¹⁸F-FDG PET/CT, ^99mTc-MIBI, and MRI in Evaluation of Patients with Multiple Myeloma

MATERIALS AND METHODS

Thirty-three patients (11 females, 22 males; mean age ± SD, 64 ± 12 y) with newly diagnosed MM according to standard criteria were enrolled in this prospective study, which had undergone institutional approval before its inception. After informed consent had been obtained, all patients underwent whole-body ¹⁸F-FDG PET/CT, whole-body ^99mTc-MIBI, and MRI of the spine and pelvis in a random order within a maximum interval of 10 d. None of the patients had undergone chemotherapy or radiotherapy before the study.

¹⁸F-FDG PET/CT scans were acquired after fasting for 8 h and 60–90 min after intravenous administration of ¹⁸F-FDG (350–370 MBq). The blood glucose level, measured just before tracer administration, was <120 mg/dL in all patients. ¹⁸F-FDG PET/CT images were obtained using a PET/CT Discovery LS8 scanner (GE Healthcare). All scans were performed in 2-dimensional mode. An emission scan was performed in the caudocranial direction, from the upper thigh to the base of the skull (4 min/each bed position) and from the feet to the base of the thigh (2 min/each bed position). Iterative image reconstruction was completed with an ordered-subset expectation maximization (OSEM) algorithm (2 iterations, 28 subsets). CT with a 4-slice multidetector helical scanner was used (detector row configuration, 4 × 5 mm; pitch, 1.5; gantry rotation speed, 0.8 s per revolution; table speed, 30 mm per gantry rotation; 140 kV and 80 mA). Attenuation-corrected emission data were obtained using filtered backprojection CT reconstructed images (gaussian filter with 8-mm full width at half maximum) to match the PET resolution. Transaxial, sagittal, and coronal images and coregistered images were examined using Xeleris software (GE Healthcare). Focal areas visible on at least 2 contiguous PET slices—showing a maximum standardized uptake value (SUVmax) ≥ 2.5 and corresponding to CT abnormalities not attributable to benign bone pathologies—were considered to be sites of active disease. In particular, hypermetabolic sites corresponding to spondylopathy, osteoarthrosis, joint disease, or traumas were carefully excluded from the analysis, whereas those corresponding to CT abnormalities—such as lytic lesions, minor lytic changes, osteopenic areas, morphologic changes not clearly attributable to degenerative disease, and minimal asymmetry of bone marrow attenuation likely due to plasma cell infiltration—were included.

^99mTc-MIBI imaging studies were performed by acquiring planar anterior and posterior whole-body scans (lasting about 10 min) 10 min after intravenous injection of 555 MBq of ^99mTc-MIBI using a dual-head γ-camera (ECAM; Siemens), equipped with a low-energy, high-resolution collimator.

MRI studies were performed at 1.5 T (Achieva; Philips) along sagittal planes covering the whole spine with 3 partially overlapping slabs and along coronal planes for the study of the pelvis. MRI sequences included T1- and T2-weighted turbo-spin-echo images (with and without fat suppression by a preparatory pulse with spectral inversion [SPIR]) and postcontrast T1-weighted fat-suppressed turbo-spin-echo images (5 min after intravenous administration of 0.1 mmol/g gadopentetate dimeglumine [Magnevist]; Schering). The sequence parameters (repetition time/echo time/echo train length) used for the spine were 477/13/4 for T1-weighted images and 3,500/120/43 for T2-weighted images. The sequence parameters used for the pelvis were 550/14/5 for T1-weighted images and 3,500/120/43 for T2-weighted images with SPIR fat suppression. The whole study lasted approximately 35 min, including patient positioning.

¹⁸F-FDG PET/CT, ^99mTc-MIBI, and MRI were read and interpreted by 2 independent nuclear medicine physicians or 2 independent radiologists who were unaware of the imaging results. The data obtained were compared by using a χ² test or a Fisher exact test as appropriate. A probability value ≤ 0.05 was considered statistically significant. When a focal pattern was detected, the number and site of focal bone or soft-tissue lesions were reported. The number of focal lesions detected in each patient by each one of the 3 imaging techniques was compared by using the nonparametric paired-data Kendall's coefficient-of-concordance (W) test. A probability value ≤ 0.01 was considered statistically significant.

