Epilepsy surgery of focal cortical dysplasia–associated tumors



The goal of the present study was to evaluate the clinical characteristics and postoperative seizure outcome of epileptogenic tumors associated with focal cortical dysplasias (FCDs) compared to both solitary FCD type I and solitary tumors. Particular attention is given to FCD type IIIb (tumors associated with FCD type I), which have been recently classified as a separate entity. We retrospectively reviewed the clinical charts of 1,109 patients who were operated on for drug-resistant focal epilepsy, including 492 patients with a histologic diagnosis of solitary FCD I and II (83 and 157 cases, respectively), solitary tumors (179 cases), and FCD-associated tumors (73 cases, 58 of which met the criteria of FCD IIIb of the new International League Against Epilepsy [ILAE] classification). The different subgroups were evaluated for clinical characteristics and postoperative surgical outcome. Clinical variables and postoperative seizure outcome of patients with coexisting tumor and FCDs (FCD IIIb and tumor associated FCD II) were similar to those of patients with a solitary tumor and differed significantly from patients with solitary FCDs. Nevertheless, tumors associated with FCDs are characterized by a striking male predominance and a higher seizure frequency as compared to solitary tumors. Patients with drug-resistant focal epilepsy secondary to a solitary tumor or with a tumor-associated FCD have similar basic clinical presentation and postoperative seizure outcome. Nevertheless, the epileptogenic contribution of the associated FCDs can be crucial, and it needs to be adequately assessed. The impact of FCD on tumor-related epilepsy deserves future research in order to optimize the surgical strategies aimed at seizure relief.

Among the etiologic substrates of refractory focal epilepsies, tumors are encountered in approximately 30% of operated cases (Luyken et al., 2003). The most common tumor subtypes are gangliogliomas (GGs), dysembryoplastic neuroepithelial tumors (DNTs), and pilocytic or pleomorphic astrocytomas, all of which share relatively benign biologic behavior and long duration of epilepsy (Thom et al., 2012). Less frequently, long-term refractory epilepsy may be due to more aggressive tumors, such as diffuse astrocytomas and oligodendrogliomas (World Health Organization [WHO] grade II).

It has been noted that especially in glioneuronal tumors (GGs and DNTs), the region surrounding the neoplasm is characterized by cortical disorganization in 80% of cases (Morris et al., 1998; Kleihues & Cavanee, 2000). In addition, a strong association of these tumors with focal cortical dysplasias (FCDs) has been reported (Pasquier et al., 2002; Ferrier et al., 2006; Tassi et al., 2010), raising the question as to the pathophysiologic processes underlying the development of these distinct entities and which of the two coexisting lesions carries the leading role in seizure generation. Invasive electroencephalography (EEG) evaluations have demonstrated that in cases of glioneuronal tumors, more abundant epileptiform activities may be recorded in patients with a coexisting FCD (Ferrier et al., 2006). Other data, however, have shown that glioneuronal tumors can be intrinsically epileptogenic, even when associated with FCDs (Barba et al., 2011). The indolent behavior of glioneuronal tumors, as well as their potential epileptogenicity, has raised the hypothesis of a developmental rather than neoplastic nature of these entities (Barkovich et al., 2005).

For these reasons, the attention of neuropathologists has moved progressively from the tumor to the FCD, and in the more recent three-tiered classification of FCDs released by the ad hoc Task Force of the International League Against Epilepsy (ILAE), a new category, FCD type III was introduced, which includes FCDs associated with other pathologies (Blümcke et al., 2011). Specifically, FCD type IIIb is defined by the association of FCD type I with a glial or glioneuronal tumor. The main goal of the present study is to evaluate the clinical characteristics and postoperative seizure outcome of FCDs IIIb compared to both solitary FCD I and solitary tumors to examine the clinical significance of the new diagnostic entity.

Patients and Methods

We retrospectively reviewed the clinical charts of 1,109 patients operated on for drug-resistant focal epilepsy between 1996 and 2012 at “C. Munari” Epilepsy Surgery Center, Niguarda Hospital, Milan, Italy. Patients with a histologic diagnosis of a glial or glioneuronal tumor associated with any FCD (type I and II) were selected. Tumor-associated FCD I was defined as “FCD IIIb” and tumor-associated FCD II was defined as “double pathology,” as recommended by the ILAE Task Force (Blümcke et al., 2011). The group of patients with a histologic diagnosis of any solitary FCD (I and II) or any solitary tumor was chosen as control.

