Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors

Journal of Orthopaedic Science xxx (2017) 1e12

Contents lists available at ScienceDirect

Journal of Orthopaedic Science journal homepage: http://www.elsevier.com/locate/jos

Instructional Lecture

Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors* Shintaro Sugita 1, Tadashi Hasegawa* Department of Surgical Pathology, Sapporo Medical University School of Medicine, South 1 West 16, Chuo-ku, Sapporo, Hokkaido 060-8543, Japan

a r t i c l e i n f o

a b s t r a c t

Article history: Received 30 November 2016 Received in revised form 6 February 2017 Accepted 11 February 2017 Available online xxx

During routine pathological examination, fluorescence in situ hybridization (FISH) plays a significant role in the genetic analysis of samples. FISH can detect genetic abnormalities such as chromosomal translocations, gene amplifications, and deletions in formalin-fixed, paraffin-embedded (FFPE) specimens. Due to its practical advantages, FISH is already used in many pathology laboratories. It is especially useful for the diagnosis of translocation-related sarcomas (TRSs), which comprise about 25% of soft tissue sarcomas. Because TRSs have specific chimeric genes derived from characteristic chromosomal translocations, their diagnosis would not be possible without FISH. FISH significantly contributes to the genetic confirmation of TRS. Analysis using next-generation sequencing (NGS), the latest powerful method for comprehensive genomic analysis, has recently revealed many kinds of chromosomal translocations in various TRSs. We often use experimental results to create custom probes for FISH and have applied NOCA2 split probes and CIC split, CIC-FOXO4 fusion probes to the pathological diagnosis of soft tissue angiofibroma and CIC-rearranged sarcoma, respectively. Some chimeric fusions detected by NGS induce the expression of related proteins and their detection using immunohistochemistry is beneficial for pathological diagnosis. We previously identified characteristic FOSB expression in pseudomyogenic hemangioendothelioma (PHE) with a specific SERPINE1-FOSB fusion, revealing the usefulness of FOSB immunohistochemistry in the differential diagnosis of PHE and its mimics. Finally, we participated in a central review of a clinical trial of trabectedin monotherapy. We guaranteed an accurate diagnosis by using FISH and genetic confirmation to select appropriate TRS patients and thereby confirm the accuracy of the patient enrollment of the clinical trial. FISH is an essential tool for the pathological diagnosis of soft tissue and bone tumors. It can detect various genetic abnormalities in an “in situ” fashion using FFPE specimens on glass slides during routine examination. It is also an excellent tool for translating the latest experimental findings to practical use in routine pathological diagnosis. Further instrumental improvements in FISH will help it to become the universal method for the genetic analysis of pathological diagnoses. © 2017 The Japanese Orthopaedic Association. Published by Elsevier B.V. All rights reserved.

1. Introduction Various genetic abnormalities in soft tissue and bone tumors have recently been identified by next-generation sequencing (NGS), the latest fast and powerful method for comprehensive genomic analysis. These findings can characterize the inherent

* This Instructional lecture was presented at the 49th Annual Musculoskeletal Tumor Meeting of the Japanese Orthopaedic Association, Tokyo, July 14, 2016. * Corresponding author. Fax: þ81 11 615 1418. E-mail addresses: [email protected] (S. Sugita), [email protected] (T. Hasegawa). 1 Fax: þ81 11 615 1418.

nature of specific tumor types and contribute to the genetic confirmation of pathological diagnoses [1]. As pathologists, we often try to detect genetic abnormalities to diagnose soft tissue and bone tumors. Such tumors can be genetically diagnosed using fluorescence in situ hybridization (FISH), a convenient and efficient tool. FISH can detect chromosomal translocation, gene amplification, gene deletion, and abnormal polyploidy in formalin-fixed, paraffin-embedded (FFPE) specimens, the type of samples we universally use in routine pathological diagnostic work. Of the many possible genetic abnormalities, specific chimeric genes derived from chromosomal translocations are particularly valuable for the confirmative diagnosis of translocation-related sarcoma (TRS), which comprises about 25% of soft tissue sarcomas (STSs) [2]

http://dx.doi.org/10.1016/j.jos.2017.02.004 0949-2658/© 2017 The Japanese Orthopaedic Association. Published by Elsevier B.V. All rights reserved.

Please cite this article in press as: Sugita S, Hasegawa T, Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors, Journal of Orthopaedic Science (2017), http://dx.doi.org/10.1016/j.jos.2017.02.004

2

S. Sugita, T. Hasegawa / Journal of Orthopaedic Science xxx (2017) 1e12

(Table 1). Now that many commercial FISH probes are available, FISH can be performed in many pathology laboratories and will play increasingly significant roles in the genetic analysis of pathological diagnoses in the future. In this article, we first explain the principle behind FISH and how to interpret various signal patterns. We then describe the practical use and utility of FISH, introduce our recent studies and review the latest literature, and finally discuss the potential of FISH to play an even greater practical role in routine examinations. 2. The principle behind FISH and the interpretation of its signal patterns

specimens used for FISH with the morphology visible in hematoxylin and eosin-stained sections, we can investigate the fluorescent signals “in situ” in the specimens being examined. To interpret the signal pattern, it is imperative that we precisely understand the designs of the probes and what kinds of abnormalities can be detected. In the practical situation, we investigate fluorescent signals as red and green “spots” under the fluorescent microscope and interpret the signal patterns to assess the genetic abnormalities. Fig. 1 shows some signal patterns and the interpretations of representative abnormalities, including translocation, amplification, deletion, and polyploidy. 2.1. Chromosomal translocation