RESULTS

The results of whole-body ¹⁸F-FDG PET/CT, whole-body ^99mTc-MIBI, and MRI of the spine and pelvis performed on the 33 MM patients were compared according to the presence of a normal, diffuse, or focal (combination with or without diffuse) pattern of bone marrow involvement as shown in Table 1. Whole-body ¹⁸F-FDG PET/CT was positive in 32 patients (97%), 3 (9%) of whom had a pure diffuse pattern of bone marrow uptake, whereas 29 (88%) showed focal lesions in the presence (13 patients, 39%) or absence (16 patients, 48%) of diffuse uptake. Whole-body ^99mTc-MIBI was positive in 30 patients (91%), of whom 11 (33%) presented a diffuse pattern of uptake and 19 (57%) showed a focal pattern with (13 patients, 39%) or without (6 patients, 18%) the association of diffuse uptake. MRI of the spine and pelvis was positive in 27 patients (81%), of whom 13 (39%) had a diffuse pattern and 14 (42%) showed a focal pattern, in combination with a diffuse pattern (8 patients, 24%) or alone (6 patients, 18%).

Comparing the pattern of bone marrow involvement obtained by ¹⁸F-FDG PET/CT, ^99mTc-MIBI, and MRI, using the Fisher exact test, a significant statistical difference (P < 0.005) between the 3 imaging methods was found. In particular, ¹⁸F-FDG PET/CT detected a focal pattern, alone or combined with a diffuse pattern, with a higher frequency than ^99mTc-MIBI (P < 0.05) and MRI (P < 0.001). On the other hand, ^99mTc-MIBI and MRI were comparable and performed better than ¹⁸F-FDG PET/CT in the detection of a pure diffuse pattern. By analyzing the number and sites of focal lesions detected, we found that ¹⁸F-FDG PET/CT showed a total of 196 focal lesions, of which 75 were in the spine and pelvis (35 and 40, respectively) and 121 were in other districts, whereas ^99mTc-MIBI visualized a total of 63 focal lesions, of which only 1 was in the spine, 9 were in the pelvis, and 53 were in other districts. Eighteen of the total focal lesions visualized by ¹⁸F-FDG PET/CT and only 3 of the lesions detected by ^99mTc-MIBI were localized in soft tissues. Finally, MRI detected a total of 51 focal lesions—40 in the spine and 11 in the pelvis. Comparing the number of focal lesions per patient detected by each imaging method on the whole dataset, using the nonparametric paired-data Kendall's W test, we showed that ¹⁸F-FDG PET/CT visualized more focal lesions (5.94 ± 9.29) than ^99mTc-MIBI and MRI (1.91 ± 4.45 and 1.54 ± 2.45, respectively), with a significant statistical difference (P < 0.001) as shown in Table 2. We also performed a post hoc analysis using the Kendall's W test with Bonferroni correction and found a significant statistical difference between the number of focal lesions per patient detected by ¹⁸F-FDG PET/CT and both ^99mTc-MIBI (P < 0.001) and MRI (P < 0.005).

To compare homogeneously the 3 imaging methods used, we focused our analysis exclusively on the data obtained in the spinal and pelvic district as shown in Table 3. Comparing these data, using the Fisher exact test, we also found a significant statistical difference (P < 0.05) between the 3 imaging techniques. In particular, ¹⁸F-FDG PET/CT and MRI were comparable and performed better than ^99mTc-MIBI in the detection of a focal pattern, alone or combined with a diffuse pattern. On the other hand, ^99mTc-MIBI and MRI were comparable and performed better than ¹⁸F-FDG PET/CT in the detection of a diffuse pattern. However, these differences were statistically significant only between ¹⁸F-FDG PET/CT and ^99mTc-MIBI (P < 0.01). Comparing the number of focal lesions per patient visualized by each imaging technique exclusively in the spinal and pelvic district, using the nonparametric paired-data Kendall's W test, we found that ¹⁸F-FDG PET/CT and MRI showed more focal lesions (2.27 ± 4.64 and 1.54 ± 2.45, respectively) compared with ^99mTc-MIBI (0.30 ± 0.68), with a significant statistical difference (P < 0.005) as shown in Table 4. Moreover, the post hoc analysis performed by using the Kendall's W test with Bonferroni correction showed a significant statistical difference between the number of focal lesions per patient detected by ¹⁸F-FDG PET/CT and MRI (P < 0.001 and P < 0.01, respectively) versus ^99mTc-MIBI, whereas no significant statistical difference was found between ¹⁸F-FDG PET/CT and MRI.

Finally, ¹⁸F-FDG PET/CT, ^99mTc-MIBI, and MRI influenced the subsequent clinical management in 18% of patients. In particular, ¹⁸F-FDG PET/CT alone or in combination with ^99mTc-MIBI upstaged and, consequently, modified the clinical management in 4 patients, whereas ^99mTc-MIBI alone and MRI alone modified the clinical management in 1 patient each.

