Can a risk model survive messy screening data?
Multi-modality imaging evaluation, nationwide HER2-Low cohort analysis, breast density AI, and Wāhine Māori health equity.
Massey University & University of Auckland — one integrated cohort
Statistical methods with a line of sight to clinical impact.
Methods that survive messy data. Proven in the clinic.
Breast cancer risk modelling, endocrine therapy adherence, embryo selection AI, and reproducible research methods — from statistical problem to clinical decision.
What the lab is working on
active streamsCan machine learning rank embryos better than an embryologist?
Morphokinetic stage detection, embryo annotation and segmentation, and cohort-relative ranking behind WyldBloom's clinic tools.
Can AI and automation reduce IVF transfer burden and reclaim embryologist time?
Cohort-relative embryo ranking and lab automation tools now running in NZ fertility clinics.
Can a method be trusted if no one can reproduce it?
Open tools and workflows that make translational research auditable from raw data to published conclusion.
Research that left the building
WyldBloom
Research into practice · WyldBloom
Making embryo development measurable.
AI, imaging, and clinical decision support turn embryo development into structured, traceable observations — supporting judgement rather than replacing it.
Visit WyldBloom ↗ AI-derived morphology · developmental observations
One cohort, two institutions
selected peopleHooman MisaghiPhD · Auckland
“Artificial Intelligence Enabled IVF”
Reihaneh HosseiniPhD · Massey
“AI-driven breast cancer screening equity for Aotearoa”
Yuqi DuanPhD · Auckland
“Epidemiology of early-onset cancers with a focus on breast cancer”
Yuting YangPhD · Auckland
“Histological and molecular subtypes of breast cancer in NZ women”
Selected work
tagged by where it landedImproving Breast Cancer Detection in Higher Risk Women: A Multi-Modality Imaging Evaluation in a Private Screening ClinicJournal of Medical Imaging and Radiation Oncology
Application of landmark analysis and piecewise Cox regression to identify features associated with prognosis: A national retrospective cohort study of New Zealand womenCancer Treatment and Research Communications
Accurate machine learning model for human embryo morphokinetic stage detectionJournal of Assisted Reproduction and Genetics
Analysis of HER2-Low Breast Cancer in Aotearoa New Zealand: A Nationwide Retrospective Cohort StudyCancers
Join the lab
If a method can't survive messy data, it doesn't belong in the toolkit.
I take students who want to build things that leave the building.
How to join →