Press "Enter" to skip to content

Increased Risk of Chemo-Associated AML for Nearly All Solid Cancers

Cancer patients treated with chemotherapy for almost every type of solid tumor were found to be at increased risk for developing therapy-related myelodysplastic syndrome or acute myeloid leukemia (MDS/AML), a population-based cohort study found.

Patients receiving cytotoxic agents in the modern era for 22 of 23 different solid tumor types had anywhere from a 1.5- to 10-fold increased risk for treatment-related MDS/AML, reported Lindsay Morton, PhD, of the National Cancer Institute (NCI) in Bethesda, Maryland, and colleagues in JAMA Oncology.

And the researchers also found that risk for therapy-related MDS/AML, a group of secondary neoplasms associated with poor outcomes, was greatest among younger patients.

“We’ve known for a long time that the development of myeloid leukemia is a very rare adverse effect of some types of cancer treatments that damage cells,” Morton said in a statement from the NCI. “There have been many changes in cancer treatment over time, including the introduction of new chemotherapy drugs and drug combinations, but we didn’t know what the risk of therapy-related leukemia looked like for patients since these changes were made.”

In an editorial accompanying the study, Shyam Patel, MD, PhD, of the Stanford Cancer Institute in California, wrote that information on outcomes associated with treatment-related MDS/AML is “highly desirable to oncologists because improved survival from primary cancers is associated with increased risk for secondary neoplasms later in life.”

For the cohort study, the researchers looked at 1,619 patients with therapy-related MDS/AML from among more than 700,000 adults diagnosed with a first primary solid cancer from 2000 to 2013 who went on to receive chemotherapy and survived at least 1 year, as reported to the Surveillance, Epidemiology, and End Results (SEER) Program. Additional descriptive analyses were conducted using linked SEER and Medicare data on 165,820 older adults who received initial chemotherapy for a first primary cancer during this stretch of time.

The median overall survival after diagnosis of treatment-related MDS/AML was 7 months, with 78.4% of the patients diagnosed having died during the study period.

Therapy-related MDS/AML occurred significantly more often than expected after initial chemotherapy for all primary solid tumors examined, except for colon cancer. The highest standardized incidence ratio (SIR) occurred in cancer of the bone (SIR 39.0), soft tissue (SIR 10.4), and testis (SIR 12.3), all of which were typically diagnosed in younger patients, the researchers noted. SIRs ranged from 5 to 9 for peritoneum, smell cell lung, ovary, fallopian tube, brain, and central nervous system cancers, and from 1.5 to 4 in all remaining cancers examined.

Patel noted that the data on risk for treatment-related MDS/AML in younger patients is likely the most clinically relevant and leads one to “speculate that the diagnosis and treatment of a solid tumor at an early age permits a longer latency period during which clonal evolution can occur in a previously damaged bone marrow compartment.”

However, he also wrote that continued follow-up of these trends may reveal a relative decrease in the risk for therapy-related MDS/AML “given the recent acceleration of regulatory agency approvals of novel anticancer therapies whose clinical activity is defined by specific molecular aberrations in contrast with approval of new applications of the same cytotoxic chemotherapies.”

Patel posited several reasons for the lack of risk associated with colon cancer, including diagnosis at a later age, wherein “subclinical clonal hematopoiesis induced by chemotherapy may not have ample time to evolve into frank disease.”

There was a greater risk for therapy-related MDS/AML within 5 years of diagnosis compared with 5 or more years after diagnosis for esophageal, gastric, soft-tissue, breast, uterine corpus, and testis cancers. However, for 15 of the 23 solid tumors examined, risk for MDS/AML remained statistically significant more than 5 years after the primary diagnosis.

The linked SEER-Medicare data revealed that the proportion of claims that included a known leukemogenic agent increased from 57% of claims in 2000-2001 to 81% in 2012-2013. The biggest driver of this increase was platinum agents — with an increase from 35% to 59% during that time period — specifically among gastrointestinal cancers such as esophageal, stomach, colon, and rectum cancers.

Looking at U.S. cancer statistics, the researchers estimated that in 2018 about 360,000 adults will be diagnosed with one of these 23 solid tumors, will undergo initial chemotherapy, and survive at least 1 year. Of these patients, they estimated that three-quarters (73.0%) of the 714 expected cases of therapy-related MDS/AML occurring within 5 years of diagnosis would be attributable to chemotherapy, with others due to radiation therapy and other factors.

“While advances in cancer treatment approaches have improved the prognosis for many types of cancer, the number of patients at risk of developing rare, therapy-related leukemia after cancer chemotherapy in the modern treatment era has markedly expanded,” Morton said. “Assessments of treatment risks and benefits should balance these risks and other adverse effects of chemotherapy against potential gains in survival following treatment for the initial solid cancer.”

Patel suggested four criteria for appropriate risk assessment of chemotherapy administration for solid tumors: the probability of a chemotherapeutic agent inducing a disease-initiating event, likelihood that pathogenic events will occur at a later time, the benefit gained with respect to solid tumor regression, and the consequences of forgoing chemotherapy.

“Because patients who receive chemotherapy for solid tumors are expected to live longer, the incidence of treatment-related myeloid neoplasms is likely to increase, pending the further availability of non-DNA-damaging agents such as targeted therapy,” Patel wrote. “For these reasons, a holistic approach to risk assessment should be undertaken.”

Morton and colleagues reported no conflicts of interest.

Patel reported no conflict of interests.