Skip to content
Bending the Curve Trial by Trial
?

Bone, soft tissue and gastrointestinal stromal tumours (connective-tissue cancers)

Sarcoma and GIST

How chemotherapy, limb-sparing surgery and kinase inhibitors shaped care of bone sarcoma, soft-tissue sarcoma and GIST.

Soft-tissue cancer including heart only · 5-year relative survival · United States, SEER 8 registries

60.06%→68.18%

1975 (observed value; first year in the table) to 2018 (observed value; last observed year in the table). Read these with the caveats under “Then and now”.Source

26 milestones, by decade

  • In a randomised trial reported in 1986, 36 patients with osteosarcoma of the extremity were assigned, after definitive surgery, to intensive multi-agent adjuvant chemotherapy or to observation. The abstract reports actuarial relapse-free survival at two years of 17 percent without adjuvant treatment and 66 percent with chemotherapy (P less than 0.001), with similar results in 77 additional patients who were not randomised. The randomised group was small, and the abstract does not name the agents.
  • In 1998, sequencing of the KIT gene in five gastrointestinal stromal tumours found mutations that left the receptor active without its ligand, and the authors suggested that the tumours might arise from interstitial cells of Cajal. On 1 February 2002 the FDA granted accelerated approval to imatinib for KIT-positive unresectable or metastatic GIST; the approval letter states that effectiveness rested on objective response rates and that no controlled trials had shown clinical benefit. In the 147-patient randomised comparison of 400 and 600 mg daily published in 2002, 79 patients (53.7 percent) had a partial response and none a complete response, pooled across both doses. FDA's list of accelerated approvals with verified clinical benefit gives 26 September 2008 as the traditional approval date for this indication.
  • Olaratumab with doxorubicin received FDA accelerated approval for soft-tissue sarcoma on 19 October 2016. The confirmatory randomised phase 3 trial ANNOUNCE (509 patients) reported median overall survival of 20.4 months with olaratumab plus doxorubicin and 19.7 months with placebo plus doxorubicin (hazard ratio 1.05; 95% CI 0.84 to 1.30; P = .69). The EMA EPAR page prints a revocation of the EU authorisation on 22 July 2019, and the FDA table of withdrawn accelerated approvals lists 25 February 2020; the two agencies acted separately, so the dates differ.

Evidence current to October 2026 · Checked against primary sources to the extent possible

A guided tour of 12 milestones. You can pause or step through it yourself.

Filter milestonesshowing all 26
Treatment type
Setting

The age of rays

  1. 1921 · Origin · UnderstandingEwing describes 'diffuse endothelioma of bone'In 1921 the American pathologist James Ewing reported several cases of a bone tumour under the name 'diffuse endothelioma of bone'. The tumour later took his name. A 2011 history article says he described the first cases of what is now called Ewing sarcoma.Ewing's report was published in the Proceedings of the New York Pathological Society under the title 'Diffuse endothelioma of bone' and reprinted in 2006 as a 'Classic'. A 2011 history article states that Ewing sarcoma is named after Ewing, who described the first cases in 1921. The reprint was cited by title only here; no clinical data are recorded.

Molecular origins

  1. 1981 · Treatment · ChemotherapyIntergroup Ewing's Sarcoma Studies: multimodal therapyTwo intergroup studies tested chemotherapy for Ewing sarcoma of bone, given with radiation and surgery. In the first, reported in 1981, adding doxorubicin or lung radiotherapy to a three-drug regimen did better. In the second, reported in 1990, a high-dose intermittent four-drug schedule beat a moderate-dose continuous one at five years, with more heart toxicity.IESS-II (214 patients, localized nonpelvic Ewing's sarcoma): high-dose intermittent vs moderate-dose continuous doxorubicin, cyclophosphamide, vincristine, dactinomycin. At 5 years, disease-free 68% vs 48% (P=.02), relapse-free 73% vs 56% (P=.03), surviving 77% vs 63% (P=.05); more severe cardiovascular toxicity with the intermittent schedule. IESS-I (first reported 1981; 1990 follow-up of 342 children): 5-year relapse-free survival 60% with doxorubicin added, 24% without, 44% with pulmonary radiotherapy.
  2. 1982 · De-escalation · SurgeryRandomised trial: limb-sparing surgery plus radiotherapy vs amputationIn a randomised trial of 43 adults with high-grade sarcoma of an arm or leg, limb-sparing surgery with radiotherapy was compared with amputation; everyone also had chemotherapy. Five-year survival was similar (83% and 88%). Cancer returned locally in four people who kept the limb and in none who had amputation.Randomised 2:1 (N=43, high-grade extremity STS): limb-sparing resection plus radiotherapy (n=27) vs amputation (n=16); all received postoperative doxorubicin, cyclophosphamide and high-dose methotrexate. Local recurrence 4 vs 0 (p1 = 0.06). Five-year DFS 71% vs 78% (p2 = 0.75); OS 83% vs 88% (p2 = 0.99). Multivariate analysis: final margin was the only correlate of local recurrence.
  3. 1982 · Treatment · ChemotherapyPreoperative chemotherapy tailors osteosarcoma treatmentIn a 1982 series of 57 people with osteosarcoma of a limb, chemotherapy was given before surgery. The removed tumour showed how well it had responded, and that result guided the chemotherapy given afterwards. Most people stayed free of recurrence during follow-up of up to about three years.Non-randomised series (n=57, primary osteogenic sarcoma of an extremity): preoperative high-dose methotrexate with citrovorum rescue, doxorubicin and BCD for 4-16 weeks; postoperative regimen chosen by histologic response. Of 35 patients without a good response, given a cisplatin-containing regimen, 32 (91%) were continuously disease-free at 6-34 months. Overall 53 of 57 (93%) were continuously disease-free at 6-35 months (median 20).
  4. 1986 · Treatment · ChemotherapyRandomised trials: adjuvant chemotherapy for osteosarcomaIn 1986 a randomised study of 36 people with osteosarcoma of a limb compared chemotherapy after surgery with observation alone. Two years later, about two in three people given chemotherapy had no relapse, against about one in six who were observed. A second randomised study reported a similar direction.Randomised trial (36 patients, nonmetastatic high-grade osteosarcoma of the extremity, after definitive surgery): adjuvant multi-agent chemotherapy vs observation. Two-year actuarial relapse-free survival 66% vs 17% (P<0.001); similar results in 77 further patients who declined randomisation. A separate randomised trial (Eilber 1987; 59 patients, 32 vs 27) reported significantly better disease-free and overall survival with chemotherapy at median 2 years.
  5. 1988 · Treatment · ChemotherapyIntergroup Rhabdomyosarcoma Studies I to IV: successive protocolsFour successive cooperative-group studies enrolled children and young people with rhabdomyosarcoma between 1972 and 1997. In the first, 5-year survival was 55% among 686 patients. In the fourth, three chemotherapy combinations gave similar 3-year failure-free survival (75% to 77%) in patients without metastases.Four sequential Intergroup Rhabdomyosarcoma Studies (IRS-I enrolled patients younger than 21 years). IRS-I (N=686): 5-year OS 55%; 52% in clinical group III, 20% in group IV. IRS-II (N=999): 5-year OS 63%. IRS-III (N=1062): 5-year PFS 65% vs 55% in IRS-II (historical comparison, P<.001). IRS-IV, nonmetastatic (N=883): 3-year FFS 75% (VAC), 77% (VAI), 77% (VIE); P=.42.

