Extramedullary plasmacytoma

Tumor occurrence and therapeutic concepts




Extramedullary plasmacytoma (EMP) is a rare entity belonging to the category of non-Hodgkin lymphoma. EMPs make up 4% of all plasma cell tumors and occur mainly in the upper aerodigestive tract (UAD). Seven patients with EMP included in this evaluation were under the authors' care and have been clinically followed since 1990. Because there are no general guidelines for the treatment of patients with EMP, the authors tried to obtain detailed data about the occurrence of this disease and also reviewed the therapies that have been used. To do so, they evaluated all EMP cases published in the medical literature until now and included their own experience.


Based on the clinical course and follow-up of their own EMP patients, the authors evaluated and reinvestigated all EMP cases cited in MEDLINE, Index Medicus, DIMDI (Deutsches Institut für medizinische Dokumentation und Information, Cologne, Germany), and the reference lists of the publications found through these sources.


In a detailed literature search, more than 400 publications between 1905 and 1997 were found, and these revealed that EMP mainly occurs between the fourth and seventh decades of life. Seven hundred fourteen cases (82.2%) were found in the UAD, and 155 cases (17.8%) were found in other body regions. The following therapeutic strategies were used to treat patients with EMP of the UAD: radiation therapy alone in 44.3%, combined therapy (surgery and radiation) in 26.9%, and surgery alone in 21.9%. The median overall survival or recurrence free survival was longer than 300 months for patients who underwent combined intervention (surgery and radiation). This result was statistically highly significant (P = 0.0027, log rank test) compared with the results for patients who underwent surgical intervention alone (median survival time, 156 months) or radiation therapy alone (median survival time, 144 months). In most cases of non-UAD EMP, surgery was performed (surgery alone, 55.6%; surgery and radiation combined, 19.8%; radiation alone, 11.1%), but there were no statistical differences in survival (P = 0.62). Overall, after treatment for EMP in the UAD, 61.1% of all patients had no recurrence or conversion to systemic involvement (i.e., multiple myeloma, MM); however, 22.0% had recurrence of EMP, and 16.1% had conversion to MM. After treatment for EMP in non-UAD areas, 64.7% of all patients had no recurrence or MM, 21.2% had recurrence, and 14.1% had conversion to MM.


The current investigation provides evidence that surgery alone gives the best results in cases of EMP of the UAD when resectability is good. However, if complete surgical tumor resection is doubtful or impossible and/or if lymph node areas are affected, then combined therapy (surgery and radiation) is recommended. These results, which were obtained from retrospective studies, should be confirmed in randomized trials comparing surgery with combined radiation therapy and surgery. Cancer 1999;85:2305–14. © 1999 American Cancer Society.

Plasmacytoma is an immunoproliferative, monoclonal disease of the B-cell line and is classified as non-Hodgkin lymphoma. It originates as a clone of malignant transformed plasma cells. The plasma cell is a highly mature form of B-cell that is capable of immunoglobulin synthesis and secretion. Plasma cell diseases originate from pathologic plasmablasts that dedifferentiate during the maturation process from primary and secondary B blasts to plasmablasts into malignantly transformed plasmablasts situated in the bone marrow. They migrate and then return to establish themselves in the bone marrow. In rare instances, with the assistance of adhesion molecules, they also may settle in soft tissue or in an extracellular connective tissue area.1 This is the origin for monoclonal plasma cell foci located outside the bone marrow, called extramedullary plasmacytoma (EMP).

The different types of plasma cell tumors are: 1) EMP: a) primary (true) plasmacytoma of the mucosa, unifocal/multifocal, with/without affected lymph nodes; and b) extramedullary manifestations of multiple myeloma (MM); 2) solitary plasmacytoma of the bone (SPB); 3) the multifocal form of MM; 4) MM; and 5) plasmablastic sarcoma (modified classification of Bartl and Deicher, 19952). Isolated EMPs are rare tumors and comprise 4% of all plasma cellular diseases. The first report of the occurrence of EMP was made in 1905.3 Since then, EMP has been described continuously in numerous case reports. Seven patients with EMP (primary/true EMP) have been under our clinical care since 1990. In general, there are no international guidelines for the treatment of EMP. We reviewed EMP of the upper aerodigestive tract (UAD) as well as non-UAD EMP, as published in medical literature and included our own clinical experiences.


