Haematology decision support

Better software for the work around a diagnosis.

We work with clinical teams on the difficult parts of haematology: response assessment, connected diagnostic workflows, and usable data in settings where infrastructure is limited.

Built as clinical workflow support—not autonomous diagnosis.

Fig. 01WHO diagnostic traceSynthetic example
Diagnostic trace showing a WHO 2022 decision map, the selected criterion, evidence used, and resulting pathway
Source visibleCriteria explicitClinician reviewed
Current work

Three problems we are working on now.

Different clinical settings, one approach: make the source, rules and human review points explicit.

01
The ChristieInternal evaluation

AML treatment response

A deterministic ELN 2017 and 2022 response calculator that makes every threshold, missing value and blood-count source visible for review.

From assessment data to an auditable response summary and PDF.Read about the project
02
haemOSSandbox proof of concept

A modern diagnostic workflow

An exploration of what comes after HODS: connecting referral, validated laboratory results, classification, reporting and final sign-off in one case workspace.

Designed around Epic and NHS sandbox integration boundaries.Read about the project
03
CambodiaPrototype

AML records and cohort exploration

Source-linked extraction from Cambodian Tumor Board and CanReg proformas, followed by human review, a shared case register and governed cohort queries.

The core extraction and query workflow is deterministic.Read about the project
How we build

The source and the reasoning stay visible

From source-linked findings to visible classification logic and framework outputs.

01

Review findings against their source

Structured findings remain alongside the original report text, so clinicians can inspect what was recorded and identify items that still need review.

  • Source report and structured findings shown together
  • Blast, molecular, cytogenetic, and immunophenotype review
  • Clinician-entered values clearly distinguished
  • Uncertain parser matches surfaced for review
Classification results input summary showing structured findings beside highlighted source report evidence
02

See the complete classifier map

Explore the configured WHO, ICC, ELN risk, validation-and-gates, and downstream MRD logic as a connected system rather than a hidden calculation.

  • Complete WHO 2022 classifier view
  • Explicit terminal results and labelled branches
  • Selectable decisions and outcomes
  • Separate views for AML, CML, MDS, CMML, and MPN rules
Complete classifier map showing WHO 2022 AML rules, labelled branches, and terminal results
03

Compare framework outputs

WHO 2022 and ICC 2022 outcomes are presented separately, with direct access to the clinical reasoning steps and diagnostic trace behind each result.

  • WHO 2022 and ICC 2022 outcomes kept distinct
  • Clinical reasoning available step by step
  • Diagnostic trace linked from each framework result
  • Outputs remain available for clinician review
WHO 2022 and ICC 2022 classification outputs with links to clinical reasoning and diagnostic traces
Clinician feedback

What clinicians have said

Selected feedback shared with Haem.io.

Haem.io has the potential to transform diagnosis and decision making for patients with haematological malignancies across the NHS.

Prof Charles Craddock CBEChair, UK AML Research NetworkUniversity of Warwick

Haem.io has the potential to revolutionise the quality of care that patients receive.

Dr John ChadwickConsultant HaematologistThe Christie NHS Foundation Trust

A real step forward in what technology can offer in supporting clinicians to make accurate clinical diagnoses.

Dr Tom CoatsHaematology ConsultantRoyal Devon & Exeter NHS Trust

I would wholeheartedly recommend this platform. This will be of significant use for clinicians in the front line.

Dr P A CahalinConsultant HaematologistBlackpool Teaching Hospitals NHS Trust

Building with clinical teams

We are developing focused tools with collaborators in the UK and Cambodia.