OncoSignal

Case study · Ewing sarcoma (paediatric) · IGF1R · ganitumab, R1507 (teprotumumab)

IGF-1R antibodies in Ewing sarcoma

Every quote on this page was matched against its source by machine. The analysis has not yet been checked line by line by a person.

Failed trial NCT02306161 · set against NCT00563680, NCT00642941

How obvious was the failure: intermediate. Rule: A reason was stated (for example lack of efficacy or futility), but the sources do not say why a supported target gave no efficacy. The trial group reported no EFS difference (hazard ratio 1.00) and an early closure recommended on the observed outcomes and a possible pneumonitis increase, but it offered only possible explanations for why the target gave no benefit (inadequate tumour exposure, no predictive biomarker for enrichment) and established none of them.[20][21][24][25]

A five-minute explainer of this page, with captions. Every number in it comes from the checked sources below. The analysis has not yet been checked line by line by a person.

The short version

What the whole page adds up to, in 8 points. Every number is from a checked source quote; follow the link to its chart and the numbers to their sources.

  1. 1
    The hope came from small studies in patients whose disease had come back. R1507 shrank tumours in 10% of patients with recurrent or refractory Ewing sarcoma family tumours, and ganitumab gave partial responses in 6% of a pooled group.[1][8][11]See the chart
  2. 2
    The decisive trial asked a different question in different patients: would adding ganitumab to chemotherapy keep people with newly diagnosed, metastatic disease free of relapse for longer? It was sized to detect a hazard ratio of 0.67, about a third fewer events.[19][62]See the chart
  3. 3
    The case for the target rested on other evidence than human genetics. Open Targets 26.09 scores IGF1R in Ewing sarcoma on literature (0.83), with no genetic association or somatic mutation score.See the chart
  4. 4
    It found no difference. At 3 years, 37.4% of patients on chemotherapy alone and 39.1% with ganitumab were alive without relapse or a new cancer (hazard ratio 1.00, 95% CI 0.76 to 1.33). Overall survival was 59.5% against 56.7%.[20]See the chart
  5. 5
    The trial closed early, in March 2019, on the observed outcomes and a possible rise in pneumonitis (0.7% against 2.7%), before its formal inefficacy boundary had been crossed.[21][22]See the chart
  6. 6
    In the blood, the drug met its target level in 9 of 10 and then 10 of 10 patients tested. The trial group said it may still not have reached the tumour at an active level.[23][24]See the chart
  7. 7
    No validated biomarker existed to pick the patients most likely to respond, and a later commentary noted that the diagnosis had not been confirmed molecularly in 14% of patients.[25][46]See the chart
  8. 8
    No cause was established. The trial group wrote that the reasons are not clear, and offered possible ones rather than a finding.[34]See the chart

What a go/no-go reader can take from it. Our analysis. Not established fact.

  • A response rate in relapsed patients is not a forecast of an add-on survival benefit in newly diagnosed patients. Before a trial is sized on it, ask to see how one was turned into the other.
  • Blood levels that meet a target do not show the target was hit in the tumour. Ask for drug levels or receptor blockade in tumour tissue.
  • A target backed by the literature but not by human genetics in this cancer carries more risk, and that risk belongs in the price.
  • Without a biomarker, a trial tests the average patient. If early responses were rare, the average can show nothing even if a few patients benefit.

Ganitumab was meant to cut the rate of relapse or death by about a third. The rate did not change.

Hazard ratio: how often an event happened in patients given ganitumab with chemotherapy, compared with chemotherapy alone, at any point during follow-up. 1.00 means the same rate in both groups; 0.67 would mean about 33% lower.

How to read this: Each filled dot is the trial's estimate. The bar around it is the 95% confidence interval: how precisely the trial pinned the ratio down, not how much patients differed. Left of the dashed line favours ganitumab; right of it favours chemotherapy alone. The hollow ring is what the trial was designed to detect.

Relapse, new cancer or death
event-free survival
Planned 0.67Result 1.0095% CI 0.76 to 1.33
Death
overall survival
Result 1.0395% CI 0.75 to 1.42
0.500.751.001.251.50Fewer events with ganitumabMore events with ganitumab

298 eligible patients with newly diagnosed metastatic Ewing sarcoma: 148 given chemotherapy alone, 150 given chemotherapy plus ganitumab, in AEWS1221 (NCT02306161).

In absolute terms, 37.4% of patients on chemotherapy alone and 39.1% with ganitumab were alive without relapse or a new cancer at 3 years (next chart).

Source: COG AEWS1221 investigators, reporting the trial's design assumptions [19] (2023-01-20); Steven G. DuBois et al.; Children's Oncology Group, overall-survival result [33] (2023-01-20). Chart: OncoSignal.

After 3 years, about the same share of patients were alive and free of relapse in both groups

Percentage of patients in each group alive without relapse or a new cancer, and percentage alive, 3 years after joining the trial. These are estimates for each group as a whole, not a prediction for any one patient.

