Lecture: Innovations and Advances in Retinoblastoma Management

>> JESSE BERRY: Thank you all for joining this morning. It is early here, as you can see, when I give these lectures it’s always sometime reasonable around the world and very early in LA, and I always hope not to wake my children up. I’m so grateful to all of you who are joined for this discussion. I’m Jess Berry. I’m going to talk to you about updates over the last several years that are moving retinoblastoma towards more precise and personalized medicine. These are my financial disclosures. I do not have any interactions with any companies that sell products to patients. Do I however have research funding as well as a provisional patent on some of my aqueous humor research which I will discuss today. Today we’re going to go over diagnosis, management of disease in retinoblastoma, which is still a hotly debated topic across multiple centers across the U.S. and around the world, and some innovations in the field, including a more focal delivery method for chemotherapy in the intravitreal space, using OCT for retinoblastoma, its applications within ocular oncology are evolving. A brief review of retinoblastoma, as you know, it’s the most common primary intraocular tumor in children. Thankfully it’s a rare disease, there are about 300 cases in the U.S. Between one in 15,000 to 20,000 live births. In general it presents with either leukocoria or strabismus. Generally one eye often sees well, the child appears to have normal visual development and a parent or caregiver notices something going on in one of the eyes. There’s no male-female predilection or racial predilection. In general the average age at diagnosis is 18 months. There have been rare cases of five, six, seven-year-olds or even adults, although again, quite rare, we think that’s a somewhat different clinical scenario. And retinoblastoma is a genetic disease. It is in fact one of the main genetic diseases that ushered the idea of cancer genetics. The RB1 tumor suppress sore gene, my director played a huge role in that, Lynn Murphy, as well as others. The thought was that honestly, this was just nuts and that cancer was only an activating mutation and that there was no idea of something like a cell cycle regulator, which is what the RB1 tumor suppresser is. Now of course we know there are multiple tumor suppresser genes and in fact while the retinoblastoma tumor suppresser gene is the initiating event for retinoblastoma in the vast majority of eyes, it also plays a role in a vast majority of human cancers. Now, all children who have bilateral disease, meaning retinoblastoma in the right and the left eye, carry a germ line RB1 mutation which means the cells in their body have this alteration and they are predisposed to retinoblastoma as well as to other cancers throughout life. But 15% of unilateral patients also carry a germ line mutation. And so testing patients, but particularly testing unilateral patients, is critical for the management of this disease. Even now, even in the United States, I have patients who are told you were unilateral disease, there’s no way you can carry the germ line mutation and then of course they go on to have a child with retinoblastoma or they go on to develop a second cancer and no one can quite understand why. If you take nothing else from this talk except for that, I think that’s one of the most important and critically often missed aspects of the care of retinoblastoma patients. Now, as I mentioned, anyone who has a germ line mutation can develop a tumor in the other eye and they can also develop tumors in their body throughout life. They can pass that alteration down to future offspring. And they also, we think, have increased risk of second tumors in there’s radiation. Thankfully we don’t use radiation quite as much anymore with retinoblastoma care, but it does remain a concern for some patients. And the way that we at this point in time, in the United States and around the world, still diagnose retinoblastoma is based on clinical features. So these children are being examined under anesthesia for clinical diagnosis of disease. There are several features we’re looking for, including buphthalmos. We’re looking for seeding, small pieces of viable tumor that break off into the vitreous or subretinal space, and can break off and develop other tumors in the eye, and then calcification. Retinoblastoma is not the only thing that causes calcification in the eye, but in a young child, in the setting of a dome-shaped or retinal-shaped lesion, absolutely retinoblastoma should be at the top of your differential. This also varies center to center. But the two most common classifications for retinoblastoma are the international intraocular retinoblastoma classification or the IIRC. This is the ABCDE classification that is used most commonly. Many centers are either adding on or progressing to add the AJCC, so that we as people who take care of retinoblastoma are talking in the same way as those who take care of other cancers, using the AJCC for staging. As you can see, the disease progression goes from A through E where A eyes are very small tumors and are at restricted sites, not close to the phobia or the optic nerve. B are bigger or closer tumors. C tumors have seeding, so localized seeding either in the vitreous or the subretinal space. Group D tumors have diffuse seeding to the vitreous or subretinal space. Group E eyes are generally considered eyes that are filled with tumor where the tumor has destroyed most of the structures responsible for vision. Again, it is arguable, but many people feel that an E eye is an unsalvageable eye and there is some danger to the child by trying to save these eyes. Now, you’ll see the vast majority of children, even in the United States, about 80% present with group D and E. It’s quite rare to get an eye that has a smaller tumor burden. And of course in other places, children may present far more advanced, so beyond an E where the tumor has actually already escaped the eye and is in the or bit or extraocular space. And so currently, our prediction of which eyes will respond to therapy generally prediction about 50% success in group D and E. And again, you can see here, A, small and far, B, bigger or closer, C for seeding, and D for diffuse. That’s how my residents often remember the ABCDEs. We love to talk about eye and vision but the number one important thing to consider when taking care with a child with retinoblastoma is of course their life. Thankful in the U.S. and Europe we have a greater than 90% likelihood of being able to save their life but as children progress with disease outside the eye in the extraocular space, that means their prognosis for life changes. Retinoblastoma is a global disease and it’s important