Comparison of 11C-4′-thiothymidine, 11C-methionine, and 18F-FDG PET/CT for the detection of active lesions of multiple myeloma

Purpose The aims of this study were to evaluate the possibility of using 11C-methionine (11C-MET) and 11C-4′-thiothymidine (11C-4DST) whole-body PET/CT for the imaging of amino acid metabolism and DNA synthesis, respectively, when searching for bone marrow involvement in patients with multiple myeloma (MM) and to compare these findings with those for 18F-FDG PET/CT and aspiration cytology. Methods A total of 64 patients with MM, solitary plasmacytoma, monoclonal gammopathy of undetermined significance, or an unspecified diagnosis were prospectively enrolled. All the patients underwent three whole-body PET/CT examinations within a period of 1 week. First, the tracer accumulation was visually evaluated as positive, equivocal, or negative for 55 focal lytic lesions visualized using CT in 24 patients. Second, the percentages of marrow plasma cells as calculated using a bone marrow aspiration smear and tracer accumulation were evaluated in the posterior iliac crests of 36 patients. Results Among the 55 lytic lesions, the 11C-MET and 11C-4DST findings tended to reveal more positive findings than the 18F-FDG findings. Based on the standard criteria for the diagnosis of active myeloma using the percentage of marrow plasma cells, significant differences were found between the 18F-FDG and 11C-MET findings and between the 18F-FDG and 11C-4DST findings, but no significant difference was observed between the 11C-MET and 11C-4DST findings. Conclusion The addition of 11C-MET and 11C-4DST to 18F-FDG when performing PET/CT enabled clearer evaluations of equivocal lesions. Based on cytological diagnostic criteria, 11C-MET and 11C-4DST were more sensitive than 18F-FDG for the detection of active lesions. 11C-MET and 11C-4DST were more useful than 18F-FDG for the detection of active lesions, especially during the early stage of disease.


