Cytogenetics of extramedullary manifestations in multiple myeloma


Correspondence: Georgia Schilling, Department of Oncology, Haematology and BMT with Section Pneumology, Hubertus Wald Tumorzentrum - University Cancer Centre Hamburg (UCCH), University Medical Centre Hamburg-Eppendorf, 52 Martinistraße, Hamburg 20246, Germany.



Extramedullary disease in patients with multiple myeloma is a rare event, occurring mostly in advanced disease or relapse. Outcome is poor and prognostic factors predicting the development of extramedullary disease have not been defined. We investigated cytogenetic alterations of myeloma cells in different extramedullary manifestations by adapting the fluorescence in situ hybridization (FISH) technique in combination with cytoplasmic immunoglobulin staining to study the cytogenetics of plasma cell tumours on paraffin embedded material. Thirty six patients were investigated: 19 with extramedullary disease, 11 with skeletal extramedullary disease and six with solitary extramedullary plasmacytoma. The first two groups showed the following results: del(17p13) 32% vs. 27%, del(13q14) 35% vs. 27%, MYC-overrepresentation 28% vs. 18% and t(4;14) 37% vs. 18%. We detected an overall higher incidence of del(17p13) in both groups compared to data from bone marrow samples of multiple myeloma reported to date (range 7–16%). The solitary extramedullary plasmacytomas presented overall less cytogenetic aberrations than the other groups. Most important, three patients with extramedullary disease and one with skeletal extramedullary disease presented different FISH findings in the extramedullary tumour compared to their bone marrow plasma cells. del(17p13), occurring additional in three of four cases, seems a strong marker for extramedullary progression of myeloma.

Multiple myeloma (MM) is a haematological malignancy with clonal proliferation of plasma cells in the bone marrow (BM). Extramedullary disease (EMD) is a rare event in MM patients, mostly occurring in patients with advanced disease or during relapse.

Definition of EMD is controversial as some authors consider extramedullary organ manifestations arising from haematogenous spread as well as soft tissue infiltration originating from a bone lesion. The latter manifestation, skeletal EMD, is common in MM and the most frequent mechanism of myeloma spread (Varettoni et al, 2010; Bladé et al, 2011). Following recent studies, EMD in our series includes only plasma cell tumours originating from haematogenous dissemination and comprises various locations, such as lymph nodes, subcutaneous tissue, central nervous system (CNS), skin and other organs (Bladé et al, 2011; Detweiler Short et al, 2011). Masses arising from bone lesions, growing per continuitatem, are classified as skeletal EMD. In contrast, primary extramedullary plasmacytomas (EMP) are defined as isolated extraosseous plasma cell tumours without BM infiltration, following an indolent clinical course and only sporadic conversion to MM (Galieni et al, 2000).

Overall, EMD has been reported to be associated with a poor prognosis, with high mortality and a short overall survival time (Terpos et al, 2005; Varettoni et al, 2010; Detweiler Short et al, 2011; Rasche et al, 2012). A recent study, analysing patients with intracranial involvement, reported a shorter overall survival of MM patients with CNS involvement due to haematological spread compared to those patients with intracerebral tumour masses resulting from contiguous bone lesions (Gozzetti et al, 2012).

Organ infiltration by plasma cells at initial diagnosis of MM is very uncommon. In a large study of 1003 myeloma patients, extramedullary manifestations (EM) were observed in 7%, 85% were of skeletal type whereas only 15% were of haematogenous type (Varettoni et al, 2010).

Cytogenetic alterations are considered to have an important prognostic value in MM, helping to identify high-risk patients. Using fluorescence in situ hybridization (FISH), chromosomal changes have been observed in BM plasma cells in about 90% of MM patients at initial diagnosis (Avet-Loiseau et al, 2007). del(17p13) and t(4;14)(p16:q32) have been identified as independent negative prognostic markers for survival of MM patients (Avet-Loiseau et al, 2007).

Although novel effective agents are available for myeloma treatment today, patients with EMD often do not respond to therapy, indicating biological differences between intramedullary and extramedullary disease. Moreover, a rising incidence of EMD in the era of novel-agent based therapy (with thalidomide, lenalidomide, or bortezomib) has been discussed (Varettoni et al, 2010; Detweiler Short et al, 2011). However, our knowledge about the biology of extramedullary myeloma is limited. Only a few studies have addressed the cytogenetic make-up of EMD (Chang et al, 2004; Katodritoua et al, 2009; López-Anglada et al, 2010; Billecke et al, 2012). In these studies del(17p13) and del(13q14) were suggested as markers for progression to extramedullary disease (Minnema et al, 2008; Katodritoua et al, 2009; Billecke et al, 2012). Regarding EMP, a recent study did not find any significant cytogenetic differences between EMP and MM (Bink et al, 2008).