that we always remember the life of the child is more important than the limb of the child and that is our number one go home, to save that child’s life. Let’s talk modern treatments. Enucleation, as long as I’m invited to give these talks I’ll leave enucleation on this list because it remains important for retinoblastoma both in the primary and secondary setting. It is a life-saving surgery. Attempts to salvage the eye from retinoblastoma require one or more forms of chemoreduction. Wall talk about both systemic and intra-arterial chemotherapy. Tumor consolidation may include laser or cryo or brachytherapy. Intravitreal chemotherapy targets seeding. And external beam therapy is rarely given, not never but quite rare. All of this of course presumes intraocular disease. Once you have disease in the orbit, outside the eye its or in the CNS space, you’re discussing a much more intensive multimodal Reg. Enucleation, one-hour surgery, it’s thought to be curative in 96% or greater of cases with a low risk of recurrence in the first 12 months. These days as treatments have become more successful, this is often reserved for failed treatment or secondary enucleation. In very bad group D’s or E’s, we will on any offer enucleation, it avoids a lot of anesthesia for the children and they’re happy both cosmetically and functionally with the result after an enucleation. For salvaging treatment, again, it requires one or more forms of chemotherapy. And the reason I say one or more is that very frequently you may start with one form and due to response or due to things that happen in the other eye, you may transition to another form at least at our site. This is either systemic or intraarterial. Generally alongside this treatment you have examinations under anesthesia every four to eight weeks and as I mentioned, multiple local therapies. Retinoblastoma tumors will show recurrence regardless of treatment. The aim is to identify these recurrences when they’re small and amenable to local therapy. And while there is debate in the retinoblastoma community regarding which modality is better and when to initiate them, both systemic chemotherapy and intraocular chemotherapy have pros and cons. This is old now, but I like the ability to show this graph, a review of outcomes by group A through E. As you can see, again, 80% of patients do present with more advanced D and E eyes. This is the area in which systemic chemotherapy starts to have lesser efficacy and intraarterial has better efficacy with more advanced eyes. There are local complications. For systemic chemotherapy we hear a lot about systemic complications. These children have to be monitored. Interarterial chemotherapy, which uses generally different forms of chemotherapy can also cause neutropenia. Local complications are vascular, when you’re delivering the chemotherapy through the artery into the eye. You can see bradycardia that needs to be management. Thankfully rare complications such as avascular retinopathy. Metastatic disease, the risk of metastatic therapy is still an issue that we are grappling with, because again, our number one goal is the life of the child, specifically over trying to save the eye which may have very little function. All of these photos is me giving laser to a child, reproduced with permission from the parents. This is showing an RB1 mutated child. You can tell that by all of the separate tumors coming up. This child had multiple, multiple, multiple sessions of laser to control each of those small tumors. Does the mode of treatment matter? Yousef et al. found a 2.1% rate of metastatic disease for patients who had been treated for retinoblastoma. Other reviews suggested that it’s possibly higher in children who were having treatment with intraarterial chemotherapy, in the range of 4, 5, 6% of these children. That being said, we also looked at the risk of metastasis in children treated with systemic chemoreduction, and the rate was similar. Even after enucleation, the risk of metastatic disease is never zero. As a community, our goal should be zero. While we might not attain that goal, it’s very clear that if retinoblastoma is removed from the child before it spreads, we can save the child’s life. Now, knowing exactly when to remove that eye or getting the child into our care when we’re still able to do that, that’s a work in progress and it’s a moving target. At least for me, that’s our goal, zero percent. Because of that, with a goal in mind, then you need to start thinking about what is the treatment goals versus the risk to these children. This is especially relevant with new treatment modalities. What is the risk to the child of treating an eye like this that is full of tumor? You can see there’s neo-vascularization of the iris, the pressure is over 40. Is this an eye we should treat or remove? Part of the complicating issue with retinoblastoma is all of White House have treated this disease have eyes like this, an eye that’s filled with tumor and responded very nicely to treatment and the child is now 20/25, good visual function, and alive at well, even though at diagnosis there was quite a burden of tumor. How do we identify which eyes do better and which may actually have some functional vision? Not only just a diagnosis but we also have to think about recurrence. This is a child, a familial case, came in very early to us, diagnosed with group B, did very well with treatment, you can see some waxy changes, definitely an area you would need to watch. Very quickly there was an aggressive recurrence about six weeks later. This is a fair amount of tumor growth within a six-week period. You can also have a recurrence in the setting of vitreous seeding. Thankfully we have a much better treatment for this even than when I was in training. You can see nice calcified tumor, this is a good treatment effect here, but at some point started to develop a mass. These seed quickly. This eye is filled with several active vitreous seeds, all of which can lay down on the retina and develop new tumors. But outside of even intraocular recurrences, there’s extraocular recurrences. This is a child with an intraocular recurrence and developed an orbital tumor. We published on this, but we don’t know for sure, maybe, you can see some thinning of the sclera, maybe the tumor growth itself combined with the aggressive laser therapy helped the tumor escape into the orbital space, you can see that mass right there. And this was a scary case for me, this was another patient who, again, had a nice treatment effect, then developed a pretty large recurrence. Obviously this is in the macula, so the central visual function was not very good. There’s actually a clear space clinically between this tumor and the optic nerve. The decision was made to stop treatment