Regarding the standard practice in pathological diagnostic work, we usually perform DNA FISH using DNA probes that detect certain abnormalities in genomic DNA in chromosomes [3]. FISH probes are designed to include certain parts of a specific base sequence that are complementary to the target chromosomal DNA and such base sequences are then usually labeled with red and green fluorescent dyes. Thus, the presence of certain genetic abnormalities in tumors can be confirmed by the detection of red and green fluorescent signals. Therefore, we should recognize the following essential points when making a custom probe. First, we have to clarify what kinds of genetic abnormalities we will detect. This enables us to select the correct type of probes for the research objective(s). Next, we decide which complementary base sequences would be most effective for detecting target abnormalities and then choose bacterial artificial chromosome (BAC) clones including the base sequences to generate specific probes. For example, when we design split-signal probes, we have to select BAC clones that cover objective gene loci and should label them with red and green dyes that do not overlap a break point (Fig. S1). Because, when using alternate serial sections on glass slides, we can compare the FFPE

Table 1 Representative chromosomal and genetic abnormalities of soft tissue and bone tumors (mainly those including translocation-related sarcoma). Histologic type

Translocation

Chimeric gene

Ewing sarcoma and Ewing-like sarcoma

t(11;22)(q24;q12) t(21;22)(q22;q12) t(7;22)(q22;q12) t(17;22)(q21;q12) t(16;21)(p11;q22) t(6;22)(p21;q12) t(4;19)(q35;q13) t(X;19)(q13;q13.3) t(12;16)(q13;p11) t(12;22)(q13;q12) t(X;18)(p11;q11) t(2;13)(q35;q14) t(1;13)(q36;q14) t(9;22)(q22;q12)

EWSR1-FLI1 EWSR1-ERG EWSR1-ETV1 EWSR1-ETV4 FUS-ERG EWSR1-POU5F1 CIC-DUX4 CIC-FOXO4 FUS-DDIT3 EWSR1-DDIT3 SS18-SSX PAX3-FOXO1A PAX5-FOXO1A EWSR1-NR4A3

t(11;22)(p13;q12) t(12;22)(q13;q12) t(2;22)(q33;q12) t(17;22)(q22;q13) inv(12)(q13q13) t(12;15)(p13;q25) t(X;17)(p11;q25) t(2;19)(p23;q13.1) t(1;2)(q22-23;p23) t(7;16)(q34;p11) t(11;16)(p11;p11) t(1;3)(p36.3;q25) t(7;19)(q22;q13)

EWSR1-WT1 EWSR1-ATF1 EWSR1-CREB1 COL1A1-PDGFB NAB2-STAT6 ETV6-NTRK3 ASPL-TFE3 TPM4-ALK TPM3-ALK FUS-CREB3L2 FUS-CREB3L1 WWTR1-CAMTA1 SERPINE1-FOSB

t(5;8)(p15;q13) del(8)(q13q21)

AHRR-NCOA2 HEY1-NCOA2

CIC-rearranged sarcoma Myxoid liposarcoma Synovial sarcoma Alveolar rhabdomyosarcoma Extraskeletal myxoid chondrosarcoma Desmoplastic small round cell tumor Clear cell sarcoma Angiomatoid fibrous histiocytoma Dermatofibrosarcoma protuberans Solitary fibrous tumor Infantile fibrosarcoma Alveolar soft part sarcoma Inflammatory myofibroblastic tumor Low-grade fibromyxoid sarcoma Epithelioid hemangioendothelioma Pseudomyogenic hemangioendothelioma Soft tissue angiofibroma Mesenchymal chondrosarcoma

Chromosomal translocation results in chimeric genes. Their detection often helps us to confirm the pathological diagnosis of TRS. Translocations are detected using dual-color, split-signal or dual-color, fused-signal probes [3,4]. Because split-signal probes are designed to encompass a break point in the chromosome between the red and green dyes, we see two pairs of fused signals in normal diploid cells (Fig. 1a). When a chromosomal translocation has occurred, we observe a pair of fused and split signals because the chromosome breakage creates a distance between the red and green spots (Fig. 1b). On the other hand, fused-signal probes are designed to encompass a prospective fusion point on the derived chromosome between the red and green dyes [3,5]. Therefore, the interpretations of positive translocation signals are completely different between split- and fused-signal probes unless both probes show the same signal pattern (Fig. 1c). In our lab, we count 50 tumor cell nuclei and consider the sample to be translocation positive if more than 10% of tumor cells show a positive signal pattern [6]. 2.2. Gene amplification Amplification of specific genes occurs in various tumors and its detection has diagnostic value for some STSs. For example, we often investigate amplification of MDM2 and CDK4 genes for the confirmative diagnosis of well-differentiated and dedifferentiated liposarcoma [3]. To estimate amplification, we use probes that include a control locus without gene amplification and a target locus with gene amplification on the same chromosome, labeled by green and red dyes, respectively. We examine the ratio of green (control) to red (amplification) signals and consider the specimen to be “amplification positive” if the ratio is 2.0 or more (this is the amplification signal cutoff value). Schematically, nuclei with gene amplification exhibit few green signals and numerous red signals (Fig. 1d). 2.3. Gene deletion Deletion of a specific locus on a chromosome sometimes has significant effects on pathological diagnosis and disease treatment. For example, we detected an NF1 deletion in malignant peripheral nerve sheath tumor [7] and an INI1 deletion in epithelioid sarcoma [3]. Moreover, homozygous p16 deletion may help us to diagnose some difficult cases of malignant mesothelioma [8]. To estimate deletions, we use probes that include a control locus without gene deletion and a target locus with a gene deletion, labeled by green and red dyes, respectively. A deletion results in a decreased number of red signals compared with green signals. Schematically, nuclei with a deletion exhibit two green signals and one red signal (Fig. 1e). We calculate the percentage of cells with a deletion signal because nuclear truncation often influences the detection of deletion signals. Therefore, the definitive cutoff value for deletions should be assessed for each histological type.

