4. Discussion
HCL is a rare indolent B-CLPD and represents 2% of all leukemias (2). Typical clinical features of HCL concern patients with a median age at diagnosis of 54 years with cytopenia (60-80%) and splenomegaly (80-90%) (3). However, HCL can have rare articulation, skin, hepatic, auto-immune or bone involvements (4-8). HCL was first described in 1958 by Bouroncle et al , and the first reported bone involvement (osteolytic type) occurred almost 20 years later, in 1977 by Rhyneret al (9,10). Recently, a literature review of the MEDLINE database looking for articles relating only to bone lesions in HCL was performed by Robak et al (1). Bone involvement was observed in about 3% (0-13%) of patients and very rarely observed at presentation. Thirty-six cases were reported: one-third had symptomatic bone lesions at the time of diagnosis, but the majority of bone involvement was demonstrated later, up to several years (from two months to 22 years). After review of the literature including known cases of HCL with skeletal involvement, clinical features could be more specified. Indeed, the data of Robak et al confirm the preponderant involvement of the femur (head, neck or both) estimated at 75% of the bone lesions in HCL (1). The second main localization of bone involvement is distributed in the axial skeleton (6). There are only a few similar case reports in the literature (6). Bone damage was mainly represented by osteolytic lesions but rare cases of severe osteoporosis and aseptic necrosis have also been described (1). Osteolytic lesions are nonspecific on imaging and the first etiologies considered are solid tumor metastases or myeloma, but the clinician must be vigilant about other diagnoses.
However, isolated bone presentation has been observed in HCL, with bone lesions strongly associated with high tumor burden and bone marrow infiltration (11). The main peculiarity of our two cases consisted in a very symptomatic clinical presentation with exclusive bone pain of increasing intensity that contrasted with a poorly informative biological workup (absence of cytopenia) in relation to a mild marrow infiltration. However, persistent bone pain of increasing intensity can be considered as typical of skeletal involvement in HCL, estimated at 89% of patients (1).
Although our case is an uncommon clinical disease presentation, the phenotype of hairy cell leukemia was typical with an immunological HCL score of 4/4 (12). Moreover, the BRAFV600E mutation was positive, excluding an HCL variant disorder (13). To note, a case mimicking multiple myeloma has also been reported in a patient presenting a painful upper left thigh with several hypermetabolisms at18F-FDG PET/CT and hypergammaglobulinemia with positive immunofixation for monoclonal immunoglobulin A. Neither hairy cells nor BRAFV600E mutation was identified in blood, bone marrow aspirate and bone marrow trephine biopsy from the iliac crest. The diagnosis of HCL was on the core biopsy of the left femoral neck showing hairy cells with the positive BRAFV600Emutation (14). All these unusual clinical presentations exposed here should induce hematologists not to neglect HCL in diagnostic hypothesis in case of bone lesions and, above all, to carry out biopsy.
Concerning 18F-FDG PET/CT, its place in care of HCL patients remains unclear and needs to be determined. Its use in HCL is quite recent and is still being developed. First of all, radiography is insufficient, many cases including ours had normal x-ray. Most of the time, a CT scan or MRI is performed as a second step and then sometimes a 18F-FDG PET/CT. Only a few cases of patients with HCL who underwent 18F-FDG PET/CT have been published and all of them showed hypermetabolic localization whether classical or not (5, 6, 15-17). Some studies indicate that 18F-FDG PET/CT is an interesting tool for the evaluation of the extension of HCL. Indeed, all 18F-FDG PET/CT of HCL patients with bone involvement show hypermetabolisms not only in areas of pain but also in areas without clinical translation, suggesting other existing or early asymptomatic bone lesions. MRI, while often used, is a less specific and less sensitive technology than 18F-FDG PET/CT to detect multiple and/or early localizations. In light of these observations, it can be assumed that bone lesions at the time of diagnosis of HCL are underestimated since imaging is only performed in case of pain and not systematically at the time of discovery of the hemopathy. The delay between the appearance of a bone lesion related to HCL and the development of symptoms is therefore completely unknown. Other studies indicate that 18F-FDG PET/CT is of interest in the evaluation of response to treatment (18). In HCL patients, 18F-FDG PET/CT showed normalization of FDG uptake at all previously pathological sites following cladribine treatment (17). However, it is unclear whether 18F-FDG PET/CT is sufficiently sensitive to determine residual disease, which is currently evaluated in bone marrow and blood (19). Based on these observations and our cases, we think that 18F-FDG PET/CT seems to be quite reliable to assess the disease. An observational trial could be led first for epidemiological purposes in order to better characterize disease presentation, especially at diagnosis by the systematic 18F-FDG PET/CT to better specify the frequency of bone involvement in HCL. All reported cases of HCL with bone involvement have been treated because of pain and not because of cytopenia or splenomegaly. However, Do asymptomatic non-threatening bone lesions found by routine 18F-FDG PET/CT at diagnosis represent treatment criteria on their own? Due to low reported incidence, there exists no established recommendation for this group of patients with HCL with bone involvement. It is clear that this particular presentation needs to be better clarified as to its actual frequency, pathophysiology, prognosis, and response to treatment. Routine 18F-FDG PET/CT would seem to be a first step to answering these questions. Moreover, this examination can guide a possible biopsy if necessary (notably to eliminate a differential diagnosis).
