Imaging of Soft-Tissue Tumors Using l-3-[Iodine-123]Iodo-α-methyl-tyrosine Single Photon Emission Computed Tomography: Comparison with Proliferative and Mitotic Activity, Cellularity, and Vascularity¹

Soft-tissue sarcomas are a heterogeneous group of malignant tumors that can arise from mesenchymal structures at any site of the body. These tumors constitute 1% of all cancers. A large proportion (60%)is located in an extremity. They often reach a large size before a diagnosis is established. They grow locally aggressive, frequently invade surrounding tissues, and often disseminate to distant sites (1). Treatment planning depends on information regarding the presence or absence of metastases, the local situation, and histological parameters. Chest CT³ is commonly used to screen for pulmonary metastases and magnetic resonance imaging to assess the local situation and resectability. The most important prognostic factor is tumor malignancy grade, i.e., a higher tumor grade, is associated with a worse prognosis (1, 2, 3, 4).

Considerable interest exists in noninvasive determination of the malignancy grade of sarcomas (5, 6, 7, 8, 9). For this purpose,various nuclear medicine techniques, such as SPECT and more recently PET, using various tracers, have been applied. Radiolabeled amino acids are potentially suitable tracers; because of the increased protein metabolism of malignant tumors, the uptake in malignancy is increased (10). Furthermore, amino acid uptake is less disturbed by uptake in inflammatory tissue than the frequently used tracer FDG (11, 12, 13, 14, 15). However, most carbon-11-labeled amino acids are difficult to synthesize and require the infrastructure of PET.

The radiolabeled amino acid IMT is avidly taken up in many tumors (16, 17). The agent has been introduced for imaging of brain tumors where tumor uptake was shown to represent amino acid transport activity, an important step in protein metabolism (18, 19, 20, 21). Because at present IMT is the only amino acid tracer suitable for SPECT, imaging with this tracer might combine the specificity of amino acids with the wide availability of SPECT.

The goals of this study were to determine whether: (a) IMT SPECT can visualize soft-tissue sarcomas; (b) IMT uptake can discriminate between benign and malignant tumors; and (c)uptake is related to tumor grade and the degree of proliferation, using the mitotic index and the Ki-67 proliferation index as markers.

Patients referred to the University Hospital Groningen between October 1997 and February 1999 for management of a suspected soft-tissue sarcoma were asked to participate in this study. Suspicion for malignancy was based on rapid growth, size, consistency at palpation,location, and radiographic appearance of the tumor. Selection criteria included age >18 years and planned histological confirmation within 2 weeks after the IMT study.

Thirty-two patients [16 males and 16 females; median age, 45 years(range, 18–82)] were included in the study. Patient and tumor data are presented in Table 1. All patients were studied 1–2 weeks prior to biopsy or definitive resection. All patients were clinically staged with a chest CT, magnetic resonance imaging, or CT of the soft-tissue mass as well as ultrasound of the liver and bone scintigraphy. The study was approved by the Medical Ethics Committee of the University Hospital Groningen. Written informed consent was obtained from all patients.

Synthesis of IMT was carried out slightly modified from Krummeich et al. (22). Briefly, Iodogen iodination with Na¹²³I [specific activity, >185 TBq/mmol (5000 Ci/mmol); Amersham Cygne, Eindhoven, the Netherlands] of the precursor l-α-methyl-tyrosine was performed in a borate buffer. IMT was purified by elution with saline containing 5% ethanol over a C-18 SepPak cartridge (Waters, Milford, MA) preconditioned with 100% ethanol, followed by saline containing 5% ethanol. After filtration through a sterile 0.22 μm Millex GV filter (Millipore,S. A., Molsheim, France), a colorless ready-to-inject solution was obtained. Samples were demonstrated to be sterile and pyrogen free. Quality control was performed by high-performance liquid chromatography on a RP-18 column (Multosorb 100 4.6) using H₂O:ethanol:acetic acid (87.5:10:2.5, v/v/v %)as eluent. Radiochemical purity was >99% in all cases. The overall synthesis time, including purification and quality control, was <1 h. Radiochemical yield was 50–65%.

