Cancer Treatment Reviews 41 (2015) 235–246
Contents lists available at ScienceDirect
Cancer Treatment Reviews journal homepage: www.elsevierhealth.com/journals/ctrv
Tumour Review
Primary and secondary bone lymphomas Carlo Messina a,1, David Christie b,1, Emanuele Zucca c,1, Mary Gospodarowicz d,1, Andrés J.M. Ferreri a,1,⇑ a
Unit of Lymphoid Malignancies, Department of Onco-Haematology, San Raffaele Scientific Institute, Milan, Italy Genesiscare and Bond University, Inland Dr., Tugun, QLD, Australia c Oncology Institute of Southern Switzerland, Bellinzona, Switzerland d Department of Radiation Oncology, Princess Margaret Hospital, Ontario Cancer Institute, Toronto, ON, Canada b
a r t i c l e
i n f o
Article history: Received 15 December 2014 Received in revised form 20 January 2015 Accepted 1 February 2015
Keywords: Primary bone lymphoma Osteolymphoma Extranodal lymphoma Diffuse large B-cell lymphoma Polyostotic lymphoma
a b s t r a c t Recent studies have contributed to the enhancement of clinical and molecular knowledge on bone lymphomas, a group of rare malignancies with particular characteristics. Nevertheless, several questions remain unanswered and the level of evidence supporting some diagnostic and therapeutic decisions remains low. Currently, three different forms of bone lymphomas can be distinguished: the primary bone lymphoma, consisting of a single bone lesion with or without regional lymphadenopathies; the polyostotic lymphoma, consisting of multifocal disease exclusively involving the skeleton; and the disseminated lymphoma with secondary infiltration of the skeleton. The first two forms exhibit a good prognosis, requiring treatments similar to those commonly used for nodal lymphomas of the same category, but several issues regarding the role of surgery and local control of the disease, the sequence of treatment, radiation volumes and doses, management of pathological fractures and prevention of late sequelae deserve particular attention. Due to its rarity, prospective trials exclusively focused on bone lymphomas appear unrealistic, thus, critical revision of our own experience and analyses of large cumulative series as well as molecular studies on archival cases remain valid alternatives to improve our knowledge on this obscure lymphoproliferative malignancy. The present review is based on the analysis of the largest available database of bone lymphomas established under the sponsorship of the International Extranodal Lymphoma Study Group (IELSG) as well as on the critical revision of related literature. We provide recommendations for diagnosis, staging, treatment, and response assessment of these patients in everyday practice as well as for the management of special conditions like pathological fractures, indolent forms and central nervous system prophylaxis. Ó 2015 Elsevier Ltd. All rights reserved.
Introduction Every lymphoma category can involve the skeleton, as an exclusive lesion or as a part of a disseminated disease. Although skeletal involvement is relatively common in non-Hodgkin lymphomas, the available literature on diagnostic and therapeutic management of primary bone lymphomas, that is lymphomas exclusively involving the skeleton, is sparse and fragmentary, mostly reported before worldwide use of rituximab and positron emission tomography (PET). The level of evidence supporting therapeutic decisions in primary bone lymphomas is very low as no prospective trials have been published. The relevant literature is almost exclusively ⇑ Corresponding author at: Unit of Lymphoid Malignancies, Division of OncoHematological Medicine, Department of Onco-Hematology, San Raffaele Scientific Institute, Via Olgettina 60, 20132 Milan, Italy. Tel.: +39 02 26437649; fax: +39 02 26437625. E-mail address: [email protected] (A.J.M. Ferreri). 1 International Extranodal Lymphoma Study Group. http://dx.doi.org/10.1016/j.ctrv.2015.02.001 0305-7372/Ó 2015 Elsevier Ltd. All rights reserved.
constituted by small, retrospective series often furnishing conclusions on unreliable subgroup analyses, and with important interpretation biases due to stage migration and use of obsolete histopathological classifications. An additional bias regards the use of radiotherapy as exclusive treatment in unfit patients, whereas recent advances in supportive care have extended the number of patients treated with curative intent. As a consequence of these methodological caveats and the impossibility of conducting large prospective trials, several therapeutic questions remain open: the role of surgery and radiotherapy, the best radiation volumes and doses, the most effective chemoimmunotherapy combination, and prognostic factors, among others. In this complicated context, large, retrospective studies of cumulative, unselected series remain a valid tool to improve our knowledge on primary bone lymphomas. This review is based on the analysis of the largest available database of bone lymphomas, established under the sponsorship of the International Extranodal Lymphoma Study Group (IELSG),
236
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246
as well as on the critical analysis of related literature. It provides recommendations for the diagnosis, staging, treatment, and response assessment of these patients, and addresses the management of special conditions like pathological fractures, indolent bone lymphomas and CNS dissemination risk in everyday practice. Definition, incidence and epidemiology Criteria used to define and classify primary bone lymphomas changed several times in the last decades. While there is general agreement that cases with a solitary lesion arising in a bone should be considered as a primary bone lymphomas, there is no consensus over the best categorization of cases with multifocal osseous disease or cases with concomitant soft tissue, visceral and/or lymph nodal infiltration [1–5]. In the previous version of the World Health Organization (WHO) classification of tumours of soft tissue and bone, primary bone lymphoma was defined by (1) a single skeletal tumour without regional lymph node involvement, or by (2) multiple bone lesions without visceral or lymph node involvement [6,7]. Conversely, the last versions of the WHO Classification does not provide definition criteria for these disorders [8]. The opinion of the authors is that only cases with a clear bone origin should be considered as primary bone lymphomas, that is, primary bone lymphomas should include cases with a single bony lesion, with or without involvement of regional lymph nodes as well as cases with multiple bony lesions, but without lymph nodal or visceral disease. The latter subgroup is usually called ‘‘multifocal osseous lymphoma’’ or ‘‘polyostotic lymphoma’’, and represents an entity with particular clinical and prognostic characteristics [9]. Disseminated lymphomas with concomitant involvement of the skeleton should be defined as ‘‘secondary bone lymphoma’’. In these cases, bone involvement counts as a systemic extra-nodal site and the disease should be considered to be stage IV [10]. A lymphoma that has arisen in soft tissues, lymph nodes or other organs and infiltrates an adjacent bone secondarily should not be considered to be a primary bone lymphomas. However, this is a common issue in many types of extranodal lymphomas and, similarly, in bone lymphomas the differences are not so clear cut in practice and it may be very difficult to separate these two situations. Special difficulties arise in specific anatomical locations; for instance, it is difficult to distinguish lymphomas primarily arising in nasal-paranasal bones from lymphomas arising in the mucosal surfaces of paranasal sinuses. Similarly, it is often difficult to distinguish the primary site of disease in lymphomas of the spine (i.e., bone or nearby soft tissues) [7]. In many cases, a subjective judgement will be required about whether a case should be categorised as primary bone lymphomas or lymphoma secondarily affecting the bone. The exact incidence of primary bone lymphomas is difficult to define, but it seems to account for about 5% of extranodal lymphomas, 1 High LDH serum level B symptoms Pain Swelling Bulky disease Fracture
15% 34% 9% 82% 40% 23% 15%
38% 30% 24% 92% 45% 15% 25%
62% 65% 30% 90% 34% 32% 29%
Sites of involvement (%) Skull Spinal cord Pelvis Humerus Forearm Femur Forefoot Lymph nodes Cerebrospinal fluid Bone marrow Other
15% 17% 17% 7% 7% 20% 13% – – – 4%
32% 65% 32% 13% 16% 38% 19% – 3% – –
19% 51% 33% 17% 8% 24% 14% 28% 1% 35% –
DLBCL = diffuse large B-cell lymphoma; MB-DLBCL = multifocal bone diffuse large B-cell lymphoma; ECOG PS: Eastern Cooperative Oncology Group Performance Status; LDH lactate dehydrogenase.
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246
The standard contrast-enhanced computed tomography (CT) scan is the primary modality for staging, restaging, and follow-up of lymphoma patients. CT demonstrates the boundaries of any extraosseous extension as well as indicating cortical breakthrough by the tumour [31], and may detect osteolysis, osteosclerosis and fragments of bone sequestra [32]. Magnetic Resonance Imaging (MRI) reveals the extent of disease in more detail, particularly by identifying cortical changes such as linear channels of cortical destruction, as well as intratumoural fibrosis, replacement of trabecular bone and bone marrow by tumour [33,34]. Abnormal signal intensity areas are visible on both T1 and T2 weighted images with minimal contrast enhancement, hypointense in T1weighted and hyperintense in T2-weighted [35]. Accordingly to its relatively high cellularity [35], primary bone lymphomas exhibit restricted diffusion with low apparent diffusion co-efficiency value on diffusion-weighted imaging [36]. An increased post-treatment apparent diffusion co-efficiency value correlates with higher tumour necrosis [37] and is usually consistent with decreased 18 fluorodeoxyglycose (18FDG) uptake on 18FDG-PET scan [38], suggesting tumour response (see below). Functional imaging Bone lesions usually result in increased uptake on technecium scanning, while functional imaging (18FDG-PET) also reveals associated soft tissue involvement. 18FDG-PET display a higher specificity and sensitivity than conventional bone scintigraphy in identifying lymphomatous infiltration of skeleton [39]. 18FDG-PET-computed tomography scan (PET-CT) is a hybrid imaging technique that simultaneously provides functional and anatomical information, with a higher sensitivity and specificity than standard CT scan in lymphoma patients [40–44]. In staging of extranodal lymphomas, sensitivity, specificity and accuracy of PET-CT are 97%, 100% and 98% compared to 87%, 85% and 84% of CT [44]; diagnostic sensitivity is not correlated to lymphoma category [45]. Disagreements between PET-CT and CT usually regard spleen involvement or extranodal lesions, like bone marrow, bone, thyroid, and prostate [45]. In fact, diagnostic sensitivity of CT in lymphoma patients with bone lesions is 17%, and addition of this procedure to PET/CT does not further improve diagnostic sensitivity [46]. Conversely, the addition of 18 FDG-PET to CT scan results in upstaging in 42% of lymphoma patients, with a low rate of false positive scans [46]. Although it is supported only by small, retrospective studies, PET-CT is recommended by the recent Lugano Classification as a standard tool for initial evaluation, staging and response assessment of FDG-avid lymphomas, whereas CT plays a central role in non-avid lymphoma categories [47]. 99
Diagnosis A clinic-radiological suspicion of bone lymphoma must be confirmed by histopathological examination (morphological exam, immunohistochemistry and molecular analyses) of a diagnostic sample obtained by surgical procedure. This is mandatory to confirm lymphomatous nature and to define histotype since current knowledge and available radiologic tools do not allow a distinction among the different lymphoma categories involving the skeleton. Excision biopsy should be avoided; biopsies should be limited in size to reduce the risk of pathological fracture. Some bony sites such as the skull base can be particularly difficult to biopsy [48], with the risk of false negative results and delayed diagnosis [49]. In cases with stage-IIE disease, lymphadenectomy is advisable because it is associated with a lower risk of
237
orthopaedic sequelae and facilitates pathologist’s diagnostic performance. Interestingly, a role of serum soluble interleukin-2 receptor (sIL-2R) levels to achieve a non-surgical diagnosis of primary bone lymphomas was recently suggested [50,51]; normalized serum sIL-2R levels seems to be consistent with a therapeutic response in primary bone lymphomas [52]. These findings need for further investigation and confirmation in well-designed studies. Pathology, morphology and immunophenotype Histopathological diagnosis and categorization of primary bone lymphomas can be difficult due to some frequent problems like crush artefacts in biopsy specimens [53], the presence of reticulum or hyalinised fibrosis, inflammation and some specific diagnostic traps, including sarcoma-like spindle cells and carcinoma-like clustering [6]. Diffuse large B-cell lymphoma (DLBCL) is the most common histological subtype of lymphoma, either primarily or secondarily infiltrating the skeleton. It accounts for 70–80% of all bone lymphomas [1,5,29,30,54], with rare to anecdotal occurrences of follicular, marginal zone, lymphoplasmacytic, anaplastic large cell, NK/T-cell, Burkitt, and Hodgkin lymphomas [55,56]. Morphologically, tumour cells are large sized and consistent with follicle centre or centroblastic cell type, often with nuclear cleavage [4,57,58]; large multilobated cells are reported in around half the cases. Evidence of germinal centre (GC) derivation has been noted in at least 50% of cases [59]. Tumour cells are immunoreactive for B-cell markers: CD45, CD20, CD21, CD45, CD79a [53,60,61]; immunoreactivity for CD75 and CD10 is variable [62]. T-cell markers are usually negative, but small CD3+ cells are often present. BCL-2 and BCL-6 immunoreactivity has been reported in 35% and 69% of cases, respectively [62]. Monotypic IgG, IgH and HLA-DR have been noted [61]. Available data on primary bone T-cell lymphomas are sparse; primary bone T-cell lymphomas do not exhibit distinctive features than extra-osseous T-cell lymphomas. Most reported cases of primary bone T-cell lymphomas are anaplastic large-cell lymphoma (CD3 /+; CD43+/ ; CD30+), often associated with t(2;5)(p23;q35) and ALK-1 expression [62–64]. Cytogenetic and molecular abnormalities Cytogenetic and molecular studies of primary bone lymphomas are limited and usually relate to DLBCL. A monoclonal pattern assessed by PCR is detected in 54% of cases, IgH gene rearrangement in 72% [60] and BCL-2 translocation in 5% of cases [62]. Discrepancy between high BCL2 protein expression (55–70%) and low prevalence of BCL2/IgH rearrangements (4%) [65] indicates that other mechanisms (e.g., gene amplification) may be responsible for the BCL2 protein over-expression in primary bone DLBCL. Studies using fluorescence in situ hybridization has shown BCL2 and cMYC translocations respectively in 28% and 9% of primary bone DLBCL; none of these cases showed BCL6, PAX5, ALK, and CCND1 rearrangements [66]. These features seem to provide specific characteristics to primary bone DLBCL in comparison with other extranodal B-cell lymphomas. Most primary bone lymphomas are ‘‘de novo’’ DLBCL with a follicle-centre origin, which is suggested by frequent MUM1 positivity and less common BCL-6 mutations [67–69]. To date, there are no published studies evaluating gene expression profiles in primary bone DLBCL. Increased expressions of osteoclast-activating factors, such as MIP-1alpha, MIP-1beta and RANKL, in primary bone DLBCL suggest a potential causative role in osteolysis and hypercalcemia [33,70].
