AI for Oncology Research

Surface cancer hypotheses
researchers can act on

RehmaAI synthesizes evidence across biomedical literature, clinical trial registries, and genomics databases — returning structured, source-traced hypotheses with a human gate at every step.

Request access See how it works
Data sources PubMed 40M+ · ClinicalTrials.gov · TCGA · UniProt · ChEMBL
Capabilities

Built for oncology research, not generic search

RehmaAI is scoped to cancer biology. It retrieves, cross-references, and synthesizes evidence across specialized oncology sources — not the open web.

Literature synthesis

PubMed oncology journals and cancer biology publications. Queries return structured summaries with citation trails — not ranked document lists.

Trial intelligence

ClinicalTrials.gov and oncology-specific registries. Surfaces active, completed, and terminated trials relevant to a research question with endpoint data.

Genomics & datasets

TCGA, COSMIC, and curated cancer genomics databases. Biomarker and mutation data linked to relevant clinical and translational evidence.

Hypothesis dossiers

Structured evidence packages with supporting findings, counterevidence, and uncertainty flags. Every dossier requires researcher review and approval before any experimental step.

Discovery registry

Longitudinal tracking of research questions, findings, and outcomes — including negatives. Institutional knowledge accumulates across research cycles.

No auto-retraining

The model does not retrain on clinical data automatically. Learning is deliberate: reviewed, approved, and versioned. Every AI decision is permanently logged.

The problem

Most cancer signals are buried in data researchers cannot connect

Oncology research produces evidence faster than any team can systematically review. PubMed indexes more than 800,000 biomedical articles annually. ClinicalTrials.gov lists over 460,000 registered studies. TCGA and COSMIC together hold genomic profiles across dozens of cancer types.

The bottleneck is not data volume — it is structured synthesis across sources. A signal in COSMIC may already have an active trial in ClinicalTrials.gov and three contradicting papers in PubMed. Finding that constellation manually takes weeks. Missing it means duplicated effort, or a hypothesis that was quietly disproved years ago.

RehmaAI is built to close that gap: a structured, traceable path from a research question to a dossier a researcher can evaluate and act on.

"The challenge is not generating hypotheses. It is knowing which ones are supported, which are contradicted, and which have already been tested."

— A recurring theme in translational oncology research
Source-traced, not summarized Every claim in a dossier links to its source document, trial record, or dataset entry. Researchers can inspect the evidence directly.
Counterevidence included Contradicting findings are surfaced alongside supporting ones. Uncertainty is a first-class output, not an edge case.
Negatives are recorded Hypotheses that did not hold up are logged in the Discovery Registry, preventing redundant investigation across cycles.
Workflow

Six steps. A human decision at every one.

Intelligence without governance is liability. No step in the pipeline advances without a deliberate human action.

01 — SUBMISSION
Research question intake

A researcher submits a cancer research question or hypothesis candidate through the workspace. The system structures it into a retrievable query.

02 — RETRIEVAL
Multi-source evidence retrieval

The RAG engine queries PubMed, ClinicalTrials.gov, TCGA, and COSMIC in parallel, pulling primary source records against the structured query.

03 — SYNTHESIS
Evidence dossier generation

Retrieved records are synthesized into a structured dossier: supporting evidence, counterevidence, uncertainty flags, and a full citation trail.

04 — REVIEW GATE
Researcher review & approval

The dossier reaches the researcher's queue. No hypothesis advances to any experimental consideration without an explicit researcher decision.

05 — REGISTRY
Discovery Learning Registry

Every outcome — confirmed, refuted, or inconclusive — is recorded with the full dossier. Negative findings are logged with the same rigor as positive ones.

06 — CONTINUITY
Institutional knowledge loop

Validated signals accumulate as citable discovery outputs. The registry compounds over research cycles, reducing redundant effort across teams.

Data sources

Scoped to oncology. No general-purpose search.

RehmaAI queries curated sources in cancer biology and translational oncology — not the open web. Each source is indexed, versioned, and attribution-tracked.

Literature
PubMed / MEDLINE

Oncology-filtered journal articles, preprints via Europe PMC, and cancer biology publications with citation-level retrieval.

Clinical trials
ClinicalTrials.gov

Active, completed, and terminated oncology trials with endpoint data, eligibility criteria, and sponsor information.