The aim of our analysis was to investigate cytogenetic aberrations in myeloma cells of EMD, skeletal EMD and EMP patients, assuming that EMD represents the most aggressive course of disease. Therefore we suspected differences in the incidence of poor prognostic genetic alterations between the three myeloma manifestations.

We used the FISH method in combination with immunofluorescence for cytoplasmic kappa and lambda light chains (cIg-FISH) to study paraffin-embedded EMD tissue. To the best of our knowledge, this is the first systematic cytogenetic cIg-FISH study that investigated chromosomal aberrations directly in the plasma cells of extramedullary manifestations of MM patients, not in BM cells as previously described (Chang et al, 2004; Rasche et al, 2012).

Materials and methods


Thirty six cases were collected retrospectively from the archives of participating institutes and haematological departments. Baseline data were retrieved from databases and medical records. Biopsies of extramedullary tumours were sent as paraffin-embedded tissue samples to our institution.

Following our definition of EMD we subdivided the cases into two groups: The first group comprised 19 patients with true EMD, namely MM patients with extramedullary plasma cell tumours resulting from haematogenous spread. Sites involved were pleura, pleural effusion, subcutaneous tissue, lymph nodes, skin, liver, CNS, ovary, muscles, larynx, lung and omentum majus (Table 1). The second group consisted of 11 cases, classified as skeletal EMD, with plasma cell tumours arising from bone disease. Affected sites were infiltrated soft tissue surrounding an osteolysis and various bones (Table 2). Histopathological and immunohistochemistry findings confirmed myeloma in all cases. In addition, we studied a third group of six patients with EMP, occurring as isolated plasma cell tumours without signs for systemic disease dissemination, all but one typically derived from the upper aerodigestive tract (UAD) (Table 3).

Table 1. Clinical and FISH data of MM patients with EMD
CasesSiteSexAge (years)TypeBMStagea del(17p) (%) del(13q) (%) MYC (%) t(4;14) (%)
  1. The percentage of plasma cells that show the indicated aberration are reported. Positive scored results are highlighted in bold.

  2. F, female; M, male; BM, bone marrow; NA, not applicable; ND, not determined.

  3. a

    Stage according to Durie and Salmon (1975).

Occurrence at relapse
1PleuraF68IgG lambdat(11;14)III60 99 0
2Subcutaneous tissueM80IgA kappamycIII 79 4 80 17
3Lymph nodeM68IgG kappaNormalIII07100
4SkinM45IgG kappaNormalIII11 83 0
5Pleural effusionF67IgG kappadel(13q)III 100 90 22
6Pleural effusionM55IgA kappadel(13q), others NDIII6NANA 84
7SkinM63IgD lambdaNDIII3911 36
8LiverM58IgG lambdaNormalIII4053
9CNSNANAIgG kappaNormalIII04 100 5
10LiverM58IgG kappadel(13q), del(17p13)III 46 83 00
11OvaryF41IgG lambdaNormalIII 34 53 105
12M. iliopsoasM56IgA kappamycIII44 92 0
13M. psoasF55IgD lambdaNAPCL 68 92 2 56
14LarynxM74IgG kappaNAIII10521
Occurrence at initial diagnosis
15Subcutaneous tissueM72NANANA0 86 44
16LiverM56IgL kappadel(13q), del(17p13)III 85 NA20
17LungF58IgM lambdaNAIII453 64
18Omentum majusM81IgM kappaNAIII0 27 0 15
Time of occurrence unknown
19Thyroid glandM74IgL kappaNANA6410 50
Table 2. Clinical and FISH data of MM patients with skeletal EMD
  1. For abbreviations see Table 1.

Occurrence at relapse
20SpineM51IgL kappanormalIII0%0% 95% 0%
21SpineM70IgG kappaNDIII 96% 93% 1%0%
22SpineM59IgA kappaNDIII2%1%2%0%
23Chest wallM73IgG kappaNAIII 90% 94% 2% 18%
24OrbitaF72IgG kappat(4;14)III10%3%7% 78%
25Soft tissueM71NANAIII 38% 0%11%2%
Occurrence at initial diagnosis
27Soft tissueF70IgL kappanormalIII0% 96% 3%0%
28BoneF65IgA kappahyperdiploidIII0%0%0%0%
29BoneM80NANDI2%3% 88% 8%
30Chest wallF78IgL lambdanormalII0%7%3%2%
Table 3. Clinical and FISH data of patients with EMP
CasesSiteSexAge (years)Type del(17p) (%) del(13q) (%) MYC (%) t(4;14) (%)
  1. The percentage of plasma cells that show the indicated aberration are reported. Positive scored results are highlighted in bold.

