Dose Escalation Study of Lithium With Oxaliplatin and Capecitabine in Advanced Oesophago-Gastric or Colorectal Cancer

1. Is the biology right?

• Somatic driver evidence: no · None of IMPA1, GSK3A, GSK3B has somatic-mutation evidence in colorectal cancer

  Rule: Open Targets has somatic-mutation evidence for the target in this cancer. Source: Open Targets: IMPA1 in colorectal cancer, 2026-11-11
• Human genetic evidence: no · No germline or somatic evidence for IMPA1, GSK3A, GSK3B in colorectal cancer

  Rule: Any germline or somatic human genetic evidence for the target in this cancer, counted apart from animal-model and literature evidence. Source: Open Targets: IMPA1 in colorectal cancer, 2026-11-11
• Pediatric target: no · Not on the FDA list

  Rule: Target is on FDA's Pediatric Molecular Target List (relevant targets); membership alone neither requires nor waives a study. Source: FDA Pediatric Molecular Target List (Listed as February 2026 on the FDA Pediatric Oncology page; the file served is 9.7.22_the_relevant_molecular_target_list.xlsx, whose Summary of Changes sheet is headed 'SEPTEMBER 2022 UPDATE' (footnote date 2022-09-01)), 2026-11-11
• Haematologic cancer: no · Solid tumour

  Rule: From the indication, or the registry conditions when no indication matches. Source: ClinicalTrials.gov NCT03153280, 2026-11-11

  Published context: Likelihood of approval from phase 1: 7.5% haematologic vs 4.6% solid (BIO / Informa / QLS 2021, 2011–2020). Population-level, not a prediction.

• Background therapy: Oxaliplatin, Capecitabine: standard agents, not counted in these flags.
• IMPA1 (via Lithium): Evidence types in colorectal cancer: none. · Open Targets: IMPA1 in colorectal cancer
• GSK3A (via Lithium): Evidence types in colorectal cancer: literature; literature listed separately from human genetics. Datasources: europepmc. · Open Targets: GSK3A in colorectal cancer
• GSK3B (via Lithium): Evidence types in colorectal cancer: literature, clinical; literature listed separately from human genetics. Datasources: clinical_precedence, europepmc. · Open Targets: GSK3B in colorectal cancer

• Not loaded on this app yet: Normal-tissue expression in vital organs. Source: Human Protein Atlas (ticket 36).
• Not loaded on this app yet: Pan-essential in cell lines. Source: DepMap (ticket 36).
• Not loaded on this app yet: Approved drug on this target. Source: Drugs@FDA (ticket 30).

2. Can the result be trusted?

• Randomised: no · Single group

  Rule: Registry allocation field. Source: ClinicalTrials.gov NCT03153280, 2026-11-11
• Small denominator: yes · 2 enrolled; each patient moves a rate by 100 ÷ 2 = 50.0 points

  Rule: Enrolment under 30; the arithmetic is shown, not a benchmark. Source: ClinicalTrials.gov NCT03153280, 2026-11-11
• Primary endpoint is overall survival: no · Incidence of dose limiting toxicity (DLT) within the two first cycles at each dose level.

  Rule: A registered primary outcome measure names overall survival. Source: ClinicalTrials.gov NCT03153280, 2026-11-11

  Published context: Non-OS primary endpoint: odds ratio 2.79 for a positive phase 3 (Otsuka 2021, 400 phase 3 oncology trials, to 2020). Shown with the next figure: a surrogate is easier to hit, and often a weak guide to survival.

  Published context: 52% of surrogate-to-OS correlations are low (Prasad 2015, published to 2014). Shown with the figure above.
• Results overdue: yes · Primary completion 2023-09-07, 38 months ago; no results posted

  Rule: Primary completion (actual) more than 12 months ago and no results posted to the registry. Source: ClinicalTrials.gov NCT03153280, 2026-11-11

  Published context: 40.9% of applicable trials report results within a year (Results-reporting studies summarised in spec §11, all therapy areas). All areas, not oncology only.

• Not loaded on this app yet: Treated and evaluable counts, assessor, confirmed response, follow-up and data cutoff. Source: Readout extraction (ticket 32).
• Not loaded on this app yet: Primary endpoint changed between registry versions. Source: Registry versions in the store (ticket 32).