Targeted therapy

  1. 1998 · Understanding · UnderstandingActivating KIT mutations found in gastrointestinal stromal tumoursIn 1998 researchers sequenced the KIT gene in five gastrointestinal stromal tumours and found mutations that kept the KIT growth signal switched on without its usual trigger. A separate 1998 study found KIT on all 78 tumours it examined. Both groups suggested the tumours may arise from the gut's pacemaker cells, the interstitial cells of Cajal.Hirota et al. (Science 1998) sequenced KIT cDNA from five GISTs: mutations between the transmembrane and tyrosine kinase domains; mutant proteins constitutively activated; transfection transformed Ba/F3 cells. Kindblom et al. (Am J Pathol 1998): all 78 stromal tumours strongly KIT-immunoreactive, 56 of 78 CD34-positive; 30 of 30 control tumours (including leiomyomas and leiomyosarcomas) KIT-negative. Both proposed an origin from interstitial cells of Cajal.
  2. 2001 · Treatment · TargetedImatinib in GIST: case report, phase 2 and FDA approvalA 2001 case report described the drug STI571, later called imatinib, in one person with metastatic GIST. In February 2002 the US regulator approved imatinib for KIT-positive GIST that could not be removed or had spread, based on response rates. A randomised trial published that August reported partial responses in about half of 147 patients.Joensuu (NEJM 2001): single-patient report (no abstract). FDA accelerated approval 1 Feb 2002 (21 CFR 314 subpart H) for KIT-positive unresectable/metastatic GIST on objective response rates; no controlled trials of clinical benefit. Demetri (NEJM 2002), randomised, 147 patients, 400 or 600 mg daily: partial response 53.7% (79 patients), stable disease 27.9%, no complete responses, no significant difference between doses. Long-term: median survival 57 months, median follow-up 63 months.
  3. 2003 · Understanding · UnderstandingPDGFRA mutations and KIT genotype shape imatinib responseIn 2003 researchers reported that about 35 percent of GISTs without KIT mutations carried activating mutations in a related gene, PDGFRA. A second 2003 study linked the mutation type to imatinib response: partial responses were most common with KIT exon 11 mutations and did not occur in tumours with no detected KIT or PDGFRA mutation.Heinrich (Science 2003): PDGFRA activating mutations in 14 of 40 (about 35%) GISTs lacking KIT mutations; KIT and PDGFRA mutations appeared mutually exclusive. Heinrich (J Clin Oncol 2003): 127 patients from a phase II imatinib study; KIT mutations 88.2%, PDGFRA 4.7%; partial response 83.5% (KIT exon 11), 47.8% (exon 9; P=.0006), 0.0% (no detected mutation; P<.0001).
  4. 2003 · Treatment · ChemotherapyINT-0091: adding ifosfamide and etoposide in Ewing sarcomaA randomised trial in children and young adults with Ewing sarcoma added ifosfamide and etoposide to a four-drug regimen. When the cancer had not spread, about seven in ten were free of events at five years, against about half on the standard regimen. When it had spread, the added drugs made no significant difference.Randomised trial (518 eligible patients aged 30 or younger): standard doxorubicin, vincristine, cyclophosphamide, dactinomycin vs the same drugs alternating with ifosfamide and etoposide. Nonmetastatic stratum (n=398): 5-year EFS 69% vs 54% (P=0.005); OS 72% vs 61% (P=0.01). Metastatic stratum (n=120): no significant difference in 5-year EFS (P=0.81).
  5. 2004 · Treatment · TargetedImatinib 400 mg vs 800 mg daily in advanced GISTTwo randomised trials compared imatinib 400 mg once daily with 400 mg twice daily in advanced GIST. The EORTC trial reported longer progression-free survival on the higher dose; the North American S0033 trial found no significant differences. A pooled analysis of 1,640 patients found a small progression-free survival advantage, mainly with KIT exon 9 mutations, and no survival difference.EORTC 62005 (946 randomised, 400 mg once vs twice daily): progression 56% vs 50%, HR 0.82 (0.69-0.98), p=0.026. S0033 (746 enrolled): median PFS 18 vs 20 months, median OS 55 vs 51 months, no significant differences. MetaGIST (1,640 patients, median follow-up 45 months): small PFS advantage for the high dose, identical OS, PFS benefit essentially in KIT exon 9 mutants (772 genotyped).
  6. 2005 · Treatment · ChemotherapyINT-0133: mifamurtide added to chemotherapy; EU authorisationA Children's Oncology Group trial added ifosfamide, a liposomal immune-stimulating drug (muramyl tripeptide, now called mifamurtide), or both to standard chemotherapy for osteosarcoma. The first report could not be read as a simple comparison because the two additions interacted. A later analysis reported better survival with the immune-stimulating drug, and the European Commission authorised it in 2009.INT-0133, 2 x 2 factorial (677 patients in 2005; 662 in 2008): ifosfamide and/or MTP added to cisplatin, doxorubicin, high-dose methotrexate. 2005: 3-year EFS 71% standard, 68% +MTP, 61% +ifosfamide, 78% both; interaction precluded factorial analysis. 2008 (no interaction): MTP improved 6-year OS from 70% to 78% (P=.03; HR 0.71, 95% CI 0.52-0.96); ifosfamide did not improve EFS or OS.
  7. 2006 · Treatment · TargetedSunitinib after imatinib failure in GISTIn January 2006 the US regulator approved sunitinib for GIST that had progressed on, or could not be treated with, imatinib. In the randomised trial reported that October, the time before tumours grew was a median of about 27 weeks with sunitinib and about 6 weeks with placebo.Double-blind, placebo-controlled trial in 312 patients resistant to or intolerant of imatinib, randomised 2:1 to sunitinib (n=207) or placebo (n=105); stopped early at a planned interim analysis. Median time to tumour progression 27.3 weeks (95% CI 16.0-32.1) vs 6.4 weeks (4.4-10.0), HR 0.33, p<0.0001. FDA approval of NDA 21-938 on 26 January 2006.
  8. 2008 · Treatment · TargetedAdjuvant imatinib after GIST resection: 1 year, then 3 yearsIn December 2008 the US regulator approved imatinib after surgery for KIT-positive GIST. A 2009 placebo-controlled trial found fewer recurrences with one year of imatinib. A 2012 trial in high-risk disease found longer recurrence-free and overall survival with three years than one year. A third trial against observation did not meet its main endpoint.Z9001 (placebo, 1 year of imatinib; KIT-positive GIST of at least 3 cm): 1-year RFS 98% vs 83%, HR 0.35 (0.22-0.53). SSG XVIII/AIO (high risk; 36 vs 12 months): 5-year RFS 65.6% vs 47.9%, HR 0.46 (0.32-0.65); 5-year OS 92.0% vs 81.7%, HR 0.45 (0.22-0.89). EORTC trial (2 years vs observation): 5-year IFFS 87% vs 83%, HR 0.87, P=0.31 (final). FDA accelerated approval 19 Dec 2008; conversion 31 Jan 2012.

Immunotherapy

  1. 2012 · Treatment · ChemotherapyAEWS0031: interval-compressed chemotherapy in Ewing sarcomaA Children's Oncology Group trial compared giving the same Ewing sarcoma chemotherapy every two weeks or every three weeks. At five years, roughly three in four people given it every two weeks were free of events, against about two in three every three weeks. Side effects were reported as similar.AEWS0031: 587 enrolled, 568 eligible (284 per arm), age under 50, localized extradural Ewing sarcoma; vincristine/doxorubicin/cyclophosphamide alternating with ifosfamide/etoposide every 14 vs 21 days. 5-year EFS 73% vs 65% (HR 0.74, 95% CI 0.54-0.99; P=.048); toxicity similar. Update: 10-year EFS 70% vs 61% (P=.03), OS 76% vs 69% (P=.04); no difference in second malignant neoplasm incidence (P=.5).
  2. 2012 · Treatment · TargetedPALETTE: pazopanib after chemotherapy in non-adipocytic sarcomaIn 369 people with metastatic non-adipocytic soft-tissue sarcoma that had progressed after chemotherapy, the tablet pazopanib delayed progression compared with placebo: median 4.6 months versus 1.6 months. Overall survival was not significantly different (12.5 versus 10.7 months). Diarrhoea, nausea and high blood pressure were more common with pazopanib.Double-blind phase 3 (PALETTE; N=369, 2:1): pazopanib 800 mg daily (n=246) vs placebo (n=123), no crossover, after failure of standard chemotherapy. Median PFS (primary) 4.6 vs 1.6 months (HR 0.31; 95% CI 0.24-0.40; p<0.0001). Median OS 12.5 vs 10.7 months (HR 0.86; 95% CI 0.67-1.11; p=0.25). FDA approval 26 April 2012 for advanced STS after prior chemotherapy (limitation of use: adipocytic STS).
  3. 2012 · Treatment · TargetedGRID: regorafenib after imatinib and sunitinibThe GRID trial gave regorafenib or placebo to people with metastatic GIST that had progressed on imatinib and sunitinib. Median progression-free survival was 4.8 months with regorafenib and 0.9 months with placebo. The US regulator approved regorafenib for this use in February 2013.GRID, double-blind phase 3 (199 patients, 57 hospitals, 17 countries): regorafenib 160 mg daily (3 weeks of each 4; n=133) vs placebo (n=66), 2:1, plus best supportive care. Median PFS by independent central review 4.8 vs 0.9 months, HR 0.27 (95% CI 0.19-0.39), p<0.0001; 56 placebo patients (85%) crossed over. FDA approval 25 Feb 2013 under NDA 204369.
  4. 2014 · Treatment · ChemotherapyEORTC 62012: doxorubicin vs doxorubicin plus ifosfamideIn 455 adults aged 18 to 60 with advanced soft-tissue sarcoma, adding ifosfamide to doxorubicin (given on an intensified schedule) did not significantly lengthen overall survival (median 14.3 versus 12.8 months). It lengthened progression-free survival (7.4 versus 4.6 months) and tumours responded more often (26% versus 14%), with more blood-count toxicity.Phase 3 (EORTC 62012; N=455; age 18-60, WHO PS 0-1, high-grade): doxorubicin 75 mg/m2 (n=228) vs intensified doxorubicin plus ifosfamide 10 g/m2 with pegfilgrastim (n=227), up to six cycles. Median OS (primary, ITT) 12.8 vs 14.3 months (HR 0.83; 95.5% CI 0.67-1.03; p=0.076). Median PFS 4.6 vs 7.4 months (HR 0.74; 95% CI 0.60-0.90; p=0.003). Response 14% vs 26%. Febrile neutropenia 13% vs 46%.
  5. 2015 · Treatment · ChemotherapyTrabectedin vs dacarbazine in liposarcoma and leiomyosarcomaIn 518 adults with advanced liposarcoma or leiomyosarcoma that had progressed after earlier treatment, trabectedin delayed progression compared with dacarbazine (median 4.2 versus 1.5 months). Overall survival did not differ at the first analysis (12.4 versus 12.9 months) or at the final analysis, published later (13.7 versus 13.1 months).Phase 3 (N=518, 2:1): trabectedin (n=345) vs dacarbazine (n=173) after an anthracycline and at least one further systemic regimen. Median PFS (final analysis) 4.2 vs 1.5 months (HR 0.55; P<.001). Interim OS, the primary endpoint (64% censored): 12.4 vs 12.9 months (HR 0.87; P=.37). Final OS in a later report (577 randomised): 13.7 vs 13.1 months (HR 0.93; P=.49). FDA approval 23 October 2015.
  6. 2016 · Treatment · ChemotherapyEribulin vs dacarbazine in liposarcoma and leiomyosarcomaIn 452 adults with advanced liposarcoma or leiomyosarcoma who had already had at least two treatments, eribulin gave longer overall survival than dacarbazine (median 13.5 versus 11.5 months). Side effects of grade 3 or higher were more common with eribulin (67% versus 56%). The US approval covers liposarcoma only.Open-label phase 3 (N=452; intermediate/high-grade liposarcoma or leiomyosarcoma; at least two prior regimens including an anthracycline), 1:1: eribulin 1.4 mg/m2 on days 1 and 8 (n=228) vs dacarbazine (n=224) every 21 days. Median OS (primary, ITT) 13.5 vs 11.5 months (HR 0.77; 95% CI 0.62-0.95; p=0.0169). Grade 3 or higher adverse events 67% vs 56%. FDA approval 28 January 2016 for liposarcoma only.
  7. 2016 · Reversal · ChemotherapyEURAMOS-1: response-adapted escalation did not improve EFSEURAMOS-1 tested changing osteosarcoma treatment after surgery according to how well the tumour had responded to earlier chemotherapy. For poor responders, adding two more drugs did not improve event-free survival and caused more side effects. For good responders, adding maintenance interferon was not statistically different from chemotherapy alone.EURAMOS-1 (2260 registered, 1334 randomised). Poor responders (618): MAP vs MAPIE, EFS HR 0.98 (95% CI 0.78-1.23), non-proportional hazards; grade 4 non-haematological toxicity 24% vs 12% with MAPIE. Good responders (716): MAP vs MAP plus pegylated interferon alfa-2b, EFS HR 0.83 (0.61-1.12; P=.214). JCOG0905 (51 vs 52 poor responders): DFS HR 1.05 (0.55-1.98), OS HR 1.48 (0.68-3.22).
  8. 2019 · Reversal · ImmunotherapyOlaratumab: phase 2 survival benefit not confirmed in ANNOUNCEA randomised phase 2 trial in 133 people reported median survival of 26.5 months with olaratumab plus doxorubicin versus 14.7 months with doxorubicin alone, and the drug received US accelerated approval in 2016. The phase 3 ANNOUNCE trial in 509 people found no difference (20.4 versus 19.7 months), and the European authorisation was revoked in 2019.Phase 2 (N=133 randomised): median OS 26.5 vs 14.7 months (HR 0.46; 95% CI 0.30-0.71); PFS HR 0.67 (95% CI 0.44-1.02; p=0.0615). Phase 3 ANNOUNCE (N=509; doxorubicin plus olaratumab vs plus placebo): median OS 20.4 vs 19.7 months (HR 1.05; 95% CI 0.84-1.30; p=.69); leiomyosarcoma 21.6 vs 21.9 (HR 0.95). FDA accelerated approval 19 October 2016; EC revocation 22 July 2019; FDA withdrawal listed 25 February 2020.