All operated patients presented with seizures that were resistant to medical treatment. They underwent an individualized presurgical evaluation, which included history, careful evaluation of seizure semiology, interictal EEG, high-resolution (1.5 T) magnetic resonance imaging (MRI) (Colombo et al., 2009), scalp video-EEG monitoring as needed, and in cases with inconsistent findings, invasive evaluation with stereo-EEG (SEEG) (Cossu et al., 2005; Cardinale et al., 2013). Microsurgical resections were conducted with the aim of removing the epileptogenic zone (EZ), including the MRI-visible lesion. Histologic diagnosis was obtained by processing the surgical specimens as specified elsewhere (Tassi et al., 2002). Postoperative seizure outcome was assessed according to the classification system proposed by Engel (Engel et al., 1993).

The different subgroups of patients were compared for possible differences in gender, age at seizure onset, age at surgery, epilepsy duration before surgery, seizure frequency (categorized as high frequency, i.e., more than 30 spells/month, and low frequency, i.e., <30 spells/month), findings at MRI (categorized as presence or absence of a structural lesion), use of SEEG recording, site of surgery (for statistical purpose: temporal vs. extratemporal), and postoperative seizure outcome (Engel's class I vs. class II–IV). Bivariate statistical analysis was performed by using two-tailed Fisher's exact test for categorical and two-tailed Student's t-test for continuous variables. p-Values < 0.05 were considered statistically significant.


Four hundred ninety-two patients fulfilled the selection criteria and were assigned to the following subgroups according to their pathologic features.

Tumor-associated FCDs

FCD IIIb (FCD I plus tumor)

This subgroup included cases with FCD I associated with a tumor (Fig. 1). There were 58 patients (79% of all tumor-associated FCDs), whose main features are reported in Table 1. The tumor types associated with the FCD I are reported in Table 2.

Table 1. Clinical characteristics and postoperative seizure outcome for solitary FCD I and tumors compared to FCDs IIIb
 FCD I (n = 83)Tumors (n = 179)FCD IIIb (n = 58)FCD IIIb versus FCD IFCD III b versus tumor
  1. a

    Comparison performed between temporal versus other localizations.

  2. SD, standard deviation.

  3. b

    Comparison performed between class I versus class non-I (according to Engel classification).

Males (%)52 (63)95 (53)44 (76)p = 0.1034 p = 0.0022

Age (years) at seizure onset

Mean ± SD (range)

6.8 ± 8.8 (0–53)11.4 ± 9.0 (0–38)10.7 ± 7.6 (0–33)p = 0.0070p = 0.5941

Age (years) at surgery

Mean ± SD (range)

21.3 ± 12.0 (2–57)23.5 ± 12.8 (1–51)25.4 ± 14.3 (1–58)p = 0.0673p = 0.3410

Duration of epilepsy (years)

Mean ± SD (range)

14.5 ± 9.2 (1–42)12.1 ± 10.5 (0–35)15.3 ± 12.5 (0–44)p = 0.6622p = 0.0558
Seizure frequency (%)     
Low45 (54)56 (75)51 (88) p < 0.0001 p = 0.0450
High38 (46)44 (25)7 (12)
Structural lesion at MRI (%)58 (70)179 (100)57 (98) p < 0.0001 p = 0.2447
SEEG performed (%)56 (67)23 (13)11 (19) p < 0.0001 p = 0.2818
Site of surgery (%)     
Temporal19 (23)99 (55)43 (74) p < 0.0001 a p = 0.0134 a
Frontal27 (33)28 (16)7 (12)
Posterior quadrant19 (23)36 (20)7 (12)
Rolandic/perirolandic8 (10)11 (6)
Other multilobar10 (11)5 (3)1 (2)
Outcome (%)     
Class I39 (47)146 (81)48 (83) p < 0.0001 b p = 1b
Class II10 (12)16 (9)4 (7)
Class III14 (17)12 (7)4 (7)
Class IV20 (24)5 (3)2 (3)
Table 2. Oncotypes in the subgroups of FCD IIIb and solitary tumors
Oncotype of tumorsFCD III b (n = 58) (%)Solitary tumors (n = 179) (%)
DNT19 (33)59 (33)
Ganglioglioma26 (4572 (40.2)
Gangliocytoma2 (3.4)3 (1.7)
Desmoplastic infantile ganglioglioma1 (0.6)
Neurocytoma1 (0.6)
Pilocytic astrocytoma2 (3.4)13 (7.3)
Pleomorphic xanthoastrocytoma6 (10.3)6 (3.4)
Oligodendroglioma2 (3.4)13 (7.3)
Astrocytoma5 (2.8)
Oligoastrocytoma1 (1.7)1 (0.6)
Astroblastoma1 (0.6)
Anaplastic oligodendroglioma2 (1.1)
Anaplastic astrocytoma1 (0.6)
Anaplastic oligoastrocytoma1 (0.6)
Figure 1.

