Curing cancer
Many cancers are curable, but spread-out cancers still kill nearly 10 million people a year.
Open in the interactive tree →Cancer is hundreds of diseases in which cells evolve to escape control, so there is no single cure. Surgery, radiation, chemotherapy and, since the 2010s, immunotherapy cure many early cancers, but cancers that have spread usually return as tumors evolve resistance.
As of October 2026
The American Cancer Society's 2026 report says about 70 percent of people diagnosed with cancer in the US now survive five years and credits 4.8 million lives saved, but incidence is rising for many common cancers. WHO counts nearly 10 million cancer deaths a year, about one in six deaths worldwide, and estimates that about 38 percent of cancers are preventable. On 19 August 2026 Merck and Moderna reported the first positive phase 3 trial of an individualized mRNA cancer vaccine (INTerpath-001, 1,137 patients with resected melanoma, added to pembrolizumab; company press release, full data not yet public).
What is missing
- Ways to stop or eliminate metastases, which cause most cancer deaths
- Understanding and countering how tumors evolve resistance to each drug
- Tests that find cancers early and reliably across organs without many false alarms
- Treatments that reach and work inside the hostile environment of solid tumors
- Affordable access: many people lack screening and modern drugs
Becomes possible once solved
- Cancer vaccines that prevent disease
- Treating cancer as a manageable chronic condition
- Early detection from a blood test
Open steps
- Stopping metastasis Low AI leverageUnderstand how cancer cells leave a tumor, survive in blood and seed other organs, and find ways to stop or clear metastases.
- Beating drug resistance Medium AI leveragePredict and counter how a tumor evolves resistance to each drug, ideally before treatment starts.
- Early detection across organs High AI leverageReliable tests that find many cancers early without many false alarms and are shown in trials to save lives.
- Treating inside solid tumors Medium AI leverageMake drugs and immune cells reach and work inside the hostile environment of solid tumors.
- Choosing vaccine targets Medium AI leveragePick, for each patient, the tumor mutations that the immune system will actually attack, to build individual vaccines.
Where AI could help
Medium AI leverage. AI already helps find cancers and propose targets, but metastasis, drug resistance and slow trials are limited by biology and time.
- Reading mammograms, pathology slides and scans to catch cancers earlier and cut reader workload
- Generating and ranking drug combinations and targets for resistant tumors
- Designing personalized neoantigen vaccines and antibodies faster
- Matching patients to trials from medical records
Shown so far
- In January 2026 The Lancet published the MASAI trial of about 106,000 women: AI-supported mammography screening had fewer interval cancers (1.55 versus 1.76 per 1,000) than standard double reading. source
- In October 2025 Google and Yale reported that the C2S-Scale 27B model predicted a drug-plus-interferon combination that raised antigen presentation by roughly 50% in human cells in the lab (an early lead, according to Google). source
- In February 2025 Google described an AI co-scientist whose proposed acute myeloid leukemia drug candidates inhibited leukemia cell lines in laboratory tests (Google-authored report). source
Prerequisites
- Chemotherapy1942
- Oncogenes1976Cancer as damaged genes underlies targeted drugs and the search for cures
- Checkpoint Immunotherapy2014–2018
- CAR-T cell therapy2017
- AI in medical diagnosis2018–2026
- mRNA technology2020