Precision & engineered

  1. 2020 · Treatment · TargetedAvapritinib (PDGFRA exon 18) and ripretinib (fourth line)In 2020 the US regulator approved two GIST drugs. Avapritinib, for PDGFRA exon 18 mutations, was tested in a single-arm phase 1 study in which 49 of 56 people with the D842V mutation had a response. Ripretinib, after three or more kinase inhibitors, was tested against placebo: median progression-free survival 6.3 versus 1.0 months.NAVIGATOR (phase 1, open-label), PDGFRA D842V-mutant GIST: overall response 88% (95% CI 76-95), 49 of 56 patients, 5 complete and 44 partial responses; recommended phase 2 dose 300 mg. INVICTUS (phase 3, 129 patients, 2:1, ripretinib 150 mg daily vs placebo): median PFS by BICR 6.3 months (95% CI 4.6-6.9) vs 1.0 (0.9-1.7), HR 0.15 (0.09-0.25), p<0.0001. FDA approvals 9 January and 15 May 2020.
  2. 2020 · Reversal · RadiationSTRASS: preoperative radiotherapy in retroperitoneal sarcomaIn 266 adults with a newly diagnosed retroperitoneal sarcoma that could be operated on, adding radiotherapy before surgery did not lengthen the time free of abdominal recurrence: median 4.5 years with radiotherapy plus surgery and 5.0 years with surgery alone. The authors concluded that preoperative radiotherapy should not be standard.Open-label phase 3 (EORTC-62092 STRASS; N=266; 133 per arm, 1:1): preoperative radiotherapy (50.4 Gy in 28 fractions) plus surgery vs surgery alone. Median follow-up 43.1 months. Median abdominal recurrence-free survival (primary, ITT) 4.5 vs 5.0 years (HR 1.01; 95% CI 0.71-1.44; log-rank p=0.95). Grade 3-4 lymphopenia 77% vs 1%; serious adverse events 24% vs 10%.
  3. 2024 · Treatment · ImmunotherapySU2C-SARC032: pembrolizumab added to radiotherapy before surgeryIn a randomised trial of people with stage III sarcoma of the arm or leg (undifferentiated pleomorphic sarcoma, or dedifferentiated or pleomorphic liposarcoma), adding the immunotherapy pembrolizumab to radiotherapy and surgery was linked to longer disease-free survival: 67% versus 52% at two years. Grade 3 or higher side effects were more common (56% versus 31%).Open-label randomised trial (registered as phase 2; 143 randomised; modified ITT n=127, median follow-up 43 months): preoperative radiotherapy then surgery (n=63) vs preoperative pembrolizumab with radiotherapy, surgery and adjuvant pembrolizumab (n=64). DFS (primary): HR 0.61 (90% CI 0.39-0.96; one-sided log-rank p=0.035); 2-year DFS 52% vs 67% (90% CI 42-64 and 58-78). Grade 3 or higher adverse events 31% vs 56%.
  4. 2026 · Reversal · TargetedTazemetostat: epithelioid sarcoma accelerated approval withdrawnTazemetostat, a tablet that blocks the EZH2 protein, received US accelerated approval in January 2020 for advanced epithelioid sarcoma after a single-arm study in which 9 of 62 people (15%) responded. In March 2026 the manufacturer withdrew it after a safety signal in a lymphoma trial; the FDA lists the sarcoma approval as withdrawn on 22 June 2026.Single-arm phase 2 basket study, epithelioid sarcoma cohort (N=62, INI1/SMARCB1 loss, tazemetostat 800 mg twice daily): objective response 15% (9 of 62; 95% CI 7-26; primary); median PFS 5.5 months; median OS 19.0 months; median follow-up 13.8 months. FDA accelerated approval 23 January 2020. Ipsen withdrew the product 9 March 2026 (safety signal in SYMPHONY-1); FDA table: withdrawal 22 June 2026.

Then and now

  • 5-year relative survival · United States, SEER 8 registries

    • 60.06%

      1975 (observed value; first year in the table)

    All Races, Both Sexes (the chart note's label); the page states no age range for the survival series; SEER Cancer Stat Facts category 'Soft Tissue including Heart' (page heading 'Soft Tissue Cancer'). SEER 8 five-year relative survival series, 1975-2018. The page lists soft tissue sarcoma, childhood soft tissue sarcoma and the Ewing sarcoma family of tumors among its types of soft tissue cancer and gives no figure by histology.

    Modeled trend value for 1975: 63.48%. The observed yearly values are variable (lowest 55.18% in 1979, highest 71.95% in 2014). The SEER registry groupings page describes SEER 8 as the grouping available for cases diagnosed from 1975. Relative survival excludes deaths from other causes. 'Observed' is the table's label for the actual yearly value, as opposed to the Joinpoint modeled-trend value in the adjacent column. The page headings and chart note ('SEER 8 5-Year Relative Survival Percent from 1975-2018') do not label the 'Year' column as year of diagnosis. The page does not describe changes over time in case mix, diagnostic criteria, staging or treatment and attributes the change to no cause; read beside the mortality values. Compare only with other values from this same SEER 8 series, not with the SEER 21 values. Group defined by tumour site, not histology: SEER's site recode table (ICD-O-3/WHO 2008; the Stat Facts page does not say which recode it uses) lists C38.0, C47.0-C47.9 and C49.0-C49.9 for 'Soft Tissue including Heart', excluding some morphology codes, and lists retroperitoneum (C48.0), stomach and small intestine as separate site groups; SEER publishes no separate figure for GIST on a Cancer Stat Facts page (the gist page returned HTTP 404 on 21 October 2026).

    Source

  • 5-year relative survival · United States, SEER 8 registries

    • 71.95%

      2014 (observed value; highest observed value in the table)

    All Races, Both Sexes (the chart note's label); the page states no age range for the survival series; SEER Cancer Stat Facts category 'Soft Tissue including Heart' (page heading 'Soft Tissue Cancer'). SEER 8 five-year relative survival series, 1975-2018. The page lists soft tissue sarcoma, childhood soft tissue sarcoma and the Ewing sarcoma family of tumors among its types of soft tissue cancer and gives no figure by histology.

    Modeled trend value for 2014: 67.81%. Shown because the observed series is noisy: single-year values move by several percentage points. The SEER registry groupings page describes SEER 8 as the grouping available for cases diagnosed from 1975. Relative survival excludes deaths from other causes. 'Observed' is the table's label for the actual yearly value, as opposed to the Joinpoint modeled-trend value in the adjacent column. The page headings and chart note ('SEER 8 5-Year Relative Survival Percent from 1975-2018') do not label the 'Year' column as year of diagnosis. The page does not describe changes over time in case mix, diagnostic criteria, staging or treatment and attributes the change to no cause; read beside the mortality values. Compare only with other values from this same SEER 8 series, not with the SEER 21 values. Group defined by tumour site, not histology: SEER's site recode table (ICD-O-3/WHO 2008; the Stat Facts page does not say which recode it uses) lists C38.0, C47.0-C47.9 and C49.0-C49.9 for 'Soft Tissue including Heart', excluding some morphology codes, and lists retroperitoneum (C48.0), stomach and small intestine as separate site groups; SEER publishes no separate figure for GIST on a Cancer Stat Facts page (the gist page returned HTTP 404 on 21 October 2026).

    Source

  • 5-year relative survival · United States, SEER 8 registries

    • 68.18%

      2018 (observed value; last observed year in the table)

    All Races, Both Sexes (the chart note's label); the page states no age range for the survival series; SEER Cancer Stat Facts category 'Soft Tissue including Heart' (page heading 'Soft Tissue Cancer'). SEER 8 five-year relative survival series, 1975-2018. The page lists soft tissue sarcoma, childhood soft tissue sarcoma and the Ewing sarcoma family of tumors among its types of soft tissue cancer and gives no figure by histology.

    Modeled trend value for 2018: 68.23%; the table shows no observed five-year value for 2019 onward. The SEER registry groupings page describes SEER 8 as the grouping available for cases diagnosed from 1975. Relative survival excludes deaths from other causes. 'Observed' is the table's label for the actual yearly value, as opposed to the Joinpoint modeled-trend value in the adjacent column. The page headings and chart note ('SEER 8 5-Year Relative Survival Percent from 1975-2018') do not label the 'Year' column as year of diagnosis. The page does not describe changes over time in case mix, diagnostic criteria, staging or treatment and attributes the change to no cause; read beside the mortality values. Compare only with other values from this same SEER 8 series, not with the SEER 21 values. Group defined by tumour site, not histology: SEER's site recode table (ICD-O-3/WHO 2008; the Stat Facts page does not say which recode it uses) lists C38.0, C47.0-C47.9 and C49.0-C49.9 for 'Soft Tissue including Heart', excluding some morphology codes, and lists retroperitoneum (C48.0), stomach and small intestine as separate site groups; SEER publishes no separate figure for GIST on a Cancer Stat Facts page (the gist page returned HTTP 404 on 21 October 2026).

    Source

  • 5-year relative survival · United States, SEER 21 registries (excluding IL)

    • 65.7%

      2016-2022

    SEER Cancer Stat Facts category 'Soft Tissue including Heart'; the headline value in the page's At a Glance panel and survival section, which does not label age, sex, race or stage (the stage table beneath it is labelled All Races, Both Sexes). Stage unknown (8% of cases) has a five-year relative survival of 62.7%, not entered as a record.

