Histomorphological responses after therapy with pegylated interferon α-2a in patients with essential thrombocythemia (ET) and polycythemia vera (PV)

Essential thrombocythemia (ET) and polycythemia vera (PV) are BCR-ABL–negative myeloproliferative neoplasms (MPN) characterized by an overproduction of mature blood elements, tendencies toward thrombosis and hemorrhage, extramedullary hematopoiesis (mild splenomegaly), and transformation to overt myelofibrosis (MF)/acute myeloid leukemia (AML) [1, 2]. The therapeutic approach for patients at high risk of thrombosis include cytoreductive therapy such as hydroxyurea [3] or recombinant interferon-alfa [4‐6]. Several clinical studies have reported that pegylated interferon-alfa can induce complete hematologic remission (CHR) in majority of patients with ET and PV, and is even capable to induce complete molecular remission (CMR) in some patients [6‐8]. Similarly, clinical benefit and high HR were noticed in patients with early stages of MF, where interferon-alfa has been shown to possibly reverse bone marrow (BM) fibrosis [9, 10].

Recently, we reported a long-term follow-up of a prospective, phase 2 clinical trial of pegylated interferon alfa-2a (PEG-IFN-α-2a) in 83 patients with PV or ET. After a median follow-up of 83 months, overall HR and MR were noticed in 80 and 63%, respectively, and were shown durable in some patients, with overall median response durations of 66 and 53 months, respectively [8]. Our findings confirmed observations by others that neither baseline molecular status nor achievement of a MR on therapy, are a prerequisite for achieving a HR, and some patients continued to derive significant clinical benefit despite losing their HR or best MR. Histomorphologic evaluation of BM is an important part of establishing a diagnosis [11] and in response assessment in patients with ET/PV [12]. According to the most recent WHO classification, a major criterion for diagnosis of ET is a BM biopsy showing proliferation of the megakaryocyte lineage, with no significant increase in granulopoiesis or erythropoiesis, and very rarely minor (grade 1) increase in reticulin fibers; and for PV, a BM biopsy showing hypercellularity with panmyelosis [11]. The development of BM fibrosis is considered part of the natural evolution of these diseases, and has been reported in up to 20 and 51% of patients with ET and PV during their disease duration, respectively [13‐15]. Although there is an increasing evidence of direct correlation between a higher BM fibrosis grade in patients with ET/PV and signs of progressive disease, such as splenomegaly, cytogenetic aberrations, increased risks of vascular complications, and transformation to overt MF [14‐17], the prognostic value of improvement in BM fibrosis (or other BM characteristics) in patients with PV and ET while on therapy, remains unknown. Despite promising preclinical data showing that the robust immunomodulatory and anti-proliferative effects of interferon-alfa may reverse BM fibrosis via its stimulatory effects on hematopoietic stem cells and inhibitory effects on megakaryopoiesis [18‐20], morphologic remission with interferon-alfa therapy has been rarely reported [21‐23]. Herein, we report the histomorphological BM responses of patients with ET/PV treated with PEG-IFN-α-2a in a prospective, open-label, single-center, phase 2 clinical trial.

Overall, 83 patients (43 PV, 40 ET) were enrolled in a clinical trial of PEG-IFN-α-2a between 2005 and 2009 at our institution. Detailed inclusion and exclusion criteria, treatment schedule, response assessment, statistical considerations and primary analyses have been previously reported [6, 24].

Hematologic and molecular responses, as defined by the European LeukemiaNet [12], were assessed every 3 and 6 months, respectively. Histomorphology of the BM, including fibrosis grading, and morphological responses according to European LeukemiaNet, International Working Group-Myeloproliferative Neoplasms Research and Treatment (IWG-MRT) and European Consensus [12, 25] were assessed by expert hematopathologists at our institution before study enrollment, every 6–12 months during the first 3 years while on study, and then every 12–24 months during the subsequent years. Some patients continued to have BM evaluated every 12–24 months after treatment discontinuation. Complete BM remission (BM-CR) was defined by ELN and IWG-MRT criteria [12] as follows: absence of more than grade 1 reticulin fibrosis and disappearance of megakaryocyte hyperplasia in ET or trilinear hyperplasia with age-adjusted normocellularity in PV. An incomplete, partial response (BM-PR) was defined as fibrosis grading that had improved by at least one grade level on at least 2 consecutive samples taken ≥ 12 months apart, but with persistent morphological features of ET or PV.

Responses and clinical data were analyzed using descriptive statistics. Fisher’s exact and Wilcoxon rank-sum tests were used to compare categorical and continuous variables, respectively. Time to BM-responses and its duration was estimated by using Kaplan–Meier method. GraphPad Prism and SPSS v.23 were used for all analyses.

