番茄社区

Expert Critique

FROM THE ASCO Reading Room

Amit Reddy, MBBS Postdoctoral Fellow Mitchell Cancer Institute, University of South Alabama

It is not completely understood how to identify patients at highest risk or how to prevent therapy-related myeloid neoplasms (tAML), and at present, the available treatments are not curable. Some might say that the risk factors may include high-dose chemotherapy, older age, longer duration of therapy, and combination treatment regimens. Recently there have been some studies for the possible origin and prevention of tAML. The first study noted below included using molecular barcode sequencing of 32 genes and targeted gene sequencing on bone marrow samples. The investigators scrutinized the clonal evolution from clonal hematopoiesis to the development of therapy-related myeloid neoplasms. The researchers suggested, based on the results, that clonal hematopoiesis could be used as a predictive marker to identify patients with cancer who are at risk for therapy-related myeloid neoplasms. The second study was similarly done, but there were different results in the distribution of clonal hematopoiesis gene mutations between the cases and controls, suggesting that the therapy-related malignancies may vary by type of mutation.

In summary, the real significance of the two studies will be when therapeutic regimens exist that can eliminate nascent clonal hematopoiesis, and as a result prevent tAML and other myeloid neoplasms.

Full Critique

One of the greatest ironies of modern oncology is that potentially curative therapies also have the small but very real potential for causing great harm years or decades later.

Or as , of Dana-Farber Cancer Institute in Boston put it: "Therapy-related myeloid neoplasms are an iatrogenic tragedy.

"Almost no one is cured of therapy-related myeloid neoplasms with presently available treatment approaches; even for patients who are young enough and fit enough to undergo allogeneic hemopoietic stem cell transplantation, long-term overall survival at 3 to 5 years is less than 25%," Steensma wrote in a recent editorial in accompanying two studies that shed light on the possible origin and prevention of therapy-related acute myeloid leukemia (tAML) and other myeloid neoplasms.

In those studies, investigators from the University of Texas MD Anderson Cancer Center in Houston and Moffitt Cancer Center in Tampa, Fla., report that pre-leukemic clones that escape cytotoxic therapies for non-myeloid cancers may increase the risk for therapy-related myeloid neoplasms. The findings indicate that it may be possible to identify and eradicate the clones in at-risk patients, Steensma suggested.

Clonal Hematopoiesis

In the MD Anderson study, , and colleagues performed a comparing patients who had been treated for a primary non-myeloid cancer at their center from 1997 through 2015 who later developed tAML or therapy-related myelodysplastic syndromes (tMDS) (cases) and age-matched patients treated with chemotherapy for lymphoma who did not develop tAML or tMDS during a minimum of 5 years of follow-up.

The investigators examined peripheral blood samples taken at the time of a primary cancer diagnosis and bone marrow samples collected from cases at the time of a therapy-related myeloid neoplasm diagnosis. They used molecular barcode sequencing of 32 genes to detect clonal hematopoiesis in pre-treatment peripheral blood, and targeted gene sequencing on bone marrow samples from cases to investigate clonal evolution from clonal hematopoiesis to the development of therapy-related myeloid neoplasms.

In 10 of 14 cases, the baseline peripheral blood samples contained small, preleukemic clones -- clonal hematopoiesis -- prior to cytotoxic chemotherapy. In contrast, clonal hematopoiesis was detected in only 17 of 54 controls prior to chemotherapy.

Takahashi and colleagues also examined clonal hematopoiesis in an external cohort of patients treated with cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) with or without melatonin at their center from 1999 through 2001 as part of a clinical trial.

Of the 74 patients in this cohort, five (74%) developed tAML or tMDS. Of these five patients, four (80%) had baseline clonal hematopoiesis. Of the remaining 69 patients in the external cohort who did not develop therapy-related myeloid neoplasms, only 11 (16%) had baseline clonal hematopoiesis.

The investigators created a multivariate model based on data from the external cohort, which showed that clonal hematopoiesis was significantly associated with risk for therapy-related myeloid neoplasms (hazard ratio 13.7, P = .013).

"Our results suggest that clonal hematopoiesis could be used as a predictive marker to identify patients with cancer who are at risk of developing therapy-related myeloid neoplasms."

Unhealthy CHIPs

In the Moffitt research, , and colleagues conducted a similar, showing that patients with therapy-related myeloid neoplasms were more likely than controls to have clonal hematopoiesis of indeterminate potential, or CHIP. As in the study by Takahashi and colleagues, CHIP was often present before exposure to chemotherapy.

To do this, the team first identified cases from an internal biobank of 123,357 patients. Included were all patients who were diagnosed with a primary cancer, treated with chemotherapy, and developed a subsequent therapy-related myeloid neoplasm. The patients had to be age 70 or older at the time of either a primary or therapy-related cancer diagnosis with peripheral blood or mononuclear samples collected before the diagnosis of the second cancer.

For controls, Gillis et al identified patients diagnosed with a primary malignancy at age 70 or older who had chemotherapy but did not develop tAML or tMDS. Each case was matched with four or more controls selected for age at diagnosis, sex, primary tumor type, smoking status, chemotherapy drug class, and duration of follow-up.

In cases where paired CHIP and therapy-related myeloid neoplasm samples were available, the investigators used sequential targeted and whole-exome sequencing to assess clonal evolution.

Among all patients (13 cases and 56 controls), CHIP was seen at baseline in 23 (33%). This contrasts with findings from previous studies showing a prevalence of CHIP among older patients without cancer of about 10%, Gillis and colleagues noted.

CHIP was present in eight of 13 cases (62%) versus 15 of 56 controls (27%; P = .024) The odds ratio for therapy-related myeloid neoplasms with CHIP was 5.75 (P = .013).

"We recorded a significantly higher prevalence of CHIP in individuals who developed therapy-related myeloid neoplasms (cases) than in those who did not (controls); however, around 27% of individuals with CHIP did not develop therapy-related myeloid neoplasms, suggesting that this feature alone should not be used to determine a patient's suitability for chemotherapy," the investigators emphasized.

They noted that that there were differences in the distribution of CHIP-related gene mutations between cases and controls, suggesting that the risk for therapy-related malignancies may vary by the type of mutation.

In his editorial, Steensma wrote that the real importance of the two studies will come when therapies exist that can effectively eradicate nascent clonal hemopoiesis, thereby preventing therapy-related myeloid neoplasm evolution in at-risk patients."