Mesenchymal tumours represent one of the most challenging field of diagnostic pathology and refinement of classification schemes plays a key role in improving the quality of pathologic diagnosis and, as a consequence, of therapeutic options. The recent publication of the new WHO classification of Soft Tissue Tumours and Bone represents a major step toward improved standardization of diagnosis. Importantly, the 2020 WHO classification has been opened to expert clinicians that have further contributed to underline the key value of pathologic diagnosis as a rationale for proper treatment. Several relevant advances have been introduced. In the attempt to improve the prediction of clinical behaviour of solitary fibrous tumour, a risk assessment scheme has been implemented. NTRK-rearranged soft tissue tumours are now listed as an “emerging entity” also in consideration of the recent therapeutic developments in terms of NTRK inhibition. This decision has been source of a passionate debate regarding the definition of “tumour entity” as well as the consequences of a “pathology agnostic” approach to precision oncology. In consideration of their distinct clinicopathologic features, undifferentiated round cell sarcomas are now kept separate from Ewing sarcoma and subclassified, according to the underlying gene rearrangements, into three main subgroups (CIC, BCLR and not ETS fused sarcomas) Importantly, In order to avoid potential confusion, tumour entities such as gastrointestinal stroma tumours are addressed homogenously across the different WHO fascicles. Pathologic diagnosis represents the integration of morphologic, immunohistochemical and molecular characteristics and is a key element of clinical decision making. The WHO classification is as a key instrument to promote multidisciplinarity, stimulating pathologists, geneticists and clinicians to join efforts aimed to translate novel pathologic findings into more effective treatments.


The publication of the new WHO classification always generates within the sarcoma community great expectations. Mesenchymal tumours are in fact regarded as one of the most challenging fields of diagnostic pathology and refinement of classification schemes is perceived as the cornerstone around which improving the quality of both pathologic diagnosis and therapeutic options 1. Published data indicate in sarcoma a rate of diagnostic inaccuracy ranging between 20 and 30% 2-4. However, this is not at all due to pathologists’ negligence. As a matter of fact, there exist objective factors that appear to impact negatively over both the accuracy and the reproducibility of pathologic diagnosis, factors the pathologist have always tried to overcome through specific strategies, most of all by implementation of diagnostic second opinion in expert centers or within collaborative networks 5. Four main sources of challenge can be identified.

  • Rarity. Sarcomas as a whole are characterised by an incidence of approximately 5 cases/100,000 thus matching the formal definition of a rare tumor 6. However, soft tissue malignancies are further subclassified in approximately 70 subtypes, each characterized by a distinct morphology, that often translates into a specific clinical behaviour as well as into specific therapeutic approaches. Moreover, many histotypes are exceedingly rare (in the range of 0.1 cases/100,000), to the extent that a pathologist not working in a high-volume centre may not encounter them for years. In this scenario, achieving specific expertise represents undoubtedly a challenge.
  • Intrinsic complexity. Mesenchymal tumours are characterised by specific diagnostic peculiarities. First of all, the mere application of morphologic criteria of malignancy used for epithelial cancer are not always applicable. The best example is represented by nodular fasciitis, a benign condition most often occurring in the forearm of a young adults, who exhibits clinicopathologic features (rapid growth, hypercellularity and high mitotic activity) that in the context of an epithelial cancer would be highly supportive of a diagnosis of malignancy. The relatively frequent violation of diagnostic criteria of malignancy appears to represent a unique feature of mesenchymal tumours and represents one of the major reasons for diagnostic inaccuracy.
  • Technological complexity. The diagnostic process of mesenchymal tumours relies upon a complex combination of conventional microscopic morphology, immunohistochemistry, and molecular genetics 7. This requires state of the art molecular technology that more and more includes Next Generation Sequencing approaches 8. In contrast with popular belief molecular genetics requires high professional expertise coupled with even higher quality control standards (as compared for example to immunohistochemistry). In this perspective, it is quite intuitive that centralisation of molecular diagnostics in high volume centres is mandatory on order to maintain high analytic quality 9.
  • Lower impact of educational efforts. Even if education still plays a fundamental role in increasing diagnostic expertise, it has to be admitted that in the field of rare cancers its efficacy is somewhat hampered. Unless a pathologist is continuously exposed to soft tissue tumours diagnostics, the expertise developed through educational efforts is at risk of being gradually lost.
  • The field of sarcoma oncology is rapidly evolving through the development of an even close correlation between pathologic diagnosis and tailored treatments 10-16. The WHO classification of soft tissue tumours, since 1999 17-19. has introduced a profound change in its methodological approach aimed to support a more rational therapeutic approach. Major changes can be summarised as follows:

  • Integration of morphology with immunohistochemistry and molecular genetics. The marriage between morphology and genetics by direct involvement of cytogeneticists has certainly represented a major step forward.
  • Involvement of a broad number of sarcoma expert pathologists. This approach has minimised the risk of generation of “opinion-leader” bias and has led to a broader diffusion of the classification among pathologists.
  • Precise definition of clinicopathological categories. It is now broadly accepted that in between benign and malignant categories there exists an ”intermediate” category of lesions that can be either locally aggressive (i.e. desmoid fibromatosis) or rarely metastasizing (i.e. plexiform fibrohistiocytic tumour).
  • Involvement of clinicians. For the first time the 2020 WHO classification of soft tissue and bone tumours, representatives of clinical disciplines such as medical, surgical and radiation oncology have been directly involved. This close interaction has strongly enhanced the clinical value of pathological diagnosis.
  • The aim of this paper is to review the main advances contained in the current classification and also discuss new perspectives that most likely will generate some debate in the years to come. In consideration of the complexity of the new classification and despite the fact that some soft sarcomas can rarely arise primarily in bone 20, with the exception of undifferentiated round cell sarcoma that can occur both in soft tissue and bone, we will focus on tumours of the soft tissue.

    Advances in classification


    The entities characterised by adipocytic differentiation are listed in Table I. Newcomers are represented by atypical spindle cell/pleomorphic lipomatous tumour and myxoid pleomorphic liposarcoma. Importantly, both lesions are now regarded as benign and therefore treated with simple surgical excision.

    The label atypical spindle cell lipomatous tumour represents a new name for the entity formerly known as spindle cell liposarcoma 21,22 and at the time of first description regarded as a variant of well differentiated liposarcoma. The entity is now defined as an ill-circumscribed, moderately atypical spindle cell tumour featuring the presence of a variable number of lipoblasts. Matrix can vary from fibrous to myxoid (Fig. 1A). A remarkable tendency to occur superficially is observed. The new label is justified by the distinctively indolent clinical course. Separation from the group of well differentiated liposarcoma/atypical lipomatous tumour is also justified by the fact that atypical spindle cell tumour is not driven genetically by the amplification of MDM2 and/or CDK4 genes. In further contrast dedifferentiation (i.e. transition from low grade to high grade histology) seems not to occur.

    Atypical pleomorphic lipomatous tumour has been recently introduced to recognise the existence of examples of pleomorphic lipoma that, while keeping some key diagnostic features (presence of florette-like multinucleated giant cells as well as of coarse eosinophilic collagen bundles) exhibit higher cellularity, increased mitotic activity and presence of numerous lipoblasts, however without reaching histologic criteria consistent with the diagnosis of pleomorphic liposarcoma 23 (Fig. 1B). Atypical pleomorphic lipomatous tumour, despite alarming morphology, is characterised by a benign clinical behaviour that contrasts with the remarkable aggressiveness of pleomorphic liposarcoma.

    Both spindle and pleomorphic atypical lipomatous tumours occur most often in the limbs and limb girdles of middle-aged patients. Importantly, they can rarely arise at visceral locations including the retroperitoneum. From a molecular standpoint RB1 gene loss is frequently observed that translates into loss of nuclear RB1 immunoreactivity. As mentioned, both lesions tend to exhibit a low rate of recurrence that seems to be limited to lesions excised incompletely.

    Myxoid pleomorphic liposarcoma is an exceedingly rare adipocytic malignancy, and is characterised by occurrence in children and adolescents with female predominance 24. Most common anatomic sites are represented by the mediastinum followed by the limbs and the head and neck region. Morphologically myxoid pleomorphic liposarcoma shows features of both myxoid (presence of a rich capillary size vascular network set in myxoid background) and pleomorphic liposarcoma (presence of pleomorphic lipoblasts (Fig. 1C). No specific genetic aberrations have been so far detected. Clinical behaviour is aggressive with a high recurrence rate and early metastatic spread to the lungs, bone, and soft tissues. It should be remembered that true myxoid liposarcomas almost never occurs in children, those reported case most likely representing examples of lipoblastoma 25,26.


