Thymus is a primary lymphoid organ with a crucial role in T cell maturation and the development of immune tolerance. The structure of the gland consists of an outer capsule, a cortex and a medulla. Immature T cells (thymocytes) move through the thymic cortex and corticomedullary junction, undergoing serial phenotypic modifications, which eventually result in their maturation through a positive and negative selection process (8,9). During positive selection, only those thymocytes that have a T cell receptor (TCR) capable of binding the Major Histocompatibility Complex (MHC) expressed on thymic epithelial cells are preserved and enter the thymic medulla, where a negative selection process takes places. More specifically, the medulla contains medullary thymic epithelial cells (mTECs) expressing various tissue-specific self-antigens (TSAs). Those T cells that interact with TSAs with high affinity undergo apoptosis (8,9). Expression of TSAs by mTECs is controlled by the transcription factors auto immune regulator (AIRE) and forebrain embryonic zinc finger-like protein 2 (Fezf2) (9,10). Nevertheless, negative selection is not completely efficient, since some auto reactive T cells might escape thymic selection and be released into the bloodstream and, therefore, additional peripheral tolerance mechanisms are critical to avoid autoimmunity (11,12).

Approximately 30% of patients with TM present with autoimmune and paraneoplastic syndromes at the time of diagnosis, the most common being Myasthenia Gravis (MG) and less frequently pure red cell aplasia, systemic lupus erythematosus and hypogammaglobulinaemia (13,14). In contrast to TM, TC, which is a more aggressive tumor, with local invasiveness, early nodal dissemination, and a higher metastatic potential, is not associated with autoimmune disorders, possibly because of the absence of immature T cells within the tumor (13).

It has been suggested that the disruption of thymic architecture and the dysfunctional mTECs in patients with TETs may result in impaired maturation of thymocytes and release of autoreactive T cells into the bloodstream (13). Defective expression of transcription factors AIRE and Fezf2 has also been implicated in impaired negative selection of autoreactive T cells. In addition, downregulated expression of MHC class II molecules by thymoma cells has been implicated in the depletion of central immune tolerance and predisposition towards autoimmunity (13–18).

Of note, data from The Cancer Genome Atlas analysis (TCGA) raise the hypothesis that deficient central immunotolerance and immunosuppression are unlikely to be the sole mechanism of MG in TM (15). In this study, the molecular aberrations of 24 TM associated with Myasthenia Gravis (MG+) were compared to those of 72 TM without Myasthenia Gravis (MG-). Genes that are characteristically implicated in immunotolerance mechanisms were not differentially expressed between MG+ and MG- thymomas. Moreover, MG+ thymomas were not associated with mutations in any single gene or with any specific DNA methylation signature or miRNA expression profile. MG+ thymomas, however, overexpress genes coding for mid-sized neurofilament and ryanodine receptors type III proteins, which share sequences with major antigens associated with MG, such as the acetylcholine receptor (AChR) and titin epitopes. Hence, it was proposed that an additional mechanism of autoimmunity in MG+ Thymomas could rely on molecular mimicry of antigens between tumor cells and the target organ (15).

Immune checkpoint inhibitors (ICIs) have changed the natural history of many types of cancer and have achieved durable responses in a subset of patients. The identification of predictive biomarkers to define the patients that are more likely to respond to checkpoint inhibition is an ongoing challenge (16). The two most well-studied biomarkers, yet with several limitations, are the programmed cell death ligand-1 (PD-L1) expression on tumor cells and/or lymphocytes and the tumor mutational burden (TMB) representing the number of single nucleotide variants in a tumor genome coding area and putatively indicating the ‘immunogenicity’ of the tumor (16).

It should be noted that PD-L1 is normally expressed in the non-neoplastic thymus (17) and PD-1/PD-L1 interaction negatively regulates the beta-selection and modulates the positive selection as well. Studies report that PD-L1 is generally highly expressed in neoplastic epithelial cells in TETs (18), but correlations with clinicopathological data and survival remain ambiguous with controversial results among studies (19–23). Also, TETs are characterized by a low TMB, which is the lowest among adult cancers, but it is significantly higher in TC compared to TM (15).