    Separate series from the SEER 8 values: SEER 21 (excluding IL), a different grouping of registries; the SEER registry groupings page says SEER 21 has a more limited set of available variables and is only available for limited statistics. Not comparable with the SEER 8 yearly values. Relative survival excludes deaths from other causes. The page does not say whether the period is year of diagnosis. Group defined by tumour site, not histology: SEER's site recode table (ICD-O-3/WHO 2008; the Stat Facts page does not say which recode it uses) lists C38.0, C47.0-C47.9 and C49.0-C49.9 for 'Soft Tissue including Heart', excluding some morphology codes, and lists retroperitoneum (C48.0), stomach and small intestine as separate site groups; SEER publishes no separate figure for GIST on a Cancer Stat Facts page (the gist page returned HTTP 404 on 21 October 2026).

    Source

  • 5-year relative survival · United States, SEER 21 registries (excluding IL)

    • 83.3%

      2016-2022; localized stage at diagnosis (57% of cases)

    All Races, Both Sexes (the stage table's label); age range not stated; 'Soft Tissue including Heart', SEER Combined Summary Stage 'Localized' (confined to the primary site); 57% of cases in the stage table (the page text gives 56.8%).

    Separate series from the SEER 8 values: SEER 21 (excluding IL), a different grouping of registries; the SEER registry groupings page says SEER 21 has a more limited set of available variables and is only available for limited statistics. Not comparable with the SEER 8 yearly values. Relative survival excludes deaths from other causes. The page gives the stage-specific values for the whole group and no breakdown by histology or site. Group defined by tumour site, not histology: SEER's site recode table (ICD-O-3/WHO 2008; the Stat Facts page does not say which recode it uses) lists C38.0, C47.0-C47.9 and C49.0-C49.9 for 'Soft Tissue including Heart', excluding some morphology codes, and lists retroperitoneum (C48.0), stomach and small intestine as separate site groups; SEER publishes no separate figure for GIST on a Cancer Stat Facts page (the gist page returned HTTP 404 on 21 October 2026).

    Source

  • 5-year relative survival · United States, SEER 21 registries (excluding IL)

    • 58.2%

      2016-2022; regional stage at diagnosis (18% of cases)

    All Races, Both Sexes (the stage table's label); age range not stated; 'Soft Tissue including Heart', SEER Combined Summary Stage 'Regional' (spread to regional lymph nodes); 18% of cases.

    Separate series from the SEER 8 values: SEER 21 (excluding IL), a different grouping of registries; the SEER registry groupings page says SEER 21 has a more limited set of available variables and is only available for limited statistics. Not comparable with the SEER 8 yearly values. Relative survival excludes deaths from other causes. The page gives the stage-specific values for the whole group and no breakdown by histology or site. Group defined by tumour site, not histology: SEER's site recode table (ICD-O-3/WHO 2008; the Stat Facts page does not say which recode it uses) lists C38.0, C47.0-C47.9 and C49.0-C49.9 for 'Soft Tissue including Heart', excluding some morphology codes, and lists retroperitoneum (C48.0), stomach and small intestine as separate site groups; SEER publishes no separate figure for GIST on a Cancer Stat Facts page (the gist page returned HTTP 404 on 21 October 2026).

    Source

  • 5-year relative survival · United States, SEER 21 registries (excluding IL)

    • 17%

      2016-2022; distant stage at diagnosis (17% of cases)

    All Races, Both Sexes (the stage table's label); age range not stated; 'Soft Tissue including Heart', SEER Combined Summary Stage 'Distant' (cancer has metastasized); 17% of cases.

    Separate series from the SEER 8 values: SEER 21 (excluding IL), a different grouping of registries; the SEER registry groupings page says SEER 21 has a more limited set of available variables and is only available for limited statistics. Not comparable with the SEER 8 yearly values. Relative survival excludes deaths from other causes. The page gives the stage-specific values for the whole group and no breakdown by histology or site. Group defined by tumour site, not histology: SEER's site recode table (ICD-O-3/WHO 2008; the Stat Facts page does not say which recode it uses) lists C38.0, C47.0-C47.9 and C49.0-C49.9 for 'Soft Tissue including Heart', excluding some morphology codes, and lists retroperitoneum (C48.0), stomach and small intestine as separate site groups; SEER publishes no separate figure for GIST on a Cancer Stat Facts page (the gist page returned HTTP 404 on 21 October 2026).

    Source

  • Age-standardised mortality · United States

    • 0.9per 100,000

      1975 (observed value; first year in the table)

    All Races, Both Sexes (the table note); age-adjusted death rate for 'Soft Tissue including Heart' (page heading 'Soft Tissue Cancer'); U.S. Mortality series, 1975-2024.

    Modeled trend value for 1975: 0.83. Cause-of-death data from U.S. Mortality (National Center for Health Statistics data as used by SEER), not SEER registry data; the page does not describe how its cause-of-death category maps to the registry site group. Rates are age-adjusted per 100,000. The page does not describe screening for this cancer; read it beside the SEER 8 survival values, which are a separate series.

    Source

  • Age-standardised mortality · United States

    • 1.55per 100,000

      1997 (observed value; highest observed value in the table)

    All Races, Both Sexes (the table note); age-adjusted death rate for 'Soft Tissue including Heart'; U.S. Mortality series, 1975-2024.

    Modeled trend value for 1997: 1.52. The page text states that age-adjusted death rates have been falling on average 0.6% each year over 2015-2024. Cause-of-death data from U.S. Mortality (National Center for Health Statistics data as used by SEER), not SEER registry data; the page does not describe how its cause-of-death category maps to the registry site group. Rates are age-adjusted per 100,000. The page does not describe screening for this cancer; read it beside the SEER 8 survival values, which are a separate series.

    Source

  • Age-standardised mortality · United States

    • 1.26per 100,000

      2024 (observed value; last year in the table)

    All Races, Both Sexes (the table note); age-adjusted death rate for 'Soft Tissue including Heart'; U.S. Mortality series, 1975-2024. The page's 2020-2024 average is 1.3 per 100,000.

    Modeled trend value for 2024: 1.27. Cause-of-death data from U.S. Mortality (National Center for Health Statistics data as used by SEER), not SEER registry data; the page does not describe how its cause-of-death category maps to the registry site group. Rates are age-adjusted per 100,000. The page does not describe screening for this cancer; read it beside the SEER 8 survival values, which are a separate series.

    Source

  • 5-year relative survival · United States, SEER 8 registries

    • 53.4%

      1975 (observed value; first year in the table)

    All Races, Both Sexes (the chart note's label); the page states no age range for the survival series; SEER Cancer Stat Facts category 'Bone and Joint Cancer'. SEER 8 five-year relative survival series, 1975-2018.

    Modeled trend value for 1975: 47.05%, which is lower than the observed value. The observed yearly values are variable (lowest 47.59% in 1980, highest 75.53% in 1993); the page calls bone and joint cancer rare (estimated 4,110 new cases in the U.S. in 2026) and does not explain the year-to-year variation. The SEER registry groupings page describes SEER 8 as the grouping available for cases diagnosed from 1975. Relative survival excludes deaths from other causes. 'Observed' is the table's label for the actual yearly value, as opposed to the Joinpoint modeled-trend value in the adjacent column. The page headings and chart note ('SEER 8 5-Year Relative Survival Percent from 1975-2018') do not label the 'Year' column as year of diagnosis. The page does not describe changes over time in case mix, diagnostic criteria, staging or treatment and attributes the change to no cause; read beside the mortality values. Compare only with other values from this same SEER 8 series, not with the SEER 21 values. Group defined by tumour site, not histology: SEER's site recode table (ICD-O-3/WHO 2008; the Stat Facts page does not say which recode it uses) lists C40.0-C41.9 for 'Bones and Joints', excluding some morphology codes. The page names osteosarcoma, chondrosarcoma and the Ewing sarcoma family as common types and gives no figure by histology.

    Source

  • 5-year relative survival · United States, SEER 8 registries

    • 75.53%

      1993 (observed value; highest observed value in the table)

    All Races, Both Sexes (the chart note's label); the page states no age range for the survival series; SEER Cancer Stat Facts category 'Bone and Joint Cancer'. SEER 8 five-year relative survival series, 1975-2018.

    Modeled trend value for 1993: 69.24%. Shown because the observed series is noisy: single-year values move by many percentage points. The SEER registry groupings page describes SEER 8 as the grouping available for cases diagnosed from 1975. Relative survival excludes deaths from other causes. 'Observed' is the table's label for the actual yearly value, as opposed to the Joinpoint modeled-trend value in the adjacent column. The page headings and chart note ('SEER 8 5-Year Relative Survival Percent from 1975-2018') do not label the 'Year' column as year of diagnosis. The page does not describe changes over time in case mix, diagnostic criteria, staging or treatment and attributes the change to no cause; read beside the mortality values. Compare only with other values from this same SEER 8 series, not with the SEER 21 values. Group defined by tumour site, not histology: SEER's site recode table (ICD-O-3/WHO 2008; the Stat Facts page does not say which recode it uses) lists C40.0-C41.9 for 'Bones and Joints', excluding some morphology codes. The page names osteosarcoma, chondrosarcoma and the Ewing sarcoma family as common types and gives no figure by histology.

    Source

  • 5-year relative survival · United States, SEER 8 registries

    • 68.57%

      2018 (observed value; last observed year in the table)

    All Races, Both Sexes (the chart note's label); the page states no age range for the survival series; SEER Cancer Stat Facts category 'Bone and Joint Cancer'. SEER 8 five-year relative survival series, 1975-2018.

    Modeled trend value for 2018: 70.89%; the table shows no observed five-year value for 2019 onward. The SEER registry groupings page describes SEER 8 as the grouping available for cases diagnosed from 1975. Relative survival excludes deaths from other causes. 'Observed' is the table's label for the actual yearly value, as opposed to the Joinpoint modeled-trend value in the adjacent column. The page headings and chart note ('SEER 8 5-Year Relative Survival Percent from 1975-2018') do not label the 'Year' column as year of diagnosis. The page does not describe changes over time in case mix, diagnostic criteria, staging or treatment and attributes the change to no cause; read beside the mortality values. Compare only with other values from this same SEER 8 series, not with the SEER 21 values. Group defined by tumour site, not histology: SEER's site recode table (ICD-O-3/WHO 2008; the Stat Facts page does not say which recode it uses) lists C40.0-C41.9 for 'Bones and Joints', excluding some morphology codes. The page names osteosarcoma, chondrosarcoma and the Ewing sarcoma family as common types and gives no figure by histology.

    Source

  • 5-year relative survival · United States, SEER 21 registries (excluding IL)

    • 68.7%

      2016-2022

    SEER Cancer Stat Facts category 'Bone and Joint Cancer'; the headline value, which the page does not label with age, sex, race or stage. The page gives no stage breakdown for this site.