Please cite this article in press as: Sugita S, Hasegawa T, Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors, Journal of Orthopaedic Science (2017), http://dx.doi.org/10.1016/j.jos.2017.02.004

S. Sugita, T. Hasegawa / Journal of Orthopaedic Science xxx (2017) 1e12

3

Fig. 1. Schematic representations and cytological images of various signal patterns detected by FISH. a. Normal pattern in diploid cells showing two pairs of fused signals. b. Translocation pattern obtained by a split-signal probe showing a pair of fused (yellow) and split (separated red and green) signals. In the case of the latter, chromosome breakage increased the distance between the red and green signals. c. Translocation pattern obtained by a fused-signal probe also showing a pair of fused (yellow) and split (separated red and green) signals. d. Amplification pattern consisting of numerous red (target locus) and several green (control locus) signals. This pattern suggests gene amplification at the target locus. e. Deletion pattern showing one red (target locus) and two green (control locus) signals. One red signal disappeared due to gene deletion at the target locus. f. Polyploidy pattern showing three pairs of fused signals. This pattern suggests an abnormal triploid karyotype.

2.4. Polyploidy and aneuploidy Signs of polyploidy and aneuploidy supplement the information provided by analyses of the above signals. In general, STSs do not tend to show supplemental abnormalities, although we have shown that rhabdomyosarcoma characteristically

exhibits polyploidy [3,9]. Normal cells usually show a diploid karyotype and thus show a pair of fused (yellow) signals if translocations are analyzed using split-signal probes. However, tumor cells exhibit polyploidy more often than diploidy and show three pairs of signals if the tumor cells are triploid (Fig. 1f).

Please cite this article in press as: Sugita S, Hasegawa T, Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors, Journal of Orthopaedic Science (2017), http://dx.doi.org/10.1016/j.jos.2017.02.004

4

S. Sugita, T. Hasegawa / Journal of Orthopaedic Science xxx (2017) 1e12

2.5. Important notes on the practice of FISH To obtain valuable results, surgical materials should be immediately fixed with formalin to avoid tissue degeneration, although the fixation time should not be excessive. Both formalin hypo- and hyperfixation often influence the sensitivity of FISH signals and the signals in such materials are often diminished or absent. Moreover, the decalcification process severely affects signal sensitivity and we often cannot detect any signals in decalcified material. Therefore, we should separate tumor specimens without bony tissues for FISH use to avoid a decalcification process when we handle bone tumor materials, for example, for Ewing sarcoma (ES). Importantly, FISH results using inappropriate materials are in danger of leading us to false-negative findings.

Genetic confirmation of TRS is the most typical use for FISH. While we theoretically can use both split- and fused-signal probes for the detection of chimeric translocations, we prefer to use splitsignal probes. Because some TRSs, especially in soft tissue sarcoma, sometimes have several chimeric genes consisting of a common gene and a different partner gene, even if from an identical tumor type, fused-signal probes particular for a certain fusion may fail to detect positive signals. Thus, we often use two split probes that correspond to an individual gene of chimeric fusion. We occasionally detect positive rearrangement signals in a small percentage of tumor cells in non-specific tumor types. This minor positivity should not be considered indicative of positive rearrangement because the percentage of positive signals is still far below the

Fig. 2. Radiological and pathological findings of soft tissue angiofibroma. a. T2-weighted MRI findings of STA. Coronal T2-weighted image showing a multinodular and heterogeneous intermediate-intensity mass in the subcutaneous tissue of the left medial thigh. The irregular low-intensity areas consisted of hemosiderin deposit and a signal void reflecting intrinsic dilated vessels. Also visible around the mass is a linear signal void indicating peripheral dilated vessels. b. The tumor characteristically consists of random and fascicular proliferations of spindle-shaped tumor cells with collagenous stroma and prominent small branching vessels (200). c. STA exhibiting fibrovascular tumor morphology showing fibroblastic tumor cell proliferation with various types of intermingled vessels. Hemangiopericytoma-like vessels and middle-sized vessels with perivascular fibrinoid necrosis are characteristically observed (40). d. The tumor cells are focally positive for desmin on IHC (200). e. The tumor cells are focally positive for EMA on IHC (200). f. FISH using dual-color and split-signal NCOA2 probes detects a pair of fused (arrow) and split (arrowhead) signals, revealing NCOA2 rearrangement (1000).

Please cite this article in press as: Sugita S, Hasegawa T, Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors, Journal of Orthopaedic Science (2017), http://dx.doi.org/10.1016/j.jos.2017.02.004

S. Sugita, T. Hasegawa / Journal of Orthopaedic Science xxx (2017) 1e12

general cutoff values. Thus, this situation means a non-specific positivity that represents one of the complicated genetic abnormalities. 3. Application of FISH: from research to pathological diagnosis 3.1. Diagnostic utility of NCOA2 FISH in soft tissue angiofibroma Soft tissue angiofibroma (STA) is a recently identified benign soft tissue tumor with a characteristic fibrovascular morphology that tends to affect the lower leg of middle-aged women [10] (Fig. 2a). Histologically, STA consists of random and focally

5

fascicular proliferations of bland, spindle-shaped tumor cells with an abundant collagenous-to-myxoid stroma and prominent, small, thin-walled branching vessels (Fig. 2b). Floret-type multinucleated giant cells are sparsely observed. The tumor also shows slit-like or hemangiopericytoma-like vessels and medium-to-large vessels with perivascular hyalinosis or fibrinoid necrosis (Fig. 2c) [10]. On immunohistochemistry (IHC), STA characteristically shows focal desmin and epithelial membrane antigen (EMA) expression (Fig. 2d and e) [10]. Because of its fibrovascular morphology, the differential diagnosis of STA includes solitary fibrous tumor (SFT), cellular angiofibroma (CAF), low-grade fibromyxoid sarcoma, myxoid liposarcoma, and myxofibrosarcoma. The differential diagnosis of SFT

Fig. 3. Radiological and pathological findings of solitary fibrous tumors. a. MRI findings of SFT. Axial T2-weighted image showing a multinodular heterogeneous intensity mass with striped low-intensity pattern in contact with the posterior surface of the right femur. The mass has a signal void reflecting intrinsic dilated vessels. Vascular pedicles associated with the mass are also noted (arrow). b. SFT consists of patternless or fascicular proliferation of fibroblastic spindle tumor cells with many slit-like or hemangiopericytoma (HPC)-like vessels (40). c. Typical HPC-like vessel showing a stag-horn appearance. HPC-like vessels are a characteristic but not specific histological feature of SFT (100). d. Bland fibroblastic tumor cells surrounded by abundant fibrous stroma intermingled with band-like collagen fibers (200). e. The tumor cells are diffusely positive for CD34 on IHC (200). f. The tumor cells show diffuse nuclear expression of STAT6 on IHC (200).

