AUTHORS
Jan Philipp Bewersdorf, MD, Rong Wang, PhD2, Aram Bidikian, MD3, Rohan K Achar, MD, MSCR4, Henry Dumke, MD5, Lourdes Mendez, MD, Ph D6, Alain Mina, MD7, Luis E. Aguirre, MD, MPH8, Maximilian Stahl, MD9, Amer M. Zeidan, MD10, Xiaomei Ma, MD, PhD11, Nikolai Podoltsev, MD, PhD12
AFFILIATIONS
1 Yale University School of Medicine, New Haven, Connecticut, United States
2 Yale University School of Public Health, New Haven, Connecticut, United States
3 Yale University School of Medicine, New Haven, Connecticut, United States
4 Yale University School of Medicine, New Haven, Connecticut, United States
5 Yale University School of Medicine, New Haven, Connecticut, United States
6 Yale University School of Medicine, New Haven, Connecticut, United States
7 Yale University School of Medicine, New Haven, Connecticut, United States
8 Yale University School of Medicine, New Haven, Connecticut, United States
9 Yale University School of Medicine, New Haven, Connecticut, United States
10 Yale University School of Medicine, New Haven, Connecticut, United States
11 Yale University School of Public Health, New Haven, Connecticut, United States
12 Yale University School of Medicine, New Haven, Connecticut, United States
ABSTRACT
Introduction: All-trans retinoic acid (ATRA) plus arsenic trioxide (ATO) transformed the treatment of APL (Lo-Coco et al., N Engl J Med, 2013), yet real-world early mortality remains approximately 15%. We evaluated concordance with National Comprehensive Cancer Network (NCCN) recommendations and factors associated with non-concordant treatment and differentiation syndrome (DS).
Methods: We used the Flatiron Health Research database, an electronic health record-derived deidentified database from 220+ oncology practices in the United States, to assess treatment patterns and outcomes of patients (pts) newly diagnosed with APL during 2014-2024. Adult pts (≥18 years) were included if they met the following criteria: 1) confirmation of t(15;17) by FISH or karyotype within a time window of -30 to +21 days relative to the date of APL diagnosis (Day 0), 2) a documented risk group, defined as low- (white blood cell [WBC] count ≤10x109/L) or high-risk (WBC count >10x109/L) at treatment initiation, and 3) receipt of APL therapy. NCCN concordance was assessed by applying a prespecified treatment classification algorithm that mapped frontline induction regimens to the NCCN recommendations in effect at the time of treatment initiation according to disease risk. Guideline-permitted alternative regimens, clinical trial therapy, and use of additional cytoreductive agents were considered concordant. In the absence of a formal diagnosis code, DS was defined as the receipt of dexamethasone and concurrent cytoreductive therapy with hydroxyurea, gemtuzumab ozogamicin (GO), anthracycline and/or cytarabine within 30 days of the first dose of ATRA or ATO. We conducted multivariable logistic and Cox regression analyses with death as a competing risk to identify variables associated with the receipt of NCCN-concordant treatment and the development of DS, respectively.
Results: Among 316 pts (52.5% male, 59.2% Non-Hispanic White, 31.3% aged 65–85), 227 (71.8%) had low-risk and 89 (28.2%) had high-risk disease. Low-risk pts were older than high-risk pts (median age: 56 vs. 49 years, p<0.01). Compared with low-risk pts, high-risk pts received treatment sooner (median 0 vs 2 days, p<0.01). Treatment began prior to diagnosis in 13.5% of low-risk and in 9.3% of high-risk pts (p=0.31). Among low-risk pts, the most common regimen was ATRA+ATO (n=173; 76.2%), with ATRA+ATO-based regimens increasing from 64.7% in 2014 to 100% in 2024 (ptrend = 0.02). Among high-risk pts, the most common regimen was ATRA+ATO+GO (n=27; 30.3%); the proportion of pts receiving ATRA+ATO+GO and/or anthracycline increased from 13.3% in 2014-2016 to 70.2% in 2020-2024 (ptrend <.01). End-of-induction bone marrow evaluation was performed in 80.6% of low-risk and 86.5% of high-risk pts.
Overall concordance with NCCN recommendations was 86.1%, varying from 71.4% in 2014 to 89.7% in 2024 (ptrend = 0.08). Low-risk pts received NCCN-concordant treatment more frequently than high-risk pts (88.6% vs. 79.8%; p=0.04). Pts receiving NCCN-concordant treatment were younger, with a higher proportion aged <50 years compared to those receiving non-concordant care (42.7% vs. 29.6%; p=0.04). In a multivariable model, high-risk disease (compared with low-risk disease, odds ratio [OR] = 0.35, 95% confidence interval [CI]: 0.17-0.73, p=0.01), and age 65–85 years (compared with 18-49 years, OR=0.34, 95% CI: 0.13-0.89, p=0.03) were associated with lowered odds of receiving NCCN-concordant treatment.
DS occurred in 65 pts (20.8%; 15.1% low-risk and 35.6% high-risk). Among the 65 pts, median time from treatment initiation to onset was 7.5 days for low-risk and 2 days for high-risk pts (p<.01). Pts with high-risk disease had higher risk of DS (hazard ratio [HR] = 3.22, 95% CI: 1.79-5.95, p<0.01), while pts 50-64 years had lower risk of DS (compared with 18-49 years, HR=0.49, 95% CI: 0.27-0.90, p=0.02).
Conclusion: In this real-world cohort of pts with APL, adoption of ATRA+ATO-based frontline therapy increased over time; however, concordance with NCCN treatment recommendations remained suboptimal, particularly among high-risk pts. High-risk disease and older age were independently associated with lower odds of receiving NCCN-concordant therapy, and high-risk pts had an elevated risk of DS. These findings highlight persistent gaps in the implementation of evidence-based treatment strategies and support targeted efforts to improve guideline-concordant care, especially for pts with high-risk APL.
DISCLOSURES
Nikolai Podoltsev, MD, PhD: No relevant conflicts of interest to declare.