AUTHORS
Kana Lucero, MD, Erika Ren2, Katayoun Elizabeth Khalilian, BSc3, Bailey Brooks, BS4, Revant Gattamaraju, MD5, Jimmy Lee, MD6, Mutlu Mete7, Carmelo Gurnari, MD,PhD8, Jaroslaw P. Maciejewski, MD, PhD, FACP9, Eileen Hu, MD10, Taha Bat, MD11
AFFILIATIONS
1 University of Texas Southwestern, Dallas, Texas, United States
2 University of Texas Southwestern, Dallas, Texas, United States
3 University of Texas Southwestern, Dallas, Texas, United States
4 University of Texas Southwestern, Dallas, Texas, United States
5 University of Texas Southwestern, Dallas, TX, United States
6 University of Texas Southwestern, Dallas, TX, United States
7 University of North Texas, Denton, Texas, United States
8 Cleveland Clinic Foundation, Cleveland, Ohio, United States, University of Rome Tor?Vergata, Rome, Italy
9 Taussig Cancer Institute, Cleveland Clinic, Cleveland, Ohio, United States
10 University of Texas Southwestern, Dallas, TX, United States
11 University of Texas Southwestern, Dallas, Texas, United States
ABSTRACT
Introduction: Over 50% of patients with acquired immune aplastic anemia (AA) present with paroxysmal nocturnal hemoglobinuria (PNH) clones at diagnosis. A clinically significant subset of patients with PNH clones will undergo PNH clonal expansion and develop overt PNH disease, requiring lifelong monitoring. No method exists to reliably stratify AA patients who are at a higher risk of clonal expansion and clinical PNH. This study aims to address this gap by developing a logistic regression model capable of predicting the binary study outcome of PNH clonal expansion in AA patients.
Methods: This was a multi-institution retrospective study of 129 pediatric and adult patients with concurrent AA and detectable PNH granulocyte clones <20% at the University of Texas Southwestern and Cleveland Clinic. PNH clonal expansion was defined by the development of a PNH granulocyte clone ≥20% within 6 years of follow up. Patients with less than 6 years of follow up were excluded due to the empirical distribution of clonal expansion timing within the cohort.
Results: Fifty-six patients met criteria for final analysis. Of these 56 patients, 50% were male with a median age of 44 at AA diagnosis. Among these patients, 16 (28.6%) experienced PNH clonal expansion, with a median time to expansion of 25.5 months. Initial PNH granulocyte clone size was the sole predictor of clonal expansion retained by the L1-penalized logistic regression model that assessed 10 clinical variables at the time of AA diagnosis. Patients who experienced clonal expansion had a significantly higher median initial PNH granulocyte clone size compared to those without clonal expansion (6.31% vs. 0.6%, p < 0.001). An initial PNH granulocyte clone size of 3.61% or greater provided 75% sensitivity and 85% specificity for predicting expansion within 6 years of initial clone size detection. Kaplan-Meier curve analysis of expansion free survival (EFS) stratified by initial granulocyte clone sizes of 3.61% demonstrated a significant difference between the two groups (log rank p-value <0.0001). Univariable cox regression showed that patients with an initial granulocyte clone above 3.61% had a 9.7-fold higher hazard of clonal expansion compared to patients below this threshold (HR 9.66, 95% CI 3.09-30.17; p-value <0.001).
Conclusion: L1-penalized logistic regression identified a baseline PNH granulocyte clone size of 3.61% as the optimal threshold for predicting subsequent clonal expansion within 6 years in patients with AA. As current fixed-interval surveillance does not account for individual risk, applying this threshold for risk stratification could help identify patients who may benefit from more frequent monitoring.
DISCLOSURES
Kana Lucero, MD: No relevant conflicts of interest to declare.