Waldenström macroglobulinemia/lymphoplasmacytic lymphoma (WM/LPL) is a lymphoplasmacytic lymphoma in which the bone marrow is infiltrated by immunoglobulin (Ig)M-producing clonal lymphoplasmacytic cells. Similar to other indolent lymphomas, treatment is indicated for patients with symptomatic WM/LPL, for example, systemic symptoms, systemic or bulky lymphadenopathy, anemia or thrombocytopenia, hyperviscosity, secondary amyloidosis, paraneoplastic neuropathy, etc. [1]. As for assessing treatment response of WM/LPL, consensus-based uniform response criteria were proposed by the International Working Group on WM in 2002 and were then revised in the subsequent International Workshops [2]. These uniform response criteria are based mainly on the change of serum monoclonal IgM protein level after treatment, since IgM serves as a good marker for tracking disease burden in a particular patient with WM/LPL [3, 4]. If there is significant lymphadenopathy or hepatosplenomegaly at baseline, a CT scan of the chest/abdomen is needed to determine the resolution of extramedullary disease.

18F-FDG PET/CT has been widely used for staging and response assessment of FDG-avid nodal lymphomas. The International Working Group criteria for reviewing post-treatment 18F-FDG PET scans are based on visual interpretation using the 5-point Deauville scale with mediastinal blood pool and liver as the comparator [5]. However, 18F-FDG PET/CT is not recommended in WM/LPL neither for staging or treatment response, as this indolent lymphoma is usually not FDG-avid [6], and classification of diffuse bone marrow involvement of lymphoma (which is the characteristic of WM/LPL) is difficult with 18F-FDG PET/CT. A study on the role of 18F-FDG PET/CT in WM/LPL showed a sensitivity of only 43% in detection of bone marrow involvement; moreover, this study found there was no statistical correlation between 18F-FDG PET/CT response and monoclonal protein response [7].

As chemokine receptor 4 (CXCR4), a key factor for tumor growth and metastasis, is overexpressed in the malignant B-cells of patients with WM/LPL at a high level [8,9,10], we previously conducted a prospective cohort study and reported that 68Ga-pentixafor, a CXCR4-targeted PET probe, was obviously more sensitive than 18F-FDG in detecting tumor involvement of WM/LPL (100% vs. 58.8%) [11,12,13]. We also found 68Ga-pentixafor is superior to 18F-FDG in determining disease response and residual disease after treatment [11, 14]. Recently, we reported the results of visual response assessment of 68Ga-pentixafor and 18F-FDG PET/CT based on 5-point scale in 15 patients with WM/LPL, and found that 68Ga-pentixafor PET-based response was consistent with clinical response categories, but 18F-FDG PET/CT missed the response in nearly half of the patients [15]. Based on the above evidence, we intend to further investigate if the quantitative tumor burden measurements on 68Ga-pentixafor PET/CT can be a precise and objective biomarker to assess the treatment response of WM/LPL, and 18F-FDG PET/CT was also analyzed as a reference.

Methods

Study design and patients

This is a retrospective analysis of the data from our prospective cohort study on the role of 68Ga-pentixafor PET/CT in WM/LPL approved by the Institutional Review Board of Peking Union Medical College Hospital (protocol ZS-1113) and registered at ClinicalTrials.gov (NCT 03436342). A total of 15 patients with newly diagnosed symptomatic WM/LPL at the Department of Hematology, Peking Union Medical College Hospital, were consecutively recruited from March 2018 to June 2020. Written informed consent was obtained from each patient. Laboratory tests and bone marrow evaluation for WM/LPL were done at enrollment. Patients were then referred for 18F-FDG and 68Ga-pentixafor PET/CT for baseline evaluation that were performed within 1 week after enrollment. Chemotherapy against WM/LPL was started within 2 weeks thereafter. After completion of chemotherapy, all the patients underwent follow-up 18F-FDG and 68Ga-pentixafor PET/CT. In the meantime, clinical response was evaluated according to the consensus response criteria on WM/LPL [3]. The response category was classified as complete response (CR), very good partial response (VGPR), partial response (PR), minimal response (MR), stable disease (SD), and progressive disease (PD) mainly based on monoclonal IgM protein level: CR is defined as absence of serum monoclonal IgM protein by immunofixation; VGPR is monoclonal IgM protein detectable ≥ 90% reduction in serum IgM level from baseline; PR is defined as monoclonal IgM protein detectable ≥ 50% but < 90% reduction in serum IgM level from baseline; MR means monoclonal IgM protein detectable ≥ 25% but < 50% reduction in serum IgM level from baseline; SD is monoclonal IgM protein detectable < 25% reduction and < 25% increase in serum IgM level from baseline; PD is defined as ≥ 25% increase in serum IgM level from lowest nadir.