After several case reports were published, trials are now being conducted to assess the efficacy and safety of PD-1/PD-L1 inhibitors in patients with ΤΕΤs (Table I).

The PD-1 inhibitor pembrolizumab was investigated by Giaccone et al in 40 patients with recurrent TC in a phase II trial (24). Patients with prior history of autoimmune disease were ineligible. An overall response rate (ORR) of 22.5% was observed. Disease control was achieved in 30 (75%) patients with a median duration of response of 3 years. Median progression-free survival (mPFS) was 4.2 months and median overall survival (mOS) was 24.9 months. High PD-L1 expression (>50% of tumor cells), was found in ten patients, six of whom had a complete response (CR) or partial response (PR). Out of 27 patients with low or negative PD-L1 expression, 85% had progression of disease (PD) as the best response. IFN-γ signature expression was correlated with response to pembrolizumab. Targeted exome sequencing showed that TP53 mutations were associated to lower expression of PD-L1 and shorter OS, while mutations in CYLD, another tumor suppressor gene, were associated with high PD-L1 expression. When four patients were rechallenged with pembrolizumab, two responses were recorded, one of them two years after completing therapy. Six patients (15%) developed severe immune-related adverse events (irAEs), including myocarditis and polymyositis (24,25).

Another phase II trial evaluated pembrolizumab in 26 patients with recurrent TC and 7 patients with recurrent TM (26). Patients with active autoimmune disease were excluded. The ORR was 19.2% in patients with TC and 28.6% in patients with TM. The median duration of response was not reached in patients with TM and was 9.7 months in patients with TC. Median PFS was 6.1 months in both groups. Median OS was 14.5 months for TC and not reached in patients with TM. High PD-L1 levels were confirmed as significant predictive biomarker of response: 35.7% of patients with high PD-L1 expression achieved a PR, whereas none of the patients with low PD-L1 expression responded. Patients with TM experienced significantly more grade 3 or 4 irAEs as compared to TC (71% vs. 15%, respectively) including hepatitis, myocarditis and MG (26).

Rajan et al conducted a phase I trial with the PD-L1 inhibitor avelumab in 7 patients diagnosed with TM and one with TC (27). Almost 30% of the patients had an objective response, while 3 patients had response after a single dose of avelumab. A particularly higher rate of irAEs was described compared to other solid tumors treated with avelumab. The incidence of grade 3 AEs was 38%, with the same rate for grade 4 AEs: 5 patients developed severe irAEs including myositis and respiratory muscle insufficiency. Interestingly, all responders who developed irAEs had previously been treated with sunitinib, a multikinase inhibitor with an anti-angiogenic effect and with immunomodulatory properties. Only one of the patients that did not respond developed an irAE. PD-L1 expression predictive value was not evaluated due to inadequate number of patients for analysis.

Further analysis of the trial uncovered that pretreatment absolute lymphocyte count was higher in responders compared to non-responders, while percentage of B cells, regulatory T cells and natural killer cells were lower in responders vs. non-responders. Intratumoral immune infiltrates were also evaluated in two patients with TM and showed that the immune pre-treatment infiltrate mainly composed of immature T cells, shifted towards predominantly mature CD8+ T cells infiltrate in the responder, while it kept an immature T cells predominance in the non-responder. Moreover, pre-therapy peripheral blood mononuclear cells exhibited higher T-cell receptor diversity in patients who responded and developed irAEs compared to those who did not (27).

The same group also demonstrated that pre-existing anti-acetylcholine receptor (anti-AchR) autoantibodies and B cell lymphopenia confer risk for developing myositis after treatment with avelumab even in the absence of autoimmune clinical history (28). These findings need additional evaluation because they might represent biomarkers of pre-existing autoimmunity in patients without a clinical history of autoimmune disease, at higher risk of irAEs. These markers are under evaluation in an ongoing trial of avelumab in patients with advanced TETs (NCT03076554).