How to read this: Grey is chemotherapy alone and orange is chemotherapy plus ganitumab. The thin bars are 95% confidence intervals, which show how precisely each percentage was estimated. They overlap almost entirely.

Alive with no relapse or new cancer
3-year event-free survival
37.4% Chemotherapy alone (95% CI 29.3 to 45.5)
39.1% Chemotherapy plus ganitumab (95% CI 31.3 to 46.7)
Alive
3-year overall survival
59.5% Chemotherapy alone (95% CI 50.8 to 67.3)
56.7% Chemotherapy plus ganitumab (95% CI 48.3 to 64.2)
0%25%50%75%100%

298 eligible patients: 148 on chemotherapy alone, 150 with ganitumab. Kaplan-Meier estimates at 3 years.

Source: COG AEWS1221 investigators, randomized phase III report [20] (2023-01-20). Chart: OncoSignal.

Why it was expected to work

  • Xenograft regressions and chemotherapy enhancement supported the target and the combination strategy.[3][26]
  • R1507 produced a 10% phase II response rate and a median response duration of 29 weeks. Roche had explicitly assigned the program high priority after the earlier Ewing responses.[1][4][5][7][8][13][61]
  • Ganitumab produced an Ewing partial response in the ASCO 2010 report. Its later pooled phase II ORR was 6%; COG cited 11% Ewing-specific partial response or prolonged stable disease.[9][10][11][13]
  • NantCell acquired ganitumab rights for broad oncology development and anticipated accelerated development after the Ewing orphan designation.[14][15][16][17]
  • AEWS1221 planned 300 patients to detect an EFS hazard ratio of 0.67 or smaller.[19][62][63]
  • Early ganitumab results included a 6% ORR and 17% clinical benefit rate in a pooled EFT/DSRCT cohort, not an Ewing-only population.[28][29]
  • The protocol model assumed median EFS of 16.6 months on comparator therapy and 24.4 months on experimental therapy.[32]

What happened

  • NCT02306161 tested ganitumab plus VDC/IE against VDC/IE, with subsequent ganitumab maintenance in the experimental arm.[16][18]
  • Among 298 eligible patients, 3-year EFS was 37.4% with standard treatment and 39.1% with ganitumab. The EFS hazard ratio was 1.00.[20][64][65]
  • Ganitumab did not reduce EFS events or death. Three-year overall survival was 59.5% versus 56.7%, standard versus experimental.[20][27][33]
  • In March 2019, the DSMC recommended closure and stopping ganitumab because of the observed outcomes and potential increased pneumonitis risk, before the formal inefficacy boundary had been crossed.[21][22]
  • AEWS1221/NCT02306161 was the randomized phase III trial in newly diagnosed metastatic Ewing sarcoma.[59]
  • Median relative FDG-PET SUV was 0.34 in both randomized arms.[42]
  • Roche's business halt concerned R1507. The separate robatumumab report attributes its early termination to sponsor portfolio reprioritization.[52][56]
  • The authors proposed inadequate tumor exposure, absent biomarker enrichment, and reduced chemotherapy intensity as possible explanations. They did not establish a definitive mechanism.[24][25][35][34]
  • In Ewing sarcoma, CABONE reported cabozantinib responses, while TK216 advanced through early dose-escalation cohorts.[67][68]

The early signal: tumours shrank in about 1 in 10 relapsed patients, or fewer

Response rate: the share of patients whose tumours shrank by a set amount (a partial or complete response).

How to read this: Bars start at zero. The thin line on the R1507 bar is its 95% confidence interval, which shows how precisely the rate was measured. The ganitumab report gave no interval.

R1507, phase 2
relapsed or refractory Ewing sarcoma, published 2011; 95% CI 4.9% to 16.5%
10%
Ganitumab, phase 2
Ewing family tumours and DSRCT pooled, 35 patients assessed, published 2012
6%
0%5%10%15%20%

The phase 3 trial asked a different question in different patients: whether adding ganitumab to chemotherapy kept newly diagnosed patients with metastatic disease free of relapse for longer.

Source: Pappo et al., SARC R1507 phase II investigators, Journal of Clinical Oncology [8] (2011-10-24); William D. Tap et al., ganitumab phase II report; these results pool EFT and DSRCT [11] (2012-04-16). Chart: OncoSignal.

What was said about why

Stated is the sponsor or trial group's own words; Reported is a named third party. Categories are our label from a fixed list.

79 checked sources: 43 are the trial group's or sponsor's own words

Who is speaking in each checked source. Context items are dated facts from the same years that nobody gave as a reason.

  • 43 Stated the trial group's or sponsor's own words
  • 13 Reported a named third party
  • 23 Coincident dated context, not given as a reason

Source: OncoSignal evidence pack for this case (2026-11-11). Chart: OncoSignal.

How strong was the case for the target?