Please cite this article in press as: Sugita S, Hasegawa T, Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors, Journal of Orthopaedic Science (2017), http://dx.doi.org/10.1016/j.jos.2017.02.004

6

S. Sugita, T. Hasegawa / Journal of Orthopaedic Science xxx (2017) 1e12

and CAF may one of the most difficult and important challenges faced by pathologists, particularly with small specimens [11]. SFT is a fibroblastic/myofibroblastic tumor of rarely metastasizing intermediate malignancy that affects individuals of all ages and both sexes. SFT often occurs in superficial and deep soft tissue in the extremities, head and neck region, thoracic and abdominal cavities, and sometimes visceral organs (Fig. 3a). SFT consists of a patternless proliferation of fibroblastic spindle tumor cells with intermingled collagen bands and is typically accompanied by hemangiopericytoma-like vasculature (Fig. 3bed). On IHC, SFT is positive for CD34 (Fig. 3e) and STAT6 (Fig. 3f), which has recently been established as a specific marker of SFT. CAF is a benign fibroblastic/myofibroblastic tumor that usually occurs in the superficial soft tissue of the vulvovaginal and

inguinoscrotal region in middle-aged and older patients (Fig. 4a). CAF consists of short, intersecting fascicles of bland spindle tumor cells in the fibromyxoid stroma with small-to-mediumesized thick-walled vessels and band-like collagen fibers (Fig. 4b and c). CAF occasionally contains mature adipose tissue (Fig. 4d). On IHC, the tumor cells are positive for CD34 and exhibit loss of retinoblastoma protein expression (Fig. 4e) caused by a 13q14 deletion, which contains the retinoblastoma gene (Fig. 4f). However, STA has a specific fusion of AHRR-NCOA2 derived from t(5;8)(p15;q13), which has been detected using NGS by several researchers [12]. Using the findings of the previous study, we designed a custom dual-color, split-signal probe for NCOA2 FISH and examined its utility in the differential diagnosis of STA and mimicking tumors. Only STA cases showed a NCOA2 split signal in more than 10% of

Fig. 4. Radiological and pathological findings of cellular angiofibroma. a. MRI findings of CAF. Fat-suppressed axial T2-weighted image showing a markedly high-intensity mass surrounded by a thickened low-intensity capsule in the subcutaneous tissue of the anterior pubis. The tumor contains wavy-stripe low-intensity lines and exhibits no expansile growth. b. CAF showing the same fibrovascular tumor histology as STA and SFT. The tumor consists of fascicular proliferation of fibroblastic tumor cells with abundant fibrous stroma and small-to-mediumesized vessels (100). c. The tumor characteristically has thick-walled vessels and band-like collagen fibers (200). d. The tumor focally intermingles with foci of mature adipose tissue (100). e. The tumor cells are negative for retinoblastoma protein because of the deletion of the retinoblastoma locus on 13q14 (200). f. FISH for 13q14 deletion reveals a heterozygous deletion pattern with two green (control) signals and one red (including retinoblastoma gene locus, arrow) signal (1000).

Please cite this article in press as: Sugita S, Hasegawa T, Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors, Journal of Orthopaedic Science (2017), http://dx.doi.org/10.1016/j.jos.2017.02.004

S. Sugita, T. Hasegawa / Journal of Orthopaedic Science xxx (2017) 1e12

7

Fig. 5. Pathological findings of CIC-DUX4 fusion sarcoma of the small intestine. a. The tumor proliferates in a nodular fashion in the submucosa to the seroma of the small intestine (40). b. The tumor consists of a solid proliferation of small-to-mediumesized round tumor cells with a coarse nuclear chromatin pattern and moderate nuclear atypia (400). c. The tumor cells are focally positive for CD99 on IHC (200). d. The tumor cells are positive for ETV4 on IHC (200). e. The tumor cells are focally positive for WT1 on IHC (200). f. FISH using dual-color and split-signal EWSR1 probes fails to detect EWSR1 split signals. g. FISH using dual-color and split-signal CIC probes detects a pair of fused (arrow) and split (arrowhead) signals, revealing CIC rearrangement (1000).

tumor cells (Fig. 2f). We also examined STAT6 IHC: only SFT cases exhibited strong and diffuse nuclear expression of STAT6 and none of the STA cases were positive for STAT6. We concluded that NCOA2 FISH was useful for the confirmative diagnosis of STA and that its combination with STAT6 IHC is a powerful approach for the differential diagnosis of STA and SFT [11]. 3.2. CIC-FOXO4 fusion sarcoma: a rare distinct variant of CICrearranged sarcoma The classification of the so-called “small round cell sarcoma (SRCS)” represented by classical ES has dramatically changed and we cannot discuss SRCSs without reviewing the latest genetic background findings. In addition to ES/Ewing-like sarcoma, several

studies have reported that CIC-rearranged sarcomas show a small round cell morphology somewhat resembling ES/Ewing-like sarcoma without any types of TET/FET (EWSR1 and FUS)-ETS fusions [13]. CIC-rearranged sarcomas account for up to 68% of noneEWSR1-rearranged SRCSs [14]. CIC-rearranged sarcoma tends to emerge in various anatomic sites of patients of all ages and shows lobular proliferation of small round tumor cells with round-to-oval nuclei and scant pale eosinophilic cytoplasm (Fig. 5a). The nuclei exhibit fine-to-coarse nuclear chromatin and one or several small nucleoli (Fig. 5b). On IHC, the tumor cells focally express CD99 (Fig. 5c); this finding is different from the diffuse and strong CD99 expression in ES. Genetically, most CIC-rearranged sarcomas have a CIC-DUX4 fusion derived from t(4;19)(q35;q13) (Fig. 5f and g). The most important clinical feature of CIC-rearranged sarcomas is their