The PRIMER study was a two-stage single arm phase II trial that investigated the activity of the PD-1 inhibitor nivolumab in 15 patients with TC (29). No objective responses were observed in this trial and accrual closed early for futility at the first stage. The disease control rate (DCR) was 73%, the mPFS was 3.8 months and the mOS was 14.1 months, while the toxicity profile was manageable (29). Ak and Aydiner retrospectively tested the efficacy of nivolumab at four TMs, three TCs, and one mixed histology (30). Two patients' evaluation of best response was not applicable. Among the five available patients, the ORR was 66.7%, and the DCR was 100%. The median follow-up time was 16.1 months. The mPFS and mOS were 6.5 months and 7.4 months, respectively (30).

The NIVOTHYM trial (NCT03134118) is the first international multicenter phase II trial evaluating the use of nivolumab with or without ipilimumab in patients with advanced/relapsed type B3 TM or TC after platinum-based chemotherapy. The primary endpoint is PFS rate at 6 months. The results of the nivolumab monotherapy cohort demonstrate a manageable safety profile and objective activity; however, they are insufficient to meet the trial primary objective. The second cohort is currently ongoing to assess the combination of nivolumab plus ipilimumab and the results are eagerly awaited.

Angiogenesis plays a pivotal role in tumor progression and is regulated by several pro-angiogenic molecules, such as vascular endothelial growth factor (VEGF), platelet-derived growth factor (PDGF) and transforming growth factor beta (TGFβ), which have been found overexpressed in several cohorts of TETs (31–33). Recently, it has been reported that the deregulation of the equilibrium between activine A and its natural inhibitor follistatin is also a pro-angiogenic pathogenic mechanism in TETs. In patients with TC, high follistatin levels were observed and correlated with advanced tumor stage, and tumor microvessel density (34).

Sunitinib is an anti-angiogenic multikinase inhibitor targeting VEGFR, PDGFR and c-KIT. Sunitinib was administered as 50 mg orally once daily for six-week cycles (4 weeks on/2 weeks off treatment) in a phase II trial in patients with TETs (35). The study met its primary endpoint in the TC cohort with an ORR of 26% (DCR was 91%), while ORR in TM was only 6% leading to early close of the TM cohort per protocol rule. Also, mPFS was 7.2 and mOS was not reached within the TC cohort, while mPFS was 8.5 months and mOS was 15.5 months within the TM cohort. Lower baseline levels of circulating tumor cells and lower or stable levels of circulating endothelial cells after administration of sunitinib were associated with improved OS. An upregulation of immune checkpoint receptors was reported in most patients, which was associated with improved OS. Overall, toxicity profile was well tolerated, with a rate of adverse events similar to what is described in other cancers, but many patients in both cohorts required dose reductions (35).

Further studies have been conducted to evaluate alternative schedules of sunitinib and/or its administration in heavily pretreated TETs. A phase II study evaluated sunitinib at a modified dose of 50 mg once daily using a 2-weeks-on/1-week-off schedule. However, only 8% of the patients with TC responded (36). Another phase II trial conducted in 25 patients with metastatic TC after platinum-based chemotherapy reported 22% ORR and 70% SD with a PFS of 15.2 months (37). An ongoing phase II trial (NCT01621568) is also investigating sunitinib toxicity and efficacy using a schedule of 3-week cycles (50 mg daily for 2 weeks with 1 week off) in patients with advanced TETs with at least one prior line of platinum-based chemotherapy. In a real-world retrospective study of 28 patients from the French RYTHMIC network, 15 patients received sunitinib as ≥fourth-line treatment, with initial daily dose of sunitinib of 50 mg in 11 patients, 37.5 mg in 16 patients, and 25 mg in one patient (38). Sunitinib-related AEs seem to be tolerable. In the overall population, ORR was 22% (29% for TM and 20% for TC) and median PFS was 3.7 months (5.4 months for TM and 3.3 months for TC) (38).

