Rechallenge with PD-1/PD-L1 inhibitors after a (un)planned interruption may offer a potential safety net for patients with disease progression. At ESMO Congress 2019, results from a study systematically investigating this approach for the first time confirmed the possibility of successful rechallenge.Results from this phase I/II study presented yesterday reported that among 70 patients with metastatic cancer who were receiving durvalumab for 1 year—to treat a variety of advanced solid tumours—rechallenge on disease progression was associated with responses (all partial) in 11.4% of patients and stable disease in 60% (Abstract 1175O). Although encouraging, the study serves to highlight that there are still so many questions left to answer.Firstly, what is the optimal duration of initial treatment? Currently, many patients with response or stable disease upon anti-PD-1/PD-L1 treatment in clinical trials receive therapy for around 2 years. In clinical practice, this duration may be even longer. Being able to safely reduce this treatment duration would relieve patients from treatment, perhaps lower the risk of side effects and also reduce associated costs. In the study reported yesterday, treatment was discontinued after 1 year. While a proportion of the patients responded to rechallenge, the rates seemed lower than what can be expected for primary response to the same agent (11.6% compared with up to 50%, depending on the disease type) and we do not know whether outcomes would have been better if patients had simply continued with treatment beyond 1 year. Results from the CheckMate 153 study in non-small-cell lung cancer suggest that this might be the case, showing that progression-free survival is better with continuous nivolumab than with a fixed 1-year treatment duration.1 The safety of early treatment discontinuation in patients achieving a complete or partial response is being investigated in the ongoing Safe-Stop melanoma study in The Netherlands, but stopping safely at an earlier timepoint may be cancer specific. Case reports of patients across cancer types requiring early immunotherapy discontinuation due to immune-related toxicity have shown durable responses, suggesting that early stopping can be safe.Rechallenge is also complicated by the fact that different tumour types respond differently to treatment. For instance, around 30–40% of patients with melanoma may respond to initial anti-PD-1 treatment, whereas for lung cancer, the proportion is around 20–25%. More data on the true nature of responses to rechallenge are needed: what is their duration? Are they as deep as initial treatment responses? Is there a higher chance of tumour escape? For example, it is surprising that the study presented yesterday reported zero responses to rechallenge in microsatellite instability-high disease, which usually responds very well to this treatment.It appears to me that rechallenge with anti-PD-1/PD-L1 agents after planned treatment interruption may have a role in the management of some tumours. However, there is still so much we need to know before the application of this strategy can be confirmed.1. Spigel D, et al. Ann Oncol 2017;28(Suppl 5):Abstract 1297OESMO Congress 2019 abstract:1175O - Durvalumab activity in previously treated patients who stopped durvalumab without disease progression