European Journal of Radiology 83 (2014) 970–974
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Infiltration patterns in monoclonal plasma cell disorders: correlation of magnetic resonance imaging with matched bone marrow histology Mindaugas Andrulis a , Tobias Bäuerle b , Hartmut Goldschmidt c , Stefan Delorme d , Ola Landgren e , Peter Schirmacher a , Jens Hillengass c,d,∗ a
Institute of Pathology, University of Heidelberg, Heidelberg, Germany Department of Diagnostic and Interventional Radiology, University of Hamburg, Hamburg, Germany c Department of Hematology and Oncology, University of Heidelberg, Heidelberg, Germany d Department of Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany e Multiple Myeloma Section, Metabolism Branch, National Cancer Institute, Bethesda, USA b
a r t i c l e
i n f o
Article history: Received 19 August 2013 Received in revised form 27 February 2014 Accepted 7 March 2014 Keywords: Paraproteinemias Multiple myeloma Plasma cells Magnetic resonance imaging Bone marrow Histology
a b s t r a c t Objectives: To investigate how plasma cell infiltration patterns detected by MRI match the plasma cell distribution in bone marrow biopsy. Methods: We assessed 50 patients with monoclonal plasma cell disorders of all clinical stages. MRI infiltration pattern was compared with matched BM histology from the same anatomic region. Results: MRI revealed a minimal (n = 11, 22%), focal (n = 5, 10%), diffuse (n = 14, 28%) and mixed (n = 20, 40%) infiltration pattern. Diffuse MRI pattern was predominant in smoldering myeloma patients whereas the MRI patterns with “focal component” (i.e. focal and mixed) were most common in symptomatic myeloma (p < 0.01). In histology an interstitial (n = 13, 26%), nodular (n = 23, 46%) and packed marrow (n = 14, 28%) was found respectively. All three histological types of infiltration were observed in patients with diffuse and mixed MRI patterns. Minimal MRI pattern was found in all MGUS patients and was associated with an interstitial BM infiltration. In two patients with minimal MRI pattern an extensive micro-nodular BM infiltration was found in histology. Conclusions: Infiltration patterns in MRI represent different histological growth patterns of plasma cells, but the MRI resolution is not sufficient to visualize micro-nodular aggregates of plasma cells. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Multiple myeloma (MM) is an incurable neoplastic disorder characterized by the infiltration and proliferation of monoclonal plasma cells in bone marrow (BM). According to a current disease progression model [1–3] the symptomatic disease is preceded by preclinical stages of monoclonal gammopathy of undetermined significance (MGUS) and smoldering multiple myeloma (SMM). The number of neoplastic plasma cells in the bone marrow gradually increases with disease progression. However, the distribution of plasma cells within the bone marrow is irregular. As described by Bartl et al. [4], 6 different microscopic infiltration patterns, ranging from interstitial infiltration to packed marrow, can be found in bone marrow biopsies.
∗ Corresponding author at: Department of Hematology and Oncology, University of Heidelberg, Im Neuenheimer Feld 410D-69120 Heidelberg, Germany Tel.: +49 6221 5639397; fax: +49 6221 565647. E-mail address: [email protected] (J. Hillengass). http://dx.doi.org/10.1016/j.ejrad.2014.03.005 0720-048X/© 2014 Elsevier Ireland Ltd. All rights reserved.
Due to uneven distribution of plasma cells various infiltration patterns (minimal, focal, diffuse, mixed) are also observed by modern imaging techniques [5,6]. Magnetic resonance imaging (MRI) is the most sensitive method for detection of both focal and diffuse infiltrates [6]. In contrast to histology, MRI patterns represent “macroscopic” aggregates of plasma cells and differ from histological patterns in scale by an order of magnitude. Nevertheless, MRI patterns correlate significantly with the number of plasma cells in the bone marrow [5] and provide important information for patient management [7,8]. Even thought the infiltration patterns in the biopsy as wells as in imaging are widely used in daily clinical practice, no direct comparison of histological and imaging infiltration patterns has been performed yet. The aim of the present work was to investigate how infiltration patterns detected by MRI match the plasma cell distribution as seen in bone marrow biopsy. For this, we performed a detailed comparison of an MRI of the pelvis with almost synchronous bone marrow histology from the same anatomic region in a cohort of patients with different stages of monoclonal plasma cell disorders.
M. Andrulis et al. / European Journal of Radiology 83 (2014) 970–974 Table 1 Patient characteristics. Median age, years (range) Sex, n (%) Male Female IMWG Category, n (%) MGUS SMM MM Ig Isotype, n (%) Asecretory IgG IgA BJ
971
2.4. Bone marrow samples 57
(38 74)
24 26
(48) (52)
4 19 27
(8) (38) (54)
1 37 8 4
(2) (74) (16) (8)
IMWG, international myeloma working group; BJ, Bence Jones protein.
