Soft tissue tumours represent a diverse group of mesenchymal lesions which often present diagnostic challenges to clinicians and pathologists. Histological classification of these tumours is based on their degree of differentiation and metastatic potential: benign, intermediate (locally aggressive), intermediate (rarely metastasising), and malignant.1 Recent advances in molecular cytogenetics (fluorescence in-situ hybridisation [FISH]) and molecular assays (reverse transcription–polymerase chain reaction [RT-PCR]) have contributed to the ever-evolving nature of classification and diagnosis of soft tissue sarcomas. Over the past two decades, conventional karyotyping has demonstrated diagnostic utility in detecting a wide range of recurring numerical and structural chromosomal aberrations in soft tissue tumours.

Unlike the newer molecular techniques such as FISH, knowledge of the expected genetic change is not required and this enables karyotyping to function as a genome-wide screening tool. Furthermore, karyotyping can detect any further clonal progression in the event of a tumour relapse. The drawbacks of karyotyping include the dependency on sterile tumour specimens, success of growth culture, and being time-consuming.2

Histology, immunohistochemistry, and electron microscopy may sometimes show borderline or non-specific features. An example is that of malignant peripheral nerve sheath tumours which have been historically difficult to distinguish from other spindle cell sarcomas such as synovial sarcomas.3 Karyotyping has shown the main difference to be the presence of the (X;18) translocation.3 Many previous studies4 5 6 have also demonstrated the role of conventional karyotyping in the detection of clonal aberrations in 68% of malignant fibrous histiocytomas, and 38 to 48% of heterogeneous soft tissue sarcomas. Our study aimed to highlight the use of conventional karyotyping as a genome-wide screening tool, and also as an adjuvant diagnostic tool in the validation of histological diagnosis for soft tissue tumours.

Cytogenetic analysis involves a coordinated effort between surgical pathologists and cytogenetic laboratory technicians.6 In our study, fresh tumour samples were collected in sterile bottles from the surgical theatre and transported immediately to the cytogenetics laboratory. Next, the tumour specimens were washed 3 times with media containing Hank’s balanced salt solution, and 2% penicillin and streptomycin. After washing, the tissue was minced finely with scalpels and digested in collagenase II (GIBCO, Gaithersburg [MD], US) at a concentration of 1400 units/mL for 1 hour. The disaggregated tissue was then transferred into a centrifuge tube and washed twice with 1X Hank’s balanced salt solution and then with Roswell Park Memorial Institute complete medium (culture medium). The cells were centrifuged and transferred to a culture medium containing RPMI 1640, 20% fetal bovine serum, 2% 200 mmol/L L-glutamine, and 2% 5000 U penicillin and 5000 µg streptomycin.

Cells were cultured and harvested according to standard cytogenetic preparations and procedures. The cultures were set up in a 37°C incubator with 5% CO₂. The time of harvesting the cells depended on the degree of cell proliferation in culture. At harvest, 50 µL colcemid (10 µg/mL) was added to the cultures for 3 hours to arrest the cells at metaphase. Cultured cells were detached by treatment with 1X trypsin EDTA and then treated with 0.075 mol/L KCl-0.6% trisodium citrate solution (1:2) for 20 minutes at 37°C. After fixation in two changes of methanol-acetic acid (3:1), chromosome spreads were made by the air-drying method. Chromosomes were stained using the GTG banding method. A total of 20 cells were analysed in each case and karyotype results were designated according to International System of Human Cytogenetic Nomenclature (ISCN 2005, 2009).7 8

Conventional karyotyping of 35 consecutive fresh soft tissue tumour specimens was performed in a cytogenetic laboratory at our institution over a period of 4 years from 2005 to 2009. Medical records and histopathology reports for each patient case were reviewed and diagnoses were formulated based on the World Health Organization classification of soft tissue tumours.1 Recurrent chromosomal abnormalities were identified using the Mitelman Database of Chromosome Aberrations in Cancer,9 and with relevant literature search. Any novel chromosomal aberrations were also noted. This research protocol was approved by the ethics committee of our institution, and informed consent from the patients was obtained by the surgeon.