Forty-six year-old woman with seizure onset at 16 years of age, presenting with a tumor associated with FCD type 1a (FCD IIIb). (A) and (B) T2-weighted fluid-attenuated inversion recovery (FLAIR) sagittal (A) and axial (B) MRI sequences, showing a small-sized pseudo-cystic lesion surrounded by increased signal in the left basal temporal region. The patient underwent a tailored left anteromesial temporal lobectomy with gross total resection of the MRI-visible abnormal tissue. (C) and (D) High-power photomicrographs showing a pleomorphic xanthoastrocytoma with meningeal infiltration (C, hematoxylin-eosin, 200×) and a FCD type 1a (D, Klüver-Barrera, 40×).

“Double pathology” (FCD II plus tumor)

There were eight patients (11% of all tumor-associated FCDs, Fig. 2) whose main features are detailed in Table 3. FCD II with tumors are not included in the FCD IIIb variant of the ILAE classification, which recommends the definition of “double pathology” for FCD II associated with another lesion.

Table 3. Main clinical characteristics and postoperative seizure outcome for solitary FCDs II, tumor- associated FCDs II, and solitary tumors, and statistical comparison between subgroups
 FCD II (n = 157)Tumor (n = 179)FCD II+ tumor (n = 8)FCD II+ tumor versus FCD IIFCD II+ tumor versus tumor
  1. a

    Comparison performed between temporal versus other localizations.

  2. b

    Comparison performed between class I versus class non-I (according to Engel classification).

  3. SD, standard deviation.

Males (%)91 (58)95 (53)5 (63)p = 1p = 0.7261

Age at seizure onset

Mean ± SD (range)

5.9 ± 5.9 (0–30)11.4 ± 9.0 (0–38)11.6 ± 8.5 (2–23)p = 0.0100p = 0.9509

Age at surgery

Mean ± SD (range)

23.3 ± 13.5 (1–60)23.5 ± 12.8 (1–51)26.3 ± 14.4 (7–45)p = 0.5418p = 0.5477

Duration of epilepsy

Mean ± SD (range)

17.3 ± 11.6 (1–53)12.1 ± 10.5 (0–35)14.6 ± 8.5 (1–26)p = 0.5175p = 0.5080
Seizure frequency (%)     
Low74 (47)135 (75)8 (100 p = 0.0031 p = 0.2014
High83 (53)44 (25)
Structural lesion at MRI (%)134 (75)179 (100)8 (100)p = 0.6014p = 1
SEEG performed88 (56)23 (13) p = 0.0018 p = 0.5986
Site of surgery (%)     
Temporal10 (6)99 (55)6 (75) p < 0.0001 a p = 0.4692a
Frontal85 (54)28 (161 (12)
Posterior quadrant32 (20)36 (20)1 (12)
Rolandic/perirolandic20 (14)11 (6)
Other multilobar10 (6)5 (3)
Outcome (%)     
Class I132 (84)146 (81)6 (76)p = 0.6186bp = 0.6449b
Class II6 (4)16 (9)1 (12)
Class III10 (6)12 (7)1 (12)
Class IV9 (6)5 (3)
Figure 2.