2. Materials and methods 2.1. Patient characteristics The current group of 50 patients with monoclonal plasma cell diseases consisted of 4 individuals with monoclonal gammopathy of undetermined significance (MGUS), 19 patients with smoldering multiple myeloma (SMM) and 27 patients with symptomatic multiple myeloma (MM) according to the classification of the international myeloma working group [9]. Of the 27 patients with MM 3 were symptomatic because of anemia, 3 because of anemia and bone disease and 21 because of bone disease alone. Patient characteristics are summarized in Table 1. 2.2. Magnetic resonance imaging
The matched bone marrow biopsies were taken directly after the MRI in 48 patients. In two patients the biopsies were taken 19 days before and 20 days after the MRI, respectively. A subset of the biopsies (n = 22) was fixed in formol–methanol solution, embedded in methacrylat as described previously [4]. The remaining core biopsies were fixed in 4% buffered formaldehyde and embedded in paraffin according to standard histology protocols. 3–4 m thin sections were cut and stained for Giemsa to evaluate the histological infiltration pattern. To confirm the accuracy of histological pattern evaluation the paraffin embedded biopsies were additionally immunostained for CD138 as described previously [14]. 2.5. Evaluation of histological parameters The histological infiltration pattern was classified as (1) interstitial, (2) interstitial and sheet-like, (3) interstitial and nodular, (4) nodular, (5) sarcomatoid and (6) packed marrow as described [15]. To facilitate the comparative analysis histological patterns were grouped as follows: the cases with exclusively interstitial infiltration were classified as “interstitial”, the cases with interstitial and sheet-like or nodular infiltration were grouped into the category “nodular”, cases with sarcomatoid infiltration pattern and packed marrow were grouped into the category “packed marrow”. The bone marrow cellularity and the number of plasma cells were assessed in a semi-quantitative manner as described [16]. The histological analysis was performed by an experienced haematopathologist (MA) who was blinded to clinical and imaging data. Examples of histological infiltration patterns are depicted in Fig. 1A. 2.6. Statistical analysis
All whole-body MRI scans from which the pelvic images were used for the current analysis were performed at the University Hospital of Heidelberg and the German Cancer Research Center, Heidelberg, Germany on one of two similar 1.5 Tesla MRI systems (MAGNETOM Avanto, Siemens Medical Solutions, Erlangen, Germany) according to a protocol published previously [10,11]. Focal lesions were defined as lesions 5 mm (slice thickness of the MR images) or larger in diameter with low signal intensity in T1weighted and corresponding high signal intensity in T2-weighted images (Fig. 1a). Diffuse infiltration was defined as homogeneously reduced signal intensity detectable in T1-weighted images alone or in combination with increased signal intensity in T2-weighted images according to the definitions published previously [12,13]. This retrospective analysis of MR images, histology and clinical parameters was approved by the institutional ethics committee. Due to the retrospective nature of the analysis, an informed consent was waived. Both examinations had been performed for clinical work-up. Investigators were blinded to clinical parameters and stage of the disease.
A minimal pattern was found in 11 patients (22%), a focal pattern in 5 patients (10%), a diffuse pattern in 14 patients (28%) and a mixed pattern in 20 patients (40%). All 4 MGUS patients showed a minimal infiltration in MRI. However, 3 patients with SMM and also 4 patients with symptomatic myeloma presented with a minimal pattern. Examples of infiltration patterns in MRI are shown in Fig. 1A (upper and middle row).
2.3. Evaluation of imaging parameters
3.2. Histological findings
Radiological review was performed by two investigators in consensus (JH, TB). Infiltration patterns were classified according to the definitions by Stäbler et al. [13] as minimal (homogeneous hyperintensity in T1- and hypointensity in T2-weighted images, i.e. similar appearance as normal bone marrow), diffuse (homogeneous hypointensity in T1 alone or in combination with hyperintensity in T2-weighted images compared to intervertebral discs) or focal (hypointense lesions in T1- on a hyperintense background corresponding to hyperintense lesions on a hypointense background in T2-weighted images). A mixed pattern was defined as focal lesions on a diffuse background. Examples of the different infiltration patterns are shown in Fig. 1A.
Histology revealed an interstitial pattern in 13 patients (26%), nodular infiltrates in 23 patients (46%) and packed marrow in 14 patients (28%). The length of bone marrow biopsies was highly variable and ranged from 1 mm to 30 mm (median = 15 mm). Importantly, the biopsy size was not different among the three histological patterns, so it is unlikely that the assessment of histological pattern was biased by sampling error due to small biopsy size (Fig. 2a). The amount of fatty marrow was not different in cases with interstitial and nodular infiltration, but significantly reduced in cases with packed marrow (Fig. 2b). The number of plasma cells was significantly associated with histological infiltration pattern (Fig. 2c).
All analyzes were conducted in R [17]. The plots were generated using ggplot2 library [18]. For analysis of discrete variables the Chi-square or Fishers exact test was employed. The comparison of continuous variables was performed using Mann–Whitney U test. 3. Results 3.1. Imaging findings
972
M. Andrulis et al. / European Journal of Radiology 83 (2014) 970–974
Fig. 1. Correlation between MRI pattern, histological pattern and clinical stage. Representative MRI images (a; T1 and T2 rows) and corresponding pictures of bone marrow biopsies (a; CD138 row, immunoperoxidase staining, original magnification × 100) from the same patients representing the three MRI and histological patterns of bone marrow infiltration by plasma cells. Distribution of MRI and histological patterns in relation to the clinical disease stage (b).
Fig. 2. Relationship of biopsy size, fatty marrow, and plasma cell count to histological pattern. The biopsy sizes (a) does not differ for the three histological patterns; the cases with packed marrow have significantly higher cellularity in the biopsy (b); positive correlation of plasma cell count with the histological infiltration pattern (c).
3.3. Correlation between MRI and histological patterns, clinical parameters and disease stage We observed a significant correlation between MRI pattern and histological pattern (Fig. 1b and Fig. 3a, p < 0.001, Chi-square test). The majority of cases with minimal MRI pattern (8/11) had interstitial bone marrow infiltration. Only 3 patients with minimal MRI pattern revealed nodular infiltration in the biopsy (Fig. 1b). Remarkably, in two cases of MM MRI failed to detect an extensive BM-infiltration – in fact they were classified as minimal by MRI (Fig. 1b). Both of these MM cases had more than 80% of plasma cells and showed nodular infiltration in the bone marrow biopsy. A possible sampling error due to small biopsy core was unlikely since the length of both biopsy cores was more the 10 mm thus being greater than used MRI slice thickness of 5 mm. The packed marrow pattern was not observed among patients with a minimal pattern in MRI. All three histological types of infiltration were observed with equal frequency in patients with diffuse MRI pattern (Fig. 3a). The cases with diffuse MRI pattern had the lowest fat content and the widest range of plasma cell count in the biopsy but these findings did not reached statistical significance (Fig. 3c and d). Interestingly, the diffuse MRI pattern was significantly more frequent in SMM patients
compared to MM. In contrast, the MRI patterns with “focal component” (i.e. focal and/or mixed) were most frequent in patients with clinically manifest disease (Fig. 1b, p < 0.01 Fishers exact test). Among cases with mixed MRI infiltration histology revealed nodular infiltration and packed marrow with equal frequency. Only two cases with mixed MRI pattern showed interstitial infiltration in the biopsy. The percentage of plasma cells in the biopsy and extent of anemia were most prominent in patients with mixed MRI pattern (Fig. 3d and e). 4. Discussion In the present study we provide a detailed comparison of the MRI infiltration pattern and the distribution of monoclonal plasma cells assessed on a bone marrow biopsy. In detail, our data demonstrate several important relations between imaging and histology and provide clinically important information for interpretation of MRI findings. We demonstrate that a minimal MRI pattern can be observed in only few patients with extensive bone marrow infiltration and nodular histological pattern. Interestingly, in these cases the fatty marrow amounted to 20% and 40% and the plasma cells were grouped in somewhat smaller nodules. Thus the detection
M. Andrulis et al. / European Journal of Radiology 83 (2014) 970–974
973
Fig. 3. Correlation of MRI patterns with histological parameters and clinical findings. The frequency of histological patterns within every MRI pattern (a); MRI patterns are independent of biopsy core length (b); the cases with diffuse MRI pattern have the lowest fat content (c), but this finding did not reach statistical significance; MRI patterns are associated with the number of plasma cells in the biopsy (d) and cases with diffuse MRI exhibit the broadest range of plasma cell count (d); a mixed MRI pattern is significantly associated with clinically manifest disease stage as demonstrated by relationship of MRI patterns to anemia (e).
