Quality‐of‐life outcomes in patients with advanced melanoma: A review of the literature

For patients with metastatic melanoma, the emergence of immune checkpoint inhibitors and targeted BRAF and MEK inhibitors has markedly enhanced clinical outcomes compared with chemotherapy. However, these novel agents are also associated with unique sets of adverse events, and increased overall survival can lead to prolonged exposure to some novel agents. Therefore, clinical evaluation of these therapies has now included the analysis of health‐related quality of life (HRQoL) in addition to more traditional efficacy and safety outcomes as a measure of patient perception of benefit. The current review focuses on HRQoL outcomes in clinical trials of immune checkpoint inhibitors and targeted therapies in patients with advanced and metastatic melanoma to inform healthcare providers about patient perception of HRQoL as a new perspective in treatment decision making.

becoming a realistic prospect and treatment duration for some therapies has been extended by favorable outcomes, consideration of HRQoL and symptoms is of particular importance.
This review aims to provide an overview of HRQoL in the era of novel therapies in the treatment of advanced melanoma as a means of providing insight for clinical decision making. We will discuss some of the currently available tools used to assess HRQoL in patients with melanoma and examine the recent clinical trials regarding HRQoL outcomes in patients with advanced and metastatic melanoma.
The QLQ-C30 form is a self-reported, 30-item questionnaire and includes five scales that address patients' level of functioning (physical, role, cognitive, emotional, and social) as well as nine symptom scales or single-item questions (assessing fatigue, nausea and vomiting, pain, dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial difficulties; Aaronson et al., 1993;European Organisation for Research and Treatment of Cancer, 2016). The overall score ranges from 0 to 100 points and has demonstrated consistent responses across populations with various cancers (Aaronson et al., 1993;Hjermstad et al., 1995;Kaasa et al., 1995;Osoba et al., 1994Osoba et al., , 1997. In a validation study, a high level of overall agreement was observed between patient answers in the questionnaire and observer interpretations from detailed interviews using the QLQ-C30 questions in an open-ended format (Groenvold et al., 1997). The QLQ-C30 questionnaire has shown meaningful correlations with Eastern Cooperative Oncology Group (ECOG) performance status and degree of weight loss in patients with lung cancer (Ko et al., 2003).
Like the QLQ-C30, the Functional Assessment in Cancer Treatment-General (FACT-G) is considered a core questionnaire generalized for all cancers. Its modification, the FACT-M form, is a validated assessment tool that contains an additional block of melanoma-specific questions (Askew et al., 2009;Cormier, Davidson, Xing, Webster, & Cella, 2005;Cormier et al., 2008). Of 24 melanoma subscale questions, nine items are clearly related to surgical complications (Cormier et al., 2008). The FACT-M form differs from the QLQ-C30 in the way the items are formulated as statements rather than questions. These statements are emotionally colored and encourage patients to "reflect on their thoughts and feelings throughout" (Luckett et al., 2011). Another commonly used HRQoL questionnaire is the EuroQol EQ-5D form, which comprises five questions about pain and physical, social, and emotional well-being, as well as a visual analogue scale, which asks the patient to rate his or her health perception on a 0-to-100 scale.
Regardless of the assessment tool used, however, the magnitude of the change in score that is indicative of a clinically significant difference in QoL requires quantification to facilitate clinical interpretations.
In a study that assessed minimally important differences (MIDs) in the QLQ-C30, researchers compared the changes in a given patient's QLQ-C30 score and the same patient's responses to a separate subjective significance questionnaire that used a 7-category scale ranging from "much worse" through "no change" to "much better." The results indicated that a difference of 5-10 points was perceived by the patients as "a little change," 10-20 points as "moderate change," and >20 points as "very much" change in perceptions of patients in physical, emotional, and social functioning (Osoba, Rodrigues, Myles, Zee, & Pater, 1998). Regarding the FACT-M, MID ranges were between 1 and 9 points depending on the subscale analyzed (Askew et al., 2009). The MID for the generic EQ-5D instrument (Kind, 1996;Pickard, Neary, & Cella, 2007;Rabin & Charro, 2001)
The randomized, double-blind, Phase III MDX010-20 trial compared the efficacy of ipilimumab, gp100 vaccine, and a combination of both agents in patients with previously treated advanced melanoma (Revicki et al., 2012). The results from this trial demonstrated that ipilimumab could prolong survival while minimally impacting QoL, as mea-
Nivolumab was compared with dacarbazine in the randomized, double-blind, Phase III CheckMate 066 trial of patients with treatment-naive BRAF wild-type metastatic melanoma (Long et al., 2016). Completion rates for the QLQ-C30 questionnaire (secondary outcome) and the EQ-5D (exploratory outcome) were approximately 65%-70% for both arms throughout the assessment period. Although The EQ-5D utility scores showed normalization of initial decline back to baseline by week 13 in the nivolumab and combination arms; however, in the ipilimumab arm, recovery of EQ-5D scores did not occur until week 19 (after the four cycles of ipilimumab therapy were already completed). Prior treatment with IL-2, surgery, or radiotherapy was allowed. b Included assessments at the time of disease progression and 4-6 weeks after progression.
who discontinued treatment early were not included in the presented interim QoL analysis.