PET/CT study

The DOTA-CPCR4-2 peptide was purchased from CSBio Co (CA 94025, USA). The radiolabeling of 68Ga-pentixafor was performed manually before injection according to the procedures as previously published. 18F-FDG was synthesized in house with an 11 MeV cyclotron (CTI RDS 111, Siemens, Germany). The PET scans were performed with dedicated PET/CT scanners (Biograph 64 Truepoint TrueV, Siemens, Germany; Polestar m660, SinoUnion, China) from the tip of the skull to the middle thigh. For 18F-FDG PET/CT, the patients fasted for at least 6 h, and the blood glucose levels were monitored (4.7–6.9 mmol/L) before an injection of 18F-FDG (5.55 MBq/kg). The PET/CT images (2 min/bed) were acquired with an uptake time of 75.0 ± 13.2 (mean ± SD) min. For 68Ga-pentixafor PET/CT, imaging was performed (2–4 min/bed) with an uptake time of 45.9 ± 19.7 min after an injection of 85.1 ± 27.4 MBq of 68Ga-pentixafor. The acquired data were reconstructed using the ordered-subset expectation maximization method (Biograph 64: 2 iterations, 8 subsets, Gaussian filter, image size of 168 × 168; Polestar m660: 2 iterations, 10 subsets, Gaussian filter, image size of 192 × 192).

Imaging analysis

Two experienced nuclear medicine physicians (YL and QP) visually assessed the PET/CT images and were in consensus for image interpretation. The presence and sites of tumor involvements, and the intensity of the uptake in the lesions were recorded. Semi-quantitative measurements of whole-body tumor burden were done in 68Ga-pentixafor and 18F-FDG PET/CT both at baseline and follow-up, measured as metabolic tumor volume (MTV, defined as the sum of the metabolic volumes of all tumors) and total lesions glycolysis/uptake (TLGFDG and TLUCXCR4, defined as the sum of individual MTV multiplied by its mean SUV) [16, 17]. PET/CT data were transferred in DICOM format to MIM workstation (version 6.6.11, MIM Software, USA). Then, a rectangular volume of interest was drawn including bone marrow and all focal lesions previously localized in the PET/CT images. Subsequently, tumor contours were first semiautomatically segmented with a SUV cutoff of 2.5. The contours were then checked and manually adjusted to exclude the physiological uptakes in heart, urinary tract, brain, vocal cords, liver, etc. Afterward, volumetric parameters of MTV, TLG FDG/TLUCXCR4, and SUVpeak were automatically obtained from the statistics generated with the final volumetric extraction.

Statistical analysis

The percentage of change in semi-quantitative PET/CT parameters between the baseline PET/CT (Upre) and post-treatment PET/CT (Upost) was described by ∆U% (∆U% = (Upre − Upost)/Upre × 100%, U referred to TLGFDG/TLUCXCR4, MTV and SUVpeak). The percentage of change in serum monoclonal protein (M-protein) and total serum IgM level between the baseline and the time of follow-up PET/CT was recorded as ∆M-pro% and ∆IgM%, respectively. The clinical response categories of CR, VGPR, PR, MR, SD and PD were assigned as score 1–6, respectively, for correlation analysis. The correlations between ∆U% and clinical response, ∆M-pro%, ∆IgM% were analyzed by Spearman’s rank correlation coefficients (for skewed data). A p value < 0.05 was considered statistically significant. Statistical analyses were done with the MedCalc software (version 19.6.4).