level of plasma cell infiltration in MRI is not directly dependent on the number of neoplastic cells. These results suggest that the plasma cell nodules have to reach a certain size in order to be visualized as focal infiltrates by MRI. It also appears that the MRI pattern previously annotated as “normal” [12,13,19–21] might in fact be found despite an interstitial or micro-nodular infiltration. This hypothesis is further supported by our finding that focal MRI pattern was exclusively observed in patients who also had a nodular infiltration at histology (Fig. 1b). Among cases with diffuse MRI pattern all three types of histological infiltration were observed. This indicates that diffuse MRI pattern may be more determined by the amount of fatty marrow or the distribution of plasma cells rather than only the cell count itself. This is further supported by the fact that the cases with diffuse MRI pattern had the highest cellularity in the biopsy, even if this finding did not reached statistical significance (Fig. 3c; p = 0.056, Mann–Whitney U test). This finding is in line with previous observation [19]. On the other hand, the diffuse MRI pattern was significantly associated with SMM hence stressing the clinical importance to recognize this type of infiltration. Previous studies have also shown that the sensitivity of unenhanced MRI for the detection of diffuse infiltration by myeloma is limited for low-grade infiltration, whereas high-grade infiltration can be detected reliably [22,23]. However, quantification is still very difficult. In this setting functional MRI techniques as diffusion weighted imaging (DWI) might be of additional value [7]. The mixed MRI pattern was not associated with any histological infiltration type. This appears like a discrepancy between histology and MRI, but it has to be taken into account that while spatial
resolution of the histological analysis is a single plasma cell (ca. 20 m) the MRI resolution is limited by the voxel size. In the whole body MRI protocol used in our clinical routine the voxel size was 1.25 × 1.25 × 5 mm3 . Therefore, if a biopsy needle is put into the upper iliac crest, a macroscopic focal lesion detectable by MRI may be punctured leading to a packed marrow image in histology. Histological infiltration patterns were significantly associated with the plasma cell count. This possibly indicates that plasma cells aggregate in nodules during tumor progression which is in line with earlier findings showing that the presence and number of focal lesions is of prognostic significance for disease progression not only in symptomatic but also in asymptomatic patients [8,10]. In accordance to this, we observed a major MRI difference between SMM and MM: the diffuse MRI pattern was predominant in SMM patients whereas MRI patterns with a focal component were most common in MM. This suggests that in the preclinical stages, the plasma cells are interstitially scattered or form micro-nodular infiltrates undetectable by MRI despite a high plasma cell count. Consequently such cases would by classified as minimal or diffuse infiltration. It is a valid assumption that when diseases advances, the plasma cells build nodular aggregates that are big enough to be detected as focal or mixed infiltration by MRI. This implicates that the focal MRI component corresponds to neoplastic infiltrates that drive the disease progression. A limitation of the current study is that the evaluation of histology and MRI were performed retrospectively. However, we attempted to compensate this issue by using a blinded double read to achieve maximum reproducibility. Furthermore, the biopsy was done in transversal direction and the MRI read in coronal view but
974
M. Andrulis et al. / European Journal of Radiology 83 (2014) 970–974
clinical experience allows assuming that the infiltration patterns in MRI are the same in both views and that the coronal orientation gives a better impression of the pelvic bone marrow. 5. Conclusions In summary, we present the first comparative analysis of MRI and synchronous bone marrow histology from the same anatomic region in patients with monoclonal plasma cell disorders. We conclude that infiltration patterns in MRI indeed represent different growth in the bone marrow as detected by histology. Furthermore, our findings show that even extensive bone marrow infiltration is not always detectable by MRI because the limited resolution of MRI is not sufficient to visualize micro-nodular aggregates of plasma cells. In other terms, the scale of patterns in MRI and histology are inevitably different by an order of magnitude. Taken together, the findings presented here demonstrate the strengths and limitations of conventional MRI and argues for use of advanced, contrastenhanced or DWI-based techniques for detection of interstitial or micro-nodular plasma cell infiltrates. Conflict of Interest The authors have no conflict of interest. Acknowledgements This work was supported by grants from the International Myeloma Foundation and from the German Research Foundation: grant for “initiation of bilateral cooperation” and Transregio 79. References [1] Landgren O, Kyle RA, Pfeiffer RM, Katzmann JA, Caporaso NE, Hayes RB, et al. Monoclonal gammopathy of undetermined significance (MGUS) consistently precedes multiple myeloma: a prospective study. Blood 2009;113:5412–7, http://dx.doi.org/10.1182/blood-2008-12-194241. [2] Korde N, Kristinsson SY, Landgren O. Monoclonal gammopathy of undetermined significance (MGUS) and smoldering multiple myeloma (SMM): novel biological insights and development of early treatment strategies. Blood 2011;117:5573–81, http://dx.doi.org/10.1182/blood-2011-01-270140. [3] Morgan GJ, Walker BA, Davies FE. The genetic architecture of multiple myeloma. Nat Rev Cancer 2012;12:335–48, http://dx.doi.org/10.1038/nrc3257. [4] Bartl R, Frisch B, Burkhardt R, Fateh-Moghadam A, Mahl G, Gierster P, et al. Bone marrow histology in myeloma: its importance in diagnosis, prognosis, classification and staging. Brit J Haematol 1982;51:361–75. [5] Baur A, Stäbler A, Bartl R, Lamerz R, Reiser M. Infiltration patterns of plasmacytomas in magnetic resonance tomography. RöFo Fortschritte Auf Dem Geb Röntgenstrahlen Nukl 1996;164:457–63, http://dx.doi.org/10.1055/s-2007-1015689. [6] Zamagni E, Nanni C, Patriarca F, Englaro E, Castellucci P, Geatti O, et al. A prospective comparison of 18F-fluorodeoxyglucose positron emission tomography-computed tomography, magnetic resonance imaging and wholebody planar radiographs in the assessment of bone disease in newly diagnosed multiple myeloma. Haematologica 2007;92:50–5.
