yc

Adialante has redesigned MRI so it fits in a mobile clinic and is paid per scan, aiming to bring MRI cancer screening to any medical site, starting with prostate cancer.

Adventris Pharmaceuticals, a Johns Hopkins spin-out, develops off-the-shelf cancer vaccines designed to raise immune responses against mutant KRAS, the driver mutation shared by pancreatic, colorectal and lung cancers, with the long-term aim of a yearly vaccine that prevents those cancers before they form. Its programmes are preclinical or early clinical.

Algen uses CRISPR screens read out at the single-cell level, plus machine learning, to find which RNA messages drive cancer and inflammation, and then designs drugs against them.

Alixia is developing drugs that act on the supporting cells and inflammation around a tumour as well as the cancer cells, to stop cancers from becoming resistant.

Altay Therapeutics is developing small-molecule drugs for chronic liver disease, including the scarring that leads to liver cancer and liver cancer itself.

Aluna works with hospitals to build carefully curated cancer datasets and evaluation sets so that medical AI models can be trained and tested properly.

Ananya Health makes a portable freezing device that lets a nurse in any clinic treat abnormal cervical cells before they turn into cancer, at a tenth of the usual cost.

AOA Dx is developing a blood test to detect ovarian cancer early, based on fat-sugar molecules called glycolipids that tumours shed.

Arpeggio Bio uses AI on large RNA-sequencing datasets to find drugs against transcription factors, the hard-to-hit proteins that switch cancer genes on, with a first programme in immunotherapy-resistant melanoma.

Asher Bio engineers immune-stimulating molecules that switch on only one type of immune cell, such as killer T cells, so the drug is more potent against tumours and causes fewer side effects.

Bikanta tried to use tiny fluorescent diamonds as imaging probes to spot cancer earlier. It went through Y Combinator in 2014 and is no longer operating.

BillionToOne counts DNA molecules in blood far more precisely than standard sequencing. Known first for prenatal testing, it now sells liquid biopsy tests that read a tumour's mutations and track treatment from a blood draw.

Blank Bio trains AI models on tumour RNA to predict how a disease will progress and which patients will respond, so drug companies can run smarter cancer trials.

CCG.ai built software to combine genomic and clinical data so oncologists could choose and adjust treatment sooner. The company is no longer active.

Clear Gene is developing tests that tell a surgeon, quickly, whether all of the cancer was removed during an operation.

Cofactor Genomics reads the RNA of a tumour sample to model the immune cells inside it, and uses that to predict who will respond to immunotherapy.

Data Driven Bioscience is building a faster, cheaper sequencing-based way to diagnose and subtype cancers, starting with lymphoma.

Delee, from Mexico, built a blood-testing device that catches circulating tumour cells to help diagnose cancer earlier and follow treatment.

Digistain shines infrared light through a breast tumour slide and reads its chemical fingerprint to score how likely the cancer is to return, giving a result in a day rather than weeks and at a fraction of the cost of genomic tests.

Eris Biotech is making small-molecule pills that release the brakes on the immune system inside tumours, starting with mesothelioma.

Y Combinator Spring 2026 start-up building a programmable DNA-targeting drug: a molecule that recognises a cancer-specific DNA sequence inside a cell and triggers a kill switch only there.

Granza Bio is turning the natural attack particles that killer immune cells fire at their targets into programmable medicines that can re-arm exhausted immune cells against cancer.

Guardian Bio turns a patient's stem cells into dendritic cells, the immune system's teachers, trains them against that patient's tumour and gives them back to restart the immune attack.

Harmonic Discovery designs single drugs that hit several disease-driving proteins at once, arguing that cancer is rarely caused by one broken protein.

HistoWiz is a mail-in pathology laboratory for cancer researchers: send tissue, get scanned digital slides and expert readings online within days.

Humane Genomics builds artificial cancer-killing viruses that only infect and only multiply inside tumour cells, starting with a rare liver cancer in children.

Isabl reads a tumour's whole genome and transcriptome instead of a small gene panel, and its software turns that into a report clinicians can use.

iSono Health makes a wearable ultrasound scanner that images the whole breast automatically in about two minutes, with software to flag suspicious masses, for use outside specialist imaging centres.