    The entities characterised by fibroblastic/myofibroblastic differentiation are listed in Table II. Despite representing one of the larger groups, only three new benign entities are considered in the new classification: angiofibroma of soft tissues 27, EWSR1-SMAD3 positive fibroblastic tumour 28, and superficial CD34-positive fibroblastic tumour 29.

    Angiofibroma of soft tissue is a benign lesion most often occurring in the subcutaneous soft tissues of the extremities of adult patients. Morphologically it is composed of a uniform spindle cell proliferation set in a variably myxoid and collagenous background, and associated with a remarkably rich thin-walled vascular network (Fig. 2A). Neoplastic cells express CD34 and EMA. Desmin positivity is often observed in dendritic cells. The presence of an AHRR-NCOA2 fusion gene is reported in the majority of cases 30.

    EWSR1-SMAD3 fibroblastic tumour represents one of the new entities in which the name is determined by the involved fusion gene. It is a benign neoplasm most often occurring in the hands and feet with a broad age range. Morphologically it is composed of intersecting cellular fascicles of cytologically bland monomorphic spindle cells, alternating with hypocellular, hyalinised areas (Fig. 2B). Immunohistochemically it is characterized by nuclear expression of ERG.

    Superficial CD34-positive fibroblastic tumour occurs most often skin and subcutis of the lower extremities of middle-aged patients. Morphologically is composed of spindle and epithelioid cell proliferation, often featuring striking cytologic atypia that is associated with insignificant mitotic activity. Neoplastic cells invariably express CD34 (Fig. 2C). Importantly, despite alarming morphologic features no recurrences are reported following complete excision.


    The category of fibrohistiocytic tumours is the group that through the last three editions of the WHO classification has lost most of the tumour entities. Currently recognised lesions are listed in Table III. The very concept of fibrohistiocytic differentiation is rather elusive as in most cases the histiocytic component is actually non-neoplastic. However, as happens in giant cell tumour, the histiocytic component contributes significantly to the development of the lesion. The most relevant entity that has disappeared since 2013 is represented by the family of so called malignant fibrous histiocytoma (MFH), a group of lesions that until the early 2000 accounted for approximately 50% of sarcoma diagnoses 31-33. Undifferentiated pleomorphic sarcoma currently represents the correct label for the prototypical storiform and pleomorphic variant of MFH. Giant cell MFH is currently replaced by three distinct tumour types: giant cell tumour of soft tissues, extraskeletal osteosarcoma and giant cell rich osteosarcoma. Myxoid MFH is currently recognised as a purely fibroblastic tumour identified with the original name myxofibrosarcoma 34-36. So called inflammatory MFH overlaps entirely with the inflammatory variant of dedifferentiated liposarcomas 37, and so called angiomatoid MFH (a clinically indolent lesion most often harbouring a EWSR1-CREB1 fusion gene and, more rarely a EWSR1-ATF1 or a FUS-ATF1) is currently listed within the group of soft tissue lesion of unknown differentiation 38.


    The group of neoplasms featuring endothelial differentiation is also very heterogenous both in terms of degree of vasoformative morphology and biological behaviour (Tab. IV). It is worth stressing that despite the label implying an intermediate behaviour, epithelioid haemangioendothelioma (EHE) is actually ranked among vascular malignancies. This is the consequence of an overall metastatic rate of approximately 15%, and the tendency to be remarkably aggressive in specific anatomic locations such as lungs and pleura. Two morphological and molecular variants (CAMTA1 and TFE3-related) are recognised however at the moment no statistically significant differences in terms of clinical behaviour seems to emerge 39.

    The single new entity appearing among vascular lesions is represented by a benign lesion named anastomosing haemangioma (AH). Initially reported in the male genital tract 40,41, anstamosing haemangioma most commonly occurs in the kidney and in the retroperitoneum of adults. Morphologically AH is composed of anastomosing capillary-sized vessels featuring scattered hobnail endothelial cells (Fig. 3). Anastomosing haemangioma is frequently misinterpreted as angiosarcoma, however if present endothelial atypia is mild, and is never associated with multilayering as is typically observed in angiosarcoma.