DISCUSSION

The results of the present study indicate that ¹⁸F-FDG PET/CT performs better than both ^99mTc-MIBI and MRI in the detection of focal lesions on the whole data analysis. However, in the spinal and pelvic district, ¹⁸F-FDG PET/CT and MRI were comparable, and both performed better than ^99mTc-MIBI in visualizing focal lesions. On the other hand, ^99mTc-MIBI and MRI showed a higher rate of diffuse pattern detection compared with ¹⁸F-FDG PET/CT both in the whole data and in the spinal and pelvic analysis. Because of the objective difficulty—for both practical and ethical reasons—in obtaining biopsy samples from each suspected lesion, our findings were interpreted in relation to the clinical context and analyzed to compare simultaneously the 3 imaging modalities. Moreover, previous studies showed that a whole-body radiographic survey—traditionally used in the Durie and Salmon staging system to detect lytic lesions—can significantly underestimate bone and bone marrow involvement, especially in newly diagnosed patients (4). Therefore, an independent “gold standard” of reference cannot be easily provided in this disease. Nevertheless, the 3 imaging methods singly or in combination influenced the subsequent clinical management in 18% of patients.

The higher number of focal lesions detected by ¹⁸F-FDG PET/CT compared with ^99mTc-MIBI (196 and 63 lesions, respectively) may be due to the mechanism of ¹⁸F-FDG uptake that reflects the increased glycolysis usually occurring in tumor cells and, thus, the rapid growth and invasive characteristics of focal lesions (4,16,26)—though the inflammation that can be associated with tumor proliferation may also contribute to increased ¹⁸F-FDG uptake (27). Moreover, the use of a hybrid system composed by PET and CT images allowed the detection of small or slightly active lesions that were barely distinguishable from the surrounding normal tissue on the basis of PET images alone (28). The hybrid system also allows a more precise anatomic localization of hypermetabolic lesions and, therefore, a better discrimination between bone and soft-tissue lesions (8,10). In fact, ¹⁸F-FDG PET/CT detected a total of 18 soft-tissue lesions, whereas ^99mTc-MIBI visualized only 3 of them. The resolution of ^99mTc-MIBI could be improved by performing SPECT. This technique of acquisition, though, would be time-consuming and costly, taking from ^99mTc-MIBI imaging some of its positive characteristics—technical ease, rapidity of execution (the study is performed 10 min after injection and lasts only 10 min), and low costs.

On the whole data analysis, ¹⁸F-FDG PET/CT detected more focal lesions than MRI (196 and 51 lesions, respectively) because of the presence of a consistent number of lesions outside the spine and pelvis (121 lesions detected by ¹⁸F-FDG PET/CT and 53 lesions detected by ^99mTc-MIBI). In fact, focusing our analysis exclusively on the spinal and pelvic district, the number of focal lesions visualized by ¹⁸F-FDG PET/CT and MRI became comparable (75 and 51 lesions, respectively). In this district, both ¹⁸F-FDG PET/CT and MRI performed better than ^99mTc-MIBI, which detected 10 lesions only; this finding could be due to the physiologic uptake of ^99mTc-MIBI in the liver and its excretion in the bowel, which may obscure local focal lesions (18).

^99mTc-MIBI performed better than ¹⁸F-FDG PET/CT in the detection of a diffuse pattern of bone marrow uptake both in the whole data (33% of patients by ^99mTc-MIBI and 9% by ¹⁸F-FDG PET/CT) and in the spinal and pelvic analysis (54% of patients by ^99mTc-MIBI and 18% by ¹⁸F-FDG PET/CT). The meaning of ¹⁸F-FDG diffuse bone marrow uptake in MM patients must be further investigated, as a mild and diffuse ¹⁸F-FDG uptake in the spine could be also found in young or mildly anemic patients (28,29). On the other hand, previous studies showed that ^99mTc-MIBI concentrates in malignant plasma cells and that diffuse tracer uptake correlates with the percentage of plasma cell infiltration and the amount of a monoclonal component (13,15). Moreover, it has been shown that ^99mTc-MIBI bone marrow uptake is able to identify active myeloma and that the extension and intensity of tracer uptake correlates both with the clinical status and the stage of disease (13). In fact, a previous study showed that moderate-to-intense diffuse ^99mTc-MIBI uptake or focal uptake with or without diffuse uptake, in the absence of inflammation or other pathologies, excludes the diagnosis of monoclonal gammopathy of unknown significance (MGUS) and correlates with poor prognosis (19). However, it should be noted that faint bone marrow uptake has been reported also in patients affected by pathologies other than MM (20). Nevertheless, when the intensity of diffuse ^99mTc-MIBI uptake was analyzed according to the criteria used by Pace et al. (13), specificity improved significantly (20). False-negative cases by ^99mTc-MIBI may be due, rather, to the overexpression of P-glycoprotein (Pgp) that can be associated with multidrug-resistant myeloma. ^99mTc-MIBI, in fact, is a transport substrate of the energy-dependent efflux pump Pgp, and its washout increases over time from the bone marrow of MM patients overexpressing this protein (17,30). Therefore, to overcome the action of Pgp in our study, imaging was performed no later than 10 min after the injection of ^99mTc-MIBI.