Open Targets 26.09, IGF1R and ewing sarcoma, leaving out clinical evidence, which includes the failed trial itself, so the answer is not circular. Scores run from 0 to 1.

literature0.83

The trial that raised hopes against the trial that failed

From a 38-patient study without a comparison group to a 312-patient randomised trial, measuring something different

The trial that raised hopes and the trial that failed, side by side. Randomised means patients were assigned to treatments by chance, so the groups can be compared fairly.

Phase 2 · NCT00563680
Design
Not randomised
Patients
38
Ages
16 Years and over
Main measure
Objective response rate (Partial Response [PR] or Complete Response [CR]) as determined by RECIST
Started
2007-10
Phase 3 · NCT02306161
Design
Randomised
Patients
312
Ages
up to 50 Years
Main measure
Event-free Survival
Started
2014-12-12
Patients enrolled, same scale
Phase 2
38
Phase 3
312

Patients is the enrolment each registry record reports, so it can differ from the number analysed. The full field-by-field comparison is in the table below.

Source: ClinicalTrials.gov NCT00563680; ClinicalTrials.gov NCT02306161. Chart: OncoSignal.

FieldNCT00563680NCT02306161
Phase · changedPHASE2PHASE3
Allocation · changedNON_RANDOMIZEDRANDOMIZED
MaskingNONENONE
Enrolment · changed38 (actual)312 (actual)
Ages · changed16 Years to not publishednot published to 50 Years
Arms · changedExploratory Cohort; Main CohortRegimen A (VDC/IE); Regimen B (VDC/IE + ganitumab)
Primary outcome · changedObjective response rate (Partial Response [PR] or Complete Response [CR]) as determined by RECISTEvent-free Survival
Countries · changed0: not published3: Canada; Puerto Rico; United States

From the current registry records. Earlier versions of a record are not available to automated readers.

What the trial assumed, and what happened

The plan: ganitumab would push the median event-free time from 16.6 to 24.4 months

Median event-free time: the time by which half the patients in an arm have relapsed, had a new cancer or died. The numbers are the protocol's planning assumptions, used to size the trial.

16.6 chemotherapy alone24.4 hoped with ganitumab
0612182430months

The planned gain: 7.8 months longer before half the patients had an event.

The published paper did not report the median for either arm, so the plan cannot be set against an observed median. Its 3-year results are in the chart above.

Source: Children's Oncology Group; AEWS1221 protocol/SAP, proposed-design model [32] (2019-05-01). Chart: OncoSignal.

  • Planned sample size, AEWS1221: assumed 300 patients; observed not published
  • Event-free survival hazard ratio, AEWS1221: assumed 0.67; observed 1.00
  • Protocol-model assumed comparator-arm median EFS: assumed 16.6 months; observed not published
  • Protocol-model target experimental-arm median EFS: assumed 24.4 months; observed not published
  • Planned statistical power: assumed 81%; observed not published

Where in the chain could it have broken?

Each link is answered from a cited source or marked not published. A missing answer is often the finding.

Checked sources speak to 8 of the 8 links between hitting the target and helping patients

A drug helps patients only if every link holds, top to bottom. Each link is marked by whether a checked source answers it; the answers are in the table below.

How to read this: Follow the links downwards. A filled marker means at least one checked source speaks to that link. A hollow marker means nothing was published, and that gap is often the finding.

  1. 1. target · answered by 1 source summary
    Is there human genetic or somatic driver evidence for this target in this cancer?
  2. 2. dependency · answered by 2 source summaries
    Does the tumour need the target when it is present?
  3. 3. exposure · answered by 1 source summary
    Did the drug reach the tumour at an active concentration?
  4. 4. engagement · answered by 4 source summaries
    Was the target shown to be hit in patients, at the dose used?
  5. 5. patients · answered by 4 source summaries
    Were the enrolled patients the ones whose tumours carry the target?
  6. 6. modality · answered by 2 source summaries
    Did a different way of hitting the same target later succeed?
  7. 7. design · answered by 3 source summaries
    Were dose, comparator, endpoint and size able to show the effect?
  8. 8. safety · answered by 1 source summary
    Did toxicity limit dose or duration?

Answered is not the same as fine. Several answers on this page are the trial group's own possible explanations, which they did not establish.

Source: OncoSignal evidence pack for this case (2026-11-11). Chart: OncoSignal.

In the blood, ganitumab reached its target level in 9 of 10 patients tested after the first dose and 10 of 10 after the second

Target: a trough level (the lowest level, just before the next dose) above 10,000 ng/mL, chosen from earlier testing.

After the first dose
9 of 10 met the target
After the second dose
10 of 10 met the target

These are blood levels in the patients tested. They do not show how much drug reached the tumour, which the trial group said may have been too little.