Patients were injected i.v. with 200–300 MBq IMT after at least a 4-h fast. They were given 10 drops of Lugol’s solution p.o. 15 min before injection to prevent thyroid uptake of possibly formed free ¹²³I. A large-field-of-view, double-headed gamma camera (MULTISPECT 2; Siemens, Inc., Hoffman Estates, IL) was used with a medium-energy, all-purpose collimator and a 15% window centered on the 159 kV photopeak of ¹²³I. System resolution was 12 mm full-width at half-maximum at 10-cm distance.

Fifteen min after injection, SPECT imaging of the tumor area was performed, usually followed by a planar spot view. After imaging the tumor area, additional spot views were recorded to obtain whole-body information. In 20 patients, the thorax was additionally imaged using SPECT.

All SPECT acquisitions included 64 views (2 × 32; 5.6°/step) of 30 s duration each in a 128 × 128 matrix format with a zoom factor of 1.23. This corresponds to a pixel dimension of 4.1 mm. Transaxial tomograms were reconstructed without prefiltering using filtered back-projection with a Butterworth filter of sixth order and a cutoff frequency of 0.275 Nyquist. Ten-min spot views were recorded in a 128 × 128 matrix. Total imaging time was ∼1 h. The reported radiation burden of IMT is 0.007 mSv/MBq, yielding an effective dose equivalent of 1.4–2.1 mSv (23).

Without knowledge of the staging and histopathological data, two experienced readers analyzed all images visually for tumor uptake and abnormal extratumoral uptake. ROIs were manually placed over SPECT tumor slices with maximum visibility and on planar spot views. Reference ROIs were placed over representative background muscle tissue, usually contralaterally. ROIs were drawn at 80% of the maximal pixel value around the lesion under study; in some cases, this procedure resulted in more than one ROI per tumor (24). An example is presented in Fig. 1. The T:B ratio was calculated by dividing uptake intensity (counts/pixel) in the tumor ROI by uptake intensity in the background ROI.

All IMT scintigraphic findings were compared with the results of conventional imaging. T:B ratios were compared with tumor grade,mitotic index, tumor cellularity, and vascularity and the Ki-67 proliferation index.

The histological diagnosis was made on H&E-stained paraffin sections with or without additional immunohistological stains. All tumors were classified according to Enzinger and Weiss (1) into 10 different histological types. Soft-tissue sarcomas were graded according to the grading system of Coindre et al.(25), in which points are assigned to differentiation level, mitotic index, and necrosis.

In 27 of the 32 lesions, determination of the mitotic index,Ki-67 proliferation index, cellularity, and vascularity was performed. In the five remaining cases (three benign and two malignant), there was not enough material left for these determinations. The number of mitotic figures per 2 mm² was counted in H&E-stained paraffin sections. Proliferating cells were detected using the monoclonal antibody MIB-1 (Immunotech S. A., Marseille, France),which recognizes an epitope of the Ki-67 antigen. This is a nuclear antigen present in all phases of the cell cycle, except for the G₀ phase, in which the cells are withdrawn from the cell cycle, and the early G₁ phase, the phase before the start of DNA synthesis (26). Immunohistochemistry was performed on paraffin sections according to a method modified from Shi et al. (27) and Emanuels et al. (28). In brief, after heating on a hot plate, the slides were dewaxed in xylene and rehydrated in serial ethanol washes (100, 96, and 70%). After heating twice in an autoclave for 10 min at 110°C in boiling solution (pH 6.0), the slides were incubated with a 1:400 dilution of the antibody in BSA (pH 7.4). The primary antibody was detected with a biotinylated secondary antibody (multilink), followed by a streptavidin-alkaline phosphatase conjugate (Ready-to-Use Link and Label; Biogenex, San Ramon, CA). Final color was developed using bromochloroindolyl-phosphate 4-nitroblue-tetrazolium chloride (Boehringer Mannheim, Mannheim,Germany).