Please cite this article in press as: Sugita S, Hasegawa T, Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors, Journal of Orthopaedic Science (2017), http://dx.doi.org/10.1016/j.jos.2017.02.004

8

S. Sugita, T. Hasegawa / Journal of Orthopaedic Science xxx (2017) 1e12

Fig. 6. Pathological findings of CIC-FOXO4 fusion sarcoma of the posterior neck. a. The tumor consists of solid and sheet-like proliferations of small round tumor cells and shows lobular architectures separated by desmoplastic fibrous stroma (100). b. The tumor cells have slightly irregularly shaped round nuclei with coarse chromatin and conspicuous nucleoli and scant, amphophilic-to-slightly eosinophilic cytoplasm (400). c. The tumor cells are focally positive for CD99 on IHC (200). d. The tumor cells are positive for ETV4 on IHC (200). e. The tumor cells are positive for WT1 on IHC (200). f. FISH using dual-color and split-signal CIC probes detects a pair of fused (arrow) and split (arrowhead) signals, revealing CIC rearrangement (1000). g. FISH using dual-color and fused-signal CIC-FOXO probes detects a pair of fused (arrow) signals and one green (arrowhead) signal, revealing CIC-FOXO4 fusion (1000). f and g: [17] http://journals.lww.com/ajsp/pages/articleviewer.aspx?year¼2014&issue¼11000&article¼00015&type¼abstract with permissions of Wolters Kluwer Health, Inc.

significantly worse prognosis compared with ES/Ewing-like sarcoma. Therefore, we should recognize them as a distinct entity of SRCS from ES/Ewing-like sarcoma [13e15]. Although CIC-rearranged sarcoma must be a distinct entity, it is difficult for pathologists to diagnose such SRCSs by morphology and IHC alone. Some recent studies have indicated a diagnostic utility of ETV4 and WT1 IHC in CIC-rearranged sarcoma (Fig. 5d and e) [16], although genetic analysis including FISH or RT-PCR might be necessary for confirmative diagnosis. We previously reported a CIC-FOXO4 fusion sarcoma derived from t(X;19)(q13;q13.3) as a rare distinct variant of CIC-rearranged sarcoma, which arose in the posterior neck of 63-

year-old male [17]. This tumor consisted of a solid proliferation of small round tumor cells with coarse chromatin and conspicuous nucleoli and showed lobular architectures separated by desmoplastic fibrous stroma (Fig. 6a and b). On IHC, the tumor cells were focally positive for CD99 (Fig. 6c). Moreover, they also were positive for ETV4 and WT1 (Fig. 6d and e). We first diagnosed the SRCS as a CIC-rearranged sarcoma because we detected CIC split signals by FISH using a custom CIC split probe. However, we detected no DUX4 split and CIC-DUX4 fusion signals by custom probes in the SRCS and expected an unknown partner, besides DUX4, which was fused to the CIC gene. We then performed NGS using frozen material of the

Please cite this article in press as: Sugita S, Hasegawa T, Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors, Journal of Orthopaedic Science (2017), http://dx.doi.org/10.1016/j.jos.2017.02.004

S. Sugita, T. Hasegawa / Journal of Orthopaedic Science xxx (2017) 1e12

SRCS and discovered a novel CIC-FOXO4 fusion. Subsequently, we created a custom CIC-FOXO4 fusion probe and confirmed the fusion in the specimens by FISH (Fig. 6f and g). There is one other reported case of CIC-FOXO4 fusion sarcoma and both of them should be categorized as CIC-rearranged sarcoma [17,18]. As in our case, NGS is an essential and powerful method for

9

comprehensive genomic analysis and significantly contributes to the identification of new chimeric fusions in soft tissue and bone tumor. However, as pathologists, we should determine how to apply such genetic findings to diagnosis. The simplest and most effective way to apply research findings to the practical setting is to design FISH probes for pathological diagnosis.

Fig. 7. Radiological and pathological findings of pseudomyogenic hemangioendothelioma. a. Radiological findings of PHE. Multiple bone destructive lesions were found in the femur and tibia. Individual lesions showing well-defined axially elongated bone destruction without a sclerotic rim. The medial condyle of the femur shows slightly expansile growth. b. PHE consists of a fascicular proliferation of bland, spindle-shaped cells with oval nuclei and “pseudomyogenic” eosinophilic cytoplasm. c. PHE sparsely has rhabdomyoblast-like cells with abundant eosinophilic cytoplasm and epithelioid cells. d. The tumor cells are positive for cytokeratin AE1/AE3 on IHC (200). e. The tumor cells are positive for CD31 on IHC (200). f. The tumor cells show nuclear ERG expression on IHC (200). g. The tumor cells show characteristic nuclear FOSB expression on IHC (200).

