Introduction

Translocation events involving FN1 have been described in a wide variety of neoplasms, all of which share the presence of a cartilage matrix with or without a variable extent of calcification [1,2,3]. Specifically, synovial chondromatosis, characterized by multinodular growth of mature cartilaginous tissue with clustering of chondrocytes has recently been shown as frequently harboring the FN1-ACVR2 translocation [2]. The fusion of FN1 to FGFR1 or FGFR2 has also been described in soft tissue chondroma, particularly in examples showing grungy to lacy (chondroblastoma-like) calcification [2]. Phosphaturic mesenchymal tumor, frequently harboring FN1-FGFR1 or FN1-FGF1 fusions, shows variable histologic features, and is generally characterized as a proliferation of bland spindled to stellate cells with associated flocculent-appearing calcified, chondroid or ossified matrix within a highly vascularized stroma [3, 4]. Lastly, calcifying aponeurotic fibroma, a proliferation of bland fibroblastic cells with calcified fibrocartilage-like nodules, have been shown to harbor recurrent FN1-EGF gene fusions [1]. Besides these four entities, similar histologic features may be seen in a wide spectrum of cartilaginous neoplasms of soft tissue. These include so-called chondroid tenosynovial giant cell tumor (TGCT), calcium pyrophosphate dihydrate (CPPD) deposition disease (tophaceous pseudogout), and chondroblastoma [5,6,7,8,9]. These tumors in some cases share overlap in cytomorphology and histologic features, sometimes making precise histologic classification challenging.

In this study, we present the molecular, clinical, and pathologic features of 12 cases of soft tissue tumors with a chondroid matrix showing an anatomic predilection for the temporomandibular joint (TMJ) as well as the distal extremities and define a group of tumors harboring FN1 gene fusions which we will refer to as calcified chondroid mesenchymal neoplasms. We also review the literature and discuss the relevant differential diagnosis of various entities highlighting their overlapping and distinct histologic and molecular features.

Methods and materials

Study design and case selection

Our pathology group has noted the similarity in morphologic features among tumors in our archives initially classified as various entities such as “chondroid TGCT,” “tophaceous pseudogout,” or simply “chondroid neoplasm”, for which the differential diagnosis included chondroid TGCT, tophaceous pseudogout, and chondroma. In order to investigate whether these lesions represented distinct entities or a morphologic spectrum of related entities, 12 such cases (nine from the University of Washington Department of Pathology Archives and three from the archived cases sent in consultation to CDMF at the Brigham and Women’s Hospital) were retrieved and histologic slides were independently reviewed by four pathologists (EYC, RWR, JGM, and CDMF). All cases were subjected to targeted RNA sequencing using the ArcherDX FusionPlex panel. Clinical, pathologic, and molecular findings were recorded.

This project was approved by the Institutional Review Board at the University of Washington.

Targeted RNA sequencing

All twelve specimens used in this study were archived formalin-fixed paraffin-embedded (FFPE) tissue specimens. Total nucleic acid (TNA) was extracted from the FFPE specimens using AllPrep DNA/RNA FFPE kit according to the manufacturer’s recommended protocol (Qiagen, Valencia, CA, USA). The Fusionplex RNA-sequencing assay was performed using a customized 115-gene panel covering a wide spectrum of cancer genes known for their involvement in gene fusions in neoplasia including, but not limited to, FN1, CSF1, FGFR1, and FGFR2 (ArcherDx, Inc. Boulder, CO). The methods for Fusionplex RNA-sequencing analysis have been described previously [10]. The genes analyzed by Fusionplex including their corresponding NM transcript identification numbers are listed in Table S1.

Sanger sequencing confirmation

For the positive fusions detected by targeted RNA-seq, RT-PCR and Sanger sequencing were performed to confirm the fusions and breakpoints at the RNA level. Once cDNAs were synthesized by random priming with Fusionplex reagent kit (ArcherDX, Boulder, CO, USA), they were subjected to a polymerase chain reaction (PCR) using FastStart Taq Polymerase (Roche Diagnostics, Indianapolis, IN, USA) with specific primers designed for each partner gene (Supplemental Materials and Methods). PCR products were cloned using TOPO TA Cloning Kit (Invitrogen, Carlsbad, CA, USA) and then were Sanger-sequenced using Eurofins Genomics Tube Sequencing service (Eurofins Genomics, Louisville, KY, USA).