[7] Hillengass J, Bäuerle T, Bartl R, Andrulis M, McClanahan F, Laun FB, et al. Diffusion-weighted imaging for non-invasive and quantitative monitoring of bone marrow infiltration in patients with monoclonal plasma cell disease: a comparative study with histology. Brit J Haematol 2011;153:721–8, http://dx.doi.org/10.1111/j. 1365-2141.2011.08658.x. [8] Walker R, Barlogie B, Haessler J, Tricot G, Anaissie E, Shaughnessy Jr JD, et al. Magnetic resonance imaging in multiple myeloma: diagnostic and clinical implications. J Clin Oncol Off J Am Soc Clin Oncol 2007;25:1121–8, http://dx.doi.org/10.1200/JCO.2006.08.5803. [9] International Myeloma Working Group. Criteria for the classification of monoclonal gammopathies, multiple myeloma and related disorders: a report of the International Myeloma Working Group. Brit J Haematol 2003;121:749–57. [10] Hillengass J, Fechtner K, Weber M-A, Bäuerle T, Ayyaz S, Heiss C, et al. Prognostic significance of focal lesions in whole-body magnetic resonance imaging in patients with asymptomatic multiple myeloma. J Clin Oncol Off J Am Soc Clin Oncol 2010;28:1606–10, http://dx.doi.org/10.1200/JCO.2009.25.5356. [11] Fechtner K, Hillengass J, Delorme S, Heiss C, Neben K, Goldschmidt H, et al. Staging monoclonal plasma cell disease: comparison of the Durie-Salmon and the Durie-Salmon PLUS staging systems. Radiology 2010;257:195–204, http://dx.doi.org/10.1148/radiol.10091809. [12] Baur A, Stäbler A, Nagel D, Lamerz R, Bartl R, Hiller E, et al. Magnetic resonance imaging as a supplement for the clinical staging system of Durie and Salmon? Cancer 2002;95:1334–45, http://dx.doi.org/10.1002/cncr.10818. [13] Stäbler A, Baur A, Bartl R, Munker R, Lamerz R, Reiser MF. Contrast enhancement and quantitative signal analysis in MR imaging of multiple myeloma: assessment of focal and diffuse growth patterns in marrow correlated with biopsies and survival rates. AJR Am J Roentgenol 1996;167:1029–36, http://dx.doi.org/10.2214/ajr.167.4.8819407. [14] Chatterjee M, Andrulis M, Stühmer T, Müller E, Hofmann C, Steinbrunn T, et al. The PI3K/Akt signalling pathway regulates the expression of Hsp70, which critically contributes to Hsp90-chaperone function and tumor cell survival in multiple myeloma. Haematologica 2012, http://dx.doi.org/10.3324/haematol.2012.066175. [15] Bartl R. Histologic classification and staging of multiple myeloma. Hematol Oncol 1988;6:107–13. [16] Riccardi A, Ucci G, Luoni R, Castello A, Coci A, Magrini U, et al. Bone marrow biopsy in monoclonal gammopathies: correlations between pathological findings and clinical data. The Cooperative Group for Study and Treatment of Multiple Myeloma. J Clin Pathol 1990;43:469–75. [17] R Core Team (2012) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. [18] Hadley W. ggplot2: Elegant Graphics for Data Analysis. New York: Springer; 2009. [19] Moulopoulos LA, Gika D, Anagnostopoulos A, Delasalle K, Weber D, Alexanian R, et al. Prognostic significance of magnetic resonance imaging of bone marrow in previously untreated patients with multiple myeloma. Ann Oncol Off J Eur Soc Med Oncol ESMO 2005;16:1824–8, http://dx.doi.org/10.1093/annonc/mdi362. [20] Bäuerle T, Hillengass J, Fechtner K, Zechmann CM, Grenacher L, Moehler TM, et al. Multiple myeloma and monoclonal gammopathy of undetermined significance: importance of whole-body versus spinal MR imaging. Radiology 2009;252:477–85, http://dx.doi.org/10.1148/radiol.2522081756. [21] Lecouvet FE, Vande Berg BC, Michaux L, Malghem J, Maldague BE, Jamart J, et al. Stage III multiple myeloma: clinical and prognostic value of spinal bone marrow MR imaging. Radiology 1998;209:653–60. [22] Wasser K, Moehler T, Nosas-Garcia S, Rehm C, Bartl R, Goldschmidt H, et al. [Correlation of MRI and histopathology of bone marrow in patients with multiple myeloma]. RöFo Fortschritte Auf Dem Geb Röntgenstrahlen Nukl 2005;177:1116–22, http://dx.doi.org/10.1055/s-2005-858362. [23] Schmidt GP, Baur A, Stäbler A, Schoenberg SO, Steinborn M, Baltin V, et al. [Estimation of diffuse bone marrow infiltration of the spine in multiple myeloma: correlation of MRT with histological results]. RöFo Fortschritte Auf Dem Geb Röntgenstrahlen Nukl 2005;177:745–50, http://dx.doi.org/10.1055/s-2005-857869.
