This is another transcript of my notes and thoughts from the New Horizons conference that took place February 1st, 2013 in San Francisco. Along with Part 1, which covered the keynote, this will then conclude the entire morning session from that day.
Leading off after the keynote was Dr. Ike Ahmed’s forum on “New Horizons in Glaucoma Devices”. This concentrated on minimally invasive surgical devices, so I’m opting for now to skip commentary as I’ve little noteworthy to say. But, I will offer a single sound-bite from someone else, Dr. Michael Belkin, who, in comparing the devices on offer and the decision to use one over the other, said: “The best marketer will win.”
I’m also dodging comment on the industry panel discussion on Glaucoma devices, too. The panelists were industry leaders – CEOs and VPs – and as such were forgivingly guarded in their views. My one thought here was how encouraging to hear from each of them that they have to look outside of their own confines to innovate; the main example being the need to work with academia and monitor start-ups. Dr. Ralf Kuschnereit, the CEO of Carl Zeiss Meditec Inc., made the point that clinical partners are the best innovators. As with all topics discussed, each member of the panel agreed to a large degree with the previous member’s answer to the question posed, and so it went back and forth. Greg Kunst, of Alcon Laboratories Inc., closed with some insightful remarks on Novartis’s strategy toward emerging markets, however.
And onto the subject of this post: “Why Glaucoma Biomarkers?”, from Dr. Jeffrey Goldberg, a Professor of Ophthalmology at the Shiley Eye Center, UCSD. The topic of biomarkers is obviously of great interest to us at Voxeleron as we’re in the quantitative imaging space. Indeed, Dr. Goldberg’s very interesting presentation was likely of tremendous interest to all in attendance. He framed his talk with the following question: can we identify new molecular and metabolic biomarkers for Glaucoma diagnosis and progression? The status quo here is dominated by intra ocular pressure (IOP) as an end point, but, as Dr. Goldberg so rightly pointed out, the disease is characterized by visual field loss and changes in structure at the optic nerve and not IOP (although there is plenty of evidence to show that reducing the IOP does provide effective protection against the degradation of the retinal ganglion cells). Furthermore, IOP as a risk factor is not sufficient for diagnosis or monitoring progression. Consequently, Dr. Goldberg is advocating that, clinically, we should be concentrating on what is damaged due to Glaucoma. Namely the optic nerve and more specifically the dysfunction or death of the ganglion cells. The problem then posed was that we do not know how to measure these things well and cannot distinguish between the ganglion cell death and dysfunction.
The complaints brought up against the functional measures of the visual field are clear and understood: its noise and subjectivity. Good examples were shown to illustrate this. Pattern electroretinogram (PERG) was then suggested as more objective and may precede retinal nerve fiber loss, so it is a strong candidate for a biomarker. The issue here being that, as the disease progresses, the measurements suffer dramatic signal loss, rendering the technique only useful for very early stage disease monitoring. Microperimetry was not mentioned.
Structural measures were then touched upon, very lightly, with stereo fundus photography being cited as the primary approach. Additionally, Dr. Goldberg said, there was an entire “alphabet soup” of alternative modalities, including: optical coherence tomography (OCT – in the form of RNFL thickness around the nerve and in the macula), GDx (laser polarimetry, looking at the micro-tubular organization), HRT (confocal scanning laser – reconstructs topography of the nerve). I guess here I’d comment that among this list of devices, clearly OCT is the standard of care, and Dr. Goldberg offered no preference for one modality over another. The GDx is a bit of a one-trick pony, and confocal microscopy is of limited use these days when true volumetric imaging is available; the consequence being that both devices are being rapidly squeezed out of the market by OCT. Measurements of the Ganglion Cell layer using OCT were not mentioned.
Ignoring for now the measures that were not discussed, the measures that Dr. Goldberg did cover, he said, were not enough for the development of nerve protecting drugs. Also, he confirmed that the FDA is not allowing these as measures of Glaucoma diagnosis or progression because the correlation to the loss of visual function is not established well enough. Establishing such a link is a very active field of research, and one that is advancing rapidly. In that respect, I was surprised that microperimetry was not mentioned. Quite recently, the FDA approved a combination Microperimetry / OCT devise, which will allow direct comparison of visual function to structural measurements in the macular. Researchers, such as Dr. Vas Sadda at the Doheny Eye Center, are already using this instrument to investigate the relationship between structure and function. They reported on initial findings at the recent ARVO meeting, but concluded that additional layer segmentation functionality is potentially key to these studies: “Correlation of sensitivity with morphometrics of individual retinal sublayers could potentially strengthen these associations and be a topic of future study.”
So if there’s nothing out there then, what should we look to for biomarkers in the future? Dr. Goldberg’s answer is the following: something that relies on reduced clinical expertise; something that allows for measures of progression at shorter intervals, or immediately even – i.e., more sensitive, better repeatability; and something that indicates the risk of progression – your prognosis a measure of the state of your ganglion cells and optic nerve head (ONH) are doing at this time. He wants to be able to:
- Assess Progression
- Assess Risk of progression / Impending progression (current measures are too late – after loss)
- Offer guided therapy – patient specific measures.
Assessment of risk of progression gets back to his desire to see dysfunction before cell death. Would treatment then be effective? Also hypothesized was that, while Glaucoma is typically a slowly progressing disease, it takes a very long time to perform the clinical trials required of any candidate neuro-protective drug. If dysfunction, therefore, were an endpoint, trials, he argued, could be shorter.