    Separate series from the SEER 8 values: SEER 21 (excluding IL), a different grouping of registries; the SEER registry groupings page says SEER 21 has a more limited set of available variables and is only available for limited statistics. Not comparable with the SEER 8 yearly values. Relative survival excludes deaths from other causes. The page does not say whether the period is year of diagnosis. Group defined by tumour site, not histology: SEER's site recode table (ICD-O-3/WHO 2008; the Stat Facts page does not say which recode it uses) lists C40.0-C41.9 for 'Bones and Joints', excluding some morphology codes. The page names osteosarcoma, chondrosarcoma and the Ewing sarcoma family as common types and gives no figure by histology.

    Source

  • Age-standardised mortality · United States

    • 0.86per 100,000

      1975 (observed value; first year in the table)

    All Races, Both Sexes (the table note); age-adjusted death rate for 'Bone and Joint Cancer'; U.S. Mortality series, 1975-2024.

    Modeled trend value for 1975: 0.88. Cause-of-death data from U.S. Mortality (National Center for Health Statistics data as used by SEER), not SEER registry data; the page does not describe how its cause-of-death category maps to the registry site group. Rates are age-adjusted per 100,000. The page does not describe screening for this cancer; read it beside the SEER 8 survival values, which are a separate series.

    Source

  • Age-standardised mortality · United States

    • 0.41per 100,000

      2002 (observed value; lowest observed value in the table)

    All Races, Both Sexes (the table note); age-adjusted death rate for 'Bone and Joint Cancer'; U.S. Mortality series, 1975-2024.

    Modeled trend value for 2002: 0.44. Cause-of-death data from U.S. Mortality (National Center for Health Statistics data as used by SEER), not SEER registry data; the page does not describe how its cause-of-death category maps to the registry site group. Rates are age-adjusted per 100,000. The page does not describe screening for this cancer; read it beside the SEER 8 survival values, which are a separate series.

    Source

  • Age-standardised mortality · United States

    • 0.52per 100,000

      2024 (observed value; last year in the table)

    All Races, Both Sexes (the table note); age-adjusted death rate for 'Bone and Joint Cancer'; U.S. Mortality series, 1975-2024. The page's 2020-2024 average is 0.5 per 100,000.

    Modeled trend value for 2024: 0.53. The page text states that age-adjusted death rates have been rising on average 1.7% each year over 2015-2024 (and incidence rising 0.4% a year over 2014-2023); it gives no cause. Cause-of-death data from U.S. Mortality (National Center for Health Statistics data as used by SEER), not SEER registry data; the page does not describe how its cause-of-death category maps to the registry site group. Rates are age-adjusted per 100,000. The page does not describe screening for this cancer; read it beside the SEER 8 survival values, which are a separate series.

    Source

Survival figures are affected by earlier detection (lead time) and overdiagnosis, and by changes in who is diagnosed. For screened cancers, read them alongside mortality.

Evidence chains

A lane follows one defined population and setting, where each new trial tests against what was then standard. Compare only within a lane, and read every comparison as illustrative.

Resectable osteosarcoma, chemotherapy around surgery

People with resectable high-grade osteosarcoma of the extremity or limb girdle receiving chemotherapy after surgery (the 1986 trial: nonmetastatic extremity disease after definitive surgery; INT-0133: without clinically detectable metastases; EURAMOS-1: aged 40 or younger) · Adjuvant · endpoint: EFS

Results are not comparable across trials.

The 1986 trial randomised 36 people to chemotherapy or observation after surgery and reported two-year relapse-free survival; the later trials gave chemotherapy in every arm and tested additions (INT-0133: ifosfamide and/or mifamurtide added to cisplatin, doxorubicin and methotrexate, three-year event-free survival). Endpoints and time points therefore differ, and only one can share the chart axis; the other is listed beside it. For INT-0133 the drawn pair is standard chemotherapy versus standard plus mifamurtide from the 2005 report, in which the authors could not analyse by factorial design because of an interaction; the ifosfamide arms are not drawn, and the 2008 reanalysis of mifamurtide versus no mifamurtide (overall survival) reported a hazard ratio without arm values, so it is not drawn. EURAMOS-1 and JCOG0905 reported hazard ratios without arm medians or landmark percentages and are not drawn. The single-arm 1982 neoadjuvant series is not in this lane. Eras, age limits and response-based randomisation differ. Compare only each trial's own arms.

Each pair shows one trial: its control arm and its experimental arm. Yesterday's experimental arm is often today's control. Populations, definitions and follow-up differ.

020406080100EFS at 24 months (%)1986link-osteosarcoma-adjuvant-1986
  • experimental arm
  • control arm
  • different endpoint from the lane's
  • step line through experimental values
EFS at 24 months, rate, across 1 trial in this lane. link-osteosarcoma-adjuvant-1986 (1986): experimental 66 % vs control 17 %.
Data table
TrialYearEndpointControl (%)Experimental (%)
link-osteosarcoma-adjuvant-19861986RFS (differs from lane)17 %66 %
INT-01332005EFS (not plotted: different measure or time point)71 %68 %

Localized Ewing sarcoma, first-line chemotherapy

People with previously untreated localized Ewing sarcoma (IESS-II: nonpelvic bone; INT-0091: nonmetastatic stratum, aged 30 or younger; AEWS0031: localized extradural, younger than 50) · Localised · endpoint: EFS

Results are not comparable across trials.

All three entries are percentages at five years, but the endpoint names differ: IESS-II reported disease-free survival (drawn with a distinct marker), while INT-0091 and AEWS0031 reported event-free survival, and INT-0091 gave its values as mean with standard error. The comparisons differ in kind: IESS-II compared two schedules of the same four drugs (treatment 2 is the comparator only by convention; the abstracts name no control), INT-0091 added ifosfamide and etoposide, and AEWS0031 compared the same chemotherapy every 14 or 21 days. The IESS-I randomisation pairs differed by institution group, so it is not drawn. AEWS0031's 10-year update is a separate result and is not drawn. Eras, local-therapy practice, age limits and eligibility differ. Compare only each trial's own arms.

Each pair shows one trial: its control arm and its experimental arm. Yesterday's experimental arm is often today's control. Populations, definitions and follow-up differ.

020406080100EFS at 60 months (%)198120032012IESS-IIINT-0091AEWS0031
  • experimental arm
  • control arm
  • different endpoint from the lane's
  • step line through experimental values
EFS at 60 months, rate, across 3 trials in this lane. IESS-II (1981): experimental 68 % vs control 48 %; INT-0091 (2003): experimental 69 % vs control 54 %; AEWS0031 (2012): experimental 73 % vs control 65 %.
Data table
TrialYearEndpointControl (%)Experimental (%)
IESS-II1981DFS (differs from lane)48 %68 %
INT-00912003EFS54 %69 %
AEWS00312012EFS65 %73 %

Advanced GIST, imatinib dose

People with unresectable or metastatic GIST treated with imatinib (S0033: KIT-expressing; EORTC 62005 and the 400 versus 600 mg phase 2 trial: advanced disease) · Metastatic · endpoint: PFS

Results are not comparable across trials.

Only S0033 gives median progression-free survival for each arm (400 mg daily versus 400 mg twice daily; median follow-up 4.5 years), and its abstract reports no significant difference. EORTC 62005 and the pooled analysis of both trials report hazard ratios without arm medians and the phase 2 trial of 400 versus 600 mg reported response figures pooled across doses, so none of these are drawn. The 2001 to 2002 milestone concerns response rates and an accelerated approval, not progression-free survival. Compare only each trial's own arms.

Each pair shows one trial: its control arm and its experimental arm. Yesterday's experimental arm is often today's control. Populations, definitions and follow-up differ.

0510152025PFS, median (months)2004s0033-gist
  • experimental arm
  • control arm
  • different endpoint from the lane's
  • step line through experimental values
PFS, median, across 1 trial in this lane. s0033-gist (2004): experimental 20 months vs control 18 months.
Data table
TrialYearEndpointControl (months)Experimental (months)
s0033-gist2004PFS18 months20 months

Resected GIST, adjuvant imatinib

People with KIT-positive GIST after surgery (Z9001: complete gross resection of a primary tumour of at least 3 cm; SSG XVIII/AIO: high estimated risk of recurrence; EORTC 62024: intermediate or high risk after R0-R1 surgery) · Adjuvant · endpoint: RFS

Results are not comparable across trials.

The comparisons differ: Z9001 compared one year of imatinib with placebo, SSG XVIII/AIO compared 36 with 12 months of imatinib (so the control arm also received imatinib), and EORTC 62024 compared two years of imatinib with observation. Risk definitions differ, and the EORTC 62024 primary endpoint was changed to imatinib failure-free survival, with recurrence-free survival a secondary endpoint. The drawn values are recurrence-free survival at five years for SSG XVIII/AIO and EORTC 62024; Z9001 reported one-year recurrence-free survival at a median follow-up of 19.7 months, which is a different time point and is listed beside the chart rather than drawn on it. Compare only each trial's own arms.

Each pair shows one trial: its control arm and its experimental arm. Yesterday's experimental arm is often today's control. Populations, definitions and follow-up differ.

020406080100RFS at 60 months (%)20082008ssg-xviii-aioeortc-62024-gist
  • experimental arm
  • control arm
  • different endpoint from the lane's
  • step line through experimental values
RFS at 60 months, rate, across 2 trials in this lane. ssg-xviii-aio (2008): experimental 65.6 % vs control 47.9 %; eortc-62024-gist (2008): experimental 70 % vs control 63 %.
Data table
TrialYearEndpointControl (%)Experimental (%)
ssg-xviii-aio2008RFS47.9 %65.6 %
eortc-62024-gist2008RFS63 %70 %
Z90012008RFS (not plotted: different measure or time point)83 %98 %

Advanced GIST after imatinib, placebo-controlled trials

People with advanced GIST after kinase inhibitor therapy (sunitinib trial: resistant to or intolerant of imatinib; GRID: failure of imatinib and sunitinib; INVICTUS: progression on imatinib, sunitinib and regorafenib, or intolerance) · Relapsed / refractory · endpoint: PFS

Results are not comparable across trials.

Each trial enrolled people at a different treatment line, so the populations differ; all three were placebo-controlled with a 2:1 allocation. The sunitinib trial measured time to tumour progression and its trial record has no arm values, so it is not drawn. GRID and INVICTUS reported median progression-free survival by independent central review; placebo patients could cross over in both, and INVICTUS follow-up differed between arms. The single-arm NAVIGATOR study of avapritinib has no comparator and is not drawn. Compare only each trial's own arms; the chart is not a pooled or network estimate.

