Article Text


Utility of sarcoma-specific fusion gene analysis in paraffin-embedded material for routine diagnosis at a specialist centre
  1. Khin Thway1,
  2. Sasha Rockcliffe1,
  3. David Gonzalez2,
  4. John Swansbury2,
  5. Toon Min2,
  6. Lisa Thompson2,
  7. Cyril Fisher1
  1. 1Department of Histopathology, Royal Marsden Hospital, London, UK
  2. 2Department of Haemato-Oncology, Royal Marsden Hospital, Sutton, Surrey, UK
  1. Correspondence to Dr Khin Thway, Department of Histopathology, Royal Marsden NHS Foundation Trust, 203 Fulham Road, London SW3 6JJ, UK; khin.thway{at}


Aims Diagnosis of soft tissue sarcomas can be difficult. It can be aided by detection of specific genetic aberrations in many cases. This study assessed the utility of a molecular genetics/cytogenetics service as part of the routine diagnostic service at the Royal Marsden Hospital.

Methods A retrospective audit was performed over a 15-month period to evaluate the diagnostic usefulness for soft tissue sarcomas with translocations of fluorescence in situ hybridisation (FISH) and reverse-transcriptase PCR (RT-PCR) in paraffin-embedded (PE) material. Results were compared with histology, and evaluated.

Results Molecular investigations were performed on PE material in 158 samples (total 194 RT-PCR and 174 FISH tests), of which 85 were referral cases. Synovial sarcoma, Ewing sarcoma and low-grade fibromyxoid sarcoma were the most commonly tested tumours. Myxoid liposarcoma showed the best histological and molecular concordance, and alveolar rhabdomyosarcoma showed the best agreement between methods. FISH had a higher sensitivity for detecting tumours (73%, compared with 59% for RT-PCR) with a better success rate than RT-PCR, although the latter was specific in identifying the partner gene for each fusion. In particular, referral blocks in which methods of tissue fixation and processing were not certain resulted in higher RT-PCR failure rates.

Conclusions FISH and RT-PCR on PE tissue are practical and effective ancillary tools in the diagnosis of soft tissue sarcomas. They are useful in confirming doubtful histological diagnoses and excluding malignant diagnoses. PCR is less sensitive than FISH, and the use of both techniques is optimal for maximising the detection rate of translocation-positive sarcomas.

  • Audit
  • FISH
  • molecular genetics
  • sarcoma
  • soft tissue tumours

Statistics from


Sarcomas are a heterogeneous group of mesenchymal neoplasms for which histological assessment is challenging, as the morphology and immunoprofile of different tumour types frequently overlap, and can mimic other tumour types. Many sarcomas have characteristic chromosomal translocations, resulting in fusion genes that translate into chimeric proteins.1–4 As most sarcomas are tumour specific, their detection by cytogenetic or molecular genetic techniques is diagnostically useful. Translocations can be detected by fluorescence in situ hybridisation (FISH), and the novel fusion transcripts by reverse-transcriptase PCR (RT-PCR).

The histopathology department of the Royal Marsden Hospital (RMH) has approximately 2000 soft tissue tumour accessions per year, of which about half are referral or second opinion cases, and half are from surgical procedures performed on site. All cases are reported or reviewed by a specialised soft tissue pathologist. Although immunohistochemistry (IHC) is the most frequently used ancillary diagnostic tool, molecular genetic and/or cytogenetic analysis is performed on a significant portion of cases, for diagnostic confirmation or exclusion. While the use of fresh/frozen tissue remains optimal, this is frequently unavailable, as cases are often referred from other institutions or the need for banking fresh tissue may not be suspected.

The RMH molecular diagnostic service for soft tissue sarcomas comprises two separate diagnostic laboratories offering FISH and RT-PCR. DNA sequence analysis is performed after a proportion of RT-PCR tests (those for rare or uncommon fusion genes) to ensure test specificity (figure 1B). We assessed the utility of FISH and RT-PCR on paraffin-embedded (PE) (largely formalin-fixed) material as part of the routine diagnostic investigation for patients with translocation soft tissue sarcomas in a 15-month period.