238
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246
Table 2 Staging procedures in patients with bone lymphoma.
Table 3 IELSG staging system for DLBCL of the bone.
Test/procedure Demographics and medical history* Physical examination Blood tests# Chest X-ray Contrasted CT scan of the neck, chest, abdomen, and pelvis MRI of bony lesions 18 FDG-PET Bone marrow biopsy
IELSG stage
Lymphoma extension
Ann Arbor stage
IE IIE
Single bony lesion Single bony lesion with involvement of regional lymph nodes Multifocal disease in a single bone or lesions in multiple bones in a disease exclusively limited to the skeleton (without lymph nodal or visceral disease) – called also ‘‘multifocal osteolymphoma’’ or ‘‘polyostotic lymphoma’’ Disseminated lymphoma with at least one bony lesion
IE IIE
IVE
In case of suspicion of involvement of particular organs Cerebrospinal fluid (CSF) examination§ Gadolinium-enhanced brain MRI§ Gastrointestinal tract endoscopy Blood smears
IV
* This includes family history, exposure to a toxic agent, prior malignancies, analysis of comorbidities. # This includes full blood count, alkaline phosphatase, lactate dehydrogenase, erythrocyte sedimentation rate, C-reactive protein, beta-2 microglobulin and protein electrophoresis. § In patients with high risk of CNS dissemination (see ‘‘CNS prophylaxis’’).
IV
IV
but improvement in sensitivity and specificity of radiologic and functional procedures led to increasing detection of multifocal disease. This suggests that multifocal disease, both within a single bone (monostotic disease) and in different bones (polyostotic disease), was underestimated in the past decades. Direct soft tissue infiltration is detected in about 20% of patients, but this is superfluous to staging, therapeutic and prognostic considerations (see Table 4).
Staging Diagnostic tests and procedures in the initial work-up of patients with primary bone lymphomas are listed in Table 2. Stage of disease is defined according to the Ann Arbor staging system [71]. However, this staging system has some important limitations when attempting to analyze different stage subgroups and to compare reported studies. Based on their different prognosis and outcome (Fig. 1), the IELSG staging system has been proposed (Table 3), where bone lymphomas are classified into four different stages: stage IE = single bony lesion; IIE = single bony lesion plus regional lymphadenopathy; IVE = polyostotic lymphoma; and IV = conventional stage-IV lymphoma with skeleton involvement. Most patients with bone DLBCL present with unifocal disease,
Prognosis The survival of patients with primary bone DLBCL is significantly related to disease stage, with the 5-year overall survival (OS) varying from 82% for patients with stage-IE disease to 38% for patients with disseminated DLBCL and skeleton involvement (Fig. 1). In early-stage disease, relapses occur equally within and outside the primary sites of disease, with local and systemic relapse rate of 10% and 17%, respectively [72]. In polyostotic lymphomas, most relapses exclusively involve the skeleton, with involvement of secondary sites in 21% of recurrences; the exclusive involvement of bones both at presentation and at relapse seem to
1,0
Probability OS
0,8
0,6
0,4
0,2
0,0
0
12
24
36
48
60
Months
72
84
96
108
Stage IE Stage IIE Stage IVE Stage IV
Fig. 1. Overall survival curves of bone DLBCL according to stage of disease. Stage IE: single bone lesions without regional lymphadenopathies; stage IIE: single bone lesions plus regional lymphadenopathies; stage IVE: polyostotic DLBCL (multifocal lymphoma involving exclusively the skeleton); stage IV: advanced-stage DLBCL with secondary involvement of the skeleton. Only patients managed with primary anthracycline-based chemotherapy followed or not by radiotherapy were considered. Similar results were obtained when the whole unselected subgroups were considered.
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246 Table 4 More frequent pathological fractures sites at presentation. Site of pathological fracture
Frequency at presentation (%)
Lower limb Spine Upper limbs Pelvis Skull
49 38 27 17 8
suggest a homing mechanism in polyostotic lymphoma [73,74], but molecular and biological explanations for this behaviour remain unknown. In these patients, half of the deaths are due to lymphoma [9]. Overall, prior to the IELSG-14 study [9,26,55,75], no consistent prognostic factors were noted. Several unconfirmed prognosticators have been recorded in undersized, retrospective studies with evident selection biases [1,3,5,29,59,76–79]. Conventional International Prognostic Index (IPI) plays a prognostic role in patients with advanced-stage DLBCL with skeleton involvement but not in polyostotic lymphoma and primary bone lymphomas [9,80]. This is due to the fact that stage and number of extranodal sites have no variability in primary bone lymphomas. Importantly, the IELSG14 study has revealed that age, performance status and serum LDH levels, the three remaining IPI variables, are independently associated with OS in every subgroup of bone lymphomas. The prognostic significance of the phenotypic and genetic characteristics of primary bone DLBCL has been studied in a few series. In fact, the germinal centre (GC) phenotype and related molecular features (CD10 expression, BCL-6 mutations, translocations involving 3q27) are associated with favourable outcome in primary bone DLBCL [30,62,65,81], whereas non-GC signature and related features (immunoblastic variant, MUM1 expression) are unfavourable predictors in DLBCL of the bone [82,83]. Prognosis is poorer in primary bone T-cell lymphomas than in primary bone B-cell lymphomas [30,84], in CD56-positive ALCL in particular [85].
