Abstract
(d)-18F-fluoromethyltyrosine (d-18F-FMT), or BAY 86-9596, is a novel 18F-labeled tyrosine derivative rapidly transported by the l-amino acid transporter (LAT-1), with a faster blood pool clearance than the corresponding l-isomer. The aim of this study was to demonstrate the feasibility of tumor detection in patients with non–small cell lung cancer (NSCLC) or head and neck squamous cell cancer (HNSCC) compared with inflammatory and physiologic tissues in direct comparison to 18F-FDG. Methods: 18 patients with biopsy-proven NSCLC (n = 10) or HNSCC (n = 8) were included in this Institutional Review Board–approved, prospective multicenter study. All patients underwent 18F-FDG PET/CT scans within 21 d before d-18F-FMT PET/CT. For all patients, safety and outcome data were assessed. Results: No adverse reactions were observed related to d-18F-FMT. Fifty-two lesions were 18F-FDG–positive, and 42 of those were malignant (34 histologically proven and 8 with clinical reference). Thirty-two of the 42 malignant lesions were also d-18F-FMT–positive, and 10 lesions had no tracer uptake above the level of the blood pool. Overall there were 34 true-positive, 8 true-negative, 10 false-negative, and only 2 false-positive lesions for d-18F-FMT, whereas 18F-FDG was true-positive in 42 lesions, with 10 false-positive and only 2 false-negative, resulting in a lesion-based detection rate for d-18F-FMT and 18F-FDG of 77% and 95%, respectively, with an accuracy of 78% for both tracers. A high d-18F-FMT tumor–to–blood pool ratio had a negative correlation with overall survival (P = 0.050), whereas the 18F-FDG tumor–to–blood pool ratio did not correlate with overall survival. Conclusion: d-18F-FMT imaging in patients with NSCLC and HNSCC is safe and feasible. The presented preliminary results suggest a lower sensitivity but higher specificity for d-18F-FMT over 18F-FDG, since there is no d-18F-FMT uptake in inflammation. This increased specificity may be particularly beneficial in areas with endemic granulomatous disease and may improve clinical management. Further clinical investigations are needed to determine its clinical value and relevance for the prediction of survival prognosis.
Despite the widespread use and success of 18F-FDG PET/CT for cancer imaging, the potential benefits of amino acid–based tracers with a potentially higher specificity have been discussed and investigated in the past (1). Especially, the well-known accumulation of 18F-FDG in inflammatory lesions—indistinguishable from tumor accumulation—can lead to over-staging of patients. One well-established concept for improved tumor specificity is the increased expression of amino acid transporters in transformed cancer cells, leading to increased accumulation of amino acids in tumor cells (2). The general feasibility of amino acid imaging in different tumor types has been demonstrated extensively (3). However, several studies have shown an insufficient sensitivity for some amino acid tracers, for example, O-2-18F-fluoroethyl-l-tyrosine (18F-FET) for head and neck cancer (4–6). The situation is different for brain tumors, for which 18F-FET is established and helpful in differentiating inflammatory from cancerous lesions and in tumor grading (7). Furthermore, 18F-FET has added value in therapy assessment after surgery or radiotherapy of malignant gliomas (8).
(d)-18F-fluoromethyltyrosine (d-18F-FMT), or BAY 86-9596, is a new diagnostic amino acid PET tracer. Close bioisosteres of l-amino acids labeled for imaging with 18F tend to behave like their naturally occurring counterparts, and some are integrated and metabolized by mammalian cells, resulting in a significant uptake in nontumorous cells, whereas d-isomers are characterized by a faster clearance (9). The principle of using d-amino tyrosine derivatives was introduced by Tsukada et al. (10,11). Comparison of d-18F-FMT with other “alkyl”-tyrosines showed the superiority of d-18F-FMT over d- and l-18F-FET or d- and l-fluoropropyltyrosine regarding tumor uptake rate in mice (10). These results suggest that tumor uptake decreases with an increasing length of the alkyl chain carrying the 18F label. A high tumor accumulation of d-18F-FMT along with a rapid blood pool clearance resulted in favorable images with significant tumor–to–blood pool and tumor-to-organ ratios for d-18F-FMT in tumor-bearing mice and rats (12) and could also be confirmed in a mouse bone metastasis model originating from renal cancer cells (13). d-18F-FMT was successful in monitoring the radiotherapy response of tumors in mice bearing tumors with a squamous cell carcinoma cell line (14).