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Infiltration patterns in monoclonal plasma cell disorders: correlation of magnetic resonance imaging with matched bone marrow histology Mindaugas Andrulis a , Tobias Bäuerle b , Hartmut Goldschmidt c , Stefan Delorme d , Ola Landgren e , Peter Schirmacher a , Jens Hillengass c,d,∗ a
Institute of Pathology, University of Heidelberg, Heidelberg, Germany Department of Diagnostic and Interventional Radiology, University of Hamburg, Hamburg, Germany c Department of Hematology and Oncology, University of Heidelberg, Heidelberg, Germany d Department of Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany e Multiple Myeloma Section, Metabolism Branch, National Cancer Institute, Bethesda, USA b
a r t i c l e
i n f o
Article history: Received 19 August 2013 Received in revised form 27 February 2014 Accepted 7 March 2014 Keywords: Paraproteinemias Multiple myeloma Plasma cells Magnetic resonance imaging Bone marrow Histology
a b s t r a c t Objectives: To investigate how plasma cell infiltration patterns detected by MRI match the plasma cell distribution in bone marrow biopsy. Methods: We assessed 50 patients with monoclonal plasma cell disorders of all clinical stages. MRI infiltration pattern was compared with matched BM histology from the same anatomic region. Results: MRI revealed a minimal (n = 11, 22%), focal (n = 5, 10%), diffuse (n = 14, 28%) and mixed (n = 20, 40%) infiltration pattern. Diffuse MRI pattern was predominant in smoldering myeloma patients whereas the MRI patterns with “focal component” (i.e. focal and mixed) were most common in symptomatic myeloma (p < 0.01). In histology an interstitial (n = 13, 26%), nodular (n = 23, 46%) and packed marrow (n = 14, 28%) was found respectively. All three histological types of infiltration were observed in patients with diffuse and mixed MRI patterns. Minimal MRI pattern was found in all MGUS patients and was associated with an interstitial BM infiltration. In two patients with minimal MRI pattern an extensive micro-nodular BM infiltration was found in histology. Conclusions: Infiltration patterns in MRI represent different histological growth patterns of plasma cells, but the MRI resolution is not sufficient to visualize micro-nodular aggregates of plasma cells. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Multiple myeloma (MM) is an incurable neoplastic disorder characterized by the infiltration and proliferation of monoclonal plasma cells in bone marrow (BM). According to a current disease progression model [1–3] the symptomatic disease is preceded by preclinical stages of monoclonal gammopathy of undetermined significance (MGUS) and smoldering multiple myeloma (SMM). The number of neoplastic plasma cells in the bone marrow gradually increases with disease progression. However, the distribution of plasma cells within the bone marrow is irregular. As described by Bartl et al. [4], 6 different microscopic infiltration patterns, ranging from interstitial infiltration to packed marrow, can be found in bone marrow biopsies.
∗ Corresponding author at: Department of Hematology and Oncology, University of Heidelberg, Im Neuenheimer Feld 410D-69120 Heidelberg, Germany Tel.: +49 6221 5639397; fax: +49 6221 565647. E-mail address: [email protected] (J. Hillengass). http://dx.doi.org/10.1016/j.ejrad.2014.03.005 0720-048X/© 2014 Elsevier Ireland Ltd. All rights reserved.
Due to uneven distribution of plasma cells various infiltration patterns (minimal, focal, diffuse, mixed) are also observed by modern imaging techniques [5,6]. Magnetic resonance imaging (MRI) is the most sensitive method for detection of both focal and diffuse infiltrates [6]. In contrast to histology, MRI patterns represent “macroscopic” aggregates of plasma cells and differ from histological patterns in scale by an order of magnitude. Nevertheless, MRI patterns correlate significantly with the number of plasma cells in the bone marrow [5] and provide important information for patient management [7,8]. Even thought the infiltration patterns in the biopsy as wells as in imaging are widely used in daily clinical practice, no direct comparison of histological and imaging infiltration patterns has been performed yet. The aim of the present work was to investigate how infiltration patterns detected by MRI match the plasma cell distribution as seen in bone marrow biopsy. For this, we performed a detailed comparison of an MRI of the pelvis with almost synchronous bone marrow histology from the same anatomic region in a cohort of patients with different stages of monoclonal plasma cell disorders.
M. Andrulis et al. / European Journal of Radiology 83 (2014) 970–974 Table 1 Patient characteristics. Median age, years (range) Sex, n (%) Male Female IMWG Category, n (%) MGUS SMM MM Ig Isotype, n (%) Asecretory IgG IgA BJ
971
2.4. Bone marrow samples 57
(38 74)
24 26
(48) (52)
4 19 27
(8) (38) (54)
1 37 8 4
(2) (74) (16) (8)
IMWG, international myeloma working group; BJ, Bence Jones protein.