Treatment of different clinical forms Stage IE-IIE DLBCL of the bone (or primary bone DLBCL) Different strategies like chemotherapy, immunotherapy, surgery and radiotherapy have been used to treat primary bone DLBCL, a separate entity usually displaying favourable clinical features and good prognosis. The role of surgery in primary bone lymphomas is limited to sampling for histopathological diagnosis, stabilization and internal fixation of affected bone and resolution of pathological fracture (see below). The use of surgery instead of radiotherapy has been suggested in patients with extensive destruction of weightbearing bones in order to reduce the risk of pathological fractures [86]. Excision of weight-bearing bones and amputation are no longer used to control local disease since they may result in severely impaired outcomes due to post-surgical complications, chemotherapy delay and long-term sequelae [87]. Two large cumulative studies of the Rare Cancer Network and the IELSG suggest anthracycline-based chemotherapy followed, when indicated, by involved-field radiotherapy as first-line treatment for patients with primary bone DLBCL [26,72]. With this strategy, the overall response rate (ORR) is over 90% and the 5-year OS is 84%. In line with other pre-rituximab studies (Table 5), the IELSG study demonstrates that chemo-radiotherapy produces significantly better results than the inverse sequence. CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone) is the firstchoice regimen in primary bone DLBCL; discrepancies mostly regard radiotherapy [88]. Variations in chemotherapy, like substitution of doxorubicin with mitoxanthrone or liposomal doxorubi-
239
cin to reduce cardiotoxicity and the use of infusional, thirdgeneration or 14-day regimens, should follow the same criteria used in nodal DLBCL. A survival benefit of the addition of the anti-CD20 monoclonal antibody rituximab to CHOP in primary bone DLBCL has not been demonstrated. Although a recent retrospective study suggests no survival benefit with the addition of rituximab in localized extranodal lymphomas [92], this antibody has essentially changed the natural behaviour of DLBCL, and, with a few exceptions, is an unavoidable part of first-line treatment for this lymphoma, both in nodal and extranodal forms. Moreover, a positive effect of rituximab in a low-risk malignancy like primary bone DLBCL is also suggested by favourable results reported in patients with low-risk DLBCL enrolled in the MINT trial [93]. In a recent, small series of primary bone DLBCL [89], R-CHOP was associated with a complete remission (CR) rate of 95%, and an 8-year OS of 95%. A few retrospective studies suggest a benefit with the addition of rituximab, with a 3-year PFS of 80–90% after R-CHOP and 50–60% after CHOP [29,90,91]. The survival benefit of the irradiation of affected bones after primary chemotherapy and the potential for complications are a matter of debate in patients with primary bone DLBCL. Prospective studies focused on this therapeutic issue in patients with primary bone DLBCL do not exist, thus, debate is mostly based on indirect evidence from large trials in nodal DLBCL. Before rituximab, a few randomized trials suggested that the addition of radiotherapy is superfluous in DLBCL patients who achieve a CR after primary chemotherapy [94,95]. In the rituximab era, the addition of consolidation radiotherapy significantly improved outcome in a large retrospective series of 469 DLBCL patients treated with R-CHOP combination [96]. Consolidative irradiation of bulky residual masses after R-CHOP14 has been associated with an improved outcome in a prospective series of 166 elderly patients with DLBCL [80]. However, these studies included all stages of disease, both nodal and extranodal DLBCL, and, sometimes, the irradiation volume was not prospectively defined nor described, resulting in important interpretation biases. Regarding primary bone DLBCL, a retrospective study of 161 patients treated with CHOP or derivatives suggests that the addition of consolidative post-chemotherapy radiotherapy does not further improve outcome, with both chemo-radiotherapy and chemotherapy alone producing a 5-year PFS of 74% and 67%, respectively [26]. However, selection biases related to the small size of analyzed subgroups and the management of primary bone DLBCL with more favorable features with chemotherapy alone cannot be excluded. It is clear that the use of intensive immunochemotherapy without consolidation radiotherapy requires formal testing and validation in a randomized trial before it can be used as an alternative treatment for earlystage DLBCL. In particular, the role of PET-driven consolidative radiotherapy remains an appealing open question in the management of limited-stage DLBCL. The choice of radiation volume in primary bone lymphomas patients should result from an accurate risk–benefit analysis. This should consider the risk of exposure of sensitive organs, such as lung, brain, bowel or kidney, among others, and of late bone effects. Although most radiotherapy experts favor whole-bone irradiation, with a dose of 38–40 Gy [88], evidence supporting irradiation of the whole bone and/or regional lymph nodes is limited. In line with previous series [72], the IELSG-14 study has showed a 5-year PFS of 76% and 64% respectively for primary bone DLBCL patients managed with CHOP followed by irradiation of the whole bone or of a part of the affected bone (partial-bone irradiation) [26]. However, the clinical relevance of the irradiation of a whole long bone (a.e., femur, humerus) or a whole flat bone (a.e., vertebra) is very different. In cases where the irradiation of the whole bone appears risky, a smaller volume that maintains wide margins (3–5 cm) around
240
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246
pre-chemotherapy tumour borders within the affected bone is advisable. Margins around the soft tissue or extra-osseous borders can be further restricted to 1–2 cm around post-chemotherapy volumes as the distinction between normal and abnormal tissue is not so problematic, and some individualisation of dose and volume will always be required. Radiation dose depends on the size of the irradiated volume, the anatomical area and response to primary chemotherapy. Data suggesting better results with a dose >40 Gy are outdated [97]. In the IELSG-14 study, no significant survival difference between 47 patients irradiated with a dose 636 Gy and 58 patients irradiated with >36 Gy have been shown, with a 5-year PFS of 72% and 75%, respectively [26]. This is in line with a large randomized trial that has demonstrated that, compared with previous standard doses of 40–45 Gy, a radiation dose of 30 Gy is not associated with loss of efficacy in aggressive lymphomas, with no significant differences in in-field relapse rate, PFS and OS [98]. Polyostotic lymphoma Polyostotic lymphoma is characterized by exclusive multifocal involvement of the skeleton, without affecting lymph nodes or other organs. Although patients with polyostotic DLBCL and patients with disseminated DLBCL and skeletal involvement display similar clinical presentation [9], prognosis is significantly better in patients with polyostotic DLBCL, with an ORR of 92% and 65% (p = 0.002), a 5-year PFS of 57% and 35% (p = 0.01) and 5-year OS of 75% and 37% (p = 0.008), respectively [9]. These differences are independent from the IPI score and other variables, and suggest that polyostotic lymphoma may be a different entity, that deserves to be investigated from biological and molecular points of view. An intriguing finding is that the use of post-chemotherapy radiotherapy in patients with polyostotic DLBCL has been associated with a significantly better outcome, with a 5-year OS of 83% for patients managed with chemoradiotherapy and 55% for patients managed with chemotherapy alone [9]. Accordingly, patients with polyostotic DLBCL should be treated with the same strategy used for disseminated nodal DLBCL, considering consolidation radiotherapy in patients with lesions located in adjacent bones that can be irradiated with acceptable side effects. Advanced-stage DLBCL with skeletal involvement Secondary bone involvement by systemic lymphoma is more common than primary bone lymphomas [53,99,100]. Bone involvement is frequently observed in cases of DLBCL with high tumour burdens and disseminated disease, mostly occurring in patients with an intermediate-high IPI. The prognostic value of bone involvement in DLBCL remains controversial [77,80,101,102]; related articles include heterogeneous series managed with varied strategies, and the prognostic value of skeleton infiltration may has changed after rituximab wide use [80]. In the pre-rituximab era, patients with stage-IV DLBCL and skeletal involvement managed with anthracycline-based chemotherapy, followed or not by bone irradiation, achieved a 65% ORR and a 5-year PFS of 34% [9]. In the rituximab era, R-CHOP treatment, with or without radiotherapy, is associated with a 65% CR rate, and a 5-year PFS of 54% [101,103]. Nevertheless, the clinical benefit of the addition of rituximab to CHOP chemotherapy in patients with stage-IV DLBCL and skeletal involvement remains matter of debate. A retrospective analysis of 292 DLBCL patients with skeletal involvement registered in nine consecutive prospective trials did not demonstrate an advantage with the addition of rituximab [80]. However, one-fifth of patients had primary bone lymphoma and not disseminated disease, and 80% of analyzed patients had been treated without rituximab, which may result in unbalanced comparisons and low-level
evidence. Thus, chemoimmunotherapy of bone DLBCL should be the same routinely used for nodal DLBCL, that is a combination of CHOP or CHOP-like regimen plus rituximab. The clinical benefit of irradiation of involved bone in patients with disseminated DLBCL is an important open question. A beneficial effect of this strategy was suggested by a recent retrospective study of the German High-Grade Non-Hodgkin lymphoma Study Group [80]. The analysis of the effect of radiotherapy was restricted to patients with disease responsive to anthracycline-based chemotherapy ± rituximab, with a significantly and independently better outcome among patients who received radiotherapy to sites of skeletal involvement. Contrasting results in smaller series should be considered with caution because it is possible that patients with more aggressive disease were more likely to receive irradiation [29]. The discretionary nature of the indications for bone irradiation and heterogeneity of bone involvement (bulky or not, single or multiple) preclude firm conclusions concerning the usefulness of radiotherapy in patients with advanced-stage DLBCL and skeleton involvement. Importantly, a recent study suggested a benefit for the rational use of PET-guided radiotherapy in patients with advanced DLBCL and residual abnormalities at CT scan after RCHOP [104]. In fact, 18FDG-PET is an important tool to define therapeutic response (see ‘‘Response to treatment’’), and, in that study [104], irradiation of PET-positive residual disease resulted in survival figures similar to those reported for patients with negative post-R-CHOP PET. This strategy may avoid superfluous radiation exposure of PET negative patients. However, 18FDG-PET exhibits some technical and interpretative limitations (see below), and whether this procedure can identify those primary bone lymphoma patients who can be spared from radiotherapy remains to be shown in appropriately designed studies. Indolent bone lymphomas Indolent primary bone lymphomas represents up to 8% of all bone lymphomas (Table 6), and these anecdotal cases are not specifically described in histopathological series [105,106]. The largest series of indolent bone lymphomas was recently reported by the IELSG [55]. Out of an international series of 499 patients with a diagnosis of NHL and skeleton involvement, 26 (5%) patients had an indolent bone lymphomas; ten of them had a small lymphocytic lymphoma, 10 had a follicular lymphoma and 6 had a lymphoplasmacytic lymphoma. Eleven patients had limited stage and 15 had advanced disease. No patients with Richter’s syndrome were identified [107]. With all the limitations of a pre-PET retrospective study, a significantly better outcome in patients with small lymphocytic lymphoma has been recorded; these patients frequently achieved long-term remission and rarely died of lymphoma. The prognosis of patients with limited-stage lymphoplasmacytic lymphoma or follicular lymphoma was less favourable, with a higher proportion of tumour dissemination, possibly reflecting limitations in staging sensitivity. Patients with primary indolent bone lymphomas in the IELSG series had been managed with varied first-line treatments, obtaining an ORR of 73%, and a 5-year PFS and OS of 37% and 46%, respectively. Performance status and stage of disease were independently associated with OS. All patients with limitedstage indolent bone lymphomas managed with radiotherapy achieved a CR, and 63% of them remained relapse-free; conversely, only 33% of patients managed without radiotherapy achieved a CR, and all of them experienced relapse [55]. The addition of chemotherapy was not associated with improved OS in irradiated patients with limited-stage indolent bone lymphomas. Patients with advanced-stage indolent bone lymphomas showed response and survival rates similar to those reported in patients with the same stage and histological type but without skeleton involvement. Accordingly, patients with localized indolent bone lymphomas
241
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246 Table 5 Largest available series of bone lymphomas. Reference
Median age
Years of enrollment
ORR (%)
DFS
OS (years)
Ostrowski 261 [6] Heyning 60 [81] Dosoretz 30 [57] De Camargo 24 [125] Ueda [126] 34
45
1907–1982 Femur (21%)
0
NA
NR
58
0
56
46% (5)
53% (10 – uni) 35% (10 – mul) 61% (5)
48
76
12
0
NR
40% (5)
63% (5)
37
8
43
0
NR
NR
70% (5)
44
26
10
41
0
NR
NR
80 100
81 100
30 0
14 32
56 68
6 0
NR 100
81% (5) 48% (10)
75% (5 – st I) 50% (5 – st II) 88% (5) 53% (10)
Beal [1] Marshall [127] Rathmell [5] Dubey [128] Fidias [129] Fairbanks [3] Horsman [27] Bacci [82] Baar [116] Christie [124] Stein [130]
1967–1988 NR
85
100
0
56
33
0
NR
39% (10)
40% (10)
52
1967–1992 Femur (20%)
98
100
9
11
80
0
NR
63% (10)
60% (10)
37 63
41 63
100 100
0 3
0 79
100 16
0 0
100 NR
55
100
16
41
38
0
57
73% (10) 90% (5 – CT + RT) 57% (5 – RT) NR
87% (10) NR
37
1970–1995 Appendix (75%) 100 1970–1989 Long bone 93 (57%) 1970–2003 Pelvis (24%) 73
26 17 70
NR 36 60
1972–1982 Fenur (23%) 1975–1992 Femur (29%) 1973–1999 Spine (29%)
80 100 65
100 88 80
0 30 0
0 6 44
100 64 56
0 0 0
100 94 83
88% (13) 77% (3) NR
88% (13) 77% (3) 59% (5)
19
54
1979–2000 NR
95
58
42
0
58
0
95
NR
Messina [9] 37 Govi [55] 26 161 Bruno Ventre [26] Zinzani [59] 52 Gianelli 28 [62] Barbieri 77 [76] 131 Ramadan [29] Lewis [131] 28 Ford [86] 22
53 60 55
1980–2005 Spine (65%) 1980–2005 Pelvis (47%) 1980–2005 Femur (20%)
100 0 100
0 42 100
35 30 8
0 15 14
65 55 78
0 0 0
92 73 91
90% 87% 56% 25% 68%
58 51
1982–1998 Femur (27%) 1982–1999 Femur (24%)
85 93
79 100
16 20
21 3
63 74
0 0
90 NA
84% (8 – CR pts) 75% (4)
68% (9) 78% (4)
42
1983–2001 Extremities (51%) 1983–2005 Spine (29%)
97
100
0
13
87
0
95
76% (15)
88% (15)
79
46
44
8
48
21
NR
40% (10)
41% (10)
89 91
71 77
36 18
14 0
50 82
0 0
NA NR
46% (6) 85% (10)
60% (6) 74% (10)
Bayrakci [87] Cai [72] Catlett [90]
20
48
1984–1994 Femur (39%) 1985–2003 Long bone (50%) 1986–1997 Femur (24%)
NR
70
35
0
65
0
65
NR
116 30
51 49
86 90
100 70
13 16
12 10
75 71
3 40
91 NR
62% (4) NR
78% (st I) 16% (st IV) 72% (10) 73% (5)
20 33 28
44 40 47
100 100 68
90 39 32
15 48 50
15 0 0
65 52 50
0 39 0
NA 88 89
NR NR 77% (3)
74% (5) 75% (4) 84% (3)
21
34
100
10
48
0
52
100
95
100% (8 – CR pts)
95% (8)
53 31
52 55
1999–2009 Long bone (38%) 2000–2007 Femur (24%) 2000–2007 Femur (26%)
90 97
77 68
12 0
21 0
62 100
37 19
92 96
83% (4) 64% (5)
100% (4) 90% (5)
28
41
2001–2009 Femur (25%)
100
78
30
3
66
0
NR
62% (2)
NR
de Leval[65] Kim [132] Maruyama [30] Pellegrini [89] Alencar[91] Christie [133] Nasiri [134]
No. of patients
Stage I–II (%)
CT (%)
RT (%)
77
68
6
63
22
1943–1996 Femur (24%)
92
62
NR
8
58
1950–1978 Femur (30%)
93
100
0
38
1955–1999 Spine (25%)
83
NA
56
1961–1988 Pelvis (29%)
53
82 28
48 52
1963–2003 Femur (27%) 1962–1997 Femur (18%)
27
53
45
63 45 50
Most common primary site
1987–2008 Spine (28%) 1989–2005 Long bone (57%) 1990–2000 Femur (35%) 1992–2010 Pelvis (39%) 1995–2004 Pelvis (41%)
Aggressive type (%)
assessed with complete staging procedures are candidate for radiotherapy alone [55,108], while patients with advanced-stage indolent bone lymphomas should be managed like other patients with disseminated indolent lymphomas.
