CT Scans Save Lives. But as They’re Used More Often, How Many Are Causing Cancer?

As CT scans become increasingly prevalent in U.S. health care, radiation from the often lifesaving imaging technique could come with a steep long-term cost: tens of thousands more cases of cancer, according to a UCSF study published Monday in JAMA Internal Medicine.

Researchers analyzed CT scan data and estimated that the 93 million scans performed in 2023 — nearly one for every three people in the country — could lead to approximately 103,000 future cancer cases over patients’ lifetimes. A similar analysis years earlier estimated that routine CT exposures in 2007 would lead to about 29,000 future cancer cases.

The projection reflects both the sharp rise in how often CT scans are used and the higher radiation doses many patients now receive.

“Most patients will not develop cancer from a CT scan,” said Dr. Rebecca Smith-Bindman, lead author and professor of epidemiology and biostatistics at UCSF. “But if you’re not getting any benefit, then even a small risk is unacceptable.”

During a CT scan, a patient lies on a table that slides through a large, donut-shaped machine. Inside, an X-ray tube rotates around the body, capturing images of thin slices — like layers of a loaf of bread. A computer then assembles those slices into a 3D image that doctors use to spot injuries, tumors or other abnormalities.

The process exposes the body to radiation, which can alter DNA. Most of the time, the body repairs these changes. However, occasionally, a cell keeps dividing abnormally, and over time, that can lead to cancer.

“There’s DNA damage response pathways that are active all the time that take these insults, and they repair your DNA,” said Dr. Patricia Nguyen, a cardiologist at Stanford who was not involved in the study. “But as we age, those responses aren’t as strong … and that’s when damaged cells can turn into cancer.”

The researchers used the National Cancer Institute’s Radiation Risk Assessment Tool, which estimates cancer risk based on scan type, patient age and radiation dose. The model draws on historical data, including studies of Japanese atomic bomb survivors. While the methodology is widely accepted in public health research, it has limitations.

“They don’t know what’s happening biologically to the actual patient,” Nguyen said. “They are just basing it on formulas or retrospective data … we don’t know what happens in each individual.”

CT scans are often essential — especially in emergencies like a serious car crash, when doctors need to quickly check for internal bleeding. But they’re also used far more often than necessary. Some are ordered for mild headaches or routine infections, when they’re unlikely to change a diagnosis or improve care.

“Our estimates put CT on par with other significant risk factors, such as alcohol consumption and excess body weight,” Smith-Bindman said. “Reducing the number of scans and reducing doses per scan would save lives.”

Use of the technology has climbed 35% over the last decade. A mix of pressure and convenience drives the increase: Doctors have less time with patients, imaging is typically reimbursed by insurance, and patients often expect a clear answer. Meanwhile, as medicine shifts away from open surgeries to less invasive procedures, physicians increasingly rely on CT scans to plan and guide treatment.

The researchers found that certain types of CT scans — especially of the abdomen and pelvis — were responsible for a large share of the projected cancer cases. Many involved “multiphase” imaging, where a patient is scanned two or three times in one session. While that approach can be clinically useful, the study found it’s often unnecessary.

The takeaway isn’t to avoid CT scans but to be more thoughtful about when they’re used — and why. For some patients, asking a simple question may be the best safeguard.

“Every test has potential risks,” Nguyen said. “And patients should ask: Why am I getting this? What will the result change?”