Child Health, Upper Maudlin Street, Bristol BS2 8BJ, UK; finbar.ocallaghan{at}


Objectives: (1) In a population-based study of tuberous sclerosis (TSC), to identify the number of patients presenting with symptomatic giant cell astrocytomas (GCAs); (2) within a subset of this population, to identify the number who would be diagnosed with GCAs on predetermined radiological criteria.

Methods: Patients with TSC in Wessex (a geographical region of England) were identified, and their medical history determined. A subset were invited to have a cranial MRI if they did not have a history of a symptomatic GCA and if they were likely to tolerate cranial imaging without a general anaesthetic. Scans were performed according to a standard protocol on a single scanner and were reported blindly by a neuroradiologist.

Results: 179 people were identified with TSC. Ten of these had a history of treatment for a symptomatic GCA. Forty-one of the remainder had a cranial MRI. Thirty-nine of these had subependymal nodules, of whom 24 (59%) had at least one (maximum 11) that showed enhancement with gadolinium. In seven (17%), the lesion was >1 cm, and all of these lesions showed gadolinium enhancement.

Conclusions: In this study, the proportion of patients with TSC who had a history of symptomatic GCA was 5.6%. In the subset without such a history, who underwent imaging, the number diagnosed as having a GCA on radiological criteria was much higher (59% gadolinium enhancement and 17% >1 cm in size). Screening for GCAs (performing scans on asymptomatic patients with TSC) would therefore identify large numbers of patients who had not presented with symptoms. This finding leads us to recommend that screening should not be undertaken.

View Full Text

Statistics from

What is already known on this topic

  • Subependymal nodules that line the lateral ventricles of the cerebral hemispheres are a common feature in patients with tuberous sclerosis complex (TSC).

  • Giant cell astrocytomas (GCAs), which probably develop from pre-existing subependymal nodules, can develop in patients with TSC. They are usually situated at the foramen of monro in the lateral ventricles and may cause signs and symptoms of raised intracranial pressure if they block the circulation of cerebrospinal fluid. Untreated they can cause blindness and death.

  • GCAs usually present in the second and third decade of life.

What this study adds

  • The history of symptomatic GCAs in a large population of patients with TSC is low (5.6%) but significant.

  • The prevalence of lesions on MRI scanning that satisfy published radiological criteria for the diagnosis of a GCA is high (17–59%).

  • Current radiological criteria for the diagnosis of a GCA are likely to identify many lesions that will not go on to cause symptoms.

The tuberous sclerosis complex (TSC) is a genetic disease characterised by the formation of hamartomas (benign tumours) throughout the body. The hamartomas are related to or cause the well-known clinical features of the disease, such as epilepsy, severe learning difficulties, behavioural disorders (particularly autism and hyperactivity) and skin (facial angiofibromatosis, ungual fibromas, forehead fibrous plaques, hypomelanic macules and shagreen patches), renal (angiomyolipomas) and cardiac (rhabdomyomas) lesions.14

Three types of hamartoma can occur in the brain. The cortical or subcortical tubers that gave the disease its name are usually present at birth. They are probably the cause of the symptomatic epilepsy seen in ∼75% of patients with TSC, and their number and location may be related to the cognitive and behavioural difficulties seen in TSC.5 Commonly, patients with TSC will also have lesions lining the ventricular system within the brain known as subependymal nodules. These small lesions are usually multiple and are not known to cause any clinical problems, although their numbers may correlate with the numbers of tubers and a patient’s cognitive level. Finally, there are subependymal giant cell astrocytomas (GCAs) which usually arise within the ventricular system, although they have been described elsewhere in the central nervous system.6 It is possible that they arise from pre-existing subependymal nodules, although this has not been proven conclusively. If they are large, they may block the ventricular drainage system and cause clinical symptoms of raised intracranial pressure.

Such symptomatic GCAs are a serious complication of TSC. Patients may suffer acutely or subtly from headaches, drowsiness, vomiting, disturbed vision and ataxia. If left untreated, these lesions lead to increasing debility, blindness and ultimately death. These symptoms can be difficult to recognise in patients with severe learning difficulties and behaviour disorders. The potential for poor outcome from these lesions has led to recommendations to use cranial imaging to help identify GCAs at a pre-symptomatic stage.711 If it were possible to identify asymptomatic lesions that would inevitably cause problems later, then such lesions could be removed electively. Histology is unhelpful in this situation. Under the microscope, GCAs appear identical with the subependymal nodules found in the same area and from which they are often thought to arise. Radiological criteria have been suggested to discriminate between a lesion that will continue to grow and cause symptoms and one that will not continue to grow or will not cause symptoms. Suggested radiological criteria for GCAs include arbitrary size criteria (eg, >0.5 cm or 1 cm in diameter), location within the ventricular system close to the foramen of Monro, the tendency to show enhancement with intravenous contrast media, or by demonstrable evidence of growth on serial scans.7 10 It is not known how many GCAs that are diagnosed radiologically in this way continue to grow and cause clinical problems. It is an important question, as some clinicians use radiological criteria to justify pre-emptive neurosurgical intervention and thereafter report a low recurrence rate and low morbidity.810 It is possible that these clinicians are appropriately intervening and electively removing lesions that would have progressed to cause significant problems. By removing them electively at an asymptomatic stage, they may have reduced the risk of possible complications in operations to remove symptomatic tumours. However, it may also be the case that they are removing lesions that would never have caused clinical problems, thereby subjecting patients to unnecessary neurosurgery.

