Wednesday, September 19, 2018

Why Epilepsy Surgery Needs to be Changed


-->

Most epileptic individuals are able to take medicine to combat their seizures. By most, I’m referring to 70%. For the other 30%, their seizures are medication-resistant. A beneficial option to those medication-resistant individuals is sometimes surgical resection of the brain are that is responsible for the seizure initiation. However, up to one-third of patients continue to have seizures even after the surgery. This is because the surgeries themselves are not very creative. Along with the continuation of seizures, the surgery itself carries a lot of risks. A lot of patients report post-surgical amnesia, vision defects, and psychological sequelae effects, such as depression. Given these worries, surgical candidates must first undergo a variety of tests and exams which help to map out the ictal focus, seizure initiation, and the surrounding region as accurately as possible. Epilepsy surgery then is usually saved for individuals with an easily accessible lesion.

Even individuals with an easily accessible lesion continue to have seizures. Gleichgerrcht et al. explored whether analysis of pre-existing whole brain networks could be used to predict the outcome of a surgical candidate. Using existing imaging and clinical data from previously treated patients, the authors trained a deep networking computer approach to analyze surgical outcome based on presurgical structural connectivity. These computed predictions were far superior to those arising from standard clinical variables alone. The authors were even surprised when they discovered that some of the connections linked to surgical outcome were actually far away from the seizure focus.

Though this model was very effective, it is not certain if its performance is reproducible since the original model was based on 50 individuals. If it can be reproduced on a larger scale is still unknown. Also, all of the individuals used in this model had one type of epilepsy, temporal lobe epilepsy. This is the most common type in adults and has a specific site of seizure origin. As to whether this model could be used in younger patients or different types of epilepsy is to be investigated. This study is a great example of how machine use and imaging data can be combined to enable better risk stratification of prospective surgical candidates. This study urges us to see epilepsy not just as a local brain dysfunction, but as a global one as well.

Posted by Danielle Bermingham (1)


9 comments:

  1. Of the 50 individuals studied using the deep networking computer approach did any have surgery after the study and if so how did they feel after the surgery? Did they not have the same side-effects or outcomes as other post-surgical patients? Also, was it mentioned if this approach was going to be expanded to a wider audience, ie, people with other types of epilepsy or younger patients or did the study stop with this small sample size? It is very intriguing, though, that they have found another approach to solving this issue that affects so many people.

    -Maddie Powers

    ReplyDelete
    Replies
    1. According to this article there was no further mention of the side-effects of the 50 individuals or if they continued to have seizures as well. It was mentioned that opening up this approach to a larger cohort remains to be investigated.

      - Danielle Bermingham

      Delete
  2. I had no idea there were different types of epilepsy.. so was the surgery being done before just a general surgery or was it specialized to each type? What is the cause of epilepsy in the first place? In this new model was there any serious side effects that the individuals experienced? If they were able to minimize the effects from the original surgery then this could be a huge step in finding a way to treat this issue!

    Posted by: Katarzyna Mosio

    ReplyDelete
    Replies
    1. Yes, there are four main types of epilepsy. I only have limited knowledge of the surgical techniques, but I believe that the surgery that was done before was specialized for each specific patient, not to the type of epilepsy. Though there are types, each case is different. Epilepsy is caused by irregular brain activity and can happen at any age and any point in time. The outcomes of this new model were not described in this article, but yes this could definitely be a huge step if it proves to be effective on a large scale.

      - Danielle Bermingham

      Delete
  3. This is an amazing leap in medicine and technology, this will, of course, help thousands of people suffering from this one kind of epilepsy. But, I feel like a sample size of fifty is quite small and it seems like it must go through more tests to see if it works for most other kinds of epilepsy. This also brings up the idea of how far can technology advance medicine, its not perfect and there are limitations to what computers can do, it will never be solved, but being reduced is the best we can do.

    ReplyDelete
    Replies
    1. I agree it is a very large step! I also agree that 50 is quite a small sample size. I would like to see this model be applied to a large sample size with other forms of epilepsy as well not just temporal lobe. Agreed, reducing the severity and frequency of this is all we can hope for!

      - Danielle Bermingham

      Delete
  4. I had no idea that temporal lobe resection was still used today for possible epilepsy treatment. After learning about the case of patient H.M. and his memory disorder due to the removal of both medial temporal lobes to cure seizures, I figured it was no longer performed. But that was a long time ago and it sounds like the process for determining if a patient is eligible has gotten a lot more rigorous. Also, the concept and applications of machine learning is incredible and I think it's going to play an important role in the future.

    -Matt Murdoch

    ReplyDelete
  5. This is so cool! I love hearing about these kinds of advancements because they can have such a large effect on so many people's lives. I was unaware that so many people suffered such debilitating side effects due to this surgery, but this is absolutely something that requires further investigation. Based on this study, do you think that it would be possible to extend such results to a large population and to different kinds of epilepsy? How long do you think such advancements may take place?

    Posted by Alexandra McGuire

    ReplyDelete
  6. The different types of epilepsy makes the treatment harder to be satisfying and having no side effects. It is a very normal and good strategy to follow patients after surgery detecting their comfort, pain levels and if any side effects appeared. By mentioning "A lot of patients report post-surgical amnesia, vision defects, and psychological sequelae effects, such as depression" is not really clear to me, i think if we start putting numbers and statistic calculations their, the idea could be more clear because by saying a lot it could be 40% or it could be 10% which matters to consider the procedure good or really bad. The people that were still having seizures after surgery could be due to an undetected abnormal brain activity in another location in the brain (not treated) or could be caused by the complications of the surgery but did these people go back to surgery or try anything else to stop the seizures again? If yes, did that help? No follow up was mentioned for patients after the post-surgical symptoms appeared.

    Posted by Jad Imad

    ReplyDelete