3. What must it beat?

• Not loaded on this app yet: Control-arm result and approved therapies for the same segment, with label links. Source: Drugs@FDA, FDA labels and readouts (tickets 30, 32).

4. Too crowded, or too late?

• Stopped on this target: 46 terminated, withdrawn or suspended trials of drugs against IMPA1; 5 give no reason. Reasons are quoted as the registry gives them on the target page. · IMPA1: what already stopped

• Not loaded on this app yet: Programs on the target by phase and months behind the leader. Source: Programs and Drugs@FDA (ticket 33).

5. Can it be approved, and stay approved?

• Single-country: yes · All 2 sites in Ireland

  Rule: All listed sites are in one country; the share of US sites is shown. Source: ClinicalTrials.gov NCT03153280, 2026-11-11
• EU paediatric investigation plan: no · No EMA paediatric investigation plan decision names Lithium

  Rule: An EMA PIP decision (agreement, modification or waiver) lists the drug as its active substance, exact name. Source: EMA data download (dated snapshot), 2026-11-11
• EU orphan designation: no · No EU orphan designation names Lithium

  Rule: An EU orphan designation lists the drug as its active substance (exact name) or belongs to a medicine that does; withdrawn and expired designations are listed with their status. Source: EMA data download (dated snapshot), 2026-11-11
• Authorised in the EU: no · EMA lists no centrally authorised medicine of Lithium

  Rule: An EMA medicine record lists the drug as its active substance or INN, exact name; national-only authorisations are not in this file. Source: EMA data download (dated snapshot), 2026-11-11
• FDA orphan designation: not published · FDA's orphan designation database turns away automated requests, so this app does not read it; search it by hand at the link

  Rule: Would be: an FDA orphan designation names the drug. Source: FDA orphan designations search
• FDA pediatric oncology approval: no · FDA's table lists no pediatric approval of Lithium

  Rule: FDA's Pediatric Oncology Drug Approvals table names the drug. Source: FDA Pediatric Oncology Drug Approvals (content current as of 2026-03-03), 2026-11-11
• Registered studies open to children: yes · 57 interventional studies of Lithium include patients under 18; 12 are recruiting, not yet recruiting or active

  Rule: The registry's age group includes Child (birth to 17 years), which also counts studies that set no minimum age. Source: ClinicalTrials.gov search, 2026-11-11

• Caveat: Site counts are not patients per country, and trials registered only outside ClinicalTrials.gov are not covered. Where the sites are is a fact about the record, not a judgement of its quality.
• Cited case: In February 2022 FDA asked its Oncologic Drugs Advisory Committee whether sintilimab's application, resting on ORIENT-11, a phase 3 run only in China, needed further trials showing it applies to US patients and US practice. One case, not a rule. · FDA, ODAC 10 Feb 2022, draft questions
• Rare pediatric disease voucher: FDA, content current as of 29 Apr 2026: “As of February 3, 2026, enactment of the Consolidated Appropriations Act, 2026, the rare pediatric disease PRV program will sunset after September 30, 2029. FDA may not award any PRVs under this program after September 30, 2029.” Voucher sale prices are not shown here; they appear only as company-reported news. · FDA: Rare Pediatric Disease Designation and Priority Review Voucher Programs

• Not loaded on this app yet: Precedent approvals for target, modality and endpoint; accelerated-approval status. Source: Drugs@FDA and FDA accelerated-approval tables (ticket 30).

6. Can the sponsor finish?

• Sponsor: Cancer Trials Ireland (academic or other) · ClinicalTrials.gov NCT03153280
• study director: Cancer Trials Ireland Dublin 11, Ireland, Cancer Trials Ireland · Person page

• Not loaded on this app yet: Filed cash, quarterly spend and the runway arithmetic. Source: SEC XBRL company facts (ticket 37).
• Not loaded on this app yet: Sponsor's completed vs stopped oncology trials and results-reporting record. Source: Registry, by company (ticket 37).

What public data cannot tell you

Public records cannot show: sales forecasts; a probability of success for this asset; full deal terms and net prices; FDA designations and action dates before approval (shown only as company-reported); clinical holds unless the company disclosed them; manufacturing; patents on unapproved drugs; whether preclinical findings have been replicated.