Introduction
Multiple myeloma (MM) is characterized by the neoplastic proliferation of plasma cells [1], and the majority of myeloma cells produce monoclonal immunoglobulins or immunoglobulin-related proteins and various proteins causing complications. MM may present with both osseous and extraosseous manifestations and accounts for 1 % of all malignant diseases [2] and about 10 % of hematologic malignancies [3]. Symptoms develop as a result of anemia, immunosuppression, renal failure, hypercalcemia, and bone destruction resulting in frequent pathologic fractures [4]. MM typically evolves from a premalignant condition called monoclonal gammopathy of undetermined significance (MGUS), with M (monoclonal)-protein detected in blood or urine. Typically, end-stage organ damage does not occur in MGUS or in a more progressed condition called smoldering MM. MGUS and smoldering MM are usually not treated [5,6]. The new Durie/Salmon PLUS staging system has integrated new imaging techniques into a new generation of anatomic and functional myeloma staging [7].
Morphologic imaging techniques, such as X-ray, computed tomography (CT), and magnetic resonance imaging (MRI), show the extent of tumors but not their activity or their viability; thus, these techniques have limitations when assessing the treatment response or early progression. 18 F-FDG PET/CT is a non-invasive, whole-body imaging method that has been widely used to detect malignant tissues and to monitor treatment response in patients with solid tumors and lymphoma [8,9]. In addition, the technique is useful for examining the function of the red marrow and for detecting bone marrow involvement in both benign and malignant disorders. Consequently, 18 F-FDG PET/CT is well recognized as a powerful diagnostic tool for the initial staging of patients with MM [10]. In addition, 18 F-FDG PET/CT is also useful for evaluating the response to therapy [11], such as the restaging of MM [12].
However, 18 F-FDG accumulation in the areas of inflammation or infection may obscure accurate evaluations of the therapeutic effects on tumor tissues [13]; hence, false-negative results have been reported for 18 F-FDG, especially in patients with early stage MM [11]. The use of new tracers capable of compensating for the limitations of 18 F-FDG could be very helpful for detecting active myeloma lesions more accurately than with 18 F-FDG alone.
The use of 11 C-methionine ( 11 C-MET) and 18 F-fluorodeoxy-L-thymidine ( 18 F-FLT) has been reported for the evaluation of MM. 18 F-FDG, 11 C-MET, and 18 F-FLT have different mechanisms of tracer uptake, enabling the visualization of the metabolic status of glucose, amino acid metabolism, and proliferative activity, respectively, in bone marrow and extramedullary lesions [14]. 11 C-MET is a widely used tracer for the imaging of brain tumors [15]. An increase in 11 C-MET uptake in the bone marrow reportedly reflects an increase in cellular proliferation and protein synthesis [16]. MM is characterized by the neoplastic proliferation of plasma cells and the production of monoclonal immunoglobulins. Dankerl et al. reported that the increase in methionine uptake in plasma cells is the basis for the imaging of active MM using 11 C-MET PET/CT, and peripheral bone marrow expansion has been observed in MM patients. Extramedullary MM can also be detected and localized with a high sensitivity using 11 C-MET PET/CT [16]. 18 F-FLT is a tracer that monitors the activity of thymidine kinase one and its uptake and hence is related to DNA synthesis, which is a surrogate marker for cellular proliferation. Therefore, 18 F-FLT may help in differentiating old and inactive lytic lesions from foci of rapidly proliferating MM cells, which could be potential targets for local radiation treatment [14].
Alternatively, Toyohara et al. developed [methyl-11 C] 4 0thiothymidine ( 11 C-4DST) as a novel tracer for imaging cell proliferation. 11 C-4DST is a promising tracer because after it has been incorporated into DNA, the occurrence of labeled nucleotide dephosphorylation (which can be an issue with 18 F-FLT) is relatively rare [17]. This irreversible nature of 11 C-4DST is expected to contribute to a more sensitive tumor uptake than 18 F-FLT. Indeed, the tumor uptake of 11 C-4DST was higher than that of 3 H-FLT and was correlated with the DNA synthesis level in animal models [18]. Initial clinical trials of 11 C-4DST have demonstrated its safety, radiation dosimetry, and application for brain tumor imaging [19].
Furthermore, Minamimoto et al. applied 11 C-4DST PET/CT to proliferation imaging in non-small cell lung cancer and demonstrated a strong correlation between 11 C-4DST uptake by tumor tissues before surgery and the MIB-1 index of the surgical pathology findings (a standard marker of proliferation). A linear regression analysis indicated that the SUVmax for 11 C-4DST was not significantly correlated with the microvessel density, as determined using CD31 staining [20]. Thus, 11 C-4DST and blood flow are not correlated, whereas 11 C-4DST is correlated with cell proliferation. Based on these conclusions, we hypothesized that 11 C-4DST might be capable of detecting the proliferation status of MM more sensitively than 18 F-FLT. Hence, we selected 11 C-4DST as a representative marker of in vivo proliferation.
The aim of the present study was to evaluate the potential of whole-body 11 C-4DST and 11 C-MET PET/CT imaging for the detection of bone marrow involvement in patients with MM and to compare the results with those obtained using 18 F-FDG.

