JLE

Epileptic Disorders

MENU

Type I focal cortical dysplasia: surgical outcome is related to histopathology Volume 12, issue 3, September 2010

Figures

See all figures

Auteur(s) : Laura Tassi1, Rita Garbelli2, Nadia Colombo3, Manuela Bramerio4, Giorgio Lo Russo1, Francesco Deleo2, Gloria Milesi2, Roberto Spreafico2

1C. Munari Epilepsy Surgery Centre
2Division of Epilepsy Clinic and Experimental Neurophysiology, C. Besta Neurological Institute IRCCS Foundation, Milan
3Department of Neuroradiology, C. Besta Neurological Institute IRCCS Foundation, Milan
4Department of Pathology, Niguarda Hospital, C. Besta Neurological Institute IRCCS Foundation, Milan, Italy

Article reçu le 9 Mars 2010, accepté le 16 Juin 2010

The distinctive focal structural anomalies of the cortex were first described by Taylor et al. (1971) which were referred to as “focal cortical dysplasia” (FCD). However, since then, the term has been extensively used in the literature to refer to a wide spectrum of abnormalities of cortical anatomy. The true prevalence of FCD is unknown, and is probably underestimated in studies based on imaging techniques alone due to the presence of false-negative magnetic resonance imaging (MRI) and the wide range of techniques used. However, FCD accounts for a high percentage of histologically proven findings, ranging from 20% in adults to 50% in paediatric patients undergoing epilepsy surgery (Bast et al., 2006; Luders and Schuele, 2006).

Although a number of classification methods for these complex structural abnormalities have been proposed (Kuzniecky et al., 1991; Palmini et al., 1994; Barkovich and Kuzniecky, 1996), none is considered satisfactory. Barkovich et al. (2005) have recently suggested a new classification of different types of malformations of cortical development which is mainly based on MRI findings and genetic data. This classification includes the various subtypes of FCD in different subgroups based on the assumption that they are due to impairment of ontogenetic processes which take place at different stages of embryonic development.

With the aim of developing a more clinically useful classification of FCD, and in order to assess whether the neuropathologically defined subgroups corresponded to clinically homogeneous groups, Tassi et al. (2002) re-examined histological specimens from a series of patients who were surgically treated for intractable epilepsy. The results were correlated with electroclinical findings, MRI data and post-surgical outcome. This classification was further refined by Palmini et al. (2004), who proposed two main forms of FCD type I, characterised by laminar disruption alone with no major cytological abnormalities (type IA) and the presence of hypertrophic pyramidal-like neurons outside layer V associated with cortical disorganisation (type IB). A second main group, which included cases similar to those originally recognised by Taylor et al. (1971), was also characterized by profound cortical disruption with dysmorphic neurons (type IIA) and balloon cells (type IIB). Type II FCDs are considered to be highly epileptogenic and are the easiest form to detect by means of MRI (Colombo et al., 2003; Krsek et al., 2008).

Although many papers on FCD have been published over the last few years, there is still debate concerning the factors that influence surgical outcome. Some authors have found that a favourable surgical outcome is associated with the presence of severe histopathological features (Tassi et al., 2002; Chung et al., 2005; Kim et al., 2009), whereas others have reported contradictory results (Fauser et al., 2004). Furthermore, there is evidence that patients surgically treated for intractable epilepsy associated with FCD have a worse outcome than those undergoing surgery for other focal lesional epilepsy syndromes (Yun et al., 2006). Improvements in neuropathological diagnostics and the publication of some recent molecular biology studies have clearly shown a good clinical correlation between type II FCD and post-surgical outcome, however, the same cannot be said for type I.

Neuropathological examinations of surgical specimens are essential for classifying the variants of FCD and may help to construct a classification of cortical dysplasias; however, findings need to be correlated with pre-surgical electroclinical data and, above all, post- surgical outcome (Blümcke et al., 2009). The aim of this retrospective study was to investigate correlations between neuropathologically identified type I FCD (isolated or associated with other lesions) and the pre-surgical data and post-surgical outcome of a large cohort of patients surgically treated for intractable epilepsy.

