Secondary Logo

Journal Logo

Original Articles

Assessment of Central Nervous System Lymphoma Based on CXCR4 Expression In Vivo Using 68Ga-Pentixafor PET/MRI

Starzer, Angelika M. MD; Berghoff, Anna S. MD, PhD; Traub-Weidinger, Tatjana MD; Haug, Alexander R. MD; Widhalm, Georg MD, PhD; Hacker, Marcus MD; Rausch, Ivo PhD§; Preusser, Matthias MD; Mayerhoefer, Marius E. MD, PhD∥,¶

Author Information
doi: 10.1097/RLU.0000000000003404
  • Free


Central nervous system lymphoma (CNSL) is a rare form of non-Hodgkin lymphoma that can occur as a primary tumor (primary central nervous system lymphoma [PCNSL]; 3% of all brains tumors) or as a secondary manifestation of systemic lymphoma. Contrary to other brain tumors, surgical resection does not improve survival; however, histological confirmation by neurosurgical biopsy is commonly performed.1

18F-FDG PET/CT is the recommended technique for staging and response assessment in most lymphomas, particularly aggressive subtypes such as diffuse large B-cell lymphoma (DLBCL), because they show increased glucose metabolism.2 However, because 18F-FDG shows pronounced physiologic accumulation in the brain, 18F-FDGPET is not recommended for CNS lymphomas due to reduced tumor-to-background contrast. PET tracers currently investigatedfor CNS lymphomas include 18F-fludarabine and 11C-methionine,3,4 but no convincing clinical data exist. Thus, MRI currently remains the standard technique for preoperative assessment and follow-up of CNSL, despite its limited ability to distinguish between posttreatment pseudoprogression and true progression.1

Because previous data showed that lymphomas express high levels of the G-protein–coupled chemokine receptor CXCR4 (C-X-C chemokine receptor 4), 68Ga-pentixafor was recently developed as a PET tracer that specifically targets the CXCR4 receptor and has been applied to lymphoma, leukemia, and myeloma.5–7 Although 68Ga-pentixafor cannot penetrate the intact blood-brain barrier, the latter is impaired in patients with brain tumors.8 Recently, a single study in 11 patients suggested that 68Ga-pentixafor can be used for assessment of CNSL.9

Our present PET/MRI study aims to (1) provide more evidence on the accuracy of 68Ga-pentixafor PET for CNSL detection, (2) provide the first images demonstrating CNSL treatment response assessment with 68Ga-pentixafor PET, and (3) determine the relationship between quantitative 68Ga-pentixafor PET and MRI metrics.



Patients with histologically proven primary or secondary CNSL were included in this prospective proof-of-principle study. Approval by the local ethics committee (no. 1866/2019) and written informed patient consent were obtained. Exclusion criteria were as follows: age younger than 18 years, pregnancy, breast-feeding, and known contraindications to MRI. Patients were treated according to the Good Clinical Practice guidelines.

Image Acquisition

Patients underwent 68Ga-pentixafor PET/MRI on a simultaneous hybrid system (Biograph mMR; Siemens, Erlangen, Germany) at the earliest possible time point after presentation/diagnosis of CNSL. PET/MRI of the head was performed 60 minutes after IV administration of 150 MBq of 68Ga-pentixafor, with 30 minutes acquisition time, 3 iterations and 21 subsets, a 256 × 256 matrix, and a voxel size of 1.40 × 1.40 × 2.03 mm3, using the point-spread function-based reconstruction algorithm HD-PET. 68Ga-pentixafor was synthesized as previously described.6 Attenuation correction was done using the model-based approach incorporating spatially variant bone information into MR-based attenuation correction using an axial 2-point Dixon T1-weightedVIBE SPAIR 3D sequence, as implemented in the acquisition software version VE11P.

An axial RESOLVE diffusion-weighted imaging (DWI) sequence (b-50, b-1000) with apparent diffusion coefficient (ADC) mapping and an isotropic FLAIR sequence were obtained. An isotropic T1-weighted MPRAGE sequence was acquired before and after injection of extracellular Gd-based contrast media (CE-MRI).