Seven patients with EMP of the head and neck region were treated between 1990 and 1997 in our department. The patients were diagnosed with EMP (primary/true EMP) if they met the following criteria: 1) a biopsy proven plasma cell tumor involving a single extramedullary site with or without lymph node involvement, 2) a bone marrow biopsy showing less than 5% plasma cells, 3) normal skeletal survey, and (4) no Bence Jones protein in urine. The presence of a monoclonal band in a serum protein electrophoresis did not exclude a patient from being diagnosed with EMP.


The histologic sections from each case were examined independently by two pathologists without knowledge of the patients' identities or clinical status. Four-micrometer-thick paraffin sections were cut and stained with hematoxylin and eosin (HE), periodic acid–Schiff (PAS) Giemsa for cytological detail, and Gomori stain for reticulin fibers. A Congo red stain was performed in sections from all seven patients. No evidence for amyloid deposition was found. All cases were examined immunohistochemically for kappa and lambda light chains along with immunochemistry for immunoglobulin (Ig) heavy-chains IgG, IgM, and IgA commercial antibodies (Dako, Carpenteria, CA)and using standard techniques (amino-ethylcarbazol, diaminobenzidine).

Histologic Grading

The cases were graded according to the histologic grading criteria devised by Bartl et al.4 (1987) for MM. This involves a three-tiered grading system that is summarized below.

Grade 1 (low grade)

Figure 1 shows a marchalko type in which the plasma cells are indistinguishable from normal cells, with excentric cartwheel nuclei, a perinuclear “hof,” and basophilic cytoplasm. Occasional mitotic figures can be seen.

Figure 1.

Low grade extramedullary plasmacytoma containing marchalko-type plasma cells (Giemsa stain, original magnification ×400).

Grade 2 (intermediate grade)

Figure 2 shows the asynchronous type in which there is marked discrepancy of maturation between nucleus and cytoplasm. At least 50% of all cells have enlarged nuclei with prominent nucleoli. The cytoplasmic:nuclear ratio is low, but a small perinuclear hof usually is present.

Figure 2.

Intermediate grade extramedullary plasmacytoma. Polymorphous and asynchronous cell types with large, excentric nuclei with prominent nucleoli and abundant basophilic cytoplasm (Giemsa stain, original magnification ×400).

Grade 3 (high grade)

Figure 3 shows a plasmablastic type with large nuclei and very prominent, centrally located nucleoli. The cytoplasm is restricted to a fairly narrow rim. The perinuclear hof is inconspicuous or absent.

Figure 3.

High grade extramedullary plasmacytoma consisting of plasmablasts with large nuclei and prominent nucleoli with a moderate rim of basophilic cytoplasm (Giemsa stain, original magnification ×400).

Statistical Methods

Actuarial survival analysis was performed according to the method described by Kaplan and Meier, and the groups were compared by using log rank test for this analysis.5, 6 We included all of the patients with EMP who were discussed in literature. Our own seven patients are described in Table 1. P values ≤ 0.05 were considered to indicate statistical significance. The data were analyzed with the SPSS statistical package for the social sciences (SPSS Inc., Chicago, IL).7

Table 1. Treatment and Follow-Up of the Seven Patients in the Current Study
CaseAge (yrs)/genderSites of occurrenceImmunohistochemistrySerum IgGrade (Bartl)TherapyFollow-up
  • M: male; F: female; NA: no abnormal findings; XRT: radiation therapy; Ig: immunoglobulin; Gy: gray.

  • a

    The patient died of myocardial infarction.

155/MParotid glandIgG, IgA, kappa light chainNALowSurgery, XRT 40 Gy in 4 weeksLocal control after 102 mos
256/MSphenoid sinusIgG, lambda light chainNAHighSurgery, XRT 50 Gy in 5 weeksLocal control after 75 mos
369/MLarynx/glottisKappa light chainNAHighSurgeryLocal control after 62 mos
477/MNasal cavityNo expressionNAHighSurgery, XRT refusedLocal control after 51 mos
570/MNeck sideIgG, kappa light chainNAIntermediateSurgery, XRT 60 Gy in 6 weeksLocal control after 11 mosa
645/MNasal cavity, maxillary sinusIgA, kappa light chainElevated IgA levelsIntermediateSurgery, XRT 40 Gy in 4 weeksLocal control after 30 mos
740/MAryepiglottic space and regional neck side involvementIgG, kappa light chainElevated IgG levelsLowSurgery, XRT 40 Gy in 4 weeksLocal control after 20 mos

Literature Search

The sources for our literature search were the data base MEDLINE (register of all publications in English language), Index Medicus, Deutsches Institut für Medizinische Dokumentation und Information (Cologne), and the reference lists from the respective publications (case reports, review articles, etc.).