  2. F, female; M, male; BM, bone marrow; NA, not applicable.

31UterusF44NA000 19
33Sphenoidal sinusM23NA0000
34Maxillary sinusF65IgG lambda0000
35Paranasal sinusM82IgG kappa0 93 00
36EpipharynxM40Lambda 30 000

FISH data of the BM samples were collected from participating institutes when available and dated back to the time of initial diagnosis in the majority of the cases. Baseline cytogenetics were available in 17 of 30 cases with EMD and skeletal EMD.

All patients' specimens were gathered in the context of routine diagnostics and only surplus material was analysed for the above-mentioned study. The patients gave written agreement for collecting and archiving their surplus material for further experimental studies in the participating institutes.

cIg-FISH analyses on paraffin-embedded sections

For cytogenetic investigation of the paraffin-embedded samples, we combined FISH with stains for intracytoplasmic light chains (κ or λ) by fluorescence-labelled antibodies (cIg-FISH). For melting the paraffin wax, slides were placed on a hotplate set at 60–70°C for 3 min. Slides were dewaxed by washing three times for 10 min each using xylol in a lidded coplin jar. The slides were then rehydrated in a series of ethanol (100%, 90%, and 70%). Pretreatment was performed by incubating in citrate buffer for 35 min at 80°C. After cooling to room temperature for 15–20 min, the slides were washed in 2× standard saline citrate (SSC) twice for 5 min and then digested for 10–16 min (depending on thickness of the sample and material) in pepsin solution at 37°C. From this point, cIg-FISH was performed according to standard procedures (Kröger et al, 2004). At least 100 nuclei of 7-Amino-4-Methylcoumarin-3-Acetic Acid (AMCA)-positive plasma cells were scored.

Panel of probes for cIg-FISH

To screen the chromosomal regions 8q24 and 17p13, we applied the spectrum orange probes LSI C-MYC and LSI TP53 and the dual colour translocation probe t(4;14) FGFR3/WHSC1 (MMSET);IGH for the detection of t(4;14) (all Abbott Diagnostics, Chicago, IL, USA) as recommended by the manufacturer's instructions. A non-commercial probe (BAC 58C16; RCPI-11) mapping to 13q14 was applied to test 13q deletions. A probe hybridizing to centromere region of chromosome 7 (p7t1) served as control for censor deletions of 13q14 and 17p13 respectively, caused by a deficient hybridization or not well-conserved nuclei. The probe hybridizing to 8q24 was used to detect MYC-gains, but not rearrangements of this gene.

Statistical analysis

Loglinear models were used to test the independence of the given aberrations. We followed the recommendations of Agresti (2002) to build up models by adding higher order interaction terms, if they significantly improved model fit using a Likelihood-Ratio test for hierarchical loglinear models. All analyses were done using R.2.13.2, package mass (Venables & Ripley, 2002). P values < 0·05 were considered statistically significant.


True EMD

Nineteen patients (Table 1) were included in this group. Overall, IgG-subtype and kappa-light chain expression corresponded to that usually seen in unselected MM-series. Median age was 60·5 years (range 41–81), 13 were male, five female, and one unknown. Loss of 17p13 was seen in six of 19 patients (32%) and was strongly associated with del(13q14) (4/6, = 0·012). del(13q14) was found in six of 17 patients (35%). MYC-overrepresentation was seen in 5/18 cases (28%) and never occurred together with del(13q14) (= 0·012). Translocation t(4;14) was observed in seven of 19 cases (37%) and occurred as a single abnormality in four cases. Concomitant presence of t(4;14) and del(13q14) was seen in two cases and with del(17p13) in one case.

Results of cytogenetic BM analyses were available in 12 cases. We found differences between BM and extramedullary tumours in three cases (Patients 2, 5 and 11, Table 1) with additional aberrations in EMD cells. MYC-analysis on BM was available only for Patients two and 12 and results corresponded to those found in the extramedullary tumours.