Please cite this article in press as: Sugita S, Hasegawa T, Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors, Journal of Orthopaedic Science (2017), http://dx.doi.org/10.1016/j.jos.2017.02.004

10

S. Sugita, T. Hasegawa / Journal of Orthopaedic Science xxx (2017) 1e12

3.3. Diagnostic value of specific proteins induced by chimeric fusions: specific FOSB expression in pseudomyogenic hemangioendothelioma The functions of many chimeric genes have not been identified. Nevertheless, most are proposed to play some important roles in tumorigenesis. Some chimeric genes lead to the overexpression of related proteins and their detection has diagnostic value in some

tumors. For example, recent studies using NGS have identified a specific fusion of NAB2-STAT6 derived from an inverted insertion of 12q13 in SFT [19]. This fusion leads to nuclear overexpression of STAT6 protein in tumor cells, and IHC of STAT6 is an effective diagnostic tool for SFT [11,20,21]. As is the case in SFT, epithelioid hemangioendothelioma (EHE), a malignant vascular tumor, has a specific fusion of WWTR1-CAMTA1 [22] and this fusion also leads to nuclear CAMTA1 overexpression; thus, nuclear CAMTA1

Fig. 8. Radiological and pathological findings of epithelioid hemangioendothelioma. a. Radiological findings of EHE. Multiple round-to-oval bone destructive lesions are found in the cortex and medulla in the diaphysis of the femur. The lesions show a thin sclerotic rim and internal calcification in the remaining bone. b. EHE consists of fascicular proliferation of spindle-shaped, eosinophilic tumor cells and the morphology resembles PHE. c. The tumor cells occasionally have intracytoplasmic lumina showing the appearance of primitive vessels. d. The tumor cells are positive for cytokeratin AE1/AE3 on IHC (200). e. The tumor cells are positive for CD31 on IHC (200). f. The tumor cells show characteristic nuclear CAMTA1 expression on IHC (200).

Please cite this article in press as: Sugita S, Hasegawa T, Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors, Journal of Orthopaedic Science (2017), http://dx.doi.org/10.1016/j.jos.2017.02.004

S. Sugita, T. Hasegawa / Journal of Orthopaedic Science xxx (2017) 1e12

expression on IHC is a useful diagnostic indicator of EHE [23] (Fig. 8f). We examined FOSB expression in pseudomyogenic hemangioendothelioma (PHE), which is a rare vascular tumor of rarely metastasizing intermediate malignancy that has been included for the first time in the latest WHO classification [24]. PHE affects young adults and children with a marked male predominance and mainly occurs in the soft tissue and bone of the lower limbs (Fig. 7a). Histologically, PHE shows characteristic myogenic and epithelioid morphology despite its vascular nature (Fig. 7b and c). Furthermore, PHE is usually positive for epithelial and vascular markers on IHC (Fig. 7def). Therefore, practical diagnosis should aim to distinguish PHE from EHE, epithelioid angiosarcoma, Kaposi sarcoma, and epithelioid sarcoma [25]. Of these tumors, EHE is the most important one in the differential diagnosis of PHE due to similarities in their histology and clinical presentation (Fig. 8a). They usually show a fascicular proliferation of relatively bland spindle and/or epithelioid cells with eosinophilic cytoplasm and clinically form multiple musculoskeletal lesions and multifocal lesions in the bone (Fig. 8b and c). They also are positive for some epithelial and vascular markers on IHC (Fig. 8d and e). Recent studies have revealed that PHE has a specific fusion of SERPINE1-FOSB derived from t(7;19)(q22;q13) and that this fusion leads to an elevated expression of FOSB mRNA [26]. Thus, we expected to see specific nuclear FOSB expression in PHE and conducted FOSB IHC in PHE and mimicking tumors. We observed diffuse and strong nuclear expression of FOSB only in PHE patients (Fig. 7g). We also examined CAMTA1 IHC in the same cohort and reached the conclusion that a combination of FOSB and CAMTA1 IHC was beneficial for distinguishing PHE from EHE [25]. IHC detection of specific proteins resulting from chimeric fusions can sometimes be helpful for pathological diagnosis, similar to FISH. 3.4. Diagnostic role of FISH in an expert review of a clinical trial The most important consideration when beginning a clinical trial is the careful selection of patients. Kawai et al. [27] reported the efficacy of trabectedin monotherapy after chemotherapy versus best supportive care in patients with advanced TRS. Trabectedin, which has recently been approved in Japan, is a tetrahydroisoquinoline alkaloid with anti-tumor activities in STSs that binds to the minor groove of DNA and blocks DNA repair machinery. This randomized, open-labeled, phase 2 trial demonstrated that trabectedin significantly reduced the risk of disease progression and death in patients with advanced TRS after

11

standard chemotherapy. We participated in the expert review of the patients, using FISH to confirm the primary pathological diagnosis of each study site [6]. We performed FISH with various commercially available and custom probes to review the 76 enrolled TRS cases (Table 2). Ultimately, we supported the enrollment of 73 patients after excluding 3 cases of primary diagnosed synovial sarcoma (SS): 2 of these cases were ultimately diagnosed as malignant peripheral nerve sheath tumor (MPNST) and 1 as sarcomatoid carcinoma because no SS18 split signals were observed. We realized that the differential diagnosis of MPNST and SS, especially in round cell-type MPNST and poorly differentiated SS, was difficult because the histology and IHC findings sometimes overlapped. Nevertheless, detection of a SS18 split signal by FISH helped us to exclude mimicking SS cases. Among the 73 study patients, we detected no signals in 7; these patients' samples were not suitable for FISH because the specimens had experienced some problems, such as hyperfixation and decalcification. After the exclusion of these 7 patients, we finally detected translocations in 95% (63/66) of the 73 study cases, indicating a high sensitivity (Table 2) [6]. In this way, FISH can play an important role in the genetic confirmation of pathological diagnoses and can be used as an excellent means to select patients in clinical trials. 4. Future directions for FISH: in pursuit of a more practical use in routine examinations There is no doubt that FISH is an essential tool for genetic analysis in routine diagnostic examination because of its practical convenience and reproducibility compared with histology. FISH can be easily performed in many pathology laboratories, although some improvements may be required before the universal adoption of FISH. We believe that standardization of the manually operated procedure is the most important point to ensure FISH accuracy. The automation of individual procedures has dramatically shortened the examination time and might help to reduce the burden on technicians and improve the turn-around-time of pathological diagnoses involving genetic analysis. Indeed, automatic devices for FISH have recently been commercialized, which will improve the convenience and availability of genetic analyses in routine pathological examinations. We should not forget that pathological diagnosis first depends on morphological examination of adequately sampled specimens. This principle is reasonable for not only the efficient practice of daily diagnostic work, but also to reduce the diagnostic costs of genetic studies. Therefore, we are always conscious of the