To measure the Ki-67 proliferation index and tumor cellularity, we used ocular micrometry on a Leica microscope (Rijswijk, the Netherlands)with an eyepiece grid at ×400. Fifteen fields were randomly selected throughout histologically viable areas. Tumor cellularity (cellular density) was defined as the total number of cells excluding endothelial cells, inflammatory cells, and necrosis and presented as number per 2 mm². MIB-1-positive nuclei were identified, and the number of these nuclei was divided by the total number of nuclei in each of the 15 fields to calculate an index/field. The Ki-67 proliferation index (also named Ki-67 labeling index), representing proliferative activity, was defined as the mean of the indices of the 15 fields. Vascularity in each tumor was estimated by quantification of the total number of small blood vessels or parts of large vessels in H&E-stained paraffin sections. Ten fields were randomly selected, and the total number of vessels in 10 fields at ×400 (corresponding to 2 mm²) was counted.

To determine the ability to distinguish malignant from benign tumors, a Student’s t test was performed on T:B ratios (normally distributed) in both groups. Spearman’s correlation coefficient was used to correlate T:B ratios with Ki-67 proliferation index, mitotic index, tumor vascularity, and cellularity. Kruskal Wallis nonparametric ANOVA with Dunnett’s T3 post-hoc multiple comparisons test was performed to relate tumor grade with T:B ratios. Two-tailed Ps < 0.05 were considered significant.

Eleven patients were found to have benign lesions; the remaining 21 had soft-tissue sarcomas. Three patients had been treated previously for a soft-tissue sarcoma (2, 3, and 4 years earlier) and were now diagnosed to have a local recurrency. Eighteen tumors were localized in an extremity (56%), and 14 in the trunk (44%). Tumor size was between 2 and 23 cm (median, 8.6 cm). There were 4 low-grade (grade I), 5 intermediate-grade (grade II), and 12 high-grade (grade III) sarcomas(Table 1). Benign lesions were assigned grade 0.

All 21 soft-tissue sarcomas were visualized on the IMT SPECT images. Tumors in the extremities and in the chest were well visualized because of high tumor uptake and low background uptake, whereas tumors near excretory systems (kidneys, bladder, and ureters) were more difficult to evaluate. IMT uptake in tumors was often inhomogeneous, in accordance with the typical histological appearance of sarcomas where areas of necrosis and hemorrhage are interspersed with areas containing vital tumor cells. Typical examples are presented in Figs. 2 and 3.

In 5 of the 11 benign lesions no IMT uptake was found; in four, faint uptake was found, and in two, avid uptake. No uptake was found in all five lipomas, and faint uptake in two large intramuscular hemangiomas,in an aspecific fibrotic lesion, and in an aggressive fibromatosis lesion. Avid uptake was present in a growing neurofibroma and in another aggressive fibromatosis lesion (the latter is considered“borderline malignant”).

The average T:B ratio in benign tumors was 1.52 ± 0.60, whereas IMT uptake in the malignant tumors was much higher: 3.83 ± 1.16(P < 0.001; Fig. 4). There was minimal overlap in individual values, attributable to avid uptake in the aggressive fibromatosis and in the growing neurofibroma(Fig. 4). When a T:B ratio of 2.0 was chosen as the cutoff point,sensitivity for the detection of malignancy in a soft-tissue mass was 100% with a specificity of 88%. Using 2.7 as the cutoff point,sensitivity was 89%, but specificity was 100%.

T:B ratios correlated with histological grades (r =0.82; P < 0.001); IMT uptake increased with higher tumor grade. Post-hoc analysis revealed that the difference between benign tumors versus grade 1, 2, and 3 tumors caused the significance. Grade 1 versus grade 2 and grade 2 versus grade 3 could not be distinguished from each other. However, low-grade tumors could be discriminated from benign tumors in this group (Fig. 5). In addition, benign and grade 1 tumors together had lower IMT uptake than grade 2 and 3 tumors together (P < 0.0001).