CT + RT (%)
Rituximab (%)
(6 – st I–II) (6 – st IV) (5) (10) (5)
50% (10)
74% (5) 29% (10) 75% (5)
categories, being the most common forms [84]. Skeletal involvement does not seem to play a prognostic role in these patients. Currently, there is no evidence suggesting that these patients should be treated differently to patients with the same lymphoma category and stage of disease but without skeletal involvement.
Uncommon aggressive bone lymphomas Special therapeutic issues Some anecdotal cases of aggressive lymphomas other than DLBCL primarily arising in the bones have been anecdotally reported (Table 6) [82,109]; lymphoblastic and Burkitt lymphoma, among B-cell categories, and anaplastic large cell lymphoma and peripheral T-cell lymphoma not otherwise specified, among T-cell
Response to treatment Conventional radiologic and functional procedures exhibit limitations in the assessment of response to treatment in primary
242
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246
Table 6 Pathological classification with relative frequencies in the largest series of PBL. Histology Diffuse large B-cell Diffuse mixed Follicular Peripheral T-cell Extra-nodal marginal zone Mantle cell Small lymphocytic Transformed MALT Burkitt’s/Burkitt like Anaplastic large T-cell Lymphoma NOS Plasmocitoid Immunoblastic Lymphoblastic Unclassifiable low grade Other
Zinzani [59] 44 (84%) – 2 (4%) – – – 2 (4%) – 2 (4%) 2 (4%) – – – – – –
Beal [1] 66 (80%) 4 (5%) 3 (4%) – – – 3 (4%) – 1 (1%) – 2 (2%) 1 (1%) 1 (1%) 1 (1%) – –
Ramadan [29] 103 (79%) – 7 (5%) 2 (1.5%) 4 (3%) 1 (1%) 2 (1.5%) – 2 (1.5%) 4 (3%) – – – 2 (1.5%) 1 (1%) –
Alencar [91] 44 (83%) – 3 (5.7%) 2 (3.8%) 1 (1.9%) 1 (1.9%) 1 (1.9%) 1 (1.9%) – – – – – – – –
Cai [72] 91 (78%) – 7 (6%) – – – – – – 6 (5%) – – – – – 12 (11%)
Fig. 2. CT scan (left column) and PET (right column) features in a patient with a stage-IE DLBCL of the left femur. CT scan (A) and PET (B) performed at diagnosis show evident osseous abnormalities and increased FDG uptake (arrows), respectively. CT scan (C) and PET (D) performed immediately after treatment conclusion show persistence of osseous abnormalities and reduced FDG uptake (arrows), respectively. CT scan (E) and PET (F) performed at one year of follow-up show respectively persistence of evident osseous abnormalities (arrow), which may be interpreted as residual disease (false positive), and remission of FDG uptake (arrow).
bone lymphomas patients; architectural osseous distortion can remain unchanged for many years in X-rays and CT scans, and residual uptake can persist in bone scintigraphy and 18FDG-PET (Fig. 2). Bone remodelling and artefacts due to stabilizing hardware complicate the assessment of local disease control; in particular, artefacts caused by intramedullary titanium rods interfere with interpretation of changes and limit the usefulness of post-treatment MRI and CT scans. These persisting changes prevent the designation of CR using international criteria [99]. In this case, the
designation of response is usually made upon resolution of any symptoms, particularly pain, signs such as swelling, resolution of soft tissue infiltration at CT, MRI and/or PET, evidence of bone healing on X-ray, and absence of signs of progressive disease. Replacement of intraosseous tumour and adjacent bone marrow with fatty marrow and sclerosis in follow-up MRI scans can help to confirm CR. 18FDG-PET is more useful than bone scanning in the assessment of response to treatment [110,111], and may be used to drive consolidative radiotherapy [104]. However, although 18FDG-PET
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246
has a high sensitivity for the detection of osseous disease [39], its predictive value is limited in primary bone lymphomas due to false-positive findings related to post-chemotherapy bone remodeling, inflammation, marrow hyperplasia and bone necrosis and false negatives related to the resolution of the equipment, the technique used and the variability of FDG avidity among lymphoma categories [43,112–114]. As a logical consequence of these limitations, response rates in primary bone lymphomas trials should be considered with caution. In routine practice, the current plan in the absence of definitive evidence of local or distant failure is to continue surveillance with CT scans [115], whereas the role of the PET scans as a follow-up procedure remains to be defined [99]. Management of pathological fractures Pathological fractures occur in 10–15% of patients with DLBCL of the bone, with the humerus being the most frequently affected bone (66%) [10,29,91]. The analysis of the IELSG-14 study has clearly showed that the presence of pathological fracture at presentation is associated with independently worse outcome in patients with primary bone DLBCL, with a 5-year OS of 54% for patients with pathological fracture and 68% for patients without this complication [75]. Unfortunately, this study was not able to identify pre- and post-treatment variables predicting the occurrence of a pathological fracture (i.e., lysis of the cortex, tumour size, soft tissue infiltration). The goals of initial surgical stabilization of the pathological fracture are usually to preventing further bone destruction and displacement, to enable weight-bearing, to assist pain relief or healing and to obtain a better quality of life. While these are all important goals, there is no evidence of an improved cancerrelated outcome with initial surgical stabilization of the pathological fracture, with a 5-year OS of 45% and 54%, respectively for operated and not operated patients [75]. These data suggest that any initial surgical stabilization should be kept to a minimum, and used to improve patient’s quality of life and prevent bone disintegration only if chemotherapy delays can be avoided. A pathological fracture or a high risk of fracture may lead the treating physician to start the treatment program with radiotherapy. Importantly, irradiation of a fractured bone before chemotherapy does not improve disease control or survival, when compared to the chemo-radiotherapy sequence [75]. These results should be taken into account with caution since a selection bias related to initial irradiation of patients with more destructive lesions, and, thus, with more aggressive lymphoma, cannot be excluded. Nevertheless, patients with pathological fracture should be managed like nodal DLBCL, including anthracycline-based chemotherapy followed by consolidation radiotherapy to the fractured bone with 30–40 Gy [75,76,116]. Radiation volumes and doses should follow the above-discussed recommendations for primary bone DLBCL patients. CNS prophylaxis CNS dissemination is an early and fatal event reported in 5% of DLBCL [117,118]. The risk of CNS recurrence associated with skeletal involvement is a matter of debate, with rates of 4% and 0.6%, respectively for DLBCL patients with and without skeletal involvement [80,119]. In the IELSG-14 study, CNS involvement, usually meningeal lymphomatosis, occurred in 2.5% of patients with primary bone DLBCL [26], 5% of patients with polyostotic DLBCL [9,120] and 2% of patients with stage-IV DLBCL and skeleton lesions [9,75]. Probably, these rates would be even lower in the rituximab era, since CNS dissemination has been reduced in DLBCL patients treated with this antibody [121]. Importantly, the small number of events hampers the identification of reliable predictors
243
of CNS recurrence in primary bone lymphomas studies, but available evidence suggests that CNS prophylaxis is superfluous in primary bone DLBCL. Noteworthy, an accurate CSF assessment, brain MRI and personalized prophylaxis are advised in patients with disseminated DLBCL and involvement of bones close to the CNS (skull and/or spine). CNS recurrence occurs in 7% of these patients, especially in patients with other high-risk features [9,75]. Late iatrogenic sequelae Chronic pain, limb dysfunction and late pathological fracture are the most common iatrogenic sequelae in patients with bone lymphoma. The involvement of the femur, pelvis and the spine is more commonly associated with chronic symptoms, and a second pathological fracture after anti-lymphoma treatment has been recorded in 10% of patients with secondary bone DLBCL [75]. Such fractures can occur in the absence of local recurrence and can lead to persisting non-union with subsequent disability, mostly in bones of the lower limb. Contributing factors may include architectural disturbance due to previous tumour, pathological fracture before treatment and the existence of other medical conditions such as postsurgical osteomyelitis, Paget’s disease and osteoporosis, particularly among elderly women. Both radiotherapy and chemotherapy have been implicated as incurring a higher risk of subsequent pathological fracture; in particular, this risk has been attributed to corticosteroids and large radiation fields, fractions and doses [122,123]. These and other risk factors, such as involvement of weight-bearing bones, local recurrence at the fracture site and the size of the biopsy defect should be accurately considered. Orthopaedic fixation of fractures, limitation of the use of corticosteroids, reduction of radiation doses to
Contents lists available at ScienceDirect
Cancer Treatment Reviews journal homepage: www.elsevierhealth.com/journals/ctrv
Tumour Review
Primary and secondary bone lymphomas Carlo Messina a,1, David Christie b,1, Emanuele Zucca c,1, Mary Gospodarowicz d,1, Andrés J.M. Ferreri a,1,⇑ a
Unit of Lymphoid Malignancies, Department of Onco-Haematology, San Raffaele Scientific Institute, Milan, Italy Genesiscare and Bond University, Inland Dr., Tugun, QLD, Australia c Oncology Institute of Southern Switzerland, Bellinzona, Switzerland d Department of Radiation Oncology, Princess Margaret Hospital, Ontario Cancer Institute, Toronto, ON, Canada b
a r t i c l e
i n f o
Article history: Received 15 December 2014 Received in revised form 20 January 2015 Accepted 1 February 2015
Keywords: Primary bone lymphoma Osteolymphoma Extranodal lymphoma Diffuse large B-cell lymphoma Polyostotic lymphoma
a b s t r a c t Recent studies have contributed to the enhancement of clinical and molecular knowledge on bone lymphomas, a group of rare malignancies with particular characteristics. Nevertheless, several questions remain unanswered and the level of evidence supporting some diagnostic and therapeutic decisions remains low. Currently, three different forms of bone lymphomas can be distinguished: the primary bone lymphoma, consisting of a single bone lesion with or without regional lymphadenopathies; the polyostotic lymphoma, consisting of multifocal disease exclusively involving the skeleton; and the disseminated lymphoma with secondary infiltration of the skeleton. The first two forms exhibit a good prognosis, requiring treatments similar to those commonly used for nodal lymphomas of the same category, but several issues regarding the role of surgery and local control of the disease, the sequence of treatment, radiation volumes and doses, management of pathological fractures and prevention of late sequelae deserve particular attention. Due to its rarity, prospective trials exclusively focused on bone lymphomas appear unrealistic, thus, critical revision of our own experience and analyses of large cumulative series as well as molecular studies on archival cases remain valid alternatives to improve our knowledge on this obscure lymphoproliferative malignancy. The present review is based on the analysis of the largest available database of bone lymphomas established under the sponsorship of the International Extranodal Lymphoma Study Group (IELSG) as well as on the critical revision of related literature. We provide recommendations for diagnosis, staging, treatment, and response assessment of these patients in everyday practice as well as for the management of special conditions like pathological fractures, indolent forms and central nervous system prophylaxis. Ó 2015 Elsevier Ltd. All rights reserved.