at this time and perform an enucleation. Even though there was this clear space, there was actually invasion of the optic nerve here and you can see retinoblastoma outside the nerve which of course changes the treatment protocols as well for these children. While this may be our desire, it’s definitely my desire for every single child with retinoblastoma, the reality is that sometimes these tumors are too advanced and it’s important as a community that we balance all innovation with safety of the child and the reality of this disease. So I’m going to move now to talking about some of our innovations. Again, over the past five or so years. I’m going to start with treatment for seeding. So seeding in the vitreous and/or subretinal space was previously very difficult to treat. It was the number one cause of secondary enucleation after failed attempts to save an eye. You would just get to the point where you were chasing yourself because you would have all these seeds in the vitreous and they were falling down and making all these tumors in the retina. There just wasn’t a good way to get ahead of the cancer. At that time the main treatment for vitreous seeding was external radiation. Retinoblastoma is generally very radio sensitive. You could give a curative dose to the eye, a dose that’s relatively well-tolerated by the eye. This was a time when cataract surgery was not routinely done in retinoblastoma patients. Because of the second cancer risk, it’s almost completely contraindicated in children under 12 months of age. Even in older children, there is a risk that this can be a site of second tumors in the future, and frankly they’re thought to be worse than the retinoblastoma itself. And so a little over ten years ago, Francis Meunier introduced an approach to chemotherapy in retinoblastoma. This is an example of that being done. One of the first approaches was to take the aqueous out of the eye to lower the intraocular space. You had already done a UBM to make sure there was no ciliary body invasion in this area. I’ll play this one more time. You can see you cryo at the injection site so if any active seeds are coming with you, you would kill that. Dr. Meunier was not the first to suggest this, but often there were implications. This safety-enhanced approach allowed us for the first time to be able to give chemotherapy into the intraocular space directly. Of course as ophthalmologists we’re using to giving intraocular therapy for other reasons. However all treatments have implications, as you can imagine. Let’s talk about the goals of the therapy. The intravitreal injection is meant to target the seeds. The yellow is the tumor. You can see this is part of our technique here, you’re doing the paracentesis, you’re marking the site. When you inject, the needle is visualized behind the lens and you see the cryo. The goal is pretty impressive. Removal of all visible vitreous seeds. These either go to invisible or calcified. It’s a pretty impressive things. These eyes don’t even look the same. This is a patient who had a baseball of seeds and they all went away. Salt and pepper retinopathy at the periphery is also common. This is a home run success for this eye. Unfortunately we found that in some patients, at some times, generally after the first injection, they developed a toxic hemorrhagic retinopathy, blood and thunder everywhere, retinal edema. We still don’t actually know what is causing this, whether it has to do with the dose or whether specifics about the patient population. I encourage anyone to continue to study this, several centers have. We proposed a hypothesis this, as we show, the retina, the chemo going into the eye. After the first injection, we know from adult studies that oftentimes there’s a vitreous detachment. We hypothesize that potentially after the first injection you inadvertently inject in a tighter volume, thus causing a retinal necrosis. The ocriplasmin study in adults supported that. Our review of our patients, both at the time of injection, where there was not a visible PVD, versus after where there was, in these two patients, suggested to us that maybe this played a role. Here is an example of how dramatic the toxicity is in these patients. This is a nice intact fulva. You can see the significant retinal atrophy there. We actually did not find a link to vitreous detachment and this toxicity. So we as a community are still trying to better understand and predict which children may respond very well. That’s our goal. Let’s talk about OCT for retinoblastoma. So you saw this eye before, this was a very advanced eye. Again, neo-vascular glaucoma, pressure 40. At diagnosis, other eye looked pretty good so we went forward with a primary enucleation. Blood was sent for RB1 testing. Then the patient came back. I always test my residents and tell them if I’m going to show them this picture, there’s a reason for it, and to see how many tumors they can pick up. I’ll have you test yourself there. When we looked in, it looked pretty good. Here is what the OCT showed. There’s tumor 1. Tumor 2. And tumor 3. I’ll play that one more time, I know it goes fast. Right above the optic nerve. Here we go. Tumor 1, tumor 2, and tumor 3. There are in fact three small tumors in this eye that are so small to almost be invisible to the naked eye. They’re barely distorting the retinal architecture. That makes them not so visible to the human eye clinically. But you can detect them on OCT. We were able to give just focal laser therapy for those three small tumors and that patient did very well. You can also use the OCT to monitor for scar recurrence. This was a patient that overall looked pretty good, you can see a little bit of thickening at the top of the tumor but was responding well to the therapy. You see this heterogeneous mass here, although one week later you can see that there’s an alteration there. This is becoming less homogenous. I love this picture, this is a nice vitreous, spherical seed here. These can be more resistant to therapy because the sphere itself is encapsulated, so a little less sensitive to the chemotherapy. It continues to let seeds off into the vitreous space. And finally, I’m going to end with a discussion about my work, was liquid biopsy for retinoblastoma. And so as I’ve mentioned all along in this talk, retinoblastoma is diagnosed clinically. Recurrences are diagnosed clinically. And we monitor these eyes clinically. This is not a tumor that is routinely biopsied, that’s due to concern for tumor spread, not a hypothetical concern but a real concern based on inadvertent attempts in children where they either thought it could be done safely or where this realize there was retinoblastoma and it caused the tumor to spread. While ocular oncologists are quite skilled at