Mathematical optimization models to improve radiation therapy for cancer

Engineering researchers at the University of Arkansas have developed mathematical optimization models that will make radiation treatment plans safer and more efficient than conventional plans.

Conventional radiation therapy uses a single, cumulative treatment plan that neglects changes in tumor geometry and biology over time. However, recent technological advances have made it possible to capture these changes throughout the course of treatment.

Working with geometric and biological data gathered from the most advanced technology used to capture tumor changes, the researchers achieve optimization in terms of delivering the maximum dose of energy to the tumor without undue risk to surrounding healthy tissues. By incorporating the most recent information about tumor geometry and biology, the researchers' models will help optimize radiation treatment on a per-session and cumulative basis.

"This is one of the first efforts to develop a method of radiation therapy optimization that uses biological information while maintaining both cumulative and per-fraction, or per-session, dose constraints," said Behlul Saka, who recently completed his doctorate in industrial engineering. "Our optimization models may be used to generate treatment plans based on tumor biology prior to treatment, but also may react to changing tumor biology throughout treatment."

More than one million Americans are diagnosed with cancer each year, and more than half of these patients receive radiation therapy at some point during treatment. This type of therapy destroys cancer cells or slows their rate of growth by applying high-energy rays to tumors. Conventional radiation therapy involves shooting radiation from different beam angles with a linear accelerator that rotates around the patient. The variety of beam angles limit radiation exposure and thus spare healthy tissues exposed from particular angles while concentrating their combined effect on the tumor. Care is usually taken to assure homogeneous distributions of radiation across the tumor.

Intensity Modulated Radiation Therapy, or IMRT, addresses the limitations of conventional radiation therapy by delivering small parts of rays, called beamlets, directly to specific sections or particular points on a tumor while also sparing healthy tissue. In other words, with IMRT, the radiation dose conforms more precisely to the three-dimensional shape of the tumor.

Saka's research adds to the advantages of IMRT by developing optimized treatment plans in response to changes in tumor geometry measured against cumulative and per-session doses. To build the models, he relied on anonymous, archived patient data.

Saka tested the models on two simulated cases of lung cancer. These cases demonstrated clinically significant improvements in tumor response over time. On a different simulated test case, by comparing tumor hypoxia information, the models demonstrated significant improvements in controlling a tumor. Hypoxia refers to an inadequate oxygenation of blood, which contributes to cell growth and resists radiation. These gains in tumor control and average doses were significant throughout treatment. The models also displayed the volatility of tumor control relative to emerging changes in tumor hypoxia values.

"This work suggests that a homogeneous approach to dose distributions across the tumor can be greatly improved upon," Saka said. "Non-homogeneous distributions - that is, varying doses according biological responses of the tumor - apparently can be much more effective. Of course, clinical trials will be required to confirm these simulation findings."

Saka continues his research on improved cancer treatment planning as a research scientist at Elekta, a health-care company that develops innovations and clinical solutions for treating cancer and brain disorders. His work at the University of Arkansas was completed under the direction of Ron Rardin, Distinguished Professor and holder of the John and Mary Lib White Systems Integration Chair. Rardin and his students have been working on radiation planning optimization for more than a decade.

"I believe Behlul's work points the way to much more refined treatment planning, by exploiting more of the information available as the tumor changes through 25 to 50 daily treatment sessions," Rardin says.

Source: University of Arkansas