Table 2 Comparison of pathological diagnoses between study sites and the central review by FISH analysis. Histological type

Study site

Central review

Myxoid liposarcoma Synovial sarcoma Mesenchymal chondrosarcoma Extraskeletal Ewing sarcoma Alveolar soft part sarcoma Alveolar rhabdomyosarcoma Clear cell sarcoma Extraskeletal myxoid chondrosarcoma Dermatofibrosarcoma protuberans Angiomatoid fibrous histiocytoma Desmoplastic small round cell tumor Total

24 21 6 5 5 5 5 2 1 1 1 76

24 18 6 5 5 5 5 2 1 1 1 73

FISH Probe

Positive

Negative

ND

DDIT3 SS18 NCOA2-HEY1 EWSR1 TFE3 FOXO1 EWSR1 NR4A3 PDGFB EWSR1 ERSR1

22 16 3 4 5 4 5 1 1 1 1 63

0 4a 0 0 0 1 0 1 0 0 0 6

2 1b 3 1 0 0 0 0 0 0 0 7

ND, not detected. a 4 FISH-negative cases of synovial sarcoma included 2 MPNSTs, 1 sarcomatoid carcinoma, and 1 synovial sarcoma with SS18-SSX fusion by RT-PCR. b 1 FISH ND case with SS18-SSX fusion by RT-PCR.

Please cite this article in press as: Sugita S, Hasegawa T, Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors, Journal of Orthopaedic Science (2017), http://dx.doi.org/10.1016/j.jos.2017.02.004

12

S. Sugita, T. Hasegawa / Journal of Orthopaedic Science xxx (2017) 1e12

importance of morphological diagnosis and do not hesitate to consult special pathologists if a diagnosis is challenging. [12]

5. Conclusion FISH is an essential tool for the pathological diagnosis of soft tissue and bone tumors. It can detect various genetic abnormalities in an “in situ” fashion using FFPE specimens on glass slides during routine examinations. The latest genetic findings can be applied to routine pathological practice by using experimental results to design appropriate probes. Improvements in FISH devices, such as automation of manual procedures, will help it to become a universal method for the genetic analysis of pathological diagnoses.

[13]

[14]

[15]

Conflict of interest The authors declare that they have no conflict of interest. [16]

Appendix A. Supplementary data [17]

Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jos.2017.02.004. References

[18]

[1] Guseva NV, Tanas MR, Stence AA, Sompallae R, Schade JC, Bossler AD, Bellizzi AM, Ma D. The NAB2-STAT6 gene fusion in solitary fibrous tumor can be reliably detected by anchored multiplexed PCR for targeted nextgeneration sequencing. Cancer Genet 2016 Jul-Aug;209(7e8):303e12. [2] Mertens F, Antonescu CR, Hohenberger P, Ladanyi M, Modena P, D'Incalci M, Casali PG, Aglietta M, Alvegård T. Translocation-related sarcomas. Semin Oncol 2009 Aug;36(4):312e23. [3] Miura Y, Keira Y, Ogino J, Nakanishi K, Noguchi H, Inoue T, Hasegawa T. Detection of specific genetic abnormalities by fluorescence in situ hybridization in soft tissue tumors. Pathol Int 2012 Jan;62(1):16e27. [4] Noguchi H, Mitsuhashi T, Seki K, Tochigi N, Tsuji M, Shimoda T, Hasegawa T. Fluorescence in situ hybridization analysis of extraskeletal myxoid chondrosarcomas using EWSR1 and NR4A3 probes. Hum Pathol 2010 Mar;41(3): 336e42. [5] Nakayama R, Miura Y, Ogino J, Susa M, Watanabe I, Horiuchi K, Anazawa U, Toyama Y, Morioka H, Mukai M, Hasegawa T. Detection of HEY1-NCOA2 fusion by fluorescence in-situ hybridization in formalin-fixed paraffin-embedded tissues as a possible diagnostic tool for mesenchymal chondrosarcoma. Pathol Int 2012 Dec;62(12):823e6. [6] Sugita S, Asanuma H, Hasegawa T. Diagnostic use of fluorescence in situ hybridization in expert review in a phase 2 study of trabectedin monotherapy in patients with advanced, translocation-related sarcoma. Diagn Pathol 2016 Apr 12;11:37. [7] Park JK, Sugita S, Soma T, Yamashiro K, Hasegawa T. Malignant peripheral nerve sheath tumor coexisting with congenital melanocytic nevus in an elderly man. Pathol Int 2016 May;66(5):309e10. [8] Nabeshima K, Matsumoto S, Hamasaki M, Hida T, Kamei T, Hiroshima K, Tsujimura T, Kawahara K. Use of p16 FISH for differential diagnosis of mesothelioma in smear preparations. Diagn Cytopathol 2016 Sep;44(9):774e80. [9] Matsumura T, Yamaguchi T, Seki K, Shimoda T, Wada T, Yamashita T, Hasegawa T. Advantage of FISH analysis using FKHR probes for an adjunct to diagnosis of rhabdomyosarcomas. Virchows Arch 2008 Mar;452(3):251e8. ~ o-Enríquez A, Fletcher CD. Angiofibroma of soft tissue: clinicopathologic [10] Marin characterization of a distinctive benign fibrovascular neoplasm in a series of 37 cases. Am J Surg Pathol 2012 Apr;36(4):500e8. [11] Sugita S, Aoyama T, Kondo K, Keira Y, Ogino J, Nakanishi K, Kaya M, Emori M, Tsukahara T, Nakajima H, Takagi M, Hasegawa T. Diagnostic utility of NCOA2