REMORA trial explored the activity of lenvatinib, an oral multi-kinase inhibitor that targets VEGFR, FGFR, c-KIT, and other kinases, in 42 patients with TC who progressed after at least one line of platinum-based chemotherapy and had not previously received any anti-angiogenic agents (39). The trial met its primary endpoint with an ORR of 38% and an acceptable toxicity profile. The DCR was 95%, the mPFS was 9.3 months and the mOS was not reached. The most frequent treatment-related AEs were hypertension, thrombocytopenia, diarrhea and palmar-plantar erythrodysesthesia syndrome. Serious AEs were reported in 19% of the patients, including bowel perforation, left ventricular dysfunction, pneumonitis and electrocardiogram T wave abnormalities, while there were no deaths due to AEs (39). Additionally, a post hoc subgroup analysis by histological type revealed ORRs of 46.7 and 16.7% for squamous cell carcinoma and non-squamous cell carcinoma, respectively. The clinical activity of lenvatinib with ORR of 38% is the highest, to date, that has been reported in TETs that progressed after first-line chemotherapy, and therefore, lenvatinib is a promising therapeutic option.

Novel, small-molecule tyrosine kinase inhibitors (TKIs) with anti-angiogenic activity are explored for patients with TETs. Anlotinib is a new oral, broad spectrum TKI, which can strongly inhibit VEGFR, PDGFR, FGFR, c-kit. Anlotinib is highly selective for VEGF receptors VEGFR2 and VEGFR3 (40,41). A recent case report described the efficacy of anlotinib in a patient with refractory TC after multiple lines of chemotherapy as well as anti-angiogenic therapy with another multi-target TKI, apatinib (42). Apatinib is a highly selective TKI, which mainly competes with the ATP-binding site of VEGFR-2. The patient was unable to tolerate the toxicity associated with apatinib and by the 13th month disease progressed (43). After the first anti-angiogenic TKI failed, anlotinib was well tolerated, without obvious AEs and the patient achieved a PFS of over 23 months; therefore, further clinical investigation of anlotinib is warranted (42). Moreover, lucitanib is an oral, potent selective inhibitor of VEGFR, PDGFR and fibroblast growth factor receptor (FGFR), which was evaluated in a phase Ib trial enrolling 15 patients with TM or TC. Two patients had PR, while 11 patients had SD. PFS was 7.5 months in thymic carcinomas (44). Finally, a phase II trial with TKI regorafenib, in patients with TETs who progressed after at least one line of chemotherapy recently reported results with a disease control rate of 78.9% (45).

Less favorable results have been reported with bevacizumab, a humanized monoclonal antibody against VEGF, which was investigated in a phase 2 trial in combination with erlotinib. The trial enrolled 18 patients with recurrent TM or TC. No objective responses were observed, SD was observed in 11 patients (60%), while in 7 patients (40%) PD was the best response (46).

Currently, several phase II trials are ongoing with anti-angiogenic drugs either as single-agent therapies or in combination with other types of systemic treatment (chemotherapy or immunotherapy). A phase II trial assesses the activity of sunitinib in patients with type B3 TM or TC who have received at least one prior platinum-based chemotherapy (Style Trial-NCT03449173). RELEVENT is a phase II trial that will investigate the combination of ramucirumab, an anti-VEGFR2 monoclonal antibody, with carboplatin and paclitaxel in the first-line setting for relapsed or metastatic TETs of any histological type (NCT03921671).

c-KIT (CD117) is a transmembrane receptor with tyrosine kinase activity encoded by the proto-oncogene KIT. Overexpression of c-KIT is associated with the development of gastrointestinal stromal tumors (GIST), melanomas and certain types of leukemias and lymphomas; however, apart from the above, activating mutations of KIT are uncommon in most solid tumors (47). In TETs, overexpression of c-KIT is quite often in TC (46–80%), but KIT mutations are rare and are found in less than 10%. On the other hand, c-KIT overexpression is rare in TM (2–4%) (15,48). Most KIT-mutated TCs are poorly differentiated squamous cell carcinomas (48). Petrini et al evaluated a large cohort of 120 TETs specimens (13 TC and 107 TM) and observed that KIT overexpression was much higher in TC than TM, there was no association with the stage of the disease, but KIT overexpression was a negative prognostic marker. In this study, no KIT mutations were identified by sequencing the gene from exons 1 to 20 (49).