A significant limitation of 18F-FDG PET imaging is the accumulation in inflamed tissues, leading to false-positive results (15). d-18F-FMT proved to have no accumulation in turpentine-induced inflammatory tissue in a mouse model, whereas 18F-FDG displayed significant uptake and retention (11).
Therefore d-18F-FMT could supersede the currently used tyrosine derivatives because of its significant tumor uptake and rapid clearance, potentially resulting in an earlier image acquisition and improved tumor-to-background levels. Herein, the first results, to our knowledge, are reported on the tracer safety and diagnostic potential of d-18F-FMT in patients with head and neck squamous cell cancer (HNSCC) and patients with non–small cell lung cancer (NSCLC).
MATERIALS AND METHODS
Patients
This prospective study was performed in 2 centers and approved by the Institutional Review Board of each. The conduct of the study met all local legal and regulatory requirements and was in accordance with the ethical principles originating from International Conference on Harmonization guideline E6: Good Clinical Practice. To be included in the study, patients had to give written informed consent; be between 30 and 80 y old; if female, either have no childbearing potential or not be pregnant or nursing; have an Eastern Cooperative Oncology Group performance status of 0–2; have a life expectancy of more than 3 mo; and have confirmation of adequate function of major organ systems. In addition, inclusion in the study required that patients have undergone clinically indicated 18F-FDG PET/CT for staging, restaging, or therapy response assessment that showed a tumor mass with 18F-FDG uptake and a high certainty for NSCLC or HNSCC, followed within 21 d by d-18F-FMT PET/CT, with no chemotherapy, radiotherapy, or immune/biologic therapy or biopsy performed or scheduled between the 2 scans. The final inclusion criterion was histopathologic or cytologic confirmation of malignancy.
Safety data were acquired for each patient 1 d before the scan, after the scan, and 1 d after d-18F-FMT PET/CT injection. Clinical follow-up data were collected for all patients.
Image Acquisition and Analysis
18F-FDG PET/CT Scan
All patients were first examined using a routine clinical protocol on an integrated PET/CT scanner (GE Healthcare DSTX, 64-slice CT [GE Health Systems], or Biograph mCT 4-64 PET/CT [Siemens]) with injection of 350 MBq of 18F-FDG 60 min before the examination. PET/CT was acquired from the mid thigh to the vertex of the skull with scan times of 2–3 min per frame. Every patient fasted for at least 4 h, and blood glucose levels were not elevated (<8 mmol/L). A low-dose unenhanced CT scan was obtained (80 mAs, 140 kVp, 0.5 s/tube rotation, pitch of 1.7, and slice thickness of 3.75 mm) (16).
d-18F-FMT PET/CT Scan
All patients were injected with 300 MBq of d-18F-FMT. For the first 6 patients, a 20-min dynamic PET acquisition was started at the tumor region right after injection. Afterward, 3 whole-body PET images were acquired after 30, 45, and 60 min. These data were used to determine the optimal imaging time point for d-18F-FMT. In an interim assessment before patient 7, it was decided to obtain 2 whole-body d-18F-FMT scans, the first at 30 min after injection and the second at 45 min, for patients 7–20.
Image Analysis
18F-FDG images were analyzed by 2 dual-board-certified nuclear medicine physicians and radiologists in consensus. d-18F-FMT images were analyzed by the same readers, with a direct comparison of both scans.
