Flomics posts 92% multi-cancer test
What happened
- Flomics Biotech said its cfRNA blood test detected five cancers from a single draw in a 1,000-plus-sample study presented in Spain in January 2025. (flomics.com) - The headline numbers were AUC 0.92 and 83% sensitivity at 90% specificity, with 80% sensitivity reported for Stage I disease. (flomics.com) - It matters because cfRNA could widen multi-cancer screening beyond ctDNA, but this is still pre-validation conference-stage evidence. (flomics.com)
Why it matters
A multi-cancer blood test lives or dies on one question — can it catch cancer early enough to matter, without flooding clinics with false alarms? That is the w(flomics.com)A. In a study presented in Santiago de Compostela in late January 2025, the Barcelona company reported that its cfRNA test picked(flomics.com)00 samples. (flomics.com) ##(flomics.com)ters because RNA is a readout of what cells are doing right now, not just what mutations they carry. Flomics combines cfRNA sequencing with machine-learning models to classify whether a sample looks cancer-like and, separately, what cancer type it may be. The five cancers in this dataset were colorectal, lung, breast, prostate, and pancreatic. (flomics.com) ### Why use RNA instead of DNA? Most of the multi-cancer screening buzz has centered(flomics.com)uld, at least in theory, surface earlier biological disruption because it captures gene-expression changes. That is the pitch behind cfRNA more broadly, and it is why people in liquid biopsy keep revisiting it despite the technical headaches. (flomics.com) ### So what did Flomics show? The company said the study used samples from more than 1,000 patients colle(flomics.com)thy classifier posted mean AUC 0.92 ± 0.01, with 83% sensitivity at 90% specificity. For Stage I cancers, Flomics said sensitivity was 80%. Those are strong-looking numbers for an early detection claim, especially because pancreatic cancer is in the mix. (flomics.com) ### How good is the cancer typing? This is where the story gets more pra(flomics.com) cancer to 99% for prostate cancer. In other words, the model is not just trying to wave a red flag. It is also trying to point clinicians toward the likely organ system. That matters because every extra diagnostic branch after a positive screen costs time, money, and patient anxiety. (storage.unitedwebnetwork.com) ### What is the catch? The catch is that cfRNA is hard. The field still lacks standardized(flomics.com)ntrolled. That makes single-company performance claims interesting but not decisive. A good conference poster is not the same thing as a locked clinical assay. (sciencedirect.com) ### Is this a published validation? Not yet in the usual sense. What is public here is a company announcement plus conference abstracts and posters, not a full peer-reviewed clinical validation paper. Flomics has said it is prepa(storage.unitedwebnetwork.com)would tell you whether these numbers hold up outside the original development set. (flomics.com) ### Why would labs care? Because a real cfRNA screen would change the front end of cancer workups. More blood-first triage means more reflex imaging, more confirmatory(sciencedirect.com)at only happens if validation shows the test can stay accurate in routine populations, not just curated study cohorts. (flomics.com) ### Bottom line? Flomics did not prove that cfRNA has won multi-cancer screening. But it did put up numbers strong enough to keep the idea alive — and maybe make people take RNA a lot more seriously. The next step is simple to say and hard to do: bigger, messier, real-world validation. (flomics.com)
Key numbers
- Flomics Biotech said its cfRNA blood test detected five cancers from a single draw in a 1,000-plus-sample study presented in Spain in January 2025.
- (flomics.com) The headline numbers were AUC 0.92 and 83% sensitivity at 90% specificity, with 80% sensitivity reported for Stage I disease.
- In a study presented in Santiago de Compostela in late January 2025, the Barcelona company reported that its cfRNA test picked(flomics.com)00 samples.
- The company said the study used samples from more than 1,000 patients colle(flomics.com)thy classifier posted mean AUC 0.92 ± 0.01, with 83% sensitivity at 90% specificity.
What happens next
- Flomics combines cfRNA sequencing with machine-learning models to classify whether a sample looks cancer-like and, separately, what cancer type it may be.
- The next step is simple to say and hard to do: bigger, messier, real-world validation.
- (flomics.com) It matters because cfRNA could widen multi-cancer screening beyond ctDNA, but this is still pre-validation conference-stage evidence.
Quick answers
What happened in Flomics posts 92% multi-cancer test?
Flomics Biotech said its cfRNA blood test detected five cancers from a single draw in a 1,000-plus-sample study presented in Spain in January 2025. (flomics.com) The headline numbers were AUC 0.92 and 83% sensitivity at 90% specificity, with 80% sensitivity reported for Stage I disease. (flomics.com) It matters because cfRNA could widen multi-cancer screening beyond ctDNA, but this is still pre-validation conference-stage evidence. (flomics.com)
Why does Flomics posts 92% multi-cancer test matter?
A multi-cancer blood test lives or dies on one question — can it catch cancer early enough to matter, without flooding clinics with false alarms? That is the w(flomics.com)A. In a study presented in Santiago de Compostela in late January 2025, the Barcelona company reported that its cfRNA test picked(flomics.com)00 samples. (flomics.com) ##(flomics.com)ters because RNA is a readout of what cells are doing right now, not just what mutations they carry. Flomics combines cfRNA sequencing with machine-learning models to classify whether a sample looks cancer-like and, separately, what cancer type it may be. The five cancers in this dataset were colorectal, lung, breast, prostate, and pancreatic. (flomics.com) Why use RNA instead of DNA? Most of the multi-cancer screening buzz has centered(flomics.com)uld, at least in theory, surface earlier biological disruption because it captures gene-expression changes. That is the pitch behind cfRNA more broadly, and it is why people in liquid biopsy keep revisiting it despite the technical headaches. (flomics.com) So what did Flomics show? The company said the study used samples from more than 1,000 patients colle(flomics.com)thy classifier posted mean AUC 0.92 ± 0.01, with 83% sensitivity at 90% specificity. For Stage I cancers, Flomics said sensitivity was 80%. Those are strong-looking numbers for an early detection claim, especially because pancreatic cancer is in the mix. (flomics.com) How good is the cancer typing? This is where the story gets more pra(flomics.com) cancer to 99% for prostate cancer. In other words, the model is not just trying to wave a red flag. It is also trying to point clinicians toward the likely organ system. That matters because every extra diagnostic branch after a positive screen costs time, money, and patient anxiety. (storage.unitedwebnetwork.com) What is the catch? The catch is that cfRNA is hard. The field still lacks standardized(flomics.com)ntrolled. That makes single-company performance claims interesting but not decisive. A good conference poster is not the same thing as a locked clinical assay. (sciencedirect.com) Is this a published validation? Not yet in the usual sense. What is public here is a company announcement plus conference abstracts and posters, not a full peer-reviewed clinical validation paper. Flomics has said it is prepa(storage.unitedwebnetwork.com)would tell you whether these numbers hold up outside the original development set. (flomics.com) Why would labs care? Because a real cfRNA screen would change the front end of cancer workups. More blood-first triage means more reflex imaging, more confirmatory(sciencedirect.com)at only happens if validation shows the test can stay accurate in routine populations, not just curated study cohorts. (flomics.com) Bottom line? Flomics did not prove that cfRNA has won multi-cancer screening. But it did put up numbers strong enough to keep the idea alive — and maybe make people take RNA a lot more seriously. The next step is simple to say and hard to do: bigger, messier, real-world validation. (flomics.com)
