Study Design
PreCycle is a multicenter, randomized, parallel-group, Phase IV clinical trial with the primary objective of testing the hypothesis of superiority for time to deterioration (TTD) in patients using the ePRO system „CANKADO active“ over „CANKADO inform“ version. „CANKADO active“ is the fully functional CANKADO-based eHealth treatment support service, including documentation of daily drug intake, daily documentation of QoL, feedback functions (PRO-React) and On-site surveys. „CANKADO inform“ stands for a CANKADO-based eHealth service with a personal login. On-site surveys without feedback functions for the patient and documentation of daily drug intake will be available.
See Figure 1 and Figure 2: Precycle study design
Participants
Eligible patients have histologically or cytologically proven diagnosis of HR+ HER2- locally advanced or metastatic breast cancer and are either candidates to receive palbociclib in combination with aromatase inhibitor or candidates to receive palbociclib in combination with fulvestrant for their locally advanced or metastatic disease. All anticancer treatments used in this study are approved drugs and therapy is in accordance to German treatment guidelines (17). The trial compares two different ways of eHealth support and documentation of patient reported quality of life data.
For Patients who are candidates for palbociclib in combination with aromatase inhibitor or fulvestrant, one prior line of chemotherapy for locally advanced or metastatic breast cancer is allowed in addition to a maximum of two lines of endocrine therapy. For inclusion and exclusion criteria refer to Table 1.
Patients will be stratified according their eligibility of receiving palbociclib with endocrine therapy (AI or fulvestrant) as first or later lines.
Treatment
Patients allocated to the combination of palbociclib with aromatase inhibitor will receive:
- Palbociclib, 125 mg, orally once daily on Day 1 to Day 21 of every 28-day cycle followed by 7 days off treatment and
- Aromatase inhibitor, orally once-daily (continuously).
- Pre- or peri-menopausal patients should additionally receive a LHRH-agonist
Patients allocated to the combination of palbociclib with fulvestrant will receive:
- Palbociclib, 125 mg, orally once daily on Day 1 to Day 21 of every 28-day cycle followed by 7 days off treatment and
- Fulvestrant , 500 mg, intramuscularly on Days 1 and 14 of Cycle 1, every 28 days (± 7 days) thereafter starting.
- Pre- or peri-menopausal patients should additionally receive a LHRH-agonist
Patients of each treatment group (palbociclib / aromatase inhibitor and palbociclib/fulvestrant) will randomized 2:1 in the Intervention Arm A using „CANKADO active“ and in the control Arm B using „CANKADO inform) (see:Figure 2).
Patients will continue to receive study treatment together with the assigned ePRO assessment until investigator assessed disease progression, symptomatic deterioration, unacceptable toxicity, death, or withdrawal of consent, whichever occurs first.
Patients discontinuing the active treatment phase will enter a follow-up period phase during survival further progression and new anti-cancer therapy information will be collected once a year up to 48 months after randomization.
In addition, biomarkers will be assessed as a scientific translational program within this study. Tumor material from available primary tumor and/or available biopsies from recurrent disease will be collected. Blood samples will be collected at four time points during the study when also routine blood samples are mandatory (see: Figure 3).
Statistical considerations
The primary objective is to test the hypothesis of superiority of TTD (time to deterioration) in terms of DQoL (deterioration of quality of life) in patients supported by eHealth therapy management using the FACT-G scale. The sample size is estimated as follows. The trial tests the hypothesis of superiority of DQoL in the test collective (with CANKADO active, 2/3 of patients randomized) in comparison to the control collective (CANKADO inform, 1/3 of patients randomized) according to the standard stratified log-rank test (for a discussion see: Schoenfeld, D. A., & Tsiatis, A. A. (1987). A modified log rank test for highly stratified data. Biometrika, 167-175; Kalbfleisch, J. D., & Prentice, R. L. (2002). The statistical analysis of failure time data (2nd Ed.). John Wiley & Sons, p. 224).
Two strata are determined by
1) first-line patients
2) later-line patients.
First-line patients are assumed to comprise about 5/8 (62.5%) of the entire trial population; the remaining patients (3/8, i.e. 37.5%) are assumed to be treated in later lines. See figure 10.1.1 for a visualization of the patient distribution across strata and arms. The test is calibrated with respect to an alternative hypothesis that asserts a hazard ratio of 0.8 between the two arms (lower hazard in CANCADO active).
To estimate a lower bound for the expected number of deterioration events we assume that at least a progress will generate a DQoL. Therefore, the median PFS reported in the PALOMA 1, and 2 trials (first line patients treated with palbociclib and letrozol), as well as PALOMA 3 (2nd line patients treated with palbociclib and fulvestrant) may serve as model for first-line and later-line strata here. Consequently, we used the upper confidence limit for median PFS reported in Paloma 1 (27.5 months) and Paloma 3 (11 months) to compute conservative estimates for the expected number of events in the first-line and later-line stratum respectively.
In first-line patients, the proposed hazard ratio between CANCADO arms of 0.8 corresponds to about 4 months superior TTD for CANCADO active; in later-line patients, it corresponds to about 2 months superior TTD. Such an increase is assumed to have a clinically relevant benefit.
The sample size was estimated using a validated Monte-Carlo simulation implemented in Python 3.5. In all 4 groups (2 arms with 2 strata each) exponential survival was used as parametric sampling distribution with hazard rate computed from median PFS estimates as indicated above. In addition, an independent exponential censoring process was used to simulate loss to follow-up with 48-month probability of censoring calibrated at 10%.
If 960 patients are recruited (assuming 10% loss to follow-up), we can expect to reject the null-hypothesis with 80% power if a stratified two-sided test of equal hazards between (CANKADO active) and (CANKADO inform) is performed at alpha = 0.05. The corresponding expected number of events across groups is 693.
- To show superiority w.r.t. TTD of QoL in CANCADO active vs. CANCADO inform
- Stratified log-rank test, two-sided, α = 0.05, β < 0.2
- Hazard ratio of 0.8 assumed as alternative hypothesis in sample size calculation
- Sample size of n=960 required to reach 80% power, assuming 10% loss to follow-up
- In total, 693 observed events expected
- Expected distribution of patients across strata and treatment arms given in figure below