Results

Clinical and pathologic findings

Our series of 12 patients includes six women and six men with a mean age of 55.7 years (range 22–72 years). Anatomic locations include the TMJ and/or temporal bone in six cases (50%), hand/digits in five cases (42%), and foot in one case (8%). Tumor size, when available, ranged from 0.5 to 4.0 cm (mean 3.1 cm). All tumors were removed by surgical excision. Follow-up information was available for three cases, with a mean duration of 1.7 years. No recurrence or metastasis was reported in any case. See Table 1 for complete clinicopathologic information.

Table 1 Clinicopathological features and molecular characterization of gene fusions.
Full size table

Histologically, all tumors showed multinodular architecture and chondroid to the cartilaginous matrix (Fig. 1A, B, cases 2 and 4) with increased cellularity towards the periphery of the nodules. The matrix frequently showed calcification that was coarse or grungy to lacey (chondroblastoma-like) (Figs. 1C, D and 5A, cases 2, 4, and 11). Remarkably, the grungy calcifications in cases 5 and 12 were intensely basophilic, appeared crystalline (Case 5 shown in Fig. 2A, B), and examination using polarized light revealed refractive rhomboid crystals consistent with CPPD deposition (Fig. 2C). The tumor cells within the chondroid/cartilage matrix were polygonal to stellate with abundant eosinophilic cytoplasm and eccentrically placed nuclei with small nucleoli, while the cells in fibrous septa were more often smaller and spindled with fibroblastic features (Figs. 3A–C, 4A–D, and 5A, B, cases 2, 3, 8–11). Osteoclast-like giant cells were present in all cases and careful examination could often find at least focal areas resembling TGCT (epithelioid to histiocytoid cells with eccentric nuclei and occasionally hemosiderin deposition) in most cases (Fig. 4A–D). TGCT-like features were particularly prominent in cases 9 and 10 (Fig. 4A–D), which also had the least amount of chondroid stroma.

Fig. 1: Histologic features of case 2 and case 4 with FN1-FGFR2 fusion.
figure 1

A, B Characteristic lobular architecture. C, D Grungy to lacey (chondroblastoma-like) calcifications. E, F Polygonal to spindled cells within the chondroid matrix frequently associated with osteoclast-like giant cells, while septa separating lobules contain spindled, fibroblastic cells. Panels (A), (C), and (E) represent case 2. Panels (B), (D), and (F) represent case 4.

Full size image
Fig. 2: Case 5 (FN1-FGFR2 fusion) with CPPD crystal deposition.
figure 2

A, B Case 5 with extensive basophilic grungy calcification at low power (A) and epithelioid to differentiated chondrocytes within crystalline deposits at high power (B). C Refractile rhomboid crystals (some indicated by arrows) when viewed using polarized light consistent with CPPD.

Full size image
Fig. 3: Histologic features of case 7 with FN1-MERTK fusion.
figure 3

A Vaguely lobular architecture with circumscribed tumor border [2]. B Cells within the variably chondroid matrix and collagenous stroma are polygonal to oval with eosinophilic cytoplasm and eccentrically placed nuclei.

Full size image
Fig. 4: Calcified chondroid mesenchymal neoplasm in this study with features resembling tenosynovial giant cell tumor (TGCT).
figure 4

A, B Case 9 predominantly shows the histologic features of epithelioid cells with eccentric nuclei admixed with osteoclast-like giant cells, resembling some features of TGCT (A) as well as a focal area of epithelioid to stellate cells in a chondroid matrix with calcification (B). C, D Case 10 occurring in the TMJ for comparison with case 9 showing mononuclear epithelioid/histiocytoid cells with eccentric nuclei and ring-like distribution of cytoplasmic hemosiderin as well as multinucleated giant cells (C). The focal area of chondroid metaplasia is present (D).

Full size image
Fig. 5: Histologic features of case 11 with FN1-NTRK1 fusion.
figure 5

A Lobular growth pattern with chondroid matrix and grungy calcifications. B Ovoid to spindled cells with eccentric nuclei and eosinophilic cytoplasm within the chondroid matrix. Occasional osteoclast-like giant cells are seen.