Lesions suggestive of or unequivocal for malignant involvement were visually judged for 18F-FDG and d-18F-FMT activity, with − indicating no increased uptake, + indicating just slightly increased uptake (more than background), and ++ indicating markedly increased uptake.
Furthermore, a quantitative analysis of 18F-FDG versus d-18F-FMT accumulation was performed using the maximum standardized uptake value (SUVmax) ratios of tumor to blood pool to determine the relative uptake into malignant tissue.
d-18F-FMT Preparation
The radionuclide 18F and the drug substance (R)-2-amino-3-(4-18F-fluoromethoxyphenyl)-propionic acid hydrochloride were produced at both clinical sites, starting from a precursor compound that had been produced following good manufacturing practice guidelines. d-18F-FMT was synthesized as previously described (11).
Statistical Analysis
Quantitative data are expressed as mean ± SD if normally distributed or as median with ranges if nonnormally distributed. Diagnostic performance, detection rate, and accuracy were determined by comparing the results of d-18F-FMT PET/CT and 18F-FDG PET/CT with the standard of reference on a per-patient and per-lesion base using the χ2 test for related variables. Patients were divided into groups above and below the median of the tumor-to-background ratio for d-18F-FMT and 18F-FDG, and Kaplan–Meier graphs were generated to determine associations between tracer uptake and overall survival for NSCLC and HNSCC (significance was determined using the log rank [Mantel-Cox] test).
RESULTS
Patient Characteristic
Thirty-nine patients were screened, of which 18 were not randomized. One patient was excluded after randomization before d-18F-FMT injection because of acute renal failure. Two patients were excluded after the d-18F-FMT scan, because the requested study dose of 300 MBq of d-18F-FMT was not achieved; both patients received only 200 MBq of d-18F-FMT (Fig. 1). The mean delay between the scans was 7 ± 5.1 d (with a range of 1–18 d).
Patient selection.
A total of 18 patients (13 men and 5 women), 10 with NSCLC and 8 with HNSCC, were included. They ranged in age from 44 to 77 y (median, 64 y). In total, 54 suggestive lesions were identified; 52 were 18F-FDG–positive, with a median tumor–to–blood pool ratio for SUVmax of 3.2 (range, 1.1–7.3). The absolute median uptake for 18F-FDG was an SUVmax of 8.9 ± 4.3 (range, 2.2–20.5). Thirty-four malignant lesions were d-18F-FMT–positive, with a median tumor–to–blood pool ratio for SUVmax of 1.5 (range, 1–3.2) and absolute values for SUVmax of 4.6 ± 1.6 (range, 2.1–8.7). Only 2 lesions were negative on 18F-FDG but detected on d-18F-FMT. In 37 lesions, pathologic findings were available as a standard of reference (27 lesions with histologic results and 10 with cytologic results). For 9 lesions, follow-up imaging was used as a standard of reference (3 with MR imaging of the spine, 5 with CT follow-up examinations, and 1 with an 18F-FDG PET/CT follow-up examination). Four lymph nodes, highly suggestive of metastatic involvement on the initial 18F-FDG PET/CT study, did not undergo further evaluation before radiotherapy (SUVmax median, 6.1; range, 5.1–11.5). Three lesions could be identified as inflammatory changes in tonsils on physical examination (Tables 1 and 2).
Clinical Data for NSCLC
Clinical Data for HNSCC
Safety
The injection of 307 MBq (SD, 10.2 MBq; range, 288–325 MBq) of d-18F-FMT was well tolerated. Safety data did not show any alteration in the sequential pre- and postdosing blood values, electrocardiograms, urine testing, or physical examinations.