    This group of lesions share morphologically the presence of a distinctive perivascular pattern of growth (Tab. V). Cell shape varies form epithelioid (such as in glomus tumour) to spindle as observed in myopericytoma/myofibroma and angioleiomyoma. It seems likely the whole group represents a spectrum of perivascular neoplasia exhibiting a variable contractile phenotype. Most of the lesions tend to exhibit a benign clinical course. The diffuse variants (glomangiomatosis and myofibromatosis) however are associated with significant local morbidity. Malignancy represents an extremely rare condition.

    Since 2013 the major advance is represented by the abolition of haemangiopericytoma 42. The original intuition of Arthur Purdy Stout represented by the recognition of neoplasms composed of perivascular contractile cells, remains absolutely valid, and is currently encompassed by myopericytoma. Unfortunately, through the decades hemangiopericytoma became the preferred label for unrelated lesions sharing an hemangiopericytic vascular pattern, namely the presence of dilated, branching, thin walled blood vessels. Most haemangiopericytomas are now recognized as examples of solitary fibrous tumours, however the list of lesions featuring haemangiopericytomas-like vessels is broad and includes among others myopericytoma/myofibroma, synovial sarcoma, malignant peripheral nerve sheath tumours, endometrial stroma sarcoma, and mesenchymal chondrosarcoma. Solitary fibrous tumour still represents a diagnostic challenge because as mentioned above can mimic many mesenchymal and non-mesenchymal unrelated tumour entities 43. The availability of STAT6 immunostaining has certainly contributed to improve diagnostic accuracy 44 and, most importantly, for the first time the prediction of the outcome of is not simply determined by mitotic count 45, but by a risk assessment scheme that considers, patient’s age, tumour size, depth of location, and mitotic index 46.


    The changes introduced in the category of soft tissue tumours featuring smooth muscle differentiation are represented by the inclusion as distinct entities of both EBV-associated smooth muscle tumours, and inflammatory leiomyosarcoma (Tab. VI).

    EBV-associated smooth muscle tumours are intimately associated with EBV infection most often in the context of immunosuppression 47. Age range is broad as anatomic location. HIV-associated lesions however tend to occur most often in the central nervous system whereas post-transplant proliferations feature a remarkable tropism for the liver, followed by the lungs and the gastrointestinal tract. Prognosis tends to be more related to the evolution of the state of immunosuppression

    Most cases of inflammatory leiomyosarcoma occur in the deep soft tissues of the lower limbs, trunk and retroperitoneum of adult patients, with a peak incidence between the third and the fourth decade 48. Microscopically smooth muscle cells are most often low grade and associated with a lymphoplasmacytic infiltrates that tend to be so prominent to overshadow the neoplastic component. Less often a histiocytic population predominates sometimes assuming a xanthomatous appearance. Inflammatory leiomyosarcoma seems to behave less aggressively than ordinary leiomyosarcoma however follow-up data are still very limited.


    No major changes are reported within this subgroup (Tab. VII). A notable exception is however the recognition that rhabdomyosarcoma can occur primarily in bone as a spindle cell and epithelioid cell neoplasm (Fig. 4) that is most often associated with fusion of the TFCP2 gene with FUS or EWSR1 and expression of ALK (in absence of ALK gene rearrangement) 49-51. Recently an alternative MEIS1-NCOA2 gene fusion has been reported 52. With the bias determined by their extreme rarity, the clinical behaviour appears to be very aggressive.


    One major step forward is that in the 5th series GIST is covered by the same authors in both the gastrointestinal and soft tissue fascicles, therefore ensuring a more homogenous diagnostic approach (Tab. VIII). In comparison with previous editions SDH-deficient GISTs are now discussed in greater detail. This distinct subset (accounting for approximately 5 to 10% of all GIST) is characterised by occurrence in children and adolescents, with female predominance 53. Relatively often SDH-deficient GISTs feature a distinctive multinodular pattern of growth associated with predominantly epithelioid morphology (Fig. 5A). The presence of SDH subunit genes mutations is observed in approximately 60% of cases (SDHA mutations are the most frequently observed) and, whatever the subunit involved is predicted by immunohistochemical loss of SDHB expression 54 (Fig. 5B). In 40% of cases SDHC promoter methylation (epimutation) is detected. Importantly SDH-deficient GIST appears to be resistant to tyrosine kinase inhibitors. In this context the risk assessment schemes utilised for KIT/PDGFRA mutated GIST is not applicable.