So what are the candidates? Here, Dr. Goldberg worked his way through his own collaborations. The first was work led by Dr. Alf Dubra in the department of ophthalmology at the Medical College of Wisconsin, Dr. Andrew Huberman, of UCSD, and Dr. Vivek Srinivasan. I’ll itemize these and the comments from Dr. Goldberg, and then add my own closing commentary at the end.
- We can measure electrical activity in the retinal ganglion cells (RGCs) and have shown that, following an optic nerve injury, activity goes down before the RGCs die. Measure, therefore, their actual electrical activity. This, he thinks, can help answer his question of “how happy are the RGCs?
- There are many different RGC subtypes. This collaboration has been able to identify specific subtypes using genetic markers. They are looking for a particular subtype that might flag impending damage. They also have different morphological shapes in the retina (seen using adaptive optics), and that some of them degenerate earlier than others. If you can identify and follow these then there is the possibility that these could be end points.
- Can we resolve structures in the living retina at the cellular level? This, says Dr. Goldberg, has been a major problem for all currently available imaging modalities. Using adaptive optics rod and cone imaging is possible, and now they are looking in detail at inner retinal vessels that are responsible for nourishing the RGCs. He also showed images of axons and also talked about how you could see degenerative changes that are not normally appreciated and could now potentially measure.
- Metabolism: here he cited two interesting approaches. The first related to the use of mitochondrial dynamics within RGC axons, which show change due to Glaucomatous insult. Alternatively, they are also measure blood flow using Doppler OCT to gauge metabolic activity.
- Lastly, mention was given to using different nano-particles to develop selective [fluorescent?] signaling to hone in on injured axons in the optic nerve. Here, I assume, such that the extent of injury can be measured.
This was a well-received talk that covered a lot of ground in a short period of time. My initial impression was similar to that of the keynote talk in that it rather ignored the current state of the art and chose instead to over-reach and talk about topics many years out. Anything measurable shown – e.g., images or signals – were perhaps 5 years out in terms of getting put in a device and then also a ways away from the automated measurements using image and signal analysis. We’ve all seen first-hand how device manufacturers have struggled getting automated measurements good enough to be able to say that measured structural change is real and not an artifact of the measurement algorithm. Knowing well how the vision algorithms work, someone will have their hands full accurately measuring, for example, morphometry of different RGC subtypes and then change in only one of those subtypes (based at least on the images I saw).
So what was missing from the survey of existing modalities from Dr. Goldberg’s talk? This question allows me to add now the alternative opinion in that layer thickness measures of the ganglion cells in the macula and their axons at the optic nerve – the retinal nerve fibers – using OCT are good first order measures of atrophy that have been reliably studied (here and here, for example.). Furthermore, Harwerth et al. established a quantitative relationship between visual sensitivity and ganglion cell density, publishing their findings back in 2004. But it is true, atrophy may not indicate functional loss (there will be some compensation), and can also be hindered by other bodies within those layers, such as glial cells, that may remain unaffected by the cellular apoptosis. It is, however, generally assumed that structural loss precedes visual loss, and Quigley’s seminal work showing this holds true. So would a structural measurement be the prognosis required? That all said, if you were new to the imaging field and hearing this lecture for the first time, you would be led to assume that there is no structural measurement of the ganglion cell layer, which is certainly not the case. And combining ganglion cell layer thicknesses with microperimetry could really further our understanding in this area.
As was the case for the keynote, we have a Glaucoma clinical researcher, a leader in his field, seemingly not so interested in looking at the inner retinal layers of the macula. Additionally, the fact that we can do this is overlooked, and our direction turns to either very expensive devices (e.g., adaptive optics), or devices that are many years away from their translation to clinical practice. We discard, therefore, the vast sources of data that is available and the associated low costs. Might we hypothesize, for example, that the dendrites of the RGCs swell well ahead of apoptosis? Indeed, morphological change is evident (see Liu et al., for example). And how might the nuclear layers be affected? And when?
The idea of using Doppler OCT for metabolism is a good one as this data can be taken from any OCT device. Dr. Huang’s lab at the Casey Eye Institute in Oregon have been advocates for this for some time now and it does make tremendous sense. I suspect the difficulty in normalizing the measures is an issue along with an inter-individual variation in overall metabolism makes this harder. Also, the measures might vary wildly by external factors such as body temperature, as the blood to some extent acts as a heatsink to the eye.
Dr. Goldberg’s lecture was very interesting. I applaud the visionary aspects of the topics, citing very forward thinking research. But I again advocate for pragmatism and the need to understand what is currently to hand, while at the same time steering our reach for the future. So to close, I’ll just comment briefly to the five categories of potential end points:
- [Electrical activity in the RCGs] The data looked very noisy, and change has only been shown to occur after optic nerve injury, so perhaps a little too late?
- [Retinal morphology] Image analysis algorithms will need to be developed to support this, which is fine by us!
- [Cellular imaging] Even using adaptive optics, the RGCs are very, very hard to see. Some good background reading can be found in the paper from Werkmeister et al., who among other things describes the difficulty in the imaging process and the need for contrast agents to image at the cellular level.
- [Metabolism] I can only comment on Doppler OCT, which is given above.
- [Molecular imaging] This was a closing remark, and I didn’t really understand it. It’s likely that this is many years out, but the subject area is in the hands of many brilliant people, so perhaps closer than I might appreciate. It was interesting to learn that application to the retina is being done.
In summary, this was an excellent and enjoyable talk. It seemed however to skip the opportunities that are current, and look instead to what is really on the horizon. Perhaps this is what people were more interested in hearing.