Source: DuBois et al., Children's Oncology Group [23] (2023-01-20). Chart: OncoSignal.

target
Is there human genetic or somatic driver evidence for this target in this cancer?
Open Targets 26.09, IGF1R in ewing sarcoma, without clinical evidence: literature 0.83
dependency
Does the tumour need the target when it is present?
Preclinical IGF-1R blockade caused Ewing xenograft regressions and enhanced chemotherapy activity. Laboratory combination data explicitly helped motivate the randomized trial.[3][26]
Investigators questioned whether IGF-1R was sufficiently critical, while noting that responses in relapsed disease argued against that explanation. Separate cell studies reported switching to IGF-2/IR-A dependence.[34][76]
exposure
Did the drug reach the tumour at an active concentration?
Ganitumab serum trough benchmarks were generally met, but investigators considered inadequate intratumoral concentrations possible. R1507 xenograft studies also identified schedule-dependent activity and an insufficient trough target.[23][24][51]
engagement
Was the target shown to be hit in patients, at the dose used?
The investigators said intratumoral concentrations might nevertheless have been inadequate. This was a proposed explanation, not an established finding.[24]
Earlier R1507 investigators described how insufficient dosing could permit incomplete ligand blockade. This was a pharmacologic concern, not an established explanation for AEWS1221.[50][51]
No validated predictive biomarker was available, so the trial could not enrich for likely beneficiaries.[25]
The earlier ganitumab study reported a postdose IGF-1 rise. It found no apparent relationship between IGF-1 levels and tumor response.[77][78]
patients
Were the enrolled patients the ones whose tumours carry the target?
Early R1507 testing involved recurrent or refractory disease; AEWS1221 involved newly diagnosed metastatic Ewing sarcoma. No validated predictive biomarker permitted enrichment.[1][16][25]
The trial group identified possible diagnostic misclassification, absent predictive enrichment, and the unfavorable metastatic population as potential limitations.[39][25][40]
Earlier activity was reported in relapsed disease. AEWS1221 tested newly diagnosed metastatic disease.[13][62]
Investigators said lack of patient selection may have contributed. Molecular confirmation was not required, and no validated predictive biomarker was available for enrichment.[79][25]
modality
Did a different way of hitting the same target later succeed?
The SARC study tested R1507. Robatumumab is identified separately as 19D12/MK-7454/SCH717454; it should not be substituted for R1507 in this case study.[1][2]
Earlier investigators and commentators identified insulin-receptor signaling and alternative receptor tyrosine kinases as possible resistance mechanisms. These were not demonstrated causes of the randomized result.[49][54][55]
design
Were dose, comparator, endpoint and size able to show the effect?
Early response signals led to a different test: adding ganitumab to chemotherapy and measuring EFS. The randomized trial was designed to detect an EFS hazard ratio of 0.67 or smaller.[13][18][19][26][63][64]
The authors considered reduced chemotherapy intensity, local-control differences, and response definitions. They also considered early ganitumab discontinuation, but their post hoc analysis did not support that explanation.[35][36][37][38]
The trial compared VDC/IE with or without ganitumab, using 18 mg/kg ganitumab and targeting an EFS hazard ratio of 0.67 or lower. Investigators proposed reduced consolidation intensity and possible local-control differences as explanations. Their post hoc analysis did not support dilution of benefit from early discontinuation.[18][75][19][35][36][37]
safety
Did toxicity limit dose or duration?
Potential increased pneumonitis risk contributed to the closure recommendation. All four experimental-arm patients with grade 3 or worse pneumonitis had received thoracic radiotherapy.[21][22]

Our analysis. Not established fact.

  1. The phase 3 may have asked a question the early trials could not answer: the expectation came from single-agent response or prolonged stable disease in about 6 to 11% of relapsed patients, while AEWS1221 was powered for an event-free survival hazard ratio of 0.67 and observed 1.00 when ganitumab was added to intensive first-line chemotherapy. If a drug's effect shows up mainly as occasional tumour shrinkage in refractory disease, that signal could be too small and too differently measured to predict a one-third reduction in events added to an intensive backbone.

    Rests on
    assumed.Event-free survival hazard ratio, AEWS1221, design.Primary outcome, design.Allocation, design.Arms, chain.design
    Would confirm
    A published model or an earlier add-on trial in a paediatric sarcoma showing that a single-agent response rate near 10% in relapsed disease typically corresponds to an add-on event-free survival hazard ratio much closer to 1 than to 0.67; or a protocol document showing the 0.67 target was chosen for feasibility rather than derived from the early data.
    Would refute
    Documented cases where comparable single-agent response rates in relapsed paediatric sarcoma were followed by add-on randomised hazard ratios of 0.67 or better, or a protocol rationale deriving the 0.67 target directly from the ganitumab and R1507 results.
    Rival explanations
    The effect size was reasonable and the drug is simply inactive in first-line metastatic disease.; The trial was stopped at the second interim analysis, so a smaller true benefit could not be excluded.
    Novelty
    Also noted by DuBois et al., Journal of Clinical Oncology (2023), AEWS1221 trial report, Discussion. This is a substantially similar qualitative observation about insufficient relapse-setting activity—not an explicit analysis of the 0.67 target or the response-versus-EFS endpoint mismatch., checked 2026-11-11

  2. The early signal and the randomised test may have sampled different diseases: responses were seen in recurrent or refractory tumours, in a phase 2 that enrolled patients aged 16 and over, while AEWS1221 enrolled newly diagnosed metastatic patients with no lower age limit. If IGF-1R dependence is acquired or enriched after first-line therapy, or differs with age, the responding subset could be largely absent from the population randomised.

