4. Diagnostics, prognostic markers and scores – Soft Tissue Sarcoma

Background

When a malignant soft tissue tumor is suspected, contrast-enhanced magnetic resonance imaging (MRI) is favored as the primary imaging modality because it provides the best visualization of tumor margins and differentiation of fatty tumors as well as cystic and necrotic and non-necrotic tumor components.

Soft tissue tumors often have a heterogeneous structure, with the most aggressive tumor portion usually being the least differentiated (i.e., having the highest grading) and therefore determining the biological behavior of the tumor. These are often regions of highest tumor cell density, highest vascularization, and therefore highest glucose uptake. Imaging prior to biopsy should primarily be performed with MRI including intravenous (i.v.) contrast and diffusion weighting to optimise assessment of tumor margins. If there are contraindications to i.v. contrast, i.v. (and if necessary, enteral) contrast-enhanced CT or PET-CT with i.v. contrast should be performed as an alternative.

In MRI, the axial plane is the most important for determining the tumor boundary. The MRI protocol should include at least one T1w sequence (SE or TSE) in a long axis (coronary or sagittal), two T2w sequences in axial and long axis planes, and two T1w sequences after i.v. contrast agent in the axial slice and a long axis plane. A T1w sequence without fat saturation should be measured before and after contrast administration with identical acquisition parameters to detect “true“ contrast uptake. The spatial resolution should be approximately 0.5 x 0.5 mm in the slice direction, and the slice thickness should be 3 to 5 mm. An axial T2 sequence without fat saturation is obligatory. A DIXON sequence is acceptable as an alternative if the spatial resolution is appropriate. 

In addition to the above sequences, a diffusion-weighted sequence should be performed. This is useful for determining diagnosis, highest cell density to guide biopsy, and differentiation from hematoma.

A conventional radiograph or native CT can be performed to detect calcification, bony involvement, or to estimate fracture risk. The calcific component of a soft tissue tumor may be underestimated or undetectable by MRI.

In retroperitoneal tumors, CT with i.v. contrast may be equal to MRI for local staging and can be used as an alternative. CT can distinguish between well-differentiated and poorly differentiated sarcomas relatively well [81]. However, MRI is more suitable for delineating sarcomas from neurovascular bundles and for identification of marrow infiltration [82], [83], [84].

Background

Early staging of soft tissue tumors suspected of malignancy can in some cases facilitate interdisciplinary diagnosis and contribute to assessing whether a malignant lesion is present when metastatic disease has already been detected. This allows prompt treatment planning and the initiation of further diagnostic/molecular investigations (e.g. mutation analysis in metastatic gastrointestinal stromal tumor).

Background

Due to the increasing demands on the number and complexity of diagnostic techniques to be performed on a biopsy sample, the regular use of a fine needle biopsy for initial diagnosis is discouraged. Fine needle biopsy also limits reference pathological second opinions due to limited material. Reliable grading of sarcomas is usually not possible on fine needle biopsy, especially in highly differentiated lesions or spindle cell sarcomas with low proliferative activity [88]. Although fine needle biopsy is more successful for the diagnosis of high-grade sarcomas, „accuracy“ remains limited to 91 % [89]. If neoadjuvant therapy is planned, an exclusively fine-needle biopsy diagnosis is not recommended. By using higher caliber core biopsies, significantly more diagnostic material can be obtained, which is then also available for immunohistochemical and molecular pathological investigations.

Although fine-needle aspiration biopsy may have significant limitations in definitive histologic assignment, it is capable of separating malignant from benign lesions with high sensitivity and specificity [90]. Fine needle aspiration biopsy can therefore be considered for histological confirmation of local recurrence or metastases of previously histologically confirmed soft tissue sarcomas.

Background

The access route should be determined by an interdisciplinary discussion prior to biopsy.

Core needle biopsy is now a frequently recommended procedure in guidelines for obtaining samples from soft tissue tumors. Studies have shown that this technique is comparable [91] or at most slightly inferior to open biopsy in terms of diagnostic accuracy in experienced hands [92]. Percutaneous core biopsies require a smaller access route than open biopsies, which facilitates resection of the puncture tract during subsequent definitive surgery [93], [94], [95]. Multiple core biopsies are typically performed through one access point. This is relevant because tumor cell metastases can be found in the biopsy tract during both open biopsy and percutaneous biopsy. It has been shown that cell metastases are significantly more likely to be present in the biopsy tract after open biopsy than after percutaneous biopsy [96]. As a consequence, core biopsy in combination with optimal imaging is favoured to guide diagnosis [97], [98]. A systematic review with meta-analysis proved that core biopsy has a high accuracy for soft tissue tumor diagnosis and subtyping. The lower complication rates associated with core biopsies also contributes to its use as the first-line biopsy technique [99].