Results for NSCLC
On a patient-based analysis, 18F-FDG PET/CT correctly staged 5 of 10 patients (50%) and under-staged 1 patient (10%): histopathology confirmed metastases in ipsilateral mediastinal lymph nodes, which were considered reactive on 18F-FDG PET/CT because they were only mildly enlarged and 18F-FDG–negative (Fig. 2). Four patients were over-staged (40%): in 2 patients increased 18F-FDG uptake in osseous lesions was considered suggestive, but follow-up MR imaging proved degenerative changes; in 1 patient pneumonia could not be differentiated from a secondary tumor (Fig. 3); and in 1 patient an 18F-FDG–active contralateral mediastinal lymph node was considered suggestive of metastasis but was negative on pathologic examination (Table 3).
Non–small cell lung cancer in right hilum with mediastinal lymph node metastases. (A and B) 18F-FDG PET with mild active primary tumor (large arrow) and 18F-FDG–negative lymph node metastases (arrowhead and small arrow). (C and D) d-18F-FMT PET with higher uptake in primary tumor (large arrow) and positive uptake in ipsilateral and infracarinal mediastinal lymph node metastases (arrowhead and small arrow). A and C are coronal views; B and D are axial views.
Non–small cell lung cancer in left lower lobe with ipsilateral mediastinal lymph node metastasis. (A and B) 18F-FDG PET/CT with primary tumor (large arrow) and 1 mediastinal lymph node metastasis in level 5 (small arrow). (B) Round 18F-FDG–avid consolidation in left superior segment with high 18F-FDG-activity interpreted as metastasis or secondary carcinoma (arrowhead). (C and D) d-18F-FMT PET with uptake in primary tumor (large arrow) and mediastinal lymph node metastasis (small arrow). (D) Consolidation in superior-lower-lobe segment is not d-18F-FMT–avid; on CT follow-up scan this lesion resolved within 2 wk after antibiotic therapy, consistent with pneumonia. A and C are coronal views; B and D are axial views.
NSCLC: 10 Patients
Six of 10 patients (60%) were correctly staged on the basis of the d-18F-FMT scan. Three patients were under-staged (30%): in those patients the primary tumor did not show any d-18F-FMT accumulation above the background level. Only 1 patient (10%) was over-staged, having a d-18F-FMT–positive lesion in the lumbar spine that was considered unequivocal by both readers.
In 10 patients with NSCLC, a total of 24 lesions were identified. On a lesion-based analysis, 18F-FDG PET/CT correctly identified 16 lesions as true-positive, and 13 lesions were true-positive on d-18F-FMT scans. Six lesions were false-positive with 18F-FDG: 3 were degenerative bone changes, 1 was a reactive lymph node, and 2 were pulmonary infiltrates. Two lesions were false-positive on d-18F-FMT, both corresponding to degenerative bone changes. Two lymph nodes were false-negative on 18F-FDG PET, and 5 lesions were false-negative on d-18F-FMT: in 3 cases the primary tumor showed no d-18F-FMT accumulation above the blood pool level, and in 2 cases lymph nodes were missed on d-18F-FMT PET/CT.
For patients with NSCLC, the overall lesion-based analysis for 18F-FDG/d-18F-FMT showed a detection rate of 89%/72% (χ2, P = 0.453), and the accuracy was 67%/71% (χ2, P = 0.751).
Results for HNSCC
On a patient-based analysis, 18F-FDG PET/CT correctly staged 4 of 8 patients (50%) and over-staged 4 patients (50%). In 3 patients high 18F-FDG uptake in the tonsils was considered suggestive, but clinical examination confirmed only inflammatory changes. These tonsils were all d-18F-FMT–negative. On d-18F-FMT, 5 of 8 patients (63%) were correctly staged. Three patients were under-staged (38%): in 1 patient the primary tumor did not show any d-18F-FMT accumulation above the blood pool level, and in 2 patients contralateral or mediastinal lymph nodes were d-18F-FMT–negative.