Figure 1

(A) Clear cell sarcoma. This tumour occurs in young adults and has a nested architecture and similar immunoprofile to melanoma, which may cause diagnostic confusion. (H&E ×400). (B) Example of DNA sequencing analysis in a case of clear cell sarcoma showing an unusual variant of the EWSR1–ATF1 fusion between exon 10 of EWSR1 and exon 5 of ATF1. (C) Ewing sarcoma. This tumour comprises sheets of monomorphic small round cells. A proportion of tumours lack CD99 expression and, as this is itself not a wholly specific immunohistochemical marker for Ewing sarcoma, the differential diagnosis includes a variety of small round cell tumours. In this example, the tumour has a nested distribution with central discohesion, and may be mistaken morphologically for alveolar rhabdomyosarcoma (H&E ×400). (D) Fluorescence in situ hybridisation showing Ewing sarcoma. Dual-colour break-apart probes are used for the EWSR1 gene on chromosome 22. There are split red and green signals, indicating abnormal clones with translocations involving the EWSR1 gene.

Materials and methods

A retrospective audit was performed for all soft tissue tumours examined at RMH that had RT-PCR and/or FISH performed on PE material, over a 15-month period (1 April 2007 to 30 June 2008). The material assessed was that of (a) tumours biopsied or excised at the RMH Sarcoma Unit and processed locally, and (b) tumours in paraffin blocks from other institutions referred with the patient or sent for diagnostic second opinion. Molecular tests performed on fresh material were excluded. Tests were requested to confirm or exclude a histological diagnosis. Tumours with an obvious histological diagnosis (eg, biphasic synovial sarcoma) were not analysed by molecular means, and most cases tested were those with a degree of diagnostic uncertainty. All small round cell sarcomas were tested. Sarcomas (excluding gastrointestinal stromal tumours (GISTs)) for which routine molecular diagnostic tests are available in the unit are listed in table 1. For FISH, PE sections (2 μm thick) were dewaxed overnight at 60 C, treated with hot buffer wash (2–3 h), and then with proteolytic enzyme treatment at 37 C, and washed in distilled water, and then an alcohol series, before addition of DNA probes (Abbott Laboratories, Maidenhead, UK). Hybridisation was performed overnight according to manufacturer's protocols. Standard methods were used for RT-PCR.5–9 For each patient, results of RT-PCR and FISH were compared, and these findings were compared with the histology.

Table 1

Molecular diagnostic tests available for soft tissue tumours at the Royal Marsden Hospital


Molecular tests were requested on PE material from 158 soft tissue lesions (141 adults, 17 patients <18 years): 85 were from blocks prepared elsewhere (49 for second opinion; 36 for diagnostic review), and 73 were from those processed locally. In total, 194 RT-PCR and 174 FISH tests were performed. In 165 instances, both methods were used to assess for a specific tumour (including those having multiple tests for differential diagnoses), while in the remaining cases only one method was used. RT-PCR was positive (ie, identified an appropriate fusion transcript) in 45/194 cases (23.2%), negative in 100 cases (51.5%), and unsuccessful in 49 cases (25.2%). FISH was positive (ie, identified an appropriate chromosomal breakpoint) in 52 of 174 cases (29.9%), negative in 112 cases (64.4%) and unsuccessful in 10 cases (5.7%).

The most frequent requests were for synovial sarcoma (SS) (for 46 patients), Ewing sarcoma (ES) (37 patients) (figure 1C), low-grade fibromyxoid sarcoma (35 patients) and myxoid/round cell liposarcoma (MLPS) (25 patients), followed by extraskeletal myxoid chondrosarcoma (EMC) (19 patients) and alveolar rhabdomyosarcoma (ARMS) (17 patients). Less frequent requests were for desmoplastic small round cell tumour (seven patients), alveolar soft part sarcoma (four patients), clear cell sarcoma (three patients) (figure 1A) and endometrial stromal sarcoma (two patients). No tests were requested for angiomatoid fibrous histiocytoma or inflammatory myofibroblastic tumour. The tumour most frequently tested to confirm a histological diagnosis was SS, followed by ES and MLPS. The tumour most frequently tested for exclusion was low-grade fibromyxoid sarcoma, followed by SS and ES. The tumour for which molecular results showed the highest match to the histological diagnosis was MLPS (84.6% of cases tested for confirmation), followed closely by SS (84.2%). FISH had a total sensitivity of 72.9% (43/59 assessable cases), compared with 58.7% (37/63 assessable cases) for RT-PCR.