Qualitative Image Analysis

A board-certified radiologist and a board-certified nuclear medicine physician evaluated all PET/MRIs in random order. The 68Ga-pentixafor PET and diagnostic MRI components were read separately, blinded to each other and to pathology and clinical reports. For PET, focal tracer accumulations within the brain that were visibly distinct from the surrounding background uptake were rated as PET positive and annotated. Similarly, lesions with contrast enhancement on CE-MRI were rated as positive and annotated; in case of subtle/equivocal enhancement, other MRI sequences were used for confirmation.

Quantitative Image Analysis

68Ga-pentixafor PET-based tumor volumes (PTV) were generated for each individual lesion. Because, contrary to 18F-FDG PET where a 41% SUVmax threshold is recommended by the European Association of Nuclear Medicine, no threshold is established for PTV generation on 68Ga-pentixafor PET, 3 thresholds (41%, 50%, and 70% SUVmax) were evaluated. PTV41%, PTV50%, and PTV70% were tested against morphological tumor volumes determined by manual lesion segmentation on CE-MRI (VOLMRI). SUVmean was calculated for each PTV, and also for a 2-cm cubic volume within the contralateral cerebral hemisphere (in a normal-appearing area on MRI), to calculate tumor-to-background ratios (TBRs: lesion SUVmax/contralateral SUVmean). Mean ADCs (ADCmean, ×10–6 mm2/s) were measured based on VOLMRI. All measurements were performed using LIFEx 6.0 (

Statistical Analyses

The accuracy of 68Ga-pentixafor PET was defined as the rate of agreement with MRI. Pearson correlation coefficients were used to determine relationships between PTVs and VOLMRI, and paired sample t tests were used to compare means. Correlations of SUVmean and TBR with ADCmean were also investigated. Because of our hypothesis-generating study design, no correction for multiple testing was applied. The specified significance level was P < 0.05. Statistical tests were performed using SPSS 24.0 (SPSS Inc, Chicago, IL).


Seven patients (4 women and 3 men; 54.9 ± 15.0 years) were enrolled: 5 with PCNSL and 2 with secondary CNSL (Table 1). Patient 1 underwent 5 68Ga-pentixafor PET/MRIs within 16 months; patient 3 underwent 68Ga-pentixafor PET/MRI twice. Therefore, 12 PET/MRI examinations were available for analysis.

TABLE 1 - Patient Characteristics
ID Sex/Age Tumor PET/MRI Scans Histological Diagnosis Therapy Regimen
1 M/49 PCNSL, 1 lesion (1) Baseline: post 2 cycles of CHT; (2) 8-wk follow-up: post 4 cycles of CHT; (3) 25-wk follow-up: post-WBRT after disease progression; (4) 35-wk follow-up: no therapeutical intervention since last follow-up; (5) 54-wk follow-up: no therapeutical intervention since last follow-up Primary CNS high-grade B-cell lymphoma CHT with high-dose MTX, WBRT10
2 F/56 PCNSL, no lesion (1) Baseline: 4 cycles of CHT; (2) 15-wk follow-up: post 5 cycles of CHT and autologous stem cell transplantation Primary CNS high-grade B-cell lymphoma CHT with high-dose MTX11
3 F/30 SCNSL, 2 lesions (1) Baseline at second recurrence: postsurgery + CHT at primary DLBCL manifestation, CHT at first recurrence, and 1 cycle of CHT at second recurrence; (2) 5-wk follow-up: post 2 cycles of CHT Secondary CNS high-grade B-cell lymphoma CHT with R-CHOP, high-dose MTX
4 M/65 PCNSL, no lesion (1) Baseline: post 1 cycle of CHT Primary CNS high-grade B-cell lymphoma CHT with high-dose MTX
5 F/54 SCNSL, 1 lesion (1) Baseline at first recurrence: post CHT at primary DLBCL manifestation, 1 cycle of CHT at first recurrence Secondary CNS high-grade B-cell lymphoma CHT with R-CHOP, high-dose MTX
6 F/51 PCNSL, 7 lesions (1) Baseline: pretreatment Primary CNS high-grade B-cell lymphoma CHT with high-dose MTX
7 M/79 PCNSL, 2 lesions (1) Baseline: post 1 cycle of CHT Primary CNS high-grade B-cell lymphoma CHT with high-dose MTX
M, male; F, female; CHT, chemotherapy; MTX, methotrexate; SCNSL, secondary central nervous system lymphoma.