Although mutations in KIT are rare, when arising, they might be targetable with TKIs, such as imatinib mesylate, which is already widely used in chronic myeloid leukemia and GISTs. A number of studies have been performed to evaluate the association of specific mutations with sensitivity to imatinib or other inhibitors.

Girard et al investigated seven samples of TC and sequenced exons 10 and 14 in addition to the more frequently-mutated exons 9, 11, 13, and 17 (48). Interestingly, one of the mutations, H697Y, was in exon 14 and showed higher in vitro sensitivity to sunitinib than imatinib (48). Other mutations that have been reported in TC and show differential sensitivity to TKIs are V560del at exon 11, L576P at exon 11, Y553N at exon 11, D820E at exon 17, V559G at exon 11, V577-579del at exon 11, and K642E at exon 13 (47). Mutations at exon 11 confer sensitivity to imatinib. Strobel et al reported a V560del kit mutation in a patient with metastatic poorly differentiated epidermoid carcinoma, which was sensitive to imatinib and achieved a PFS of 6 months (50). Yoh et al identified the L576P kit mutation in exon 11 of a TC. This mutation was previously described in GIST to be sensitive to imatinib (51). V559G and Y553N mutations at exon 11 are susceptible to imatinib as well (52,53). Mutations at exons 13, 14 and 17 seem to be associated with primary resistance to imatinib. D820E mutation at exon 17 and K642E mutation at exon 13 confer resistance to imatinib but are sensitive to sorafenib (54,55). Bisagni et al reported a case of a TC harboring the mutation D820E at exon 17. The patient was treated with sorafenib, and the authors reported a partial response of more than 15 months (54). Another case with a 577-579del in exon 11 also conferred sensitivity to sorafenib (56).

Although responses to imatinib have been reported in those KIT-mutated tumors previously described, no responses were reported in two phase II trials evaluating the activity of imatinib in unselected patients or selected only based upon histologic type (B3 TM or TC), or KIT staining by immunohistochemistry and not upon genotyping (57,58).

Sorafenib is a multi-target TKI of KIT and other kinases. As previously described, it demonstrated efficacy in KIT-mutated tumors (54–56). It also showed antitumor activity in patients with refractory TCs, irrespective of the presence of KIT mutations (59–61). A case series of 5 patients with metastatic pre-treated TC reported DCR of 80% (PR in 2 patients, SD in 2 patients), and PD in 1 patient (20%) (59). The mPFS and mOS were 6.4 and 21.2 months, respectively. Of note, the tumor of only one of the two responding patients harbored a KIT mutation (D820E at exon 17).

Schirosi et al proposed an interesting and practical therapeutic algorithm based on the type of KIT mutations in order to choose the most effective TKI. It seems that clinical activity of imatinib in TC significantly depends on the presence and type of c-KIT mutation. On the other hand, sorafenib and sunitinib seem to be less selective than imatinib and can be effectively used in TC harboring imatinib-resistant c-KIT mutations (i.e. in exons 13, 14 and 17) or in wild-type TC due to their anti-angiogenic activity. Those data are based on small number of patients and further trials should be designed including genetically well-characterized populations (62).

Mutations at different levels of the phosphatidylinositol-3 kinase (PI3K)/AKT/mammalian target of rapamycin (mTOR) pathway have been observed in both TM and TC and despite their rarity if taken singularly, when taken together, they are found in more than 5% of TETs according to the TCGA PanCancer Atlas (15,63,64). Therefore, it has been suggested that compounds targeting proteins of the pathway, such as mTOR or PI3K might have clinical efficacy in TETs (63–65).