Fourteen out of 19 patients (74%) had developed EMD during disease progression or relapse. Nine patients relapsed with EMD after peripheral blood stem cell transplantation (PBSCT) and five patients during conventional chemotherapy. Median time from initial diagnosis to the occurrence of EMD was 23 months (range 8–101 months). Seven of 14 patients received a novel agent-based therapy with bortezomib, lenalidomide and/or thalidomide. All patients showed progressive disease and median time from EMD presentation to death was 5 months (range 1–22) (Table 4).

Table 4. Treatment modalities and disease history of patients with EMD
CaseNumber of previous regimensTherapy regimens before EMD (novel-agents in bold)Time initial diagnosis to EMD (months)OutcomeTime from EMD to death (months)
  1. EVC, Epirubicin, Etoposide, Cyclophosphamide; IEV, Epirubicin, Etoposide, Ifosfamide; MP(T), Melphalan, Prednisone, (Thalidomide); Rd, Lenalidomide, low dose Dexamethasone; PBSCT, peripheral blood stem cell transplantation; auto, autologous; allo, allogeneic; (T)CED, (Thalidomide), Cyclophosphamide, Etoposide, Dexamethasone; (C)VD, (Cyclophosphamide), Bortezomib, Dexamethasone; VAD, Vincristine, Doxorubicin, Dexamethasone; ID, Idarubicin, Dexamethasone, (V)TD, (Bortezomib), Thalidomide, Dexamethasone, CE, Cyclophosphamide, Etoposide; VACOP-B, Etoposide, Doxorubicin, Cyclophosphamide, Vincristine, Prednisone, Bleomycine; VIP-E: Etoposide, Ifosfamide, Cisplatin, Epirubicin; PD, progressive disease.

Occurrence at relapse
14EVC, IEV, autoPBSCT, Thalidomide ±Melphalan, Vaccination23PD7
22MPT, Rd15PD10
41Dexamethasone, autoPBSCT89PD3
63CED, autoPBSCT, VD24PD3
72VAD, Cyclophosphamide mono, autoPBSCT8PD3
85ID, autoPBSCT, Bendamustin/Thalidomide, VTD, VD, Rd101PD
102ID, IEV, autoPBSCT, alloPBSCT,20PD1
113ID, IEV, auto-alloPBSCT, Rd, VD39PD22
121CVD, CE, autoPBSCT16PD5
135VACOP-B, VIP-E, autoPBSCT, Melphalan, Dexamethasone mono, ID, Doxorubicin mono24PD8

Cytogenetic aberrations were found in 79% of the patients (11/14); del(17p13) in 33% (5/14), del(13q14) in 31% (4/13), MYC-overrepresentation in 38% (5/13) and t(4;14) in 29% (4/14). Patients 2, 5, and 11 showed the most interesting results as cytogenetics revealed different findings in the BM and EMD. Patient 2 showed a MYC-overrepresentation in the BM and, in addition, the extramedullary tumour presented a t(4;14) and a del(17p13). In Patient 5, in addition to a pre-existent del(13q14) in the BM, a del(17p13) was found in the plasma cell tumour. Patient 11 had normal BM cytogenetics and showed a del(13q14) and del(17p13) in the infiltrated ovary.

Four of 19 MM patients (21%) had EMD at initial diagnosis (Patients 15–18, Table 1). Two patients showed an IgM subtype (Patient 17 with lambda and Patient 18 with kappa light chain expression), a very rare variant in MM. Interestingly, both cases revealed a t(4;14) and Patient 18 also showed a further del(13q), which are rare events in IgM MM (Owen et al, 2011; Schuster et al, 2011). All four patients showed cytogenetic aberrations: del(17p13) and del(13q14) were detected as single aberrations in two cases (Patients 16 and 15). Translocation t(4;14) was present in two cases and occurred along with del(13q14) in one case (Patient 18). No MYC-overrepresentation was detected in this group. No information regarding time of development of EMD could be obtained for one of the 19 cases.

Skeletal EMD

The group with skeletal EMD included 11 patients, seven with disease progression or relapse, four with skeletal EMD at initial diagnosis. As shown for the previous EMD patients, IgG was also the main subtype and kappa light chain expression was common. Six patients were male and 5 female, median age was 71 years (range 51–73). FISH analysis revealed del(17p13) and del(13q14) in 27% of the cases (3/11), MYC-overrepresentation and t(4;14) was identified in 18% (2/11) (Table 2). del(13q14) occurred together with del(17p13) in 2 cases (Patients 21 and 23) (= 0·081). As in the EMD group, MYC-overrepresentation never occurred concomitantly with del(13q14). Cytogenetic BM results were available in five cases. Only one patient (Patient 27) had different cytogenetic findings in the BM (normal) versus skeletal EMD [del(13q14)].