A total of 30 suggestive lesions were identified for the 8 patients with HNSCC. On a lesion-based analysis, 18F-FDG PET/CT correctly identified 26 lesions as true-positive; on d-18F-FMT scans, 21 lesions were true-positive (Fig. 4). Four lesions were false-positive for 18F-FDG: they consisted of 3 inflammatory changes in the tonsils and 1 reactive lymph node. Five lesions were false-negative on d-18F-FMT: in 1 case the primary tumor showed no d-18F-FMT accumulation above the blood pool level, and in 4 cases lymph node metastases were missed on d-18F-FMT PET/CT (Table 4).
Squamous cell carcinoma of left tonsil with 1 large lymph node metastasis (4 cm), left level IIA. (A and B) 18F-FDG PET with active primary tumor (arrow) and clearly positive lymph node metastasis (arrowhead). (C and D) d-18F-FMT PET with high uptake in primary tumor (arrow) and equally high uptake in cervical lymph node metastasis (arrowhead). A and C are coronal views; B and D are axial views.
HNSCC: 8 Patients
For patients with HNSCC, the overall lesion-based analysis for 18F-FDG/d-18F-FMT showed a detection rate of 100%/81% (χ2 rejected because there were no false-negative cases for 18F-FDG) and an accuracy of 83%/81% (χ2, P = 0.337).
Taking all 54 lesions into consideration, a detection rate and accuracy of 95% and 78%, respectively, for 18F-FDG and 77% and 78%, respectively, for d-18F-FMT (χ2, P = 0.432/P = 0.600) were achieved.
PET and Survival
After a mean follow-up of 21.6 mo (range, 4.2–40.7 mo), of the 18 patients 8 were alive with disease and 10 had died of disease. The median tumor–to–blood pool ratio was 3.2 for 18F-FDG and 1.5 for d-18F-FMT. Using this as a cutoff, Kaplan–Meier analysis showed a significant negative association between overall survival and d-18F-FMT (P < 0.05) but not between overall survival and 18F-FDG (P = 0.93) (Table 5; Fig. 5).
Kaplan–Meier Survival for d-18F-FMT and 18F-FDG
Kaplan–Meier curves for overall survival using median tumor–to–blood pool ratio as cutoff for d-18F-FMT (A) and 18F-FDG (B) in 18 patients with NSCLC and HNSCC, illustrating that high d-18F-FMT uptake in both tumor entities might have higher association with adverse outcome than 18F-FDG.
DISCUSSION
This study proved that tumor imaging with d-18F-FMT is safe and feasible for NSCLC and for HNSCC.
The hypothesis that d-18F-FMT would lead to an increased tracer accumulation compared with its l-isomer was further supported clinically in NSCLC patients, with a mean SUVmax of 4.5 ± 1.5 for d-18F-FMT compared with 1.8 ± 1.0 for l-18F-FMT in the literature (17). SUVs, however, are highly variable and depend on scanner properties, dose application, patient weight, and image reconstructions. Therefore, the value of a direct comparison with published l-18F-FMT data is limited. d-18F-FMT could not reach as high a tumor–to–blood pool level as that achieved with 18F-FDG. d-18F-FMT was able to detect more malignant lesions than 18F-FDG in only 1 patient (5%), whereas relevant lesions were missed in 6 patients (33%). On the other hand, 18F-FDG results were false-positive in 10 lesions overall, whereas d-18F-FMT was false-positive in only 2 osseous lesions. MR imaging showed degenerative changes in both areas, with no evidence of malignancy, suggesting a tendency toward better differentiation between malignant and inflamed tissue. The reason for increased d-18F-FMT uptake in sclerotic bone lesions is unclear.