Thirty-one year-old woman who experienced onset of seizures at 14 years of age, with a tumor associated with an FCD IIa. (A) T2-weighted FLAIR coronal MRI sequence showing a neoplastic lesion in the right frontal lobe. A lesionectomy, extended to surrounding tissue, was performed. (BD) High-power photomicrographs show a neoplastic nodule in white matter, characteristic of a dysembryoplastic neuroepithelial tumor (DNT) (B, Klüver-Barrera 40×), DNT floating neuron in myxoid matrix (C, Hematoxylin-Eosin, 400×), and a dysmorphic neuron (D, Bielschowsky 400×).

FCD I plus tumor plus hippocampal sclerosis (HS)

This group of patients (seven cases, 10% of all the tumor-associated FCDs) did not receive further consideration in the present study, due to its uncertain attribution to FCD IIIa, FCD IIIb, or as “double pathology” according to the ILAE classification (Fig. 3).

Figure 3.

Thirty-nine year-old woman, who presented with a history of febrile convulsions in infancy. A few years later she had onset of afebrile partial seizures. (A, B) T2-weighted FLAIR axial and coronal MRI sequences showing a neoplastic lesion in the right temporal lobe surrounded by an extended area of abnormal signal, suggesting focal cortical dysplasia. The patient received a tailored right anteromesial temporal lobectomy, which included gross total resection of MRI-visible abnormal tissue. (CF) High power photomicrographs showing a ganglioglioma (C, Hematoxylin-Eosin, 400×), an FCD type 1a (D, Klüver-Barrera, 10×), neuronal depletion in sector 1A of the Ammon's horn with adjacent neoplastic tissue (E, Klüver-Barrera, 40×). and neuronal loss in Ammon's horn (F, Neu-N, 20×).

Solitary FCDs

Solitary FCD I

This group included 83 patients. Table 1 reports the main demographic and clinical features of these patients.

Solitary FCD II

One hundred fifty-seven patients were selected, whose clinical features are detailed in Table 3.

Solitary tumors

This group consisted of 179 patients. Their main clinical characteristics and histologic subtypes are summarized in Tables 1 and 2. The distribution of glioneuronal tumors and other oncotypes in this group and in that of tumors with FCD IIIb did not differ significantly (p = 0.3649, two-tailed Fisher's exact test).

Statistical evaluation

The clinical characteristics of FCDs IIIb were compared with those of both solitary FCDs I and solitary tumors (Table 1). FCDs IIIb significantly differ from solitary FCDs I in age at seizure onset, seizure frequency, presence of MRI findings, need for SEEG evaluation, site of surgery (temporal vs. extratemporal), and seizure outcome. FCD IIIb differed from solitary tumors in terms of gender, seizure frequency, and site of surgery (temporal vs. extratemporal).

The same statistical procedure was conducted to compare the main features of tumor-associated FCD II (double pathology) with those of solitary FCD II and solitary tumors (Table 3). Tumor-associated FCDs II differed significantly from solitary FCD II in age at seizure onset, seizure frequency, use of SEEG evaluation, and site of surgery. No statistically significant differences were observed between tumor-associated FCD II and solitary tumors. Finally, statistical comparison did not show any significant difference between the clinical features of tumor-associated FCD II and FCD IIIb (Table 4).

Table 4. Comparison of the main clinical characteristics and postoperative seizure outcome between FCD II–associated tumors and FCD IIIb
 FCD II+ tumor (n = 8)FCD IIIb (n = 58)p-Value
  1. a

    Comparison performed between temporal versus other localizations.

  2. b

    Comparison performed between class I versus class non-I (according to Engel classification).

Males (%)5 (63)44 (76)0.4146

Age at seizure onset

Mean ± SD (range)

11.6 ± 8.5 (2–23)10.7 ± 7.6 (0–33)0.7577

Age at surgery

Mean ± SD (range)

26.3 ± 14.4 (7–45)25.4 ± 14.3 (1–58)0.8681

Duration of epilepsy

Mean ± SD (range)

14.6 ± 8.5 (1–26)15.3 ± 12.5 (0–44)0.8788
Seizure frequency (%)   
Low8 (100)51 (88)0.5842
High7 (12)
Structural lesion at MRI (%)8 (100)57 (98)1
SEEG performed (%)11 (19)0.3338
Site of surgery (%)   
 Temporal6 (75)43 (74)1a
Frontal1 (12)7 (12)
Posterior quadrant1 (12)7 (12)
Other multilobar1 (2)
Outcome (%)   
Class I6 (76)48 (83)0.6300b
Class II1 (12)4 (7)
Class III1 (12)4 (7)
Class IV2 (3)