68Ga-pentixafor PET demonstrated 18 focal CNSL lesions with increased tracer uptake in 5/7 patients (10/12 PET/MRI examinations), all of which were confirmed by CE-MRI (Fig. 1). The remaining 2 patients had undergone biopsy before imaging and showed only linear 68Ga-pentixafor uptake along the trepanation defects, as confirmed by MRI. No additional lesions without 68Ga-pentixafor uptake were visible on MRI; consequently, 68Ga-pentixafor PET accuracy was 100%.

A 54-year-old woman (no. 5) with secondary CNSL adjacent to the left interventricular foramen (arrows). The lesion shows marked focal tracer uptake on 68Ga-pentixafor PET and contrast enhancement on T1-weighted CE-MRI, as well as moderately hyperintense signal on the DWI and FLAIR images. On the ADC map, diffusivity is similar to that of the surrounding brain tissue.

Quantitative PET/MRI metrics are provided in Table 2. The t tests revealed no significant difference between VOLMRI and PTV41% (mean difference, −0.24 cm3; 95% confidence interval [CI], −0.91 to +0.42 cm3; P = 0.45), but significant differences between VOLMRI and PTV50% (mean difference, −1.36 cm3; 95% CI, −2.71 to −0.10 cm3; P = 0.048) and between VOLMRI and PTV70% (mean difference, −3.68 cm3; 95% CI, −7.0 to −0.37 cm3; P = 0.032). Correlations between ADCmean and SUVmean41% (r = 0.68), SUVmean50% (r = 0.68), and SUVmean70% (r = 0.68), and between ADCmean and TBR (r = 0.61) were moderate.

TABLE 2 - 68Ga-Pentixafor PET/MRI Metrics of Patients With PCNSL Lesions According to MRI
ID Scan Lesion SUVmax SUVmean 41% PTV 41% SUVmean 50% PTV 50% SUVmean 70% PTV 70% TBR VOLMRI ADCmean
1 1 1 4.5 2.85 5.57 3.12 4.19 3.7 1.86 34.62 7.47 204
1 2 1 5.93 3.95 30.97 4.19 26.15 4.68 14.16 49.42 33.66 818
1 3 1 1.44 0.88 1.73 0.98 1.21 1.17 0.48 8.00 3.568 151
1 4 1 1.59 0.96 1.36 1.07 0.94 1.3 0.35 7.57 2.965 94
1 5 1 1.42 0.82 1.32 0.94 0.93 1.16 0.31 9.47 0.738 218
3 1 1 5.18 3.27 2.89 3.56 2.19 4.22 0.92 25.90 2.42 792
3 1 2 4.95 3.14 2.57 3.43 1.95 4.07 0.86 24.75 1.89 729
3 2 1 6.49 4.11 6.08 4.5 4.57 5.3 2.01 34.16 4.48 757
5 1 1 2.1 1.29 0.595 1.42 0.43 1.73 0.160 14.00 0.43 424
6 1 1 9.83 6.32 31.52 6.77 25.31 7.9 10.60 51.74 34.14 820
6 1 2 5.5 3.38 2.08 3.75 1.47 4.5 0.59 28.95 2.23 799
6 1 3 3.86 2.38 1.42 2.6 1.05 3.14 0.39 20.32 1.29 562
6 1 4 2.75 1.63 1.3 1.84 0.87 2.24 0.31 14.47 1.03 712
6 1 5 2.29 1.38 1.7 1.51 1.24 1.84 0.42 12.05 2.12 687
6 1 6 4.95 3.03 1.77 3.33 1.28 4.04 0.49 26.05 1.55 442
6 1 7 1.5 0.91 0.91 0.99 0.66 1.23 0.22 7.89 0.93 566
7 1 1 0.46 0.27 1.53 0.3 1.15 0.37 0.36 2.56 0.23 322
7 1 2 0.39 0.25 1.81 0.27 1.45 0.32 0.72 2.17 0.35 289
Mean ± SD 3.62 ± 2.50 2.27 ± 1.62 5.40 ± 9.51 2.48 ± 1.74 4.28 ± 7.88 2.94 ± 2.01 1.96 ± 3.88 20.78 ± 14.87 5.64 ± 10.43 521.44 ± 256.18