Kernal Bio designs mRNA medicines that switch on only in the intended cell type, for example programming a patient's T cells against cancer inside the body without taking them out first.

Known Medicine breaks a patient's tumour into thousands of tiny three-dimensional tumours, treats each with a different drug, and uses AI to predict which drugs will work. It has been acquired.

Kopra Bio engineers viruses that teach the immune system to kill cancer, starting with glioblastoma, the most aggressive brain cancer.

Liyfe built a digital nurse that checks in with cancer patients between visits, scores their symptoms and flags who needs a call, to cut emergency visits.

Luminate makes a wearable device that presses gently on the scalp during chemotherapy to stop the drug reaching the hair follicles, so patients keep their hair.

theMednet is an online question-and-answer network where oncologists ask hard clinical questions and get answers from named experts, with AI now helping to search the archive.

Modulari-T designs synthetic genes that let engineered immune cells sense their surroundings and change behaviour, so cell therapies can find and kill cancer more reliably.

NearWave makes a handheld near-infrared scanner that shows within days whether a breast cancer is responding to chemotherapy before surgery, so patients are not left on a treatment that is failing.

Notable Labs ran an automated laboratory in Foster City that exposed a patient's own leukaemia or lymphoma cells to combinations of approved drugs and measured which killed them, to predict which treatment would work. It listed on Nasdaq in 2023, but its website was offline in September 2026, so it is recorded as wound down.

Nucleo builds AI agents that help oncologists and radiologists pull measurements and insights out of CT scans to characterise tumours and guide treatment.

Numen is working on earlier cancer detection, with the stated goal of a future without cancer deaths.

Omanta runs a personal research programme for one patient at a time, for people with cancer who have exhausted standard options, coordinating tissue, testing, multi-omic analysis and a treatment strategy with their doctors.

Oncobox profiles a tumour's genes and the activity of its signalling pathways, then ranks approved targeted drugs by how likely they are to work for that patient.

Origin is building the largest dataset of how individual cancer cells from real patients respond to drugs and genetic medicines, to license to pharma and AI companies.

Persephone studies the gut bacteria of cancer patients and designs live bacterial medicines meant to make immunotherapy work for more people.

REPROSENT is a patient app that helps people with cancer and their families track symptoms and treatment, and turns what they record into alerts for their care team.

ResistanceBio set out to design drugs that stop cancers becoming resistant to treatment. The company is no longer active.

Reverie Labs used machine learning to design cancer drugs that could cross into the brain. It is no longer operating.

Serinus uses an explainable AI model of cancer cell biology to work out how tumours evade a drug, and to design combinations that block the escape routes.

Shasqi puts an inert form of a chemotherapy drug into the bloodstream and switches it on only inside the tumour, using a chemical reaction called click chemistry, so the rest of the body sees far less of the drug.

Signatur turns complex gene-expression tests, starting with breast cancer prognosis, into simple PCR tests that a local hospital lab can run in days.

SynsoryBio builds protein drugs with built-in if-then logic: they stay switched off until they sense they are in diseased tissue, a way to make immune-activating and cell-killing biologics usable at doses that would otherwise harm healthy tissue. It is preclinical, with a small team and no recorded financing.

TareBio reads the immune cells already fighting a cancer in a patient's body and works backwards to design a vaccine that makes more of them, in weeks rather than years.

Tradewind is developing a single antibody for ovarian cancer that both attacks the tumour cells directly and helps the immune system attack them too.

Triomics builds AI agents that read the unstructured parts of a cancer patient's records, such as pathology reports and scanned faxes, to match patients to trials and prepare visits automatically.

Valink designs antibody-drug conjugates that grab two targets at once, aimed at solid tumours that immunotherapy does not reach.

Velorum is developing cancer drugs that exploit heme, the iron-carrying molecule tumours depend on in unusual ways.

X-Zell isolates rare abnormal cells from a blood sample and stains and images them so a pathologist can look for signs of cancer without a tissue biopsy.

Yemaachi collects and sequences cancer samples from across Africa, the most genetically diverse population on earth, so that cancer tests and drugs work for African patients too.

zPREDICTA grew patient-derived tumour tissue in three-dimensional cultures that reproduce the tumour's surroundings, so drug companies could test cancer drugs on something closer to a real tumour. It was later acquired.