Full size image

Molecular findings

An in-frame gene fusion was detected in 10 of the 12 tumors tested. The fusion genes of 10 tumors were comprised of FN1 as the 5′ partner gene and various 3′ partner genes including FGFR2 in 5 cases (cases 1–5), FGFR1 in one case (case 7), and novel partner genes MERTK, TEK, and NTRK1 in three cases (cases 8, 9 and 11, respectively) (Table 1, Fig. 6). An FGFR1-PLAG1 fusion gene was detected in 1 tumor with FGFR1 as the 5′ partner and PLAG1 as the 3′ partner (case 6). All fusion transcripts were further verified by RT-PCR and Sanger sequencing (Fig. S1). No gene fusion was detected in case 10 and case 12.

Fig. 6: Schematic of chimeric transcripts and proteins resulting from gene fusions.
figure 6

A RNA transcripts with exon structures and related protein domains in the genes involved in the fusions. B Chimeric transcripts and proteins detected in the cohort of soft tissue chondroma of this study and related tumors in the literature, showing the retained exons and functional domains of the fusion genes of FN1, FGFR2, FGFR1, MERTK, NTRK1, TEK, and PLAG1. Case 102 is the only case with FN1-FGFR2 characterized by RNA-seq previously [2]. *PMTs = phosphaturic mesenchymal tumors with FN1-FGFR1 fusions and range of breakpoints reported in the literature [3, 13, 26, 28, 29]. *PAs = pleomorphic adenoma of salivary gland origin with FGFR1-PLAG1 fusions and range of breakpoints reported in the literature [41]. Untranslated regions (5′ UTR and 3′ UTR) are shown as narrow bars. Exons are shown as boxes with numbers. Protein domains are represented by shapes with keys shown in the box: SP = signal peptide, FN I = FN type 1 domain, FN II = FN type 2 domain, FN III = FN type 3 domain, Ig = immunoglobulin-Like loop domain, TM = transmembrane domain, and TK = tyrosine kinase, EGF = EGF-like domain, LRRCT = leucine-rich repeat C-terminal domain, C2H2 Zn = C2H2 Zn finger, and NLS = nuclear localization signal.

Full size image

Gene fusions involved various breakpoints within the coding sequence of the 5′ partner gene FN1, ranging from the 3′ end of exon 11 to exon 48, retaining the signal peptide domain (SP), all FN type 1 domains (FN1 binding assembly domain), and FN type 2 domains, and up to sixteen FN type 3 domains (Fig. 6). The various 3′ partner genes of the fusion transcripts had breakpoints at the 5′ ends of exon 5 to exon 7 of FRGR2 in 5 cases, exon 9 of FGFR1, exon 2 of MERTK, exon 7 of NTRK1, and exon 13 of TEK (Fig. 6, Table 1), retaining transmembrane domain (TM) and tyrosine kinase (TK) domain in all eight cases. However, up to two extracellular FGF-binding (Ig-like) domains were retained in the fusions, including two of three Ig-like domains (Ig2 and Ig3) of FGFR2 in cases 2–5, Ig3 only of FGFR2 in case 1, Ig1 and Ig2 of MERTK in case 8, no Ig-like domain of FGFR1 in case 7, and no Ig-like domain of TEK in case 9, and no Ig-like domain of NTRK1 in case 11 (Fig. 6, Table 1). Of note, two alternatively spliced in-frame fusion transcripts of FN1-FGFR2 were detected in case 2, and three alternatively spliced fusion transcripts of FN1-NTRK1 were detected in case 11 with one in-frame fusion (3′ exon 21 of FN1 to 5′ exon 7 of NTRK1) and apparently two shorter out-of-frame fusions (3′ exon 21 of FN1 to 5′ exon 8 or 10 of NTRK1). One tumor harbored an FGFR1-PLAG1 fusion gene with breakpoints between exons 1 and 2 of FGFR1 and between exons 2 and 3 of PLAG1.

Production of FGF23 in phosphaturic mesenchymal tumors, a subset of which harbors the FN1-FGFR1 fusion, has been demonstrated in cases with or without osteomalacia [11]. To assess whether FGF23 is expressed in the neoplasms of our case series, we performed RT-PCR on the eight tumor samples with sufficient RNA (cases 1–4, 7, 9, and 11) using primers against regions spanning exons 1 and 2 as well as exons 2 and 3 of FGF23. None of the tumor samples tested expressed FGF23 (Fig. S2).