From January 2005 to March 2009, 35 consecutive fresh tissue specimens were harvested from soft tissue tumour surgical specimens. Histopathology results revealed 20 distinct morphologies. There were 29 malignant tumours, five benign tumours, and one of uncertain malignant potential. In our study, there was an almost equal gender representation with 18 males and 17 females, and age ranging from 15 to 81 years. Table 1 shows an overview of the patient’s age at diagnosis, tumour site, and tissue type for all 35 cases. The majority of our patients (37%) were in the age-group of 41 to 60 years. The most common tumour location was in the extremities (60%), and adipose tissue (34%) was the most common type. As shown in Table 2, conventional cytogenetic analysis revealed an abnormal karyotype detection rate of 63% (22 of 35 cases). Diagnostic abnormal karyotype was seen in nine (26%) cases—Ewing’s sarcomas (n=2), desmoplastic small round cell tumour (DSRCT) [n=1], synovial sarcomas (n=3), myxoid liposarcomas (MLPSs) [n=2], and lipoma (n=1). A normal karyotype (ie 46, XX or 46, XY) was seen in 11 (31%) cases. There were also two (6%) cases of culture failure. Table 3 shows the diagnosis, full karyotype results, and diagnostic utility for all 22 cases with abnormal karyotype.

A wide range of structural and numerical chromosomal abnormalities exists. These aberrations may be characterised by chromosomal gains or losses, balanced or unbalanced translocations, deletions or insertions, ring or marker chromosomes, or multiple complex karyotypes.6 Sarcomas may be categorised into two major cytogenetic groups: (i) sarcomas with tumour-specific chromosomal alterations and simple karyotypes2 10 11 or (ii) sarcomas with non-specific chromosomal alterations and complex unbalanced karyotypes.2 For group (i), karyotypes are considered to be tumour-specific or recurrent if the abnormality is found in two or more cases. For group (ii), a complex karyotype abnormality will not be specific for the diagnosis but is supportive of the diagnosis of malignancy. A ring chromosome may also indicate some form of malignancy. While a marker chromosome is diagnostically non-specific, it is an indicator of clonal progression and further testing by whole chromosome painting (CP) FISH may aid the diagnosis. Chromosome painting refers to the hybridisation of fluorescently labelled chromosome-specific probe pools for the detection of chromosomal aberrations.12 The simultaneous hybridisation of multiple CP probes, each tagged with a specific fluorochrome, enables the coloured display of all 24 human chromosomes also known as multicolour FISH.12 The advantages of CP include its ability to detect subtle telomeric translocations and small chromosomal markers, barely the size of a chromosomal band.12 Despite showing some utility as a genetic screening tool, CP is more straightforward only when used in conjunction with conventional cytogenetics which provide information on the specific chromosomes involved. This is because CP alone requires the iterative hybridisation of multiple CP probes, which is not always practical due to time constraints and limited specimens.12

Of the 22 cases with abnormal karyotype results, nine (41%) cases showed tumour-specific chromosome abnormalities. These nine cases had abnormal karyotypes which were consistent with the Mitelman Database of Chromosome Aberrations in Cancer9 and previously published literature.2 11 A normal karyotype was seen in 11 (31%) cases where three tumour tissues were of fibrous origin. One study in our literature search demonstrated a normal karyotype in 42% of cases; the majority of these were soft tissue tumours with a fibrous component or grossly dense matrix.6 The study rationalised that tumour cells embedded in a dense matrix were more difficult to culture.6 The two culture failure cases could have been due to specimen contamination or insufficient sample. A study conducted in a single institution in the United States (n=48) had documented an abnormal karyotype detection rate of 48% and a 10% culture failure rate in patients with soft tissue tumours.6 In contrast, our Asian cohort study had a higher detection rate of 63% (n=35) and a lower culture failure rate of 6%. The small sample size of this study was limited by the disease prevalence (rarity of sarcomas) as well as the logistics of obtaining fresh specimens from the surgical operating room. We intend to conduct future studies with a bigger sample size and explore other cytogenetic aberrations in soft tissue sarcomas using FISH in conjunction with conventional karyotyping.