In patient 1, changes on the 4 posttreatment 68Ga-pentixafor PET scans relative to the respective prior scan were +456.0% (consistent with clinical progression after 2 chemotherapy cycles), −94.4%, −21.4%, and −2.9% for PTV41% (consistent with response after chemotherapy completion and whole-brain radiotherapy [WBRT]), and +350.8%, −89.4%, −16.9%, and −75.1% for VOLMRI, respectively. In patient 3, follow-up PTV41% and VOLMRI increased by +110.4% and +85.1% for lesion 1, respectively, whereas the second lesion had resolved.


The results of our study, the second to apply 68Ga-pentixafor PET to CNSL, provide further evidence that CXCR4 imaging may have a role for assessment of this type of brain tumor. Similar to results by Herhaus et al,968Ga-pentixafor PET showed perfect accuracy for CNSL detection, with high lesion-to-background contrast, and without false-positive findings within the brain. Apart from tumor and physiologic uptake (eg, venous sinuses), only linear 68Ga-pentixafor uptake along the trepanation defects of the skull was noted, probably due to an inflammatory reaction.

Because there are no established criteria for generation of PTVs on 68Ga-pentixafor PET, we compared 3 relative SUVmax thresholds in analogy to those listed in the European Association of Nuclear Medicine guideline 1.0 for 18F-FDG PET-based metabolic tumor volumes: 41%, 50%, and 70%. Contrary to PTV50% and PTV70%, there was no significant difference between PTV41% and VOLMRI, suggesting that PTV41% reflects the actual CNSL volume on 68Ga-pentixafor PET. Although this means that, for pretherapeutic CNSL assessment, there is no rationale for using PTV41% over VOLMRI, this may possibly differ for treatment response assessment; in patients 1 and 3, changes on follow-up 68Ga-pentixafor PET/MRI were more pronounced for PTV41% than for VOLMRI. Also, in clinical practice, SUVs rather than PTVs are used for response assessment, and future studies need to focus on their clinical value to distinguish between radiation-induced inflammation and viable residual tumor.

Contrary to Herhaus et al,9 our PET/MRI protocol included DWI, a technique that relies on tissue diffusivity for signal generation. Diffusion-weighted imaging–based ADC measurements indirectly reflect tumor cell density6 and enable assessment of treatment-induced necrosis. One limitation of DWI is the sensitivity of ADC measurements to intralesional hemorrhage, which can occur in CNSL particularly after treatment. Contrary to previous findings in the bone marrow of chronic lymphocytic leukemia patients,6 we observed significant, moderate correlations between ADCmean and 68Ga-pentixafor SUVmean as well as TBR, suggesting a possible relationship between cellularity and CXCR4 receptor density in CNSL, but also suggesting that 68Ga-pentixafor PET and DWI provide complementary information.

CXCR4 is activated through its ligand CXCL12, which activates mitogen-activated protein kinase and phosphatidylinositol 3 kinase pathway.5 The CXCR4/CXCL12 axis mediates tumor cell migration.6 Novel therapies such as the Bruton’s tyrosine kinase inhibitor ibrutinib, the CXCR4 inhibitor plerixafor, and CXCR4-directed endoradiotherapy with 177Lu-pentixather are designed to target the CXCR4/CXCL12 pathway.12 A CXCR4-specific diagnostic imaging test therefore has high clinical potential, not just for imaging but also for treatment response assessment. Findings in patients 1 and 3 suggest that 68Ga-pentixafor PET may possibly also be useful for response assessment (Fig. 2), although more data are clearly needed.

