Skeletal Radiol DOI 10.1007/s00256-014-1919-1

SCIENTIFIC ARTICLE

Utility of quantitative FDG-PET/CT for the detection of bone marrow involvement in follicular lymphoma: a histopathological correlation study Hugo J. A. Adams & Thomas C. Kwee & Rob Fijnheer & Stefan V. Dubois & Peter E. Blase & Rutger A. J. Nievelstein & John M. H. de Klerk

Received: 3 March 2014 / Revised: 30 April 2014 / Accepted: 14 May 2014 # ISS 2014

Abstract Objective To determine the value of visual and quantitative 18 F-fluoro-2-deoxy-D-glucose positron emission tomography/ computed tomography (FDG-PET/CT) for the detection of bone marrow involvement in follicular lymphoma, using direct histopathological examination at the right posterior iliac crest as reference standard. Materials and methods This retrospective study included 22 patients with newly diagnosed follicular lymphoma who had undergone FDG-PET/CT before BMB of the right posterior iliac crest. FDG-PET/CT images were visually evaluated for bone marrow involvement in the right posterior iliac crest. Volumes of interest were placed in the right posterior iliac crest to calculate the 3D partial volume corrected mean standardized uptake value (cSUVmean), maximum standardized uptake value (SUVmax) and peak standardized uptake value (SUVpeak). Results Sensitivity and specificity of visual FDG-PET/CT analysis for the detection of bone marrow involvement in the right posterior iliac crest were 0.0 % (95 % confidence H. J. A. Adams : T. C. Kwee (*) : R. A. J. Nievelstein Department of Radiology and Nuclear Medicine, University Medical Center Utrecht, Heidelberglaan 100, 3584 CX Utrecht, The Netherlands e-mail: [email protected] R. Fijnheer Department of Hematology, Meander Medical Center, Amersfoort, The Netherlands S. V. Dubois Department of Pathology, Meander Medical Center, Amersfoort, The Netherlands P. E. Blase : J. M. H. de Klerk Department of Nuclear Medicine, Meander Medical Center, Amersfoort, The Netherlands

interval (CI): 0–32.4 %) and 100 % (95 % CI: 78.5–100 %), respectively. Areas under the receiver-operating characteristic curve of cSUVmean, SUVmax and SUVpeak for the detection of bone marrow involvement in the right posterior iliac crest were 0.85 (95 % CI: 0.63–0.96), 0.89 (95 % CI: 0.68–0.98) and 0.87 (95 % CI: 0.65–0.97), respectively. Optimal cutoff values for cSUVmean, SUVmax and SUVpeak were 1.3, 2.1 and 1.7, and yielded sensitivity and specificity combinations of 75.0 % and 85.7 %, 87.5 % and 85.7 % and 87.5 % and 85.7 %, respectively. Conclusion This histopathological correlation study shows that, unlike visual interpretation of FDG-PET/CT images, quantitative FDG-PET/CT analysis may be beneficial in diagnosing bone marrow involvement by follicular lymphoma. Keywords Biopsy . Bone marrow . Diagnosis . FDG-PET/ CT . Follicular lymphoma Abbreviations cSUVmean 3D partial volume corrected mean SUV SUVmax maximum SUV SUVpeak peak SUV

Introduction Follicular lymphoma is the second most frequent nonHodgkin lymphoma subtype, and it accounts for about 1020 % of all lymphomas in western countries [1]. Bone marrow biopsy (BMB) is a standard staging procedure in follicular lymphoma [2], but it is an invasive procedure and prone to sampling errors. 18 F-fluoro-2-deoxy-D-glucose positron emission tomography/computed tomography (FDG-PET/ CT), which is strongly encouraged in follicular lymphoma to

Skeletal Radiol

facilitate the interpretation of posttreatment FDG-PET/CT examinations [3], may be a non-invasive alternative to BMB. Although 95 % of follicular lymphomas are FDG-avid [4], the role of FDG-PET/CT for the detection of bone marrow involvement in these patients is still unclear. A recent metaanalysis, which included both studies using stand-alone PET and studies using integrated PET/CT, reported FDG-PET to have pooled sensitivity and specificity of 46 % and 93 % in indolent lymphomas, but a separate analysis in follicular lymphoma was not performed [5]. One recent study in follicular lymphoma reported that visual FDG-PET/CT analysis had a sensitivity of only 31 % and a specificity of 92 % [6]. Interestingly, quantitative FDG-PET/CT analysis [standardized uptake value (SUV) measurements] resulted in an improved sensitivity and specificity of 83 % and 96 % [6]. One major limitation of that study, however, was the lack of a direct correlation analysis between FDG-PET/CT findings and histology. More scientific evidence is needed to clarify the role of FDG-PET/CT for bone marrow assessment in follicular lymphoma. The purpose of this study was to determine the value of visual and quantitative FDG-PET/CT for the detection of bone marrow involvement in follicular lymphoma, using direct histopathological examination at the right posterior iliac crest as reference standard.