Introduction Every lymphoma category can involve the skeleton, as an exclusive lesion or as a part of a disseminated disease. Although skeletal involvement is relatively common in non-Hodgkin lymphomas, the available literature on diagnostic and therapeutic management of primary bone lymphomas, that is lymphomas exclusively involving the skeleton, is sparse and fragmentary, mostly reported before worldwide use of rituximab and positron emission tomography (PET). The level of evidence supporting therapeutic decisions in primary bone lymphomas is very low as no prospective trials have been published. The relevant literature is almost exclusively ⇑ Corresponding author at: Unit of Lymphoid Malignancies, Division of OncoHematological Medicine, Department of Onco-Hematology, San Raffaele Scientific Institute, Via Olgettina 60, 20132 Milan, Italy. Tel.: +39 02 26437649; fax: +39 02 26437625. E-mail address: [email protected] (A.J.M. Ferreri). 1 International Extranodal Lymphoma Study Group. http://dx.doi.org/10.1016/j.ctrv.2015.02.001 0305-7372/Ó 2015 Elsevier Ltd. All rights reserved.
constituted by small, retrospective series often furnishing conclusions on unreliable subgroup analyses, and with important interpretation biases due to stage migration and use of obsolete histopathological classifications. An additional bias regards the use of radiotherapy as exclusive treatment in unfit patients, whereas recent advances in supportive care have extended the number of patients treated with curative intent. As a consequence of these methodological caveats and the impossibility of conducting large prospective trials, several therapeutic questions remain open: the role of surgery and radiotherapy, the best radiation volumes and doses, the most effective chemoimmunotherapy combination, and prognostic factors, among others. In this complicated context, large, retrospective studies of cumulative, unselected series remain a valid tool to improve our knowledge on primary bone lymphomas. This review is based on the analysis of the largest available database of bone lymphomas, established under the sponsorship of the International Extranodal Lymphoma Study Group (IELSG),
236
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246
as well as on the critical analysis of related literature. It provides recommendations for the diagnosis, staging, treatment, and response assessment of these patients, and addresses the management of special conditions like pathological fractures, indolent bone lymphomas and CNS dissemination risk in everyday practice. Definition, incidence and epidemiology Criteria used to define and classify primary bone lymphomas changed several times in the last decades. While there is general agreement that cases with a solitary lesion arising in a bone should be considered as a primary bone lymphomas, there is no consensus over the best categorization of cases with multifocal osseous disease or cases with concomitant soft tissue, visceral and/or lymph nodal infiltration [1–5]. In the previous version of the World Health Organization (WHO) classification of tumours of soft tissue and bone, primary bone lymphoma was defined by (1) a single skeletal tumour without regional lymph node involvement, or by (2) multiple bone lesions without visceral or lymph node involvement [6,7]. Conversely, the last versions of the WHO Classification does not provide definition criteria for these disorders [8]. The opinion of the authors is that only cases with a clear bone origin should be considered as primary bone lymphomas, that is, primary bone lymphomas should include cases with a single bony lesion, with or without involvement of regional lymph nodes as well as cases with multiple bony lesions, but without lymph nodal or visceral disease. The latter subgroup is usually called ‘‘multifocal osseous lymphoma’’ or ‘‘polyostotic lymphoma’’, and represents an entity with particular clinical and prognostic characteristics [9]. Disseminated lymphomas with concomitant involvement of the skeleton should be defined as ‘‘secondary bone lymphoma’’. In these cases, bone involvement counts as a systemic extra-nodal site and the disease should be considered to be stage IV [10]. A lymphoma that has arisen in soft tissues, lymph nodes or other organs and infiltrates an adjacent bone secondarily should not be considered to be a primary bone lymphomas. However, this is a common issue in many types of extranodal lymphomas and, similarly, in bone lymphomas the differences are not so clear cut in practice and it may be very difficult to separate these two situations. Special difficulties arise in specific anatomical locations; for instance, it is difficult to distinguish lymphomas primarily arising in nasal-paranasal bones from lymphomas arising in the mucosal surfaces of paranasal sinuses. Similarly, it is often difficult to distinguish the primary site of disease in lymphomas of the spine (i.e., bone or nearby soft tissues) [7]. In many cases, a subjective judgement will be required about whether a case should be categorised as primary bone lymphomas or lymphoma secondarily affecting the bone. The exact incidence of primary bone lymphomas is difficult to define, but it seems to account for about 5% of extranodal lymphomas, 1 High LDH serum level B symptoms Pain Swelling Bulky disease Fracture
15% 34% 9% 82% 40% 23% 15%
38% 30% 24% 92% 45% 15% 25%
62% 65% 30% 90% 34% 32% 29%
Sites of involvement (%) Skull Spinal cord Pelvis Humerus Forearm Femur Forefoot Lymph nodes Cerebrospinal fluid Bone marrow Other
15% 17% 17% 7% 7% 20% 13% – – – 4%
32% 65% 32% 13% 16% 38% 19% – 3% – –
19% 51% 33% 17% 8% 24% 14% 28% 1% 35% –
DLBCL = diffuse large B-cell lymphoma; MB-DLBCL = multifocal bone diffuse large B-cell lymphoma; ECOG PS: Eastern Cooperative Oncology Group Performance Status; LDH lactate dehydrogenase.
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246
The standard contrast-enhanced computed tomography (CT) scan is the primary modality for staging, restaging, and follow-up of lymphoma patients. CT demonstrates the boundaries of any extraosseous extension as well as indicating cortical breakthrough by the tumour [31], and may detect osteolysis, osteosclerosis and fragments of bone sequestra [32]. Magnetic Resonance Imaging (MRI) reveals the extent of disease in more detail, particularly by identifying cortical changes such as linear channels of cortical destruction, as well as intratumoural fibrosis, replacement of trabecular bone and bone marrow by tumour [33,34]. Abnormal signal intensity areas are visible on both T1 and T2 weighted images with minimal contrast enhancement, hypointense in T1weighted and hyperintense in T2-weighted [35]. Accordingly to its relatively high cellularity [35], primary bone lymphomas exhibit restricted diffusion with low apparent diffusion co-efficiency value on diffusion-weighted imaging [36]. An increased post-treatment apparent diffusion co-efficiency value correlates with higher tumour necrosis [37] and is usually consistent with decreased 18 fluorodeoxyglycose (18FDG) uptake on 18FDG-PET scan [38], suggesting tumour response (see below). Functional imaging Bone lesions usually result in increased uptake on technecium scanning, while functional imaging (18FDG-PET) also reveals associated soft tissue involvement. 18FDG-PET display a higher specificity and sensitivity than conventional bone scintigraphy in identifying lymphomatous infiltration of skeleton [39]. 18FDG-PET-computed tomography scan (PET-CT) is a hybrid imaging technique that simultaneously provides functional and anatomical information, with a higher sensitivity and specificity than standard CT scan in lymphoma patients [40–44]. In staging of extranodal lymphomas, sensitivity, specificity and accuracy of PET-CT are 97%, 100% and 98% compared to 87%, 85% and 84% of CT [44]; diagnostic sensitivity is not correlated to lymphoma category [45]. Disagreements between PET-CT and CT usually regard spleen involvement or extranodal lesions, like bone marrow, bone, thyroid, and prostate [45]. In fact, diagnostic sensitivity of CT in lymphoma patients with bone lesions is 17%, and addition of this procedure to PET/CT does not further improve diagnostic sensitivity [46]. Conversely, the addition of 18 FDG-PET to CT scan results in upstaging in 42% of lymphoma patients, with a low rate of false positive scans [46]. Although it is supported only by small, retrospective studies, PET-CT is recommended by the recent Lugano Classification as a standard tool for initial evaluation, staging and response assessment of FDG-avid lymphomas, whereas CT plays a central role in non-avid lymphoma categories [47]. 99
Diagnosis A clinic-radiological suspicion of bone lymphoma must be confirmed by histopathological examination (morphological exam, immunohistochemistry and molecular analyses) of a diagnostic sample obtained by surgical procedure. This is mandatory to confirm lymphomatous nature and to define histotype since current knowledge and available radiologic tools do not allow a distinction among the different lymphoma categories involving the skeleton. Excision biopsy should be avoided; biopsies should be limited in size to reduce the risk of pathological fracture. Some bony sites such as the skull base can be particularly difficult to biopsy [48], with the risk of false negative results and delayed diagnosis [49]. In cases with stage-IIE disease, lymphadenectomy is advisable because it is associated with a lower risk of
237
orthopaedic sequelae and facilitates pathologist’s diagnostic performance. Interestingly, a role of serum soluble interleukin-2 receptor (sIL-2R) levels to achieve a non-surgical diagnosis of primary bone lymphomas was recently suggested [50,51]; normalized serum sIL-2R levels seems to be consistent with a therapeutic response in primary bone lymphomas [52]. These findings need for further investigation and confirmation in well-designed studies. Pathology, morphology and immunophenotype Histopathological diagnosis and categorization of primary bone lymphomas can be difficult due to some frequent problems like crush artefacts in biopsy specimens [53], the presence of reticulum or hyalinised fibrosis, inflammation and some specific diagnostic traps, including sarcoma-like spindle cells and carcinoma-like clustering [6]. Diffuse large B-cell lymphoma (DLBCL) is the most common histological subtype of lymphoma, either primarily or secondarily infiltrating the skeleton. It accounts for 70–80% of all bone lymphomas [1,5,29,30,54], with rare to anecdotal occurrences of follicular, marginal zone, lymphoplasmacytic, anaplastic large cell, NK/T-cell, Burkitt, and Hodgkin lymphomas [55,56]. Morphologically, tumour cells are large sized and consistent with follicle centre or centroblastic cell type, often with nuclear cleavage [4,57,58]; large multilobated cells are reported in around half the cases. Evidence of germinal centre (GC) derivation has been noted in at least 50% of cases [59]. Tumour cells are immunoreactive for B-cell markers: CD45, CD20, CD21, CD45, CD79a [53,60,61]; immunoreactivity for CD75 and CD10 is variable [62]. T-cell markers are usually negative, but small CD3+ cells are often present. BCL-2 and BCL-6 immunoreactivity has been reported in 35% and 69% of cases, respectively [62]. Monotypic IgG, IgH and HLA-DR have been noted [61]. Available data on primary bone T-cell lymphomas are sparse; primary bone T-cell lymphomas do not exhibit distinctive features than extra-osseous T-cell lymphomas. Most reported cases of primary bone T-cell lymphomas are anaplastic large-cell lymphoma (CD3 /+; CD43+/ ; CD30+), often associated with t(2;5)(p23;q35) and ALK-1 expression [62–64]. Cytogenetic and molecular abnormalities Cytogenetic and molecular studies of primary bone lymphomas are limited and usually relate to DLBCL. A monoclonal pattern assessed by PCR is detected in 54% of cases, IgH gene rearrangement in 72% [60] and BCL-2 translocation in 5% of cases [62]. Discrepancy between high BCL2 protein expression (55–70%) and low prevalence of BCL2/IgH rearrangements (4%) [65] indicates that other mechanisms (e.g., gene amplification) may be responsible for the BCL2 protein over-expression in primary bone DLBCL. Studies using fluorescence in situ hybridization has shown BCL2 and cMYC translocations respectively in 28% and 9% of primary bone DLBCL; none of these cases showed BCL6, PAX5, ALK, and CCND1 rearrangements [66]. These features seem to provide specific characteristics to primary bone DLBCL in comparison with other extranodal B-cell lymphomas. Most primary bone lymphomas are ‘‘de novo’’ DLBCL with a follicle-centre origin, which is suggested by frequent MUM1 positivity and less common BCL-6 mutations [67–69]. To date, there are no published studies evaluating gene expression profiles in primary bone DLBCL. Increased expressions of osteoclast-activating factors, such as MIP-1alpha, MIP-1beta and RANKL, in primary bone DLBCL suggest a potential causative role in osteolysis and hypercalcemia [33,70].
