I’m going to talk about this phase I study of nivolumab, which is one of the anti-PD1 antibodies, in patients with relapsed refractory classical Hodgkin lymphoma. The background of this, briefly, is that classical Hodgkin lymphoma is a little bit unusual among lymphomas in that under the microscope the malignant cells, the tumour cells, are rare, the so-called  Reed–Sternberg cells, and they’re surrounded by an extensive inflammatory infiltrate that doesn’t seem able to eradicate the cancer.
So people have wondered for a long time how it was that Hodgkin lymphoma could attract such a brisk immune response and yet that this immune response failed to kill the tumour. Genetic analyses that were performed a few years ago showed that classical Hodgkin lymphoma frequently has amplification of genetic material at this area on the ninth chromosome called 9p24 and that this amplification mostly results in upregulation, so increased expression of the PD1 ligands PDL1 and PDL2.
PD1 sits on the surface of immune cells, most notably T-cells, and when it’s engaged by its ligands, PDL1 and PDL2, this results in a downregulation of the immune response so it temporarily shuts down the immune response and all of this together led to the hypothesis that classical Hodgkin lymphoma may have a genetically determined dependence on the PD1 pathway for survival. This is unusual in that other tumours that have been targeted with PD1 blockade don’t have the same known genetic abnormality to support their dependence on PD1.
So based on this, classical Hodgkin lymphoma was included as an independent expansion cohort in a phase Ib study of nivolumab in patients with relapsed refractory hematologic malignancies. The study schema is shown on this slide; there are 105 patients who were enrolled on this study and this proceeded through a dose escalation phase and a dose expansion phase. On the dose expansion phase at a dose of 3mg/kg of nivolumab 23 patients with classical Hodgkin lymphoma were treated. The other patients are presented in a separate presentation at this meeting.
The primary endpoint of this study was the safety and tolerability of nivolumab and the secondary endpoint included the overall response, the duration of response of progression free survival and then biomarker assessments.
So this slide shows some of the characteristics of the 23 patients who participated in this study. As you can see they’re relatively young, the median age was 35 and, of note, they were quite heavily pre-treated. So 78% of the patients had received previously an autologous stem cell transplant and relapsed following transplantation and 78% had received prior brentuximab which we know is one of the most active drugs in Hodgkin lymphoma. The average number of prior lines of therapy was four and, as you can see at the bottom of this slide, about a third of patients had received six or more lines of prior therapy. So this is a population of patients that had received extensive pre-treatment for their Hodgkin lymphoma with for many of them very few alternative options available.
This slide summarises the safety data on this study. Without going through all the details you can see that five patients, or 22% of the total, had a grade 3, so a severe, related adverse event but there were no life-threatening drug-related adverse events and no treatment related deaths on the study. There are two patients that stopped treatment because of toxicity, one patient who developed myelodysplastic syndrome who actually had been very heavily pre-treated, had received a lot of prior chemotherapy, so the MDS was possibly related to the prior therapy but because it was diagnosed while the patient was on study it was attributed as study related. Then one patient stopped for pancreatitis. The other drug-related grade 3 adverse events are shown here. Overall nivolumab has been used in thousands of patients already on clinical trials in solid tumours and overall this safety profile mirrors that from what we expected in solid tumour. But the interesting thing about that from our standpoint is that there was no apparent increase in the incidence of lung toxicity which is something we were worried about for those patients because many of them had had radiation or other drugs that can cause lung injury like brentuximab, carmustine etc. And on this study there was no apparent increase in lung toxicity.
The best response achieved by patients is shown here and, as you can see, among the 23 20 achieved a response of partial remission or complete remission which is 87% of the total. That included 17% of patients who achieved a complete remission and 70% who achieved a partial remission. The other 13% achieved stable disease as their best response. For this cohort the 24 week progression free survival was 86%.
Also shown on this slide are results for three subgroups of interest: the patients that had had a prior auto-transplant and had had prior brentuximab, so these were the most heavily pre-treated patients and had a response rate of 87%; patients who were not able to get to transplantation, 3 patients only but all of them responded to drugs; then finally the patients that had never had brentuximab with a response rate of 80% and among the 4 patients 3 of them had a complete remission.
This slide, which is a little busy, shows the response duration for the 20 patients who achieved a partial or complete remission. The salient points that I would make is that most of the responses, 60% of the responses, occurred within 8 weeks of starting study treatment and at the time of database lock, which was in June of this year, 48% of the responses are ongoing. 43% of patients, so 10 out of the 23, are still on treatment at the time of database lock and there are patients now that have been in remission for over a year.
This slide summarises the results of the correlative studies, so we were able to obtain ten tumours from patients who enrolled on this study and study them with a number of techniques to try to get at the biology that underlies the clinical responses. As you can see, all ten tumours had their genetic amplification at 9p24 which is what we thought formed the basis for vulnerability to this type of treatment. All tumours in the malignant cells in the Hodgkin Reed-Sternberg cell, or HRS, had increased expression on the cell surface of the PD1 ligands PDL1, PDL2 and phosphor-STAT3 which is a pathway that leads to further PDL1 expression.
So what we conclude at this point from the study is that nivolumab can be safely administered to patients with relapsed refractory classical Hodgkin lymphoma, that this results in these preliminary results in a response rate of 87% which is quite high for a group of patients who are very heavily pre-treated. All the tumours that we were able to study harboured genetic abnormalities in 9p24 leading to over-expression of the PD1 ligands. This supports that hypothesis that classical Hodgkin lymphoma is a tumour with a genetically determined vulnerability to PD1 blockade. We hope that PD1 blockade in the future can become an important part of the treatment for patients with classical Hodgkin lymphoma.