(Biograph 40 TruePoint PET/CT, Siemens Healthcare). All patients ingested oral contrast material and fasted for 6 h before intravenous administration of 3 MBq/kg body weight of FDG. Blood glucose levels were checked to be less than 11 mmol/l (i.e., less than 198 mg/dl) before FDG administration. PET/CT images were acquired 60 min after FDG injection. First, low-dose, unenhanced CT images were acquired with the following settings: 120 kV, 26–30 mAs (automatic dose modulation), 0.8-s tube rotation time, pitch of 1.2 and 1.5-mm slice width (reconstructed to contiguous 5-mm axial slices to match the section thickness of the PET images). PET scanning from the mid femur to the base of the skull was done in five or six bed positions, with 3 min per bed position. Lowdose CT data were used for attenuation correction of the PET images, which were reconstructed with an orderedsubsets expectation maximization algorithm for 14 subsets and four iterations. The image reconstruction matrix was 128×128. Finally, all patients were administered an intravenous non-ionic iodinated contrast agent (Xenetix 300, Guerbet; 3 ml/s with bolus tracking; portal venous phase scanning) and underwent a full-dose CT scan of the neck, chest, abdomen and pelvis, with the following settings: 120 kV, 60–160 mAs (automatic dose modulation), 0.8-s tube rotation time, a pitch of 1.2 and 1.5mm slice width.

Materials and methods

Visual FDG-PET/CT analysis

Study design and patient population

An experienced reader (T.C.K.), who was blinded to the BMB results, evaluated all FDG-PET/CT images for bone marrow involvement, both at the right posterior iliac crest (which was the site of subsequent BMB) and at other locations. Bone marrow FDG uptake higher than liver FDG uptake (a criterion that is often used in Hodgkin’s lymphoma [7] and aggressive lymphoma [8]) or bone marrow FDG uptake equal or higher than FDG uptake of extramedullary lymphomatous lesions (i.e., any extramedullary lymphoid tissue or only lymphoid tissue that is believed to be pathologic at FDG-PET/CT) in the same patient (which takes into account the variable FDG avidity of follicular lymphoma [4]) was considered positive for bone marrow involvement. Concomitant CT images were used to localize foci of pathological FDG uptake. Diffusely increased bone marrow FDG uptake (i.e., homogeneously increased FDG uptake in a large volume of the axial bone marrow, exceeding liver FDG uptake, or equal to or higher than FDG uptake of extramedullary lymphomatous lesions in the same patient) was also considered positive for lymphomatous bone marrow involvement. Bone marrow involvement, if present, was classified as unifocal, multifocal, diffuse, or focal and diffuse (i.e., focal areas of disproportionately increased activity superimposed on generalized increased activity in the bone marrow).

The local institutional review board approved this retrospective single-center study, and written informed consent was waived. All patients with newly diagnosed follicular lymphoma routinely undergo pretreatment FDG-PET/CT and BMB at the right posterior iliac crest, and these patients were retrospectively identified in the hospital’s database between September 2007 and December 2013. Inclusion criteria were patients with newly diagnosed grade I-IIIA follicular lymphoma who had undergone both pretreatment FDG-PET/CT and BMB. Exclusion criteria were grade IIIB follicular lymphoma, previously treated/relapsed lymphoma, coexistence of another lymphoma subtype in the same diagnostic biopsy, another previous malignancy within the past 5 years, patients in whom BMB was done before FDG-PET/CT, interval between FDG-PET/CT and BMB >30 days, non-diagnostic BMB, and start of therapy (chemotherapy, radiation therapy) and/or hematopoietic growth factor injections before FDG-PET/CT or BMB. FDG-PET/CT acquisition FDG-PET/CT imaging was performed using an integrated EARL-accredited 40-detector row PET/CT system