looking at the eye, a lack of biopsy actually has impact. This is why we biopsy other solid tumors including all pediatric solid tumors and even other ocular tumors because biopsy can be helpful. In retinoblastoma, our lack of biopsy means we have a lack of eye-specific diagnostic and prognostic markers for this cancer. Because we don’t have those markers, we have a lack of personalized treatment plans. Our treatment plans are based on the age of the patient, the group of the disease and the laterality of the disease. We know a lot about how tumors are formed from tumor specimens in eyes that have been enucleated. But how tumors are responding to a treatment at the molecular level and at time of recurrence is not well understood. So we’ve already gone over this, but intravitreal chemotherapy injection started in 2012. There was a lot of anxiety as it related to starting to put a needle into an eye with active retinoblastoma even with the safety-enhanced mechanisms. But it’s been well-studied, and in this pretty large review done now back in 2017, there were no documented case of extraocular spread using a safety-enhanced approach. And so this really led for us to a new opportunity. For the first time, we were putting needles in eyes with retinoblastoma and we were extracting aqueous humor from those eyes at that time. Initially we were looking for cells in the eyes or after, just throwing the aqueous away, until we had the idea, hey, maybe there’s something in there, since we’re already taking the aqueous, maybe we can study this to see whether or not it may overcome our lack of tumor tissue in these eyes. So liquid biopsy is a term that’s thrown around a lot in the scientific and also public media. What really is a liquid biopsy? A biospecimen that is a mix of cellular or noncellular components that are derived from the tumor. And from those components, you can identify specific things like cells, proteins, nucleotides, various things. DNA, which is I’m going to focus the rest of my talk on, is the main one. Most people talk about blood as a liquid biopsy. The idea that it’s shedding into the blood. But there are also proximal liquid biopsies like CSF, saliva or plural effusions, urine has gained hold as a liquid biopsy. I joke all the time that they need to update this slide now to include aqueous humor for the eye. While a tumor biopsy is one little piece of this large tumor, it gives you a snapshot, you can see pieces of it but you might not see what the whole picture is. A liquid biopsy gives you the whole picture, all the molecules from the tumor that are shedding into that biospecimen. Importantly, liquid biopsies are repeatable. A liquid biopsy can be done over time and you can see how that tumor is changing under therapeutic pressure at the time of recurrence and even what might be driving those changes. Liquid biopsy, at least as it relates to DNA, generally are looking for DNA in one or two mechanisms. And I’m just going to touch on it briefly so that we all know the lingo as I move forward through some of our slides. So whole genome sequencing means you’re looking at all the chromosomes on the genome, generally you’re looking at whole pass sequencing, you’re not sequencing the whole genome but .1 or .3 of it. Because we have a reference genome, you know the sequences, we can bin them, put the DNA that’s been sequence income those bins, you can count those reads. You can identify if there are too many or too few copies of chromosomes in those regions. SCNAs is a biologic mechanism, these chromosomal alterations can identify tumor DNA. You can look for specific genetic alterations and a variety of cancer genes, even from one sample, you can look at more than one. For retinoblastoma, of course targeting sequencing makes sense for the retinoblastoma tumor suppresser gene. Here he need to do really deep sequencing, we’re talking like 500 times sequencing in that gene region. To do that, you can get down to a very low tumor fraction. You can detect these alterations to about .1. Again, in this case you’re looking for single nucleotide variants or alterations in the gene. These are also ways to detect tumor DNA in the same way we have two ways to detect retinoblastoma. We do an EUA, we look at how the whole eye has been altered, and what compartments in the eye are affected by retinoblastoma. But then we also combine that with MRI imaging to look at the optic nerve or the CNS areas outside the eye that we cannot see. So for us at Children’s Hospital Los Angeles, we developed a profiling workflow whereby we took the ocular humor, we were able to identify cell-free DNA in the sample and then do low pass gene sequencing. Here is an example of a normal cell, noncancer cell. This is an example of a cancer cell that has this altered genome with this altered genomic profile. You can see areas of too many or too few copies. From this we were able to answer some really important questions as it relates to liquid biopsy. The first is quite simple. Is there tumor DNA in the aqueous tumor? The answer is yes. So we and others showed, we were the first but others followed suit, which is great, to show as proof of principle that you can detect tumor DNA in the aqueous humor. There was lots of press around this initial discovery. And so this was an example of three eyes, two underwent primary enucleation, tumor is shown in red. The alterations in the aqueous are shown in blue. Case 3 was the center for us, this is an eye where we were attempting salvage, we were taking the aqueous humor at each injection. Just look at the shape, this is a relatively complex profile of gains and losses. But that shape is stable over time. Something is contributing that DNA. Of course it’s the tumor in the eye. You’re getting the stable profile over time. Unfortunately this child did not respond well to treatment so enucleation was done. It then allowed us to show that that profile we had been detecting all along was in fact the same profile that’s at the level of the tumor. And in fact, we were also able to show that this is consistent with a signature for retinoblastoma. So this signature had previously been shown from tumor from an enucleated eyes, this is nice work from Erskine Koi, showing areas of common gains and losses in retinoblastoma. Again, pay attention to the shape. You can see we were showing similar areas of gains and losses in our work. And so with that work, we started to move forward into making this more applicable for testing, discussing how we could use this more frequently in children with retinoblastoma, and to really study what kinds of information you could find from these samples to move it towards the clinical space. And what we found is less than 5% of aqueous humor samples are removed for quality