Methods

We retrospectively evaluated 215 of 784 (27%) consecutive patients who underwent surgery for intractable epilepsy between May 1996 and May 2007 at the C. Munari Epilepsy Surgery Centre in Milan, Italy. The patients were selected on the basis of a proven histopathological diagnosis of type I FCD, whether isolated or associated with other lesions, and regardless of site. The lesions were classified and then related to the clinical findings, history of risk factors, MRI data, and post-surgical outcome.

All of the patients underwent an extensive pre-surgical evaluation which included the recording of: a detailed personal and family history, age at seizure onset, type of presentation, and the characteristics and frequency of seizures. The clinical records of each patient were carefully reviewed for these data, and only precise descriptions of pathological events were accepted, particularly in the case of febrile seizures. Information was not recorded when the source of the information or the description of the anamnestic data was considered to be unreliable.

All of the patients also underwent a neurological examination, a complete neuropsychological work-up, MRI, and comprehensive electroencephalographic (EEG) or video-EEG (VEEG) examinations (or both). When it was not possible to identify the epileptogenic zone with sufficient precision using electroclinical data and MRI, pre-surgical invasive stereo-EEG (SEEG) was performed (Cossu et al., 2005).

MRI

Pre-operative MR images of all of the patients were obtained using a 1.5 T machine (Philips ACS-III-NT). The MRI protocol (Colombo et al., 2003) included the following transverse double-echo spin-echo (SE) sequences of the entire brain; T2-weighted (w) coronal turbo spin-echo (TSE) T2-WI; coronal TSE fluid-attenuated inversion-recovery (FLAIR) T2-WI; and coronal inversion recovery (IR) T1-WI. Additional FLAIR T2-WI was obtained in the sagittal plane. In patients with temporal lobe epilepsy, images parallel (transverse) and perpendicular (coronal) to the major hippocampal axis were acquired. Intravenous contrast was injected when a tumoural lesion was suspected. The MRI data were retrospectively reviewed and an assessment was made of all anatomical abnormalities and focal lesions (e.g. tumours, vascular malformations and malformations of cortical development), gyration anomalies, focal thickenings of the cortex, blurring of the grey-white matter junction, abnormal signal intensity in the cortex and white matter, and atrophic changes. Hippocampal alterations were assessed visually with attention to atrophy, increased signals on T2-WI, decreased signals on T1-WI, and the loss of definition of internal structures.

Stereo-EEG procedure

Ninety patients (42%) underwent surgery after invasive SEEG; however the number of patients undergoing SEEG progressively decreased as less invasive approaches became more refined.

SEEG recordings were considered mandatory in patients when determination of the epileptogenic zone was not considered to be sufficiently reliable by means of video-EEG recording, MRI and based on the clinical aspects of the seizures.

The presurgical invasive recording and final surgical strategy were planned only after informed consent had been obtained from the patient or the parents of under-aged or intellectually impaired subjects.

SEEG was tailored to individual anatomical and electroclinical characteristics following the procedure described by Talairach and Bancaud (1966), and later refined by Munari et al. (1985) and Cossu et al. (2005).

Surgery

After obtaining informed consent, surgical resection was performed for strictly therapeutic reasons aimed at removing the cortical areas involved in seizure generation. The objective was to remove both the anatomical lesion (when identified) and the cortical epileptogenic zone identified by the electroclinical and MRI investigations, however, in some patients, an incomplete resection was made in order to avoid damaging nearby critical structures. When MRI was uninformative, the resection was planned on the basis of the electroclinical data. The extent of the resection was planned pre-operatively considering the severity of epilepsy and the risk of post-surgical deficits in each case.

Outcome was determined on the basis of Engel's classification (Engel, 1987), and all of the patients were followed for at least two years. Particular attention was paid to patients in Class Ia and Ic (who were completely seizure-free) with respect to the entire Class I [including class Ib; patients with auras, and class Id; patients with seizures after AED (antiepileptic drug) withdrawal]. AEDs were withdrawn only after a 12-month seizure-free period for class I patients. The reported follow-up and AED data were collected during visits or by telephone.