    As exemplified in Table IX no new entries have been included. Both soft tissue chondroma and extraskeletal osteosarcoma represent exceedingly rare lesions. In approximately 50% of soft tissue chondroma FN1 gene rearrangements have been documented. Extraskeletal osteosarcoma remains an extremely aggressive neoplasms of elderly patients, often presenting with lung metastases at diagnosis.


    The single major change introduced by the 2020 WHO classification is the recognition that so called melanotic schwannoma actually represents a clinically aggressive neoplasm (not belonging anymore to the intermediate category) being consequently relabelled as malignant melanotic nerve sheath tumour 55 (Tab. X). It most often occurs in a midline location within spinal and autonomic nerves of adult patients. A subset of lesions occurs in context of Carney complex. Microscopically is composed by spindle and epithelioid neoplastic cells organised in sheets and short fascicles. Cytoplasm vary from eosinophilic to amphophilic and contain round nuclei often featuring pseudoinclusions. The amount of pigment varies from focal to massively abundant (Fig. 6). Approximately half of cases contain psammoma bodies. Presence of necrosis and mitotic activity can be observed that however does not seem to predict outcome. Immunohisochemically, malignant melanotic nerve sheath tumour express S100, SOX10, HMB45 and MelanA. Pathogenetically, the vast majority of tumour are associated with inactivating mutation of the PRKAR1A gene leading to the loss of expression of the protein thereof 56. It has become evident that this lesion is associated with high risk of both local recurrence and metastatic spread, even many years from resection of the primary lesion. As a consequence, long term follow-up is strongly advised.


    The inclusion within this tumour category (Tab. XI) as an emerging entity of the category of NTRK-rearranged spindle cell neoplasm (excluding infantile fibrosarcoma that represent a distinct clinicopathologic entity defined molecularly by the presence of NTRK3-ETV6 fusion gene), represents one of the most relevant innovation. The availability of a new class of drugs characterised by NTRK inhibitory action (Entrectinib and Larotrectinib) has in fact generated a remarkable interest towards any neoplasm harbouring a rearrangement of any of NTRK 1, 2 and 3 genes 57.

    NTRK-rearranged spindle cell neoplasms appear to form a morphological spectrum that occurs most often in the superficial or deep soft tissue of the extremities of children and adolescents. The molecular pathogenesis is most often related to rearrangement of the NTRK1 gene with a variety of partners. More rarely NTRK2 and NTRK3 aberrations have been detected. At one end of this spectrum is the so-called lipofibromatosis-like neural tumour (LLNT), composed of monomorphic spindle cells co-expressing S100 and CD34, featuring a highly infiltrative pattern within subcutaneous fat, therefore closely resembling lipofibromatosis (Fig. 7A). From the clinical standpoint LLNT can recur locally but seems not to possess metastatic potential 58.

    A second subset of cases features a predominantly solid pattern of growth, and is composed of a cellular proliferation of uniform spindle cells that in higher grade examples resemble malignant peripheral nerve sheath tumours and therefore labelled as NTRK- rearranged neoplasm resembling peripheral nerve sheath tumours (Fig. 7B). Prominent deposition of hyalinised collagen is often observed 59,60. Rarely, a myopericytoma-like architecture is detectable 61. Outcome seems to correlate with morphology. Lesions with high grade features tend to behave aggressively often featuring distant spread to the lungs. All these entities share the expression (that is not of course specific) of anti- pan-TRK antibodies. Hopefully, the accrual of more cases will allow in the future to better define the clinicopathologic boundaries of this group of lesions.