All of our patients were male, with an average age of 59 years (range, 40–77 years). One patient showed clinical involvement of cervical lymph nodes at presentation. One patient (patient 6, Table 1) had an IgA and one patient (patient 7, Table 1) an IgG monoclonal band by serum protein electrophoresis. Three months after therapy (surgery followed by radiation [XRT]), the elevated Ig serum levels decreased to a normal range in the two patients. In one patient, there was evidence of adjacent bony destruction at the time of diagnosis. One patient died of myocardial infarction 11 months after treatment for EMP. After biopsy confirmation and myeloma screening, five patients were treated with a combined therapy of surgery followed by XRT. Two patients underwent tumor excision alone (one refused XRT). Doses between 40–60 gray (Gy) were administered over 4–6 weeks with five fractions per week. The median follow-up for all patients was 50 months (range, 11–102 months). The detailed follow-up of the seven patients is described in Table 1. Informed consent was obtained from all patients.

Literature Search

According to the literature search, 714 cases of EMP have been described in the submucosa of the UAD (excluding our own seven cases), and 155 cases have been described in non-UAD areas: 82.2% in the UAD and 17.8% in other body regions. Male patients have EMP in the UAD region three times more frequently than females (74.1% vs. 25.9%), and 59.1% of males and 40.9% of females (male-to-female ratio, 1.4:1) have had EMP in regions other than the UAD. Most cases of EMP in the UAD area occur between the fifth and seventh decade of life compared with the fourth and seventh decade in non-UAD areas.

Sites of EMP

The results of our search revealed a distribution for EMP in the upper aerodigestive tract (UAD),8–20 as described in Table 2. EMPs in non-UAD regions were in the gastrointestinal tract in 40% of cases (n = 62),21 in the urogenital region in 25.2% of cases (n = 39),22 in the skin in 16.8% of cases (n = 26),23 in the lung in 9.7% of cases (n = 15),24 in the breast in 3.9% of cases (n = 6),25 in the conjunctiva in 2% of cases (n = 3),26 and in the retroperitoneum in 1.3% of cases (n = 2).27 Two single EMP cases were found in the skeletal muscle28 and the pituitary gland29 (Table 3). In 7.6% of the cases reported in the UAD (n = 55) and in 2.6% (n = 4) reported in non-UAD areas (n = 4), the regional lymph nodes were involved.

Table 2. Sites of Occurrence of Extramedullary Plasmacytoma in the Upper Aerodigestive Tract According to the Literature Search
Site of occurrence/subunit% (no.)
  1. UAD: upper aerodigestive tract.

Nasal cavity or paranasal sinus43.8 (313)
 Cavity of the nose21.5 (154)
 Maxillary sinus15.2 (108)
 Frontal sinus1.3 (9)
 Sphenoidal sinus2.0 (14)
 Ethmoidal sinus1.7 (12)
 Maxilla2.2 (16)
Nasopharnyx18.3 (131)
Oropharynx17.8 (127)
 Palatine tonsil or soft palate10.5 (75)
 Oral cavity (no further details described)3.1 (22)
 Pharynx2.5 (18)
 Lingual region1.7 (12)
Larynx11.0 (79)
 Subglottis area1.4 (10)
 Epiglottis area1.3 (9)
 Aryepiglottic area1.1 (8)
 Glottis area1.1 (8)
 Morgagni sinus (no further details described)0.1 (1)
 Larynx (no further details described)6.0 (43)
Hypopharynx0.6 (4)
Glands of the UAD4.1 (29)
 Parotid gland2.0 (14)
 Submandibular gland0.7 (5)
 Thyroid1.4 (10)
 Mandibular region0.6 (4)
Trachea1.0 (7)
Esophagus0.1 (1)
Cervical lymph nodes1.7 (12)
Middle ear0.9 (6)
Mastoid0.1 (1)
Table 3. Sites of Extramedullary Plasmacytoma Outside the Upper Aerodigestive Tract Areas According to the Literature Search
Sites of occurrence%No. of cases
Urinary bladder9.014
Portio uteri2.64
Pituitary gland0.61