Results

Clinical characteristics

Fifteen patients with newly diagnosed symptomatic WM/LPL (12 men and 3 women; 60.9 ± 8.6 [range 48–76] y old) were analyzed in the present study. The median level of M-protein in recruited patients was 23.7 g/L (range 2.1–62.8 g/L), and the median level of β2-microglobulin was 5.5 mg/L (range 2.9–12.6 mg/L). According to the International Scoring System for WM (ISS-WM) [18], 3 patients were classified as being at high risk, 10 patients were classified as being at intermediate risk and one patient was classified as being at low risk. One patient with IgD LPL (patient 6) had unknown risk stratification because the serum M-protein and β2-microglobulin levels were not measured. Mutation of myeloid differentiation primary response 88, which has been identified in more than 90% of WM/LPL patients [19], was identified in all patients in the present study. Four patients (26.7%) were found to have a CXCR4 mutation.

All patients had bone marrow involvement confirmed by bone marrow aspiration and biopsy. Bone marrow involvement of WM/LPL showed markedly increased uptake of 68Ga-pentixafor (SUVmax 7.9 ± 2.5, range 5.1–14.8), but was less FDG-avid (SUVmax 3.4 ± 0.9, range 2.1–5.3). Besides bone marrow, WM/LPL involvement was seen in lymph nodes (12/15 patients, including cervical, subclavian, axillary, mediastinal, hepatic/splenic hila, peri-pancreas, para-aortic, bilateral iliac and inguinal nodes; short axis diameter of the largest lymph node, 20.8 ± 9.2 mm, range 11–36 mm), paramedullary space (3/15 patients, including soft tissues around the sternum, thoracic and lumbar vertebrae, and presacral space), liver (2/15), pancreas (1/15), and spleen (2/15). The median TLUCXCR4 and MTVCXCR4 of 68Ga-pentixafor PET/CT at baseline were 4036.3 (mean ± SD 5891.6 ± 5374.6; range 273.0–22,032.3) and 1189.1 (mean ± SD 1528.7 ± 1129.4; range 94.3–4480.8), respectively; and the median TLGFDG and TMVFDG were 672.9 (mean ± SD 672.9 ± 892.1; range 0–2902.6) and 232.4 (mean ± SD 232.4 ± 302.9; range 0–983.3), respectively. The baseline clinical characteristics and tumor involvement are summarized in Table 1.

Table 1 Patients’ clinical characteristics and treatment response
Full size table

Quantitative response assessment with PET/CT after chemotherapy

After baseline PET/CT, all of the 15 patients received chemotherapy, and follow-up PET/CT was performed after completion of treatment. The intervals between the last cycle of chemotherapy and the follow-up PET/CT were 2 weeks to 10 months (median 7 weeks). According to the consensus response criteria for WM/LPL [3], 5 patients achieved CR or VGPR after chemotherapy with a reduction of IgM and M-protein greater than 90% (range 90.4 to 100%); specifically, the 2 patients with CR also had a complete resolution of the lymph node involvement and other extramedullary disease (short axis diameter of the largest lymph node at baseline, 31 mm and 26 mm, respectively). Eight patients were evaluated as PR or MR with a reduction of IgM and M-protein less than 90% (range 32.1–85.7%), and 2 patients had PD with an increase in IgM and M-protein from the lowest nadir at follow-up (example in Fig. 1).