control. Greater than 95% of samples at all times have some DNA at a level that’s detectable and can be analyzed. At diagnosis, 98% proven DNA. Now, if you take out group A eyes, the tiniest eyes, and include only tumors that are 3 millimeters or larger, we actually found 100% of eyes at diagnosis have tumor DNA that’s detectable in the aqueous humor. So the smaller volume eyes might not but group 3 is 3 millimeter eye, that’s a small eye. Here are examples of this aqueous humor. You can see these are the kinds of alterations you might see in aqueous humor from retinoblastoma. And also that they are different, right? You’re not getting the exact same molecular profile each time. That’s important. That tells us that these molecular alterations are in all likelihood, just like other cancers, driving these tumor behavior and tumor activity. So our understanding of them, as it grows, we may be able to more clearly personalize medicine for these patients. Now, these patients also underwent RB1 testing. If we were able to identify them, they were here. This is a report after our first year of doing this with a diagnostic aqueous humor, and very soon we will have — we have had a five-year report released and very soon we’ll have an update to that report. So I’ve already given you the answer to this, but can the aqueous humor play a role in diagnosis? Of course the answer is yes. Not only is the DNA detectable, but it can be used as a companion diagnostic, both through alterations but also the more specific RB1 mutations. You can detect these RB1 mutations in the aqueous humor and specifically test for them in the blood. That makes your test for germ line RB1 alterations more sensitive because you know the alterations in the tumor. The aqueous has high enough fidelity that you can identify if there are two mutations, or the common mechanism whereby you get the loss of functional protein. There are lots of clinical questions this can help us with. We, again, as a community are pretty good at identifying retinoblastoma and retinoblastoma recurrences when we can see into the eye. But when we cannot see into the eye, this is more difficult. So whether it’s from cataract or for hemorrhage, it can be quite difficult. And in all of our hands there have been eyes that have presented somewhat in an atypical fashion, maybe there’s blood, no history of trauma, there’s mass in the eye with a little bit of calcification but not a ton. And so in the past some of these eyes have come to enucleation simply to be safe and to make sure that there wasn’t retinoblastoma lurking inside, but now I think we have a test that can help us. Can the aqueous humor play a role in prognosis for us understanding which eyes may be more likely to respond to therapy and more likely to be saved? The answer again here is yes. And this really came from the operating room for me. As a surgeon, I thought it was so frustrating that two kids would come in and look exactly alike, they would have the same burden of disease and we would give them the same treatment. One child would do really well. This tumor, almost after one cycle, would calcify and look amazing. The other child would have persistent and recurrent disease again and again until we ultimately had to do an enucleation. I wondered if we could augment therapy so there was more aggressive therapy for the more aggressive phenotype or to choose an up-front enucleation. And so we were able to, in one of our very early papers, for the very first time, compare DNA from eyes that were saved and tumor DNA from eyes that were enucleated. Now of course the reason this had never been done before was prior to using the aqueous humor, you only had DNA once the eye was enucleated and you could access the tumor. But with the ability to compare these two groups, the red are the eyes that were enucleated and the blue were the eyes that were saved, you can quickly identify there is a difference here chromosome 6P, we found that gain of chromosome 6P was not only more common in eyes that were enucleated but the amplitude, the number of copies of the 6P gain were higher in these eyes. You can see here with the blue. Not that it’s not present but the number of copies, you might have one extra copy or with a mosaic or clones within the tumor, I should say, different clones, let’s say you get one extra copy or less, but in these eyes that did poorly, you’re talking two, three, four extra copies. And so does that matter in terms of our ability to save the eye? Well, at our two-year outcome, we showed that the likelihood of saving the eye was far higher, if 6P gain was not present. So these were the eyes that in our hands tended to do well versus eyes with 6P gain that did not. Doesn’t mean we never saved a 6P eye or that we only lost eyes that had 6P but rather that there was quite a clear difference in the prognosis of these two groups. And in fact, if 6P gain was present, it was associated with a ten times increase in odds of that eye failing treatment and needed enucleation. Can the aqueous play a role for monitoring therapeutic response within the eye? Again, we found the answer is yes. I love this, people who have come to my talks have seen my present this case before, but I really like this case. This is a patient with an active disease here, active seedings, retinal tumor and active seeding. We’re treating with injections as well as laser therapy. You can see a very altered chromosomal profile. This is your second take-home point from this talk. You see a profile like this, cancer is active, there’s cancer DNA here. You have a very altered profile. As we continued the injections, the profile starts to normalize. This is tumor fraction. The tumor fraction or the percent of tumor DNA in the aqueous humor starts to plummet. Clinically the eye starts to do better. Everything is looking good. At injections 4 and 5, clinically the child looked quite good although if you look, the tumor fraction is starting to go up slightly. We are starting to see some of these alterations pop back up. At visit 6, you can see a large recurrent retinal based tumor and a highly altered genomal profile. 