Assessing the Contribution of Duffy Null Phenotype to Disparities in Eligibility Criteria for Clinical Trials for Newly-Diagnosed Colorectal Cancer

Introduction:

Black individuals with cancer are poorly represented in cancer clinical trials for many reasons including overly restrictive eligibility criteria (Awidi, JCO OP). The Duffy null (DN) phenotype is an erythrocyte phenotype associated with lower normal absolute neutrophil counts (ANC). This is seen in 66% of non-Hispanic Black (NHB) and <1% of non-Hispanic White (NHW) individuals in the United States. A recent analysis (Hibbs, JAMA NO, in press) found that ANC eligibility criteria for the colorectal cancer (CRC) trials which form the basis for National Comprehensive Cancer Network (NCCN) recommended front-line treatment regimens were 1000, 1500 (median value), or 2000/uL, the latter two of which are above the DN lower limit of normal (1200/uL). We aimed to assess the impact of ANC criteria on racial disparities in eligibility among a cohort presenting with newly diagnosed CRC.

Methods:

We identified NHB and NHW adults with newly diagnosed CRC who received curative intent chemotherapy therapy from 2006-2024 using a national, real-world database (TriNetX; Cambridge, MA) comprised of 83 healthcare organizations from the Research network. A common set of eligibility criteria were derived from the CRC trials cited in NCCN guidelines supporting front-line treatment regimens; criteria used in ≥25% of trials were included. Median values were used for criteria with thresholds. Primary objectives were to determine bivariate differences in eligibility proportions by common ANC criterion (1000, 1500, and 2000/uL) between NHB and NHW patients and to assess the contributions of ANC criteria to eligibility proportions for these groups. A secondary objective was to characterize the thresholds of ANC criteria at which eligibility proportions are different between these groups.

Results:

There were 645 patients identified of whom 422 (65.4%) had complete data and were included in the analytic cohort. Median age was 59 years (interquartile range [IQR]: 51-68) and 43.1% (n=182) were female. 80.8% (n=341) were NHW and 19.2% (n=81) were NHB. There were no differences in sex or age by race. There were significant differences in baseline hemoglobin (Hgb; NHW: 12.4 [10.7-13.6]; NHB: 11.0 [9.7-12.2]; p<0.001) and ANC (NHW: 4.5 [3.4-5.7]; NHB: 3.8 [2.6-5.7]; p=0.014). There were no other significant differences by race among the other baseline laboratory values or diagnoses assessed.

At the 1000, 1500, and 2000/uL cutoffs, 69.7% (294/422), 69.4% (293/422), and 68.5% (289/422) patients met criteria for inclusion in a CRC clinical trial. There were no significant differences by race/ethnicity in the proportion of patients excluded by ANC thresholds of 1000 (100% eligible in both groups) or 1500/uL (98.8% NHB eligible vs 99.7% NHW eligible, p=0.8). There were significantly fewer NHB than NHW patients eligible at the 2000/uL threshold (91.4% vs 98.5%, p=0.002). When all exclusion criteria were assessed, 3.7% fewer NHB (60.5% vs 64.2%; p=0.25) and 0.6% fewer NHW patients (70.4% vs 71.0%; p=0.48) were eligible for participation due to not meeting ANC threshold of 2000/uL. Bivariate differences in eligibility proportions were seen for all ANC criteria above 1900/uL (all p<0.05). Notably, differences in eligibility between NHB and NHW patients were also seen for Hgb criteria (13.6% [95% CI: 7%-23%] vs 6.2% [3.9%-9.3%]; p=0.04) but no other clinical or laboratory criteria.

Discussion:

In this counterfactual analysis of a national cohort presenting for frontline CRC therapy, baseline ANC values were significantly different between NHB and NHW patients, likely reflective of differences in Duffy status by race. ANC led to significant differences in trial eligibility at the 2000/uL but not 1500 or 1000/uL thresholds. When other eligibility criteria were applied, an ANC cutoff of 2000/uL did not lead to significant differences in eligibility for NHB patients, but 3.7% fewer NHB patients were eligible compared to only 0.6% of NHW patients. Future studies should evaluate the impact of ANC on clinical trial eligibility for other cancers as well as impact of ANC on eligibility by Duffy status rather than race or ethnicity, which is an imperfect proxy Duffy status and a limitation of this analysis. These data suggest that CRC trials should ensure ANC eligibility criteria are at least <1500/uL and consider including Duffy status in screening criteria and assessment.