Gerhardt Attard: Silke, Aurelius, thank you for asking me to speak about this topic. So a question I’ve been interested in since I realized that in the mHSPC trials, we were using a dose of prednisone 5mg daily in combination with abiraterone.

Those are my disclosures, all very relevant to this talk.

So the aim of the next ten minutes is to reach a consensus in tomorrow’s question, or if we don’t achieve that, at least in future have the information we require to make the right choice on the dose and regimen of prednisone to combine with abiraterone and indications for combining dexamethasone with abiraterone. We can split that question into two clinical dilemmas, as you probably know the abiraterone acetate label for combination prednisone is different in mHSPC, based on those aforementioned trials that the label is for 5mg daily. Of course, in contrast to the mCRPC setting where we use 10mg daily, usually 5mg BID, and there are, the main potential concern with that is the half-life of prednisone is less than 16 hours. So of course, we may not achieve control of ACTH rise for the full daily duration. And there’s some indication from the LATITUDE trial, the main trial that lead to the extension of the license of abiraterone and the space, that there’s an increased incidence of grade three or four hypokalemia, 11%, compared to mCRPC studies with 5mg BID was used respectively, 2%, and chemo-naïve population, 3%, and the post docetaxel COUGAR studies.

The second dilemma is when should we use dexamethasone. Dexamethasone is particularly appealing, especially used in several UK practices. Firstly, because it has a longer half-life, so we could get away with a once-daily dose, it’s got a higher ratio of glucocorticoid to mineralocorticoid activity when compared to prednisone. There’s a randomized phase two study called PoD that Chris Parker’s team ran that suggested greater activity with single-agent dexamethasone when compared to prednisone. There’s a number of small phase two studies retrospective analysis that have reported a decrease in PSA by 50% or greater in about 25% of patients with some of those men having confirmed radiological responses. When that patient was progressing on prednisone and then the prednisone was changed to dexamethasone. And some of those patients had previously progressed on dexamethasone single-agent prior to receiving abiraterone or prednisone, abiraterone with prednisone. 

So background slide, this is pertinent to understanding the rest of the talk. Abiraterone inhibits CYP17A1 which has two mechanisms of action. 17 alpha-hydroxylation of pregnenolone to 17 alpha-hydroxy pregnenolone and then the C17, 20-lyase conversion to androgens and estrogens. And the abiraterone inhibits both those activities. So when we use abiraterone on its own, cortisol decreases but due to a rise in ACTH that drives an increase in steroids upstream of CYP17, patients do not develop hydrenocortical insufficiency. So corticosterone is a weak glucocorticoid and that restores or maintains the gluco requirements of men. However, DOC and other metabolites of deoxycorticosterone are potent mineralocorticoids. So that causes hypokalemia, hypertension, fluid overload, et cetera. And that’s why abiraterone is given in combination with a steroid.

Now with the support of Janssen, we conducted a multi-center, international Phase 2 randomized clinical trial. Several of my co-investigators are in the room, and we were fortunate to include, or be allowed to include, or have the support to include multiple correlative studies. We hypothesized that a lower dose of steroids would be beneficial in terms of long term physiological effects. So we designed the study based around the primary endpoint which was the proportion of patients with no hypertension or hypokalemia as really the robust manifestations of mineralocorticoid excess, we could integrate within the first six cycles. You had discussions about measuring lower limb edema, for example, having patients put their legs in buckets of water and we ultimately decided on these two relatively robust endpoints. 

164 men were randomly assigned to one of four regimens. Either prednisone 5mg BID, prednisone 5mg once daily, prednisone 2.5mg BID, or dexamethasone 0.5mg daily. See the number of patients assigned to treatment, the number who received treatment, and then the number of patients who discontinued within those six months, not due to hypertension or hypokalemia. Of course, if they had either of those two last side effects, they would have counted towards our primary endpoint. Giving us the numbers on the lower row who are valuable for the primary endpoint. We hypothesized that 70% of men and the standard dose of prednisone 5mg BID, or least is was the standard dose when we designed the trial, would have no hypertension or hypokalemia and we were pretty spot on and we designed the study to say either of the regimens did not effectively control mineralocorticoid excess. If the 95% confidence intervals were below the lower limit of those 95% CIs was below 50%. And that was the case for the two lower doses of prednisone, either the 5mg once daily or the 2.5mg BID. And you’ll see the hypokalemia proportions there, I guess of note, two out of the 38 patients randomized to prednisone 5mg once daily had grade three hypokalemia. The difference in the number between hypokalemia and the proportion who did not meet the primary endpoint are those patients who had hypertension. 