Of the two cases of Ewing’s sarcoma in this study, one was a 42-year-old female (case 3; Table 3) and one a 26-year-old male (case 4; Table 3). This is an unusual clinical age-group for this sarcoma and the histological diagnosis was confirmed by karyotyping. In the male patient, the variant t(2;11;22)(q35;q24;q12) was demonstrated. For case 5 (Table 3), trisomy 12, a non-random secondary aberration, was demonstrated. One study found that the majority of chromosomal aberrations in Ewing’s sarcoma appear to be trisomy 8 and trisomy 12, occurring in 44% and 16% of Ewing’s sarcoma cases, respectively.13 14 15

Our study showed two diagnostic cases of synovial sarcoma (cases 19 and 20) with the hallmark translocation t(X;18) seen16 together with complex cytogenetic aberrations (Table 3). Another two cases of synovial sarcoma (cases 18 and 21) showed structure rearrangement on 2p/18p and translocation t(3;7)(q21;p13), respectively. These abnormalities were uncharacteristic. Histological biopsy of the left distal tibia showed a soft tissue tumour measuring 4 x 3 x 1 cm, composed of large sheets of malignant cells displaying high nuclear cytoplasmic ratio, round or irregular nuclei, nucleoli, scanty cytoplasm, and frequent mitoses. Tumour cells were positive for CD99 (MIC2 gene product), cytokeratin AE1+3 (especially epithelial-like areas), and vimentin. Further immunohistochemical staining with epithelial membrane antigen showed focal positivity. The soft tissue tumour had also invaded the distal tibia on the anteromedial and posteromedial aspects of the left leg with metastasis to the left groin lymph node. Case 21 was reviewed by various histopathologists and the general consensus was that of a high-grade undifferentiated synovial sarcoma. The representative karyogram for case 21 is shown in Figure 1.

In the study by Saboorian et al,17 there was one case of ambiguous histological results; the stained tissue smears showed densely cellular and tightly cohesive malignant spindle cells without discernible epithelial differentiation. A few differential diagnoses were formulated which included synovial sarcoma and karyotyping confirmed the diagnosis of synovial sarcoma by revealing the presence of t(X;18)(p11.2;q11.2).17 Another study by Akerman et al,18 which involved the cytogenetic evaluation of 15 surgical specimens, confirmed the (X;18) translocation as both a specific and sensitive marker for synovial sarcoma. Our study and the above studies serve to highlight the essential supportive role of conventional karyotyping in the confirmation of the diagnosis of synovial sarcoma.

Liposarcoma is the most common soft tissue sarcoma, accounting for 20% of mesenchymal neoplasms.19 It can be categorised into three subtypes: myxoid and round cell, well-differentiated, and pleomorphic.19 All three subtypes that were included in our study are discussed below.

Myxoid liposarcoma is the second most common liposarcoma subtype in which two thirds of the cases arise from the thigh musculature.19 The characteristic translocation t(12;16)(q13;p11) has been well documented in more than 90% of MLPS cases.19 20 21 22 This translocation leads to formation of a TLS-CHOP fusion gene (located at 12q13 and 16p11 respectively) which is highly sensitive and specific for MLPS.19 A possible trisomy 8 as an additional secondary change has also been reported.22 Our study demonstrated two cases of MLPS showing the t(12;16)(q13;p11) translocation. As shown in Table 3, this translocation was diagnostic of MLPS in cases 1 and 2. A study by the CHAMP group in which cytogenetic analysis was carried out in 28 MLPS specimens reported the t(12;16)(q13;p11) translocation in 26 cases; this further confirmed its consistency as a genetic marker for MLPS.20 Conventional karyotyping for t(12;16)(q13;p11) in MLPS was also shown to be useful as an adjunct diagnostic tool in poorly differentiated myxoid neoplasms in another study.21

Pleomorphic liposarcoma (PLPS) is the rarest (5% of liposarcoma) and most aggressive (highly metastatic) form of liposarcoma.19 It commonly affects the extremities in the elderly (>50 years) with an equal distribution in both genders.19 The complex structural abnormalities (unidentified marker chromosomes) and high chromosome counts (polyploidy) make it difficult to detect PLPS-specific aberrations.19 Our study demonstrated the case of a 77-year-old female (case 8; Table 3) with PLPS which showed complex, structural aberrations on karyotyping which, though not diagnostic, was indicative of a malignant clonal process.