    Rests on
    design.Ages, chain.patients, design.Phase, design.Enrolment
    Would confirm
    Paired diagnostic and relapse Ewing sarcoma samples showing higher IGF-1R pathway activity or receptor expression at relapse, or a response analysis of the ganitumab and R1507 phase 2 trials by age or prior therapy showing responders concentrated in a group AEWS1221 rarely enrolled.
    Would refute
    Similar IGF-1R pathway activity at diagnosis and relapse, and phase 2 responses spread evenly across age and prior-treatment groups.
    Rival explanations
    The pathway matters equally at diagnosis, and intensive chemotherapy masks a modest additive effect.; The phase 2 responses were rare enough to reflect chance or patient selection rather than a real subset.
    Novelty
    Not found in the sources searched on 2026-11-11 (News and trade press: The ASCO Post and Cancer Therapy Advisor; searches also targeted Endpoints, Fierce Biotech, STAT and BioCentury., Newsletters and Substacks., Blogs, including searches targeting In the Pipeline., Publicly indexed social posts, including searches targeting X/Twitter and LinkedIn., Research papers, trial reports, editorials and reviews concerning ganitumab, IGF-1R inhibition and AEWS1221.)

  3. The exposure benchmark may have measured the wrong thing: serum trough targets were reported as met, but no source reports receptor blockade in tumour, and the earlier ganitumab trial found a post-dose rise in IGF-1 with no relation to response. If circulating ligand rises when the receptor is blocked, a blood level that looks adequate could still leave tumour receptors only partly occupied.

    Rests on
    chain.exposure, chain.engagement, timeline
    Would confirm
    Tumour biopsies on treatment showing incomplete IGF-1R occupancy or persistent downstream signalling at the AEWS1221 dose, or modelling that shows the post-dose IGF-1 rise is large enough to compete with the antibody at tumour concentrations.
    Would refute
    On-treatment biopsies showing near-complete receptor occupancy and pathway shutdown with no clinical benefit, which would point away from exposure and towards dependence.
    Rival explanations
    Exposure was adequate and the tumour does not depend on the receptor in this setting.; Engagement was adequate but escape signalling bypassed it.
    Novelty
    Not found in the sources searched on 2026-11-11 (News and trade press, including searches targeting Endpoints, Fierce, STAT, and BioCentury, Newsletters and Substacks, Blogs, including searches targeting In the Pipeline, Publicly indexed social posts, including searches targeting X/Twitter and LinkedIn, Peer-reviewed trial reports, papers, reviews, and editorials)

  4. Blocking IGF-1R alone may leave a second route open: cell studies in the pack report Ewing sarcoma switching to IGF-2 signalling through the insulin receptor A isoform under IGF-1R antibodies, and reviews list insulin-receptor signalling among resistance mechanisms. If that route carries the survival signal, the result would be consistent with a correct pathway and an incomplete blockade, which a ligand-neutralising or dual receptor approach would test and a second IGF-1R antibody would not.

    Rests on
    chain.dependency, chain.modality, timeline
    Would confirm
    A clinical or patient-derived model result in Ewing sarcoma where neutralising IGF-1 and IGF-2 together, or blocking both IGF-1R and the insulin receptor, works where an IGF-1R antibody alone does not; or AEWS1221 tumour samples showing high IGF-2 or insulin receptor A expression in non-responders.
    Would refute
    Ligand-neutralising or dual-receptor agents failing in Ewing sarcoma as completely as the IGF-1R antibodies, or no link between IGF-2 or insulin receptor A levels and outcome.
    Rival explanations
    The pathway as a whole is a modifier rather than a driver in this cancer, so any blockade would give small effects.; Other receptor tyrosine kinases, not the insulin receptor, provide the escape route.
    Novelty
    Also noted by Garofalo et al., Oncogene (2011), and de Groot et al., Cancers (2020): the IGF-2 and insulin receptor A escape mechanism is already published. Only our framing is new: reading the AEWS1221 null as consistent with incomplete blockade, and proposing a ligand-neutralising or dual-receptor test., checked 2026-11-11

  5. Today's Open Targets release (26.09) lists only literature support (0.83) linking IGF1R to Ewing sarcoma, with no germline, somatic or functional-genomic evidence. If a similar gap existed when AEWS1221 was designed, and if programmes on targets without such evidence succeed less often, that flag could have tempered the expected effect size; it also suggests checking the same pattern in the other IGF-1R failures.

