Diagnostic usefulness of an amino acid tracer, α-[N-methyl-11C]-methylaminoisobutyric acid (11C-MeAIB), in the PET diagnosis of chest malignancies

Objectives Although positron emission tomography (PET) using [18F]-fluoro-2-deoxy-d-glucose (18F-FDG) is established as one of the first-choice imaging modalities in the diagnosis of chest malignancies, there are several problems to solve in clinical practice, such as false positive uptake in inflammatory diseases. The aim of this study was to evaluate the clinical usefulness of an amino acid tracer, α-[N-methyl-11C]-methylaminoisobutyric acid (11C-MeAIB), in the diagnosis of chest malignancies, in combination with 18F-FDG. Setting Fifty-nine cases (57 patients, 66 ± 12 years old) who consulted to our institution for the wish to receive differential diagnosis of chest diseases were included. Purpose of the studies were as follows: differential diagnosis of newly developed lung nodules, n = 22; newly developed mediastinal lesions, n = 20; and both, n = 17 (including lung cancer: n = 19, lymphoma: n = 1, other cancers: n = 2, sarcoidosis: n = 15, non-specific inflammation: n = 18, other inflammatory: n = 4, respectively). Whole-body static PET or PET/CT scan was performed 20 and 50 min after the IV injection of 11C-MeAIB and 18F-FDG, respectively. Results 11C-MeAIB uptake of malignant and benign lesions was statistically different both in pulmonary nodules (p < 0.005) and in mediastinal lesions (p < 0.0005). In visual differential diagnosis, 11C-MeAIB showed higher results (specificity: 73 %, accuracy: 81 %), compared to those in 18F-FDG (60, 73 %, respectively). In cases of sarcoidosis, 11C-MeAIB showed higher specificity (80 %) with lower uptake (1.8 ± 0.7) in contrast to the lower specificity (60 %) with higher uptake of 18F-FDG (7.3 ± 4.5). Conclusions 11C-MeAIB PET/CT was useful in the differential diagnosis of pulmonary and mediastinal mass lesions found on CT. 11C-MeAIB PET or PET/CT showed higher specificity than that of 18F-FDG PET/CT in differentiating between benign and malignant disease. Our data suggest that the combination of 18F-FDG and 11C-MeAIB may improve the evaluation of chest lesions, when CT and 18F-FDG PET/CT are equivocal.

A variety of new tracers has been introduced in order to overcome these limitations [9][10][11]. Amino acid tracers are one promising category. 11 C-methionine ( 11 C-MET) has been extensively investigated, though with several drawbacks being revealed [12]. 11 C-MET is not stable in vivo, becoming trans-methylated, and losing its 11 C moiety as it is metabolized [13]. Several studies have shown that 11 C-MET is not tumor-specific, and accumulates in some inflammatory diseases, such as sarcoidosis [14,15].
[N-methyl- 11 C]a-methylaminoisobutyric acid ( 11 C-MeAIB) is an artificial amino acid PET tracer, which, unlike 11 C-MET, is metabolically stable in vivo [16,17]. It has been shown that 11 C-MeAIB is a useful tracer for measurement of amino acid uptake by skeletal muscle, and in the diagnosis of malignant lymphoma and head and neck cancers [16,18,19]. The utility of 11 C-MeAIB PET in the diagnosis of chest malignancies has not yet been evaluated.
The purpose of this study is to investigate the efficacy of 11 C-MeAIB PET or PET/CT as a diagnostic tool for distinguishing between malignancies and inflammatory diseases, when CT or 18 F-FDG PET or PET/CT shows equivocal findings.  20 patients had mediastinal lesions only, and 18 patients had both lesions) were selected. Two patients received this study protocol twice with 1 year interval or more. Each patient gave written informed consent. The tracer study was approved by our institutional review boards, the Human Study Committee (approval number: #36-04, Mar. 25,2009) and the Committee for the Clinical Use of Short-Half Life Radioactive Materials (approval number: #2008-01, Nov. 28,2008). Inclusion criteria were: (1) a new chest lesion detected on CT scan, suspicious for malignancy, (2) CT (and 18 F-FDG PET or PET/CT) showing equivocal findings, with further evaluation requested by referring pulmonologists, and (3) disease confirmed by pathology, or clinical follow-up for more than 12 months after PET studies. Exclusion criteria were: (1) malignant or inflammatory lesions which received treatment within 6 months before 18 F-FDG PET or PET/CT, (2) apparent direct invasion of neighboring organs, (3) apparent extra-pulmonary metastatic lesions, and (4) patients who refused to undergo 11 C-MeAIB PET or PET/CT. Baseline diseases or detailed background of clinical conditions are shown in Table 1 (left).