Each pair shows one trial: its control arm and its experimental arm. Yesterday's experimental arm is often today's control. Populations, definitions and follow-up differ.

0246810PFS, median (months)20122020GRIDINVICTUS
  • experimental arm
  • control arm
  • different endpoint from the lane's
  • step line through experimental values
PFS, median, across 2 trials in this lane. GRID (2012): experimental 4.8 months vs control 0.9 months; INVICTUS (2020): experimental 6.3 months vs control 1 months.
Data table
TrialYearEndpointControl (months)Experimental (months)
GRID2012PFS0.9 months4.8 months
INVICTUS2020PFS1 months6.3 months

Advanced soft-tissue sarcoma, first-line doxorubicin

Adults with unresectable locally advanced or metastatic soft-tissue sarcoma without prior anthracycline treatment (EORTC 62012: aged 18 to 60 with high-grade disease; olaratumab trials: anthracycline-naive adults) · Metastatic · endpoint: OS

Results are not comparable across trials.

The control arm in every entry is doxorubicin (alone, or with placebo in ANNOUNCE), but the experimental arms differ in kind: an intensified doxorubicin plus ifosfamide regimen in EORTC 62012 and an antibody added to doxorubicin in the olaratumab trials. The olaratumab phase 2 trial was open-label with 133 randomised patients and progression-free survival as its primary endpoint, with performance status 0 to 2; ANNOUNCE was double-blind, enrolled 509 people with performance status 0 to 1 and had overall survival as a dual primary endpoint; EORTC 62012 was restricted to ages 18 to 60. ANNOUNCE's leiomyosarcoma subgroup is a separate result and is not drawn. Compare only each trial's own arms.

Each pair shows one trial: its control arm and its experimental arm. Yesterday's experimental arm is often today's control. Populations, definitions and follow-up differ.

05101520253035OS, median (months)201420192019EORTC 62012olaratumab-phase-1b-2ANNOUNCE
  • experimental arm
  • control arm
  • different endpoint from the lane's
  • step line through experimental values
OS, median, across 3 trials in this lane. EORTC 62012 (2014): experimental 14.3 months vs control 12.8 months; olaratumab-phase-1b-2 (2019): experimental 26.5 months vs control 14.7 months; ANNOUNCE (2019): experimental 20.4 months vs control 19.7 months.
Data table
TrialYearEndpointControl (months)Experimental (months)
EORTC 620122014OS12.8 months14.3 months
olaratumab-phase-1b-22019OS14.7 months26.5 months
ANNOUNCE2019OS19.7 months20.4 months

Liposarcoma and leiomyosarcoma after anthracycline, versus dacarbazine

Adults with advanced liposarcoma or leiomyosarcoma previously treated with an anthracycline (trabectedin trial: and at least one further regimen; eribulin trial: at least two previous systemic regimens) · Relapsed / refractory · endpoint: OS

Results are not comparable across trials.

Dacarbazine is the control arm in both trials, but the trabectedin trial allocated 2:1 and its abstract gives no dacarbazine dose, while the eribulin trial allocated 1:1 with centre-dependent dacarbazine dosing and was open-label. The trabectedin figure is the interim overall survival analysis (the primary endpoint, 64% of observations censored); the later final analysis (577 randomised) and the progression-free survival result are separate and not drawn. The FDA indication for eribulin is liposarcoma only, and the eribulin abstract gives no results by histology. The placebo-controlled pazopanib trial in non-adipocytic sarcoma (PALETTE) has a different control and population and is not in this lane. Compare only each trial's own arms; the two trials' medians should not be set side by side.

Each pair shows one trial: its control arm and its experimental arm. Yesterday's experimental arm is often today's control. Populations, definitions and follow-up differ.

05101520OS, median (months)20152016ET743-SAR-3007Study 309
  • experimental arm
  • control arm
  • different endpoint from the lane's
  • step line through experimental values
OS, median, across 2 trials in this lane. ET743-SAR-3007 (2015): experimental 12.4 months vs control 12.9 months; Study 309 (2016): experimental 13.5 months vs control 11.5 months.
Data table
TrialYearEndpointControl (months)Experimental (months)
ET743-SAR-30072015OS12.9 months12.4 months
Study 3092016OS11.5 months13.5 months

Standard of care, then and now

How the usual sequence of treatment changed for one setting. Choose a point in time; each step links to the milestone behind it.

Advanced GIST: kinase inhibitor sequence

People with unresectable or metastatic gastrointestinal stromal tumour · Metastatic

  1. Targeted1ImatinibImatinib in GIST: case report, phase 2 and FDA approval (2001)
  2. Targeted2Sunitinib (after imatinib)Sunitinib after imatinib failure in GIST (2006)
  3. Targeted3Regorafenib (after imatinib, sunitinib)GRID: regorafenib after imatinib and sunitinib (2012)
  4. Targeted4Ripretinib (after three or more inhibitors)Avapritinib (PDGFRA exon 18) and ripretinib (fourth line) (2020)
  5. Targeted5Avapritinib (PDGFRA exon 18, incl. D842V)Avapritinib (PDGFRA exon 18) and ripretinib (fourth line) (2020)
2020. Based only on the cited trials; guidelines vary. The FDA approved avapritinib on 9 January 2020 (single-arm study) and ripretinib on 15 May 2020 (placebo-controlled trial).
Text version
  • 2002: Imatinib
  • 2006: Imatinib → Sunitinib (after imatinib)
  • 2013: Imatinib → Sunitinib (after imatinib) → Regorafenib (after imatinib, sunitinib)
  • 2020: Imatinib → Sunitinib (after imatinib) → Regorafenib (after imatinib, sunitinib) → Ripretinib (after three or more inhibitors) → Avapritinib (PDGFRA exon 18, incl. D842V)

Localized Ewing sarcoma: chemotherapy schedule

People with newly diagnosed localized Ewing sarcoma · Localised

  1. Chemotherapy1Four-drug chemotherapy with radiotherapyIntergroup Ewing's Sarcoma Studies: multimodal therapy (1981)
  2. Chemotherapy2Alternating ifosfamide and etoposideINT-0091: adding ifosfamide and etoposide in Ewing sarcoma (2003)
  3. Chemotherapy3Same cycles every 14 days, not 21AEWS0031: interval-compressed chemotherapy in Ewing sarcoma (2012)
2012. Based only on the cited trials; guidelines vary. AEWS0031 compared the same alternating chemotherapy every 14 or 21 days; the authors reported higher five-year event-free survival every 14 days with similar toxicity.
Text version
  • 1990: Four-drug chemotherapy with radiotherapy
  • 2003: Four-drug chemotherapy with radiotherapy → Alternating ifosfamide and etoposide
  • 2012: Four-drug chemotherapy with radiotherapy → Alternating ifosfamide and etoposide → Same cycles every 14 days, not 21

Reversals and dead ends

Approaches that were abandoned or withdrawn, and trials whose results were later questioned, are part of the story.

  • 2016

    EURAMOS-1: response-adapted escalation did not improve EFS

    EURAMOS-1 tested changing osteosarcoma treatment after surgery according to how well the tumour had responded to earlier chemotherapy. For poor responders, adding two more drugs did not improve event-free survival and caused more side effects. For good responders, adding maintenance interferon was not statistically different from chemotherapy alone.

    EURAMOS-1 (2260 registered, 1334 randomised). Poor responders (618): MAP vs MAPIE, EFS HR 0.98 (95% CI 0.78-1.23), non-proportional hazards; grade 4 non-haematological toxicity 24% vs 12% with MAPIE. Good responders (716): MAP vs MAP plus pegylated interferon alfa-2b, EFS HR 0.83 (0.61-1.12; P=.214). JCOG0905 (51 vs 52 poor responders): DFS HR 1.05 (0.55-1.98), OS HR 1.48 (0.68-3.22).

  • 2019

    Olaratumab: phase 2 survival benefit not confirmed in ANNOUNCE

    A randomised phase 2 trial in 133 people reported median survival of 26.5 months with olaratumab plus doxorubicin versus 14.7 months with doxorubicin alone, and the drug received US accelerated approval in 2016. The phase 3 ANNOUNCE trial in 509 people found no difference (20.4 versus 19.7 months), and the European authorisation was revoked in 2019.

    Phase 2 (N=133 randomised): median OS 26.5 vs 14.7 months (HR 0.46; 95% CI 0.30-0.71); PFS HR 0.67 (95% CI 0.44-1.02; p=0.0615). Phase 3 ANNOUNCE (N=509; doxorubicin plus olaratumab vs plus placebo): median OS 20.4 vs 19.7 months (HR 1.05; 95% CI 0.84-1.30; p=.69); leiomyosarcoma 21.6 vs 21.9 (HR 0.95). FDA accelerated approval 19 October 2016; EC revocation 22 July 2019; FDA withdrawal listed 25 February 2020.

  • 2020

    STRASS: preoperative radiotherapy in retroperitoneal sarcoma

    In 266 adults with a newly diagnosed retroperitoneal sarcoma that could be operated on, adding radiotherapy before surgery did not lengthen the time free of abdominal recurrence: median 4.5 years with radiotherapy plus surgery and 5.0 years with surgery alone. The authors concluded that preoperative radiotherapy should not be standard.

    Open-label phase 3 (EORTC-62092 STRASS; N=266; 133 per arm, 1:1): preoperative radiotherapy (50.4 Gy in 28 fractions) plus surgery vs surgery alone. Median follow-up 43.1 months. Median abdominal recurrence-free survival (primary, ITT) 4.5 vs 5.0 years (HR 1.01; 95% CI 0.71-1.44; log-rank p=0.95). Grade 3-4 lymphopenia 77% vs 1%; serious adverse events 24% vs 10%.

  • 2026

    Tazemetostat: epithelioid sarcoma accelerated approval withdrawn

    Tazemetostat, a tablet that blocks the EZH2 protein, received US accelerated approval in January 2020 for advanced epithelioid sarcoma after a single-arm study in which 9 of 62 people (15%) responded. In March 2026 the manufacturer withdrew it after a safety signal in a lymphoma trial; the FDA lists the sarcoma approval as withdrawn on 22 June 2026.