    Rests on
    chain.target, chain.dependency, assumed.Event-free survival hazard ratio, AEWS1221
    Would confirm
    An Open Targets or literature snapshot from around 2014 showing the same absence of genetic or somatic evidence for IGF1R in Ewing sarcoma, together with other paediatric sarcoma programmes on literature-only targets failing at a similar rate.
    Would refute
    Successful paediatric oncology phase 3 trials whose targets had only literature support in Open Targets at the time, showing the flag carries little weight in this field.
    Rival explanations
    Genetic databases under-represent signalling dependencies that are not mutated, so the gap may say little about this target.; Preclinical xenograft regressions are real evidence of dependence that a genetic database does not capture.
    Novelty
    Not found in the sources searched on 2026-11-11 (News and general web results, Trade-press searches, including Endpoints, Fierce Biotech, STAT, and BioCentury, Newsletters and Substacks, Blogs, Publicly indexed social posts, Papers and reviews, including the AEWS1221 trial report and IGF1R/Ewing sarcoma literature, Open Targets Platform 26.09 materials)

What else was happening. Not shown to be a cause.

About 15 years from the first trial to the published result

Trial milestones from the registry and the publications, with the other dated events found (listed below) as grey ticks. Nothing here is shown to be a cause.

20092011201320152017201920212023
  1. 2007-10Early trial NCT00563680 starts
  2. 2014-12-12Randomised trial NCT02306161 starts
  3. 2019-03Readout: the trial is stopped and the result becomes known
  4. 2023-01-20The randomised result is published

Source: ClinicalTrials.gov NCT02306161; Trial publication (2023-01-20). Chart: OncoSignal.

What a go/no-go reader could have seen beforehand

Only facts already true before 2019-03, each with the date it became true.

What is not public

  • A definitive biological explanation for the randomized null result was not established in the cited excerpts. Inadequate intratumoral exposure was presented as a possibility.
  • A validated predictive biomarker for selecting IGF-1R responders was unavailable for AEWS1221.
  • The complete cross-indication program budget and an Ewing-specific acquisition valuation were not found in the reviewed sources. The 2015 ganitumab licensing announcement explicitly left financial terms undisclosed.
  • An Ewing-specific orphan or expedited regulatory designation for R1507 was not found in the reviewed sources; the documented Ewing orphan designation was for ganitumab.
  • A randomized CR-plus-PR objective response rate comparable to the early single-agent signal was not located in the AEWS1221 publication or posted results reviewed.
  • Observed arm-specific median EFS and median OS were not reported in the primary paper reviewed.
  • The protocol's assumed control-arm three-year EFS is tabulated rather than stated in a numerical sentence.
  • No public FDA or EMA assessment, complete response letter, or advisory-committee document explaining this Ewing-sarcoma trial result was located. This is not evidence that a regulatory rejection occurred.
  • No separate Amgen statement establishing why AEWS1221 lacked benefit, or standalone public DSMC deliberation beyond the trial group's account, was located.
  • A demonstrated biological cause of the null randomized result, including measured inadequate intratumoral ganitumab exposure or inadequate tumor receptor occupancy, was not established in the sources reviewed. The trial authors' explanations remain qualified hypotheses.
  • No located source attributes the AEWS1221 efficacy result to funding, competition, manufacturing, drug supply, or portfolio reprioritization. The quoted Roche and Schering-Plough/Merck business decisions concern different antibody studies.
  • A definitive biological cause of the negative result was not established. The trial authors explicitly described several possible explanations.

