Diagnosis of Giant Cell Tumor of the Tendon Sheath Using Multiecho Gradient-Echo Sequence: the “ Superblooming Artifact ”

Giant cell tumor of the tendon sheath is a tumor, which a ﬀ ects mainly the hands of people aged 30 to 50 years with a female prevalence. Magnetic resonance imaging (MRI) is essential for diagnosis and treatment planning. The aim of this study is to evaluate the e ﬃ cacy of multiecho gradient-echoes (MeGE) sequence in detecting hemosiderin which is the hallmark of this tumor. MRIs were performed in a sample of 11 patients with a mean age of 45. With the proposed protocol, all readers were able to detect the susceptibility artifacts due to the presence of hemosiderin. MeGE sequence allows to highlight the presence of hemosiderin, and the use of 3 echo times (ET) (8, 16, and 24 ms) is suggested.


Introduction
Giant cell tumor of the tendon sheath (GCTTS) is the second most common lesion of the hand after ganglion cysts [1,2]. The tumor affects subjects aged between 30 and 50 years old, with a female preponderance [3][4][5][6][7]. GCTTS is a slowly growing, usually painless benign lesion with a not trivial rate of recurrence (about 10%) and small risk to develop a malignant behavior [8,9].
The microscopic features of GCTTS are foamy histiocytes, giant cells, large synovial cells, and hemosiderin deposits both intracellular and extracellular. Magnetic resonance imaging (MRI) is the most useful imaging modality for both diagnostic and surgical planning purposes [10], in particular, the MRI detection of intralesion artifacts due to hemosiderin deposits is the diagnostic hallmark of this tumor.
Gradient-echo (GE) sequences are the best diagnostic acquisitions in demonstrating the presence of hemosiderin, which appears as a loss of signal (the so-called blooming artifact). It is a susceptibility artifact, a distortion of the local magnetic field that makes small lesions more conspicuous [11][12][13][14][15][16][17][18].
However, to the best of our knowledge, no previous study has investigated the optimal GE parameters to use for improving hemosiderin detection. The aims of this work are the evaluation of a multiecho GE (MeGE) sequence effectiveness in detecting hemosiderin within GCTTSs and the assessment of the best echo time for GCTTS's diagnosis.

Study Population.
The present study retrospectively reviewed the MR images of 11 patients affected by surgically treated and histologically proven GCTTSs of the hands.
The 11 patients included two men and nine women, aged between 17 and 62 years (median 45 years). All patients manifested clinically as painless, soft, slow-growing tissue lumps without local redness, swelling, or ulceration. MRIs were performed over a 10-year period, between January 2010 and January 2020.

Images Analysis.
After a previous anonymization and randomization process, MR images were evaluated by two radiologists (M.G., A.B., both with more than 20 years experience in MRI and musculoskeletal radiology), who were aware of the diagnosis of GCTTSs but were blind to the surgical findings.
The evaluation was divided in three different time points at intervals of one month to reduce bias due to memory of the previous analyses: (i) First, only nonmultiecho GE (non-MeGE) were studied (ii) Afterwards, only MeGE, in random order, were evaluated (iii) Finally, MeGE images of each exam, ordered from lowest to highest echo time, were analyzed.
The readers were asked to assess the following items: (1) Presence of hemosiderin in non-MeGE sequences (yes/no) (2) Presence of hemosiderin in MeGE images, considering separately each echo time (yes/no) (3) Confidence in distinguishing hemosiderin from fibrosis using increasing "blooming artifact" at different echo times (yes/no) on MeGE scans. This blooming increasing effect has been named "superblooming sign." In addition, two different board-certified radiologists (G.C., S.M.) independently performed image quality assessment (good/inadequate) of MeGE acquisitions at each ET in a fourth separate session.
Image quality assessment was based on tumor signal intensity, identification of tumor edges, and low resolution due to susceptibility effect.

Statistics.
The agreement of the two readers has been evaluated for each point 1, 2, and 3 and for image quality.

