06/30/2026
Our Moonshot Is MO: Building a System of Hope
"When evidence doesn't exist, every patient becomes a N-of-1."
Marlo is two years old.
He has an extraordinarily rare, high-grade spindle cell sarcoma harboring an EGFR kinase domain duplication (EGFR-KDD)—a molecular alteration that has only rarely been described in pediatric sarcomas. Like many children with ultra-rare cancers, there is no established treatment pathway and very little published evidence to guide therapy.
The Starting Point
Marlo underwent one of the most comprehensive molecular workups available, including whole-genome sequencing, whole-transcriptome sequencing, DNA/RNA cancer panels, fusion testing, methylation profiling, and chromosomal microarray analysis.
The primary genomic finding was an EGFR-KDD mutation, leading to treatment with erlotinib (50 mg daily) beginning in November 2025. Erlotinib has remained his only targeted systemic therapy through June 2026.
While his clinical team continued standard multidisciplinary care, one question remained:
Was EGFR still the biology driving the tumor, or had the disease evolved?
Looking Beyond the Mutation
Most precision oncology today focuses on identifying mutations.
But mutations don't always tell us which pathway is active today.
To answer that question, we developed a new blood-based workflow using circulating tumor cells (CTCs).
Starting from just 5 mL of blood, we isolated a very small population of circulating tumor cells and expanded them over approximately two weeks, increasing tumor cell content from less than 1% to approximately 10%—still well below the purity typically preferred for bulk RNA sequencing, but sufficient to begin transcriptomic analysis.
This represented an important technical milestone. Most commercial sequencing workflows are optimized for samples with much higher tumor content.
What the Biology Revealed
Transcriptomic analysis produced an unexpected finding.
Although Marlo's tumor carried an EGFR-KDD mutation, there was limited evidence of active EGFR pathway signaling.
Instead, the analysis suggested possible activation of the ALK signaling pathway—a pathway well known in certain lung cancers and several pediatric malignancies, and one that has proven responsive to ALK-targeted therapies when biologically relevant.
This did not prove that ALK was driving Marlo's cancer.
It did, however, generate a new, biologically plausible hypothesis that could not have been inferred from DNA sequencing alone.
Rather than anchoring treatment decisions on the original genomic alteration, Oncology CoPilot suggested evaluating whether therapy should instead be guided by the tumor's current biological activity.
The Next Step: Turning a Hypothesis into Evidence
Scientific hypotheses must be tested.
Our next phase is designed to independently validate these findings through complementary approaches, including:
Integrated pellet and plasma molecular profiling from blood samples
Comprehensive 200-gene DNA sequencing (including EGFR, ALK, KRAS, and other actionable alterations)
Independent confirmation of ALK pathway activation
Functional drug screening using patient-derived organoids to determine whether Marlo's tumor cells are actually sensitive to ALK inhibitors
Together, these studies will help determine whether the transcriptomic findings represent a clinically meaningful therapeutic opportunity.
Building the Next Generation of Precision Oncology
Marlo's case has also accelerated development of technologies that could improve care for future children with rare cancers.
Current priorities include:
Increasing circulating tumor cell enrichment to greater than 50% tumor purity, enabling higher-quality RNA sequencing from blood
Integrating single-cell RNA sequencing to resolve tumor heterogeneity and reduce background noise
Developing advanced size-based CTC isolation methods to improve sensitivity and support longitudinal monitoring over the course of treatment
Each advancement moves us closer to replacing invasive repeat biopsies with minimally invasive blood-based molecular profiling.
Why This Matters
This single case does not prove that erlotinib caused Marlo's response.
His treatment has included standard multidisciplinary care, and any clinical response is likely the result of multiple therapies working together.
What makes this case important is something different.
It demonstrates that artificial intelligence can help clinicians generate biologically informed hypotheses in situations where published evidence is scarce and conventional precision oncology reaches its limits.
For common cancers, medical evidence is built from thousands of patients.
For rare pediatric cancers, progress often begins with one carefully studied child at a time.
Every patient contributes to the next.
Every data point becomes part of a learning system.
Our goal is bigger than helping one child.
Our moonshot is building a system of hope—where every rare pediatric cancer teaches us how to better treat the next one.