Many of the false-positive findings on 18F-FDG scans were due to inflammatory or infectious changes in tonsils, pulmonary infiltrates, or lymph nodes. Although further investigation (clinical examination or biopsy) could help to elucidate the cause of these false-positive findings, further examinations and more tests would be required for those end-stage patients. Especially in areas with endemic granulomatous disease, high uptake in mediastinal lymph nodes can lead to false-positive findings, thereby over-staging patients with NSCLC. A PET tracer with higher specificity would therefore be especially desirable in these areas. The usefulness of l-18F-FMT in sarcoidosis patients has been suggested by Kaira et al., emphasizing the higher specificity of LAT-1 imaging than of 18F-FDG PET for malignant versus inflammatory disease (18).
l-18F-FMT has not yet been used clinically in HNSCC patients. The introduction of l-18F-FET for HNSCC, however, yielded promising results for the primary tumors (6). l-18F-FET yielded a mean SUVmax of 4.4 ± 2.2 in the primary tumor but only 1.4 ± 0.3 in lymph node metastases. Our results showed a mean d-18F-FMT SUVmax of 6.2 ± 1.6 in primary HNSCC tumors. d-18F-FMT was false-negative in only 4 lymph node metastases (size, 1.6 ± 1.4 cm; range, 0.7–3 cm) and positive in 15 lymph node metastases, with a mean SUVmax of 4.6 ± 1.4. These results further support the hypothesis of a favorable tumor uptake for d-18F-FMT, compared with l-isomers and extended alkyl chains.
Interestingly, we also observed a tendency for NSCLC, with fewer false-negative cases for squamous cell cancer than for adenocarcinoma; this finding was also observed for l-18F-FMT in a larger cohort of 50 NSCLC patients studied by Kaira et al. (19). In a subsequent study, they further investigated l-18F-FMT in 98 patients with NSCLC. In this study group, SUVmax for l-18F-FMT was an independent and significant predictor of overall survival, whereas 18F-FDG did not reach significance in a multivariate analysis (17). This study could confirm a significant negative correlation between a high tumor–to–blood pool ratio for d-18F-FMT and overall survival (P = 0.05), whereas the tumor–to–blood pool ratio for 18F-FDG did not reach significance (P = 0.93). However, these results have to be viewed with caution and may show only a possible trend in this heterogeneous cohort of only 18 patients, taking into account that univariate analysis using dichotomous cutoffs might further reduce the significance of the results.
Additional studies with larger patient cohorts will help to further substantiate the hypothesis of a direct correlation between tumor–to–blood pool ratio and survival prognosis. The main limitation for studies in patients with multiple metastases and for this study in particular is the limited access to histopathologic samples on a lesion-by-lesion basis to define accurate staging.
CONCLUSION
This first-in-humans study showed the feasibility and safety of d-18F-FMT imaging in NSCLC and HNSCC. The results showed a trend toward a lower sensitivity for d-18F-FMT than 18F-FDG along with a higher specificity due to the lack of uptake in inflammatory tissues. This finding may lead to improved clinical management of patients in areas with endemic granulomatous disease.
Furthermore, a significant correlation between high d-18F-FMT uptake and overall survival could be shown. Future studies with larger cohorts are needed to show the clinical value of d-18F-FMT for survival prediction and other tumor entities, particularly in comparison to 18F-FDG.
DISCLOSURE
The costs of publication of this article were defrayed in part by the payment of page charges. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734. Matthias Friebe, Andrew Stephens, and Ludger Dinkelborg are affiliated with Piramal Imaging. Kristin Kowal, Sabine Zitzmann-Kolbe, and Keith Graham are employed by Bayer HealthCare. Research support was provided by Bayer Healthcare. No other potential conflict of interest relevant to this article was reported.
Acknowledgments
We thank the nuclear pharmacy staff, the technologists, and the administrative staff at the University Hospital of Zurich and University Medical Center Groningen for their help in acquiring the data and freeing up resources. Specifically, we thank Josephine Trinckauf, Verena Weichselbaumer, and Miriam de Bloehme for the data acquisition. Furthermore, we thank Prof. Tsukada of Hamamatsu Photonics for his help and support.
Footnotes
Published online Sep. 25, 2014.
- © 2014 by the Society of Nuclear Medicine and Molecular Imaging, Inc.
REFERENCES
- Received for publication June 7, 2014.
- Accepted for publication August 14, 2014.
Jump to section
Related Articles
Cited By...
- No citing articles found.