The following therapeutic concepts were applied in patients with EMP of the UAD: surgical intervention alone in 21.9% of all cases, XRT alone in 44.3% of cases, and combined treatment (surgery and XRT) in 26.9% of cases (Fig. 4, top). The median overall survival or recurrence free survival was greater than 300 months for patients who had combined treatment (surgery and XRT). This was statistically significantly longer (P = 0.0027; log rank test) than those who underwent surgical intervention (median survival, 156 months) or XRT alone (median survival, 144 months) (Fig. 5). In non-UAD areas, surgery was performed in 55.6% of cases, XRT in 11.1%, and combined therapy with surgery followed by XRT in 19.8% of cases (Fig. 4, bottom), but no statistically significant differences in survival were seen (P = 0.62) (Fig. 6). Overall, in 61.1% of the cases with treatment for EMP in the UAD, no MM or recurrence developed; in 22.0%, there was a recurrence, and, in 16.1%, there was a conversion to a MM (Fig. 7, top). Six patients (0.8%) with EMP developed solitary plasmacytoma of the bone (SPB). For patients who were treated for EMP in non-UAD areas, there was no recurrence or MM in 64.7%, recurrence in 21.2%, and conversion to MM in 14.1% of the described cases (Fig. 7, bottom).

Figure 4.

Most common treatment of extramedullary plasmacytoma (EMP) in the upper aerodigestive tract (UAD) and in the non-UAD area according to the literature search. In the UAD, 0.5% of the patients with EMP received chemotherapy, 4.4% received radiation and chemotherapy, 0.2% received surgery and chemotherapy, and 1.8% received surgery, radiation, and chemotherapy. In the non-UAD area, 3.2% of the patients with EMP received a chemotherapy, 3.2% received radiation and chemotherapy, 4.0% received surgery and chemotherapy, and 1.6% received surgery, radiation, and chemotherapy. No comment was found concerning the treatment in 1.5% of the cases.

Figure 5.

Overall survival curve (Kaplan–Meier) of patients with extramedullary plasmacytoma in the upper aerodigestive tract versus treatment according to the literature search.

Figure 6.

Overall survival curve (Kaplan–Meier) of patients with extramedullary plasmacytoma outside the upper aerodigestive tract area versus treatment according to the literature search.

Figure 7.

Overall progression of extramedullary plasmacytoma (EMP) in the upper aerodigestive tract (UAD) and outside the UAD during the observation period according to the literature search (0.8% of patients with EMP located in the UAD later developed a solitary skeletal lesion).


According to the preferred site of EMP in the UAD, the most frequent clinical symptoms are obstruction of the respiratory tract due to tumor growth, epistaxis, rhinorrhea, pain, and swelling of the soft tissue of the neck.30 If there is suspicion of EMP, then a series of examinations is required to establish the diagnosis.

Evaluation of a primary (true) EMP

Diagnostic procedures should be carried out in two steps: An extensive imaging with computed tomography and magnetic resonance imaging31 and a complete endoscopic examination (including a tracheobronchoscopy and endoscopy of the main nasal cavities and the nasopharynx) of the ear, nose, and throat under general anesthesia are required to determine the exact extent of the tumor. Similar endoscopic examinations (i.e., gastroscopy, etc.) are required for EMP in other anatomic regions. Due to the submucosal growth of extramedullary plasmacytomas, deep biopsies, open biopsies, or, depending on the location, complete excision of the tumor are required to histologically confirm the diagnosis. The differential diagnostic evaluation often is very difficult because of the similarity of the histologic appearance of EMP to that of other tumors (i.e., lymphoma, undifferentiated carcinoma, esthesioneuroblasts, etc.) or because the benign inflammatory plasma cell granuloma is large and requires immunohistochemical demonstration of monoclonal kappa or lambda light chains or heavy chains found in plasma cell tumors (IgA, IgD, IgG, IgM).14, 32 Congo red stains are positive in amyloid deposits, which are the result of deposition of light chains in soft tissue.33

