01Research
Urothelial carcinoma
Combining gene expression, molecular features, and clinical data to study anti-PD-1 treatment outcomes.
SITC 2022 · Exploratory studyView research
Oncology · Data science · Patient stratification
Understanding
who responds.
Machine learning research at the intersection of molecular biology, clinical data, and cancer treatment response.
AlghorismusResearch grounded in evidence.
The research framework01 / 03
Different data.
A shared question.
01Molecular
profilesRNA · Genomics
profilesRNA · Genomics
02Clinical
contextDisease · Treatment
contextDisease · Treatment
03Observed
outcomesResponse · Follow-up
outcomesResponse · Follow-up
Multimodal model developmentTestable hypotheses for patient stratification
Independent validationThe next step from a signal to reliable evidence
Conceptual workflow · No patient predictions shown
The question behind the science
Can we better understand treatment response before the next decision?
Alghorismus studies how molecular and clinical information can inform patient stratification. Our published projects explore the potential of machine learning—and the validation needed to make those findings useful.
Our research
Three disease areas.
One focus on response.
02Research
Metastatic melanoma
Investigating circulating CD8+ T-cell transcriptomics as a potential source of response information.
SITC 2022 · Exploratory studyView research
03Research
Breast cancer
Evaluating multimodal models of pathological complete response using I-SPY2 trial data.
ASCO 2023 · Exploratory studyView research
These are exploratory research projects. They do not establish a validated test for choosing an individual patient’s treatment.
For research teams & life sciences
Start with your
scientific question.
Explore a collaboration around existing datasets, a biomarker hypothesis, or an independent evaluation of a model.
Explore collaborationData & study feasibility
Clarify the cohort, available features, outcomes, and limits of the data.
Multimodal analysis
Discuss how transcriptomic, molecular, imaging, and clinical data could be studied together.
Evidence & evaluation
Define the right comparisons, uncertainty measures, and independent validation needs.
Scientific transparency
A promising idea is the
beginning of the evidence.
We present study context, reported results, and limitations together. Performance in a small retrospective cohort is not a guarantee of performance in a new population.
How to read the evidenceResearch record
Publications & updates
Breast cancer response modeling in the ASCO 2023 abstract
I-SPY2 imaging and clinical data · Journal of Clinical Oncology
Urothelial carcinoma and melanoma research at SITC
Two exploratory studies of treatment response
Start a scientific conversation
What question is your
data ready to answer?
Tell us about your research objective, available data, and the decision your study needs to inform.
Discuss a project