Skeletal Radiol

Quantitative FDG-PET/CT analysis Another observer (H.J.A.A.), who was also blinded to BMB results, used the ROVER (region of interest visualization, evaluation and image registration) software package (ABX Advanced Biochemical Compounds GmbH, Radeberg, Germany) for quantitative FDG-PET/CT analysis of the right posterior iliac crest (which was the site of subsequent BMB). A spherical volume of interest with a radius of 1 cm was placed in the right posterior iliac crest (Fig. 1). The software then automatically calculated 3D partial volume corrected mean SUV (cSUVmean), maximum SUV (SUVmax) and peak SUV (SUVpeak). BMB Unilateral BMB of the right posterior iliac crest was performed in all patients by different hematologists as part of standard clinical care. Obtained bone marrow samples were evaluated by different hematopathologists. All hematologists and hematopathologists were unaware of FDG-PET/CT findings at the time of BMB/BMB interpretation. Positive biopsies were reevaluated by an experienced hematopathologist (S.V.D.) who assessed the percentage of tumor involvement in the biopsy. Statistical analysis Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of visual FDG-PET/CT analysis for the detection of lymphomatous bone marrow involvement in the right posterior iliac crest were calculated, along with binomial exact 95 % confidence intervals (CIs), using BMB as reference standard. Patient-based sensitivity of visual FDG-PET/CT analysis for the detection of bone marrow involvement was also calculated, again using BMB as reference standard. Receiver-operating characteristic (ROC) analysis was performed to calculate the areas under the ROC (AUCs) of cSUVmean, SUVmax and SUVpeak for the detection of bone marrow involvement in the right posterior iliac crest, along

Fig. 1 Example of volume of interest placement (spherical volume with 1 cm radius) in the right posterior iliac crest

with 95 % confidence intervals, using BMB as reference standard. Optimal cutoff values for cSUVmean, SUVmax and SUVpeak were calculated, along with corresponding sensitivities and specificities. Statistical analyses were executed using MedCalc statistical software, version 12.6.0 (Ostend, Belgium).

Results Patient characteristics Between September 2007 and December 2013, 54 patients with newly diagnosed follicular lymphoma underwent both FDG-PET/CT and BMB. Of these, 13 patients were excluded because BMB was performed before FDG-PET/CT, 7 patients were excluded because of the coexistence of another lymphoma in the diagnostic biopsy, 7 patients were excluded because of the presence of another malignancy, 3 patients were excluded because of grade IIIB follicular lymphoma, 1 patient was excluded because the time interval between BMB and FDG-PET/CT exceeded 30 days, and 1 patient was excluded because of a non-diagnostic BMB. Thus, 22 patients (10 males and 12 females, mean age: 63.2 years, age range: 43– 86 years) were finally included. All included patients had FDG-avid follicular lymphoma. The mean time interval between FDG-PET/CT and BMB was 12.7 days (SD: 9.5 days, range: 0–30 days). Detailed patient characteristics are shown in Table 1. Diagnostic performance of visual FDG-PET/CT analysis Visual FDG-PET/CT analysis was positive for bone marrow involvement in the right posterior iliac crest in 0 of 22 patients (0 %), whereas BMB was positive in 8 (36.4 %) of 22 patients (Fig. 2). In these eight patients, the mean percentage of malignant cells in the entire bone marrow sample was 12.1 % (SD 19.5 %, range 2–60 %). FDG-PET/CT was positive for bone marrow involvement outside the right posterior iliac crest in 10 of 22 patients (45.5 %). Unifocal, multifocal, diffuse and a combination of focal and diffuse bone marrow involvement at FDG-PET/CT was identified in 3 (13.6 %), 6 (27.3 %), 0 (0 %) and 1 (4.6 %) of 22 patients, respectively. Note that the pathological FDG uptake in the single patient with both focally and diffusely increased bone marrow FDG uptake was limited to the ribs, cervical, thoracic and lumbar vertebra, and sacrum (the iliac bone marrow, including the right posterior iliac crest, did not demonstrate any pathological FDG uptake). Sensitivity and specificity of visual FDG-PET/ CT analysis for the detection of bone marrow involvement in the right posterior iliac crest were 0 % (95 % CI: 0–32.4 %) and 100 % (95 % CI: 78.5–100 %), respectively. PPV could

