 So next up we have Eliza Barnwell, she's a fourth year medical student from MUSC in Charleston, so if she's from Charleston and did her undergrad in Virginia, she's going to talk to us about pediatrics. Okay, Sarah. The title of my talk today is Developing a Model for Postoperative Axial Length in Children Undergoing Bilateral Cataract Surgery to Optimize Visual Outcomes. This is a project that I worked on with Dr. Ed Wilson and Dr. Rupal Trivedi at the Sturma Institute. So just a little background, pediatric cataracts are the largest cause of providential blindness in childhood. There are numerous etiologies that can be present at birth or form later during childhood. Estimated prevalence ranges from about 1 to 15 for 10,000 children. And like in adults, cataract removal with IOL implantation is found to be safe and effective in children beyond infancy. And so we take biometry measurements of axial length and keratometry preoperatively. And this is very straightforward in adults, but complicated in children because their eyes are still growing, so the axial length hasn't yet reached its final measurement. So what we know about axial growth is that it increases dramatically in the first 18 months of life and then proceeds to grow slowly until around age 18 to 20. Newborns have a mean axial length of 16 to 18 millimeters. Adults have a mean axial length somewhere between 22 to 25. The final axial length is difficult to predict and continued growth after cataract surgery can be influenced by many different factors. So it's hard to know which IOL to implant. So when you're implanting an IOL in children, the goal is to restore normal vision as the eye develops into its adult size. So most pediatric ophthalmologists account for their refractive shift over time by leaving the child hyperopathy by after surgery. There have been a lot of tables developed to buy surgeons in post-operative refractive goals and here are some examples. And as you can see, they're not all the same. So one of the issues is that there's not a lot of consensus in the literature on ideal post-op goals and there's a lot of variability in post-op refraction in kids. So if the eye proceeds to grow more or less than expected after surgery, refractive errors can develop and can lead to four visual outcomes. So if we could better predict how the eye will grow in an individual patient, we could potentially develop a better IOL calculation method. So the aim of this study was to develop a model to predict future axial length of individual patients undergoing pediatric cataract surgery. We wanted to focus on children undergoing surgery after 18 months of age just during the slower phase of growth. I thought this would just be an easier time to predict growth. And then we wanted to use the model to make more accurate IOL power calculations. So this was a retrospective chart review. Pediatric patients who underwent bilateral cataract surgery at the Storm Eye Institute, we included children who had surgery at greater than 18 months of age with at least two consecutive IOL measurements. So one baseline IOL at surgery and then at least one follow-up axial length. And the exclusion criteria were traumatic cataracts and autobial lentus syndrome. The data was collected in Red Cab Database, a secure online database, collected demographic information of gender and race, collected cause of cataract, type of cataract, whether or not there's family history, baseline measurements including age AL, technique of AL measurement, ACD, lens thickness, keratometry and visual acuity, and then follow-up measurements of age AL, technique of AL measurement, ACD, and visual acuity, also the IOL type of location and whether or not there was glaucoma. And this study was, MUSC was kind of well a good place for the study because we tend to get data on axial length at each follow-up appointment. So there was a lot of data for that. And for the statistics, it was done in SAS. We first did univariate associations using GE models, which is a generalized estimating equation. And this approach is an extension of linear regression that's really useful in ophthalmic studies because it accounts for the correlation of the left eye and right eye in the same patient and also the correlation of repeated visits with the same patient. And then the univariate models were used to develop a final multivariable model for predicting post-operative axial length. So here are just some of the patient characteristics that we had. We had 100 total participants or 200 eyes, 58% were male, 42% female, 58% were Caucasian, 35% were African American, and 7% other. The mean baseline age was 6.8 years and the mean final visit age was 12.7 years. The mean baseline axial length was 22.4 and the mean final axial length was 23.3. The median number of follow-up visits was 2 with a range from 1 to 13 visits and the total follow-up time was about 6 years. And this is just a scatter plot of all of the axial length measurements and H points. And you can see it kind of follows a linear trend. So this is the results from the univariate analysis. So the univariate analysis was looking at individual variables and the effect on axial length. And the ones highlighted in yellow are statistically significant. The green, too, I just highlighted race and gender and green because I thought these were kind of the most interesting to interpret. So this is saying that males on average have an axial length that is 0.97 millimeters longer than females and Caucasians have an axial length that is 0.95 millimeters shorter than the other races. So what we did is we took the univariate results and all variables with a univariate p-value less than 0.2 were considered in the multivariable model. So the variables that were considered were age, race, gender, axial length, baseline axial length type of cataract, whether or not they had glaucoma in the age of follow-up. So in the final model these are the variables that were found to be statistically significant. So baseline, axial length, age of baseline, age of follow-up and interactions between baseline age and age of follow-up. And