Sources

  1. Pappo et al., SARC investigators; NCT00642941, Journal of Clinical Oncology, 2011-10-24 · Stated · quote found in the research run's saved excerpt
  2. Authors of the robatumumab phase II report, Pediatric Blood & Cancer; a separate study, NCT00617890, 2016-06-30 · Reported · quote found in the research run's saved excerpt
  3. Pappo et al., SARC investigators, explaining the biological rationale in Journal of Clinical Oncology, 2011-10-24 · Stated · quote found in the research run's saved excerpt
  4. Roche press release, reproduced by BioSpace, 2007-10-23 · Stated · quote found on the page
  5. Kapil Dhingra, Head of Oncology Disease Biology Area at Roche, 2007-10-23 · Stated · quote found on the page
  6. Genmab announcement of Roche's phase II sarcoma trial, 2007-12-20 · Coincident · quote found on the page
  7. Pappo et al., SARC Ewing sarcoma phase II investigators, Journal of Clinical Oncology, 2011-10-24 · Stated · quote found in the research run's saved excerpt
  8. Pappo et al., SARC R1507 phase II investigators, Journal of Clinical Oncology, 2011-10-24 · Stated · quote found in the research run's saved excerpt
  9. W. D. Tap et al., AMG 479 phase II results, ASCO Annual Meeting abstract 10001, Journal of Clinical Oncology supplement, 2010-05-20 · Stated · quote found in the research run's saved excerpt
  10. William D. Tap et al.; NCT00563680, Journal of Clinical Oncology, published online April 16, 2012, 2012-04-16 · Stated · quote found in the research run's saved excerpt
  11. William D. Tap et al., ganitumab phase II report; these results pool EFT and DSRCT, 2012-04-16 · Stated · quote found in the research run's saved excerpt
  12. Author correction to Tap et al., Journal of Clinical Oncology; corrected PFS units are weeks, 2012-11-10 · Stated · quote found in the research run's saved excerpt
  13. COG AEWS1221 investigators, retrospectively describing the earlier ganitumab trial and the rationale for AEWS1221, 2023-01-20 · Reported · quote found on the page
  14. NantCell licensing announcement reproduced by Fierce Biotech; broad oncology rights, not an Ewing-specific acquisition, 2015-01-14 · Coincident · quote found in the research run's saved excerpt
  15. Patrick Soon-Shiong, founder and chief executive officer of NantCell, 2015-01-14 · Stated · quote found in the research run's saved excerpt
  16. NantCell announcement published by Pharmaceutical Processing World, 2017-04-12 · Stated · quote found on the page
  17. Patrick Soon-Shiong, Chairman and CEO of NantCell, 2017-04-12 · Stated · quote found on the page
  18. COG AEWS1221 investigators, randomized phase III report, Journal of Clinical Oncology, 2023-01-20 · Stated · quote found on the page
  19. COG AEWS1221 investigators, reporting the trial's design assumptions, 2023-01-20 · Stated · quote found on the page
  20. COG AEWS1221 investigators, randomized phase III report, 2023-01-20 · Stated · quote found on the page
  21. Steven G. DuBois et al.; Children's Oncology Group, reporting the DSMC recommendation, 2023-01-20 · Stated · quote found on the page
  22. COG AEWS1221 investigators, 2023-01-20 · Stated · quote found on the page
  23. DuBois et al., Children's Oncology Group, 2023-01-20 · Coincident · quote found on the page
  24. COG AEWS1221 investigators, offering a possible explanation rather than a demonstrated cause, 2023-01-20 · Stated · quote found on the page
  25. COG AEWS1221 investigators, discussing patient selection, 2023-01-20 · Stated · quote found on the page
  26. COG AEWS1221 investigators, retrospectively stating the combination-treatment rationale, 2023-01-20 · Stated · quote found on the page
  27. COG AEWS1221 investigators, 2023-01-20 · Stated · quote found on the page
  28. William D. Tap et al.; early ganitumab phase II investigators, NCT00563680, 2012-04-16 · Stated · quote found in the research run's saved excerpt
  29. William D. Tap et al.; pooled EFT/DSRCT ganitumab phase II results, 2012-04-16 · Stated · quote found in the research run's saved excerpt
  30. Alberto S. Pappo et al.; R1507 study registration, 2011-10-24 · Coincident · quote found on the page
  31. Peter M. Anderson et al.; robatumumab investigators, 2016-06-30 · Coincident · quote found on the page
  32. Children's Oncology Group; AEWS1221 protocol/SAP, proposed-design model, 2019-05-01 · Stated · quote found in the research run's saved excerpt
  33. Steven G. DuBois et al.; Children's Oncology Group, overall-survival result, 2023-01-20 · Stated · quote found on the page
  34. DuBois et al., Children's Oncology Group; explicitly uncertain biological explanation, 2023-01-20 · Stated · quote found on the page
  35. Steven G. DuBois et al.; Children's Oncology Group, 2023-01-20 · Stated · quote found on the page
  36. Steven G. DuBois et al.; Children's Oncology Group, explicitly qualified explanation, 2023-01-20 · Stated · quote found on the page
  37. Steven G. DuBois et al.; Children's Oncology Group, 2023-01-20 · Stated · quote found on the page
  38. Steven G. DuBois et al.; Children's Oncology Group, response-assessment limitation and qualification, 2023-01-20 · Stated · quote found on the page
  39. Steven G. DuBois et al.; Children's Oncology Group, possible diagnostic dilution, 2023-01-20 · Stated · quote found on the page
  40. Steven G. DuBois et al.; Children's Oncology Group, 2023-01-20 · Stated · quote found on the page