Skeletal Radiol Table 1 Patient characteristics

Patients (no.) Male/female Age Mean ± SD (years) Median (years) Time interval between BMB and FDG-PET/CT Mean (days) Range (days) Visual FDG-PET/CT analysis Visual iliac crest FDG-PET/CT+(no.) Visual whole-body FDG-PET/CT+(no.) Quantitative PET/CT analysis Mean cSUVmean ±SD

BMB+

BMB-

8 3/5

14 7/7

61.8±12.3 64.0±9.1 63.5 63.5 12.0 0–24

13.1 3–30

0 4

0 6

Range cSUVmean Mean SUVmax ±SD Range SUVmax Mean SUVpeak ±SD

1.7±0.4 0.9–2.4 2.5±0.54 1.4–3.0 2.0±0.4

1.1±0.3 0.6–1.8 1.7±0.4 1.1–2.4 1.4±0.4

Range SUVpeak

1.2–2.5

0.9–2.1

− negative (BMB or FDG-PET/CT bone marrow) result + positive (BMB or FDG-PET/CT bone marrow) result

not be calculated because of the absence of true- and falsepositive cases, whereas the NPV was 63.6 % (95 % CI: 43.0–

80.3 %). Patient-based sensitivity of visual FDG-PET/CT analysis for the detection of bone marrow involvement was 50 % (95 % CI: 21.5–78.5 %). Diagnostic performance quantitative FDG-PET/CT analysis Mean cSUVmean, SUVmax and SUVpeak in the right posterior iliac crest were 1.3 (SD: 0.5, range: 0.6–2.4), 2.0 (SD: 0.6, range: 1.1–3.0) and 1.6 (SD: 0.5, range: 0.9–2.5), respectively. AUCs of cSUVmean, SUVmax and SUVpeak for the detection of bone marrow involvement in the right posterior iliac crest were 0.85 (95 % CI: 0.63–0.96), 0.89 (95 % CI: 0.68–0.98) and 0.87 (95 % CI: 0.65–0.97), respectively (Fig. 3). Optimal cutoff values for cSUVmean, SUVmax and SUVpeak were 1.3, 2.1 and 1.7, and yielded sensitivity and specificity combinations of 75.0 % and 85.7 %, 87.5 % and 85.7 % and 87.5 % and 85.7 %, respectively.

Discussion The results of this histopathological correlation study at the right posterior iliac crest show that FDG-PET/CT has limited value in the detection of bone marrow involvement in follicular lymphoma if a visual analysis is performed. It should also be noted that patient-based sensitivity of FDG-PET/CT was

a

b

c

d

Fig. 2 Example of FDG-PET/CT and BMB of the right posterior iliac crest in a 65-year-old woman with follicular lymphoma. Axial FDG-PET/ CT (a) is visually negative for bone marrow involvement at the right posterior iliac crest. FDG-PET/CT was also visually negative for bone marrow involvement at other locations (not shown). cSUVmean, SUVmax and SUVpeak in the right posterior iliac crest were 1.7, 2.8 and 2.3, respectively. Bone marrow trephine biopsy of the right posterior

iliac crest shows 30 % cellularity, consistent with the patient’s age (b, H&E, original magnification 25×). Detail showing paratrabecular localization of a small cell lymphoma, encompassing 20 % of the bone marrow cellularity (c, H&E, original magnification 100×). Same detailed area, staining positively with the pan B-cell marker CD20 (d, original magnification 100×)

Skeletal Radiol Fig. 3 ROC curves of cSUVmean (a), SUVmax (b) and SUVpeak (c) for the detection of bone marrow involvement in the right posterior iliac crest. AUCs of cSUVmean, SUVmax and SUVpeak were 0.85 (95 % CI: 0.63-0.96), 0.89 (95 % CI: 0.68-0.98) and 0.87 (95 % CI: 0.65-0.97), respectively