2. Materials and methods 2.1. Patient characteristics The current group of 50 patients with monoclonal plasma cell diseases consisted of 4 individuals with monoclonal gammopathy of undetermined significance (MGUS), 19 patients with smoldering multiple myeloma (SMM) and 27 patients with symptomatic multiple myeloma (MM) according to the classification of the international myeloma working group [9]. Of the 27 patients with MM 3 were symptomatic because of anemia, 3 because of anemia and bone disease and 21 because of bone disease alone. Patient characteristics are summarized in Table 1. 2.2. Magnetic resonance imaging
The matched bone marrow biopsies were taken directly after the MRI in 48 patients. In two patients the biopsies were taken 19 days before and 20 days after the MRI, respectively. A subset of the biopsies (n = 22) was fixed in formol–methanol solution, embedded in methacrylat as described previously [4]. The remaining core biopsies were fixed in 4% buffered formaldehyde and embedded in paraffin according to standard histology protocols. 3–4 m thin sections were cut and stained for Giemsa to evaluate the histological infiltration pattern. To confirm the accuracy of histological pattern evaluation the paraffin embedded biopsies were additionally immunostained for CD138 as described previously [14]. 2.5. Evaluation of histological parameters The histological infiltration pattern was classified as (1) interstitial, (2) interstitial and sheet-like, (3) interstitial and nodular, (4) nodular, (5) sarcomatoid and (6) packed marrow as described [15]. To facilitate the comparative analysis histological patterns were grouped as follows: the cases with exclusively interstitial infiltration were classified as “interstitial”, the cases with interstitial and sheet-like or nodular infiltration were grouped into the category “nodular”, cases with sarcomatoid infiltration pattern and packed marrow were grouped into the category “packed marrow”. The bone marrow cellularity and the number of plasma cells were assessed in a semi-quantitative manner as described [16]. The histological analysis was performed by an experienced haematopathologist (MA) who was blinded to clinical and imaging data. Examples of histological infiltration patterns are depicted in Fig. 1A. 2.6. Statistical analysis
All whole-body MRI scans from which the pelvic images were used for the current analysis were performed at the University Hospital of Heidelberg and the German Cancer Research Center, Heidelberg, Germany on one of two similar 1.5 Tesla MRI systems (MAGNETOM Avanto, Siemens Medical Solutions, Erlangen, Germany) according to a protocol published previously [10,11]. Focal lesions were defined as lesions 5 mm (slice thickness of the MR images) or larger in diameter with low signal intensity in T1weighted and corresponding high signal intensity in T2-weighted images (Fig. 1a). Diffuse infiltration was defined as homogeneously reduced signal intensity detectable in T1-weighted images alone or in combination with increased signal intensity in T2-weighted images according to the definitions published previously [12,13]. This retrospective analysis of MR images, histology and clinical parameters was approved by the institutional ethics committee. Due to the retrospective nature of the analysis, an informed consent was waived. Both examinations had been performed for clinical work-up. Investigators were blinded to clinical parameters and stage of the disease.
A minimal pattern was found in 11 patients (22%), a focal pattern in 5 patients (10%), a diffuse pattern in 14 patients (28%) and a mixed pattern in 20 patients (40%). All 4 MGUS patients showed a minimal infiltration in MRI. However, 3 patients with SMM and also 4 patients with symptomatic myeloma presented with a minimal pattern. Examples of infiltration patterns in MRI are shown in Fig. 1A (upper and middle row).
2.3. Evaluation of imaging parameters
3.2. Histological findings
Radiological review was performed by two investigators in consensus (JH, TB). Infiltration patterns were classified according to the definitions by Stäbler et al. [13] as minimal (homogeneous hyperintensity in T1- and hypointensity in T2-weighted images, i.e. similar appearance as normal bone marrow), diffuse (homogeneous hypointensity in T1 alone or in combination with hyperintensity in T2-weighted images compared to intervertebral discs) or focal (hypointense lesions in T1- on a hyperintense background corresponding to hyperintense lesions on a hypointense background in T2-weighted images). A mixed pattern was defined as focal lesions on a diffuse background. Examples of the different infiltration patterns are shown in Fig. 1A.
Histology revealed an interstitial pattern in 13 patients (26%), nodular infiltrates in 23 patients (46%) and packed marrow in 14 patients (28%). The length of bone marrow biopsies was highly variable and ranged from 1 mm to 30 mm (median = 15 mm). Importantly, the biopsy size was not different among the three histological patterns, so it is unlikely that the assessment of histological pattern was biased by sampling error due to small biopsy size (Fig. 2a). The amount of fatty marrow was not different in cases with interstitial and nodular infiltration, but significantly reduced in cases with packed marrow (Fig. 2b). The number of plasma cells was significantly associated with histological infiltration pattern (Fig. 2c).
All analyzes were conducted in R [17]. The plots were generated using ggplot2 library [18]. For analysis of discrete variables the Chi-square or Fishers exact test was employed. The comparison of continuous variables was performed using Mann–Whitney U test. 3. Results 3.1. Imaging findings
972
M. Andrulis et al. / European Journal of Radiology 83 (2014) 970–974
Fig. 1. Correlation between MRI pattern, histological pattern and clinical stage. Representative MRI images (a; T1 and T2 rows) and corresponding pictures of bone marrow biopsies (a; CD138 row, immunoperoxidase staining, original magnification × 100) from the same patients representing the three MRI and histological patterns of bone marrow infiltration by plasma cells. Distribution of MRI and histological patterns in relation to the clinical disease stage (b).
Fig. 2. Relationship of biopsy size, fatty marrow, and plasma cell count to histological pattern. The biopsy sizes (a) does not differ for the three histological patterns; the cases with packed marrow have significantly higher cellularity in the biopsy (b); positive correlation of plasma cell count with the histological infiltration pattern (c).
3.3. Correlation between MRI and histological patterns, clinical parameters and disease stage We observed a significant correlation between MRI pattern and histological pattern (Fig. 1b and Fig. 3a, p < 0.001, Chi-square test). The majority of cases with minimal MRI pattern (8/11) had interstitial bone marrow infiltration. Only 3 patients with minimal MRI pattern revealed nodular infiltration in the biopsy (Fig. 1b). Remarkably, in two cases of MM MRI failed to detect an extensive BM-infiltration – in fact they were classified as minimal by MRI (Fig. 1b). Both of these MM cases had more than 80% of plasma cells and showed nodular infiltration in the bone marrow biopsy. A possible sampling error due to small biopsy core was unlikely since the length of both biopsy cores was more the 10 mm thus being greater than used MRI slice thickness of 5 mm. The packed marrow pattern was not observed among patients with a minimal pattern in MRI. All three histological types of infiltration were observed with equal frequency in patients with diffuse MRI pattern (Fig. 3a). The cases with diffuse MRI pattern had the lowest fat content and the widest range of plasma cell count in the biopsy but these findings did not reached statistical significance (Fig. 3c and d). Interestingly, the diffuse MRI pattern was significantly more frequent in SMM patients
compared to MM. In contrast, the MRI patterns with “focal component” (i.e. focal and/or mixed) were most frequent in patients with clinically manifest disease (Fig. 1b, p < 0.01 Fishers exact test). Among cases with mixed MRI infiltration histology revealed nodular infiltration and packed marrow with equal frequency. Only two cases with mixed MRI pattern showed interstitial infiltration in the biopsy. The percentage of plasma cells in the biopsy and extent of anemia were most prominent in patients with mixed MRI pattern (Fig. 3d and e). 4. Discussion In the present study we provide a detailed comparison of the MRI infiltration pattern and the distribution of monoclonal plasma cells assessed on a bone marrow biopsy. In detail, our data demonstrate several important relations between imaging and histology and provide clinically important information for interpretation of MRI findings. We demonstrate that a minimal MRI pattern can be observed in only few patients with extensive bone marrow infiltration and nodular histological pattern. Interestingly, in these cases the fatty marrow amounted to 20% and 40% and the plasma cells were grouped in somewhat smaller nodules. Thus the detection
M. Andrulis et al. / European Journal of Radiology 83 (2014) 970–974
973
Fig. 3. Correlation of MRI patterns with histological parameters and clinical findings. The frequency of histological patterns within every MRI pattern (a); MRI patterns are independent of biopsy core length (b); the cases with diffuse MRI pattern have the lowest fat content (c), but this finding did not reach statistical significance; MRI patterns are associated with the number of plasma cells in the biopsy (d) and cases with diffuse MRI exhibit the broadest range of plasma cell count (d); a mixed MRI pattern is significantly associated with clinically manifest disease stage as demonstrated by relationship of MRI patterns to anemia (e).