Similarly to ^99mTc-MIBI, MRI also performed better than ¹⁸F-FDG PET/CT in the detection of a diffuse pattern both in the whole data and in the spinal and pelvic analysis (39% of patients by MRI of spine and pelvis, 9% by whole-body ¹⁸F-FDG PET/CT, and 18% by ¹⁸F-FDG PET/CT in the spinal and pelvic regions, respectively). Previous studies, in fact, showed that a diffuse pattern of distribution detected by MRI in MM patients correlates with increased bone marrow cellularity, increased plasmacytosis (although <10% may be associated with false-negative cases), anemia, and poorer survival (31). Moreover, recent studies showed that MRI was more sensitive than ¹⁸F-FDG PET/CT in the detection of an infiltrative pattern, allowing direct visualization of the bone content with a high spatial resolution (8,28). These features can be useful, especially in the spinal and pelvic regions that have a complex anatomy and are overlaid by bowel and ribs, respectively (32)—though, the field of view of MRI excludes regions such as skull, sternum, ribs, and long bones containing a high amount of red marrow and frequently infiltrated by malignant plasma cells (23,24), as shown in Figure 1. In fact, it has been shown that in substituting a whole-body radiographic survey with MRI of spine and pelvis, 10% of MM patients would be understaged (33). In this respect, Zamagni et al. (8) reported that MRI was superior to ¹⁸F-FDG PET/CT in the assessment of bone marrow involvement of the spine and pelvis, whereas ¹⁸F-FDG PET/CT allowed the detection of myelomatous lesions that were out of the field of view of MRI. In agreement with these findings, our study showed that MRI performed better than ¹⁸F-FDG PET/CT in the evaluation of diffuse disease and performed equally well in the detection of focal disease in the spinal and pelvic regions. Also, our study showed a considerable number of focal lesions detected by ¹⁸F-FDG PET/CT that were out of the field of view of MRI. This limitation could be overcome by using whole-body MRI. Currently, though, this imaging technique is not widely available yet, and its imaging times are still too long despite the advances in MRI, such as the development of rapid data acquisition and high-performance gradient systems (9). Moreover, the spatial resolution of whole-body MRI is worse than that of focused surface-coil MRI, resulting in poorer imaging quality (34).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1. 

Coronal, sagittal, and transaxial CT, PET, and fused images of whole-body ¹⁸F-FDG PET/CT (A), anterior and posterior whole-body ^99mTc-MIBI (B), and sagittal T1- and T1-weighted fat-saturated with gadolinium MR images of spine (C). All 3 imaging modalities, performed on the same MM patient, showed a focal + diffuse pattern of distribution. In particular, focal lesions were shown in skull, humeri, pelvis, left femur, and right fibula by whole-body ¹⁸F-FDG PET/CT; in skull, right humerus, pelvis, left femur, and right fibula by whole-body ^99mTc-MIBI; and in coccygeal bone by MRI of spine and pelvis.

CONCLUSION

In whole-body analysis, ¹⁸F-FDG PET/CT and ^99mTc-MIBI provided complementary information in the diagnostic evaluation of MM patients by detecting focal and diffuse disease, respectively. In the spinal and pelvic regions, MRI was comparable to ¹⁸F-FDG PET/CT and ^99mTc-MIBI in the detection of focal and diffuse patterns, respectively. Therefore, in the diagnostic work-up of multiple myeloma, MRI—because of its ability in detecting both focal and diffuse disease in the spine and pelvis—should be reserved for the evaluation of bone marrow involvement in these regions. Until whole-body MRI with reasonably short imaging times, good spatial resolution, and standardized sequences for MM will be widely available, the main drawback of MRI of spine and pelvis will be the limited field of view that could understage newly diagnosed MM patients, by missing lesions located outside these regions. Therefore, in the whole-body evaluation of MM patients at diagnosis, ¹⁸F-FDG PET/CT can contribute to a more accurate assessment of disease—especially in a clinical context highly suggestive of focal involvement of the appendicular skeleton, such as the presence of bone pain or pathologic fractures in long bones or in the case of discrepancies between clinical status and hematologic parameters. On the other hand, despite the limited capacity in detecting focal lesions, ^99mTc-MIBI still remains the most rapid and inexpensive technique for whole-body evaluation and may be an alternative option when a PET facility is not available.