For an adequate histological diagnosis, biopsy specimens must meet minimum standards. Smaller lesions often pose a technical challenge for biopsy. Core biopsy is applicable for lesions from about 1 cm in diameter, but for lesions < 2 cm diagnostic accuracy decrease [100]. Another important influential factor is the length and, in particular, the number of specimens. It has been shown that a plateau in diagnostic accuracy is reached for soft tissue tumors when the number of specimens exceeds 4 [101]. Existing recommendations suggest that needles of at least 16G diameter should be used [85].

Background

Incisional biopsy usually yields the largest diagnostically useful volume of tissue and thus the highest diagnostic certainty. The incision should be chosen in such a way that no healthy structures, especially nerves and vessels, are contaminated. The access must be completely removed at subsequent resection. The incision should be made as longitudinally as possible. A Drainage through the incision or led out directly in extension of the incision together with a dressing should minimise hematoma formation.

Excisional biopsy as an initial diagnostic measure may be considered for superficial, well circumscribed lesions less than 3 cm in diameter, provided that safe removal of the tumor in healthy tissue is assured [86]. British surgeons analyzed data from 3018 patients with primary benign (52 %) and malignant soft tissue tumors in a hierarchical analysis, finding the thresh-old between benign/malignant in the size region between 3 and 5cm. An analysis according to the Bayesain model indicated the practically relevant tumor size, above which a malignancy could be assumed with a high probability [102]. Consequently, a golf ball was sent to all British surgeons with the text „If your lump is bigger than a golf ball and growing, think Sarcoma“.

Background

Frozen section examinations are not suitable for diagnosis, subtyping or grading of soft tissue tumors. In individual cases, frozen section examination may help in the assessment of narrow resection margins. This does not apply to differentiated lipomatous tumors, however.

Background

If possible, core biopsies should be fixed in sufficient 4% buffered formalin for at least 6 hours before pathological processing. Non fixed samples are not recommended as material dries out quickly. Dispatch in NaCl is strongly discouraged, as this has a very unfavorable effect on the morphology of the tissue. Substitution of formalin by alcohol is also not indicated, as this hardens tissue and makes further processing considerably more difficult. Special fixatives such as Bouin’s solution are strongly discouraged, as this interferes with DNA extraction for any additional molecular pathology studies that may be required. Electron microscopy now plays only a very minor role in sarcoma diagnostics, so that preservation of tumor tissue in appropriate special fixative is generally not envisaged. Preservation of tumor tissue for later RNA extraction (e.g., RNA-later) is usually reserved for research studies.

Background

The specimen received in the pathology department should - if sent na-tive/unfixed - be immediately or - if fixed - promptly inspected by a qualified pathologist and photographically documented. Before collecting native specimens, the specimen should be measured and weighed. The site of collection of native specimens should be marked appropriately so as not to compromise subsequent evaluation of resection margins. In the case of larger specimens, these should be incised several times before further detailed processing to ensure adequate penetration of the formalin into the tissue. After sufficient fixation (usually overnight), macroscopic workup should be performed according to EORTC recommendations [103].

This includes, after appropriate color marking of the various resection margins, lamination into approximately 1 cm wide slices and photo documentation of these lamellae. The corresponding photos can be used to mark where samples were taken. The rule of thumb is to collect one kerosene block per cm of tumor diameter plus separately collect tumor samples related to the resection margins. The extent of necrosis and infarct zones should be described macroscopically and quantified approximately in %. Particular attention should be paid to the minimum distances to relevant resection margins and critical anatomical as well as clinically separately identified structures. Resection margins close to the tu-mor are stained to facilitate assessment of tumor infiltration. R0 resection is defined as no tumor manifestation on the stained resection margin („no tumor on ink“) [104], [105]. If neoadjuvant therapy has preceded, workup should be performed according to the suggestions of the EORTC-STBSG guidance [103].