Histopathological methods

The surgical specimens were fixed in 10% neutral buffered formalin, embedded in paraffin, and routinely processed. Serial 7 μm sections were stained with haematoxylin and eosin, thionin, Luxol fast blue or Bielschowsky; other sections were immunostained using antibodies against glial fibrillary acid protein (GFAP, Boehringer Mannheim, Germany), neurofilaments (2F11 monoclonal, DAKO, Germany), microtubule associated protein (MAP2, Boehringer Mannheim, Germany), and neuron-specific nuclear protein (NeuN, Chemicon International, Temecula, CA). The slides were independently reviewed by two neuropathologists, one of whom had not been involved in the initial diagnoses and was unaware of the electroclinical data, MRI findings and surgical outcome. Any disagreements were discussed until an agreed diagnosis was reached.

The following criteria were used for making the final diagnosis:

- tumours were histopathologically typed according to the World Health Organisation (WHO) Classification of Tumours (Louis et al., 2007);

- hippocampal sclerosis was diagnosed in the presence of diffuse gliosis associated with pyramidal cell loss in the CA1, CA3 and CA4 (hilus) sectors of Ammon's horn (CA) (for references see Blümcke et al., 2007);

- the diagnosis of FCD was based on the two main morphological types described in the classification by Tassi et al. (2002) and refined by Palmini et al. (2004): type I and type II (Taylor-type FCD). For this study, we only considered patients with histopathologically proven type I FCD diagnosed on the basis of cortical disruption. The majority of the patients showed laminar (horizontal) alterations particularly affecting cortical layer II or IV; however, columnar (vertical) abnormalities were also observed in some (Blümcke et al., 2009; Tassi et al., 2009). The presence of numerous heterotopic neurons in the white matter was also considered when associated with cortical alterations. The patients with these histopathological characteristics were considered to have type IA FCD (otherwise called architectural dysplasia); the patients also presenting the hypertrophic pyramidal-like neurons described by Tassi et al. (2002) were diagnosed as type IB FCD (cytoarchitectural dysplasia).

Statistical analysis

Age at epilepsy onset, seizure frequency and age at the time of surgery were compared according to histopathological subtype using analysis of variance (ANOVA) and Scheffé's post hoc analysis when necessary. The χ2 test was used to compare gender, FS, MRI findings, temporal lobe resection and surgical outcome between the patients with isolated FCD and those with other additional lesions. The duration of epilepsy in seizure-free (Engel class I) and other patients was analysed using a t test. A multivariate logistic regression model was used to evaluate the associations between temporal lobe resection, histopathological diagnosis of isolated FCD I and class I outcome. P values of < 0.05 were considered significant. All of the analyses were performed using SPSS software (version 17 for Windows).

Results

The general characteristics of the population subgroups are depicted in figure 1. According to Palmini's classification (Palmini et al., 2004) 173 of our 215 patients (80%) were classified as type IA, and 42 (20%) as type IB. However, because there was an insufficient number of type IB patients to perform statistical analysis, and the assessment of this subtype is still debated, we evaluated the cohort as a whole without any further subclassification.

The population included 132 males (61%) and 83 females (39%). Their mean age was eight years (SD 8; range 0-53) at epilepsy onset and 27 years (SD 13; range 2-60) at the time of surgery. A mean duration of epilepsy was 19 years (SD 11; range 1-59); mean seizure frequency at the time of surgery was 47 per month (SD 120; range 1-1,000). MRI was uninformative in 25 patients (12%) (table 1). The epileptogenic zone was revealed by electroclinical and imaging data in the temporal lobe in 133 cases (62%), the frontal region in 30 (14%), multilobar including the temporal region in 38 (18%), and multilobar not including the temporal region in 11 (5%) (table 2). Follow-up data was available for all of the patients, 148 (69%) were classified as Engel class I, including 123 (57%) in Engel class Ia + c (tables 3 and 4).

Type I FCD was “isolated” in 66 (31%) patients who showed no neuropathological signs of other lesions, and was associated with other lesions in 149 (69%) (table 1).

Isolated FCD I

The 66 patients with isolated FCD I included 41 (62%) males and 25 (38%) females. Their mean age at epilepsy onset was eight years (SD 10; range 0-53) and 23 years at the time of surgery (SD 12; range 2-57) with a mean duration of epilepsy of 15 years (SD 9; range 1-42). Mean seizure frequency at the time of surgery was 115 per month (SD 200; range 1-1,000); only three (4.5%) patients reported febrile seizures. MRI was uninformative in 22 (33%) subjects, and SEEG was considered mandatory in 47 (71%) (table 1).