FISH and RT-PCR were equally sensitive (ie, positive FISH or RT-PCR obtained in equal numbers, although not necessarily synchronously) in detecting SS, EMC and ARMS. FISH and RT-PCR were most often both positive in ARMS. FISH was more sensitive than RT-PCR in detecting ES and MLPS (nine FISH positive, six RT-PCR positive, for each of 14 and 13 cases, respectively), and clear cell sarcoma (one FISH positive, one unsuccessful RT-PCR, of one case). The two methods were equally sensitive in detecting desmoplastic small round cell tumour. Of the unsuccessful tests, all 10 FISH cases were referrals (100%), whereas for RT-PCR, 26 were referrals (56.5%). Summaries of the findings for each tumour tested are shown in tables 2 and 3.

Table 2

Tumours tested for molecular diagnostic confirmation, with totals for positive findings by RT-PCR and FISH

Table 3

Tumours tested for diagnostic exclusion


The detection of tumour-specific genetic aberrations is becoming an integral part of the diagnostic investigation of soft tissue tumours. Many sarcomas have characteristic genetic aberrations including (a) chromosomal translocations resulting in chimeric fusion genes, or (b) oncogenic mutations (eg, KIT/PDGFRA mutations in GISTs). Translocations can be detected by FISH (figure 1D), and the novel fusion transcripts can be detected by RT-PCR, while oncogenic mutations in GISTs can be detected by DNA sequencing analysis. RT-PCR and FISH are performed in conjunction at RMH, although occasionally only one method is requested, usually FISH when material is limited (requiring 1–2 μm sections, compared with 2×20 μm for RT-PCR).

Using the two methods ensures a higher success rate, as both have limitations. RT-PCR assays are designed to detect specific fusion transcripts, but this also means rare fusion transcripts or unusual translocation breakpoints cannot be detected. Conversely, the FISH assays utilise ‘break-apart’ probes that only identify a breakpoint in one of the common genes (eg, EWSR1), without providing information on the translocation partner. Therefore, using both methodologies provides the highest sensitivity and specificity for the detection of fusion genes in samples.

Although snap-frozen/fresh material remains optimal, the application of molecular techniques to PE tissue allows investigation of a much greater range of patient material, especially referral or archival cases. In a tertiary or specialist centre, the proportion of referral cases is large, with most molecular tests performed on PE tissue. In the study period, not all ‘translocation’ sarcomas diagnosed histologically had molecular confirmation. Cases in which diagnosis was clear cut by morphology±IHC were not analysed by molecular means, and those tested fell into two categories, those for: (a) diagnostic confirmation (diagnosis suspected but possible morphological/immunohistochemical discrepancies), (b) diagnostic exclusion (eg, where a case was thought probably not to represent a particular tumour, but excluding the diagnosis was warranted (such as ES with cytokeratin expression, tested to exclude poorly differentiated SS)).

The success of a test is not reflected only by the percentage of positive results; many tests were performed to exclude specific diagnoses so that negative results were equally relevant. Positive/negative rates were also influenced by case selection, which was strongly biased towards diagnostically difficult cases (31% of all cases tested were second opinions), unusual tumour variants, or those with limited lesional tissue. The positive result rate of both methods would undoubtedly have been higher if all translocation sarcomas reported histologically had been analysed. Reasons for negative RT-PCR results, other than absence of fusion genes, include the presence of variant/uncharacterised rare fusion transcripts for which primers are not available. Genes can have multiple fusion partners and breakpoints, such as EWSR1 in ES, which can fuse with several members of the ETS family of transcription factors.10–15 Negative findings with FISH include variant translocations with submicroscopic insertions or exchanges of material. Also, only a portion of each tumour nucleus may be represented on the slide in the very thin sections taken.