what's kind of interesting is that race and gender were not statistically significant in the multivariable model and we think this is the case because we think that by including the baseline AL and the baseline age we're accounting for those differences in gender and ethnicity. And it's really easier to make sense of this if it's written as a linear equation. So this is kind of the prediction model written out as an equation. And I just came up with some hypothetical examples just to show you how this equation would work. So if we have patient 1, the light blue line, this is a patient, say at 2.5 years has a baseline axial length of 20.5 and then we find the predicted axial length at age 18 to be 23.1 millimeters. And you can see the other two patients have different ages and different baselines and they have different final axial lengths at the same final age. So kind of the usefulness of this equation is that it can be used in a IOL calculation, IOL calculator instead of our current method which is that most pediatric ophthalmologists enter the preoperative AL and K into IOL power calculation formula and then they adjust the power based on the age of the patient to account for the myopic shifts so kind of like the table I showed at the beginning. But an alternative method could be to use this model to predict the future axial length of the patient and then enter that predicted axial length into the IOL power formula and we think this could potentially be a more accurate and customized method of IOL calculation in children. And just one thing to note, the keratometry is the less important part because it doesn't change that much after one year of age. You kind of reach your adult keratometry around one, so that was not as important as predicting the final axial length. So some of the limitations of the study is a retrospective chart review so a lot of factors couldn't be controlled. The follow-up time for patients and time between visits were variable and also not all of the patients were followed into adult. So only 13 had final measurements taken greater at 18 years of age. And it would also be useful to know the refractive error of a child's parents because there's definitely a genetic component to myopia in axial length so that would also be useful information. I think the strengths of the study are just one of the main strengths is the amount of data we've collected on axial length over time. MUSC has really done a good job of making sure to record the axial length at follow-up visits and that data is very useful. This is the first study to attempt to predict individual post-operative axial length to improve idle power calculations. So just some future directions. We want to perspective test this model by comparing the predicted AL measurements to actual AL measurements in patients following cataract surgery. And we also, if any other institutions have data on axial length, we could validate the model that way. And it would also be nice to go back to our data set and look at the patients who are now old at the adult age just to have more complete knowledge of the axial length once they reach adulthood. Are there any questions? Yes. Very nice presentation. I don't know if anybody's looked at this. I just got a curiosity. Have there been any studies where they've actually gone back and looked at how many of these children go on to have lenses and or refractive circuit down the road that they want to like due to an incorrect power calculation? That's a good question. I think that's something we could maybe look at in our data set. We haven't done that yet. Dr. Ed Wilson has done that extensively with kind of within the infant aphakea treatment trial but also on his own. Scotland Works published a lot on that too. And while the general trend for those tables that Eli has nicely put up is that most people don't require a lens exchange we have these outliers, these patients that, you know, they'll be minus 20. And they started the same way that the other patients did. And so I think that is why there's this need for a more individualized formula for children. But I'd like to point to you about knowing the parent's refraction. And I think one strength of this study was to consider bilateral cases because in a unilateral situation you get inappropriate axial growth from like deprivation or ambiopia. So this is more representative of the typical course that the child would have. So let's leave to see. So why would those outliers like minus 20, why do they speculate? And would they be minus 20 anyway? Or was it somehow impacted? Any idea? Well, in some cases there is increased axial length outside of a normal adult range. And in other cases it might be within, so that's most of the cases and so it's just unclear why that is or how we could have known that initially to better power the IOL. But then in some cases you'll still see axial lengths within the appropriate range but because they started with a very short axial length we just were not able to predict appropriately what power it was in. And you'll see, I mean we do have challenges too. I'm doing a two-year-old today, a unilateral case and I'm going to make one eye plus five. You know, because I'm trying to prevent myopia when he's older but then he's so inisomatropic in the short term. So if we could limit that in the short term if possible that would be helpful too. Just curious how they measured AL there. Is it all optical or immersion every time the kids get anywhere? So this was immersion in some of the older kids because IOL master. So these kids were all coming back for EUAs and then having immersion done again? Some of them were not there in Asija. It's corporate kids. They do immersion in the corporate fixing now. Did you study actually look at that or did you sort of say it didn't change? Yes. Some studies have actually shown that it's a fair amount of change. We did. It's not the first engine of ours. We did look at it. But in this study since we were looking at kids older than 18 months of age it was kind of if there wasn't a significant change we didn't collect it. Thank you.