only 50 % when using BMB as reference standard. In contrast, the diagnostic performance of quantitative FDG-PET/CT analysis in this setting is higher, particularly when using SUVmax or SUVpeak. The poor visibility of bone marrow involvement in follicular lymphoma at FDG-PET/CT may be explained by the fact that approximately 76 % of follicular lymphomas show a paratrabecular pattern of bone marrow involvement (i.e., lymphoid aggregates with broad bases aligned alongside bony trabecula without intervening fat cells) [9], whereas diffuse bone marrow involvement (i.e., areas of effacement by a dense infiltrate of lymphocytes not forming nodules) and nodular bone marrow involvement (i.e., circumscribed aggregates of lymphoid cells) are less common in this entity [9]. In addition, 41.0 % and 26.1 % of follicular lymphomas have been reported to demonstrate less than 10 % and 10-30 % of tumor involvement in the BMB, respectively [9], which is in line with the tumor percentages in the positive BMBs that were found in the present study (mean: 12.1 %). Quantitative FDG-PET/CT analysis may allow for the detection of follicular lymphoma deposits in the bone marrow that can be missed visually because of the aforementioned reasons. Most previous studies on the utility of FDG-PET/CT for the detection of bone marrow involvement in lymphoma included a heterogeneous group of either aggressive or indolent lymphomas [5] rather than focusing on a specific lymphoma subtype such as follicular lymphoma. In contrast, one recent retrospective study by El-Najjar et al. [6] investigated the value of FDG-PET/CT-based bone marrow assessment in 41 patients with grade I-IIIA follicular lymphoma. In that study, FDG-PET/CT images were visually and quantitatively evaluated for bone marrow involvement. For quantitative analysis, patients with focal bone marrow disease visually identified at FDG-PET/CT were excluded, and SUVmax was measured at the sternum, at both iliac blades and at the 12th thoracic vertebra. An average of these four measurements was recorded as SUVav. The single highest overall SUVmax for the four bone sites, the SUVav and the ratios SUVav/mediastinal blood pool (MBP) and SUVav/liver were compared with the BMB result. Focal bone marrow uptake was identified at FDG-PET/CT by visual analysis in six patients, including two cases in which the BMB was negative. Assessment of diffuse bone marrow involvement at FDG-PET/CT by visual

analysis had a sensitivity and specificity of 31 and 92 %, respectively. Quantitative analysis resulted in an improved sensitivity and specificity of 58 % and 96 %, respectively, when using SUVav greater than or equal to 2 as the cutoff. Using the ratio SUVav/MBP greater than or equal to 1, the sensitivity of FDG-PET/CT to detect bone marrow involvement improved to 83 %. El-Najjar et al. [6] concluded that visual analysis is useful in determining focal bone marrow involvement, whereas quantitative analysis using SUVav/ MBP has a high sensitivity and specificity for predicting bone marrow involvement in patients lacking focal bone lesions. Besides the fact that El-Najjar et al.’s study also included patients with previously treated lymphoma (which may cause persistent metabolic changes in the bone marrow), its main limitation was the lack of direct correlation between FDGPET/CT findings and histopathology [6]. The present study only included patients with newly diagnosed follicular lymphoma who had undergone BMB after FDG-PET/CT, thus allowing for such an analysis. Nevertheless, the results of the present study are in line with El-Najjar et al.’s findings [6], in that visual interpretation of FDG-PET/CT images has limited value for the detection of bone marrow involvement in follicular lymphoma, but quantitative FDG-PET/CT analysis has a considerably higher diagnostic value. Magnetic resonance imaging (MRI) is another imaging modality that may be used to evaluate the bone marrow in lymphoma. Of interest, a recent study reported that, using BMB as reference standard, patient-based sensitivity of whole-body MRI in indolent lymphoma was only 12.5 % and significantly lower (P=0.026) than the 83.3 % sensitivity in aggressive lymphoma [10]. In addition, whole-body MRI detected only three of eight follicular lymphoma cases with a positive BMB included in that study [10]. Thus, although PET/MRI systems will increasingly be used in clinical practice, it is questionable whether the MRI component will improve the diagnostic performance of (quantitative) FDG-PET alone in this combination. The present study has several limitations. First, the number of enrolled patients was relatively small, but this was due to the fact that several patients who had undergone BMB before FDG-PET/CT had to be excluded in order to achieve direct histopathological correlation in all patients. Second, this study

