Primary tab navigation

Pin-point cancer therapy
Targeting only bad cells with proton radiation

Proton radiation therapy is one of the more successful treatments for cancer. It works by disrupting cancer cells’ DNA, preventing them from reproducing. The therapy uses beams generated by accelerating protons to two-thirds the speed of light. Protons can penetrate deep into a body without disrupting healthy cells – and only impact the cancer cells as they slow down at the end of their path.

There were an estimated 12.7 million cancer cases around the world in 2008 (latest year available). This number is expected to increase to 21 million by 2030.

World Cancer Research Fund International

But due to cost and the need for pinpoint accuracy, the number of patients treated annually at the 10 proton beam therapy treatment centers in the United States is rather modest at only a few thousand per year.

IBM Research scientists in Austin, with its partners at the University of Texas M.D. Anderson Medical Cancer Center in Houston, are working to reduce the cost of identifying and treating cancer patients with proton radiation therapy by shortening the time required for treatment planning from several days to a few minutes. By creating advanced algorithms running on IBM POWER7 systems and utilizing IBM ILOG optimization software, they can simulate how far the proton beams travel in a body 1,000 times faster than conventional methods; and can solve the complex optimization problem of treatment planning 500 times faster than current practice.

How proton radiation therapy works

Today, doctors review MRI scans and manually outline the tumor in a patient’s body, which requires a high degree of experience and can take many hours. Once the tumor is found, a treatment plan is formulated to direct the proton beam at the correct angle and direction to hit the tumor.

current state of proton radiation therapy

Proton radiation beams are generated from a particle accelerator, similar to what physicists use to study the properties of matter, and a series of electromagnets steer the beams to a rotating nozzle that moves around a patient’s bed. Based on the treatment plan, the nozzle delivers a number of beams from different angles and with different energies.

The entire complicated planning, review and quality assurance process takes specialists and highly trained physicists days to calculate... but the cancer doesn't sit still, and can change shape or move in the body – which would make the plan unworkable.

Sani Nassif, IBM Research scientist

The beam energy has to be such that the protons stop at the correct location and depth in the patient's body. Due to the impractically long time for today's computers to calculate accurate models, radiation oncologists are forced to use approximate models to generate these treatment plans.

Because treatment plans use these approximate models, doctors first test the beam settings in a tank of water. Only after beam measurements on the water satisfy the requirements for a particular patient is the treatment applied. This additional quality assurance process still takes quite a bit of time and wastes radiation machine time that could be used for treating patients.

Making treatment planning faster

The goal is to complete these treatment planning calculations in the time it takes to move a cancer patient from an MRI or CT scan into the radiation therapy treatment room – mere minutes. And so far, scientists at IBM Research - Austin have simulated the automatic creation of treatment plans that cut the entire planning process down to about 15 minutes. Continued work with partners such as MD Anderson will soon make this a reality in the fight against cancer.

proton beam therapy with IBM technology

Share this story

Explore this topic

Meet the researchers

  • Sani Nassif thumbnail image

    Sani Nassif

    Research Scientist,
    IBM Research - Austin

  • Anne Gattiker thumbnail

    Anne Gattiker

    Research Scientist,
    IBM Research - Austin

  • Damir Jamsek thumbnail

    Damir Jamsek

    Research Staff Member,
    IBM Research - Austin

  • Thomas Osiecki thumbnail

    Thomas Osiecki

    System Software Research,
    IBM Research - Austin

  • Evan Speight thumbnail

    Evan Speight

    Research Staff Member,
    Master Inventor,
    IBM Research - Austin

  • Cliff Sze thumbnail

    Cliff Sze

    Research Scientist,
    IBM Research - Austin

Share this story