The aims of this study were twofold. First, we wished to identify, within a large population-based study of TSC, the number of individuals who had a history of treatment for a symptomatic GCA. Second, we wanted to obtain magnetic resonance images of the brains of a subset of the TSC population who did not have a history of a symptomatic GCA and identify how many within the subset had predetermined radiological features of a GCA.


Using standard diagnostic criteria, we identified patients with TSC living in the Wessex region of England on 1 August 1998 through an appeal to specialists to whom such patients might be referred and to general practitioners and through a search of hospital discharge diagnoses. Figures from the Office for National Statistics show the population denominator to be 3 679 162, giving a prevalence of 4.9 per 100 000. A history of neurological symptoms or of surgery was obtained either by direct interview or from the general practitioners.

All patients identified with the potential to have cranial MRI without sedation or anaesthetic and who did not have a history of treatment for a GCA were invited to undergo MRI. All scans were performed at the Royal United Hospital, Bath using a Siemens Magnetom Impact Expert 1.0 T MRI scanner. All patients underwent the same imaging protocol:

  1. Spin echo T1 sagittal sequence using 5 mm slices

  2. Fast spin echo T2 axial sequence using 5 mm slices

  3. Fast spin echo dual echo coronal sequence using 5 mm slices

  4. Spin echo T1 axial sequence using 5 mm slices

  5. Fast FLAIR (fluid-attenuated inversion recovery) axial sequence using 5 mm slices

  6. MPRAGE (magnetization-prepared rapid gradient echo; ie, three-dimensional T1 sequence) coronal sequence using 3 mm slices

  7. Post-gadolinium T1 axial sequence using 5 mm slices

All scans were reported blind by a neuroradiologist (SR). The numbers, enhancement characteristics, site and size (whether >1 cm diameter) were recorded for all lesions lining the ventricular system using a standardised reporting format. On the basis of previous publications, our predetermined criteria for the radiological diagnosis of a possible GCA were enhancement and diameter >1 cm.


A total of 179 people with TSC were identified; 149 were seen by FJKO’C. Details of the 30 not seen were obtained from their general practitioners. Ten of the 179 (5.6%) had a history of treatment for a symptomatic GCA (table 1). Six were female, and seven of the 10 had an IQ in the normal range. Six of the seven patients with normal intelligence are known to have presented with classical symptoms and signs of raised intracranial pressure. In five of these seven patients, there had been no diagnosis of TSC before the development of the GCA. Three patients had severe learning difficulty, and raised intracranial pressure had caused blindness in one of them before the diagnosis of GCA was made. The remaining two presented with deterioration in behaviour. Ages ranged from 10 to 41 at presentation with the GCA.

Table 1 Characteristics of the 10 patients who had been treated previously for a giant cell astrocytoma (GCA)

Forty-one (24 women) of the 149 patients who agreed to be seen by the investigator also agreed to, and tolerated, a cranial MRI scan for research purposes. The age range was 9–75, and the distribution of ages was skewed to the right, with a median of 25 years (interquartile range 17–46). Twenty-six of the 41 (63%) had a positive history of epilepsy, and 11 (27%) had suffered from infantile spasms.

Thirty-nine of the 41 patients imaged (95%) had evidence of subependymal nodules. The number of nodules ranged from 0 to 24 (median 6, interquartile range 3–9). Twenty-four patients (59%) had at least one nodule that was enhanced after administration of intravenous gadolinium. Nineteen patients (46%) had more than one enhanced nodule, and one patient had 11 enhanced lesions. Seven patients (17%) had evidence of a subependymal lesion close to the foramen of Monro that had a diameter of >1 cm, and all of these were enhanced (table 2).

Table 2 Characteristics of the seven patients with lesions >1 cm in diameter


The proportion of people with TSC in Wessex who had a history of a symptomatic GCA was 5.6%. In 8 out of 10, the GCA presented symptomatically in the second decade of life. However, symptomatic lesions have been described in the neonatal period, and in two of our population they did not present until the fourth or fifth decade.12 Patients with normal IQ presented with classical signs and symptoms of raised intracranial pressure. In five out of seven patients with normal IQ who had symptomatic GCAs, the discovery of the GCA led to the initial diagnosis of TSC. Presumably because they were able to articulate their symptoms to their doctors, they were all treated successfully, without subsequent evidence of recurrence and without any evidence of significant visual deficit. Patients with learning difficulty, who were unable to articulate their symptoms, presented with behavioural deterioration or visual loss.

