Post hoc analysis of CPX-351 versus 7 + 3 to determine quality of patient survival

CPX-351 is a combined drug containing daunorubicin and cytarabine in a 1:5 molar ratio that is used for the treatment of newly diagnosed secondary acute myeloid leukemia. A phase III open-label randomized study, which we reported on previously, demonstrated that CPX-351 significantly improved overall survival, event-free survival, and remission rates in comparison to the standard of care (cytarabine plus daunorubicin chemotherapy 7 + 3 regimen) with a similar safety profile.¹ Although improved overall survival with CPX-351 was demonstrated, the quality of that time without toxicities or relapse was not determined.

Q-TWiST analysis of CPX-351 trial

Recently, Jorge Cortes and colleagues published their post hoc analysis of the CPX-351 trial (NCT01696084) in which they performed a Quality-adjusted Time Without Symptoms of disease or Toxicity (Q-TWiST) analysis to provide information on how much of that survival benefit was ‘valuable’.²

Q-TWiST assesses how much of a patient’s survival time is spent living with toxicities, following disease progression or relapse, or is ‘valuable’ time. This enables the study of the value of prolonged survival to patients in the absence of direct quality-of-life measures.

Study design and patient characteristics²

The study design and patient characteristics of this study were detailed in our previous report. Briefly, patients aged 60–75 years newly diagnosed with high-risk or secondary AML were randomized to receive either

• two cycles of induction with CPX-351 (100 units/m²; daunorubicin 44 mg/m² plus cytarabine 100 mg/m²) given as a 90-minute fusion on Days 1, 3, and 5 (Days 1 and 3 for second induction); or
• the 7 + 3 regimen: cytarabine (100 mg/m²/day) continuous infusion for 7 days, plus daunorubicin (60 mg/m²) on Days 1, 2, and 3 (5 + 2 for second induction).

Patients in complete remission (CR) or CR with incomplete neutrophil or platelet recovery (CRi) were eligible for two cycles of consolidation treatment with either

• CPX-351 (65 units/m²; daunorubicin 29 mg/m² plus cytarabine 65 mg/m²) given as a 90-minute fusion on Days 1 and 3; or
• the 5 + 2 regimen: cytarabine (100 mg/m²/day) continuous infusion for 5 days, plus daunorubicin (60 mg/m²) on Days 1 and 2.

All patients included in the trial were included in the Q-TWiST analysis, and the survival time for each patient was categorized into three health states:

• TOX: time with a Grade 3 or 4 adverse event prior to remission;
• REL: time after relapse; or
• TWiST: time in remission without relapse of Grade 3 or 4 adverse events.

Within each treatment group, the time spent in each health state was determined and the differences between the time spent in these health states was calculated and compared between the arms. To calculate the gain in Q-TWiST, the health state of each arm was weighted according to published quality-of-life standards, represented by health utility (U, scale of 0.0 [death] to 1.0 [perfect health]). In the base case scenario, TOX and REL were assigned health state utility weights of 0.5, and TWiST 1.0. Q-TWiST was calculated as follows:

Q-TWiST = (U_TWiST × TWiST) + (U_TOX × TOX) + (U_REL × REL)

In cases where remission was not achieved, the TWiST and REL states were not reached, and the survival time was assumed to be only in the TOX health state. Q-TWiST gain was calculated to determine the difference in quality of patient survival between the treatment arms and was calculated as follows:

Q-TWiST gain = (Q-TWiST_CPX-351) − (Q-TWiST_{7 + 3})

This absolute value may enable patients to make treatment decisions, but relative Q-TWiST gains may be better at determining clinical benefit. The relative Q-TWiST gain percentage was calculated as follows:

Relative Q-TWiST gain = (Q-TWiST gain ÷ mean survival of the control arm) × 100

A total of 309 patients were included in the intent-to treat population, each arm had a mean age of ~68 years, and around 60% male participants (see Table 1).

Table 1. Patient characteristics*

AML, acute myeloid leukemia; CMML, chronic myelomonocytic leukemia; ECOG, Eastern Cooperative Oncology Group; HMAs, hypomethylating agents; MDS, myelodysplastic syndrome; SD, standard deviation. *Adapted from Cortes, et al.2 †A total of 151 patients received treatment in the 7 + 3 arm.
Characteristic, % (unless otherwise stated)	CPX-351 (n = 153)	7 + 3 (n = 156)
Demographic characteristics
Age
Mean (SD), years	67.8 (4.2)	67.7 (4.1)
60 to 69 years	63	65
70 to 75 years	37	35
Male	61	62
ECOG Performance Status
0	24	29
1	66	57
2	10	14
Clinical characteristics
AML subtype
Therapy-related AML	20	21
AML with antecedent MDS
With prior HMAs	33	35
Without prior HMAs	14	12
AML with antecedent CMML	7	8
de novo AML with MDS karyotype	27	24
Number of induction cycles received, n†  1 %  2 %	153 69 31	151 66 34
Number of consolidation cycles received, n†  1 %  2 %	153 17 15	151 13 8

Results²

In the base case scenario, the relative Q-TWiST gain for CPX-531 vs 7 + 3 was 53.6%, which is above the published clinically important difference of ≥15% (see Table 2).

Table 2. Q-TWiST base case analysis of CPS-351 vs 7 + 3 treatment for AML*

CI, confidence interval; Q-TWiST, Quality-adjusted Time Without Symptoms of disease or Toxicity; REL, time after relapse; SD, standard deviation; TOX, time with any Grade 3 or 4 adverse events or before relapse; TWiST, time in remission without relapse or Grade 3 or 4 adverse events. *Adapted from Cortes, et al.
	Mean (SD) duration of health state, days		Means difference (95% CI), days
	CPX-351 (n = 153)	7 + 3 (n = 156)
Health state
TOX	192 (356)	185 (273)	7 (−63, 78)
TWiST	356 (635)	174 (451)	183 (60, 306)
REL	31 (96)	9 (34.5)	22 (5, 38)
Q-TWiST	468 (623)	271 (449)	197 (76, 319)
Relative Q-TWiST gain	—	—	53.6%

In an analysis looking at only those patients who achieved CR/CRi during the trial, the relative Q-TWiST gain was 39.8%.

Sensitivity analyses, which analyzed patients in both the intent-to-treat and safety populations, across different weights of TOX and REL demonstrated constant relative Q-TWiST gains (48.0–57.6%).

Conclusion

Utilizing Q-TWiST to assess quality of any prolongation of survival offers important information for clinicians and patients when deciding on treatments. Cortes and colleagues report that the Q-TWiST analysis of the CPX-351 trial demonstrate a quality survival gain from CPX-351 treatment, and that the sensitivity analysis shows the benefit of CPX-351 to be robust. The group conclude that these data highlight CPX-351 as an effective therapeutic option for older adults with newly diagnosed high-risk/secondary AML.