Results
The agreement for detection of blooming between the two readers is summarized in Table 1.
Non-MeGE images were not sensitive in detecting of blooming artifact. Only Reader 2 was confident enough to detect blooming on T2-weighted images in 2 patients. In all the cases, both readers were able to detect blooming on MeGE 24 ms, 32 ms, and 40 ms. The best echo comparison for superblooming was between GE 8 ms and GE 24 ms which allowed to detect this sign in 100% of the patients for both readers. For image quality, GE with ET at 32 ms and 40 ms were evaluated as inadequate by both readers due to large blooming artifact and low signal-tonoise ratio.
The lesion appears strongly hypointense in the 8 ms ET image. Also, the readers considered this image positive for the presence of blooming. However, both agreed that the clear increase of the blooming effect (superblooming) in higher ET images make the diagnosis more confident. For comparison, look at the collagen of the tendons which do not show superblooming effect making differential diagnosis between GGTTS and a possible fibrous containing tumor very reliable.
In addition, note that images at 32 and 40 ms are affected by a very low signal-noise ratio and by a large blooming artifact, both degrading the images in a severe way.
In the inset (A'), the presence of a subtle peripheral capsule (empty arrow) and thin fibrous septa (arrowheads) within the nodule could be appreciated. An area of slight low signal intensity can be seen in the left corner of the lesion (white arrow). Both the readers agreed that it was not possible to obtain the difference between fibrosis and blooming artifact. However, the other echoes show with greater advantage by increasing the blooming artifact in this area (superblooming) making diagnosis of hemosiderin deposition very confident. Note that the more ventral septa show a slight superblooming effect on images at 24, 32, and 40 ms, demonstrating that small amounts of hemosiderin can be detected only evaluating the presence of this sign.

Discussion
Magnetic susceptibility "blooming" artifact within GCTTS on GE images is related to the local magnetic field created by intra-and extracellular hemosiderin deposition [19]. Hemosiderin is a superparamagnetic material consisting of conglomerates of denatured proteins, lipids, and particles containing iron released by the breakdown of hemoglobin. Superparamagnetism is more closely related to ferromagnetism than to paramagnetism. GE scans are based on only a single RF pulse, with a flip angle lower than 90°, in combination with a readout gradient reversal. This results in short TR and TE and consequent fast scan times [20].
Hemosiderin causes intense spinning nuclei dephasing with a marked T2/T2 * shortening and a consequent dark appearance within accumulation areas. When longer TEs are used on GE scans, despite a general drop of signal intensity, a higher susceptibility to field inhomogeneity is determined, and therefore, hemosiderin deposits within GCTTSs can be easily detected [20].
However, the amount of hemosiderin is usually a much less prominent feature of tendon sheath disease than of its intra-articular counterpart also known as pigmented villonodular synovitis [21].
In addition, deposition variability of hemosiderin throughout the lesion is common in GCTTSs, which can contain foci of brighter T1 and T2 signals that indicate relatively low concentrations of hemosiderin.
On T2-weighted images, blooming may be difficult to appreciate, unless the amount of hemosiderin is massive. GCTTSs are surrounded by a collagen capsule, which spread into the tumor through septa dividing the lesion in lobules.

BioMed Research International
The presence of collagenous stroma can cause hypointensity within the tumor on MRI [11,22].
The histological diagnosis of GCTTSs is rarely difficult but may be confused with features of other soft tissue tumors, particularly fibroma of the tendon sheath (FTS). Similarly, the clinical and imaging characteristics of FTS closely resemble those of GCTTSs.
However, the main drawback of this work is the lack of GE sequences in the MRI protocol. Since multiecho GE images can easily distinguish between fibrosis and hemosiderin, it could be supposed that differential diagnosis accuracy could reach 100% using this sequence. Nevertheless, considering that only GCTTs and not fibromas were evaluated in our work, further studies are necessary to confirm this hypothesis.

Conclusion
In conclusion, MeGE sequences allow detection of hemosiderin with great accuracy. The comparison between low ET and higher ET GE T2 * sequences allowed in every case to confirm the presence of hemosiderin with complete agreement (high confidence) between readers. Since high ET GE T2 * images beyond 24 ms showed a low image quality, the use of 3 ETs (8, 16, and 24 ms) is the better choice for assessment of suspected GCTTSs of the hands.

Data Availability
The manuscript data used to support the findings of this study are included within the article.

Conflicts of Interest
The authors declare that they have no conflicts of interest.