Genomics
TCGA & COSMIC

The Cancer Genome Atlas and Catalogue of Somatic Mutations in Cancer — mutation frequency, gene expression, and biomarker data across cancer types.

Expanding
Additional sources

GEO, cBioPortal, and curated oncology trial registries are in scope for subsequent data-access agreements. All sources require institutional review before activation.

Validation
Pilot — Q1 2027

Stage IV NSCLC retrospective — cohort design complete

RehmaAI's initial validation study is built around a 124-patient retrospective cohort of Stage IV Non-Small Cell Lung Cancer, covering first-line systemic therapy episodes. Primary endpoints are overall survival and time to next treatment. EGFR, ALK, PD-L1, and KRAS biomarker status are included for each patient in scope.

Data feasibility review is being finalized with an oncology center. Pilot launch is scheduled for Q1 2027. The study design is available for academic review on request.

Request study brief
124
Patients in retrospective cohort
Stage IV NSCLC
OS + TTNT
Primary endpoints tracked per patient
Q1 '27
Pilot launch target
Data feasibility review underway
10+
Biomarkers in scope incl. EGFR, ALK, PD-L1, KRAS
Design principles

What RehmaAI will not do

These are constraints, not aspirations. They are encoded into the system architecture.

01
No treatment recommendations

RehmaAI is a research tool. It surfaces evidence for researcher evaluation. It does not generate, rank, or suggest treatment pathways for individual patients.

02
No automatic model retraining

The AI model does not update from clinical data automatically. Retraining is a deliberate decision — reviewed, approved, and version-logged by the research team.

03
No hypothesis advances without researcher approval

Every dossier requires an explicit researcher decision before any experimental consideration. The system cannot escalate a signal to an action.

04
No unattributed claims

Every statement in a dossier links to its primary source. Synthetic summaries that cannot be traced to a source document are flagged as uncertain, not presented as findings.

05
Validated drug candidates require external confirmation

RehmaAI does not produce validated drug candidates. Discovery outputs are research signals that require experimental confirmation through standard scientific process.

The Team

Built by clinicians, engineers,
and domain experts

A multidisciplinary team combining oncology expertise, AI research, and health systems experience.

Leadership
Giscard Minlend
Giscard Minlend
Founder & CEO
Founder & CEO of ViroVita HealthGlobal, and the driving force behind RehmaAI — bringing cancer discovery intelligence to researchers worldwide.
Subhash Reddy
Subhash Reddy
Chief Medical Officer
Physician and clinical leader overseeing RehmaAI's medical accuracy, evidence standards, and oncology research alignment.
Jorge Sebastiao
Jorge Sebastiao
Chief Technology Officer
Technology architect with deep expertise in AI systems, cloud infrastructure, and enterprise health technology platforms.
Khadijah Barro
Khadijah Barro
VP Strategy & Partnerships
Leads institutional partnerships and strategic development, connecting RehmaAI with academic oncology research programs.
Team
Keron Burrell
Keron Burrell
Lead AI & Software Engineer
Chichi Nnyamah
Chichi Nnyamah
Head of Data Strategy & Integration
Dulaj Lakshitha
Dulaj Lakshitha
Hardware Product Engineer
Advisors
Dr. Mazin Gadir
Dr. Mazin Gadir
Healthcare Strategy & Digital Health Advisor
Director at Alvarez & Marsal Healthcare & Life Sciences (Dubai). 19+ years spanning PwC Middle East, IQVIA, SEHA, and Cerner across the UK, GCC, and USA. PhD in Quantum Electronics from University of Leeds. Advises on digital health strategy, health informatics, and organisational transformation.
Prof. Yunkap Kwankam
Prof. Yunkap Kwankam
Advisor
Former eHealth Director at the World Health Organization. Brings global health systems expertise and institutional credibility.
Anthony Ritossa
Anthony Ritossa
Business Advisor
Chairman of the Global Family Office Investment Summit (GFOIS). Advises on global investment strategy and institutional growth.
Get in touch

Partner with us on
oncology research.

RehmaAI is in active development with a limited number of academic oncology research partners. If you are building a research program in oncology or cancer biology and want to explore a collaboration, we would like to hear from you.

We are particularly interested in partnerships around retrospective cohort data access, endpoint validation methodology, and translational oncology research design.

research@rehmaai.com
We respond to all inquiries within 2 business days.
Thank you

We have received your request and will be in touch within 2 business days.