    Single-arm phase 2 basket study, epithelioid sarcoma cohort (N=62, INI1/SMARCB1 loss, tazemetostat 800 mg twice daily): objective response 15% (9 of 62; 95% CI 7-26; primary); median PFS 5.5 months; median OS 19.0 months; median follow-up 13.8 months. FDA accelerated approval 23 January 2020. Ipsen withdrew the product 9 March 2026 (safety signal in SYMPHONY-1); FDA table: withdrawal 22 June 2026.

Today and tomorrow

Standard of care today

Adjuvant

Resectable high-grade osteosarcoma (EURAMOS-1 enrolled patients aged 40 years or younger). Postoperative chemotherapy adapted to histologic response was tested in randomised trials. In EURAMOS-1, poor responders (618 randomised) had an event-free survival hazard ratio of 0.98 (95% CI 0.78 to 1.23) for adding ifosfamide and etoposide to MAP (methotrexate, doxorubicin, cisplatin), with more toxicity and no event-free survival gain; good responders (716 randomised) had a hazard ratio of 0.83 (95% CI 0.61 to 1.12; P = .214) for adding maintenance pegylated interferon alfa-2b to MAP, not statistically different at the preplanned analysis. In JCOG0905 (51 and 52 poor responders randomised), disease-free survival hazard ratio for adding ifosfamide was 1.05 (95% CI 0.55 to 1.98) and overall survival was numerically inferior (hazard ratio 1.48; 95% CI 0.68 to 3.22). Regulatory action read:

  • the EMA Mepact (mifamurtide) page shows a marketing authorisation issued on 6 March 2009 for high-grade resectable non-metastatic osteosarcoma in children, adolescents and young adults
  • the page does not state the authorisation type, and no FDA document on mifamurtide was opened. Guideline texts were not opened
  • they are cited by identifier only and no recommendation is attributed to them. ESMO-EURACAN-GENTURIS-ERN PaedCan bone sarcomas 2021 (Ann Oncol, PMID 34500044) and NCCN Bone Cancer version 2.2025 (J Natl Compr Canc Netw, PMID 40203873
  • a later version may be current) are the identifiers
Marina NM, Smeland S, Bielack SS, Bernstein M, Jovic G, Krailo MD, et al 2016Bielack SS, Smeland S, Whelan JS, Marina N, Jovic G, Hook JM, et al 2015Hiraga H, Machida R, Kawai A, Kunisada T, Yonemoto T, Endo M, et al 2025European Medicines Agency 2009Guideline: Strauss SJ, Frezza AM, Abecassis N, Bajpai J, Bauer S, Biagini R, et al 2021Guideline: Biermann JS, Hirbe A, Ahlawat S, Bernthal NM, Binitie O, Boles S, et al 2025
Localised
Ewing sarcoma, localized extradural disease in patients younger than 50 years (AEWS0031). The same alternating chemotherapy (vincristine, doxorubicin and cyclophosphamide with ifosfamide and etoposide) was given every 14 days or every 21 days after randomisation of 587 patients (568 eligible). Event-free survival at a median of 5 years was 73% with 14-day cycles and 65% with 21-day cycles (P = .048), with similar toxicity; the 10-year update, which excluded one patient, reported event-free survival of 70% versus 61% (P = .03) and overall survival of 76% versus 69% (P = .04). No FDA or EMA action for this setting was read, and no European trial result is stated here. Guideline texts were not opened; they are cited by identifier only and no recommendation is attributed to them. The ESMO bone sarcomas guideline (PMID 34500044) is the identifier.
Womer RB, West DC, Krailo MD, Dickman PS, Pawel BR, Grier HE, et al 2012Cash T, Krailo MD, Buxton AB, Pawel BR, Healey JH, Binitie O, et al 2023Guideline: Strauss SJ, Frezza AM, Abecassis N, Bajpai J, Bauer S, Biagini R, et al 2021Guideline: Biermann JS, Hirbe A, Ahlawat S, Bernthal NM, Binitie O, Boles S, et al 2025
Neoadjuvant

Localised soft-tissue sarcoma, preoperative treatment. Retroperitoneal sarcoma:

  • in the randomised phase 3 trial STRASS (266 patients), median abdominal recurrence-free survival was 4.5 years with preoperative radiotherapy plus surgery and 5.0 years with surgery alone (hazard ratio 1.01
  • 95% CI 0.71 to 1.44
  • P = 0.95), so the trial did not show a benefit of preoperative radiotherapy on its primary endpoint
  • the randomised question of preoperative chemotherapy is open (STRASS 2, see the horizon). Extremity disease: SU2C-SARC032, an open-label randomised trial registered as phase 2 in stage III undifferentiated pleomorphic sarcoma or pleomorphic or dedifferentiated liposarcoma (143 randomised
  • primary analysis in 127), reported a disease-free survival hazard ratio of 0.61 (90% CI 0.39 to 0.96
  • one-sided log-rank P = 0.035) for adding pembrolizumab to preoperative radiotherapy and surgery, with 2-year disease-free survival of 67% versus 52% and grade 3 or higher adverse events in 56% versus 31%. No FDA or EMA action for these uses was read. Guideline texts were not opened
  • they are cited by identifier only and no recommendation is attributed to them. ESMO-EURACAN-GENTURIS soft tissue and visceral sarcomas 2021 (PMID 34303806) and NCCN Soft Tissue Sarcoma version 1.2025 (PMID 41671462) are the identifiers
Bonvalot S, Gronchi A, Le Péchoux C, Swallow CJ, Strauss D, Meeus P, et al 2020Mowery YM, Ballman KV, Hong AM, Schuetze SM, Wagner AJ, Monga V, et al 2024Guideline: Gronchi A, Miah AB, Dei Tos AP, Abecassis N, Bajpai J, Bauer S, et al 2021Guideline: von Mehren M, Kane JM 3rd, Armstrong SA, Balach T, Bishop AJ, Buehler D, et al 2025
Metastatic

Advanced or metastatic soft-tissue sarcoma, first-line systemic therapy. In EORTC 62012 (455 patients aged 18 to 60 years with high-grade disease), median overall survival was 12.8 months with doxorubicin and 14.3 months with intensified doxorubicin plus ifosfamide (hazard ratio 0.83; 95.5% CI 0.67 to 1.03; P = 0.076); median progression-free survival was 4.6 versus 7.4 months and overall response 14% versus 26%, with more toxicity in the combination arm, and the authors did not support the combination for palliation unless tumour shrinkage is the specific goal. In ANNOUNCE (509 anthracycline-naive patients) olaratumab added to doxorubicin gave median overall survival of 20.4 versus 19.7 months (hazard ratio 1.05; 95% CI 0.84 to 1.30; P = .69). FDA actions read:

  • olaratumab accelerated approval on 19 October 2016 (BLA 761038 letter)
  • the FDA table of withdrawn accelerated approvals lists withdrawal on 25 February 2020. EMA EPAR page: conditional authorisation issued 9 November 2016, revocation printed 22 July 2019. Guideline texts were not opened
  • they are cited by identifier only and no recommendation is attributed to them. ESMO-EURACAN-GENTURIS 2021 (PMID 34303806) and NCCN Soft Tissue Sarcoma version 1.2025 (PMID 41671462) are the identifiers
Judson I, Verweij J, Gelderblom H, Hartmann JT, Schöffski P, Blay JY, et al 2014Tap WD, Wagner AJ, Schöffski P, Martin-Broto J, Krarup-Hansen A, Ganjoo KN, et al 2020US Food and Drug Administration 2016US Food and Drug Administration 2026European Medicines Agency 2016Guideline: Gronchi A, Miah AB, Dei Tos AP, Abecassis N, Bajpai J, Bauer S, et al 2021Guideline: von Mehren M, Kane JM 3rd, Armstrong SA, Balach T, Bishop AJ, Buehler D, et al 2025
Relapsed / refractory

Previously treated advanced soft-tissue sarcoma. FDA actions read in the approval letters:

  • pazopanib on 26 April 2012 for advanced soft tissue sarcoma after prior chemotherapy (the letter's limitation of use: efficacy in adipocytic soft tissue sarcoma or gastrointestinal stromal tumour not demonstrated)
  • trabectedin on 23 October 2015 for unresectable or metastatic liposarcoma or leiomyosarcoma after a prior anthracycline-containing regimen
  • eribulin on 28 January 2016 for unresectable or metastatic liposarcoma after a prior anthracycline-containing regimen. None of the three letters uses accelerated-approval wording. The EMA Yondelis page shows trabectedin authorised on 17 September 2007 under exceptional circumstances, which it states ended on 27 May 2015. Tazemetostat received FDA accelerated approval on 23 January 2020 for epithelioid sarcoma
  • Ipsen announced voluntary withdrawal of Tazverik in all indications from all Ipsen markets on 9 March 2026, and the FDA table lists the epithelioid sarcoma withdrawal on 22 June 2026 (the table gives no reason). Afamitresgene autoleucel received FDA accelerated approval on 2 August 2024 for unresectable or metastatic synovial sarcoma after prior chemotherapy in adults positive for HLA-A*02:01P, -A*02:02P, -A*02:03P or -A*02:06P whose tumours express MAGE-A4, on the basis of the single-arm SPEARHEAD-1 cohort 1 (44 patients infused
  • overall response 43.2%, 95% CI 28.4 to 59.0, by independent review)
  • the FDA lists of verified-clinical-benefit and withdrawn accelerated approvals, read on 21 October 2026, show no row for it. Guideline texts were not opened
  • they are cited by identifier only and no recommendation is attributed to them. ESMO 2021 (PMID 34303806) and NCCN version 1.2025 (PMID 41671462) are the identifiers
US Food and Drug Administration 2012US Food and Drug Administration 2015US Food and Drug Administration 2016European Medicines Agency 2007US Food and Drug Administration 2020Ipsen 2026US Food and Drug Administration 2026US Food and Drug Administration 2024US Food and Drug Administration 2026Guideline: Gronchi A, Miah AB, Dei Tos AP, Abecassis N, Bajpai J, Bauer S, et al 2021Guideline: von Mehren M, Kane JM 3rd, Armstrong SA, Balach T, Bishop AJ, Buehler D, et al 2025
Adjuvant

Resected KIT-positive GIST at significant risk of recurrence. FDA actions read:

  • imatinib adjuvant accelerated approval on 19 December 2008 (letter for supplement S-025) and a corrected letter of 31 January 2012 stating that approval of the supplement fulfils the accelerated-approval requirements ('SUBPART H FULFILLED'
  • the corrected letter's opening describes the supplement as proposing 'conversion of accelerated approval to full approval')
  • FDA's verified-clinical-benefit list (content current as of 16 October 2026) gives 12/19/2008 and 1/31/2012 as the accelerated and traditional approval dates. Randomised evidence: in the double-blind placebo-controlled trial Z9001, one-year recurrence-free survival was 98% with 1 year of imatinib and 83% with placebo (hazard ratio 0.35
  • 95% CI 0.22 to 0.53)
  • in SSG XVIII, 36 versus 12 months of imatinib gave 5-year recurrence-free survival of 65.6% versus 47.9% (hazard ratio 0.46
  • 95% CI 0.32 to 0.65
  • P < .001) and 5-year overall survival of 92.0% versus 81.7% (hazard ratio 0.45
  • 95% CI 0.22 to 0.89
  • P = .02), and the 10-year update reported hazard ratios of 0.66 (95% CI 0.49 to 0.87) for recurrence-free survival and 0.55 (95% CI 0.37 to 0.83) for overall survival. The duration wording of the current label was not read. Guideline texts were not opened
  • they are cited by identifier only and no recommendation is attributed to them. ESMO-EURACAN-GENTURIS GIST (Ann Oncol 2022, PMID 34560242) and NCCN Guidelines Insights GIST version 2.2022 (PMID 36351335
  • a later version may be current) are the identifiers
Dematteo RP, Ballman KV, Antonescu CR, Maki RG, Pisters PW, Demetri GD, et al 2009Joensuu H, Eriksson M, Sundby Hall K, Hartmann JT, Pink D, Schütte J, et al 2012Joensuu H, Eriksson M, Sundby Hall K, Reichardt A, Hermes B, Schütte J, et al 2020U 2008U 2012US Food and Drug Administration 2026Guideline: Casali PG, Blay JY, Abecassis N, Bajpai J, Bauer S, Biagini R, et al 2022Guideline: von Mehren M, Kane JM, Riedel RF, Sicklick JK, Pollack SM, Agulnik M, et al 2022
Metastatic

Advanced (unresectable or metastatic) GIST, by line. FDA actions read in the approval letters:

  • imatinib on 1 February 2002 (accelerated approval
  • the FDA verified-clinical-benefit list gives 9/26/2008 as the traditional approval date, and a 26 September 2008 letter states that its approval fulfilled a postmarketing commitment of the 2002 letter without using conversion wording)
  • sunitinib on 26 January 2006 after progression on or intolerance to imatinib
  • regorafenib on 25 February 2013 after imatinib and sunitinib
  • avapritinib on 9 January 2020 for PDGFRA exon 18 mutation including D842V
  • ripretinib on 15 May 2020 after three or more kinase inhibitors including imatinib. The regorafenib, avapritinib and ripretinib letters contain no accelerated-approval wording
  • the sunitinib letter approves the GIST application (NDA 21-938) without it and approves the separate renal cell carcinoma application (NDA 21-968) under accelerated approval. Randomised placebo-controlled results: GRID (199 patients) median progression-free survival 4.8 versus 0.9 months (hazard ratio 0.27
  • 95% CI 0.19 to 0.39) and INVICTUS (129 patients) 6.3 versus 1.0 months (hazard ratio 0.15
  • 95% CI 0.09 to 0.25)
  • avapritinib in PDGFRA D842V-mutant GIST rests on an open-label phase 1 study whose abstract names no comparison group (NAVIGATOR
  • objective response 88%, 95% CI 76 to 95). EMA EPAR pages: Ayvakyt (avapritinib) conditional marketing authorisation, issued 24 September 2020 (the page also covers systemic mastocytosis
  • its GIST indication is PDGFRA D842V-mutant disease)
  • Qinlock (ripretinib) authorised 18 November 2021 (the page does not state a type). EU decision dates for the imatinib, sunitinib and regorafenib GIST indications were not read. Guideline texts were not opened
  • they are cited by identifier only and no recommendation is attributed to them. ESMO GIST 2022 (PMID 34560242) and NCCN GIST version 2.2022 (PMID 36351335) are the identifiers
U 2002U 2008US Food and Drug Administration 2026U 2006U 2013U 2020U 2020Demetri GD, Reichardt P, Kang YK, Blay JY, Rutkowski P, Gelderblom H, et al 2013Blay JY, Serrano C, Heinrich MC, Zalcberg J, Bauer S, Gelderblom H, et al 2020Heinrich MC, Jones RL, von Mehren M, Schöffski P, Serrano C, Kang YK, et al 2020EMA 2026EMA 2026Guideline: Casali PG, Blay JY, Abecassis N, Bajpai J, Bauer S, Biagini R, et al 2022Guideline: von Mehren M, Kane JM, Riedel RF, Sicklick JK, Pollack SM, Agulnik M, et al 2022

Trials to watch

  • PEAK NCT05208047

    In locally advanced, unresectable or metastatic gastrointestinal stromal tumour, how does bezuclastinib (registry code CGT9486) plus sunitinib compare with sunitinib alone in progression-free survival (the registry's Part 2 primary outcome, time frame approximately 48 months)? Registry record: multi-part phase 3 study whose Part 2 is a randomised, open-label comparison, sponsor Cogent Biosciences, 482 estimated; its primary outcomes also include pharmacokinetic measures in Part 1 and a drug-interaction substudy. Registry status: recruiting; registry start 2022-04-14; registry primary completion date 2025-09-30 (listed as actual); last update posted 2026-06-16; no results posted.

  • StrateGIST 3 NCT07218926

    In metastatic or unresectable gastrointestinal stromal tumour after imatinib, how does IDRX-42 (GSK6042981) compare with sunitinib in progression-free survival (the registry's primary outcome, up to approximately 130 weeks)? Registry record: randomised, open-label phase 3, 450 estimated, sponsor GlaxoSmithKline. Registry status: recruiting; registry start 2025-12-03 (actual); registry primary completion date 2028-05-30 (estimated); last update posted 2026-10-05; no results posted.

  • STRASS 2 NCT04031677

    In high-risk retroperitoneal sarcoma (the registry summary names dedifferentiated liposarcoma and leiomyosarcoma), how does preoperative chemotherapy followed by surgery compare with surgery alone in disease-free survival (the registry's primary outcome, time frame 7 years from first patient in)? Registry record: randomised, open-label phase 3, 250 estimated, sponsor EORTC. Registry status: recruiting; registry start 2021-01-20 (actual); registry primary completion date 2027-04-21 (estimated); last update posted 2026-10-07; no results posted.

  • SaLuDo NCT06088290

    As first-line treatment of metastatic leiomyosarcoma, how does lurbinectedin plus doxorubicin compare with doxorubicin alone in progression-free survival by independent review committee (the registry's primary outcome)? Registry record: randomised, open-label phase 3 with two lurbinectedin-plus-doxorubicin dose arms and a doxorubicin arm, 469 enrolled (actual), sponsor PharmaMar. Registry status: active, not recruiting; registry start 2023-09-21 (actual); registry primary completion date 2029-08-30 (estimated); last update posted 2026-06-18; no results posted.

  • FOSTER NCT07871864

    In newly diagnosed high-grade osteosarcoma, the FOSTER evolving study platform registers patients at diagnosis (Part 1; the registry's first primary outcome is event-free survival at 3 years across treatment sequences) and its first randomised question, FOSTER-CabOS, compares maintenance cabozantinib with best supportive care after first-line chemotherapy in event-free survival. Registry record: phase 3, randomised, open-label, 530 estimated, sponsor Gustave Roussy. Registry status: recruiting; registry start 2026-10 (estimated); registry primary completion date 2035-01 (estimated); last update posted 2026-10-13; no results posted.

  • ARST2531 NCT07466316

    In newly diagnosed intermediate-risk rhabdomyosarcoma (patients 50 years or younger), how does VAC (vincristine, dactinomycin, cyclophosphamide) with a higher cyclophosphamide dose and intensity compare with lower-dose VAC/VI (adding irinotecan) plus maintenance therapy in event-free survival (the registry's primary outcome, assessed up to 5 years)? Registry record: randomised phase 3, 342 estimated, sponsor Children's Oncology Group. Registry status: recruiting; registry start 2026-07-14 (actual); registry primary completion date 2031-03-31 (estimated); last update posted 2026-08-25; no results posted.

References

Show all 95 references

Loading references… Open the list as a page

Every figure was checked against its primary source to the extent possible. Items that could not be fully confirmed are marked “needs review”. Educational history, not medical advice.

Report an error

Links open the primary source. A flag marked ⚑ needs review means a value or claim has not yet been independently confirmed.

2015TreatmentChemotherapy

majorRelapsed / refractory

Trabectedin vs dacarbazine in liposarcoma and leiomyosarcoma

In 518 adults with advanced liposarcoma or leiomyosarcoma that had progressed after earlier treatment, trabectedin delayed progression compared with dacarbazine (median 4.2 versus 1.5 months). Overall survival did not differ at the first analysis (12.4 versus 12.9 months) or at the final analysis, published later (13.7 versus 13.1 months).

Phase 3 (N=518, 2:1): trabectedin (n=345) vs dacarbazine (n=173) after an anthracycline and at least one further systemic regimen. Median PFS (final analysis) 4.2 vs 1.5 months (HR 0.55; P<.001). Interim OS, the primary endpoint (64% censored): 12.4 vs 12.9 months (HR 0.87; P=.37). Final OS in a later report (577 randomised): 13.7 vs 13.1 months (HR 0.93; P=.49). FDA approval 23 October 2015.

Standard before

In the EU, trabectedin had been authorised since 17 September 2007, under exceptional circumstances, for advanced soft-tissue sarcoma after anthracyclines and ifosfamide or when these were unsuitable; the EMA page states those circumstances ended on 27 May 2015.

What changed

On 23 October 2015 the FDA approved trabectedin for unresectable or metastatic liposarcoma or leiomyosarcoma in patients who received a prior anthracycline-containing regimen.

Why it mattered

The randomised trial showed longer progression-free survival against an active comparator but no overall survival difference; in the later final analysis the authors' sensitivity analyses suggest that post-study anticancer therapy, used in most patients in both arms, confounded overall survival.

Caveats

  • Overall survival was the primary endpoint and was not improved; the disease-control (PFS) benefit is a secondary result.
  • The interim OS hazard ratio carries no confidence interval in the abstract, and 64% of observations were censored at that analysis.
  • The primary paper reports 518 randomised patients and the final overall survival report 577, reflecting different data cutoffs; the registry enrolment figure differs again.
  • The eribulin trial used the same comparator drug with 1:1 allocation and centre-dependent dacarbazine dosing; the two trials' medians should not be set side by side.