Patients
The CT scans had been performed as routine clinical studies with a multi-detector row CT scanner, Aquilion 16 (Toshiba, Tokyo, Japan). All patients subsequently underwent both 18 F-FDG and 11 C-MeAIB PET or PET/CT within 2 weeks of the CT.
Production of 11 C-MeAIB was based on the method proposed by Någren et al. [20]. The patients fasted for more than 5 h before the injection of 18 F-FDG or 11 C-MeAIB. Blood glucose levels were measured before the injection of 18 F-FDG, and all the patients showed levels of B150 mg/dL (95.2 ± 13.1 mg/dL). All subjects underwent two separate scans, one following an intravenous injection of 18 FDG (298 ± 68 MBq), and another after the intravenous injection of 11 C-MeAIB (512 ± 50 MBq). Wholebody PET image acquisition commenced 50 min after 18 F-FDG and 20 min after 11 C-MeAIB injection. PET scans were performed either by a whole-body PET scanner, GE Advance (GE Healthcare, Waukesha WI, USA), or by a whole-body PET/CT scanner, Siemens True Point Biograph 16 (Siemens/CTI, Erlangen, Germany).

Image analysis
Visual analysis of each lesion on the two PET scans was performed by two experienced nuclear medicine physicians (RN, TH) provided with clinical information including CT scans and tumor markers. All lesions were graded as positive or negative by consensus of two readers. If a nodule showed similar or lower uptake than that in the upper to middle normal mediastinal tissues, uptake was defined as negative. If a nodule showed higher uptake than that of the normal mediastinal tissues, its uptake was defined as positive. However, high 18 F-FDG uptake was sometimes defined as negative in cases of sarcoidosis and other benign entities with a characteristic pattern by consensus of two readers.
Semi-quantitative analysis of 18 F-FDG and 11 C-MeAIB uptake was also performed. Regions of interest (ROIs) were defined on the target lesions in the transaxial tomograms of PET-only images. PET-to-CT co-registration was performed using automatic rigid/non-rigid body-deformable fusion software: Quantiva/BodyGuide (Tomographix IP Ltd., Toronto, Canada). In PET/CT scans, ROIs were defined and confirmed on the fused PET and CT images (hereafter all scans will be referred to as PET/CT scans, since PET images were always fused to CT images). The standardized uptake value (SUV) was calculated as follows: where C represents tissue activity concentration measured by PET and ID represents the injected dose. The mean SUV of the normal tissue (lung field and mediastinum) was defined as the SUVmean. The highest SUV of the lesion was defined as the SUVmax.

Statistics
All values are expressed as mean ± SD. All the statistical analyses were performed using statistical software, JMP 8J version (SAS Institute, Cary NC, USA), in which p values \0.05 were considered statistically significant. A comparison between each group was analyzed by the Wilcoxon score for unpaired data. Receiver operating characteristic curve (ROC) analyses for the diagnostic accuracy in 11 C-MeAIB PET/CT and 18 F-FDG PET/CT were generated using GraphPad Prism ver. 5.0 (GraphPad software, San Diego CA, USA). A comparison of SUVmax between 11 C-MeAIB and 18 F-FDG in each lesion was analyzed by the Logistic regression. Table 1 summarizes the patient characteristics. The final diagnosis was confirmed pathologically by surgical resection in 17 cases, while biopsy at bronchoscopy, thoracoscopy, and CT-guided biopsy confirmed diagnosis in 10 cases. Resection or biopsy of lymph nodes at neck or supraclavicular area was performed in seven cases. Clinical diagnosis was determined by follow-up for at least 12 months in 24 cases. Of the 59 cases in the present study, there were 22 malignant and 20 benign pulmonary nodules in the lungs and 10 malignant and 28 benign mediastinal lesions. Figure 1 shows a typical malignant case, while Fig. 2 shows a typical benign case; in this instance, sarcoidosis. In Fig. 3, a 18 F-FDG-strongly positive lung nodule was diagnosed as sarcoidosis. Equivocal findings of metastatic colon cancer are shown in Fig. 4.