    The creation of a separate chapter encompassing round cell sarcoma of soft tissue and bone also represents a major step forward of the 2020 WHO classification (Tab. XII). This new section contains not only the prototypical round cell sarcoma named Ewing’s sarcoma, but also three distinct subsets that differs from Ewing’s sarcoma clinically, pathologically and molecularly 49,62: 1. Round cell sarcomas with EWSR1 gene fusion with non-ETS family members 63-65, 2. CIC-rearranged sarcomas 66-68, and 3. BCOR-rearranged sarcomas 69. Interestingly, despite significant morphologic overlap, most of these entities tend to exhibit some morphologic features predictive of the underlying molecular alteration. If Ewing sarcoma represents the prototype of round cell sarcoma, CIC sarcomas always exhibits focal pleomorphism, and at times epithelioid morphology can predominate (Fig. 8A). BCOR sarcomas (despite being allocated to the round cell sarcoma family) more often tend to exhibit a spindled morphology (Fig. 8B). NFATC2 sarcoma may exhibit remarkable epithelioid features (Fig. 8C), and PATZ1 sarcomas are composed of a rather undifferentiated round to ovoid cell population (Fig. 8D), often characterised by the presence of a sclerotic background. The differential diagnosis for these tumours is rather broad, and among round cell sarcomas includes alveolar rhabdomyosarcoma, desmoplastic small round cell tumour, poorly differentiated round cell synovial sarcoma, small cell osteosarcoma, and mesenchymal chondrosarcoma. A combination of morphologic, immunohistochemical, and molecular findings allows accurate classification in most cases. A granular diagnostic approach to Ewing sarcoma and Ewing-like sarcomas is justified by significant differences in terms of both response to chemotherapy and overall survival 68,69. As all these entities are in part defined by specific fusion genes, a molecular diagnostic approach based on NGS technology should be preferred 8. In consideration of the extreme rarity of many of these tumour entities, referral to expert rare cancer centres or to rare cancer networks represents the best strategy in order to minimise diagnostic inaccuracy, and allow proper patient management.

    Standing Issues


    The integration of morphology and genetics represents one of the major advances introduce by WHO classification since 2000, to the extent that the “blue Book” was labelled as “Pathology and Genetics” 17. In the last 20 years descriptions of genetic aberrations (particularly new fusion genes) have increased exponentially. Many of these genetic abnormalities are linked with specific histotypes (i.e. MDM2 gene amplification and well differentiated/liposarcoma; SYT gene rearrangement in synovial sarcoma etc.) whereas many others (particularly fusion genes) seem to represent non-driver (stochastic) molecular events the detection of which has been enhanced by extensive use of NGS-based molecular approaches 70. Moreover, some histotypes are currently defined on the basis of their genetics (i.e. NTRK, CIC and BCOR sarcomas). Whether molecular analysis should be mandatorily associated with conventional microscopic examination is still source of debate. The WHO expert panel has decided to report in the “blue book” all pertinent genetic aberrations not making them diagnostically obligational. Two main reasons are: 1. Accurate pathologic diagnosis can be reached (in expert hands) without the support of molecular diagnostics. 2. WHO classification is meant to be used globally, including those countries in which molecular pathology is not implemented. Both arguments are of course valid. There exists however one contradiction represented by the introduction of entities labelled based on “specific” fusion genes. In our opinion, it sounds for instance somewhat illogical to make a diagnosis of EWSR1-SMAD3 positive fibroblastic tumour in absence of a formal demonstration of the associated genetic aberration.


    An unavoidable consequence of using molecular genetics as the main driver for tumour classification is to generate a debate regarding the principles behind the definition of a tumour entity. Moreover, the advent of “precision oncology” has generated the (mis) concept that extensive molecular profiling of cancer is the only way to improve patients’ outcome 71. The underlying ambition would be the generation of a new taxonomy of human cancer based on molecular genetics, aimed to replace histology-based classifications. Certainly, molecular characterisation can allow to predict sensitivity to treatments, especially molecularly targeted therapies. In the sarcoma field, this has proved extremely effective with tyrosine kinase inhibition in GIST. The principle of precision oncology is however based on the concept that medical therapies should focus on the predictive molecular biomarker rather than on the histoype 72,73. This “histological agnosticism” has led for the first time to the approval of anticancer agents based solely on the availability of a molecular target. The prototypical example is in fact represented by neurotrophic tyrosine receptor kinase (NTRK) inhibitors, the concept being that a NTRK-inhibitor may be effective against different malignancies as long as they harbour a NTRK gene fusion 57. As discussed previously, the 2020 WHO classification has introduced an “emerging entity” named NTRK-rearranged spindle cell neoplasm 19. This seems apparently a timely decision however it should be considered that several factors will necessarily influence the actual use of these new agents:

  • The clinical positioning of the drug that may vary based on the natural history of the cancer type.
  • The actual need for a medical therapy when the stage of the disease is considered.
  • The availability of alternative effective agents.
  • An NTRK-inhibitor will be used for instance as a last-line therapy in one malignancy and as a first-line therapy in another one. In other words, the regulatory approval of a drug may be histologically agnostic, but its clinical use will never be. Irrespective of its predictive power, a molecular biomarker aimed to select medical therapies should never become the key element to define a disease. In the case of NTRK rearranged spindle cell sarcomas one may suppose that these tumours are sensitive to the new NTRK-inhibitors, however their epidemiology, clinical presentation, natural history and prognosis seems to be extremely variable and somewhat unspecific. If the only common denominator, is the potential activity of a class of drugs, actually many other unrelated non-mesenchymal cancers will be sensitive to the same drugs. On the contrary, a time-honoured entity such as infantile fibrosarcoma, in addition to be pathogenetically related to a gene fusion involving NTRK3, clearly features epidemiological, clinical and pathologic characteristics which are highly specific. Its sensitivity to NTRK-inhibitors is just a complement to all that. Thus, while there is no reason to challenge the existence of a disease called infantile fibrosarcoma, one could for instance consider premature (this may change in the future as more data are collected) to consider NTRK-rearranged spindle cell neoplasm as a distinct nosological entity, and this may in principle apply to other “new” sarcoma entities. For these reasons, the concept that a new entity in the sarcoma field should never reflect just the mere presence of a predictive molecular biomarker has been recently proposed to the sarcoma community 74.


    The publication of a new WHO classification represents a major step towards standardization of cancer diagnostics. In the case of rare diseases such as soft tissue tumours its value is even greater as it contributes significantly to improve diagnostic accuracy. Pathologic diagnosis of cancer represents the result of a complex integration of microscopic, immunophenotypic and molecular features and, with rare exceptions is the cornerstone of clinical decision making. WHO classification appears more and more as the field wherein the joint efforts of pathologists, geneticists and clinicians can translate novel findings into more rationale as well as more effective treatments.

    Figures and tables

    Figure 1A..

    Figure 1B..

    Figure 1C..

    Figure 2A..

    Figure 2B..

    Figure 2C..

    Figure 3..

    Figure 4..

    Figure 5A..

    Figure 5B..

    Figure 6..

    Figure 7A..

    Figure 7B..

    Figure 8A..

    Figure 8B..

    Figure 8C..

    Figure 8D..

    Table I..

    Lipoma and lipomatosis
    Lipomatosis of nerve
    Lipoblastoma and lipoblastomatosis
    Myolipoma of soft parts
    Chondroid lipoma
    Spindle cell/pleomorphic lipoma
    Atypical spindle cell/pleomorphic atypical lipomatous tumor
    Intermediate (locally aggressive)
    Atypical lipomatous tumor
    Malignant adipocytic tumours
    Well differentiated liposarcoma: lipoma-like, sclerosing, inflammatory
    Dedifferentiated liposarcoma
    Myxoid liposarcoma
    Pleomorphic liposarcoma
    Myxoid pleomorphic liposarcoma
    Adipocytic tumours.

    Table II..

    Nodular fasciitis
    Proliferative fasciitis and proliferative myositis
    Myositis ossificans and fibro-osseous pseudotumor of digits
    Ischaemic fasciitis
    Fibrous hamartoma of infancy
    Fibromatosis colli
    Juvenile hyaline fibromatosis
    Inclusion body fibromatosis
    Fibroma of tendon sheath
    Desmoplastic fibroblastoma
    Mammary-type myofibroblastoma
    Calcifying aponeurotic fibroma
    EWSR1-SMAD3-positive fibroblastic tumour (emerging)
    Cellular angiofibroma
    Angiofibroma NOS
    Nuchal fibroma
    Acral fibromyxoma
    Gardner fibroma
    Intermediate (locally aggressive)
    Palmar/plantar-type fibromatosis
    Desmoid-type fibromatosis
    Giant cell fibroblastoma
    Dermatofibrosarcoma protuberans
    Intermediate (rarely metastasising)
    Dermatofibrosarcoma protuberans, fibrosarcomatous
    Solitary fibrous tumour
    Inflammatory myofibroblastic tumour
    Low-grade myofibroblastic sarcoma
    Superficial CD34-positive fibroblastic tumour
    Myxoinflammatory fibroblastic sarcoma
    Infantile fibrosarcoma
    Solitary fibrous tumour, malignant
    Fibrosarcoma NOS
    Low grade fibromyxoid sarcoma
    Sclerosing epithelioid fibrosarcoma
    Fibroblastic/myofibroblastic tumours.