In this sample of patients who underwent MRI scanning, the prevalence of lesions that some would diagnose as GCAs on radiological grounds because of size or enhancement with gadolinium was 59% on the basis of enhancement and 17% on the basis of >1 cm in size. It is impossible to know how many of the lesions diagnosed radiologically as GCAs will progress to cause symptoms. It is unlikely that all lesions >1 cm in size will cause future problems given what we know about the prevalence of symptomatic GCAs in this population. It is even less likely that all the smaller gadolinium-enhanced lesions will cause problems as some authors have suggested.13 At least one other author has come to similar conclusions.14

We report a cross-sectional study and are therefore unable to make detailed comments about the future growth of the lesions identified. Previous investigators have suggested that interval growth might also predict future symptoms. However, for several reasons, we do not believe that interval growth is a reliable predictor of future symptoms. Firstly, too many adults in our study had lesions that must have previously shown interval growth because they were of sufficient size to have obstructed the foramen of Munro if they had been the same size in infancy. Secondly, we have evidence from our clinic population that large lesions may stop growing. One such patient was included in this study, and for 6 years from the age of 35 there had been no interval growth of a lesion of significant size situated at the foramen of Monro (figs 1 and 2). It may be that calcification overtook the lesion preventing further growth. Whatever the reason, this observation implies that notable growth of a lesion is not in itself sufficient to predict that it will cause symptoms.

Figure 1 CT scan of brain showing large gadolinium-enhanced lesion at the foramen of Monro (arrow). The patient had no symptoms related to this lesion.
Figure 2 MRI scan of the brain of same patient as in fig 1 showing the large gadolinium-enhanced lesion 6 years later. The lesion had not increased in size. The patient still had no symptoms related to this lesion.

In addition, it is not known how quickly a GCA can grow. At one TSC conference, a verbal report was made of a symptomatic GCA arising within 8 weeks of an MRI scan that showed no evidence of a GCA (M Gomez, Mayo Clinic, personal communication).

These findings have important implications for imaging patients with TSC who have no signs of raised intracranial pressure, or for implementing screening programmes aimed at identifying lesions before symptoms arise. To screen for a particular type of lesion, it is important to be able to recognise, at an asymptomatic stage, those lesions that are going to cause future problems.15 16 It is not yet possible to identify such lesions in TSC. This study illustrates that a small percentage of patients with TSC have a history of symptomatic GCAs, yet there are many patients—as in our asymptomatic subset—who will have lesions that are categorised by some as representing the radiological diagnosis of a GCA. We consider on the basis of our findings that prophylactic removal of either gadolinium-enhanced lesions or lesions >1 cm in diameter would result in many patients undergoing unnecessary cranial surgery. In addition, a significant number of people are likely to present with raised intracranial pressure before the diagnosis of TSC has been made, and these people would not be helped by a screening programme for those already diagnosed with TSC. This leaves us with a dilemma, as surgery in patients with signs of raised intracranial pressure is believed to carry increased risks. However, operating on lesions that will never cause symptoms is clearly inappropriate. Therefore, we do not recommend screening for GCAs. Patients with signs or symptoms that might be due to raised intracranial pressure should have a cranial scan urgently. MRI is the imaging modality of choice because radiation may increase the risk of DNA damage, resulting in new hamartoma formation. When a scan is performed for diagnostic purposes or for the evaluation of epilepsy surgery, and a large or enhanced lesion is found, we recommend that this should not be removed immediately unless there is evidence of raised intracranial pressure. The suggestion that surgery for these lesions would carry lower morbidity if performed before the onset of raised intracranial pressure has to be weighed against the fact that many such lesions may not need to be removed.

Finally, it cannot be overstressed that clinicians looking after patients with TSC who have learning difficulties should have a high index of suspicion that a GCA may be developing and causing problems if there are significant behavioural or clinical changes. We advocate that such patients are clinically assessed as soon as possible, including ophthalmological review and consideration of the need for cranial scanning. However, in our experience, behavioural change is usually due to other factors, and, if a GCA is responsible for the behavioural change, other symptoms or signs of raised intracranial pressure are normally present.


View Abstract


  • Funding: FJKO’C was supported by the Wellcome Trust. This study was also supported by the Bath Unit for Research in Paediatrics (BURP).

  • Competing interests: None.

  • Ethics approval: Ethics approval was obtained.

  • Patient consent: Obtained.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.