Does Baseline Health-Related Quality of Life or Symptom Burden Predict Clinical Outcomes in Patients with Newly Diagnosed CML-CP Treated with Nilotinib or Imatinib?

Nilotinib is a more potent and more selective inhibitor of BCR-ABL in-vitro than imatinib. A randomized Phase III study was conducted comparing these two therapies in adult patients with newly diagnosed Philadelphia Chromosome positive (Ph+) chronic myelogenous leukemia (CML) in chronic phase (CP).

To evaluate whether baseline health-related quality of life (HRQL) or symptom scores were predictive of clinical outcome in this phase III trial.

A total of 846 patients were randomized to receive nilotinib 300 mg BID (n=282), nilotinib 400 mg BID (n=281) or imatinib 400mg QD (n=283). For these analyses to evaluate the impact of baseline HRQL on clinical outcomes, the treatment arms were combined. HRQL was assessed using the Functional Assessment of Cancer Therapy–Leukemia (FACT-Leu). The FACT-Leu consists of four general subscales measuring physical, social/family, emotional, and functional well-being and a 17-item leukemia-specific subscale (LeuS). The FACT-G score is the sum of the four general well-being subscales. The questionnaire was administered at baseline, and at 3, 12, and 24 months. Baseline FACT-G and LeuS scores were divided into three equal-sized groups (i.e., tertiles). Patients with high, mid, and low baseline scores were compared on several clinical outcomes: best molecular response by Month 24, best cytogenetic response by Month 24, treatment discontinuation, hospitalization, dose modification of any kind (interruption, increase, or reduction), grade 3 or 4 adverse events, and missed cytogenetic tests. Chi-square tests were used to compare these dichotomous and categorical outcomes between baseline HRQL groups. The relationship between baseline HRQL scores and Sokal Risk groups was evaluated using analysis of variance. Higher scores on the FACT-G and LeuS indicate better HRQL and less symptom burden.

Mean baseline FACT-G scores were 97.8, 85.0, and 66.3, and mean baseline LeuS scores were 61.6, 53.8, and 42.2, for the high/mid/low tertile groups, respectively. There was no significant association between best molecular response by Month 24 or cytogenetic response by Month 24 and baseline FACT-G scores (p=0.149 and p=0.094, respectively). There was an association between best molecular response by Month 24 and baseline LeuS scores (p=0.043) but no significant association with best cytogenetic response by Month 24 (p=0.316). There were no significant associations with either dose modifications (p=0.252 for FACT-G, p=0.643 for LeuS), grade 3 or 4 adverse events (p=0.531 for FACT-G, p=0.831 for LeuS), or missed cytogenetic tests (p=0.722 for FACT-G, p=0.374 for LeuS). There was a significant association between treatment discontinuation and baseline FACT-G scores (p=0.007). Only 18% of patients with the highest baseline FACT-G scores discontinued treatment compared to 26% in the middle group and 31% in the group with the lowest baseline FACT-G scores. This relationship was not statistically significant for baseline LeuS scores (p=0.070). Fourteen percent of patients with high baseline FACT-G scores were hospitalized at some point during the study, compared to 15% of patients with mid FACT-G scores, and 22% with low baseline FACT-G (p=0.099). However, 11% of patients with high LeuS scores were hospitalized compared to 20% of patients with mid-range LeuS scores, and 21% of patients with low LeuS scores (p=0.018). The mean baseline FACT-G scores were 81.4, 83.5, and 83.4 (p=0.288), and the mean baseline LeuS scores were 51.1, 53.4, and 52.5 (p=0.042), for patients with high, intermediate, and low Sokal scores, respectively.

In patients with newly diagnosed CML-CP, worse general HRQL at baseline was predictive for treatment discontinuation, but not predictive for best molecular response. Leukemia related symptoms at baseline were associated with a greater likelihood of subsequent hospitalization and moderately associated with molecular response. Baseline HRQL was not clearly associated with Sokal scores. These findings suggest that among patients with newly diagnosed CML-CP, examination of baseline HRQL and symptoms may allow patients and clinicians to better anticipate outcomes such as hospitalizations and continuation of therapy.

Beaumont:Novartis: Research Funding. Nowinski:Novartis: Research Funding. Coombs:Novartis: Employment, Equity Ownership. Szczudlo:Novartis: Employment, Equity Ownership. Blakesley:Novartis: Employment, Equity Ownership. Gallagher:Novartis: Employment, Equity Ownership. Burns:Novartis: Research Funding. Cella:Novartis: Research Funding.