4 and 5, at this point it wasn’t just me, there were two ocular oncologists, we both thought the eye looked stable and was responding to treatment. But at the molecular level, the aqueous humor had information that we as clinicians just looking in the eye did not, and it was portending relapse. You can see that right there. Now, does blood play a role as a liquid biopsy? Absolutely. There’s been a lot of good work on this. This is a really nice review from Amy Garish. She put together this review talking about liquid biopsy in general with a focus on aqueous but also blood. Our colleagues at Sloan Kettering Cancer Center have been leaders in the space as it relates to blood. Now, blood based sequencing, again, in this case you’re really looking for a needle in a haystack. The whole genome sequencing, this low pass whole genome sequencing is not very functional in this setting but instead you can do targeted sequencing. What’s been found and presented so far is that in advanced eyes, there can be some detection of mutated tumor derived RB1 in the blood. You can detect that. And that detection plummets after treatment. Now, whether that’s enucleation or whether that’s one cycle of in this case interarterial therapy, after that you can generally no longer detect it. If you can, it often means that that child is at risk for development of metastatic disease, so persistently elevated signal in the blood should suggest there may be something going on at the level of the body. The question always posed to me, which is good, is can’t you just use the blood? Can we not put a needle, even a small needle the size of a human hair, into a child’s eyes, can we just use the blood? So far, my answer to this is no, and there’s two reasons. The first is that we did a head to head comparison of aqueous humor in blood in 20 patients. We found only the aqueous humor had high enough tumor fraction for this whole pass low genome sequencing. While you may be able to detect there is tumor DNA in the blood, from detection of an RB1 alteration, if you want to look at these more prognostic and cancer-deriving alterations, you need a liquid biopsy for that. In this case it would be the aqueous. Furthermore, more than 40% of children have disease in both eyes. So here is one of — we now have many examples of where a child has a right and a left eye involved with retinoblastoma. And the alteration than you see in those eyes are different. This likely is why these patients have a different response overall to therapy. And you cannot find these in the blood. Are there other eye specific biomarkers in the aqueous that can teach us about cancer? The answer is yes. Of course there is RNA, other molecules that can be identified. One that my partner in the lab is really excited about and really working on is these extracellular vesicles. This is just aqueous humor unprocessed. You see all these vesicles floating around. The kinds and types of vesicles we see in cancer eyes appear to be different. Likely these vesicles are also playing a role. You can see here about the size and quantity of these vesicles. And you use antigen profiling, and we found there was a difference in patients with retinoblastoma and also found that CB63 vesicles seem to be predominant in the eye although they are not present in many other body fluids, so it’s likely an eye-specific population of vesicles. So I believe that analysis of cell-free DNA in the aqueous as well as the blood will really be the basis of the future of RB. And I hope that it will let us catch up in all of the ways in which personalized medicine has benefitted other children with solid tumors, where you identify an actionable alteration and you change the way you treat the child based on that. Here is just an example of the future of RB care with the aqueous humor. You can see here, this child had a large group D tumor. And at diagnosis, this is presenting funny, sorry, at diagnosis there is 6P gain here which is quite large. This 6P gain persisted. This was a child that did not respond very well to therapy. They were treated with intraarterial and then intravitreal and they required enucleation. This child was treated with systemic chemotherapy and also intravitreal and responded very well to treatment and this eye was saved. So we now have a clinically available laboratory test at CHLA. We get samples from around the country and the world to answer clinical questions as it relates to retinoblastoma and our ultimate goal of course is to improve patient outcomes through discoveries and cancer care. This goal is in full effect and we’re making progress for these children. I have a huge aqueous humor biopsy team and a clinical team that helps me do this research work as well as the clinical work. And so I’m speaking on behalf of all of them today, as well as CHLA and USC, our institutions who have been super supportive of this. Here are my resources which I’m happy to share with anyone. And thank you very much. Happy now to take questions after that big tour de force on retinoblastoma, from any of you. I can’t see the questions right now. Do you want to read if there’s any available? >> The first question is, now we need to do diagnosis between retinoblastoma and Coats’ disease, yes, we’ve published on that, I can email that to you, if you want to email me. We’re able to show for retinoblastoma and Coats’ disease, we are able at a molecular level to distinguish them. So no patient who had Coats’ or fever or any simulating lesion had any DNA in the aqueous humor. So we have very high fidelity for helping diagnose Coats’. The surgeons really want to know that this is not a retinoblastoma eye, so it’s been very effective here. Let’s see. The next question, in Vietnam, ophthalmologists may still not be able to identify the difference between retinoblastoma and Coats’ disease. That’s not true just in Vietnam, that’s true around the country and the world. We’ve had times where, you know, there’s not one ocular oncologists, we’ve had two, and two pediatric surgeons, and we’ve had cases where we couldn’t surely tell, we weren’t entirely sure. It can be difficult in some cases. Someone asked, what type of cells do you have in an retinoma. If by that you mean a benign retinoblastoma, retinomas are not yet cancer, they will not show the same alterations. To be clear, we’re not identifying cells 2009 aqueous humor, we’re identifying cell-free DNA that we can sequence to see those alterations. So how can we make the risk zero? That’s a great question. You could come to the international society of ocular oncology and we could debate this forever. I think the way we make the risk zero is twofold. One, we need to get these kids in to us earlier. Even in the U.S. we’re seeing advanced disease in the intraocular space. In other places these kids aren’t coming in until it’s too late. That’s one, these kids need to come in earlier. That involves public health, education campaigns targeting parents and caregivers. That’s one. Number two is, we need to, as a community, decide that if we treat an eye and then that child develops metastatic disease, that was a failure on us. We need to be honest about it. And I think we need to say our goal is zero. And we’re all going to learn together and do better to not treat eyes that are potentially dangerous and have the potential of developing metastatic disease. I’m not sure we’ll get to zero. We