ASCO 2018 was another whirlwind of science, meetings, connections, and exhaustion. It is never possible to attend everything you want, and it is fruitless even to try: Star Trek transporter technology is not yet available, and even if it were, the quality of the McCormick Place wifi suggests it might be most unwise to use it. Some of the biggest messages in the genitourinary streams were also reflected in some of the other streams: more is sometimes not better; the science you thought you understood is usually only part of the truth; and immunotherapy, while it can be amazing, can also be disappointing and still has a way to go.
The science you thought you understood is usually only part of the truth
Abstract 5008: Sternberg CN et al. A randomized phase 2 study investigating 3 dosing regimens of radium-223 dichloride (Ra-223) in bone metastatic castration-resistant prostate cancer (mCRPC).
Radium-223 dichloride is approved in Australia but not yet reimbursed, although hopefully this situation will change in the near future. One challenge will be to work out where in the treatment sequence this agent might best fit. The current evidence suggests it should be used early in the disease course, when patients are fitter, marrow reserve is better, and the probability of soft tissue disease is lowest.
The conventional dosing regimen for Ra-223 is 55 kBq/kg every four weeks for up to six doses (the nuclear medicine alchemists have decreed that 55 kBq/kg is equivalent to what the 50 kBq/kg dose was in the original ALSYMPCA study; ordinary mortals just nod and smile, and back away slowly).
Phase 2 studies suggested higher doses might be of benefit, and the original regimen of six doses was arbitrary, so this trial was designed to compare the standard regimen with a high dose regimen (88 kBq/kg q4w for up to 6 doses) or an extended regimen (55 kBq/kg q4w for up to 12 doses).
It is important to note that the standard treatment course of six doses runs over 21 weeks, so selection of patients is critical: there needs to be some level of confidence that treatment of extraosseous disease will not be necessary over that time frame. A planned twelve doses may have been somewhat problematic given that Ra-223 works only in bone and so these patients with CRPC had no other effective therapy acting to suppress growth of cancer outside the bones.
The primary endpoint of the trial was symptomatic skeletal event-free survival (SSE-FS), comparing high with standard dose (powered to detect a 50% improvement in SSE-FS), and extended dosing against standard duration at the conventional dose (powered to detect 65% improvement in SSE-FS). Randomisation was stratified against pain (≤4 vs >4 on BPI), number of prior chemotherapy regimens, and ALP. At least two prior systemic therapies had been given to 84-90% of patients, mainly docetaxel and abiraterone. The median number of cycles received in each group was 6; and 65% completed treatment in the standard arm, 52% in the high dose arm, and 22% in the extended dosing arm. Failure of therapy was mainly due to progressive disease. No differences in median SSE-FS were detected, which across the board was of the order of 11-13 months. Treatment-emergent adverse events were similar in nature although events of grade 3 or higher occurred more frequently in the high dose or extended arms. Median overall survival was 15.8mo (standard) vs 16.0mo (high dose) vs 14.4mo (extended dosing).
It is striking and admirable that this “negative” trial was offered an oral presentation. The presenter and the discussant both concluded that the standard dosing regimen was an appropriate one in mCRPC, and this in itself is useful information, but several questions still remain, especially in the light of recent concerns regarding the combination of Ra-223 with abiraterone: When should this agent ideally be used in the treatment sequence? Are there appropriate agents to partner with it? With how much disease outside the bone and with what pattern of progression is it reasonable to use Ra-223 as a single agent? There is still much to be done to understand how best and when to use this agent.
Abstract 5003: Clarke N et al. Olaparib combined with abiraterone in patients (pts) with metastatic castration-resistant prostate cancer (mCRPC): A randomized phase II trial.
The floodgates have opened recently with the realisation that somatic mutations in genes affecting DNA damage repair occur frequently in prostate cancer and other malignancies. My good friend and colleague Carmel Pezaro gave a masterful presentation on this in one of the education sessions, and I encourage you to look at this on the virtual meeting.