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

fluorescence in situ hybridization and Stat6 immunohistochemistry staining for soft tissue angiofibroma and morphologically similar fibrovascular tumors. Hum Pathol 2014 Aug;45(8):1588e96. €ller E, Nord KH, Mandahl N, Von Steyern FV, Domanski HA, Marin ~ oJin Y, Mo Enríquez A, Magnusson L, Nilsson J, Sciot R, Fletcher CD, Debiec-Rychter M, Mertens F. Fusion of the AHRR and NCOA2 genes through a recurrent translocation t(5;8)(p15;q13) in soft tissue angiofibroma results in upregulation of aryl hydrocarbon receptor target genes. Genes Chromosomes Cancer 2012 May;51(5):510e20. Yoshida A, Goto K, Kodaira M, Kobayashi E, Kawamoto H, Mori T, Yoshimoto S, Endo O, Kodama N, Kushima R, Hiraoka N, Motoi T, Kawai A. CIC-rearranged sarcomas: a study of 20 cases and comparisons with Ewing sarcomas. Am J Surg Pathol 2016 Mar;40(3):313e23. Italiano A, Sung YS, Zhang L, Singer S, Maki RG, Coindre JM, Antonescu CR. High prevalence of CIC fusion with double-homeobox (DUX4) transcription factors in EWSR1-negative undifferentiated small blue round cell sarcomas. Genes Chromosomes Cancer 2012 Mar;51(3):207e18. Choi EY, Thomas DG, McHugh JB, Patel RM, Roulston D, Schuetze SM, Chugh R, Biermann JS, Lucas DR. Undifferentiated small round cell sarcoma with t(4;19)(q35;q13.1) CIC-DUX4 fusion: a novel highly aggressive soft tissue tumor with distinctive histopathology. Am J Surg Pathol 2013 Sep;37(9): 1379e86. Hung YP, Fletcher CD, Hornick JL. Evaluation of ETV4 and WT1 expression in CIC-rearranged sarcomas and histologic mimics. Mod Pathol 2016 Nov;29(11):1324e34. Sugita S, Arai Y, Tonooka A, Hama N, Totoki Y, Fujii T, Aoyama T, Asanuma H, Tsukahara T, Kaya M, Shibata T, Hasegawa T. A novel CICFOXO4 gene fusion in undifferentiated small round cell sarcoma: a genetically distinct variant of Ewing-like sarcoma. Am J Surg Pathol 2014 Nov;38(11):1571e6. Solomon DA, Brohl AS, Khan J, Miettinen M. Clinicopathologic features of a second patient with Ewing-like sarcoma harboring CIC-FOXO4 gene fusion. Am J Surg Pathol 2014 Dec;38(12):1724e5. Chmielecki J, Crago AM, Rosenberg M, O'Connor R, Walker SR, Ambrogio L, Auclair D, McKenna A, Heinrich MC, Frank DA, Meyerson M. Whole-exome sequencing identifies a recurrent NAB2-STAT6 fusion in solitary fibrous tumors. Nat Genet 2013 Feb;45(2):131e2. Yoshida A, Tsuta K, Ohno M, Yoshida M, Narita Y, Kawai A, Asamura H, Kushima R. STAT6 immunohistochemistry is helpful in the diagnosis of solitary fibrous tumors. Am J Surg Pathol 2014 Apr;38(4):552e9. Doyle LA, Vivero M, Fletcher CD, Mertens F, Hornick JL. Nuclear expression of STAT6 distinguishes solitary fibrous tumor from histologic mimics. Mod Pathol 2014 Mar;27(3):390e5. Errani C, Zhang L, Sung YS, Hajdu M, Singer S, Maki RG, Healey JH, Antonescu CR. A novel WWTR1-CAMTA1 gene fusion is a consistent abnormality in epithelioid hemangioendothelioma of different anatomic sites. Genes Chromosomes Cancer 2011 Aug;50(8):644e53. Shibuya R, Matsuyama A, Shiba E, Harada H, Yabuki K, Hisaoka M. CAMTA1 is a useful immunohistochemical marker for diagnosing epithelioid haemangioendothelioma. Histopathology 2015 Dec;67(6):827e35. Hornick JL, Fletcher CDM, Mertens F. Psuedomyogenic hemangioendothelioma. In: Fletcher CDM, Bridge JA, Hogendoorn PCW, Mertens, editors. WHO classification of tumours of soft tissue and bone. 4th ed. Lyon: IARC press; 2013. p. 153e4. Sugita S, Hirano H, Kikuchi N, Kubo T, Asanuma H, Aoyama T, Emori M, Hasegawa T. Diagnostic utility of FOSB immunohistochemistry in pseudomyogenic hemangioendothelioma and its histological mimics. Diagn Pathol 2016 Aug 11;11(1):75. € rn H, Magnusson L, Nilsson J, von Walther C, Tayebwa J, Lilljebjo Steyern FV, Øra I, Domanski HA, Fioretos T, Nord KH, Fletcher CD, Mertens F. A novel SERPINE1-FOSB fusion gene results in transcriptional up-regulation of FOSB in pseudomyogenic haemangioendothelioma. J Pathol 2014 Apr;232(5):534e40. Kawai A, Araki N, Sugiura H, Ueda T, Yonemoto T, Takahashi M, Morioka H, Hiraga H, Hiruma T, Kunisada T, Matsumine A, Tanase T, Hasegawa T, Takahashi S. Trabectedin monotherapy after standard chemotherapy versus best supportive care in patients with advanced, translocation-related sarcoma: a randomised, open-label, phase 2 study. Lancet Oncol 2015 Apr;16(4):406e16.

Please cite this article in press as: Sugita S, Hasegawa T, Practical use and utility of fluorescence in situ hybridization in the pathological diagnosis of soft tissue and bone tumors, Journal of Orthopaedic Science (2017), http://dx.doi.org/10.1016/j.jos.2017.02.004