Company News

OncoHelper AI has launched Art for Hope, a new gallery of children’s artwork created to make charitable giving more personal.

Supporters can now choose a drawing when making a donation to our pediatric cancer research. After the donation is confirmed, they receive a digital reproduction of the selected artwork as a thank-you. In some cases, the original may also be available on request.

We wanted a donation to be more than a transaction. A child’s drawing can become a small but lasting reminder of why someone chose to help — something personal to keep, remember, and share.

The first artworks are now available, and the gallery will continue to grow as children contribute new drawings.

Explore Art for Hope and choose the artwork that speaks to you.

OncoHelper AI has begun development of its first working research prototype.

Despite having no external funding for this stage, we have found a practical path forward using real historical pediatric oncology data and a focused, measurable research design.

The first prototype will compare two approaches using the same patient cohort: a conventional model based primarily on static clinical information, and an OncoHelper AI model that also incorporates how the patient’s condition and response evolve over time.

Our initial research question is deliberately simple and testable:

Can longitudinal patient information improve the prediction of later clinical outcomes compared with static information alone?

To answer it, historical patient records will be transformed into structured patient trajectories. The models will be tested retrospectively using only information that would have been available at the time of each prediction, with the actual subsequent outcome kept hidden until validation.

This work creates the first practical foundation for the broader OncoHelper AI architecture — including INTERCEPTOR, TACTICIAN, SENTRY, and the Meta-AI layer.

Dataset evaluation and system design are already underway. Our current goal is to have the first working prototype and initial validation results by mid-September 2026.

For more information about the project:

OncoHelper AI has launched a new Project Opportunities section to present international research projects that are ready for collaboration with U.S. academic and scientific partners.

The first opportunity now available is Long-Read Molecular Classification of Pediatric CNS Tumors, an international pediatric neuro-oncology project combining long-read sequencing, molecular profiling, AI-supported classification, and a pilot liquid-biopsy pathway using cerebrospinal fluid.

OncoHelper AI is currently seeking a U.S. research partner able to contribute a complementary scientific component to the project and participate in the development of a joint international study framework.

The new Project Opportunities format is designed to help build focused international research collaborations around projects with an existing scientific and clinical foundation.

After a nearly three-month pause in public project updates, OncoHelper AI is returning with renewed momentum and a stronger project portfolio.

During this period, we added two promising new projects to our pipeline. We will begin publishing their descriptions on the website in the coming days, together with updates on several important decisions we have made about how we present our work.

One of those decisions concerns project teams. Early-stage teams often evolve as a project moves from concept and development into active execution. Going forward, we will present detailed project team rosters once a project has reached a sufficiently defined execution stage and the long-term team structure is clear.

Until then, our public project pages will focus on what matters most: the problem being addressed, the scientific and technical approach, current development status, and the next milestones.

More updates will follow shortly.

Building on the operational foundation announced earlier this spring, and adapting our work to the current realities, OncoHelper AI is now strengthening its role as an international oncology project office. In April and May, we shifted our focus toward grant development, project structuring, financial planning, and preparing oncology initiatives for competitive funding review. During this period, we submitted new grant applications, and they are now under review. We expect further information during the next review cycle. At the same time, we continue to develop our own research initiatives and support international scientific and clinical teams in shaping strong, realistic, and fundable oncology projects.

Building the Foundation Behind Our Work Our core projects remain the same. They are still central to our mission. But strong projects need a strong foundation behind them. Like many early-stage nonprofits, we have been carrying much of this work through volunteer effort. We now need a stronger operational base for coordination, applications, reporting, communication, and project support. That is why we have added a new project to our site: Building the Operational Foundation Behind Every Project. This effort supports all of our other work.

We have completed the core specification of Interceptor v0, defining the internal structure of the system at the current stage of development. Interceptor is being developed as a dynamic state model of a pediatric cancer patient. The system represents the patient as an evolving process described by a structured state vector and governed by mode-dependent dynamics. The current specification introduces three key components: First, a 10-component state vector that captures both tumor-related processes and the child’s physiological condition. This allows the model to track tumor progression and host response within a unified framework. Second, a mode-switching structure that separates baseline periods, active treatment phases, and recovery intervals. Each mode is associated with different system dynamics, reflecting real clinical conditions. Third, a strict separation between observations and interventions. Measured data (laboratory values, imaging-derived features, clinical events) are treated independently from clinical actions (therapy, procedures, supportive care), ensuring a consistent and interpretable modeling approach. A key design principle of Interceptor is the joint modeling of tumor dynamics and host physiology. The system integrates both domains into a single evolving representation. The goal of this specification is to establish a controlled, testable framework that can be incrementally expanded and validated.

To make our work more visible and easier to follow, the personal Facebook profile of our founder, Sergei Oleshkevich, has now been opened to the public and set to Professional Mode. In addition to updates on our website, we will now also share selected project news, milestones, and progress updates on Facebook. This gives supporters and interested readers another simple way to follow the development of OncoHelper AI and our work in pediatric oncology.

In the four months since receiving official U.S. nonprofit status, OncoHelper AI has moved from formation into execution. Research teams are now in place for our three core programs, and work is underway: our AI platform, the zinc plus iron platform, and our PARP inhibitor program. We have reached the point where small early contributions unlock real momentum. If you support better outcomes in pediatric oncology, your donation helps us keep the work moving forward.