| BRAF AND MEK INHIBITORS
The introduction of novel BRAF and MEK inhibitors and their combinations has enhanced survival in patients with BRAF V600-mutant metastatic melanoma Chapman et al., 2011;Flaherty et al., 2012;Hauschild et al., 2012;Larkin et al., 2014;Long et al., 2014Long et al., , 2015Robert et al., 2015). HRQoL has been assessed in a number of trials evaluating the efficacy and safety of BRAF inhibitors, MEK inhibitors, or their combination; results have been outlined in Table 2 and Figure 1. Interestingly, in the COMBI-d trial, the fatigue and insomnia scores trended in favor of the combination arm, while gastrointestinal symptom scores (nausea and vomiting, diarrhea, and constipation) trended in favor of the dabrafenib plus placebo arm, observations that are consistent with the known safety profile.
Similar to dabrafenib plus trametinib, the combination of vemurafenib plus cobimetinib has demonstrated improved efficacy versus vemurafenib monotherapy in the randomized, double-blind, Phase III coBRIM trial (Dreno et al., 2015;Larkin et al., 2014). Although no clinically meaningful difference was observed between arms in the mean change in global health status from baseline on the QLQ-C30 questionnaire, a higher percentage of patients in the combination arm reported clinically meaningful score improvements from baseline in social function and in the insomnia, fatigue, and pain symptom scores compared with the vemurafenib monotherapy arm (Dreno et al., 2015).
The open-label, Phase III COLUMBUS trial aimed to compare the safety and efficacy of combination encorafenib plus binimetinib with encorafenib and vemurafenib monotherapies . from these trials is complicated by the fact that patients who discontinued treatment due to side effects were not included in the reported interim QoL analysis (Boughton, 2015). This was not the case in another recent double-blinded trial of adjuvant ipilimumab in high-risk stage III melanoma, where all patients were included in the QoL analysis regardless of disease recurrence or treatment discontinuation (Coens et al., 2017). In this study, 52% of patients in the ipilimumab arm discontinued treatment due to adverse events. This discontinuation rate was associated with a statistically significant reduction in the global health subscale score by QLQ-C30 in patients in the ipilimumab arm, although this difference did not reach the predetermined MID cutoff of 10 points. To date, most studies of HRQoL in patients with advanced or metastatic melanoma have used generic instruments, such as the QLQ-C30 and EQ-5D questionnaires. One validation study revealed potential biases of the questionnaires (Groenvold et al., 1997).
Examples of these possible biases include selective reporting bias, when patients would not report pain if they believed it did not originate from cancer, and demand characteristics, which posits that cues in the research setting can affect patient responses. The latter phenomenon could be more problematic and pronounced in open-label trials (Orne, 1969).
Although these generic questionnaires provide a useful overall assessment despite aforementioned caveats, they may not address some specific symptoms associated with new melanoma treatments, such as rash and pruritus, which can substantially impact patients' HRQoL.
Use of the FACT-M instrument may help clinicians focus on HRQoL issues specific to this patient population and overcome some of the limitations of generic assessments. Acknowledging the differences in the two most commonly used HRQoL assessment tools, QLQ-C30 and FACT-M, Luckett et al. (2011) have proposed a useful algorithm for choosing between them, depending upon what aspects of QoL (specific symptoms or emotional state and social support system) a given study focuses on.
Apart from measurement tools, patients' cultural backgrounds can also influence their perception of QoL. For example, Schwarz and Hinz (2001) reported that the QLQ-C30 global health perception score (where a higher score means perception of better health) was slightly higher in randomly selected Norwegian adults compared with Germans, but the scores for constipation and diarrhea were three times higher (where a higher score means more bothersome symptoms) in Norwegians compared with Germans (Hjermstad, Fayers, Bjordal, & Kaasa, 1998). Such discrepancies might complicate interpretation and clinical application of results from multicenter trials.

| CONCLUSION
Despite the limiting factors discussed herein, we see more wide- Although in some studies, we observed an association between patients' global health perception and the burden of treatment-related adverse events, as a subjective matter, the HRQoL outcomes remain hard to measure. Researchers will need to focus on finding a unified pattern for reporting clinically meaningful HRQoL data.
Providing raw scores from questionnaires and including patients in HRQoL analysis who discontinued therapy due to toxicity would be important steps in this endeavor.