Extramedullary plasmacytoma

Six patients with EMP, which showed no signs of dissemination according to the criteria of the IMWG (The International Myeloma Working Group, 2003), were investigated. Four were male and two female with a median age of 44·5 years (range 23–82). Sites involved uterus, nose, sphenoidal sinus, maxillary sinus, paranasal sinus, and epipharynx (Table 3). Although very unusual, the EMP of the uterus was assigned to the group of EMP as no signs for systemic spread such as elevated M-protein or BM infiltration were seen. Three patients had cytogenetic aberrations, each with either a t(4;14), del(17p13) or del(13q14) (Patients 31, 35 and 36). No MYC-overrepresentation was seen in this group.

Comparison of patient groups

Median age was higher in patients with skeletal EMD compared to patients with true EMD (71 years vs. 60·5 years). Cytogenetic aberrations were found more frequently in EMD patients but did not show statistically significant differences to patients with skeletal EMD (P > 0·05). Overall, 84% of EMD patients (16/19) had cytogenetic aberrations compared to 73% of the skeletal EMD patients (7/11). Our analyses showed a del(17p13) in 32% of EMD vs. 27% of skeletal EMD, deletion of 13q in 35% vs. 27%, MYC-overrepresentation in 28% vs. 18% and translocation t(4;14) in 37% vs. 18%. For the group of patients with EMD, a loglinear model showed that only two two-way interactions are sufficient for a satisfactory fitting model. We found significant interactions between del(17p13) and del(13q14) (= 0·012) and between del(13q14) and MYC-overrepresentation (= 0·012). Three patients in the EMD group showed cytogenetic differences between BM and the extramedullary plasma cell tumour, but only one patient in the group with skeletal EMD. MYC-overrepresentation occurred always in advanced or relapsed disease, with the exception of Patient 29, a skeletal EMD that showed this aberration at initial diagnosis.

The median age of the EMP group was significantly younger compared to the other groups (44·5 years vs. 60·5/71 years, = 0·014), in line with other studies (Kremer et al, 2005). EMP patients showed overall less cytogenetic aberrations (50%) compared to EMD (84%) and skeletal EMD (73%). FISH revealed only single aberrations and showed no MYC-overrepresentation.


While FISH-analyses of the BM have helped to identify prognostic subgroups of MM patients, disease specific biological and cytogenetic features that are associated with extramedullary spread are not well defined (Owen et al, 2011). Known genetic BM aberrations reflecting MM disease progression are, for example, MYC-overrepresentation and del(17p13), both occurring as late events and secondary changes during the evolution of MM (Dierlamm et al, 2001; Kuehl & Bergsagel, 2002; Schilling et al, 2008; Fonseca et al, 2009; López-Anglada et al, 2010). In contrast, del(13q14) and t(4;14) are early events that can already be detected in precursor diseases, such as monoclonal gammopathy of undetermined significance (MGUS) or smouldering multiple myeloma (SMM) (Chiecchio et al, 2009).

Assuming EMD as an advanced and aggressive form of MM, we hypothesized a higher frequency of unfavourable cytogenetic aberrations in true EMD compared to known cytogenetic BM data, skeletal EMD and EMP. As skeletal EMD evolve from local extension of bone lesions, we expected that these tumours would demonstrate cytogenetic features of BM-derived plasma cells.

In our study, EMD and skeletal EMD showed a similar incidence of del(17p13) (EMD with 32% vs. skeletal EMD with 27%). The occurrence of del(17p13) in both groups is remarkably higher than reported in BM investigations of MM patients (Fonseca et al, 2003; Pabst et al, 2010). Furthermore, our study showed that del(17p13) occurred twice more frequently in the extramedullary manifestations of EMD patients than in the BM (32% vs. 16%), because 3 of 6 patients bearing del(17p13), showed del(17p13) in the extramedullary mass but not in the BM (Patients 2, 5 and 11). Such a different finding was not observed in the group with skeletal EMD or EMP.

Taken together, our results underline the importance of del(17p13) in extramedullary plasma cell tumour genesis and its possible relevance to the “metastatic” feature of myeloma cells.