On the basis of the electroclinical and anatomical data, 23 (35%) patients underwent frontal lobe surgery, 15 (23%) temporal lobe surgery and 26 (39%) patients underwent multilobar resection (table 2). Thirty patients (46%) were classified as class I, of whom 23 (35%) were classified as class Ia + c; 19 patients (29%) did not benefit from surgery (table 3). AED therapy was completely discontinued for nine of the class I patients (14%), reduced for two and was continued for all other patients.
Table 1 General characteristics.

Histology

No. of patients (%)

Males (%)

Females (%)

Age at epilepsy onset (SD)

Duration of epilepsy (SD)

Age at surgery (SD)

Monthly seizure frequency (SD)

Ante- cedents (%)

FS (%)

Negative MRI (%)

SEEG (%)

FCD isolated

66 (31)

41 (62)

25 (38)

8 (10)

15 (9)

23 (12)

115 (200)

25 (38)

3 (5)

22 (33)

47 (71)

FCD + MCD

16 (7)

11 (69)

5 (31)

10 (7)

18 (11)

28 (11)

21 (19)

6 (38)

1 (6)

0

12 (75)

FCD + tumour

49 (23)

35 (71)

14 (29)

8 (6)

17 (12)

25 (14)

20 (29)

13 (27)

4 (8)

1 (2)

13 (27)

FCD + HS

76 (35)

40 (53)

36 (47)

8 (8)

25 (10)

34 (11)

10 (7)

32 (42)

41 (54)

2 (3)

11 (15)

FCD + other

8 (4)

5 (62)

3 (38)

6 (3)

12 (7)

18 (10)

45 (38)

7 (88)

1 (13)

0

7 (88)

Total

215

132 (61)

83 (39)

8 (8)

19 (11)

27 (13)

47 (120)

83 (39)

50 (23)

25 (12)

90 (42)


Table 2 Site of surgery.

Histology

Temporal (%)

Frontal (%)

Parietal (%)

Occipital (%)

Multilobar except temporal (%)

Multilobar with temporal (%)

Total multilobar (%)

FCD isolated

15 (23)

23 (35)

1 (2)

1 (2)

6 (9)

20 (30)

26 (39)

FCD + MCD

8 (50)

1 (6)

0

0

1 (6)

6 (38)

7 (44)

FCD + tumour

35 (71)

6 (12)

1 (2)

0

3 (6)

4 (8)

7 (14)

FCD + HS

75 (99)

0

0

0

0

1 (1)

1 (1)

FCD + other

0

0

0

0

1 (12)

7 (88)

8 (100)

Total

133 (62)

30 (14)

2 (1)

1 (1)

11 (5)

38 (18)

49 (23)


Table 3 Surgery outcome according to Engel classification 1987.

Histology

Class Ia + Ic (%)

Class I (%)

Class II (%)

Class III (%)

Class IV (%)

Therapy stopped (%)

FCD isolated

23 (35)

30 (46)

7 (11)

10 (15)

19 (29)

9 (14)

FCD + MCD

10 (63)

14 (88)

0

1 (6)

1 (6)

1 (6)

FCD + tumour

38 (78)

40 (82)

3 (6)

5 (10)

1 (2)

17 (35)

FCD + HS

50 (66)

62 (82)

7 (9)

3 (4)

4 (5)

18 (24)

FCD + other

2 (25)

2 (25)

2 (25)

2 (25)

2 (25)

1 (12)

Total

123 (57)

148 (69)

19 (9)

21 (10)

27 (13)

46 (21)


Table 4 Outcome relative to site of surgery.

Site of surgery

No. of patients (%)

Patients in class I (%)

Patients in class Ia + Ic (%)

Temporal

133 (62)

110 (83)

91 (68)

Frontal

30 (14)

12 (40)

12 (40)

Parietal

2 (1)

1 (50)

0

Occipital.