The diagnosis of extramedullary plasmacytoma is based on the morphologic and immunophenotypical finding of a localized collection of monoclonal plasma cells in the absence of plasma cell proliferation elsewhere, especially in the bone marrow, and without the presence of malignant lymphoma. EMP must be particularly distinguished from low grade B-cell non-Hodgkin lymphoma, which also may show plasmacytic differentiation, for example, as is seen in lymphoplasmacytic lymphoma. Similar features also can be seen in follicular lymphoma, monocytoid B-cell lymphoma, and extranodal marginal zone lymphoma (low grade lymphoma of the mucosa-associated lymphoid tissue [MALT] type), which often occur in the same locations as EMP.34, 35 Other than morphologic criteria, including centrocyte-like cells, reactive follicles, or the presence of lymphoepithelial lesions, immunostaining for perinuclear or cytoplasmatic Ig expression (IgM rather than IgA or IgG), kappa or lambda light-chain restriction, or the lack of B-lymphocyte antigens (such as CD20) are helpful in differentiating MALT lymphoma from EMP.33–35 In our cases, there was no evidence of nonplasmacytic neoplastic component, and the cells showed the typical features of plasma cells. We observed centrocyte-like cells that were interpreted as infiltrating reactive lymphocytes and not as a neoplastic part of the tumor. Neither lymphoepithelial lesions nor reactive follicles were present. The Ig heavy-chain expression of our 7 cases were IgA or IgG, and all presented cases showed negative staining for the CD20 antigen.

Although there is some evidence in the literature that supports the categorization of EMP as extranodal marginal zone lymphoma (low grade lymphoma of MALT type),37, 38 our results support the classification of our cases as EMP rather than extranodal marginal zone lymphoma (low grade MALT lymphomas). Table 1 shows that immunohistochemical studies revealed the type of immunoglobulins being synthesized by the neoplastic cells in six patients of our clinical group. Mostly IgG immunoglobulins were detected. These findings are in accordance with the results of our literature search, in which 40.4% of EMP cases in the UAD and 54.3% of EMP cases in the remaining body parts produced IgG immunoglobulins. According to the study by Mock et al.,39 the development of myeloma in patients with EMP was observed to occur much more commonly in those producing non-IgG immunoglobulin classes and might support the concept that some EMP cases may be an early manifestation of MM. The criteria used for the histologic grading of extramedullary plasmacytomas were established by Bartl et al. (1987)4 for MM: Grade 1, low grade; Grade 2, intermediate grade; and Grade 3, high grade. In 1976, Wiltshaw40 classified EMP according to three stages: Stage I, EMP limited to an extramedullary site; Stage II, regional lymph nodes affected; and Stage III, multiple metastases. In the literature research, we found that, in 55 cases (7.6%) of EMP in the UAD and in 4 cases (2.6%) in non-UAD areas, regional lymph nodes were involved. If a plasmacytoma is confirmed histologically, then secondary diagnostic procedures must be carried out to exclude systemic involvement.

For the second diagnostic step they should consist of the following examinations: laboratory examinations, including blood sedimentation rate, a complete blood count and blood smear, electrolytes (including Ca2+) and enzyme determination, serum and urine protein electrophoresis, quantitative Ig determination in serum, immunoelectrophoresis and/or immunofixation in serum and in urine, and beta-2-microglobulin determination in serum. For true EMP, chemical laboratory findings are normal, except for the quantitative Ig determination. At the time of diagnosis of EMP, a monoclonal gammopathy is present in approximately 25% of cases of EMP and disappears after successful treatment of the tumor.41 A systemic X-ray examination of the skeleton (cranium, cervical spine, thoracic spine, pelvis, etc.) is definitely required for further diagnosis. An involved skeleton (osteolytic lesions) excludes true EMP and is consistent with MM or SPB. A bone marrow biopsy and aspiration is of utmost importance to determine the percentage of plasma cells. Although some authors accept that less than 10% of plasma cells in a bone marrow smear is compatible with EMP,42 others require that this should be less than 5%.43 We agree with the latter percentage, because, at the onset of MM, many patients have less than 10% of plasma cells in the bone marrow smear.2