Background

The classification of soft tissue sarcomas is based on the recognition of specific growth patterns (e.g., spindle cell, small-blue round cell, biphasic), the vascular pattern, and, if applicable, the accompanying inflammatory infiltrate, and on the detection of a lineage-specific differentiation (e.g., lipomatous, muscular, neuronal, etc.).e.g. lipomatous, muscular, neuronal, etc.), or the detection of a specific genotype, so that immunohistochemical and, if necessary, molecular investigations play a key role in the diagnosis of soft tissue sarcomas [106]. In addition, sarcomas must be differentiated from other tumors with similar morphology, such as sarcomatoid carcinomas, malignant melanomas, mesotheliomas, germ cell tumors, or even hematologic neoplasms. Many of the antibodies used (e.g., CD34, S100) are not entity-specific and must be interpreted and weighted differently for diagnosis depending on the context. For these reasons, extensive immunohistochemical analyses (so-called antibody panels) are often necessary in primary diagnostics and are now preferably used in automated systems.

For the basic diagnosis of soft tissue tumors, the following antibodies, among others, have proven successful: Broad spectrum keratins, EMA, Ki67, S100, CD45/PanLeu, CD21, CD23, SOX10, MelanA, HMB45, ER/PR, CD10, smooth muscle actin, desmin, caldesmon or calponin, myogenin, WT1, calretinin, BAP1, podoplanin/D2-40, CD34, ERG, CD117, DOG1, NSE, synaptophysin, chromogranin, CD56, CD99.

A number of relatively specific markers exist for advanced immunohistochemical diagnosis, some of which can be used to screen for underlying genetic alterations. This ever-expanding group includes beta-catenin (e.g. Desmoid fibromatosis), MDM2/CDK4 (highly differentiated/de-differentiated liposarcoma), BRAF V600E (malignant melanoma), TLE1 (synovial sarcoma), STAT6 (solitary fibrous tumor, SFT) [104], MUC4 (low grade fibromyxoid sarcoma, sclerosing epithelioid fibrosarcoma) [107], TFE3 (alveolar soft tissue sarcoma), SDHB (SDH-mutated gastrointestinal stromal tumors, GIST), HHV8 (Kaposi's sarcoma), c-MYC (radiation-induced sarcomas, especially angiosarcomas), CAMTA1 (epithelioid hemangioendothelioma), FOSB (pseudomyogenic hemangioendothelioma, epithelioid hemangioma), INI-1 (malignant rhabdoid tumor, epithelioid sarcoma, epithelioid malignant nerve sheath tumor), H3K27me (malignant peripheral nerve sheath tumor) [108], brachyury (chordoma), ALK (inflammatory myofibroblastic tumor), NKX2.2 (Ewing sarcoma), WT1/ETV4 (CIC-rearranged small cell sarcoma), BCOR (BCOR-rearranged small cell sarcoma) [107], [109]. If preceded by neoadjuvant therapy, the workup should be performed according to the suggestions of the EORTC-STBSG [103].

Background

Tumor grading has less prognostic value than histopathologic subtype for some tumor entities (including angiosarcomas, dedifferentiated and myxoid round cell liposarcomas, rhabdomyosarcomas, Ewing sarcomas, alveolar soft tissue sarcomas, epithelioid sarcomas, clear cell sarcomas) and should not be used for tumors in the intermediate dignity group (eg.E.g., dermatofibrosarcoma protuberans, solitary fibrous tumor, inflammatory myofibroblastic tumor, and others) should not be used [110], [111], [112]. For the diagnostic groups „intermediate – locally aggressive“ and „intermediate – rarely metastatic“, these terms should instead be explained in the assessment in terms of their biological meaning. Thus, if grading is not performed, the reason should be clearly explained. TNM classification is also generally not used for intermediate diagnostic groups with the exception of gastrointestinal stromal tumors).

For some tumor entities, alternative classifications for risk assessment exist instead, such as the AFIP classification by Miettinen et al. 2006 [115] or the nomograms by Joensuu et al. [116] (see Figure 3) for gastrointestinal stromal tumors (see Chapter 10.1). For PEComes, prognostic factors have been proposed by Folpe et al. and by Doyle et al. including parameters such as mitotic count, necrosis, infiltrative tumor growth, and tumor size [117], [118].

It is noted that a reported grading on core biopsy material may not be representative and may be too low, as necrosis and higher grade tumor areas may not be included. This is especially true for entities with frequent heterogeneity such as lipomatous tumors and myxofibrosarcomas. There are usually insufficient fields of view for a representative count of mitoses. The extent of necrosis cannot be adequately estimated. Imaging, which is certainly helpful in this context, is mostly not available to the pathologist (even though this should be urgently sought). The extent to which consideration of the results of radiological imaging can influence and improve grading in the future has not yet been investigated in larger prospective case series.