Interestingly, IMT uptake correlated with markers of cell proliferation; IMT T:B ratio versus the Ki-67 proliferation index gave r = 0.75, P < 0.001 (Fig. 6,A) and versus the mitotic index r = 0.63, P < 0.01 (Fig. 6,B). IMT uptake correlated also with tumor cellularity(r = 0.73 P < 0.001; Fig. 6,C), whereas no relation with tumor vascularity could be established (Fig. 6 D).

Because whole-body images were available in all patients, the ability of IMT scintigraphy to detect metastatic disease was also studied. But because of the low number of metastasized tumors, no reliable assessment of staging could be obtained. Only two patients in this series were found to have metastatic disease; one had small pulmonary metastases (2 mm on chest CT) that were negative on the IMT chest SPECT study, and the other patient had multiple small (<5 mm)intra-peritoneal metastases, negative on all preoperative investigations but detected during surgery.

This study demonstrates high uptake of IMT in soft-tissue sarcomas and low uptake in benign processes. All malignant tumors were visualized, leading to 100% sensitivity. Tumors in the direct vicinity of bladder or kidneys were somewhat more difficult to assess because of renal excretion of IMT causing high background activity. IMT uptake could differentiate benign from malignant tumors with high accuracy. Uptake increased with higher tumor grade, higher Ki-67 proliferation index, mitotic activity, and tumor cellularity, whereas no correlation with tumor vascularity was found. Therefore, this relatively simple noninvasive technique provides information on parameters of tumor activity.

To understand what processes are visualized by IMT uptake, it is important to know the mechanism of uptake. Using competition experiments in vivo, Langen et al.(19) demonstrated that IMT is a substrate for the amino acid transport systems in the blood-brain barrier. This has not been demonstrated for tumors outside the brain yet. However, in vitro data suggest that also in non-glioma tissue, uptake of IMT is almost entirely mediated by amino acid transport activity (29). Because in the present study no correlation with the amount of tumor vessels was found, tumor perfusion is not the dominating factor that governs IMT uptake, as was suggested by Deehan et al. (30) in a rat model. One could argue that uptake in the two hemangiomas must be related to blood flow and diffusion, but these tumors were also very cellular. The presumed mechanism of uptake of IMT is the accelerated protein metabolism and the resulting increased demand for amino acids. This is further supported by the relation with tumor proliferation factors, as found in this study.

To noninvasively determine the malignancy grade in soft-tissue tumors,magnetic resonance spectroscopy (9) and various nuclear medicine techniques and tracers have been used, such as ²⁰¹Tl, ^99mTc-MIBI, ^99mTc-MDP, and ⁶⁷Ga (31, 32), and more recently FDG and radiolabeled amino acids, such as l-[1-¹¹C]tyrosine (6, 7, 33, 34). Different biochemical processes in tumors are visualized by all these tracers; most likely ²⁰¹Tl uptake is related to cellular Na⁺K⁺-ATPase activity, ^99mTc-MIBI to membrane functions and mitochondrial integrity, MDP to calcium/phosphate metabolism, and ⁶⁷Ga to transferrin receptors in tumor cells or inflammatory cells. In all cases, there is more or less effect of tumor vascularization. Sensitivity of tumor detection using IMT appears better than other single photon tracers in soft-tissue sarcomas, and the presumed uptake mechanism might more directly reflect tumor viability than the above-mentioned mechanisms, but our results need to be confirmed by other studies.

Many studies have used PET to study sarcomas. FDG generally shows high uptake in tumors and has proven to be of value in the visualization and grading of soft-tissue sarcomas. In the study by Nieweg et al. (6), all 18 sarcomas were visualized using FDG PET and a correlation was found between the calculated glucose consumption and tumor grade. Schulte et al. (8)found high sensitivity using FDG PET for the detection of malignancy in 102 patients, but lower specificity caused by high uptake in aggressive benign tumors and in a patient with myositis ossificans. Uptake of l-1-[¹¹C]tyrosine also showed a correlation with tumor grade, mitotic rate, and proliferation in soft-tissue sarcoma (7, 34). Because of lower uptake in inflammatory reactions caused by radiotherapy, systemic chemotherapy,or regional cytostatic perfusion, it is suggested that radiolabeled amino acids are better suited to monitor treatment effects than FDG (7, 11, 33). The results of the present study using IMT SPECT are well in line with these PET studies, with somewhat higher correlation factors than reported for l-[1-¹¹C]tyrosine (7, 34).