Skeletal Radiol

focused on the correlation between FDG-PET/CT and BMB findings in the right posterior iliac crest (and calculated estimates of diagnostic performance at this local level accordingly), because FDG-PET/CT findings in other bone marrow sites could not be verified histologically. However, it would be practically and ethically impossible to obtain biopsies from all (FDG-PET/CT-positive) bone marrow locations. In addition, the course of FDG-PET/CT-positive bone marrow lesions outside the posterior iliac crest after treatment could not be evaluated, because follow-up FDG-PET/CT studies were not available in most patients. Third, BMBs were not imageguided, and the possibility of a mismatch between blind BMB and quantitative FDG-PET/CT measurements in some of the patients cannot be excluded. Fourth, it should be realized that the optimal cutoff values for cSUVmean, SUVmax and SUVpeak were retrospectively obtained and that their diagnostic yield may be lower when applied prospectively in another patient population. Fifth, the patient number was too low to assess which of the quantitative FDG-PET/CT metrics was diagnostically superior. Sixth, the prognostic implications of FDGPET/CT and BMB findings were not assessed. In conclusion, this histopathological correlation study shows that, unlike visual interpretation of FDG-PET/CT images, quantitative FDG-PET/CT analysis may be beneficial in diagnosing bone marrow involvement by follicular lymphoma. However, larger prospective studies are needed to confirm these findings. Funding This project was financially supported by an Alpe d’HuZes/ Dutch Cancer Society Bas Mulder Award for T.C.K. (grant no. 5409) and by a ZonMW AGIKO stipend for T.C.K. (grant no. 92003497). Data collection, data analysis and interpretation of data, writing of the paper and the decision to submit were left to the authors’ discretion and were not influenced by Alpe d’HuZes/Dutch Cancer Society and ZonMW. Potential conflicts of interest None.

References 1. Ghielmini M, Vitolo U, Kimby E, Montoto S, Walewski J, Pfreundschuh M, et al. ESMO Guidelines consensus conference on malignant lymphoma 2011 part 1: diffuse large B-cell lymphoma (DLBCL), follicular lymphoma (FL) and chronic lymphocytic leukemia (CLL). Ann Oncol. 2013;24(3):561–76. 2. Dreyling M, Ghielmini M, Marcus R, Salles G, Vitolo U. Newly diagnosed and relapsed follicular lymphoma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2011;22 Suppl 6:vi59–63. 3. Cheson BD, Pfistner B, Juweid ME, Gascoyne RD, Specht L, Horning SJ, et al. Revised response criteria for malignant lymphoma. J Clin Oncol. 2007;25(5):579–86. 4. Weiler-Sagie M, Bushelev O, Epelbaum R, Dann EJ, Haim N, Avivi I, et al. (18)F-FDG avidity in lymphoma readdressed: a study of 766 patients. J Nucl Med. 2010;51(1):25–30. 5. Chen YK, Yeh CL, Tsui CC, Liang JA, Chen JH, Kao CH. F-18 FDG PET for evaluation of bone marrow involvement in non-Hodgkin lymphoma: a meta-analysis. Clin Nucl Med. 2011;36(7):553–9. 6. El-Najjar I, Montoto S, Mc-Dowell A, Matthews J, Gribben J, Szyszko TA. The value of semiquantitative analysis in identifying diffuse bone marrow involvement in follicular lymphoma. Nucl Med Commun. 2013;35(3):311–5. 7. Adams HJ, Kwee TC, de Keizer B, Fijnheer R, de Klerk JM, Littooij AS, et al. Systematic review and meta-analysis on the diagnostic performance of FDG-PET/CT in detecting bone marrow involvement in newly diagnosed Hodgkin lymphoma: is bone marrow biopsy still necessary? Ann Oncol. 2014. 8. Adams HJ, Kwee TC, de Keizer B, Fijnheer R, de Klerk JM, Nievelstein RA. FDG PET/CT for the detection of bone marrow involvement in diffuse large B-cell lymphoma: systematic review and meta-analysis. Eur J Nucl Med Mol Imaging. 2014;41(3):565– 74. 9. Sovani V, Harvey C, Haynes AP, McMillan AK, Clark DM, O’Connor SR. Bone marrow trephine biopsy involvement by lymphoma: review of histopathological features in 511 specimens and correlation with diagnostic biopsy, aspirate and peripheral blood findings. J Clin Pathol. 2013. doi:10.1136/jclinpath-2013-201520. 10. Adams HJ, Kwee TC, Vermoolen MA, de Keizer B, de Klerk JM, Adam JA, et al. Whole-body MRI for the detection of bone marrow involvement in lymphoma: prospective study in 116 patients and comparison with FDG-PET. Eur Radiol. 2013;23(8):2271–8.