Resuming Life After Cancer Therapy: Treatment-Free Survival

Treatment-free survival is a novel endpoint in immunotherapy. TRANSCRIPT This JCO Podcast provides observations and commentary on the JCO article "Treatment-Free Survival: A Novel Outcome Measure of the Effects of Immune Checkpoint Inhibition—A Pooled Analysis of Patients With Advanced Melanoma" by Regan et al. My name is Adil Daud, and I am Professor of Medicine and Dermatology and Director of the Melanoma Program at the University of California, San Francisco. My oncologic specialty is medical oncology. Cancer therapy has achieved great success in the last 40-50 years. Where treatment with chemotherapy required inpatient hospitalization and gut-wrenching toxicity, therapy today can often be achieved with lower grade side effects and limited time in the hospital or outpatient infusion center. While these changes have brought enormous benefit to patients, many patients feel that the tug of ongoing therapy for metastatic or advanced cancer and long for a time where therapy is not continuing, and the word "cure" is not completely out of mind. Allied to these concerns is the rise and expansion of immunotherapy for cancer. The growth and spread of neoplasm often triggers the immune system, which mobilizes in response. While cancers can use a variety of adaptive mechanisms to evade the immune system, blocking these evasive mechanisms can produce lasting responses and, in some cases, durable tumor-free intervals. Cytokine therapy as exemplified by IL-2 offered this benefit since its approval in the 1990's for renal cell cancer and for melanoma for approximately 10-20% of patients treated. The CTLA4 antibody, ipilimumab has had a similar long-term disease control rate with a better toxicity profile. However, it is the anti-PD-1 monoclonal antibodies, used by themselves or with ipilimumab that have made treatment-free survival a tangible and achievable goal for many patients with cancer especially those with cutaneous melanoma. Measuring the effect of treatment is traditionally done with measures of disease control. These include tumor shrinkage as measured by an agreed upon method such as RECIST or irRECIST or by lengthening of time-either lifespan or time without progressive disease (progression-free survival). However, it is possible that a similar lifespan could be achieved in one of 2 different ways-either continuous treatment with a drug or a shorter-term treatment that stops within a few months and then the patient has no further treatment. Comparing these treatments that are profoundly different for a patient but similar in terms of standard time or disease control measures demands new measures that can help describe the benefit of one or the other treatment. In the JCO article that accompanies this podcast, Regan et al, in an analysis of 1077 patients across 2 different randomized trials, attempt to provide such a measure. They define treatment-free survival as interval between time to ICI cessation to the time to subsequent therapy or death.  With this measure, the shorter the therapy and the more delayed the need for subsequent therapy (or death) the longer this interval is. Combination immunotherapy is notoriously toxic and while treatment duration can be short, the side effect duration can be prolonged. Could TFS be contaminated with toxicity ? to answer this question, Regan et al introduce another endpoint, TFS with and without toxicity and partition this with persistent and late onset grade ≥3 toxicity. These endpoints are illustrated in Figure 1. Below the familiar Kaplan-Meier overall survival curve is a slice showing survival after subsequent therapy initiation (so factoring in progression). Below this is the TFS without toxicity. Below this slice is the TFS with toxicity and below this is the time on IO therapy. With this division, the familiar KM curve gives a lot more information and illustrates the difference between nivo-ipi combination therapy and nivolumab monotherapy. While the overall survival is not statistically different between these 2 treatments (as shown in previous publications) the TFS lets us see what is going on with patients. In Figure 4 we can see that in the combination nivo-ipi arm the TFS (mean) is 11.1 months vs 4.6 months for nivo alone and 8.7 months for ipi alone. These numbers indicate that TFS alone does not convey the full picture as ipi beats nivo in this measure due to the short duration of ipi treatment while extensive previous data including from keynote 006 show us that pd-1 therapy exceeds ctla4 therapy in virtually every other measure of health. If we look at TFS subtracting toxicity, again ipi nivo beats nivo handily with 10 months vs 4 months but again ipi with 8.5 months slots in the middle. Figures 5A, 5 B and 5C show us the TFS partitioned by grade ≥3 treatment-related adverse events, grade ≥2 TRAEs or by the use of  immune suppressants. To summarize this interesting manuscript, overall survival can be subdivided into TFS and time on therapy and time following subsequent therapy. The TFS can be further subdivided into TFS with or without toxicity (≥2 or ≥3 TRAE or use of immune suppressants). These additional  measures give some granularity to the survival and let us see what exactly survival constitutes, time off treatment and toxicity or time on treatment and with side effects. These measures have several limitations: one of the important ones is introducing differences between treatment arms which by primary outcome measures have no difference by standard PFS or OS and by favoring short duration high toxicity treatments (such as ipi) over low toxicity, low intensity chronic treatments (such as nivo). These are value judgements which matter to many patients but need to be understood as such. In oncology, we need to make these value judgements explicit and make the tradeoffs clear. Ultimately though, the importance of TFS and similar outcome measures may be that they encourage this conversation and help our field open up to what matters. This concludes this JCO Podcast. Thank you for listening.