Everolimus, an oral mTOR inhibitor, was evaluated in a phase II trial, enrolling 32 patients with TM and 19 patients with TC after at least one previous platinum-based chemotherapy (66). The study met its primary endpoint with a DCR of 88% with one patient with TC achieving a CR and the majority of patients presenting with SD (76% SD, 10% PR, 2% CR). Further evaluation by histologic type showed that DCR was 94% in TM (including 3 PRs) and 78% in TC (including 1 CR and 2 PRs). The mPFS was 16.6 and 5.6 months for TM and TC respectively, and the mOS was not reached for TM and was 14.7 months for TC. Toxicity was an important issue since 14 patients developed severe AEs and 3 patients with TM died of drug-related pneumonitis.

An additional immunohistochemical analysis in the samples of 27 patients of the aforementioned trial revealed two prognostic biomarkers (positive expression of proteins IGF1-R and p4E-BP1) but no predictive biomarker for response to everolimus was identified (66). A study of next-generation sequencing on tumor samples from a small cohort of 15 pretreated patients with TET who received everolimus did not identify any predictive biomarkers either (67). Mutations in genes including TP53, KEAP1 and CDKN2A were observed in 27% of patients, without association with time to treatment failure (TTF) (67). Recently, in a study suggesting a molecular classification of TETs using genomic information, targeted agents were evaluated and in vitro resistance was noted for everolimus in tumors of TH4 subtype (68). Taking into consideration the limited antitumor activity, mainly in the form of stabilization of disease rather than antitumor activity, and the toxicity profile, everolimus should not be considered as a standard treatment for patients with TETs.

PI3K inhibitors have been evaluated in preclinical studies and showed potential efficacy in TETs. In vitro exposure to the PI3K inhibitor pictilisib inhibited cell viability and proliferation. However, a phase II study of buparlisib, another PI3K inhibitor, in relapsed or refractory TMs was closed early because of high rate of Grade 3–4 AEs (50%) and low efficacy (ORR of 7.1%) (NCT02220855) (69).

Zucali et al performed an immunohistochemical investigation of the insulin-like growth factor 1 receptor (IGF-1R) in 132 TETs and concluded that higher expression levels of IGF-1R were significantly correlated with more aggressive histology and more advanced stage of disease (64). Expression of IGF-1R is common in all histological subtypes of TETs (70).

The clinical efficacy of cixutumumab, a fully human monoclonal antibody binding the IGF-1R, was tested in a phase II study with 49 pre-treated patients with advanced TM and TC (71). In the TM cohort, only five out of 37 patients presented a PR, while 28 had SD. In the TC cohort, none of the 12 patients responded to treatment and only five had SD. The accrual in the TC cohort closed early due to poor efficacy. Regarding the toxicity profile, 31% of patients developed severe AEs, and 2 patients died. The most common grade 3–4 AEs were hyperglycemia and increased serum lipases. Also, 24% of patients with TM developed autoimmune disorders during treatment, the most common being pure red-cell aplasia. In this trial, a significant increase in IFN-γ expressing CD4+ T cells in blood samples of patients after treatment, especially among responders, was reported (71). High toxicity and insufficient efficacy as monotherapy halted the development of IGF-1R inhibitors in most solid tumors, and although some activity was reported in TM, further investigation is unlikely.

Epidermal growth factor receptor (EGFR) upregulation is detected and successfully targeted in many solid tumors. While EGFR overexpression is common in TM and TC and is associated with poor PFS and OS, EGFR mutations are rare (48,51,72).

Even though clinical activity of anti-EGFR targeted therapy, such as cetuximab, has been reported in case reports (73,74), two phase II trials reported poor efficacy (46,75): in a phase II study of gefitinib including 26 patients with advanced TM or TC, there was only one patient with PR as best response. DNA sequencing revealed no mutations in EGFR exons 18–21 (75). In another phase II trial, 18 patients with advanced TETs were enrolled to determine the effects of combined treatment with erlotinib plus bevacizumab, but no objective responses were observed (46). Recently, Zu et al presented a patient with advanced TC who harbored an EGFR exon 19 deletion and was treated with icotinib. Eventually, the patient had a complete response for almost 40 months (76).

