Impact of Duffy-null Associated Neutropenia on Chemotherapy Dosing

The proportion of patients with chemotherapy RDI (relative dose intensity) <85% will be estimated separately for the Duffy-null and non-Duffy-null groups. The primary measure of interest is the absolute between-group difference in proportions. A one-sided 90% confidence interval (equivalently, a two-sided 80% confidence interval) will be constructed for this difference. Results will be interpreted relative to the prespecified clinically meaningful threshold of 15% to inform the go/no-go decision for Phase 2. Formal hypothesis testing will be conducted using a one-sided significance level of 10% based on Fisher's exact test.

The absolute difference in the proportion of patients with chemotherapy RDI <85% between the Duffy-null group treated under the purposed dosing algorithm and the non-Duffy-null group treated under standard of care will be estimated. A non-inferiority analysis will be performed using a prespecified margin of 7.5%. A two-sided 80% confidence interval (equivalently, a one-sided 90% confidence interval) will be constructed, and non-inferiority will be concluded if the upper bound of the confidence interval is less than 7.5%.

Data will be collected at baseline and up until 4 months after the start of chemotherapy, as well as an additional one month in order to assess treatment-related neutropenia and risks associated with therapy. Incidence of neutropenic fever, infection, hospitalization, and treatment-related mortality within the first five months of chemotherapy will be summarized descriptively within each group. Incidence proportions and corresponding two-sided 95% confidence intervals will be reported. Longitudinal patterns of neutrophil counts will be compared between the Duffy-null and non-Duffy-null groups using appropriate longitudinal methods such as mixed-effects models. In addition, exploratory time-to-first-event analyses for neutropenic fever, infection, hospitalization, and treatment-related mortality may be conducted using Cox proportional hazards models adjusted for relevant confounders (e.g., age, cancer type), and will be considered supportive of the descriptive incidence summaries.

The physician survey will be completed after each chemotherapy cycle where a dose delay or dose reduction occurs. This survey includes comments about dose changes and explaining the rationale for modifying treatment regimens. Physician survey data will be analyzed to address the exploratory objective of characterizing reasons for chemotherapy dose delays or dose reductions. Analyses will be conducted at the chemotherapy cycle level among cycles in which a dose delay or reduction occurred. The primary exploratory endpoint will be the proportion of treatment modifications attributed to neutropenia versus other reported reasons. These proportions will be summarized separately for Duffy-null and non-Duffy-null patients during Phase 1, with corresponding two-sided 95% confidence intervals. No formal hypothesis testing is planned for physician survey endpoints, and results will be interpreted as hypothesis-generating.