238
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246
Table 2 Staging procedures in patients with bone lymphoma.
Table 3 IELSG staging system for DLBCL of the bone.
Test/procedure Demographics and medical history* Physical examination Blood tests# Chest X-ray Contrasted CT scan of the neck, chest, abdomen, and pelvis MRI of bony lesions 18 FDG-PET Bone marrow biopsy
IELSG stage
Lymphoma extension
Ann Arbor stage
IE IIE
Single bony lesion Single bony lesion with involvement of regional lymph nodes Multifocal disease in a single bone or lesions in multiple bones in a disease exclusively limited to the skeleton (without lymph nodal or visceral disease) – called also ‘‘multifocal osteolymphoma’’ or ‘‘polyostotic lymphoma’’ Disseminated lymphoma with at least one bony lesion
IE IIE
IVE
In case of suspicion of involvement of particular organs Cerebrospinal fluid (CSF) examination§ Gadolinium-enhanced brain MRI§ Gastrointestinal tract endoscopy Blood smears
IV
* This includes family history, exposure to a toxic agent, prior malignancies, analysis of comorbidities. # This includes full blood count, alkaline phosphatase, lactate dehydrogenase, erythrocyte sedimentation rate, C-reactive protein, beta-2 microglobulin and protein electrophoresis. § In patients with high risk of CNS dissemination (see ‘‘CNS prophylaxis’’).
IV
IV
but improvement in sensitivity and specificity of radiologic and functional procedures led to increasing detection of multifocal disease. This suggests that multifocal disease, both within a single bone (monostotic disease) and in different bones (polyostotic disease), was underestimated in the past decades. Direct soft tissue infiltration is detected in about 20% of patients, but this is superfluous to staging, therapeutic and prognostic considerations (see Table 4).
Staging Diagnostic tests and procedures in the initial work-up of patients with primary bone lymphomas are listed in Table 2. Stage of disease is defined according to the Ann Arbor staging system [71]. However, this staging system has some important limitations when attempting to analyze different stage subgroups and to compare reported studies. Based on their different prognosis and outcome (Fig. 1), the IELSG staging system has been proposed (Table 3), where bone lymphomas are classified into four different stages: stage IE = single bony lesion; IIE = single bony lesion plus regional lymphadenopathy; IVE = polyostotic lymphoma; and IV = conventional stage-IV lymphoma with skeleton involvement. Most patients with bone DLBCL present with unifocal disease,
Prognosis The survival of patients with primary bone DLBCL is significantly related to disease stage, with the 5-year overall survival (OS) varying from 82% for patients with stage-IE disease to 38% for patients with disseminated DLBCL and skeleton involvement (Fig. 1). In early-stage disease, relapses occur equally within and outside the primary sites of disease, with local and systemic relapse rate of 10% and 17%, respectively [72]. In polyostotic lymphomas, most relapses exclusively involve the skeleton, with involvement of secondary sites in 21% of recurrences; the exclusive involvement of bones both at presentation and at relapse seem to
1,0
Probability OS
0,8
0,6
0,4
0,2
0,0
0
12
24
36
48
60
Months
72
84
96
108
Stage IE Stage IIE Stage IVE Stage IV
Fig. 1. Overall survival curves of bone DLBCL according to stage of disease. Stage IE: single bone lesions without regional lymphadenopathies; stage IIE: single bone lesions plus regional lymphadenopathies; stage IVE: polyostotic DLBCL (multifocal lymphoma involving exclusively the skeleton); stage IV: advanced-stage DLBCL with secondary involvement of the skeleton. Only patients managed with primary anthracycline-based chemotherapy followed or not by radiotherapy were considered. Similar results were obtained when the whole unselected subgroups were considered.
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246 Table 4 More frequent pathological fractures sites at presentation. Site of pathological fracture
Frequency at presentation (%)
Lower limb Spine Upper limbs Pelvis Skull
49 38 27 17 8
suggest a homing mechanism in polyostotic lymphoma [73,74], but molecular and biological explanations for this behaviour remain unknown. In these patients, half of the deaths are due to lymphoma [9]. Overall, prior to the IELSG-14 study [9,26,55,75], no consistent prognostic factors were noted. Several unconfirmed prognosticators have been recorded in undersized, retrospective studies with evident selection biases [1,3,5,29,59,76–79]. Conventional International Prognostic Index (IPI) plays a prognostic role in patients with advanced-stage DLBCL with skeleton involvement but not in polyostotic lymphoma and primary bone lymphomas [9,80]. This is due to the fact that stage and number of extranodal sites have no variability in primary bone lymphomas. Importantly, the IELSG14 study has revealed that age, performance status and serum LDH levels, the three remaining IPI variables, are independently associated with OS in every subgroup of bone lymphomas. The prognostic significance of the phenotypic and genetic characteristics of primary bone DLBCL has been studied in a few series. In fact, the germinal centre (GC) phenotype and related molecular features (CD10 expression, BCL-6 mutations, translocations involving 3q27) are associated with favourable outcome in primary bone DLBCL [30,62,65,81], whereas non-GC signature and related features (immunoblastic variant, MUM1 expression) are unfavourable predictors in DLBCL of the bone [82,83]. Prognosis is poorer in primary bone T-cell lymphomas than in primary bone B-cell lymphomas [30,84], in CD56-positive ALCL in particular [85].