Now see the numbers are very small, now how robust can this be with numbers as low as five? But as I said we integrated multiple correlative studies and we collected urine in this, in the trial, in all patients at multiple time points. Martin Gleave, earlier today mentioned analysis of androgens in blood which of course is challenging because your working with small amounts of material. With urine we collect liters of urine, we’re able to use a mass spec-based assay and measure multiple metabolites within that steroid based synthesis part, and I’ll talk through two of them and I’ll also first start with ACTH.

We used violin plots to detect the range of change. Primarily because there is a wide variability but clearly as you see for ACTH there’s a proportion of patients assigned to the lower doses of prednisone who have higher levels of ACTH on treatment. The median level for, of ACTH in those two lower doses was significantly higher when compared to baseline. Not the case for prednisone 5mg BID. And then the case of dexamethasone the level of ACTH is lower. For DOC the most, the metabolites of DOC, the most potent mineralocorticoid that rises, again we see a significant rise in DOC metabolites in the lower prednisone doses, not the case for dexamethasone and not as significantly the case for prednisone 5mg BID. And interestingly for androgen precursor metabolites we see significant decreases with prednisone 5mg BID and dexamethasone 0.5mg once daily, which we don’t with the lower doses suggesting that there is better androgen suppression when we more effectively controlled ACTH. And I’ll show you some data to support that in a couple of slides.

Now, of course, the long term physiological effects, there’s a significant rise in serum insulin and HOMA-IR with dexamethasone, not with prednisone doses, and a significant increase in body fat with prednisone 5mg BID and with dexamethasone 0.5mg daily. There is a significant decrease in lean body mass with all the steroid regimens but most significantly or the p-values is lowest with the 5mg BID sections. So again you can see a wide variability not all patients have toxicity or changes that we could detect and overall the changes were relatively minor. But there is a suggestion, as you would expect, that the higher doses of steroids have greater, long term physiological effects.

We also measured anti-tumor activity. We did not design the study to compare activity between different regimens, so treat this with caution. The blue waterfall plot refers to 5 BID, prednisone 5mg BID, the red waterfall plot to 5mg once daily, green 2.5mg BID, and the brown to 0.5mg once daily and the cam plots on the far left, again same colors used, and notably the time, the progression-free survival time was notably longer on dexamethasone although as I said, treat that with caution, because we’re not powered to compare different regimens. 

So to conclude, I don’t have the answer to which dose is best to use and I think this is a very appropriate question to this meeting. We’ll see what the rest of the panel vote on tomorrow. The physiological long term effects of abiraterone plus steroids had no discernible impact on quality of life. I didn’t show that data but in the manuscript that was published a few months ago, 5mg BID and dexamethasone 0.5mg once daily minimized mineralocorticoid excess. However that comes at a cost of increased body fat and for the case of dexamethasone, increase in HOMA-IR, increase serum insulin, and reduced bone mineral density. 5mg once daily minimizes those long term physiological side effects but very clearly results in a greater chance of mineralocorticoid excess in a proportion of patients. That can result in hypokalemia and other mineralocorticoid side effects and may also impact the efficacy of treatment and I wonder whether we should dose escalate the steroid when patients progress, especially when that’s been started in the mHSPC setting. And implied by that is that a lower steroid dose should be watched more carefully, we should ensure patients are medically optimized before we start patients of 5mg once daily.

And finally, dexamethasone, I don’t think we should start all patients on the combination of abiraterone with dexamethasone. One, you’ve seen dexamethasone is associated with more, long term physiological side effects, and of course, there’s an absence of randomized data for abiraterone plus dexamethasone. But as we do in my practice and I know several of my colleagues in the UK, I think dexamethasone can be considered for patients progressing on abiraterone and prednisone if there are no other concerns. Of course, we have no data that randomizes patients to dexamethasone alone versus abiraterone and dexamethasone. So those responses I described there may have been entirely related to the dexamethasone and the continuation of abiraterone was unnecessary, but I suspect we will not obtain level one evidence on that question either. Thank you.