Among 83 patients enrolled to this study, 58 (70%) had BM samples that were evaluable for BM response assessment. Eighteen patients treated for ≤ 12 months and 7 without consecutive BM samples were not evaluable for BM response. Demographics and clinical characteristics of 58 currently analyzed patients are detailed in Table 1. The median follow-up on study was 84 months (range 36–107), and median number of BM samples per patient was 8 (range 3–12). Median age was 52 years (range 19–75) and 29% (n = 17) were men. Sixty-two percent of patients (n = 36) were previously treated, the majority of them with hydroxyurea (n = 27, 75%). Except for abnormalities in marrow cellularity, atypical megakaryocytes characteristic of ET/PV [11], 35 patients (25 PV, 10 ET; 60%) had reticulin fibrosis of grade MF-1 or MF-2 according to European grading [25]. Median exposure to PEG-IFN-α-2a was 80 months (range 15–107). The starting doses of PEG-IFN-α-2a given subcutaneously in a weekly schedule were 450 µg in 2 patients, 360 µg in 2 patients, 270 µg in 11 patients, 180 µg in 20 patients and 90 µg in 35 patients. The dose was modified throughout the study based on toxicity or lack of efficacy. At the time of this report, out of 58 patients, 32 (55%) are still on study, 24 (75%) of whom are actively being treated.

We report on histomorphological BM responses in 58 patients with ET/PV treated with PEG-IFN-α-2a in a prospective, single-center, phase 2 clinical trial after a therapy for a median of 84 months. Our current research substantiates existing data showing that long-term therapy with PEG-IFN-α-2a is not only able to induce CHR and CMR, but could also lead to complete normalization of BM morphology in a subset of patients. Moreover, PEG-IFN-α-2a was capable to completely reverse BM fibrosis in up to 22% of patients, which is higher than previously observed by other investigators [23, 26]. BM responses were achieved after median time on therapy of 48 months, confirming the observation that a long treatment duration is necessary to achieve such a response. Our study also reports a few novel observations. In contrast to previously published results [21], we have noticed significant reduction in BM cellularity in patients with PV who achieved CMR. Patients with BM-CR seemed to derive the most benefits from the therapy with longest duration of initial responses (CHR and CMR), higher rates of CMR, and the fewest disease-related complications. Importantly, none of the patients who had achieved BM-CR progressed to overt MF, and majority of patients who had lost their initial BM-CR have maintained their hematologic and molecular responses. BM-CR was durable, as 69% of them were sustained at the time of last follow-up. Furthermore, 5 of 6 (83%) patients who continued to have BM evaluations after treatment discontinuation, have maintained their BM-CR after a median duration off therapy of 48 months. However, 2 of them have lost their best MR at the time of last follow-up, and one patients has lost his hematologic response.

Despite the fact that patients with BM-CR appeared to benefit from the therapy the most, we could not identify any predictive markers of or a correlation between the achievement of hematologic/molecular responses and BM response. Achievement of BM response or its durability was not essential for obtaining hematological or molecular response or clinical benefit (i.e., remain symptom free with no organomegaly or thrombotic event), and therefore can’t be considered as major point for a treatment decision making in a clinical practice.

One would expect that loss of HR or MR will be accompanied by the loss of BM response; and it might happen in these patients over time. However, in spite of the loss of HR and MR, these patients remain symptoms free without vascular complications while off any active therapy. The fact that their BM is not showing signs of MPN, yet their laboratory parameters (hematologic or molecular) indicate the presence of disease, further echoes our limited understanding of the dynamics of these responses, and its clinical significance. Moreover, it also introduces a question about the definition of useful objective and clinically meaningful response in these patients, as simply relying on the hematologic parameters in a patient with no clinical signs of disease may not be accurate, especially after or during therapy with a biological agents such as interferon. Hopefully, with longer follow-up and better “response” assessment tools, e.g. deep genome sequencing, we will gain better insight into this dilemma.

Notwithstanding its unclear clinical meaning, our results show the ability of PEG-IFN-α-2a to reverse BM fibrosis, confirming its significant biological activity. Furthermore, the observation that these responses could be sustained even after treatment discontinuation, reinforces its strong immunomodulatory effects with a potential for future treatment approaches. Combining PEG-IFN-α-2a with other therapies that have been shown to possibly reverse or reduce BM fibrosis in patients with MF, should be explored in clinical trials.

However, due to the fact that PEG-IFN-α-2a is capable of controlling the disease only for a limited time and progression (loss of achieved hematologic and molecular response, worsening of marrow fibrosis and even transformation to overt MF or AML) could still occur on therapy without any known biological correlations or predictive markers, we need to improve our ability to identify patients who would benefit the most from the treatment.

Major limitation of our trial is that despite its prospective character, morphologic evaluation was not predefined as one of the study endpoints, and therefore patients were evaluated retrospectively based on availability of BM samples during therapy, directed in part by them staying on therapy/responding, which could cause subjective bias.

In summary, our data confirmed the ability of PEG-IFN-α-2a to induce normalization of BM morphology in a subgroup of ET and PV patients. Although we have noticed that patients with a BM-CR may benefit from the therapy the most, with the longest disease control, no predictive markers to guide clinical decision were found. The significance of these findings as well as the prognostic value of BM response in patients with PV and ET remains incompletely understood. Our findings further highlight the need for more elaborate approach in assessing biological effects of interferon-alfa in ET and PV, to help us understand the heterogeneity and dynamics of clinical, hematologic, molecular and morphologic responses and their significance during long-term therapy with PEG-IFN-α-2a.

PEG-IFN-α-2a is capable to induce complete BM responses in a subset of ET and PV patients, but its correlation with durable clinically relevant treatment benefit warrants further investigation.