Age-Related Prognoses of Genetic Subtypes in B-Cell Acute Lymphoblastic Leukemia/Lymphoma (B-ALL): Insights from a Decade of National Data

BACKGROUND: Chromosomal rearrangements and aneuploidies are genetic hallmarks of B-cell acute lymphoblastic leukemia/lymphoma (B-ALL) and greatly impact diagnosis, risk-stratification, and treatment. However, despite advances in treatment, overall survival (OS) for older patients with B-ALL remains poor, with the 5-yr OS dropping from >90% in children to <30% in older adults. This age-related decline in OS is poorly understood and has been attributed in part to the distributions of genetic subtypes in different age groups. To address this issue, we used a decade of U.S. data to investigate whether the prognoses of different genetic subgroups also depend on patient age.

METHODS: Patients diagnosed with B-ALL between 2010-2021 were identified from the National Cancer Database, which captures >63% of leukemia diagnoses in the U.S. and includes 7 B-ALL genetic subtypes: BCR::ABL1, ETV6::RUNX1, (high) hyperdiploidy, KMT2A rearrangement, hypodiploidy, TCF3::PBX1, and IGH::IL3. Age was categorized as: infant (<1 y); toddler (1-2 y); early- (3-5 y), mid- (6-11 y), and late-childhood (12-14 y); adolescent and young adult (AYA; 15-39 y); mid- (40-65 y), and late-adulthood (66-90+ y). OS was estimated using Kaplan-Meier techniques and compared using Cox regression. p<0.005 was considered significant.

RESULTS: 35,694 patients with B-ALL were identified, including 4,716 with subtype data: BCR::ABL1 (n=2,594), ETV6::RUNX1 (n=624), high hyperdiploidy (n=581), KMT2A rearrangement (n=375), hypodiploidy (n=369), TCF3::PBX1 (n=133), or IGH::IL3 (n=40). B-ALL with TCF3::PBX1 displayed a unique race/ethnicity distribution: 15.2% arose in non-Hispanic Black patients and 23.5% arose in Hispanic patients, compared to 7.7% and 16.5% of non-TCF::PBX1 subtypes, respectively (X² p=0.001). Conversely, 52.3% of TCF3::PBX1 occurred in non-Hispanic White patients, compared to 67.6% of non-TCF::PBX1 subtypes.

B-ALL with BCR::ABL1 arose primarily during mid adulthood (50% of cases), followed by late adulthood (25% of cases). Likewise, hypodiploidy and IGH:IL3 were more common in AYA through adulthood. Conversely, ETV6::RUNX1 and hyperdiploidy were most common in early childhood. Overall, patients with B-ALL with ETV6::RUNX1 experienced the best 5-yr OS (90.3%, 95CI: 87.1-92.8), followed by hyperdiploidy (81.6%, 95CI: 77.8-84.9), TCF3::PBX1 (79.1, 95CI: 69.8-85.9), and IGH::IL3 (71.4%, 95CI: 51.8-84.1). By contrast, the 5-yr OS was 53.4% for hypodiploidy (95CI: 47.3-59.0), 52.8% for BCR::ABL1 (95CI: 50.4-55.1), and 50.8% for KMT2A rearrangement (95CI: 44.9-56.4).

Notably, across all 7 genetic subtypes of B-ALL, a consistent trend emerged: irrespective of genetic subtype, the OS of patients declined with age. For instance, although the 5-yr OS overall for B-ALL with BCR::ABL1 was 52.8% (95CI: 50.4-55.1), it ranged from 31.3% in late adulthood (95CI: 26.9-35.8) to 72.2% in AYA (95CI: 67.1-76.7) and remarkably reached 100% in the rare toddlerhood/early childhood cases (n=41; 95CI: n/a). Conversely, although the 5-yr OS overall for B-ALL with ETV6::RUNX1 was 90.3% (95CI: 87.1-92.8), it declined from 97.1% in early childhood (95CI: 92.9-98.9) to 14.1% in rare cases arising in adulthood (n=40; 95CI: 3.4-32.0). Consistently across genetic subtypes, the worst OS was experienced by patients presenting in adulthood, including after adjusting for treatments. These trends differed somewhat in patients with KMT2A rearrangements: toddlerhood/early childhood displayed the highest 5-yr OS (81.8%; 95CI: 67.8-90.1) and late adulthood displayed the shortest 5-yr OS (24.4%; 95CI: 12.9-37.8); whereas infants exhibited an intermediate 5-yr OS (55.5%; 95CI: 38.8-69.4).

CONCLUSIONS: Our findings reveal that the prognosis of B-ALL patients depends on age irrespective of genetic subtype, likely related to poor tolerance of chemoimmunotherapy in older patients. Additionally, using national data, we define the age-specific prognosis estimates within genetic subtypes of B-ALL and we identify a unique race/ethnicity distribution of B-ALL with TCF3::PBX1. This study emphasizes the importance of integrating age and genetic information into prognostic assessment and treatment planning for patients with B-ALL to improve outcomes across all age groups.