Lipomas are the most common soft tissue tumours and are benign.23 One study by Sandberg and Bridge24 had documented rearrangements affecting the 12q13~q15 region as the most common aberration (65% of 188 lipomas). Clonal chromosomal aberrations were also reported in 60% of lipomas, and of these, 70% had normal cytogenetic cells.24 The most frequent t(3;12)(q27~q28;q13~q15) translocation was seen in 25% of lipoma cases.24

Case 10 (Table 3) belonging to a 53-year-old male demonstrated the balanced t(2;12) (q23;q15) translocation which was also novel in that the breakpoint 2q23 has not been previously reported. In this patient, histology showed a large lipoma measuring 14 x 9 x 8 cm. In addition, magnetic resonance imaging suggested a malignant liposarcoma. Szymanska et al25 found that the overrepresentation of 1q and 12q sequences was a recurrent finding in lipoma-like liposarcomas but not in lipomas. This is consistent with the chromosomal aberration involving 12q15 breakpoint in case 10. The representative karyogram for case 10 is shown in Figure 2.

Atypical lipomatous tumour (ALT) is synonymous with well-differentiated liposarcoma (WDLPS) as both exhibit similar cytogenetic findings regardless of location and pathology.19 Being the most common of all liposarcomas (40%-45%), ALT/WDLPS is an intermediate (locally aggressive) soft tissue sarcoma with mature adipocyte differentiation.1 19 26 Most ALTs are characterised cytogenetically by the presence of supernumerary ring chromosomes or long marker chromosomes involving chromosome region 12q13-15.26 27

Our study demonstrated abnormal karyotypes in two cases of ALT and WDLPS each. Of the two ALTs, case 6 (Table 3) had a ring chromosome as a sole abnormality and case 7 had supernumerary ring chromosomes present in addition to the multiple complex numerical structural aberrations. Case 11 (WDLPS) showed both complex numerical and structural chromosomal rearrangements in which two dicentric chromosomes were present together with ring chromosomes and giant marker chromosomes. Case 12 (WDLPS) belonged to a 65-year-old male; a normal karyotype was seen in 19 cells, one nonclonal abnormal cell was hypodiploid which showed trisomy 12, deletion on 12p, structural rearrangement on 20q as well as a ring chromosome. It is uncertain if this nonclonal abnormal cell is of any clinical significance. Histology had showed a WDLPS measuring 19 x 12 x 4 cm infiltrating the skeletal muscle of the left thigh.

It was reported that virtually all ALT/WDLPS had abnormal cytogenetic results.26 The CHAMP group conducted a study of 59 ALT/WDLPS and evaluated their relationship and differential diagnoses with other adipose tissue tumours.28 Clonal chromosomal abnormalities were found in 55 (93%) cases and supernumerary ring or giant marker chromosomes (RGCs) were seen in 37 (63%) cases28; RGCs were also shown to have tumour progression potential. Statistical analysis demonstrated a highly significant correlation between ALTs and RGCs (P<0.0001).28 The study reaffirmed the essential role of karyotype analysis in differentiating ALTs from benign lipomas, spindle/PLPS, hibernomas, and MLPS.

Desmoplastic small round cell tumour is a rare and aggressive neoplasm that commonly affects adolescents and young adults.29 30 Our study demonstrated the case of a 27-year-old male with DSRCT showing the classic t(11;22)(p13;q12) translocation (case 17; Table 3). In this case, histopathology reports showed no evidence of malignant infiltrates in the tumour specimen but conventional karyotyping confirmed the diagnosis to be DSRCT.

Karyotype analysis detected a majority (63%) of cases with abnormal chromosomes in our Asian cohort study with nine (41%) cases showing 22 abnormal karyotypes. Our study, hence, demonstrated that conventional karyotyping played an essential supportive role in validating histological diagnosis, especially in cases with borderline or complex morphology. Newer molecular techniques such as FISH and RT-PCR techniques may be sensitive but require prior knowledge of the expected genetic change. In view of this, conventional karyotyping is useful as a genome-wide screening tool in detecting single or multiple chromosomal aberrations in each patient. The use of conventional karyotyping is highly encouraged in the pursuit of discovering more novel recurring chromosomal aberrations.

This study was internally supported by the grant from the Department of Clinical Research and Cytogenetics Laboratory, Department of Pathology, Singapore General Hospital. The authors would like to thank Dr Alvin Lim for his support and Mr Lim Ping for his technical assistance in ensuring the success of this research project.