Discussion

In this study, we identified the presence of FN1 fusion with genes encoding receptor TKs in ten calcified chondroid mesenchymal neoplasms as well as FGFR1-PLAG1 in one tumor (Table 1). Gene fusions involving FN1-FGFR2 or FN1-FGFR1 have been recently described in a subset of soft tissue chondromas, which were mostly demonstrated indirectly using break-apart FISH probes for FN1, FGFR1, and/or FGFR2 [2]. To our knowledge, only one case of soft tissue chondroma with FN1-FGFR2 has been characterized by RNA sequencing and showed a fusion between exons 1–19 of FN1 and exons 7–17 of FGFR2 [2]. The breakpoints in FN1 and FGFR2 reported, in this case, are in the range of breakpoints defined in the five tumors harboring FN1-FGFR2 fusions in our series (Fig. 6B). We also described three novel gene fusions, FN1-MERTK, FN1-TEK, and FN1-NTRK1 (Table 1, Fig. 6). FN1 encodes fibronectin-1, a glycoprotein present in a dimeric form in plasma and in dimeric/multimeric forms at the cell surface and in the extracellular matrix, which participates in cell adhesion and migration processes [12]. All fusion partners of FN1 detected in our series encode for receptor TKs; these include FGFR2 (fibroblast growth factor receptor 2), FGFR1, MERTK (MER proto-oncogene, TK), TEK (TEK TK, endothelial), NTRK1 (neurotrophic receptor TK). These fusions result in the in-frame fusion of the N-terminal region of FN1 to the intact transmembrane and TK domains of the receptor TK, and as a result, may promote dimerization of the fused receptor through the fibronectin domain and thereby lead to aberrant signal activation as previously hypothesized [3]. The presence of fusions involving these potentially targetable receptor TKs suggests potential alternative therapeutic avenues for treating such tumors if needed.

Fusions of FN1 to other receptor TKs have also been detected in a variety of neoplasms other than soft tissue chondroma, including FN1-FGFR1 and FN1–FGF1 in phosphaturic mesenchymal tumor [3, 13], FN1-EGF in lipofibromatosis, and calcifying aponeurotic fibroma [1, 14], FN1-AVCR2A in synovial chondromatosis [2], FN1-ALK in gastrointestinal leiomyoma and inflammatory myofibroblastic tumor [15,16,17], FN1-IGF1R in ALK-negative inflammatory myofibroblastic tumor [18], and FN1-ROS1 in infantile inflammatory myofibroblastic tumors [19].

Soft tissue chondromas in general are characterized by nodular/lobular growth of well-differentiated chondrocytes. However, the chondroblastoma-like variant of soft tissue chondroma as described by Cates et al. also shows cellular foci of epithelioid chondrocytes of varying size admixed with osteoclast-like giant cells within a variable amount of chondroid matrix that is frequently accompanied by lace-like calcification [20]. While most cases of soft tissue chondroma have been documented to occur in the extremities, lesions involving the parotid region and skull base have been reported [21]. The cases in our series, particularly the ones with FN1-FGFR1, FN1-FGFR2, FN1-NTRK1, and FN1-MERTK fusions, show significant morphologic overlap with previous morphologic descriptions of the chondroblastoma-like variant of soft tissue chondroma; however, we also describe TGCT-like features as well as CPPD crystal deposition, highlighting a broader morphologic spectrum in this group of entities. We also report a larger proportion of these lesions affecting the TMJ region. We, therefore, believe that this group of neoplasms may not be best classified as simply a variant of soft tissue chondroma and have chosen to apply the term “calcified chondroid mesenchymal neoplasm” for the purposes of this study.

A few of the cases in our series warrant special consideration (see a summary of differential diagnosis in Table 2). Two tumors (case 5 in the TMJ and case 12 in the finger) showed extensive deposition of calcium pyrophosphate dehydrate (CPPD) crystals. The FN1-FGFR2 fusion was identified in case 5. Tophaceous pseudogout is a mass-forming (tumoral) deposition of CPPD crystals that has been described in a number of locations near joints including the TMJ and some show erosion into the skull base [7, 22, 23]. Cellular infiltrates resembling foreign body type giant cell reaction are typically present, and cartilage/chondroid tissue (supposedly metaplastic) may also be seen in a significant subset of examples [7, 23]. Cates et al. previously observed CPPD deposition in one of their chondroblastoma-like soft tissue chondromas [19]. Given that one fusion-positive case in our series also contained CPPD deposition, these observations call into question whether tumoral (massive) tophaceous pseudogout, or at least a subset of cases with chondroid stroma, may actually represent neoplasms harboring recurrent gene fusions including FN1.