There currently are no general guidelines for the treatment of patients with EMP. However, based on the well-known radiation sensitivity of the plasma cell tumor, radiotherapy is accepted as the treatment of choice for EMP.41 The total radiation dose administered ranges between 40 Gy and 60 Gy and is given over a period of 4–6 weeks.41 Surgery or combined therapy (surgery and XRT) has been applied as often as XRT alone in the treatment of EMP in the UAD (Fig. 4, top). Single patients who underwent chemotherapy with or without combined treatment (i.e., XRT, surgery) are also described.41 Detailed evaluation of the literature revealed that, comparing the three types of therapy described above (Fig. 4) for the treatment of patients with EMP in the UAD, 61.1% of patients (n = 436) were recurrence free or without systemic involvement, 22.0% (n = 157) had a local recurrence, and 16.1% (n = 115) had disease that converted to MM (Fig. 7, top). Six patients (0.8%) with EMP developed solitary plasmacytoma of the bone (SPB). The observation period for all patients with EMP in the UAD ranged from 5 months to >350 months. In our patient group, we achieved local control without evidence of systemic involvement for 50 months (range, 11–102 months).

After treatment of EMP in non-UAD regions, 64.7% of all patients (n = 100) had no recurrence or systemic involvement, 21.2% (n = 33) had recurrence, and 14.1% (n = 22) converted to MM (Fig. 7, bottom). These data may emphasize that patients with EMP in locations other than the UAD have outcomes similar to patients with EMP in the UAD. The observation period for these cases ranged from 5 months to over 140 months. Contrary to these findings, Dimopoulos et al.44 reported that patients suffering from SPB developed a systemic involvement (i.e., MM) in two-thirds of cases: Other authors describe conversion to MM in 53% (Holland et al.45), in 48% (Knowling et al.43), and in 50 % (Corwin and Lindberg42), respectively. Patients with MM rarely have a survival time of longer than 10 years (<10%), depending on the classification of the cell type.2, 41

Of all plasma cell tumors, EMP has the best prognosis. According to Holland et al.,45 36% of patients with EMP had conversion to MM compared with 8% of the patients described by Knowling et al.43 and 17% of the patients described by Corwin and Lindberg.42 In our search, we found that EMP converted to MM in 16.1% of the patients with EMP in the UAD and in 14.1% of the patients with EMP in non-UAD areas. There is some evidence that patients with EMP who receive XRT alone tend to have a higher risk of progression to MM compared with those treated with surgical intervention or with combined therapy (surgery and XRT; Table 4). These data are consistent with our clinical group, despite the fact that we had only seven patients. All of our patients underwent surgery, and five received additional XRT (i.e., combined therapy). None developed a recurrence or a systemic involvement in the observation period (median follow-up, 50 months).

Table 4. Extramedullary Plasmacytoma of the Upper Aerodigestive Tract: Conversion to Multiple Myeloma Dependent on Treatment According to the Literature Search
Therapy% of conversion to MM after treatment
  1. MM: multiple myeloma.

Surgery and radiation13.5

We have attempted for the first time to survey all EMP cases found in the UAD and in other locations published in the medical literature. Referring to our own clinical experience with patients suffering from EMP, we recommend the following therapeutic concept. If EMP is present in the soft tissue and is locally and well operable, then surgical excision usually is sufficient.46 Of all of the patients with EMP outside the UAD, 55.6% were treated with surgery alone. However, complete removal of an EMP often is not possible, especially in the UAD, because adjacent vital organ structures may preclude a radical intervention. For such patients, we recommend surgery followed by XRT. This combined therapy seems to provide the best results for treating patients with EMP in the UAD (Fig. 5). However, because the present investigation is predominantly a retrospective study, these results should be confirmed in a multicenter randomized trial.

Patients who are diagnosed with primary EMP remain under life-long medical observation, because, even if therapy is successful, relapses or a generalized plasmacytoma still may appear years later. The use of positron emission tomography may be helpful for the detection of locally recurrent disease.


The authors thank Michael Hennig (Institute for Statistics and Epidemiology at the Technical University of Munich; Professor A. Neiβ, director) for statistical analysis.