Fig. 1
figure 1

Examples of 3 patients with clinical CR, PR, and PD after chemotherapy showing different response in 68Ga-pentixafor and 18F-FDG PET/CT. CR complete response, PR partial response, PD progressive disease

Full size image

All of the 5 patients who achieved CR or VGPR after chemotherapy showed a consistent reduction of tumor burden measured in 68Ga-pentixafor PET/CT (median ∆TLUCXCR4%, 99.6%, mean ± SD 98.4% ± 2.5%, range 93.5–100%; median ∆MTVCXCR4%, 99.6%, mean ± SD 97.7% ± 3.8%, range 90.0–100%). Complete normalization of bone marrow uptake of 68Ga-pentixafor and complete resolution of extramedullary disease was also noted in 4 patients. The remaining one patient with clinical VGPR showed significantly reduced uptake of 68Ga-pentixafor in bone marrow, but residual lymph node disease avid for 68Ga-pentixafor was also noted, although the reduction of TLUCXCR4 and MTVCXCR4 was consistent with IgM response (∆M-pro% 100%, ∆IgM% 97.3%, ∆TLUCXCR4% 93.5%, ∆MTVCXCR4% 90.0%).

In the 8 patients with clinical PR or MR after chemotherapy, a significant reduction of TLUCXCR4 and MTVCXCR4 was noted (median ∆TLUCXCR4%, 90.5%, mean ± SD 86.2% ± 11.8%, range 65.8–99.8%; median ∆MTVCXCR4%, 89.4%, mean ± SD 85.4% ± 12.4%, range 62.3–99.8%), consistent with partial reduction of lesion uptake. In the 2 patients with clinical progression, one patient showed consistent increase in tumor burden measured in 68Ga-pentixafor PET/CT (∆TLUCXCR4%, − 56.0%; ∆MTVCXCR4%, − 44.2%). The other patient had a reduction of both TLUCXCR4 and MTVCXCR4 at follow-up (∆TLUCXCR4%, 48.4%; ∆MTVCXCR4%, 44.6%), but new lesions in the lymph nodes were detected in bilateral iliac regions.

In correlation analysis, ∆TLUCXCR4%, ∆MTVCXCR4% and ∆SUVpeak% in 68Ga-pentixafor PET/CT showed a significant direct correlation with clinical response, ∆M-pro%, and ∆IgM% values (p < 0.01, Table 2). However for 18F-FDG PET/CT, the uptake values (∆TLGFDG%, ∆MTVFDG%, and ∆SUVpeak%) were independent from clinical response, ∆M-pro%, and ∆IgM% (p > 0.05, Table 2). Semi-quantitative measurements of 68Ga-pentixafor and 18F-FDG PET/CT at baseline and follow-up were listed patient by patient in Additional file 1: Table S1.

Table 2 Spearman’s rank correlation test of semi-quantitative PET/CT-based response and clinical response
Full size table

Discussion

Our study determined that 68Ga-pentixafor PET/CT-based response, measured as ∆TLUCXCR4%, ∆MTVCXCR4%, ∆SUVpeak%, was well correlated with the clinical response and M-protein response in WM/LPL after chemotherapy. However 18F-FDG PET/CT-based response was independent from the clinical response. Therefore, it is suggested that the semi-quantitative measurements of whole-body tumor burden in 68Ga-pentixafor PET/CT might serve as a good biomarker for tracking disease burden in a particular patient with WM/LPL and for assessing treatment response.

Accurate discerning the depth of treatment response is important for stratifying patients and predicting outcomes. In WM/LPL, serum IgM M-protein levels is used as a determinant of disease response to therapy and to follow disease burden for an individual patient [2]. In our study, the 5 patients with CR or VGPR after chemotherapy showed a consistent reduction of tumor burden measured in 68Ga-pentixafor PET/CT (median ∆TLUCXCR4%, 99.6%, mean ± SD 98.4% ± 2.5%, range 93.5–100%); however, we noticed that 68Ga-pentixafor PET/CT did not distinguish VGPR from CR in 2 patients (68Ga-pentixafor PET/CT presented with a reduction over 99% of TLUCXCR4 and MTVCXCR4 from baseline in 4 patients who achieved CR or VGPR). Considering there was no difference in prognosis between patients with CR and with VGPR [20], we think it may not affect our further studies in spite of the incapability of 68Ga-pentixafor PET/CT to distinguish VGPR from CR.