Only 10/176 FISH tests failed, all on referred material, while 49/194 RT-PCR tests failed (26 referrals). Unsuccessful tests were greater in referred material for FISH and PCR (100% and 56.5% of unsuccessful tests, respectively). Genetic material is more frequently degraded in blocks in which (a) the fixative is unknown, (b) fixation may be suboptimal and (c) the blocks may be significantly older. Successful RNA extraction is dependent on timing of fixation and fixative used,16 and acid fixatives such as Bouin's solutions are a major cause of DNA/RNA damage.17 Lack of sufficient RNA can be due to low sample cellularity. Although fixation and processing protocols are standardised for in-house cases, internal cases that failed may have been large specimens resected at the weekend, then left unsliced until the next working day, and subject to delayed and then overfixation. Other reasons for failure include small core biopsies with inadequate numbers of tumour nuclei, and necrotic tumour material.

‘False’ positive findings were rare, with only two examples detected (both by FISH; table 3). The genetic breakpoints identified in these neoplasms were presumably secondary to their intrinsic genetic instability, rather than representing novel/uncharacterised gene fusions central to tumorigenesis. Pertinently, further validation (by repeat FISH, by RT-PCR, or by testing in further material) was performed for each case with an unexpected positive result. Not least, the histological diagnosis was always checked and re-evaluated. Stringent precautions were also taken in all laboratories to ensure that tissue contamination did not occur.

Tumours overlapping morphologically and sharing translocation partners can pose diagnostic problems, for example pure round cell liposarcoma with a rarer EWSR1 translocation, and ES, although distinctions are crucial particularly if they impact on clinical management. Also, although most translocations are tumour specific, there is increasing evidence that morphologically and immunohistochemically different sarcomas can share the same translocation.18 Therefore, results of molecular investigations must be interpreted in conjunction with histology, other ancillary investigations and the complete clinical picture. One point of note was the necessary assumption in this study, when assessing for rate of molecular genetic confirmation, that histology was the diagnostic ‘gold standard.’ As tumours could only be triaged as requiring molecular analysis after they were assessed microscopically, the histological interpretation was used here as a reference point, for comparison with molecular findings.

In our study, FISH had an overall sensitivity of 72.9%, compared with 58.7% for RT-PCR. Previous comparative studies of FISH and RT-PCR on formalin-fixed PE tissue have found varying results. In ES, FISH sensitivity has been found to be higher (up to 100%) than RT-PCR (54%),19 whereas for SS, studies have shown almost equal detection rates of (FISH 96.7%, RT-PCR 94.7%),20 or a higher rate with RT-PCR (94% compared with 84%).21 In separate studies, FOXO1 gene rearrangements by FISH have been described in 88% of ARMS,22 but fusion transcripts detected by RT-PCR in only 55%.23

The range of tests offered in diagnostic laboratories will increase as more commercial probes become available and better validated. The fusion status of a tumour can provide useful prognostic information, for example the PAX7–FOXO1 fusion in ARMS is associated with significantly longer event-free survival than PAX3–FOXO1.24 Differences in fusion status do not currently impact on management, but microarray-based gene expression profiling has shown distinct expression patterns in many sarcoma groups,25 26 including RMS,27 28 ES29 and LPS,30 31 and as treatment becomes increasingly tailored to the individual patient and guided by specific fusion types or molecular signatures, molecular genetic analysis is likely to become more applicable, similar to the way that many haematolymphoid neoplasms are now defined.32

Take-home messages

  • Fluorescence in situ hybridisation and reverse-transcriptase PCR on paraffin-embedded tissue are practical and effective ancillary tools in the diagnosis of soft tissue sarcomas.

  • The use of both techniques is optimal for maximising the detection rate of translocation-positive sarcomas.

  • Results of molecular investigations must be interpreted in conjunction with histology, other ancillary investigations and the complete clinical picture.

Interactive multiple choice questions

This JCP article has an accompanying set of multiple choice questions (MCQs). To access the questions, click on BMJ Learning: Take this module on BMJ Learning from the content box at the top right and bottom left of the online article. For more information please go to: Please note: the MCQs are hosted on BMJ Learning—the best available learning website for medical professionals from the BMJ Group. If prompted, subscribers must sign into JCP with their journal's username and password. All users must also complete a one-time registration on BMJ Learning and subsequently log in (with a BMJ Learning username and password) on every visit.


View Abstract


  • Funding We acknowledge NHS funding to the NIHR Biomedical Research Centre.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.