Table 2 Summary of the differential diagnosis for calcified chondroid mesenchymal neoplasm (CCMN) (current study).
Full size table

For the one tumor in our series harboring an FN1-FGFR1 fusion gene (Fig. 6B, case 7), we considered the possibility of phosphaturic mesenchymal tumor (PMT). PMTs can take on a wide range of histologic appearances, but in general are characterized as having a highly vascular stroma (some with hemangiopericytoma-like pattern) with bland spindled to stellate neoplastic cells and amorphous basophilic chondro-osseous matrix with grungy calcification [4, 24, 25]. A significant subset of PMT was also shown to harbor FN1-FGFR1 fusions and are frequently associated with hypophosphatemia and tumor-induced osteomalacia secondary to the paraneoplastic secretion of fibroblast growth factor 23 (FGF23) [3, 13]. Examples of PMT without apparent tumor-induced osteomalacia, however, have also been described [25] as well as fusion-negative cases that frequently overexpress α-klotho, a transmembrane enzyme that acts as an FGF23 activator [26]. α-Klotho interacts directly with FGFR1 and forms a high-affinity binding site at the Ig3 domain for FGF23 [27]. None of the cases in our series, including the one with FN1-FGFR1, exhibited the highly vascular stroma or spindle cell proliferation typical of PMT and none of these patients had any known osteomalacia. Furthermore, unlike most FN1-FGFR1 fusions characterized by sequencing in PMT which retained mostly all 3 or 2 of Ig domains of FGFR1 [3, 13, 28, 29], the FN1-FGFR1 fusion in case 7 did not contain any Ig domains of FGFR1 (Fig. 6B). Without the Ig3 domain, the FN1-FGFR1 fusion protein could not interact with FGF23. By RT-PCR, we also did not detect the expression of FGF23 in the eight tumor samples with either FN1-FGFR1 fusion, FN1-FGFR2, FN1-TEK, or FN1-NTRK1 (Fig. S2). Given these findings, we believe case 7 in our series does not represent PMT and is best classified as a calcified chondroid mesenchymal neoplasm. A recent description of FN1-FGFR1 fusions in three cases of soft tissue chondroma was indirectly demonstrated using break-apart FISH probes for FN1 and FGFR1 without the knowledge of protein domains retained [2]. It would be informative to assess whether these three cases retain any Ig domain. The lack of significant levels of FGF23 in these three cases of soft tissue chondromas suggest a potential lack of Ig domains in their fusions. In all, our study has demonstrated the effectiveness of RNA-seq as a tool for identifying functional domains retained in gene fusions and in turn facilitating correlation of molecular findings with clinicopathological features and tumor classification.

We detected an FGFR1-PLAG1 in one case involving the TMJ. FGFR1-PLAG1 fusions have been previously detected in pleomorphic adenoma (PA) of salivary gland origin and, in such cases, the 5′-portion of FGFR1, excluding the TK domain, is fused to the entire PLAG1 coding sequence [30, 31]. This fusion gene arrangement is the same as that found in case 6 of our series. While PA is principally defined histologically by the proliferation of ductal and myoepithelial cells within hyalinized to the myxoid stroma, cartilaginous components may also be seen and can be quite extensive [32]. Our case was entirely composed of lobular cartilaginous tissue with variable, often grungy, calcification without any identifiable ductal or myoepithelial population and therefore histologically appeared essentially indistinguishable from the other chondroblastoma-like soft tissue chondromas in our series. In such an archived case, excluding the possibility of a PA with extensive cartilaginous components in which ductal/myoepithelial components were either not present or not sampled is problematic; however, we found no morphologic features to classify it as PA. We must also consider the possibility that FGFR1-PLAG1 may represent an alternate fusion gene found in non-PA soft tissue cartilaginous neoplasms.