    Table III..

    Tenosynovial giant cell tumour
    Deep benign fibrous histiocytoma
    Intermediate (rarely metastasising)
    Plexiform fibrohistiocytic tumour
    Giant cell tumour of soft parts NOS
    Malignant tenosynovial giant cell tumour
    So-called fibrohistiocytic tumours.

    Table IV..

    Synovial haemangioma
    Intramuscular haemangioma
    Arteriovenous malformation/haemangioma
    Venous haemangioma
    Anastomosing haemangioma
    Epithelioid haemangioma
    Lymphangioma and lymphangiomatosis
    Acquired tufted haemangioma
    Intermediate (locally aggressive)
    Kaposiform haemangioendothelioma
    Retiform haemangioendothelioma
    Papillary intralymphatic angioendothelioma
    Composite haemangioendothelioma
    Kaposi sarcoma
    Pseudomyogenic haemangioendothelioma
    Epithelioid haemangioendothelioma
    Vascular tumours.

    Table V..

    Benign and intermediate
    Glomus tumour NOS
    Myopericytoma, including myofibroma
    Glomus tumour, malignant
    Pericytic (perivascular) tumours.

    Table VI..

    Smooth muscle tumour of uncertain malignant potential
    EBV-associated smooth muscle tumour
    Inflammatory leiomyosarcoma
    Smooth muscle tumours.

    Table VII..

    Embryonal rhabdomyosarcoma
    Alveolar rhabdomyosarcoma
    Pleomorphic rhabdomyosarcoma
    Spindle cell / sclerosing rhabdomyosarcoma
    Skeletal muscle tumours.

    Table VIII..

    Gastrointestinal stromal tumors
    Gastrointestinal stromal tumours.

    Table IX..

    Osteosarcoma, extraskeletal
    Chondro-osseous tumours.

    Table X..

    Granular cell tumour
    Nerve sheath myxoma
    Solitary circumscribed neuroma
    Hybrid nerve sheath tumour
    Malignant peripheral nerve sheath tumour
    Melanotic malignant nerve sheath tumour
    Granular cell tumour, malignant
    Perineurioma, malignant
    Peripheral nerve sheath tumours.

    Table XI..

    Myxoma (cellular myxoma)
    Deep (aggressive) angiomyxoma
    Pleomorphic hyalinising angiectatic tumour
    Phosphaturic mesenchymal tumour
    Perivascular epithelioid tumour, benign
    Intermediate (locally aggressive)
    Haemosiderotic fibrolipomatous tumour
    Angiomyolipoma, epithelioid
    Intermediate (rarely metastasising)
    Atypical fibroxanthoma
    Angiomatoid fibrous histiocytoma
    Ossifying fibromyxoid tumour
    Phosphaturic mesenchymal tumour, malignant
    NTRK-rearranged spindle cell neoplasm (emerging)
    Synovial sarcoma
    Epithelioid sarcoma: proximal and classic variant
    Alveolar soft part sarcoma
    Clear cell sarcoma
    Extraskeletal myxoid chondrosarcoma
    Desmoplastic small round cell tumour
    Rhabdoid tumour
    Perivascular epithelioid tumour, malignant
    Intimal sarcoma
    Ossifying fibromyxoid tumour, malignant
    Myoepithelial carcinoma
    Undifferentiated sarcoma
    Spindle cell sarcoma, undifferentiated
    Pleomorphic sarcoma, undifferentiated
    Round cell sarcoma, undifferentiated
    Tumors of uncertain differentiation.

    Table XII..

    Ewing sarcoma
    Round cell sarcoma with EWSR1-non-ETS fusions
    CIC-rearranged sarcomas
    Sarcoma with BCOR genetic alterations
    Undifferentiated small round cell sarcomas of bone and soft tissue.