Aberrations of cyclin-dependent kinases (CDKs), the enzymes that strictly control the transition of the cell cycle machinery, are frequent in solid tumors, leading to abnormal cell proliferation (77). Inhibition of CDKs is actively investigated across numerous solid tumors (78). In the case of TC, deletion of CDKN2A gene, encoding for CDK2, and loss of p16 expression, an inhibitor of the cell cycle, have been described and correlated with unfavorable prognosis (79).

Milciclib, an oral CDK inhibitor, was tested in two phase II trials. The CDKO-125A-006 trial enrolled 72 pre-treated with only one line of chemotherapy patients with B3 TM (27,8%) or TC (72,2%). The CDKO-125A-007 trial included 30 patients with B3 TM (56.7%) or TC (43.3%) who had already received multiple lines of chemotherapy. ORR was less than 5% in both trials, but DCR was 75.9 and 83.3%. The mPFS and mOS were 6.83 and 24.18 months for the first study, while mPFS was 9.76 months, and OS was not reached for the second trial (80).

Recently, the results of a phase II trial of palbociclib, another CDK4-6 inhibitor, have been reported (NCT03219554). The population of the study was 48 patients with advanced TETs who had been treated with one or more lines of cytotoxic chemotherapy. After a medial follow-up of 14.5 months, the PFS at 6 months was 60% and the mPFS was 11.0 months. Six of 48 patients (12.5%) achieved PR. The mOS was 26.4 months. The toxicity profile was overall tolerable (81).

Somatostatin Receptors (SSTRs) are expressed in TETs, thus the activity of octreotide, a somatostatin analog, with and without prednisone has been evaluated by three phase II trials (82–84). The primary endpoint was the ORR in each study, and was 37, 31.6, and 88%, respectively. Notably, no responses were reported in TCs. According to these results, somatostatin analogues may represent an option in octreoscan-positive TM.

Interestingly, a comprehensive analysis of cancer-related genetic alterations among TETs reported frequent somatic mutations in epigenetic regulatory genes in TC (85). Histone deacetylase (HDAC) are enzymes that regulate gene expression by altering the chromatin accessibility state and they represent the most investigated target of epigenetic therapy.

Belinostat, a HDAC inhibitor, was evaluated alone or in combination with chemotherapy, in 67 patients with TETs in two phase II trials (86,87). In the first study, the activity of belinostat was tested in 41 pretreated patients with advanced TETs (25 TM and 16 TC). The best response was PR for two (5%) patients (both with TM), SD for 25 (61%) patients and PD for 13 patients. The mPFS was 5.8 months (86). Another phase I/II trial of belinostat, alone and in combination with CAP chemotherapy in the first-line of metastatic or recurrent TETs included 26 patients (12 TM and 14 TCs) and demonstrated an ORR of 64% in patients with TM and 21% in patients with TC (87). Of note, belinostat showed immunomodulatory activity, leading to reduction in Tregs and exhausted CD8 (+) T cell populations in blood samples of patients, which was associated with efficacy endpoints. Such immunomodulatory properties of belinostat should be further examined in combination with immunotherapy (87). Clinical efficacy in these trials was modest, but additional investigation is needed.

The inactivation of tumor suppressor proteins (TSPs) is a common pathogenic mechanism of oncogenesis. Exportin-1 (XPO1) is the main nuclear export receptor for many TSPs involved in apoptotic signaling and cell-cycle regulation. Inhibition of XPO1 has been proposed as a novel therapeutic strategy. A preclinical study in models of TETs revealed that XPO1 hyperactivation led to nuclear exclusion and inactivation of TSPs, whereas its inhibition could restore TSPs nuclear accumulation and activity (88).

Antitumor activity of selinexor, a selective XPO1 inhibitor, has been reported in 4 patients with TETs in a phase I trial. One patient presented a PR, and three patients presented SD (89). Two ongoing phase II trials are evaluating the activity of selinexor in advanced TETs (NCT03193437, NCT03466827) but the first one of them was early terminated due to slow accrual. Table II summarizes selected clinical trials of targeted agents in TETs.