Results
Semi-quantitative analysis of 18 FDG uptake and 11 C-MeAIB uptake The average SUVmax and SUVmean for 18 F-FDG and 11 C-MeAIB uptake in normal lung and mediastinum are shown in Table 2. The average SUVmax of 18 F-FDG in malignant lesions was significantly higher than that in benign lesions for 42 pulmonary nodules, while not significantly different for 38 benign and malignant mediastinal lesions. There was a wide overlap in 18  The average SUVmax of malignant lesions with 11 C-MeAIB PET/CT was significantly lower than that for 18 F-FDG PET/CT. On the other hand, 11 C-MeAIB uptake by malignant and benign lesions showed greater statistical differences both among pulmonary nodules and mediastinal lesions. Figure 5 shows the result of ROC analyses for 18 F-FDG and 11 C-MeAIB, in which the diagnostic accuracies were obtained from the SUVmax values of each tumor in both PET/CT studies. An SUVmax = 3.0 was used as the threshold for 18 F-FDG diagnosis. Other thresholds (SUVmax = 2.0, 2.5, 3.5, 4.0) gave similar or worse diagnostic results. In 11 C-MeAIB PET/CT studies, an SUVmax = 2.0 was used as the optimum threshold. 11 C-MeAIB scans showed a higher value than 18 F-FDG scans both in patientbased (Fig. 5) and lesion-based diagnoses (not shown).

Patient-based diagnostic results
Visual diagnosis of 18 FDG and 11 C-MeAIB on a perpatient basis in 59 cases is shown in Table 3 (top). The accuracy of 18 FDG was 72.9 %, which was better than that from semi-quantitative analysis. Although 18 FDG uptake was often equivocally positive, the final diagnosis was judged as true negative in many benign or inflammatory cases because of the analysis of pattern and location of the 18  and mediastinal lymphadenopathy confirmed by surgical resection (n = 6), mediastinal sarcoid lymphadenopathy by lymph node biopsy (n = 5), mediastinal IgG4-related lymphadenopathy by lymph node biopsy (n = 1), and nonspecific inflammatory change followed for more than 12 months (n = 3). The accuracy of 11 C-MeAIB was 81.4 %. Ten false positive cases with 11 C-MeAIB were as follows: granulomatous inflammatory lung nodules and mediastinal lymphadenopathy confirmed by surgical resection (n = 4), mediastinal sarcoid lymphadenopathy by lymph node biopsy (n = 3), mediastinal IgG4-related lymphadenopathy by lymph node biopsy (n = 1), and nonspecific inflammatory change followed for more than 12 months (n = 2). Nine of these 11 C-MeAIB false positive cases were also false positive with 18 F-FDG. One false negative case was of metastatic colon cancer, which was also false negative in 18 F-FDG PET images (Fig. 4).
In the semi-quantitative diagnoses (59 cases) with SUVmax = 3.0 cut-off value, the accuracy of 18 F-FDG PET/CT was 47.5 %, in which there were many false positive results as compared with those at visual diagnosis (Table 3 bottom). There were two false negative cases with 18 F-FDG PET/CT; both were papillary adenocarcinomas with ground-glass opacity (n = 2, diameter: 25 and 25 mm, SUVmax: 1.89 and 2.00). The accuracy of 11 C-MeAIB PET/CT with SUVmax = 2.0 cut-off value was 74.6 %. There were 13 false positive cases and 2 false negative cases. The false positive cases were as follows: granulomatous inflammatory mediastinal lymphadenopathy confirmed by surgical resection (n = 4), mediastinal sarcoid lymphadenopathy by lymph node biopsy (n = 3), IgG4-related syndrome (n = 1), and non-specific inflammatory change followed for more than 12 months (n = 5). The two false negative cases were papillary adenocarcinomas with ground-glass opacity (n = 2, diameter: 25 and 30 mm, SUVmax: 1.74 and 1.93). The former one was also false negative case with 18 FDG, but the latter one was true positive with 18

