At a time when oncologists are coming under pressure to prescribe checkpoint blockers there should be more – not less – scrutiny over the benefits of these revolutionary drugs. It was therefore surprising that a highly abbreviated version of data from Opdivo’s Checkmate-214 study in renal cancer was allowed to be presented at an Esmo press conference on Sunday morning.
Professor Bernard Escudier first revealed the data at this briefing, but the extent to which negative analyses were left out became evident when the meeting heard full results at an evening presidential session. After EP Vantage filed a complaint with Esmo the organisation’s president-elect, Dr Solange Peters, accepted that this might have been a concern.
However, she said Esmo did not scrutinise speakers and had no power over what they said: “The speakers are solely responsible for the information they present.”
When contacted by EP Vantage Professor Escudier, of the Institut Gustave Roussy, said he was surprised by the suggestion that relevant data were withheld. He said: “I would never withdraw important information. I selected a few slides and, as requested by Esmo, did not present the HR and p values of OS, which by the way are very positive.”
The omitted negative subgroups, however, concerned good-prognosis patients – which Professor Escudier says were in the late-breaking abstract – and PD-L1 status, which he describes as “really exploratory, totally confusing for all experts, and thus difficult to discuss”.
The timeline of events is crucial. Journalists covering the story for publication on Sunday would only have had Dr Escudier’s 8:15am press conference comments to go on; the presidential session presentation did not start until 6pm, after most major news outlets would have filed and published.
Indeed, coverage by respected outlets including Reuters, the American Journal of Managed Care and Endpoints echoed Professor Escudier’s view that Checkmate-214 was practice-changing, but did not mention negative subgroup analyses.
True, there were special circumstances around the timing of the data: Checkmate-214 had only been halted for efficacy on September 7, just as Esmo was about to get under way. Naturally all the data and analyses were very fresh, and at the press conference Dr Escudier said some analyses were being held back for the presidential session, and others for even later.
EP Vantage’s complaint to Esmo contended that attendees of the press conference were only shown positive survival curves from Checkmate-214, without hazard ratios. Only positive intent-to-treat and intermediate/poor-risk data were presented.
No subgroup data were shown to the press conference. When detailed at the subsequent presidential session these revealed that Opdivo plus Yervoy was actually detrimental in favourable-risk patients, and no more effective in PD-L1-negatives, than Sutent (Esmo 2017 – Renal cancer battlefront moves to the front line, September 11, 2017).
Not only is it staggering that Sutent control beat active drug with statistical significance, it suggests that Bristol’s combo might only be relevant in the small subgroup of poor-prognosis PD-L1-positives – representing only 20% of all-comers – with no advantage in the majority of patients.
At the press conference Professor Escudier stated that Checkmate-214 supported use of Opdivo plus Yervoy “as a new first-line standard-of-care option for patients with advanced renal cell carcinoma” – not in a specific subgroup of such patients. “Nivo/ipi will be the new standard of care in poor/intermediate patients, which represents 80% of the patients,” he told EP Vantage later.
Esmo discussants at the press conference did not critique his data and were highly supportive of the view that these were landmark results that would affect all first-line renal cell carcinoma patients equally. However, they will only have had available what was shown at the press briefing.
Element of novelty
When on Tuesday EP Vantage asked Esmo’s Dr Peters about the data omissions and partial presentation of Checkmate-214 at the Sunday morning press conference, she said Dr Escudier wanted to keep back data to preserve an element of novelty for the large audience of medical professionals attending the presidential session.
An added coincidence was that Exelixis was holding an analyst event half an hour after Sunday’s presidential session presentation started. Exelixis had updated results from Cabosun, its first-line renal cancer study of Cabometyx, which showed an equally strong PFS hazard ratio, 0.48, in the overall population as seen in the best Checkmate-214 subgroup.
However, when questioned by EP Vantage at the press conference Dr Escudier was dismissive of Cabosun, calling Checkmate-214 a “large randomised phase III compared with a small randomised phase II.” Later he told EP Vantage that it was not his role to present Cabosun data.
Bristol’s own press release detailing Checkmate-214 was highly favourable, omitting negative subgroup analyses. The group told EP Vantage that as this was the first presentation of Checkmate-214 data its release focused on the primary and secondary endpoints. It will clearly have taken legal advice to stay within investor disclosure requirements.
The issue has repercussions beyond mere investor relations, however, since renal cancer patients reading Sunday’s coverage would have got an incomplete picture of the potential of Bristol’s combination.
Dr Toni Choueiri, the lead investigator of Cabosun, who also participated in Checkmate-214, told EP Vantage: “In this day and age patients have access to Esmo, have access to Twitter, and they want ... nivo/ipi. There is a disconnect between a result like that and what you do [as a doctor] with a patient that calls you today.”
Ironically, the events occurred precisely at the time Esmo is trying to counter excessively positive coverage. In a statement on Tuesday the organisation highlighted what it called “an important problem”: that the lay press is twice as likely to report positive results than negative ones.
Dr Choueiri said that for good-prognosis patients Checkmate-214 was not the result he would have wanted to see.