1 (1)

0

0

Multilobar except temporal

11 (5)

5 (46)

2 (18)

Multilobar including temporal

38 (18)

20 (53)

18 (47)

Total multilobar

49 (23)

25 (50)

20 (41)

Total

215

148 (69)

123 (57)

FCD I and other lesions

Of the 149 patients with FCD I associated with other types of histopathologically proven lesions, 76 (35%) presented with HS, 49 (23%) tumours, 16 (7%) other types of MCD, and eight (4%) anoxic-ischaemic or post-inflammatory lesions (table 1).

FCD I and HS

The 76 patients in this group included 40 males (53%) and 36 (47%) females. Their mean age at epilepsy onset was eight years (SD 8; range 0-42) and 34 years at the time of surgery (SD 11; range 7-60) with a mean duration of epilepsy of 25 years (SD 10; range 5-59). Mean seizure frequency at the time of surgery was 10 per month (SD 7; range 1-35), and 41 (53.9%) patients reported FS; the association between a history of FS and a pathological diagnosis of HS in the entire cohort was significant (p < 0.01). Only 15% of the patients underwent invasive SEEG (table 1). Based on MRI, HS was visualised in association with suspected FCD in 40 patients and diagnosed alone in 26, FCD was suspected in eight and results were uninformative in two patients. One patient underwent multilobar resection; only the temporal lobe was involved in the other cases (table 2). The vast majority of the patients [62 (82%)] were classified as class I (66% as class Ia + c). AED therapy was terminated for 18 patients and reduced for 25 patients (table 3).

FCD I and tumours

In this group of 49 patients (35 males and 14 females), glioneuronal tumours (GNTs) were the most frequent (82%); the diagnoses included 26 gangliogliomas (WHO Grade I) (GGs; 53%), 14 dysembryoplastic neuroepithelial tumours (WHO Grade I) (DNTs; 28%), two xanthoastrocytomas (WHO Grade II), two pilocytic astrocytomas (WHO Grade I), two gangliocytomas (WHO Grade I), one meningioma (WHO Grade I), one oligoastrocytoma (WHO Grade II) and one plexus papilloma (WHO Grade I).

The patient mean age was eight years (SD 6; range 0-24) at epilepsy onset and 25 years (SD 14; range 4-58) at the time of surgery with a mean duration of epilepsy of 17 years (SD 12; range 1-44). Mean seizure frequency at the time of surgery was 20 per month (SD 29; range 1-150); only four patients (8.2%) reported FS. Thirteen (27%) patients underwent invasive SEEG recording. Based on MRI, 44 patients were shown to have only a tumour, four patients a tumour associated with cortical-subcortical signal alterations, suggesting FCD, and one patient was normal (table 1).

Surgical resection was restricted to the temporal lobe in 35 (71%) patients and the frontal lobe in six; seven patients (14%) underwent multilobar resections (table 2). Forty (82%) patients were classified as class I (78% as class Ia + c) (table 3). AED therapy was terminated for 17 patients and reduced for 10 patients.

FCD I and other MCDs

The 16 patients in this group included five (31%) females and 11 males (69%). The most frequent malformation associated with FCD I was periventricular nodular heterotopia (PNH) which was found in 11 subjects; four patients had hamartomas and one a polymicrogyria.

The patients’ mean age was ten years (SD 7; range 0-19) at epilepsy onset and 28 years (SD 11; range 2-43) at the time of surgery with a mean duration of epilepsy of 18 years (SD 11; range 1-42). Mean seizure frequency at the time of surgery was 21 per month (SD 19; range 2-60), and one patient (6%) reported FS. Invasive SEEG was performed in 12 patients (75%) (table 1). MRI was positive for MCD in 16 patients, although FCD was diagnosed in only one. Eight (50%) patients underwent temporal resection, seven (44%) multilobar resection and one frontal lobe resection (table 2). Fourteen (88%) patients were classified as class I, including 10 (63%) as class Ia + c (table 3). AED therapy was terminated for one patient and reduced for 10 patients.

FCD I and anoxic-ischaemic/inflammatory lesions

The eight patients in this group included three females and five males. Their mean age was six years (SD 3; range 2-13) at epilepsy onset and 18 years (SD 10; range 7-36) at the time of surgery with a mean duration of epilepsy of 12 years (SD 7; range 5-23). Mean seizure frequency at the time of surgery was 45 per month (SD 38; range 4-100); one patient reported FS. Invasive SEEG was performed in seven (88%) patients (table 1). MRI revealed the gliotic-inflammatory lesions in all of the patients, without any evidence of FCD (table 2).