The present study has several limitations, including its retrospective and single-center nature and the limited sample size. This prevented, for instance, subgroup analysis in order to evaluate possible differences in the clinical behavior of tumors associated with different variants of FCD I, as defined by the new ILAE classification. Within the aforementioned limitations, the key finding of our study is that FCD IIIb shows characteristics more similar to solitary tumors than to FCDs I, including excellent seizure outcome. This confirms the results from a previous study conducted in our center before the introduction of the new ILAE classification (Tassi et al., 2010). Nevertheless, some “dysplastic imprinting” is characteristic of FCD IIIb. First, the striking prevalence of male patients in the FCD I and FCD IIIb subgroups. This finding has been recently reported in FCD patients, and might be due to a still undetermined X-linked risk factor (Ortiz-González et al., 2013). Second, the slightly higher seizure frequency detected in FCD IIIb cases, in which the presence of cortical disorganization may represent anatomic substrate that enhances excitability within the EZ.

In patients with temporal lobe locations, especially mesial, our standard is to perform an anteromesial temporal lobectomy, which enhances not only seizure outcome but also the availability of peritumoral tissue for histologic evaluation. In extratemporal cases, conversely, resections are often restricted to the MRI-visible lesion, with limited opportunity to remove potentially dysplastic tissue surrounding the neoplasm. This treatment bias may explain the higher prevalence of FCD IIIb (74%) in temporal lobe resections compared to solitary tumors (55%).

Our data confirm that surgical results in solitary FCDs I are much less gratifying than those achieved with other etiologies, specifically FCD IIIb. Despite substantial histologic similarity, solitary FCD I is likely to represent a more widespread structural abnormality compared to the dysplastic component of FCD IIIb. FCD I lesions have more subtle neuroradiologic features (30% were MRI-negative in our series) and are likely part of a more diffuse and difficult-to-localize epileptogenic network. This justifies the frequent recourse to invasive EEG presurgical evaluation in FCD I cases compared to other etiologies.

In recent years, a great deal of basic and clinical research on the pathologic substrates of symptomatic drug-resistant epilepsies has been focused on malformations of cortical development, in particular FCDs, and substantial knowledge has been accumulated about the epileptogenetic properties of these malformations, which, furthermore, may now be recognized with increasing frequency by high-resolution MRI. The new classification of FCDs follows this direction, and dysplastic disorders play the leading role in the categorization of FCD-associated pathologies.

Nevertheless, the present study suggests that, at least as far as patients with FCDs IIIb are concerned, the clinical phenotype is apparently correlated with the associated pathology rather than the FCD. Furthermore, Tassi et al. (2010) have reported similar results for FCDs I associated with hippocampal sclerosis. Therefore, the “dysplasia-centered” attitude adopted in the new ILAE classification with the introduction of the FCD III variant, which relies on strict histopathologic criteria, may be less relevant from a merely clinical, “case-centered” point of view.

The ILAE classification clearly states that “the rare association between FCD Types IIa and IIb with hippocampal sclerosis, tumors, or vascular malformations should not be classified as FCD Type III variant,” and the definition of “double pathology” is recommended for the association between FCD II and other pathologies (Blümcke et al., 2011). However, in our group of tumor-associated FCDs, the association of a FCD II with a tumor was seen in 11% of cases, suggesting that this entity is not as rare as postulated. Furthermore, the clinical behavior of this subgroup is very similar to that of “true” FCDs IIIb, resembling more that of solitary tumors than that of the affiliated FCDs. Should this be confirmed by similar studies conducted in other centers, a reconsideration of the role of these entity in FCD classification could be taken into consideration.

In addition, our case series also included patients with associated FCD, tumor, and HS (10% of all cases of FCD plus tumor), which are difficult to categorize because they share the inclusion criteria of both FCD IIIa and FCD IIIb. For those cases, provided that similar data arise from studies in other centers, additional criteria for inclusion in an appropriate classification frame could be considered.


None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal's position on issue involved in ethical publication and affirm that this report is consistent with those guidelines.