Patients
Between October 2010 and October 2011, a total of 64 patients with MM or MGUS (40 men, 24 women; mean age, 58.3 years; range 33-84 years) were prospectively enrolled in this study (Table 1). All the patients were diagnosed as having MM or MGUS based on the criteria defined by the International Myeloma Working Group. Twenty-one patients were previously untreated, and 43 patients were restaged after treatment. All the patients underwent three whole-body PET-CT scans with 18 F-FDG, 11 C-MET, and 11 C-4DST within a period of 1 week.
MM is not a solid tumor but a hematological malignancy that appears as a mottled lesion distributed in the bone marrow. Consequently, biopsies of MM lesions are difficult to perform, and the tumor range can be difficult to discriminate. Thus, methods for evaluating MM require special consideration.
To compare lesion visualization, we focused on focal lytic lesions visible using CT for which the lesion localization was obvious and the tracer uptake by the lesions in the three PET studies could be easily and accurately compared. Twenty-four patients who had focal lytic lesions visible using CT were enrolled in Study 1. The remaining patients did not have focal lytic lesions, but instead had diffuse lesions or normal findings when examined using CT.
To verify tracer uptake by the lesions, we focused on lesions for which bone marrow aspiration cytology was performed. Thirty-six patients underwent bone marrow aspiration cytology within 1 week of the three PET/CT scans (Study 2). The remaining patients did not undergo bone marrow aspiration cytology within 1 week of the PET/CT scans. Fifteen patients were not enrolled in either study (Fig. 1).
The present study was approved by the institutional review board, and written informed consent was obtained from all the subjects.
PET/CT examination An in-house cyclotron and automated synthesis system (F200 Sumitomo Heavy Industries, Tokyo, Japan) produced the 18 F-FDG. 11 C-MET and 11 C-4DST were synthesized as reported previously [19,21]. The PET/CT images were obtained using two PET/CT systems (29 patients: Biograph 16; Siemens, München, Germany; and 35 patients: Discovery PET/CT 600; GE Healthcare, Fairfield, CF), with measurements obtained from the vertex  . The SUV consistency for these two PET/CT scanners was validated using a phantom study. Low-dose CT was performed first for attenuation correction and image fusion. Emission images were acquired in a threedimensional mode for 2 min per bed position using both PET/CT scanners. The PET/CT data were reconstructed using a Gaussian filter with an ordered subset expectation maximization algorithm (3 iterations, 8 subsets for Biograph 16, 3 and 16 subsets for Discovery PET/CT 600, according to the manufacturers' recommendations). The radiation exposure from 11 C-labeled radiopharmaceuticals is much lower than that from 18 F-labeled radiopharmaceuticals. The estimated effective dose for 11 C-4DST is 1.6 mSv [19], that for 11 C-MET is 2.1 mSv, that for 18 F-FDG is 7 mSv [22], and that for low-dose CT is 1.4-3.5 mSv. Therefore, the total effective dose delivered during all the PET/CT examinations was about 14.9-21.2 mSv. We felt that the radiation exposure from the PET/CT examinations performed in the present study was acceptable when their impact on the therapeutic strategy was considered.
PET data analysis Xeleris (GE Healthcare) and e-Soft (Siemens) workstations were used for image analysis. The physiological uptake of 11 C-MET is seen in the gastrointestinal tract, liver, pancreas, urinary tract, and salivary glands, as reported by Nishizawa et al. [15]. A high physiological 11 C-4DST uptake is observed in the salivary glands, liver, spleen, kidneys, bladder, and bone marrow. In contrast, the brain, lungs, myocardium, muscle, and blood pool exhibit a low physiological 11 C-4DST uptake [20]. The SUVmax of 11 C-4DST in the normal bone marrow is higher than that of 18 F-FDG (lumbar vertebrae 2-4, ilium, proximal humeri, and proximal femurs), and the SUVmax of 11 C-MET falls between these two values. Based on this normal background, the active accumulations of 18 F-FDG, 11 C-MET, and 11 C-4DST in the bone marrow lesions were evaluated.
Two experienced nuclear medicine physicians visually evaluated all the PET/CT scans for tracer accumulation in the lesions (positive, equivocal, or negative); the maximum standardized uptake value (SUVmax) was also recorded for each lesion. If the results of the two physicians differed, the physicians discussed the findings and reached a consensus. To evaluate the bone marrow lesions on the 11 C-MET and 11 C-4DST PET/CT images, the scale and window of the monitor display for these PET/CT images had to be adjusted so that the pathological uptake could be visualized with a better contrast against the high physiological uptake in the normal bone marrow. Before the start of this study, the physicians underwent training that included images from more than ten patients who had non-MM without bone diseases, since Nakamoto et al. [23] reported that suspicious lesions, including those in the bone marrow, could be clearly depicted using a proper display window and level in their 11 C-MET PET/CT study. Focal accumulation that was higher than the background was regarded as being positive, no accumulation compared with the  background was regarded as being negative, and accumulation with the same level as the background was regarded as being equivocal.

Focal lytic lesions
In 24 patients (before receiving therapy, 6 patients; after receiving therapy, 18 patients), a total of 55 focal lytic lesions (before receiving therapy, 10 lesions, after receiving therapy, 45 lesions) were detected using CT when PET/ CT was performed, but no diffuse lesions were detected. The sizes of the 55 focal lytic lesions were 23.7 ± 13.8 mm (range 6-70 mm), which was sufficiently large to be evaluated using low-dose CT and PET/CT.

