Scientists have successfully grown mini versions of patients' tumours in a lab – and then tested them against dozens of drugs to find the best possible treatment.
The technique could mean patients get effective chemotherapy sooner – and are spared unnecessary side effects from drugs that do not work.
Scientists have hailed the uniquely personalised treatment as "extremely promising".
Doctors currently treat patients with an element of guesswork, only knowing for sure whether a drug is working when they return for a CT scan several months into treatment.
But scientists at the Institute of Cancer in London were able to screen lab-grown replicas of patients' tumours against a range of drugs in just two weeks.
Dr Nicola Valeri, the study leader and a consultant oncologist at the Royal Marsden Hospital, told Sky News: "At the moment doctors are blind to whether patients will respond or not.
"Time is crucial and with this technique we get results almost in real time. It is an extremely promising way to predict whether a drug would work for a patient."
The researchers took biopsies from 71 patients and grew them in a special gel over a period of 3-4 weeks. The mini-tumours were then split into small pieces and tested against a battery of drugs.
Results in the journal Science show the technique was able to correctly identify an effective treatment in between 88% and 100% of tumours.
Significantly, the technique would avoid patients suffering side effects from ineffective drugs.
Cara Hoofe, 32, endured muscle spasms, fatigue and hairloss from a drug that failed to work against her bowel cancer.
She said: "You don't want to put yourself through all that for something that doesn't work.
"Time is everything. As soon as you can find a treatment that works for you it gives you more hope of being able to get further down the line."
The new technique could also save the NHS money by avoiding treatment of patients with tumours that will not respond to the drug.
More from Tech
New cancer drugs can cost as much as £100,000 and the NHS will not prescribe some because they are judged not to be cost effective.
Professor David Cunningham of the Royal Marsden and Institute of Cancer Research said: "This promising research moves us forward in the field of personalised medicine and should ultimately lead to smarter, kinder and more effective treatments for patients."