Chondroid TGCT (or TGCT with chondroid metaplasia) is a rare tumor with a predilection for the TMJ region and skull base that often demonstrates locally aggressive growth including bone destruction [5, 6, 33,34,35]. Histologically, chondroid TGCT shows features of conventional TGCT, such as the sheet-like proliferation of large epithelioid to histiocytoid mononuclear cells, some with hemosiderin deposition often in a ring-like deposition around the cytoplasm, and multinucleated giant cells; however, geographic or nodular areas of the chondroid matrix are also present and frequently associated with grungy, lace-like calcifications. While the majority of conventional TGCTs have been shown to harbor CSF1 fusions [36], this genetic event has not been demonstrated in chondroid TGCT. Most cases in our series showed at least focal morphologic features resembling TGCT, although these features were particularly prominent in cases 9 and 10. The finding of FN1-TEK fusion in case 9 raises the possibility that lesions previously classified as chondroid TGCT may actually fall into this group of calcified chondroid mesenchymal neoplasms with FN1 gene fusions; however, case 10, which was originally classified as chondroid TGCT, tested negative for gene fusions, including FN1, FGFR1, FGFR2, and CSF1. It remains to be determined whether other tumors described as chondroid TGCT represent a distinct tumor type or if they harbor other gene fusions that are yet to be discovered.

Additional differential diagnostic considerations include chondroblastoma and synovial chondromatosis. Chondroblastoma most often occurs in the epiphyses of long bones but may also affect the craniofacial skeleton. Histologically, it is characterized by sheets of uniform round to polygonal mononuclear cells with distinct cell borders and grooved nuclei with scattered islands of cartilaginous differentiation often showing lacey (“chicken-wire”) calcification [37]. While some of our cases show a lacey calcification pattern within the cartilaginous elements, our tumors generally lack the sheet-like cellularity and characteristic cytomorphology of chondroblastoma. Lacey (chondroblastoma-like) calcifications have also been previously described in chondroblastoma-like soft tissue chondroma [20, 21, 38]. In addition, H3F3B K36M mutation has been detected in greater than 90% of chondroblastomas, and immunohistochemical staining using a mutation-specific antibody has been shown to be highly sensitive and specific [39, 40]. Immunohistochemical staining for the K36M mutant was negative in one case with FN1-FGFR2 fusion (Case 2; data not shown). Synovial chondromatosis is characterized by a strikingly nodular proliferation of fairly uniform, mature, hyaline, cartilaginous tissue, often with distinct clustering of chondrocytes, which involves the joint synovium, tendon sheath, or bursa. A subset of cases harbors FN1-ACVR2A fusion and this finding has been described as a distinguishing feature of synovial chondromatosis from soft tissue chondroma [2]. While the tumors in our series do show lobulated growth of cartilaginous tissue, they lack the distinct clustering of chondrocytes and are negative for the FN1-ACVR2 fusion gene. Overall, our findings indicate that the tumors in our series represent distinct entities from both chondroblastoma and synovial chondromatosis.

In summary, of the 12 tumors we analyzed, we described ten examples harboring gene fusions. In nine tumors, we detected FN1 fusions with receptor TK genes including FN1-FGFR2 and FN1-FGFR1, as well as the novel fusions FN1-MERTK, FN1-TEK, and FN1-NTRK1. Unlike the FN1-FGFR1 fusions found in PMT, the FN1-FGFR1 fusion detected in our series did not retain the Ig3 domain of the FGF23 binding site. While the cases in our series showed morphologic overlap with chondroblastoma-like soft tissue chondroma, we also reported the rare finding of extensive CPPD crystal deposition in one fusion-positive case as well as TGCT-like features, which were extensive in two cases (one fusion-positive). We, therefore, are expanding the morphologic spectrum of tumors with FN1-receptor TK fusions beyond those previously described as soft tissue chondroma and therefore propose the term calcified chondroid mesenchymal neoplasm. Our findings also raise the question of whether some lesions previously classified as tophaceous pseudogout (particularly with chondroid stroma) and chondroid TGCT may actually represent neoplasms within this spectrum harboring recurrent FN1 gene fusions. The presence of FGFR1-PLAG1 in one of our cases suggests that PA of salivary gland origin should be considered in the differential diagnosis of cartilage forming lesions affecting the craniofacial region. While most of these tumors are amenable to surgical excision, the structure of these fusion genes indicates that therapeutic targeting of the receptor TKs may be a promising alternative treatment avenue if needed.