  41. Steven G. DuBois et al.; Children's Oncology Group, limits on generalizing the negative result, 2023-01-20 · Stated · quote found on the page
  42. Steven G. DuBois et al.; Children's Oncology Group, exploratory FDG-PET response measure—not CR/PR objective response rate, 2023-01-20 · Stated · quote found on the page
  43. Angel M. Carcaboso; Translational Pediatrics editorial commentary, 2023-10-10 · Reported · quote found on the page
  44. Angel M. Carcaboso; critique of the preclinical evidence supporting ganitumab, 2023-10-10 · Reported · quote found on the page
  45. Angel M. Carcaboso; balancing weak preclinical evidence against the earlier clinical signal, 2023-10-10 · Reported · quote found on the page
  46. Angel M. Carcaboso; patient-selection critique, 2023-10-10 · Reported · quote found on the page
  47. Angel M. Carcaboso; conclusion and proposed evidentiary requirements, 2023-10-10 · Reported · quote found on the page
  48. Alberto S. Pappo et al.; R1507 investigators, historical discussion of limited responses, 2011-10-24 · Stated · quote found in the research run's saved excerpt
  49. Alberto S. Pappo et al.; R1507 investigators, proposed resistance mechanisms—not findings established in AEWS1221, 2011-10-24 · Stated · quote found on the page
  50. Alberto S. Pappo et al.; R1507 investigators, reporting Roche's xenograft pharmacology, 2011-10-24 · Stated · quote found on the page
  51. Alberto S. Pappo et al.; R1507 investigators, preclinical dose-schedule explanation, 2011-10-24 · Stated · quote found on the page
  52. Alberto S. Pappo et al.; R1507 investigators, reporting Roche's decision—not the reason AEWS1221 failed, 2011-10-24 · Stated · quote found on the page
  53. Alan L. Ho and Gary K. Schwartz; historical biomarker interpretation, 2011-10-24 · Reported · quote found on the page
  54. Alan L. Ho and Gary K. Schwartz; historical hypothesis about insulin-receptor sparing, 2011-10-24 · Reported · quote found on the page
  55. Alan L. Ho and Gary K. Schwartz; historical alternative-signaling hypothesis, 2011-10-24 · Reported · quote found on the page
  56. Peter M. Anderson et al.; separate robatumumab study, not AEWS1221 or the R1507 trial, 2016-06-30 · Stated · quote found on the page
  57. Georgios M. Stergiopoulos et al.; Frontiers in Oncology case report and literature review, 2025-11-19 · Reported · quote found on the page
  58. Angel M. Carcaboso; reporting why the COG trial closed early, 2023-10-10 · Reported · quote found on the page
  59. Steven G. DuBois et al.; Children's Oncology Group, failed-trial identification, 2023-01-20 · Coincident · quote found on the page
  60. National Cancer Institute, Robatumumab (C71527) entry · Coincident · quote found in the research run's saved excerpt
  61. Pappo et al., SARC phase II R1507 investigators; NCT00642941, 2011-10-24 · Coincident · quote found in the research run's saved excerpt
  62. DuBois et al., Children's Oncology Group, AEWS1221/NCT02306161, 2023-01-20 · Coincident · quote found in the research run's saved excerpt
  63. DuBois et al., Children's Oncology Group, 2023-01-20 · Coincident · quote found in the research run's saved excerpt
  64. DuBois et al., Children's Oncology Group, 2023-01-20 · Coincident · quote found in the research run's saved excerpt
  65. DuBois et al., Children's Oncology Group, 2023-01-20 · Coincident · quote found in the research run's saved excerpt
  66. U.S. Food and Drug Administration, press release distributed through PR Newswire, 2020-01-21 · Coincident · quote found on the page
  67. Italiano et al., CABONE phase II cabozantinib investigators, The Lancet Oncology, 2020-03 · Coincident · quote found in the research run's saved excerpt
  68. TK216 first-in-human Ewing sarcoma investigators; abstract published May 30, 2017, 2017-01 · Coincident · quote found in the research run's saved excerpt
  69. Authors of “Targeting the IGF1R/PI3K/AKT Pathway Sensitizes Ewing Sarcoma to BET Bromodomain Inhibitors”, 2019-03-29 · Coincident · quote found in the research run's saved excerpt
  70. de Groot et al., Cancers; review of possible resistance mechanisms, 2020-11-29 · Coincident · quote found in the research run's saved excerpt
  71. Children's Oncology Group, AEWS1221 funding acknowledgment; retrospective disclosure, not a dated funding change, 2023-01-20 · Coincident · quote found on the page
  72. Amgen; corporate context, not the NCI trial sponsor, 2019-11-21 · Coincident · quote found on the page
  73. Amgen and Five Prime Therapeutics; acquisition agreement and oncology strategy, 2021-03-04 · Coincident · quote found on the page
  74. Anderson et al., robatumumab phase II trial group, 2016-06-30 · Stated · quote found in the research run's saved excerpt
  75. Steven G. DuBois et al., Children's Oncology Group, 2023-01-20 · Coincident · quote found on the page
  76. C. Garofalo et al., Oncogene, reporting Ewing sarcoma resistance-model findings, 2011-01-31 · Reported · quote found in the research run's saved excerpt
  77. William D. Tap et al., earlier ganitumab phase II study, NCT00563680, 2012-04-16 · Coincident · quote found in the research run's saved excerpt
  78. William D. Tap et al., earlier ganitumab phase II study, NCT00563680, 2012-04-16 · Coincident · quote found in the research run's saved excerpt
  79. Steven G. DuBois et al., Children's Oncology Group, 2023-01-20 · Stated · quote found on the page

Assembled 2026-11-11 from linked public records by research runs costing $1.2. All case studies