Machine learning approaches for imaging‐based prognostication of the outcome of surgery for mesial temporal lobe epilepsy

Abstract

Objectives Around 30% of patients undergoing surgical resection for drug‐resistant mesial temporal lobe epilepsy (MTLE) do not obtain seizure freedom. Success of anterior temporal lobe resection (ATLR) critically depends on the careful selection of surgical candidates, aiming at optimizing seizure freedom while minimizing postoperative morbidity. Structural MRI and FDG‐PET neuroimaging are routinely used in presurgical assessment and guide the decision to proceed to surgery. In this study, we evaluate the potential of machine learning techniques applied to standard presurgical MRI and PET imaging features to provide enhanced prognostic value relative to current practice. Methods Eighty two patients with drug resistant MTLE were scanned with FDG‐PET pre‐surgery and T1‐weighted MRI pre‐ and postsurgery. From these images the following features of interest were derived: volume of temporal lobe (TL) hypometabolism, % of extratemporal hypometabolism, presence of contralateral TL hypometabolism, presence of hippocampal sclerosis, laterality of seizure onset volume of tissue resected and % of temporal lobe hypometabolism resected. These measures were used as predictor variables in logistic regression, support vector machines, random forests and artificial neural networks. Results In the study cohort, 24 of 82 (28.3%) who underwent an ATLR for drug‐resistant MTLE did not achieve Engel Class I (i.e., free of disabling seizures) outcome at a minimum of 2 years of postoperative follow‐up. We found that machine learning approaches were able to predict up to 73% of the 24 ATLR surgical patients who did not achieve a Class I outcome, at the expense of incorrect prediction for up to 31% of patients who did achieve a Class I outcome. Overall accuracies ranged from 70% to 80%, with an area under the receiver operating characteristic curve (AUC) of .75–.81. We additionally found that information regarding overall extent of both total and significantly hypometabolic tissue resected was crucial to predictive performance, with AUC dropping to .59–.62 using presurgical information alone. Incorporating the laterality of seizure onset and the choice of machine learning algorithm did not significantly change predictive performance. Significance Collectively, these results indicate that "acceptable" to "good" patient‐specific prognostication for drug‐resistant MTLE surgery is feasible with machine learning approaches utilizing commonly collected imaging modalities, but that information on the surgical resection region is critical for optimal prognostication.

Interictal HFO and FDG‐PET correlation predicts surgical outcome following SEEG

“…The differences in FR × T could reflect more accurate EZ localization and subsequent higher possibility of complete resection. A recent prognostic study on mesial TLE proposed that the percent of temporal lobe hypometabolism resected was crucial for predicting outcome 46 . However, we did not replicate such results, which might be explained by the mixed seizure type cohort and differences in technique.…”

“…A recent prognostic study on mesial TLE proposed that the percent of temporal lobe hypometabolism resected was crucial for predicting outcome. 46 However, we did not replicate such results, which might be explained by the mixed seizure type cohort and differences in technique. Of interest, the specific histopathological abnormality pathologies vs nonspecific subgroup analysis also showed that the negative correlation no longer existed in the nonspecific pathologies group, which are usually considered to be associated with poorer surgical outcome.…”

Predicting seizure outcome after epilepsy surgery: do we need more complex models, larger samples, or better data?

“…Previous studies that have used both machine learning techniques and traditional statistical modelling approaches to predict post-operative seizure outcome have found that logistic regression models perform as well as, or even better than, machine learning ones. 11,15,22 To our knowledge, only one study by Yossofzai et al . 23 has found that a machine learning model outperforms a logistic regression; however, this was a 0.1-0.2 difference in AUC (0.72 versus 0.73 in the train dataset; 0.72 versus 0.74 in the test dataset).…”

“…The majority of these models have, however, been trained on relatively small sample sizes (N < 100) [7][8][9][10][11][12][13][14][15][16][17][18] and therefore have a high risk of 'overfitting' (a model overfits when it models the training dataset too closely, performing well on this dataset but consequently underperforming on new, 'unseen' datasets). 19,20 Model training sets have also been comprised almost exclusively of temporal lobe surgery patients 7,8,[10][11][12][13]15,17,18,21,22 , often relied on post-surgical factors 11,12,14,23 , and frequently utilized post-processing neuroimaging analyses that cannot be readily replicated by others. 7,[9][10][11]13,[16][17][18]21 As such, many existing models may be difficult to incorporate into routine pre-surgical evaluation.…”

“…19,20 Model training sets have also been comprised almost exclusively of temporal lobe surgery patients 7,8,[10][11][12][13]15,17,18,21,22 , often relied on post-surgical factors 11,12,14,23 , and frequently utilized post-processing neuroimaging analyses that cannot be readily replicated by others. 7,[9][10][11]13,[16][17][18]21 As such, many existing models may be difficult to incorporate into routine pre-surgical evaluation. Perhaps more importantly still, none of these models have been externally validated.…”