A quarter of patients with HCL are asymptomatic at diagnosis and require regular monitoring for a variable period of months to years before meeting at least one of the validated treatment criteria. These criteria are: absolute neutrophil count <1x109/L, hemoglobin <11g/dL, platelet count <100x109/L, symptomatic splenomegaly, infections or bleeding related to cytopenias (2). Purine nucleoside analogues (PA) (pentostatin and cladribine) remain the standard first-line treatment for HCL, achieving overall response (OR) rates of 90-100% and complete response (CR) in 70 to 90%. No significant difference between the two agents has ever been observed even if no randomized trials have compared both drugs (2,20). Remissions are durable (median 15 years), and survival is close to that for an age-matched general population (21,22).
Rituximab alone has not been formally tested in first-line therapy, but given the relatively low response rates seen in relapsed HCL, it is not recommended except in special circumstances: its use could be reserved for patients ineligible for purine analogues or who have an active infection. In contrast, it has been shown to improve remissions with a combination of rituximab and a PA (23,24). A recent publication comparing concurrent rituximab with delayed rituximab (≥6 months if minimal residual disease (MRD) is detected in blood) showed a marked improvement in MRD negative complete remission rates by concurrent rituximab. The CR rate was 100% with concurrent immunochemotherapy versus 88% with delayed administration, and the rates of undetectable MRD were 97% versus 24%, respectively. Immediate hematologic toxicity was greater (including transient grade 3-4 thrombocytopenia) but recovery was faster at 4 weeks compared with delayed rituximab (25). Chemoimmunotherapy combining cladribine (CDA) daily for five days and rituximab weekly for eight weeks (CDAR regimen) could be an alternative in high-risk HCL: splenomegaly > 3cm, leukocytosis > 1x109/L, hairy cells in the blood at 5x109/L, beta-2 microglobulin > 2N, resistance to PNA, HCL CD38 positive or IGHV unmutated (2). However, neither treatment criteria nor risk stratification take into account unusual clinical presentations, including bone involvement, which is one of the most reported atypical presentations. Our first reported case (the female patient with the threatening right frontal tumefaction) was treated in 2012. At that time, rituximab just began to be used in HCL and data on its efficacy and safety were not yet available. According to the recommendations of the time, she received a CDA in monotherapy and rituximab was delayed because of the detection of MRD in bone marrow six months after PA administration. With therapeutic sequence, she achieved an excellent and durable response. She is still currently in remission. For the second case, we considered our young man’s femur fracture as a high-risk criteria and opted for the CDAR association in light of recent studies in order to maximize the therapeutic response and eradicate the disease as long as possible in fear of a severe bone relapse even if, to date, we have no precise data on the response to the treatment and the prognosis of HCL with bone localizations compared to those with classical presentation. Radiotherapy was not applied for even if it is known that moderate doses are effective in supportive care, nevertheless, side effects are far from negligible and it is not a curative treatment of the disease.