For the determination of CR or VGPR, increased stringency is required as patients achieving VGPR or better response usually have improved progression free survival [20,21,22]. In addition to IgM response, CR or VGPR requires complete resolution of extramedullary disease (e.g., lymphadenopathy, splenomegaly) if present at baseline [3]. In the 4 patients with CR or VGPR and with a reduction over 99% of TLUCXCR4 and MTVCXCR4, there was complete normalization of bone marrow uptake of 68Ga-pentixafor and complete resolution of extramedullary disease. Interestingly, in the remaining one patient with clinical VGPR (patient 11, ∆TLUCXCR4% 93.5%, ∆MTVCXCR4% 90.0%), 68Ga-pentixafor PET/CT still detected residual lymph node disease (missed in 18F-FDG PET/CT), which was discordant with the stringent response criteria of VGPR, if using 68Ga-pentixafor PET/CT as the imaging modality to assess extramedullary disease. Considering such circumstances, we think the stringency for determination of a CR or VGPR state might be improved with the application of 68Ga-pentixafor PET/CT. However, it is not clear whether such change of treatment response categories has an impact on patients’ prognosis.

Clinical PD is defined as ≥ 25% increase in serum IgM level from lowest nadir and/or progression in clinical features attributable to the disease. As interim PET/CT was not performed during the course of chemotherapy in our study, the change of TLUCXCR4 or MTV CXCR4 at follow-up from baseline may possibly underestimate the disease progression. Therefore, visual assessment of 68Ga-pentixafor PET/CT is important for accurate interpretation of PD response. For example, in patient 15 with PD after chemotherapy, despite a nearly 50% reduction of TLUCXCR4 and MTV CXCR4 from baseline, there were new emerging lymph node diseases detected by both 18F-FDG and 68Ga-pentixafor PET/CT at follow-up, consistent with the clinical response classification of PD.

It is becoming clear that discrepancies can exist between monoclonal IgM protein responses and tumor reduction. For example, IgM responses are typically slow with monoclonal antibody-based therapy, as these agents selectively deplete the CD20+ B-cell component with sparing of the CD138+ plasma cell component [3]; transient increase in monoclonal IgM protein level (IgM flare) can occur following rituximab infusion [23]; in patients treated with bortezomib, tumor reduction in the bone marrow may not be proportional to the suppression of IgM levels [24]. Taking account of these cases, additional investigations are needed to confirm the disease response along with IgM levels, and to help with precise identification of the depth of response and early recognition of disease progression. Further studies are warranted whether 68Ga-pentixafor PET/CT could play such a role to supplement the current response criteria.

Our study had several limitations. First, it had a relatively small number of patients. However, we reported this preliminary result and found the possible significance of 68Ga-pentixafor PET/CT in response assessment of WM/LPL. Further study with a larger cohort is needed to confirm the results derived from this pilot study and to further evaluate its impact on predicting patients’ outcomes. Second, the time interval between the end of treatment and post-treatment PET/CT was variant, ranging from 2 weeks to 10 months. However, the evaluation of clinical response and laboratory tests were performed in the same period of PET/CT, so the analysis of PET/CT response was not biased. Considering the surface expression of CXCR4 is a dynamic process influenced by concomitant chemotherapy [25, 26], we did not perform interim PET/CT and scheduled the post-treatment PET/CT at least 2 weeks after completion of chemotherapy.

Conclusions

In the present study, we found that 68Ga-pentixafor PET/CT-based quantitative response after chemotherapy was well correlated with clinical response categories and monoclonal IgM protein level, a major determinant for disease response in patients with WM/LPL. Further investigation is warranted to evaluate the value of quantitative 68Ga-pentixafor PET/CT in patients’ prognosis and survival in a larger cohort of WM/LPL.