PARP represents a key pathway that cells rely upon for DNA damage repair. Inhibition of PARP, particularly in the presence of other DNA damage repair deficiencies such as BRCA mutations, causes “synthetic lethality,” ie a much greater cell kill. A large number of genes is now known to be involved in DNA repair. More recently, it has been shown that germline mutations in these genes also occur frequently in men with prostate cancer, even without a family history. On top of that, I had not appreciated that PARP function is intertwined with androgen receptor function, and that inhibition of AR signalling can induce DNA damage repair deficiency. This together with previous clinical work formed the rationale for this trial presented by Noel Clarke and published simultaneously in Lancet Oncology (http://dx.doi.org/10.1016/S1470-2045(18)30365-6).
This randomised phase 2 trial examined men with metastatic CRPC with prior docetaxel therapy but no exposure to second-generation AR-targeted agents. A limited panel of DNA damage repair genes was assessed in tumour, plasma, and germline tissue, but mutation status was not an eligibility criterion. Participants were randomised 1:1 to receive abiraterone 1000mg daily plus olaparib 300mg bd or placebo.
The primary endpoint was radiographic PFS by RECIST 1.1 or PCWG2. Adverse events were more common in the olaparib/abiraterone group, particularly anemia and nausea, consistent with the known toxicity profile of olaparib. A benefit in rPFS for the whole population was observed, (HR 0.65, 95% confidence intervals 0.44-0.97, p = 0.034; median 13.8mo for olaparib arm vs 8.2mo for placebo arm). So far there is no evidence for a survival benefit.
The phase 2 design and low numbers limit interpretation of the results especially for DNA repair status subgroups, and the study has been criticised in an editorial for its assessment of DNA repair status in multiple assays and with a relatively limited panel. Nevertheless, the results speak for themselves and, taken with other work, certainly provide a good rationale for future phase 3 studies that might not be limited only to men with known and characterised DNA damage repair deficiencies.
Immunotherapy, while it can be amazing, can also be disappointing
Abstract 5007: De Bono JS et al. KEYNOTE-199: Pembrolizumab (pembro) for docetaxel-refractory metastatic castration-resistant prostate cancer (mCRPC).
The clear message from ASCO and other meetings over the last few years is that pembrolizumab, or one of its cousins, should be added to the drinking water and can cure everything. The reality of course is quite different: it does not always work, and it tastes terrible.
Prostate cancer in any sensible universe should be expected to respond readily to immunotherapy but, aside from positive sipuleucel-T trials and some small studies with other interventions, outcomes in the modern immunotherapy era have been disappointing so far.
The reasons for this are unclear but don’t get stuck on the “mutational burden” story that so many people seem to obsess over: the truth as always will be far more complicated and other mechanisms are almost certainly involved.
The KEYNOTE-199 trial involves five cohorts of men with mCRPC: (1) patients who have received 1-2 prior chemotherapy regimens including docetaxel, PD-L1 positive, and measurable disease by RECIST 1.1 [n=131]; (2) the same, but PD-L1 negative [n=67]; (3) prior chemotherapy, bone metastases with no measurable disease by RECIST, and any PD-L1 status [n=60]; (4) chemo-naïve and on enzalutamide, with RECIST-measurable disease; and (5) as for 4 but with bone-predominant, RECIST non-measurable disease. Results for cohorts 1-3 were presented at ASCO 2018 with most information available for cohorts 1 and 2.
All men received pembrolizumab 200mg q3wk for 35 cycles or until confirmed disease progression, toxicity, or withdrawal due to investigator or patient decision. The primary endpoint was objective response by central review for cohorts 1 and 2, separately and combined.
As is so often the case, the results raise more questions than answers. The median number of doses received was 4.5 and only 47 patients (18%) stayed on treatment for 6 months or more. Most patients discontinued treatment due to radiographic or clinical progression of disease, with <10% stopping due to adverse events including two fatal events thought to be related to treatment (pneumonitis and sepsis). The majority of patients experienced progressive disease as their best response, but a small minority had good responses, some of which were striking. Response rates for RECIST-measurable patients (cohorts 1 and 2) were 6% and 3%, mostly partial responses with a few complete responses, although stable disease was seen in 17-21% and was prolonged over 6 months in 3-4%. Responses were seen in both PD-L1 positive and negative patients. Survival data unsurprisingly showed that cohort 3 (bone-only disease) had outcomes superior to cohorts 1 and 2.
Some interesting translational studies were presented but unfortunately we are no closer to knowing how best or when to use pembrolizumab or similar agents in prostate cancer. By all rights it should work: surely we just need to understand the biology and the drugs better.