level of plasma cell infiltration in MRI is not directly dependent on the number of neoplastic cells. These results suggest that the plasma cell nodules have to reach a certain size in order to be visualized as focal infiltrates by MRI. It also appears that the MRI pattern previously annotated as “normal” [12,13,19–21] might in fact be found despite an interstitial or micro-nodular infiltration. This hypothesis is further supported by our finding that focal MRI pattern was exclusively observed in patients who also had a nodular infiltration at histology (Fig. 1b). Among cases with diffuse MRI pattern all three types of histological infiltration were observed. This indicates that diffuse MRI pattern may be more determined by the amount of fatty marrow or the distribution of plasma cells rather than only the cell count itself. This is further supported by the fact that the cases with diffuse MRI pattern had the highest cellularity in the biopsy, even if this finding did not reached statistical significance (Fig. 3c; p = 0.056, Mann–Whitney U test). This finding is in line with previous observation [19]. On the other hand, the diffuse MRI pattern was significantly associated with SMM hence stressing the clinical importance to recognize this type of infiltration. Previous studies have also shown that the sensitivity of unenhanced MRI for the detection of diffuse infiltration by myeloma is limited for low-grade infiltration, whereas high-grade infiltration can be detected reliably [22,23]. However, quantification is still very difficult. In this setting functional MRI techniques as diffusion weighted imaging (DWI) might be of additional value [7]. The mixed MRI pattern was not associated with any histological infiltration type. This appears like a discrepancy between histology and MRI, but it has to be taken into account that while spatial
resolution of the histological analysis is a single plasma cell (ca. 20 m) the MRI resolution is limited by the voxel size. In the whole body MRI protocol used in our clinical routine the voxel size was 1.25 × 1.25 × 5 mm3 . Therefore, if a biopsy needle is put into the upper iliac crest, a macroscopic focal lesion detectable by MRI may be punctured leading to a packed marrow image in histology. Histological infiltration patterns were significantly associated with the plasma cell count. This possibly indicates that plasma cells aggregate in nodules during tumor progression which is in line with earlier findings showing that the presence and number of focal lesions is of prognostic significance for disease progression not only in symptomatic but also in asymptomatic patients [8,10]. In accordance to this, we observed a major MRI difference between SMM and MM: the diffuse MRI pattern was predominant in SMM patients whereas MRI patterns with a focal component were most common in MM. This suggests that in the preclinical stages, the plasma cells are interstitially scattered or form micro-nodular infiltrates undetectable by MRI despite a high plasma cell count. Consequently such cases would by classified as minimal or diffuse infiltration. It is a valid assumption that when diseases advances, the plasma cells build nodular aggregates that are big enough to be detected as focal or mixed infiltration by MRI. This implicates that the focal MRI component corresponds to neoplastic infiltrates that drive the disease progression. A limitation of the current study is that the evaluation of histology and MRI were performed retrospectively. However, we attempted to compensate this issue by using a blinded double read to achieve maximum reproducibility. Furthermore, the biopsy was done in transversal direction and the MRI read in coronal view but
974
M. Andrulis et al. / European Journal of Radiology 83 (2014) 970–974
clinical experience allows assuming that the infiltration patterns in MRI are the same in both views and that the coronal orientation gives a better impression of the pelvic bone marrow. 5. Conclusions In summary, we present the first comparative analysis of MRI and synchronous bone marrow histology from the same anatomic region in patients with monoclonal plasma cell disorders. We conclude that infiltration patterns in MRI indeed represent different growth in the bone marrow as detected by histology. Furthermore, our findings show that even extensive bone marrow infiltration is not always detectable by MRI because the limited resolution of MRI is not sufficient to visualize micro-nodular aggregates of plasma cells. In other terms, the scale of patterns in MRI and histology are inevitably different by an order of magnitude. Taken together, the findings presented here demonstrate the strengths and limitations of conventional MRI and argues for use of advanced, contrastenhanced or DWI-based techniques for detection of interstitial or micro-nodular plasma cell infiltrates. Conflict of Interest The authors have no conflict of interest. Acknowledgements This work was supported by grants from the International Myeloma Foundation and from the German Research Foundation: grant for “initiation of bilateral cooperation” and Transregio 79. References [1] Landgren O, Kyle RA, Pfeiffer RM, Katzmann JA, Caporaso NE, Hayes RB, et al. Monoclonal gammopathy of undetermined significance (MGUS) consistently precedes multiple myeloma: a prospective study. Blood 2009;113:5412–7, http://dx.doi.org/10.1182/blood-2008-12-194241. [2] Korde N, Kristinsson SY, Landgren O. Monoclonal gammopathy of undetermined significance (MGUS) and smoldering multiple myeloma (SMM): novel biological insights and development of early treatment strategies. Blood 2011;117:5573–81, http://dx.doi.org/10.1182/blood-2011-01-270140. [3] Morgan GJ, Walker BA, Davies FE. The genetic architecture of multiple myeloma. Nat Rev Cancer 2012;12:335–48, http://dx.doi.org/10.1038/nrc3257. [4] Bartl R, Frisch B, Burkhardt R, Fateh-Moghadam A, Mahl G, Gierster P, et al. Bone marrow histology in myeloma: its importance in diagnosis, prognosis, classification and staging. Brit J Haematol 1982;51:361–75. [5] Baur A, Stäbler A, Bartl R, Lamerz R, Reiser M. Infiltration patterns of plasmacytomas in magnetic resonance tomography. RöFo Fortschritte Auf Dem Geb Röntgenstrahlen Nukl 1996;164:457–63, http://dx.doi.org/10.1055/s-2007-1015689. [6] Zamagni E, Nanni C, Patriarca F, Englaro E, Castellucci P, Geatti O, et al. A prospective comparison of 18F-fluorodeoxyglucose positron emission tomography-computed tomography, magnetic resonance imaging and wholebody planar radiographs in the assessment of bone disease in newly diagnosed multiple myeloma. Haematologica 2007;92:50–5.