All of the patients underwent multilobar resection (table 2). Two (25%) patients were classified as Class Ia + c (table 3). AED therapy was terminated for one patient and reduced for another.

Statistical analysis

Statistical analysis of the subgroups did not reveal any significant difference in mean age at epilepsy onset (p = 0.80). All of the subgroups included more males than females. Scheffé's post hoc test showed that the duration of epilepsy at the time of surgery was significantly longer in the case of the patients with FCD I associated with HS than in those in all of the other groups (p < 0.001). The same test showed that seizures at the time of surgery were significantly more frequent in the patients with isolated FCD I than in those in the other subgroups (p < 0.001), as were negative MRI findings.

With the exception of the patients with FCD I and anoxic-ischaemic/inflammatory lesions, the patients with isolated FCD I had the worst outcome (46% classified as class I and 35% as class Ia + c). Isolated FCD I was significantly associated with extra-temporal resection (p < 0.01), whereas resections restricted to the temporal lobe were closely associated with a class I outcome (Pearson's χ2 test: p < 0.001). Furthermore, a multivariate logistic regression analysis designed to evaluate the possible associations between a histopathological diagnosis of isolated FCD I, temporal lobe surgery and class I outcome confirmed that a class I outcome is closely associated with surgery restricted to the temporal lobe (p < 0.001; OR: 3.69; 95% CI: 1.80-7.50), and negatively associated with isolated FCD I (p = 0.021; OR: 0.42; 95% CI: 0.21-0.88). These data suggest that the worse outcome of patients with isolated FCD I is related to both pathology and localisation.

Discussion

The advances made in imaging technologies, particularly new MRI techniques, have meant that structural brain lesions are being increasingly recognised as possible causes of chronic, drug-resistant focal epilepsies, and this has led to a greater use of surgery as an alternative means of treatment. There has also been a parallel improvement in neuropathological studies of surgical specimens, in which the most frequently encountered lesions include hippocampal sclerosis (HS), glioneuronal tumours (GNTs), vascular malformations, glial scars and MCD. However, many studies of surgical series only take predominant or major histopathological lesions into account, even though a specimen may show more than one histopathological alteration (Wolf et al., 1993; Eriksson et al., 2005). The presence of different histopathological lesions in the same specimen has raised doubts concerning the primacy of structural abnormalities in the genesis of seizures. This is particularly true in the case of FCD (especially type I), which is more prevalent in the temporal lobe where other lesions also occur (Tassi et al., 2002; Bautista et al., 2003; Fauser et al., 2004).

One of the main problems in reviewing the published data concerning MCD (and particularly FCD) in patients undergoing epilepsy surgery is the wide range of assumptions and classification systems used in the different studies. Studies based on imaging analyses use the classification proposed by Barkovich et al. (2005), whereas those based on neuropathological evaluations generally use one of two classification systems: the neuropathological grading system proposed by Mischel et al. (1995) (Kloss et al., 2002; Bocti et al., 2003; Chung et al., 2005; Lee et al., 2005; Yun et al., 2006) or the classification proposed by Palmini et al. (2004), which also considers electroclinical, imaging and post-surgical outcome data (Widdess-Walsh et al., 2005; Hildebrandt et al., 2005; Bast et al., 2006; Fauser et al., 2006; Krsek et al., 2008; Kim et al., 2009). Furthermore, most of the analyses concentrate on different types of cortical dysplasia and exclude any associated brain lesions, although the association between FCD and HS is often considered in studies of patients undergoing temporal lobe surgery.

In this retrospective evaluation of neuropathologically identified FCD I we have verified the incidence and clinical features, irrespective of localisation, and also considered other lesions observed in the same surgical specimens.

Isolated FCD I

Patients with isolated FCD I accounted for almost one-third (31%) of the entire study population. Although their mean age at epilepsy onset was not significantly different from that of the patients in the other groups, they had more frequent seizures at the time of surgery. Furthermore, MRI was completely uninformative in 33% of the cases, and 71% underwent pre-surgical invasive recordings. Routine MRI examinations are unlikely to detect subtle developmental abnormalities, particularly minor malformations of cortical development (Palmini et al., 2004), but even high-resolution MRI may not detect FCD I, particularly when outside the temporal lobe (Colombo et al., 2003, 2009).