probably won’t. But I think if our horizon is zero and if that’s our goal, the patients with retinoblastoma will only benefit from that. Which intravitreal chemotherapy is used fell melphalan is gold standard, that’s what’s used for most patient, although melphalan is what had the toxicity I showed you, and that’s pretty well described across multiple centers. In some patients that are monocular or have signs of toxicity, topotecan can be used instead, it’s thought to be a little more gentle but conversely a little less effective to the eye. Some people have hypothesized that a darkly colored fundus may have a higher risk of melphalan toxicity. But the exact eyes that are going to have it have not been well-delineated. I have not used carboplatin, it’s not been well-described. Is it possible to combine the intravitreal and the intravenous therapy? Absolutely, you start the child on systemic chemotherapy and generally after cycle 2, you start to inject in the eye so that you have two mechanisms of treating these tumors and to decrease the tumor burden. I’m not the PI for that study, although CHLA is the leading center. Even outside this study, we use that approach to therapy, particularly in children with bilateral disease, so the systemic chemotherapy goes through their body, treats both diseases and we can augment that therapy with intravitreal and we’ve had success for that. It can be more difficult to treat retinoblastoma in children who are far less likely to be intraarterial. We talk about intraarterial as this wonderful success in the treatment of retinoblastoma, and it is, but there are some procedural considerations in play for intraarterial in a way there are not quite so much for systemic. A very small premature child is one of those. Children who are premature are often undergoing ROP exams, so the premature patients I’ve had, they’re actually detected quite early. When administering intravitreal medication in infants, how far behind the limbus do you go, considering the plane is smaller than adults? The paper by Dr. Francis Meunier, he talks about that. If you have a premature child or if you were injecting a very small age, you might want to lower from 3. ROP babies are more at 2, 2.5. But in general, around 3. Which imaging is preferable for the diagnosis of retinoblastoma? And in general, MRI is preferable, that’s because, again, many of these children have a heritable RB1 mutation in their body, so we like to avoid radiation in these cases. CT of course does have some radiation in it. That being said, if CT is the only modality that’s available or sometimes we just don’t know entirely what diagnosis we’re looking for and we need to identify calcification, so CT is sometimes used, but the preferable is MRI. Would it be appropriate to liquid biopsy? That’s a great question. Right now we only biopsy eyes we see something in. If we see Coats’ or PFE or pseudo retinoblastoma and we want to rule out retinoblastoma, we would biopsy that eye. If retinoblastoma is in two ways, of course we would biopsy both of the eyes. But in a completely normal eye, we would expect a completely normal test, right now there’s no standard to do that. If there’s another followup question to that, please shoot it in there. So this is a great question, this is from Marie in Colorado. Do you follow patients with 6P gain more closely? Right now our treatment approach is not different. This is an active area of research both for my lab as well as my clinical team. We do follow these patients quite closely. We just recently actually had a patient who — this was before our use of the aqueous humor clinically, when it was still only in a research sphere, have a late recurrence of retinoblastoma with high level 6P gain. And had we been in a setting where we had the clinical test, we absolutely would have followed this child longer in the OR, right? You see these kids under EUA, you can just see the eye so much better. So we don’t move these patients to the clinic quite as quickly. If the cancer reaches the brain what will we do for the child? Again, the COG studied that, it’s multimodal high dose chemotherapy that requires a bone marrow transplant followed by radiation therapy. And for non-CNS disease, it works pretty well for CNS disease, we still struggle in that area, but that’s a whole area of neurooncology. But the COG trial is the best basis for information there. Another question is, is there any role of LDH in the AH, especially in countries where DNA biopsy is not available. So, you know, there has been some really nice work, in fact Lynn Murfree did work way back in eyes that had been enucleated and the LDH levels are definitely higher. It’s so nonspecific that in our hands that’s not been something that we have done. But I guess it is possible in lower resource settings, if you don’t have the ability to sequence the DNA, that’s something you could probably look at. That being said, it didn’t have quite the sensitivity and specificity of the DNA. So I think there would have to be larger studies in doing it now, not in enucleated eyes, to answer that question. So, good idea. It’s not a question I’m fully able to answer with our research. What is the likelihood of a child developing retinoblastoma in the second eye? So an asynchronous retinoblastoma, you detect it in one child, you might find a child has an RB1 mutation while you’re monitoring, that’s reported in about 4% of case. Another question is, is there a role for 23 or 25 age vitrectomy. Most of that work, tylectomy, was done in China. And at this time it’s not something that’s being done routinely in the United States. So I can’t comment on that. We do however have vitrectomy later, and we use the aqueous to make sure the eye is not active before our retinal surgeons go in. But there have been a case series based in China looking at vitrectomy and tylectomy. There’s been a lot of global work done, and this is why I say enucleation is a life-saving surgery. It absolutely is. For a child with unilateral disease to have one surgery and then be able to go home and be with their families and be cured, greater than 95% of the time, absolutely. Now, of course, not all children have unilateral retinoblastoma, 40% of them have bilateral, and you may be in a situation where you’re trying to save any eye and any functional vision. I’ll have the team make sure that they send out my contact, although given the number of emails I get, if you Google me, you probably will find my email pretty easily. So another question said, I sent one sample for CFDNA analysis and it didn’t show anything, does it mean it’s not diagnostic? No. When it says there are no alterations present and there’s no CTDNA identified, that means that there’s no tumor activity in in an eye. So it depends