A 49-year-old man (no. 1) with recurrent PCNSL of the right thalamus that shows high 68Ga-pentixafor uptake correlated to the pathological contrast enhancement on the T1-weighted CE-MRI. 68Ga-pentixafor PET-based metabolic tumor volumes (PTV41% shown) showed good concordance with morphologic tumor volumes (VOLMRI) both at baseline and after treatment demonstrating progression under chemotherapy with high-dose methotrexate and treatment response after WBRT.

In conclusion, our small 68Ga-pentixafor PET/MRI proof-of-principle study provides further evidence for the feasibility of this technique for CNSL lesion detection and first images of 68Ga-pentixafor PET-based response assessment. PET and MRI may possibly provide complementary information in CNSL.


The authors sincerely thank the PET/MRI staff, in particular Julia Kesselbacher and Benedikt Schmiedinger, for their technical help in conducting this study. This study was performed within the PhD thesis of Angelika M. Starzer with the title “Immune monitoring in cancer patients” in the N790 doctoral program at the Medical University of Vienna, Austria.


1. Hoang-Xuan K, Bessell E, Bromberg J, et al. Diagnosis and treatment of primary CNS lymphoma in immunocompetent patients: guidelines from the European Association for Neuro-Oncology. Lancet Oncol. 2015;16:e322–e332.
2. Cheson BD, Fisher RI, Barrington SF, et al. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. J Clin Oncol. 2014;32:3059–3067.
3. Kawase Y, Yamamoto Y, Kameyama R, et al. Comparison of 11C-methionine PET and 18F-FDG PET in patients with primary central nervous system lymphoma. Mol Imaging Biol. 2011;13:1284–1289.
4. Hovhannisyan N, Fillesoye F, Guillouet S, et al. [18F]Fludarabine-PET as a promising tool for differentiating CNS lymphoma and glioblastoma: comparative analysis with [18F]FDG in human xenograft models. Theranostics. 2018;8:4563–4573.
5. Haug AR, Leisser A, Wadsak W, et al. Prospective non-invasive evaluation of CXCR4 expression for the diagnosis of MALT lymphoma using [68Ga]Ga-Pentixafor-PET/MRI. Theranostics. 2019;9:3653–3658.
6. Mayerhoefer ME, Jaeger U, Staber P, et al. [68Ga]Ga-pentixafor PET/MRI for CXCR4 imaging of chronic lymphocytic leukemia: preliminary results. Invest Radiol. 2018;53:403–408.
7. Lapa C, Schreder M, Schirbel A, et al. [68Ga]Pentixafor-PET/CT for imaging of chemokine receptor CXCR4 expression in multiple myeloma—comparison to [18F]FDG and laboratory values. Theranostics. 2017;7:205–212.
8. Lapa C, Lückerath K, Kleinlein I, et al. 68Ga-Pentixafor-PET/CT for imaging of chemokine receptor 4 expression in glioblastoma. Theranostics. 2016;6:428–434.
9. Herhaus P, Lipkova J, Lammer F, et al. CXCR4-targeted positron emission tomography imaging of central nervous system B-cell lymphoma. J Nucl Med. 2020;jnumed.120.241703.
10. Ferreri AJ, Reni M, Foppoli M, et al; International Extranodal Lymphoma Study Group (IELSG). High-dose cytarabine plus high-dose methotrexate versus high-dose methotrexate alone in patients with primary CNS lymphoma: a randomised phase 2 trial. Lancet. 2009;374:1512–1520.
11. Ferreri AJ, Cwynarski K, Pulczynski E, et al. Chemoimmunotherapy with methotrexate, cytarabine, thiotepa, and rituximab (MATRix regimen) in patients with primary CNS lymphoma: results of the first randomisation of the International Extranodal Lymphoma Study Group-32 (IELSG32) phase 2 trial. Lancet Haematol. 2016;3:e217–e227.
12. Schottelius M, Osl T, Poschenrieder A, et al. [177Lu]pentixather: comprehensive preclinical characterization of a first CXCR4-directed endoradiotherapeutic agent. Theranostics. 2017;7:2350–2362.

central nervous system lymphoma; CXCR4; pentixafor; PET/MRI

Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.