The patients with isolated FCD I had a very poor outcome, with only 46% falling into class I (35% in class Ia + c). The distribution of FCD I was quite wide, although there was a prevalence of frontal (35%) and multilobar areas (39%). The poor surgical outcome may reflect incomplete surgical resection even in those undergoing SEEG because lesion demarcation is frequently poor in MRI-negative cases, and the dysplasia detected by MRI tended to be extensive. It can be argued that, regardless of the underlying pathology, temporal lobe surgery has the best outcome because there is more chance that the lesion(s) will be completely removed. The poor outcome observed in patients with isolated FCD I could therefore be related to prevalence outside the temporal lobe. However, multivariate logistic regression analysis showed that FCD I and a lobar location are independently correlated with surgical outcome. Similar findings have been reported by Krsek et al. (2009) showing that, in a paediatric population, the majority of patients presented multilobar lesions with only 21% seizure free after surgery.

FCD I associated with other lesions

One hundred and forty-nine of our 215 patients (69%) had FCD I associated with other lesions, of which almost 4/5 were located in the temporal lobe. Four main groups of associated lesions were identified; the vast majority were tumours (23%), HS (35%), other MCDs (7%) and anoxic-ischaemic/inflammatory lesions which were observed in only 4% of the patients, all of whom showed multilobar involvement.

It is well known that, although rare (Blümcke and Wiestler, 2002; Luyken et al., 2003), some tumours (particularly GGs and DNTs) are highly epileptogenic and mainly found in the temporal lobe of children and young adults with pharmacologically intractable focal epilepsy. The presence of dysplastic cortex near the tumour has often been observed and FCD type I has been reported in the large majority of patients (Prayson et al., 1996; Ferrier et al., 2006). In such cases, it is often difficult to differentiate between neoplastic and dysplastic lesions, even in neuropathological specimens. However, it is still debated as to whether the association of FCD with GNTs should be considered as a distinct entity (different from isolated FCD). In agreement with the current literature, we found that the majority of the alterations were in the temporal lobe (71%), although they were also frequently multilobar (14%) or frontal (12%). Perhaps surprisingly, the post-surgical outcome of these patients was very good (82% in class I and 78% in class Ia + c).

Long-term follow-up studies of large patient series have shown that patients with GNTs had favourable outcome as the large majority became seizure free. Complete resection predicts a better post-surgical outcome (Aronica et al., 2001; Luyken et al., 2003). Although no systematic studies have shown that a failure to achieve post-surgical seizure relief is due to unrecognised and unresected adjacent cortical dysplasia, the very high proportion of our patients with tumours (particularly GNTs), who became seizure free, further suggests that a large resection including the surrounding dysplastic area is the most important factor for a good post-surgical outcome and that cortical disorganisation adjacent to the tumour significantly contributes to epileptogenesis.

The largest subgroup in our population consisted of patients with FCD I and HS (76/215; 35%), in whom surgery was restricted to the temporal lobe (only one patient underwent multilobar resection including the temporal lobe). As in the case of our FCD I patients with tumours, surgical outcome was extremely good (82% in class I).

Fauser et al. (2004) reported 38 patients with mild MCD or FCD I in a cohort of 67 FCD patients (57%), of whom 91% had temporal lobe lesions; 60% of the patients were in class I and there were no differences between those with temporal and extratemporal FCD. The same group (Fauser et al., 2006) reported a similar incidence of FCD I in a more recent retrospective study of 120 patients; however, lobar distribution was not evaluated, although 40% of the patients also had HS. An association between FCD and HS has been frequently described in studies of temporal lobe surgery (Levesque et al., 1991; Raymond et al., 1994; Cendes et al., 1995; Prayson et al., 1996; Li et al., 1999), and the best surgical outcome has been obtained when both are removed (Li et al., 1999; Salanova et al., 2004; Chabardès et al., 2005), which suggests that both contribute to seizure genesis. Tassi et al. (2009) have recently reported another lesion associated with HS in 71% of patients surgically treated for temporal lobe epilepsy; although tumours were rare (8%) 83% of these patients had FCD I, thus confirming that this type of dysplasia is the most common alteration associated with HS (Eriksson et al., 2005). Interestingly, the outcome of the patients only affected by HS was very good and similar to that found in the present cohort of patients with HS associated with FCD I.