Treatment of different clinical forms Stage IE-IIE DLBCL of the bone (or primary bone DLBCL) Different strategies like chemotherapy, immunotherapy, surgery and radiotherapy have been used to treat primary bone DLBCL, a separate entity usually displaying favourable clinical features and good prognosis. The role of surgery in primary bone lymphomas is limited to sampling for histopathological diagnosis, stabilization and internal fixation of affected bone and resolution of pathological fracture (see below). The use of surgery instead of radiotherapy has been suggested in patients with extensive destruction of weightbearing bones in order to reduce the risk of pathological fractures [86]. Excision of weight-bearing bones and amputation are no longer used to control local disease since they may result in severely impaired outcomes due to post-surgical complications, chemotherapy delay and long-term sequelae [87]. Two large cumulative studies of the Rare Cancer Network and the IELSG suggest anthracycline-based chemotherapy followed, when indicated, by involved-field radiotherapy as first-line treatment for patients with primary bone DLBCL [26,72]. With this strategy, the overall response rate (ORR) is over 90% and the 5-year OS is 84%. In line with other pre-rituximab studies (Table 5), the IELSG study demonstrates that chemo-radiotherapy produces significantly better results than the inverse sequence. CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone) is the firstchoice regimen in primary bone DLBCL; discrepancies mostly regard radiotherapy [88]. Variations in chemotherapy, like substitution of doxorubicin with mitoxanthrone or liposomal doxorubi-
239
cin to reduce cardiotoxicity and the use of infusional, thirdgeneration or 14-day regimens, should follow the same criteria used in nodal DLBCL. A survival benefit of the addition of the anti-CD20 monoclonal antibody rituximab to CHOP in primary bone DLBCL has not been demonstrated. Although a recent retrospective study suggests no survival benefit with the addition of rituximab in localized extranodal lymphomas [92], this antibody has essentially changed the natural behaviour of DLBCL, and, with a few exceptions, is an unavoidable part of first-line treatment for this lymphoma, both in nodal and extranodal forms. Moreover, a positive effect of rituximab in a low-risk malignancy like primary bone DLBCL is also suggested by favourable results reported in patients with low-risk DLBCL enrolled in the MINT trial [93]. In a recent, small series of primary bone DLBCL [89], R-CHOP was associated with a complete remission (CR) rate of 95%, and an 8-year OS of 95%. A few retrospective studies suggest a benefit with the addition of rituximab, with a 3-year PFS of 80–90% after R-CHOP and 50–60% after CHOP [29,90,91]. The survival benefit of the irradiation of affected bones after primary chemotherapy and the potential for complications are a matter of debate in patients with primary bone DLBCL. Prospective studies focused on this therapeutic issue in patients with primary bone DLBCL do not exist, thus, debate is mostly based on indirect evidence from large trials in nodal DLBCL. Before rituximab, a few randomized trials suggested that the addition of radiotherapy is superfluous in DLBCL patients who achieve a CR after primary chemotherapy [94,95]. In the rituximab era, the addition of consolidation radiotherapy significantly improved outcome in a large retrospective series of 469 DLBCL patients treated with R-CHOP combination [96]. Consolidative irradiation of bulky residual masses after R-CHOP14 has been associated with an improved outcome in a prospective series of 166 elderly patients with DLBCL [80]. However, these studies included all stages of disease, both nodal and extranodal DLBCL, and, sometimes, the irradiation volume was not prospectively defined nor described, resulting in important interpretation biases. Regarding primary bone DLBCL, a retrospective study of 161 patients treated with CHOP or derivatives suggests that the addition of consolidative post-chemotherapy radiotherapy does not further improve outcome, with both chemo-radiotherapy and chemotherapy alone producing a 5-year PFS of 74% and 67%, respectively [26]. However, selection biases related to the small size of analyzed subgroups and the management of primary bone DLBCL with more favorable features with chemotherapy alone cannot be excluded. It is clear that the use of intensive immunochemotherapy without consolidation radiotherapy requires formal testing and validation in a randomized trial before it can be used as an alternative treatment for earlystage DLBCL. In particular, the role of PET-driven consolidative radiotherapy remains an appealing open question in the management of limited-stage DLBCL. The choice of radiation volume in primary bone lymphomas patients should result from an accurate risk–benefit analysis. This should consider the risk of exposure of sensitive organs, such as lung, brain, bowel or kidney, among others, and of late bone effects. Although most radiotherapy experts favor whole-bone irradiation, with a dose of 38–40 Gy [88], evidence supporting irradiation of the whole bone and/or regional lymph nodes is limited. In line with previous series [72], the IELSG-14 study has showed a 5-year PFS of 76% and 64% respectively for primary bone DLBCL patients managed with CHOP followed by irradiation of the whole bone or of a part of the affected bone (partial-bone irradiation) [26]. However, the clinical relevance of the irradiation of a whole long bone (a.e., femur, humerus) or a whole flat bone (a.e., vertebra) is very different. In cases where the irradiation of the whole bone appears risky, a smaller volume that maintains wide margins (3–5 cm) around
240
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246
pre-chemotherapy tumour borders within the affected bone is advisable. Margins around the soft tissue or extra-osseous borders can be further restricted to 1–2 cm around post-chemotherapy volumes as the distinction between normal and abnormal tissue is not so problematic, and some individualisation of dose and volume will always be required. Radiation dose depends on the size of the irradiated volume, the anatomical area and response to primary chemotherapy. Data suggesting better results with a dose >40 Gy are outdated [97]. In the IELSG-14 study, no significant survival difference between 47 patients irradiated with a dose 636 Gy and 58 patients irradiated with >36 Gy have been shown, with a 5-year PFS of 72% and 75%, respectively [26]. This is in line with a large randomized trial that has demonstrated that, compared with previous standard doses of 40–45 Gy, a radiation dose of 30 Gy is not associated with loss of efficacy in aggressive lymphomas, with no significant differences in in-field relapse rate, PFS and OS [98]. Polyostotic lymphoma Polyostotic lymphoma is characterized by exclusive multifocal involvement of the skeleton, without affecting lymph nodes or other organs. Although patients with polyostotic DLBCL and patients with disseminated DLBCL and skeletal involvement display similar clinical presentation [9], prognosis is significantly better in patients with polyostotic DLBCL, with an ORR of 92% and 65% (p = 0.002), a 5-year PFS of 57% and 35% (p = 0.01) and 5-year OS of 75% and 37% (p = 0.008), respectively [9]. These differences are independent from the IPI score and other variables, and suggest that polyostotic lymphoma may be a different entity, that deserves to be investigated from biological and molecular points of view. An intriguing finding is that the use of post-chemotherapy radiotherapy in patients with polyostotic DLBCL has been associated with a significantly better outcome, with a 5-year OS of 83% for patients managed with chemoradiotherapy and 55% for patients managed with chemotherapy alone [9]. Accordingly, patients with polyostotic DLBCL should be treated with the same strategy used for disseminated nodal DLBCL, considering consolidation radiotherapy in patients with lesions located in adjacent bones that can be irradiated with acceptable side effects. Advanced-stage DLBCL with skeletal involvement Secondary bone involvement by systemic lymphoma is more common than primary bone lymphomas [53,99,100]. Bone involvement is frequently observed in cases of DLBCL with high tumour burdens and disseminated disease, mostly occurring in patients with an intermediate-high IPI. The prognostic value of bone involvement in DLBCL remains controversial [77,80,101,102]; related articles include heterogeneous series managed with varied strategies, and the prognostic value of skeleton infiltration may has changed after rituximab wide use [80]. In the pre-rituximab era, patients with stage-IV DLBCL and skeletal involvement managed with anthracycline-based chemotherapy, followed or not by bone irradiation, achieved a 65% ORR and a 5-year PFS of 34% [9]. In the rituximab era, R-CHOP treatment, with or without radiotherapy, is associated with a 65% CR rate, and a 5-year PFS of 54% [101,103]. Nevertheless, the clinical benefit of the addition of rituximab to CHOP chemotherapy in patients with stage-IV DLBCL and skeletal involvement remains matter of debate. A retrospective analysis of 292 DLBCL patients with skeletal involvement registered in nine consecutive prospective trials did not demonstrate an advantage with the addition of rituximab [80]. However, one-fifth of patients had primary bone lymphoma and not disseminated disease, and 80% of analyzed patients had been treated without rituximab, which may result in unbalanced comparisons and low-level
evidence. Thus, chemoimmunotherapy of bone DLBCL should be the same routinely used for nodal DLBCL, that is a combination of CHOP or CHOP-like regimen plus rituximab. The clinical benefit of irradiation of involved bone in patients with disseminated DLBCL is an important open question. A beneficial effect of this strategy was suggested by a recent retrospective study of the German High-Grade Non-Hodgkin lymphoma Study Group [80]. The analysis of the effect of radiotherapy was restricted to patients with disease responsive to anthracycline-based chemotherapy ± rituximab, with a significantly and independently better outcome among patients who received radiotherapy to sites of skeletal involvement. Contrasting results in smaller series should be considered with caution because it is possible that patients with more aggressive disease were more likely to receive irradiation [29]. The discretionary nature of the indications for bone irradiation and heterogeneity of bone involvement (bulky or not, single or multiple) preclude firm conclusions concerning the usefulness of radiotherapy in patients with advanced-stage DLBCL and skeleton involvement. Importantly, a recent study suggested a benefit for the rational use of PET-guided radiotherapy in patients with advanced DLBCL and residual abnormalities at CT scan after RCHOP [104]. In fact, 18FDG-PET is an important tool to define therapeutic response (see ‘‘Response to treatment’’), and, in that study [104], irradiation of PET-positive residual disease resulted in survival figures similar to those reported for patients with negative post-R-CHOP PET. This strategy may avoid superfluous radiation exposure of PET negative patients. However, 18FDG-PET exhibits some technical and interpretative limitations (see below), and whether this procedure can identify those primary bone lymphoma patients who can be spared from radiotherapy remains to be shown in appropriately designed studies. Indolent bone lymphomas Indolent primary bone lymphomas represents up to 8% of all bone lymphomas (Table 6), and these anecdotal cases are not specifically described in histopathological series [105,106]. The largest series of indolent bone lymphomas was recently reported by the IELSG [55]. Out of an international series of 499 patients with a diagnosis of NHL and skeleton involvement, 26 (5%) patients had an indolent bone lymphomas; ten of them had a small lymphocytic lymphoma, 10 had a follicular lymphoma and 6 had a lymphoplasmacytic lymphoma. Eleven patients had limited stage and 15 had advanced disease. No patients with Richter’s syndrome were identified [107]. With all the limitations of a pre-PET retrospective study, a significantly better outcome in patients with small lymphocytic lymphoma has been recorded; these patients frequently achieved long-term remission and rarely died of lymphoma. The prognosis of patients with limited-stage lymphoplasmacytic lymphoma or follicular lymphoma was less favourable, with a higher proportion of tumour dissemination, possibly reflecting limitations in staging sensitivity. Patients with primary indolent bone lymphomas in the IELSG series had been managed with varied first-line treatments, obtaining an ORR of 73%, and a 5-year PFS and OS of 37% and 46%, respectively. Performance status and stage of disease were independently associated with OS. All patients with limitedstage indolent bone lymphomas managed with radiotherapy achieved a CR, and 63% of them remained relapse-free; conversely, only 33% of patients managed without radiotherapy achieved a CR, and all of them experienced relapse [55]. The addition of chemotherapy was not associated with improved OS in irradiated patients with limited-stage indolent bone lymphomas. Patients with advanced-stage indolent bone lymphomas showed response and survival rates similar to those reported in patients with the same stage and histological type but without skeleton involvement. Accordingly, patients with localized indolent bone lymphomas
241
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246 Table 5 Largest available series of bone lymphomas. Reference
Median age
Years of enrollment
ORR (%)
DFS
OS (years)
Ostrowski 261 [6] Heyning 60 [81] Dosoretz 30 [57] De Camargo 24 [125] Ueda [126] 34
45
1907–1982 Femur (21%)
0
NA
NR
58
0
56
46% (5)
53% (10 – uni) 35% (10 – mul) 61% (5)
48
76
12
0
NR
40% (5)
63% (5)
37
8
43
0
NR
NR
70% (5)
44
26
10
41
0
NR
NR
80 100
81 100
30 0
14 32
56 68
6 0
NR 100
81% (5) 48% (10)
75% (5 – st I) 50% (5 – st II) 88% (5) 53% (10)
Beal [1] Marshall [127] Rathmell [5] Dubey [128] Fidias [129] Fairbanks [3] Horsman [27] Bacci [82] Baar [116] Christie [124] Stein [130]
1967–1988 NR
85
100
0
56
33
0
NR
39% (10)
40% (10)
52
1967–1992 Femur (20%)
98
100
9
11
80
0
NR
63% (10)
60% (10)
37 63
41 63
100 100
0 3
0 79
100 16
0 0
100 NR
55
100
16
41
38
0
57
73% (10) 90% (5 – CT + RT) 57% (5 – RT) NR
87% (10) NR
37
1970–1995 Appendix (75%) 100 1970–1989 Long bone 93 (57%) 1970–2003 Pelvis (24%) 73
26 17 70
NR 36 60
1972–1982 Fenur (23%) 1975–1992 Femur (29%) 1973–1999 Spine (29%)
80 100 65
100 88 80
0 30 0
0 6 44
100 64 56
0 0 0
100 94 83
88% (13) 77% (3) NR
88% (13) 77% (3) 59% (5)
19
54
1979–2000 NR
95
58
42
0
58
0
95
NR
Messina [9] 37 Govi [55] 26 161 Bruno Ventre [26] Zinzani [59] 52 Gianelli 28 [62] Barbieri 77 [76] 131 Ramadan [29] Lewis [131] 28 Ford [86] 22
53 60 55
1980–2005 Spine (65%) 1980–2005 Pelvis (47%) 1980–2005 Femur (20%)
100 0 100
0 42 100
35 30 8
0 15 14
65 55 78
0 0 0
92 73 91
90% 87% 56% 25% 68%
58 51
1982–1998 Femur (27%) 1982–1999 Femur (24%)
85 93
79 100
16 20
21 3
63 74
0 0
90 NA
84% (8 – CR pts) 75% (4)
68% (9) 78% (4)
42
1983–2001 Extremities (51%) 1983–2005 Spine (29%)
97
100
0
13
87
0
95
76% (15)
88% (15)
79
46
44
8
48
21
NR
40% (10)
41% (10)
89 91
71 77
36 18
14 0
50 82
0 0
NA NR
46% (6) 85% (10)
60% (6) 74% (10)
Bayrakci [87] Cai [72] Catlett [90]
20
48
1984–1994 Femur (39%) 1985–2003 Long bone (50%) 1986–1997 Femur (24%)
NR
70
35
0
65
0
65
NR
116 30
51 49
86 90
100 70
13 16
12 10
75 71
3 40
91 NR
62% (4) NR
78% (st I) 16% (st IV) 72% (10) 73% (5)
20 33 28
44 40 47
100 100 68
90 39 32
15 48 50
15 0 0
65 52 50
0 39 0
NA 88 89
NR NR 77% (3)
74% (5) 75% (4) 84% (3)
21
34
100
10
48
0
52
100
95
100% (8 – CR pts)
95% (8)
53 31
52 55
1999–2009 Long bone (38%) 2000–2007 Femur (24%) 2000–2007 Femur (26%)
90 97
77 68
12 0
21 0
62 100
37 19
92 96
83% (4) 64% (5)
100% (4) 90% (5)
28
41
2001–2009 Femur (25%)
100
78
30
3
66
0
NR
62% (2)
NR
de Leval[65] Kim [132] Maruyama [30] Pellegrini [89] Alencar[91] Christie [133] Nasiri [134]
No. of patients
Stage I–II (%)
CT (%)
RT (%)
77
68
6
63
22
1943–1996 Femur (24%)
92
62
NR
8
58
1950–1978 Femur (30%)
93
100
0
38
1955–1999 Spine (25%)
83
NA
56
1961–1988 Pelvis (29%)
53
82 28
48 52
1963–2003 Femur (27%) 1962–1997 Femur (18%)
27
53
45
63 45 50
Most common primary site
1987–2008 Spine (28%) 1989–2005 Long bone (57%) 1990–2000 Femur (35%) 1992–2010 Pelvis (39%) 1995–2004 Pelvis (41%)
Aggressive type (%)
assessed with complete staging procedures are candidate for radiotherapy alone [55,108], while patients with advanced-stage indolent bone lymphomas should be managed like other patients with disseminated indolent lymphomas.