In accordance to reports in the literature that found an increasing frequency of t(4;14) from MGUS via SMM to MM (Hallek et al, 1998; Avet-Loiseau et al, 2002) our results revealed a higher frequency of t(4;14) in EMD compared to skeletal EMD (37% vs. 18%) and EMP (17%), emphasizing our theory that EMD evolving from haematogenous spread represents a more advanced disease stage of MM. The incidence of t(4;14) in our group with EMD was even higher when compared to primary BM data described in the literature (15–20%) which can be taken as a sign of further clonal selection and tumour evolution (Fonseca et al, 2004; Sawyer, 2011).

Deletions of 13q are detected in almost 50% of MM cases by FISH at initial diagnosis (Fonseca et al, 2009). Our FISH analyses detected a frequency of del(13q14) that was slightly lower than described in the literature (35% EMD, 27% skeletal EMD and 17% EMP). This might be due to the small study group. Deletion of 13q was formerly thought to be associated with poor prognosis but recent studies confirmed a negative prognostic value only in combination with other genetic aberrations, such as t(4;14) or del(17p13) (Avet-Loiseau et al, 2007). We observed that, when present, del(13q14) was combined with del(17p13) or t(4;14) in all relapsed cases but only in 1 skeletal EMD at diagnosis. Furthermore, 2 patients harboured a del(13q14) in the extramedullary mass but not in the BM. These results are also concordant with previous studies suggesting that deletion of 13q is a pre-condition for clonal tumour expansion because extramedullary relapses are seen more often in patients with del(13q14) (Zeiser et al, 2004; Minnema et al, 2008).

MYC-overrepresentation leads to a higher proliferation rate of myeloma cells and a more aggressive course of disease (Nobuyoshi et al, 1991). In a FISH study, our group reported that 32% of patients showed MYC-gains, which was associated with a shorter event-free survival (Dierlamm et al, 2001). The significance of MYC-gains was also shown in another study, where patients with MYC-overrepresentation (17%) showed a trend towards inferior overall survival after allogeneic stem cell transplantation (Schilling et al, 2008). This is in accordance with present findings where MYC-overrepresentation occurred in 28% of the group with EMD. In addition, this aberration was detected less frequently in the skeletal EMD group (18%) and it was absent in the EMP group, indicating genetic differences compared to EMD.

Taken together, our results suggest cytogenetic differences between extramedullary and intramedullary disease and between EMD and skeletal EMD.

A striking observation is the difference between results obtained from BM and extramedullary mass in four patients concerning del(17p13) and del(13q14), emphasizing the theory that both aberrations demonstrate markers of extramedullary myeloma spread and genetic evolution. In particular, our finding of del(17p13), which additionally occurred in extramedullary tumours in three of the four patients, points to the loss of TP53 as a promoter for extramedullary disease progression. The incidence of del(17p13) was noticeably higher in both EMD and skeletal EMD when compared to BM data reported for MM (Fonseca et al, 2003; Schilling et al, 2008; Pabst et al, 2010). These results may be interpreted as expression of a more advanced disease stage, but may also be related to clonal evolution of the plasma cells.

One further interesting finding is that Patient 2 only revealed MYC-overrepresentation in the BM at initial diagnosis. Subcutaneous lesions appeared during relapse and showed an additional t(4;14) and a del(17p13). This suggests the occurrence of t(4;14) as a secondary genetic event and is therefore contradictory to other data (Fonseca et al, 2009). As investigation of t(4;14) during relapse is not included in the IMWG recommended consensus FISH panel (Fonseca et al, 2009) it is probably not routinely detected and its frequency may therefore be underestimated.

Further prospective investigations on larger patient collectives are needed to confirm whether genetic or other features, e.g. microenvironmental interactions, modified angiogenesis or altered expression of adhesion molecules (Bladé et al, 2011), are causal for the development of EMD and its associated aggressive course.


We thank Dieter K. Hossfeld for discussion and critically reviewing our manuscript.

Author contributions

L.B. performed research, collected and analysed data, and wrote the manuscript. E.M.M.P. performed research and wrote the manuscript. A.M.M., M.E. provided material, collected, analysed and interpreted data. J.Z. performed research, provided material. A.M., J.M., M.T. provided material. A.N., M.L., Gi.S., N.K., E.G., H.H.H. provided material, collected data. E.V. performed statistical analysis. C.B. and J.D. were responsible for manuscript review. Ge.S. conceived and designed research, collected, analysed and interpreted data and wrote the manuscript. All authors approved the final version of the manuscript.

Conflict of interest

The authors report no potential conflicts of interest.