Comparison to marrow plasma cells cytology
The percentages of marrow plasma cells in the posterior iliac crests were calculated using bone marrow aspiration smears in 36 patients ( Table 2) within 1 week before or after the three PET/CT studies. Eleven of the patients were also included in the first study. According to the criteria of the International Myeloma Working Group, a bone marrow clonal cell percentage of more than 10 % is regarded as a positive pathology for active myeloma and should be regarded as the gold standard for diagnosis [24]. We evaluated the accumulation of 18 F-FDG, 11 C-MET, and 11 C-4DST in the posterior iliac crests from where the bone marrow samples were obtained. The tracer accumulation was evaluated visually as positive, equivocal, or negative uptake. Even if abnormal accumulation was visible in lesions other than the posterior iliac crests, a positive-uptake evaluation was not made. Also, when an artifact from the bone marrow puncture was observed, the artifact was carefully excluded from the evaluation. Then, we compared the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy rate of PET/ CT using each of the three tracers.

Statistical analysis
The significance of the differences in the accumulation of the three tracers was determined using the area under the curve with a receiver operating characteristic (ROC) analysis. The significance of the differences between bone marrow aspiration and the accumulation of each of the three tracers was determined using the Fisher exact test. P values less than 0.05 were considered to be statistically significant.

Focal lytic lesions
Both before and after therapy, the number of equivocal lesions observed using 18 F-FDG was larger than that observed using 11 C-MET or 11 C-4DST. For 11 C-4DST, 11 C-MET, and 18 F-FDG, the highest SUVmax values were observed, in order, for positive, equivocal and negative lesions (Table 3). Among the patients who were examined after therapy, in particular, 11 C-MET or 11 C-4DST was capable of detecting positive lesions more frequently than 18 F-FDG. 18 F-FDG was rarely able to detect skull lesions because of the high physiological accumulation in the brain, whereas 11 C-MET and 11 C-4DST were capable of clearly detecting skull lesions because of their low accumulation in the brain (Fig. 2). A typical MM patient with multiple active lesions is sh own in Fig. 3. 11 C-MET and 11 C-4DST detected positive lesions, whereas 18 F-FDG detected an equivocal lesion. The lesion was positive when evaluated using MRI and negative when evaluated using CT (Fig. 4).

Comparison to marrow plasma cells cytology
A Fisher exact test demonstrated a significant correlation between positive uptake on the PET/CT scans and a positive pathology of the bone marrow plasma cells. The P values of 11 C-MET and 11 C-4DST were lower than that of 18 F-FDG. The diagnostic potentials of the three tracers are as described below. 11 C-4DST showed the highest sensitivity. The specificity of the three tracers was comparable ( Table 4). The ROC analysis showed statistically significant differences between 18 F-FDG and 11 C-MET and between 18 F-FDG and 11 C-4DST ( Table 5). The area under the ROC curves for 11 C-MET and 11 C-4DST were greater than that for 18 F-FDG. The other three indices for 11 C-MET and 11 C-4DST were larger than those for 18 F-FDG. But no statistically significant differences in the positive predictive value, negative predictive value, and accuracy rate were observed among the three tracers when examined using a Chi square test (P [ 0.05).
In patients with more than 58 % plasma cells, all the PET/CT data showed a positive uptake. However, in patients with 10-30 % plasma cells, 11 C-MET and 11 C-4DST detected larger numbers of positive-uptake lesions than 18 F-FDG ( Table 2).
All three PET/CT scans were negative for all three MGUS patients. Extramedullary lesions were not evaluated in the present study.