Almost all studies on soft-tissue tumors have found some uptake in benign lesions, regardless of the used tracer (6, 7, 8, 31, 32). Also using IMT we found minor uptake in hemangiomas and more avid uptake in a neurofibroma and in aggressive fibromatosis lesions. These latter are actually considered to be “borderline malignant” and therefore presumably have increased amino acid metabolism leading to IMT uptake. Similar to our observations, Kuwert et al. (35) found some IMT uptake in nonneoplastic brain lesions such as inflammatory lesions and infarctions. Recently, very high uptake was described in a low-grade desmoplastic infantile ganglioglioma, another benign neoplasm (36). Therefore, it seems most likely to conclude that IMT is an aspecific tumor tracer that targets amino acid metabolism in tumors. Difference in uptake between malignant and benign disease is based on the degree of accelerated protein metabolism.

In this study, we found IMT uptake also to be related with the number of tumor cells. Whether this also holds true for other tracers is largely unknown. Because cellular density can be high in tumors, the mere fact of many cells/volume (in comparison with adjacent tissues)could already cause increased uptake. However, it seems unlikely that this effect only would result in T:B ratios between 3 and 6, as found in many high-grade tumors. Moreover, others have found no relation between IMT uptake and cellular density in brain tumors (37).

The heterogeneity of soft-tissue sarcomas is a frequently occurring difficulty in analyzing the relationship between tracer uptake and histological parameters. This heterogeneity may cause sampling errors in the histological evaluation. In the present study, IMT uptake was based on a standardized ROI analysis encompassing the entire tumor. Histopathological parameters, however, were based on a representative tissue sample obtained from the removed tumor (or large biopsy). Although we tried to reduce sampling errors as much as possible, by analyzing randomly selected microscopic areas within the sample, some degree of sampling error could still be possible. This may influence the significance of the correlations.

Although our results appear comparable with those using FDG and l-[1-¹¹C]tyrosine PET, it could be argued that the lower resolution and limited quantification possibilities of SPECT make the method inferior to PET. However, in these generally large tumors, the lower resolution is presumably not a large problem, although partial volume effects may decrease tumor heterogeneity. Limited quantification possibilities are also controversial, because some authors successfully use T:B ratios also in FDG PET studies (8), whereas others disagree as to what PET uptake parameters give the best results in separation of benign from malignant disease (6, 38, 39, 40). Furthermore, absolute quantification of protein synthesis (as possible in PET) both in tumor tissue and in “background” tissue may be disturbed by systemic influences and vary largely from day to day (33). The use of ratios might compensate for this effect.

The good correlation of IMT uptake with parameters of tumor activity may be of interest in evaluation of new treatment strategies such as angiogenesis inhibition, matrix metalloproteinase inhibition,or antisense therapy. These new drugs exert their action through new mechanisms and may require new evaluation methods. In these methods,assessment of metabolic tumor activity may complement the classic response parameter, the change in tumor size. In current routine clinical practice, however, noninvasive studies of tumor activity or its benign/malignant nature is of limited value. In general, patients and physicians rely on histological evidence, and only in exceptional cases will these studies influence diagnostic or therapeutic strategies, although the good separation between grade 1 versus higher grade tumors may assist in determination of the prognosis.

In conclusion, IMT scintigraphy clearly visualized all soft-tissue sarcomas in this group. Uptake in malignant tumors was higher than in benign lesions, leading to a sensitivity for the detection of malignancy of 100% at a specificity of 88%. Uptake increased with higher tumor grade and higher proliferation rate. Therefore, IMT SPECT noninvasively provides information about tumor activity.