Somewhere in the sprawling catalog of human cancers, scientists have been playing molecular Where's Waldo for decades. B-cell acute lymphoblastic leukemia (B-ALL) - a blood cancer that attacks the immune system's B cells - has already been sliced into 27 different subtypes based on their genetic fingerprints. Twenty-seven! You'd think we'd found them all by now.

Nope.

An international team just spotted a 28th subtype lurking in the data, and it's a nasty piece of work. They call it FOXF1/FENDRR, and it really, really doesn't want to respond to chemotherapy.

When Your Gene Gets Hijacked by a Long Non-Coding RNA

Here's what researchers found after analyzing a staggering 4,857 B-ALL patients across three cohorts: twenty patients shared a bizarre genetic signature nobody had documented before. The culprit? A rearrangement called IGH::FENDRR, where the immunoglobulin heavy chain gene (the one that normally helps make antibodies) gets fused with FENDRR, a long non-coding RNA usually associated with fetal development and - plot twist - tumor suppression in other cancers.

In most cancers, FENDRR acts like a molecular brake pedal, slowing down tumor growth in lung, gastric, and colon cancers. But in this new B-ALL subtype, the brake has somehow become the accelerator. The IGH enhancer hijacks FENDRR and cranks its expression way up, along with a transcription factor called FOXF1 that lives next door on chromosome 16.

The KRAS Connection That Nobody Expected

If the weird gene fusion wasn't enough, 17 out of 20 patients also carried a specific KRAS mutation: A146T (or its cousins A146V and A146P). KRAS mutations are cancer's greatest hits - they show up in 90% of pancreatic cancers and plenty of others. But this particular mutation at position 146 is weirdly tissue-specific, mostly appearing in colorectal cancer and blood cancers.

What makes A146T special? Unlike the more common KRAS mutations that jam the protein's "off switch," A146T works differently - it speeds up how quickly the protein cycles between active and inactive states. Think of it as a light switch that flickers really fast instead of just being stuck on.

Finding this mutation in 85% of FOXF1/FENDRR patients (versus less than 2% in other B-ALL subtypes) suggests these two genetic abnormalities are partners in crime.

Chemotherapy: "Thanks, But No Thanks"

Here's where things get grim. Of 13 patients treated with standard GMALL or GRAALL chemotherapy protocols, 8 either failed to achieve remission or still had significant disease (MRD ≥10-3) after induction. That's a 62% failure rate for treatment that usually works.

The cancer essentially shrugged at chemotherapy.

The Immunotherapy Rescue

But there's a silver lining. When patients were switched to immunotherapy - specifically blinatumomab, a bispecific antibody that functions like molecular velcro between T cells and cancer cells - the outcomes improved dramatically. Combined with stem cell transplants, 13 of 16 patients achieved ongoing molecular remission.

Blinatumomab works by grabbing a T cell with one arm and a cancer cell with the other, forcing an introduction that ends badly for the cancer. The FDA fully approved it for B-ALL patients with minimal residual disease, and for good reason: patients who achieve MRD-negative status have five-year survival rates around 75%, compared to 33% for those who don't.

Machine Learning to the Rescue

The researchers built a machine learning classifier using gene expression data that can identify FOXF1/FENDRR patients with high accuracy. This matters because catching these patients early - before wasting weeks on ineffective chemotherapy - could mean jumping straight to immunotherapy and transplant.

It's a reminder that genomic classification isn't just academic taxonomy. When different subtypes respond to completely different treatments, knowing which bucket a patient falls into can be the difference between remission and relapse.

What This Means Going Forward

Twenty patients might sound like a small number, but remember - these researchers analyzed nearly 5,000 cases to find them. If FOXF1/FENDRR represents roughly 0.4% of adult B-ALL cases, that translates to hundreds of patients worldwide who might be getting the wrong first-line treatment.

The bigger picture: even with 27 established subtypes, the genetic landscape of leukemia still has unexplored territory. Every new subtype discovered is another piece of the puzzle - another group of patients who might benefit from personalized treatment instead of one-size-fits-all chemotherapy.

And honestly, that's the promise of precision medicine in a nutshell. Not just treating cancer, but treating this specific patient's cancer, with all its quirky genetic baggage.

References

1. Bendig S, et al. IGH::FENDRR and specific KRAS mutations define a novel B-ALL molecular subtype with poor chemotherapy response. Blood. 2025. DOI: 10.1182/blood.2025031102

2. Arber DA, et al. International Consensus Classification of Acute Lymphoblastic Leukemia/Lymphoma. Blood. 2023. PMCID: PMC10646822

3. Li Z, et al. LncRNA FENDRR Expression Correlates with Tumor Immunogenicity. Genes. 2021. PMCID: PMC8226633

4. Cook JH, et al. Tissue-Specific Oncogenic Activity of KRAS A146T. Cancer Discovery. 2019. DOI: 10.1158/2159-8290.CD-18-1220

5. Gökbuget N, et al. Blinatumomab for minimal residual disease in adults with B-cell precursor acute lymphoblastic leukemia. Blood. 2018. DOI: 10.1182/blood-2017-08-798322

6. Borrero L, Papaioannou AT. Advances and Challenges in RAS Signaling Targeted Therapy in Leukemia. Molecular Cancer Therapeutics. 2025. DOI: 10.1158/1535-7163.MCT-24-0346

Disclaimer: This blog post is a simplified summary of published research for educational purposes. The accompanying illustration is artistic and does not depict actual model architectures, data, or experimental results. Always refer to the original paper for technical details.