Discussion
We demonstrated the usefulness of 11 C-4DST and 11 C-MET PET/CT imaging, compared with 18 F-FDG PET/CT imaging, in patients with MM. 11 C-4DST and 11 C-MET provided clearer findings than 18 F-FDG for lytic lesions  A Fisher exact test demonstrated a significant correlation between positive uptake on the PET/CT scans and a positive pathology of the bone marrow plasma cells visible using CT. Furthermore, 11 C-4DST and 11 C-MET had higher diagnostic accuracies than 18 F-FDG, when compared using iliac crest biopsy data.
In the first study, 11 C-4DST and 11 C-MET provided clearer findings than 18 F-FDG when evaluating whether lytic lesions detected using CT were active or inactive. 11 C-4DST and 11 C-MET showed fewer equivocal accumulations than 18 F-FDG. Osteolytic lesions are more commonly found in the axial skeleton, skull, shoulder girdle, proximal humeri, ribs, and proximal femurs [25]. 11 C-4DST and 11 C-MET were useful for evaluating the tumor activities of these lesions. As shown in Fig. 2, the disease activities of skull lesions can typically be successfully evaluated using 11 C-4DST and 11 C-MET, but not 18 F-FDG. 11 C-4DST and 11 C-MET accumulate in myeloma lesions of the peripheral long bones. Dankerl et al. [16] reported that peripheral bone marrow expansion could be observed using 11 C-MET PET/CT. In accordance with these findings, 11 C-4DST and 11 C-MET may be more useful for evaluating MM lesions in peripheral long bones than 18 F-FDG. Furthermore, Nakamoto et al. reported that 11 C-MET provided clearer positive findings in some patients, compared with 18 F-FDG. Therefore, 11 C-MET was considered to be useful for determining the therapeutic strategy, especially when the 18 F-FDG findings were equivocal or indeterminate [23]. Thus, 11 C-4DST and 11 C-MET seem to be more useful for assessing the disease activities of lytic lesions detected using CT, compared with 18 F-FDG. Furthermore, all the negative lesions were observed as such using 11 C-4DST, 11 C-MET, and 18 F-FDG.
In the second study, we demonstrated that 11 C-4DST and 11 C-MET had higher diagnostic accuracies than 18 F-FDG. An iliac crest biopsy is the standard method for determining bone marrow infiltration by plasma cells [26]. In smoldering MM (SMM), bone marrow biopsies reveal a 10-30 % diffuse infiltration of plasma cells, while the infiltration is less than 10 % in MGUS [5,27] with no evidence of MM. Therefore, we evaluated the iliac crests in patients in whom a pathological diagnosis was obtained using 18 F-FDG, 11 C-4DST, and 11 C-MET. A statistical examination was difficult to perform because the number of patients was relatively small, but 11 C-MET and 11 C-4DST seemed to be more sensitive than 18 F-FDG in patients who had not yet received therapy.
In this study, all three PET/CT scans were negative in all three MGUS patients. In patients with MGUS, marrow plasma cells account for less than 10 %, while in myeloma, the bone marrow clonal cells account for no less than 10 % [24]. Both SMM and MGUS are typically not treated, but the prognoses differ. In MGUS, the overall risk of progression is about 1 % per year [28], and the median duration of MGUS and SMM before a diagnosis of myeloma is 81 and 23 months, respectively [29]. Therefore, it is important to distinguish SMM and MGUS.
In this study, extramedullary lesions were not evaluated. However, the ability of PET/CT to evaluate the whole body in a single procedure and the potential to detect medullary and extramedullary lesions during a single examination are important advantages over standard imaging techniques, such as MRI, CT, or radiographs [11]. Dankerl et al. reported that extramedullary MM was sensitively detected and localized using 11 C-MET. The evaluation of extramedullary lesions is important because these lesions are often difficult to detect but have a major impact on the prognosis. Nakamoto et al. reported a high level of 11 C-MET uptake in normal liver and pancreas; however, this situation is unlikely to cause false-negative findings because it is unusual to have extramedullary lesions in these organs [25].
When evaluating MM, diffuse lesions are more difficult to evaluate than focal lesions. The uptake of 18 F-FDG in the skeleton is caused by the activation of hematopoietic marrow, and its pattern and amount can vary with age and with the levels of marrow function, such as the level of function during recovery after chemotherapy or when subjected to the effect of granulocyte colony stimulating factor at the time of the PET/CT examination [30]. Because 11 C-4DST can be used to evaluate DNA synthesis [20], it can accumulate in active hematopoietic marrow. Thus, it may be difficult to distinguish diffuse MM lesions from hematopoietic marrow. A means of evaluating diffuse MM lesions should be a topic of future PET/CT studies.
A whole-body survey for active lesions is a unique advantage of PET/CT, and modern image processing techniques can minimize artifacts from metal prostheses. Thus, PET/CT has the potential to become a standard modality for the staging of MM. 11 C-4DST and 11 C-MET are better at detecting active lesions than 18 F-FDG. Therefore, the Durie/Salmon PLUS staging results determined using 11 C-4DST and 11 C-MET may differ from those determined using 18 F-FDG. However, the validation of a new staging method requires prognostic observation over a long observation period. The question of which tracer is the best for evaluating the viability of MM will require further observation. We are planning to evaluate this matter in a separate study.
Conclusion 11 C-4DST and 11 C-MET are useful for detecting bone marrow involvement in patients with MM, especially at an early stage, in a manner that is more clearly and more accurately than that using 18 F-FDG.