on the time you sent it. So if you sent it at diagnosis and you knew there was an active tumor in the eye, I would say that would be a false negative test. Now, at our institution, we’ve had almost none of those. And so if you did it at the end of therapy in an eye that had been treated and you were looking to see if there was still tumor DNA, but the eye looks stable, that’s exactly the answer you would want to receive. I hope I’m being clear there. So tumor DNA is present in the aqueous. When active tumor is present in the eye. If you took an eye with retinoblastoma 20 years later and you took aqueous DNA, it would be a normal profile because there’s no active tumor to shed DNA into the aqueous humor. So at the end of therapy, that can be a very effective time to look at the profile, because you’re at the end of therapy, the eye looks stable, but like the case we showed, you want to make sure there’s not current disease lurking, and a negative profile can be very helpful there. It will say, no CTDNA or circulating tumor DNA identified in the sample, at least at our institution. So I can share a video of aqueous humor biopsy but it’s basically the beginning of the video that I showed, it’s just a very simple paracentesis. If you search my name and our laboratory, we actually have a video online that you’ll be able to see. In your experience have been cases where post enucleation, the histological diagnosis turns out to not be retinoblastoma? Thankfully in my own hands, the answer here is no, although I was trained by Dr. Lynn Murfree, one of the God fathers of retinoblastoma, he shared with me several times that had happened. We once in a sort of atypical scenario told the family that the eye, this eye should come out because it had been very damaged, it was a phthisical eye, we wanted to make sure it was not retinoblastoma. Anyone who treats enough retinoblastoma will see a phthisical eye with calcification on the ideas and they’ll take the eye out to make sure it’s not RB. Is it a contraindication to performing a biopsy? Yes. If you look up our — a great paper is the Jove paper on paracentesis procedure, all available on PubMed, has lots of great pictures. It will discuss what needs to be done including making sure that the pressure is not very high, that there’s no ciliary body invitation and no anterior invitation before you put a needle in anteriorly. Brachytherapy is also used, that’s correct. For localized seeds, a group C tumor, it can be very effective. There are some questions about can we detect chromosomal alterations or predict the possibility of occurrence during the eye during checkups. Absolutely. So when I showed that example, monitoring tumor fraction and monitoring those chromosomal alterations, that’s exactly what that’s for. So, you know, it’s hard for us often to know when to start injecting, when to stop injecting, when the eye is completely clear of disease. And so, you know, what we see visually helps us but is not the only thing. And sometimes we’re wrong, right? Sometimes we think these seeds are clear and then two months later they come back, we were wrong. Or sometimes we think, oh, man, next time I see this patient I’m going to have to start injecting this eye and they come back after another injecting of chemotherapy and they look great. It’s a companion diagnostic or additional molecular test in addition to what we see. Let’s see. Last couple of questions here are, can we only do fundoscopy of siblings? If you’re screening siblings for retinoblastoma, unless they have a known RB1 alteration, I think just looking in the eye is reasonable. If you know they have an RB1 alteration and if you’re in the OR, if you have the OCT available, you should use it, and we’ve detected small tumors that way. If you don’t, I think that’s okay, and you look in the eye to see if anything is growing, that’s very reasonable. Is intrauterine chemotherapy possible for retinoblastoma? That’s a great question. I have never heard of that. In some places they recommend early delivery of the child. But to my knowledge, no one is doing intrauterine chemotherapy. And we have not ever had a need to do intracameral chemotherapy, that’s another question. Totally different parts of the body, uterus, intracameral, AC. Of course that’s well-described, the network has been led by Frances Meunier. And people around the world now are doing intracameral, again, for us it’s just not ever been something that we’ve needed to do if we have disease that’s that advanced, usually we’re talking about enucleation. Now, if it’s the last remaining eye, that would be an indication to do that. But, you know, it’s a hard situation when you’ve got anterior chamber seeds in the last remaining eye. I have not had the problem of doing that. But I love watching the videos. And then I think I’ve answered all the questions, more or less. So the last one that I see here is, are you going to do an aqueous tap for every child with Coats’ or only those eyes with total RD? I think that depends on how clear you feel clinically. Because we have this clinical test, you know, I didn’t present this data, but this will be published soon, we’ve done — we and our collaborating partners have done over a thousand paracentesis procedures. And of those, there was one patient who developed a small self-limiting hemorrhage. There have been no significant complications from the procedure in the eye. So the risk is one in a thousand that you’re going to have anything at all. And on so far, zero that that thing will be serious. So because it’s relatively safe, because these children are already under EUA, it makes it an easy enough thing to do. Our retinal surgeons feel really good taking over the care of the patient. And the parents really love getting a piece of paper that says tumor DNA not found, right? We didn’t find evidence of retinoblastoma. Conversely, our patients with retinoblastoma, we have patients who were diagnosed first as Coats’ or diagnosed as something else, told all sort of things, told they have congenital glaucoma. So when they come to us and we can actually give them a piece of paper that says DNA found in the aqueous presented with retinoblastoma, that builds a lot of trust with the team as we proceed to treat their child for tumor in one or both eyes. So these were wonderful questions, at 6:45 to 7:10 here. I appreciate all of it. As I said, you can find my online relatively easily. I’m also on Instagram. I think the Cybersight team and all of their publications as it related to this have links to my bio. If there are questions that I didn’t answer or that you still have, please shoot me an email and I’ll do my best to get back to you over the next couple of days. Thank you so much for inviting me to speak today. [Music]