Background

The number of identified molecular pathologic aberrations is steadily increasing in soft tissue tumors [119], [120], [121], [122], [123]. These are chromosomal translocations, amplifications or specific mutations. Next-generation sequencing (NGS) technology has identified recurrent gene fusions or other genetic alterations for numerous entities in recent years. The detection of a recurrent genetic alteration can be used diagnostically, and in some cases has predictive significance [122], [123], [124]. In tumors with reciprocal chromosomal translocation or gene fusion, in the vast majority of entities, only one specific gene fusion is present that is characteristic of that entity. Rarely, the same gene fusions may be detectable in different tumor entities; more common is the involvement of one gene (e.g., EWSR1) in different gene fusions in different entities. For an overview of gene fusions, see the Mitelman database (https://mitelmandatabase.isb-cgc.org/) [125] and the COSMIC database (https://cancer.sanger.ac.uk/cosmic/fusion) [126]. The most common translocation types and genetic aberrations in soft tissue tumors are listed in Table „Selected diagnostically relevant recurrent gene fusions and driver alterations“ [119] . For each genetic aberration, different detection options are indicated (e.g., immunohistochemistry, RT-PCR, FISH, NGS).

In non-pleomorphic sarcomas, there are clustered translocations whose detection may have therapeutic consequences (e.g., NTRK). NTRK gene fusions occur in some tumor subtypes e.g. in inflammatory myofibroblastic tumor (IMT) in max. 1 %. This alteration, which can be used very well therapeutically, should be searched for in a targeted manner according to defined algorithms [127], [128].

If diagnostic evidence for NTRK fusions is secured, approved agents are available for metastatic/non-operable stage patients with larotrectinib and entrectinib.

In non-pleomorphic sarcomas, translocations are frequent and their detection may have therapeutic consequences (NTRK). NTRK gene fusions occur in some tumor subtypes, e.g. in inflammatory myofibroblastic tumor (IMT) in max. 1 %.

Background

There are several methods to detect specific gene fusions and driver alterations, which have different sensitivity and specificity [129], [130].

Hybridization-based detection (fluorescence in situ hybridization, FISH):

In FISH, detection of a translocation or gene amplification is performed by using gene-specific probes at the DNA level. Depending on the probe combination used, one or both genes involved in a translocation can be detected by split signals (e.g. EWSR1 break-apart probe) or fusion signals (e.g. COL1A1-PDGFB fusion probe). Focal amplification of a gene locus (e.g. MDM2) is detected by using a gene locus-specific probe in combination with a centromere control probe.

Advantages of FISH testing include ease of performance, low cost, high sensitivity and specificity for the corresponding break event, and a short time to result (approximately 2 working days). Disadvantages are that usually only one translocation partner is examined (e.g., EWSR1), which may lead to misinterpretation due to the involvement of this translocation partner in different, entity-specific translocations, and the lack of breadth of the procedure. Unknown or unanticipated translocations are not indicated. False-positive results, due to coexisting genetic defects (such as false-positive EWSR1-FISH for INI1–deletions, etc.), may also occur.

Sequencing-based detection (RT-PCR, NGS):

In sequencing-based detection of translocation events, detection can in principle be performed at the DNA or RNA level; however, due to the sometimes very large intronic sequences in which translocation events occur, RNA-based detection is usually used, as the examination of exonic regions is sufficient here. The enrichment of the gene fusion segments to be examined can be either PCR-based (RT-PCR, amplification-based NGS method) or via sequence-specific enrichment (hybrid capture-based NGS method). In the former, only already known gene fusions can be detected, since both translocation partners and the exact breakpoints must be known and covered by appropriate primers. With the latter method, previously unknown gene fusions can in principle also be detected via the use of gene-specific enrichment methods, if at least one of the two genes involved is appropriately covered in the enrichment approach.

The advantage of RT-PCR is the high specificity and sensitivity for the respective fusion event and the low cost. The major disadvantage is that, on the one hand, if the partner gene is previously unknown and not covered, the test is negative, and on the other hand, as with FISH testing, only known and expected fusions can be tested. The former disadvantage also applies to amplification-based NGS methods. Advantage of hybrid capture-based NGS methodology is that also previously unknown gene fusions can be detected if at least one of the partner genes is included in the assay, as well as the possibility – with appropriate panel size - to detect fusion events/entities not previously suspected in the differential diagnosis. Disadvantages of NGS methods are the higher costs, the longer processing time and the sometimes very complex validation.