CT + RT (%)
Rituximab (%)
(6 – st I–II) (6 – st IV) (5) (10) (5)
50% (10)
74% (5) 29% (10) 75% (5)
categories, being the most common forms [84]. Skeletal involvement does not seem to play a prognostic role in these patients. Currently, there is no evidence suggesting that these patients should be treated differently to patients with the same lymphoma category and stage of disease but without skeletal involvement.
Uncommon aggressive bone lymphomas Special therapeutic issues Some anecdotal cases of aggressive lymphomas other than DLBCL primarily arising in the bones have been anecdotally reported (Table 6) [82,109]; lymphoblastic and Burkitt lymphoma, among B-cell categories, and anaplastic large cell lymphoma and peripheral T-cell lymphoma not otherwise specified, among T-cell
Response to treatment Conventional radiologic and functional procedures exhibit limitations in the assessment of response to treatment in primary
242
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246
Table 6 Pathological classification with relative frequencies in the largest series of PBL. Histology Diffuse large B-cell Diffuse mixed Follicular Peripheral T-cell Extra-nodal marginal zone Mantle cell Small lymphocytic Transformed MALT Burkitt’s/Burkitt like Anaplastic large T-cell Lymphoma NOS Plasmocitoid Immunoblastic Lymphoblastic Unclassifiable low grade Other
Zinzani [59] 44 (84%) – 2 (4%) – – – 2 (4%) – 2 (4%) 2 (4%) – – – – – –
Beal [1] 66 (80%) 4 (5%) 3 (4%) – – – 3 (4%) – 1 (1%) – 2 (2%) 1 (1%) 1 (1%) 1 (1%) – –
Ramadan [29] 103 (79%) – 7 (5%) 2 (1.5%) 4 (3%) 1 (1%) 2 (1.5%) – 2 (1.5%) 4 (3%) – – – 2 (1.5%) 1 (1%) –
Alencar [91] 44 (83%) – 3 (5.7%) 2 (3.8%) 1 (1.9%) 1 (1.9%) 1 (1.9%) 1 (1.9%) – – – – – – – –
Cai [72] 91 (78%) – 7 (6%) – – – – – – 6 (5%) – – – – – 12 (11%)
Fig. 2. CT scan (left column) and PET (right column) features in a patient with a stage-IE DLBCL of the left femur. CT scan (A) and PET (B) performed at diagnosis show evident osseous abnormalities and increased FDG uptake (arrows), respectively. CT scan (C) and PET (D) performed immediately after treatment conclusion show persistence of osseous abnormalities and reduced FDG uptake (arrows), respectively. CT scan (E) and PET (F) performed at one year of follow-up show respectively persistence of evident osseous abnormalities (arrow), which may be interpreted as residual disease (false positive), and remission of FDG uptake (arrow).
bone lymphomas patients; architectural osseous distortion can remain unchanged for many years in X-rays and CT scans, and residual uptake can persist in bone scintigraphy and 18FDG-PET (Fig. 2). Bone remodelling and artefacts due to stabilizing hardware complicate the assessment of local disease control; in particular, artefacts caused by intramedullary titanium rods interfere with interpretation of changes and limit the usefulness of post-treatment MRI and CT scans. These persisting changes prevent the designation of CR using international criteria [99]. In this case, the
designation of response is usually made upon resolution of any symptoms, particularly pain, signs such as swelling, resolution of soft tissue infiltration at CT, MRI and/or PET, evidence of bone healing on X-ray, and absence of signs of progressive disease. Replacement of intraosseous tumour and adjacent bone marrow with fatty marrow and sclerosis in follow-up MRI scans can help to confirm CR. 18FDG-PET is more useful than bone scanning in the assessment of response to treatment [110,111], and may be used to drive consolidative radiotherapy [104]. However, although 18FDG-PET
C. Messina et al. / Cancer Treatment Reviews 41 (2015) 235–246
has a high sensitivity for the detection of osseous disease [39], its predictive value is limited in primary bone lymphomas due to false-positive findings related to post-chemotherapy bone remodeling, inflammation, marrow hyperplasia and bone necrosis and false negatives related to the resolution of the equipment, the technique used and the variability of FDG avidity among lymphoma categories [43,112–114]. As a logical consequence of these limitations, response rates in primary bone lymphomas trials should be considered with caution. In routine practice, the current plan in the absence of definitive evidence of local or distant failure is to continue surveillance with CT scans [115], whereas the role of the PET scans as a follow-up procedure remains to be defined [99]. Management of pathological fractures Pathological fractures occur in 10–15% of patients with DLBCL of the bone, with the humerus being the most frequently affected bone (66%) [10,29,91]. The analysis of the IELSG-14 study has clearly showed that the presence of pathological fracture at presentation is associated with independently worse outcome in patients with primary bone DLBCL, with a 5-year OS of 54% for patients with pathological fracture and 68% for patients without this complication [75]. Unfortunately, this study was not able to identify pre- and post-treatment variables predicting the occurrence of a pathological fracture (i.e., lysis of the cortex, tumour size, soft tissue infiltration). The goals of initial surgical stabilization of the pathological fracture are usually to preventing further bone destruction and displacement, to enable weight-bearing, to assist pain relief or healing and to obtain a better quality of life. While these are all important goals, there is no evidence of an improved cancerrelated outcome with initial surgical stabilization of the pathological fracture, with a 5-year OS of 45% and 54%, respectively for operated and not operated patients [75]. These data suggest that any initial surgical stabilization should be kept to a minimum, and used to improve patient’s quality of life and prevent bone disintegration only if chemotherapy delays can be avoided. A pathological fracture or a high risk of fracture may lead the treating physician to start the treatment program with radiotherapy. Importantly, irradiation of a fractured bone before chemotherapy does not improve disease control or survival, when compared to the chemo-radiotherapy sequence [75]. These results should be taken into account with caution since a selection bias related to initial irradiation of patients with more destructive lesions, and, thus, with more aggressive lymphoma, cannot be excluded. Nevertheless, patients with pathological fracture should be managed like nodal DLBCL, including anthracycline-based chemotherapy followed by consolidation radiotherapy to the fractured bone with 30–40 Gy [75,76,116]. Radiation volumes and doses should follow the above-discussed recommendations for primary bone DLBCL patients. CNS prophylaxis CNS dissemination is an early and fatal event reported in 5% of DLBCL [117,118]. The risk of CNS recurrence associated with skeletal involvement is a matter of debate, with rates of 4% and 0.6%, respectively for DLBCL patients with and without skeletal involvement [80,119]. In the IELSG-14 study, CNS involvement, usually meningeal lymphomatosis, occurred in 2.5% of patients with primary bone DLBCL [26], 5% of patients with polyostotic DLBCL [9,120] and 2% of patients with stage-IV DLBCL and skeleton lesions [9,75]. Probably, these rates would be even lower in the rituximab era, since CNS dissemination has been reduced in DLBCL patients treated with this antibody [121]. Importantly, the small number of events hampers the identification of reliable predictors
243
of CNS recurrence in primary bone lymphomas studies, but available evidence suggests that CNS prophylaxis is superfluous in primary bone DLBCL. Noteworthy, an accurate CSF assessment, brain MRI and personalized prophylaxis are advised in patients with disseminated DLBCL and involvement of bones close to the CNS (skull and/or spine). CNS recurrence occurs in 7% of these patients, especially in patients with other high-risk features [9,75]. Late iatrogenic sequelae Chronic pain, limb dysfunction and late pathological fracture are the most common iatrogenic sequelae in patients with bone lymphoma. The involvement of the femur, pelvis and the spine is more commonly associated with chronic symptoms, and a second pathological fracture after anti-lymphoma treatment has been recorded in 10% of patients with secondary bone DLBCL [75]. Such fractures can occur in the absence of local recurrence and can lead to persisting non-union with subsequent disability, mostly in bones of the lower limb. Contributing factors may include architectural disturbance due to previous tumour, pathological fracture before treatment and the existence of other medical conditions such as postsurgical osteomyelitis, Paget’s disease and osteoporosis, particularly among elderly women. Both radiotherapy and chemotherapy have been implicated as incurring a higher risk of subsequent pathological fracture; in particular, this risk has been attributed to corticosteroids and large radiation fields, fractions and doses [122,123]. These and other risk factors, such as involvement of weight-bearing bones, local recurrence at the fracture site and the size of the biopsy defect should be accurately considered. Orthopaedic fixation of fractures, limitation of the use of corticosteroids, reduction of radiation doses to
![[Treatment of primary bone lymphomas].](/assets/img/hold.png)