[7] Hillengass J, Bäuerle T, Bartl R, Andrulis M, McClanahan F, Laun FB, et al. Diffusion-weighted imaging for non-invasive and quantitative monitoring of bone marrow infiltration in patients with monoclonal plasma cell disease: a comparative study with histology. Brit J Haematol 2011;153:721–8, http://dx.doi.org/10.1111/j. 1365-2141.2011.08658.x. [8] Walker R, Barlogie B, Haessler J, Tricot G, Anaissie E, Shaughnessy Jr JD, et al. Magnetic resonance imaging in multiple myeloma: diagnostic and clinical implications. J Clin Oncol Off J Am Soc Clin Oncol 2007;25:1121–8, http://dx.doi.org/10.1200/JCO.2006.08.5803. [9] International Myeloma Working Group. Criteria for the classification of monoclonal gammopathies, multiple myeloma and related disorders: a report of the International Myeloma Working Group. Brit J Haematol 2003;121:749–57. [10] Hillengass J, Fechtner K, Weber M-A, Bäuerle T, Ayyaz S, Heiss C, et al. Prognostic significance of focal lesions in whole-body magnetic resonance imaging in patients with asymptomatic multiple myeloma. J Clin Oncol Off J Am Soc Clin Oncol 2010;28:1606–10, http://dx.doi.org/10.1200/JCO.2009.25.5356. [11] Fechtner K, Hillengass J, Delorme S, Heiss C, Neben K, Goldschmidt H, et al. Staging monoclonal plasma cell disease: comparison of the Durie-Salmon and the Durie-Salmon PLUS staging systems. Radiology 2010;257:195–204, http://dx.doi.org/10.1148/radiol.10091809. [12] Baur A, Stäbler A, Nagel D, Lamerz R, Bartl R, Hiller E, et al. Magnetic resonance imaging as a supplement for the clinical staging system of Durie and Salmon? Cancer 2002;95:1334–45, http://dx.doi.org/10.1002/cncr.10818. [13] Stäbler A, Baur A, Bartl R, Munker R, Lamerz R, Reiser MF. Contrast enhancement and quantitative signal analysis in MR imaging of multiple myeloma: assessment of focal and diffuse growth patterns in marrow correlated with biopsies and survival rates. AJR Am J Roentgenol 1996;167:1029–36, http://dx.doi.org/10.2214/ajr.167.4.8819407. [14] Chatterjee M, Andrulis M, Stühmer T, Müller E, Hofmann C, Steinbrunn T, et al. The PI3K/Akt signalling pathway regulates the expression of Hsp70, which critically contributes to Hsp90-chaperone function and tumor cell survival in multiple myeloma. Haematologica 2012, http://dx.doi.org/10.3324/haematol.2012.066175. [15] Bartl R. Histologic classification and staging of multiple myeloma. Hematol Oncol 1988;6:107–13. [16] Riccardi A, Ucci G, Luoni R, Castello A, Coci A, Magrini U, et al. Bone marrow biopsy in monoclonal gammopathies: correlations between pathological findings and clinical data. The Cooperative Group for Study and Treatment of Multiple Myeloma. J Clin Pathol 1990;43:469–75. [17] R Core Team (2012) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. [18] Hadley W. ggplot2: Elegant Graphics for Data Analysis. New York: Springer; 2009. [19] Moulopoulos LA, Gika D, Anagnostopoulos A, Delasalle K, Weber D, Alexanian R, et al. Prognostic significance of magnetic resonance imaging of bone marrow in previously untreated patients with multiple myeloma. Ann Oncol Off J Eur Soc Med Oncol ESMO 2005;16:1824–8, http://dx.doi.org/10.1093/annonc/mdi362. [20] Bäuerle T, Hillengass J, Fechtner K, Zechmann CM, Grenacher L, Moehler TM, et al. Multiple myeloma and monoclonal gammopathy of undetermined significance: importance of whole-body versus spinal MR imaging. Radiology 2009;252:477–85, http://dx.doi.org/10.1148/radiol.2522081756. [21] Lecouvet FE, Vande Berg BC, Michaux L, Malghem J, Maldague BE, Jamart J, et al. Stage III multiple myeloma: clinical and prognostic value of spinal bone marrow MR imaging. Radiology 1998;209:653–60. [22] Wasser K, Moehler T, Nosas-Garcia S, Rehm C, Bartl R, Goldschmidt H, et al. [Correlation of MRI and histopathology of bone marrow in patients with multiple myeloma]. RöFo Fortschritte Auf Dem Geb Röntgenstrahlen Nukl 2005;177:1116–22, http://dx.doi.org/10.1055/s-2005-858362. [23] Schmidt GP, Baur A, Stäbler A, Schoenberg SO, Steinborn M, Baltin V, et al. [Estimation of diffuse bone marrow infiltration of the spine in multiple myeloma: correlation of MRT with histological results]. RöFo Fortschritte Auf Dem Geb Röntgenstrahlen Nukl 2005;177:745–50, http://dx.doi.org/10.1055/s-2005-857869.
