Skip to main content

Evidence

Knowledge is power. Here’s the evidence.

The more a patient and their team know, the better they do. Medicine keeps finding it, trial after trial. Cited in full below, limits included.

higher response rate31% median response rate with biomarker selection, versus 10.5% without. An across-trial comparison, not a randomized one.570 phase II studies · 32,149 patients · Schwaederle et al., JCO, 2015This is why the deep analysis starts from your biology.
1 in 5patients found a mistake in their own recordsAmong patients who read their clinical notes, 21% reported an error; 42% of those were serious.22,889 note readers · 3 US health systems · Bell et al., JAMA Network Open, 2020It’s also why reading your record is free.
58%lower risk of developing type 2 diabetesNot cancer; what matters is what informed patients do next. Intensive lifestyle coaching versus placebo; metformin cut risk 31%.3,234 adults with prediabetes · Knowler et al., NEJM, 2002And why Triangle reads beyond one disease.

Triangle is built on this premise. Start with your own file →

What the analysis connects

One analysis. A broader view of what’s in your file.

Triangle connects your records and supported sequencing before you ask a question, then turns the results into a reading you can trace back to its sources.

Accredited laboratories do the sequencing. Triangle re-analyses supported data for research use with established tools. What runs depends on what is available in your file.

How this differs from a general AI chatbot →

01 · Your fileRecords + supported sequencing
02 · AnalysisA staged analysis with established tools
03 · Your readingA sourced reading + visible gaps

Tumour biology

  • Alterations, tiered and evidenced
  • Copy number, variants, and fusions
  • Gene expression
  • Mutational burden and MSI
  • Mutational signatures

Immune & inherited context

  • HLA type and loss
  • Immune cell estimates
  • Neoantigen candidates
  • Pathway enrichment
  • Germline signals for follow-up

Options to discuss

  • Targeted-therapy evidence
  • Relevant clinical-trial candidates
  • Drug-metabolism context
  • Supplement interactions to discuss

What’s included depends on the records and sequencing data available in your file. Evidence sources include peer-reviewed literature, FDA labelling, NCCN guidelines, and public trial registries.

See examples of the analysis tools

BWA-MEM2 · GATK · Manta · PureCN · STAR · Salmon · Arriba · OptiType · MHCflurry

The second look, measured

A second reading changes cases.

None of these studies tested Triangle. Nothing has yet; we are new. They test the premise it is built on: the second, complete look is where findings live. We will publish our own numbers as they accrue: review turnaround, corrections issued, what a run surfaced that the chart missed.

12% of first diagnoses survived a second opinion unchanged Of 286 cases carried to a second clinic, 21% emerged distinctly different and 66% better defined. 286 referrals · Van Such et al., J Eval Clin Pract, 2017
less progression when treatment matched the tumour’s biology Patients on closely matched therapy progressed more slowly and lived longer. Observational, single centre. 429 evaluable patients · Kato et al., Nature Communications, 2020
8% of cancer patients ever join a trial For 56 of every 100, no trial was open where they were treated. The barrier is the system, not the patient. 13 studies · 8,883 patients · Unger et al., JNCI, 2019

Read your own file the same way.

Get started for free