Retrospective Analysis of Residual Hairy Cell Leukemia Measured By IHC or Flow Cytometry and Frontline Treatment with Pentostatin on Time to Next Treatment

Introduction: Pentostatin (pent) and cladribine (2CdA) are purine analogs used as frontline therapy for Hairy Cell Leukemia (HCL). Both agents can induce long-term remissions, but neither is considered curative. A better understanding of the factors that predict time to next therapy (TTNT) is needed to improve prognostication and monitoring after frontline therapy. The presence of minimal residual disease (MRD), as measured by advanced molecular techniques such as PCR, IHC, and flow cytometry, has been associated with reduced relapse­‑free survival in HCL. However, there is no uniform method for detecting MRD in HCL. Furthermore, most data regarding MRD has been collected from patients treated with 2CdA and less is known about MRD status after pent. Our institution has a large cohort of HCL patients treated with pent in the frontline setting. Thus, we performed a retrospective analysis to investigate (1) the impact of persistent HCL as measured by IHC or flow cytometry on TTNT after any frontline therapy, (2) the effect of pent on TTNT, and (3) the effect of MRD status after pent on TTNT.

Methods: We conducted a retrospective analysis of HCL patients treated at Ohio State University on an IRB approved protocol. Patients with less than 1 year of follow up were excluded from the analysis. Bone marrow biopsy reports were collected at diagnosis and within 1 year of starting frontline therapy. MRD was defined as the lack of detectable HCL by morphological analysis, but the presence of HCL by IHC or flow cytometry. TTNT was calculated from the first HCL treatment to the initiation of next treatment treating death as a competing risk. The associations between TTNT and patient characteristics as well as percent bone marrow HCL measured by IHC and flow cytometry following completion of first line therapy were evaluated using Fine and Gray subdistribution hazard models accounting for competing risks.

Results: 149 patients met criteria for this analysis. Median age a diagnosis was 51, 78% were male, and 7.4% had variant HCL. There were no significant differences in baseline characteristics for patients treated with pent or 2CdA. Univariable analysis of baseline characteristics found that a palpable spleen at diagnosis (HR 2.65, 95% CI 1.51-4.63, p=0.001) and variant HCL (HR 2.62, 95% CI 1.36-5.05, p=0.004) had significantly shorter TTNT. Patients who received pent (n=39, median number of infusions 14) as frontline treatment had significantly longer (HR 0.40, 95% CI 0.20-0.79, p=0.008) TTNT compared to 2CdA (n=72, Figure 1). By our definition of MRD, there were 47 total patients whose MRD status could be accurately determined. Univariable analysis showed MRD positivity after frontline treatment, regardless of agent, trended towards a shorter TTNT but was not statistically significant (HR 2.5, 95% CI 0.75-8.28, p=0.13). This was also true following treatment with pent (n=25, HR 5.83, 95% CI 0.53-64.43, p=0.15) whereas no trend was seen after 2CdA (n=15, HR 1.05, 95% CI 0.09-11.7, p=0.97). Univariable analysis of all patients, regardless of frontline treatment, showed that IHC ranges of 1-5% (HR 3.16, 95% CI 1.12-8.90, p=0.029), and >5% (HR 9.75, 95% CI 4.12-23.09 p<0.001) had significantly shorter TTNT compared to IHC<1% (Figure 2). Similarly, residual HCL burden measured by flow cytometry of 1-5% (HR 4.63, 95% CI 1.95-10.99, p=0.001) and >5% (HR 12.26, 95% CI 4.27-35.24 p<0.001) had significantly shorter TTNT. Following pent, residual HCL ranges measured by IHC of 1-5% (HR 8.82, 95% CI 1.22-63.84, p=0.031) and >5% (HR 29.17, 95% CI 3.99-213.42, p=0.001) had significantly shorter TTNT compared to <1%. Residual HCL ranges of 1-5% (HR 5.79, 95% CI 1.28-26.12, p=0.022) and >5% (HR 21.2, 95% CI 3.09-145.64, p=0.002) by flow cytometry after pent also had significantly shorter TTNT. Similar results were observed following 2CdA.

Conclusions: Pent and 2CdA are considered efficacious in terms of achieving complete remission in HCL. Our finding that pent was associated with a significantly longer TTNT compared to 2CdA suggests there could be additional clinical differences between these agents, but drawing definitive conclusions is difficult due to the retrospective nature of this work. Finally, our data suggest that assessment of the degree of residual HCL burden measured by either IHC or flow cytometry yields important prognostic information.