Discussion
Several amino acid compounds have been suggested as feasible candidates for oncologic PET tracers which can overcome the drawbacks of 18 F-FDG. 11 C-MeAIB is considered to be one of the most promising amino acid radiotracers in clinical oncology. To our knowledge, the present study is the first to evaluate the clinical application of 11 C-MeAIB PET-to-chest lesion diagnosis.
Our principal finding is that the diagnostic results of 11 C-MeAIB PET/CT were better than those of 18 F-FDG PET/CT, especially for the identification of non-malignant lesions. Table 3 clearly reveals the higher specificity of 11 C-MeAIB PET/CT in these cases. In the evaluation of mediastinal lesions, 11 Table 4). The low positive predictive value (27.8 %) of 18 F-FDG PET/CT confirms that positive uptake of 18 F-FDG is not reliably diagnostic of malignancy. We believe that 11 C-MeAIB PET/CT would make a great contribution in the diagnosis of patients with pulmonary nodules or mediastinal lesions, when CT and 18 F-FDG PET/CT shows equivocal findings.
It should be noted that 11 C-MeAIB PET/CT displayed high diagnostic accuracy in the evaluation of sarcoidosis. There were 20 sarcoid lesions in 14 patients and the average lesion 18 F-FDG SUVmax was 7.3 ± 4.5, while that of 11 C-MeAIB was 1.8 ± 0.7 (Table 5). Visual diagnosis with 18 F-FDG PET/CT showed false positives in six patients, while those with 11 C-MeAIB PET/CT showed false positives in three patients. Although 18 F-FDG PET/CT can diagnose sarcoidosis by the specific uptake pattern of hilar and mediastinal lesions, sarcoidosis can form pulmonary nodules as well, and those lesions can be difficult to distinguish from malignancies. In addition, malignancy is often observed synchronously or metachronously in patients with sarcoidosis. In the present study, there were five pulmonary   lesions (all of them finally confirmed as benign) in five patients with sarcoidosis. In these cases, 11 C-MeAIB PET/ CT played a useful diagnostic role (Fig. 3). Since 11 C-MET accumulates in sarcoidosis [14,15], it is suggested that 11 C-MeAIB may be superior to 11 C-MET in the differentiation of sarcoidosis from malignancy. Our result is compatible with previous studies using other amino acid PET tracers, such as [ 18 F]-methyltyrosine ( 18 F-FMT) [21]. Kaira et al. suggested in their report that the use of 18 F-FMT PET in combination with 18 F-FDG PET may be effective for this purpose. In terms of biological mechanism, it is not fully understood why 11 C-MET and the other amino acid PET tracers ( 11 C-MeAIB and 18 F-FMT) show different uptake patterns in sarcoidosis. One of the conceivable mechanisms for the low uptake in sarcoidosis lesions of 11 C-MeAIB and 18 FMT is that these PET tracers, as artificial amino acids, are not metabolized in vivo [16,22]. Concerning in vivo instability of 11 C-MET, inflammatory lesion can be misdiagnosed by 11 C-MET PET because of its non-specific accumulation of free 11 C in blood when an inflammatory lesion shows hypervascularity. Comparative study of these amino acid PET tracers should be further evaluated.
In the diagnosis of malignancy, the sensitivity of 18 F-FDG and 11 C-MeAIB based on semi-quantitative patient-based diagnosis showed the same values (90.9 %) ( Table 4). In addition, the uptake of 11 C-MeAIB correlated well with 18 F-FDG uptake and there were basically no discrepant cases (Fig. 6). 18 F-FDG SUVs in malignant cases was usually two to three times higher than those of 11 C. In previous studies using 11 C-MET and 18 F-FMT, 11 C-MET and 18 F-FMT SUVs were also two to three times lower than those of 18 FDG [21,[23][24][25]. This may be a common drawback of amino acid PET tracers. Although our group included several different types of lung cancers, such as adenocarcinoma, squamous cell carcinoma, and small cell carcinoma, there was no significant difference in the uptake intensity of 11 C-MeAIB among the different histological types. It is not what we anticipated for 11 C-MeAIB PET/CT's use as a predictor of therapeutic effect, because amino acid transporters are known to work as carriers of chemotherapeutic agents, such as cisplatin, methotrexate, taxol, and melphalan [26][27][28]. The role of 11 C-MeAIB PET/CT as an imaging modality for patienttailored medicine is unknown. Further study of pre-and post-chemotherapeutic 11 C-MeAIB PET or PET/CT is needed.
Another drawback of 11 C-MeAIB is its high physiological uptake by liver. It means that 11 C-MeAIB PET/CT cannot be performed as a first-choice diagnostic modality in the evaluation of chest malignancies, because liver metastasis is common in lung cancer. Therefore, 11 C-MeAIB PET or PET/CT cannot be performed as a study for staging of advanced lung cancer. This is why we focused our study only on the differential diagnosis in chest diseases, and excluded cases with apparent distant metastasis and direct invasion of neighboring organs.
Conclusions 11 C-MeAIB PET/CT was useful in the differential diagnosis of pulmonary and mediastinal mass lesions found on CT. 11 C-MeAIB PET or PET/CT showed higher specificity than that of 18 F-FDG PET/CT in differentiating between benign and malignant disease. Our data suggest that the combination of 18 F-FDG and 11 C-MeAIB may improve the evaluation of chest lesions, when CT and 18 F-FDG PET/CT are equivocal.
Acknowledgments This study was supported by two categories of the Grants-in-Aid for Scientific Research programmed by the Japan Society for the Promotion of Science (JSPS), Challenging Exploratory On the other hand, that of 11 C-MeAIB also shows a linear correlation with that of 18 F-FDG but not significant in benign lesions (p = 0.055, R 2 = 0.078)