Dual pathology generally refers to the presence of HS with an additional macroscopic extrahippocampal lesion; the published incidence in patients with refractory temporal lobe epilepsy ranges from 5% to 30%, with the commonest extra-hippocampal lesions being cortical developmental abnormalities (Levesque et al., 1991; Cascino et al., 1993; Raymond et al., 1994; Cendes et al., 1995; Li et al., 1999). This wide range is largely due to differences in patient selection and neuropathological procedures which only consider predominant and macroscopic histopathological features.

Thom et al. (2009) have recently found that widespread changes involving neocortical structures can be present in patients with temporal lobe epilepsy and hippocampal sclerosis; in particular, 11% of the patients showed a peculiar disorganisation of the supragranular layers previously diagnosed as FCD I (Thom et al., 2000; Garbelli et al., 2006). The authors introduced the term “temporal lobe sclerosis” to describe these patients and suggested that the neocortical disorganisation is probably an acquired, non-developmental process representing part of the spectrum of mesial temporal sclerosis rather than a second or dual pathology. We also detected the same neocortical lesion associated with HS, however, we included these patients in our FCD I cohort because, although we generally agree with the hypothesis of Thom et al., we feel that further studies should confirm that such cases are indeed different from FCD I. Furthermore, the other patients with HS had neuropathological features of FCD I that do not fulfil the criteria for a diagnosis of temporal lobe sclerosis as defined by Thom et al. (2009).

The majority of our patients had FCD I associated with tumours or HS, but we also found FCD I associated with other types of MCD (7%) or other acquired lesions due to vascular or inflammatory processes (4%). The most frequent MCD associated with FCD I was nodular heterotopia, as previously reported (Tassi et al., 2005; Meroni et al., 2009). Surgical outcome in this group of patients was also very good (88% in class I, 63% in class Ia + c), once again suggesting that a favourable outcome mainly depends on complete resection of the epileptogenic zone even when the malformation involves more than one lobe.

The worst surgical outcome was found in the FCD I patients with anoxic-ischaemic lesions (25% in class I), although it was not possible to perform any statistical evaluation or particular physiopathological speculation due to the small patient number. However, it seems reasonable to assume that a pathological event led to the profound disorganisation of a large portion of the cortical mantle, presumably during pre-natal or early post-natal development. The epileptogenic zone could not be precisely identified in these cases even by means of invasive recordings, and other undetected epileptogenic areas may explain the persistence of seizures after surgery. In the case of these lesions, the compromised cortex (currently defined as FCD I) is more likely to be an acquired alteration, secondary to the development of the principal lesion which some authors have called “progressive cortical dysplasia” (Marin-Padilla et al., 2002).

In conclusion, our findings suggest that isolated FCD I should be evaluated differently from FCD I in the presence of another principal lesion. The epilepsy history and surgical outcome of FCD I patients with tumours, hippocampal sclerosis or other MCDs was similar to those of patients with tumours, HS or MCD without FCD I, whereas patients with isolated FCD I not only had a different epilepsy history but also had a poor surgical outcome. The fact that surgical outcome in patients with FCD I depends on the presence/absence of a second pathology suggests that each component in “combined” cases may contribute to the genesis of seizures. New MRI and electroclinical data, as well as new histopathological and biomolecular markers, are needed to further define this type of dysplasia.

Our data also support the use of a comprehensive clinicopathological approach to FCD I that takes into account different parameters in order to: i) develop a more reliable classification system as a prerequisite for the pre- and post-surgical management of epileptic patients; and ii) offer more reliable predictions of post-surgical outcome.

Financial support

This study was supported by the Italian Ministry of Health, Fondazione Banca del Monte di Lombardia, Associazione Paolo Zorzi per le Neuroscienze, and EU grant “Functional Genomics and Neurobiology of Epilepsy” (EPICURE, contract No. LSHM-CT-2006-0373315).

Disclosure.

None of the authors has any conflict of interest to disclose.