Paper Details

ABSTRACT

Background: Genomic TP53 alterations in acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) are associated with poor outcomes.  Point mutations (TP53^mut) and chromosomal deletions, as indicated by loss of 17p [del(17p)] each impair TP53 activity, but the relative impacts of these mechanisms are not well defined. Given uncertainty regarding the relative prognostic impact of these alterations, we sought to compare outcomes among patients with TP53^mut and del(17p), while accounting for clinical covariates, treatment intensity, and receipt of hematopoietic cell transplantation (HCT).

Methods: We retrospectively analyzed 388 adults with newly diagnosed TP53-altered AML or MDS/AML treated at the University of Washington/Fred Hutchinson Cancer Center. TP53^mut variant allele frequency (VAF) was defined by next-generation sequencing (NGS), while del(17p) and complex karyotype (CK) were defined as present if identified in ≥2 metaphase karyotypes. Overall survival (OS) and event-free survival (EFS, events = failure to achieve remission, relapse, or death) were evaluated by Cox models. Additional abstracted covariates included diagnosis, age, performance status (ECOG), white blood cell (WBC) count, platelet count, treatment (IC vs hypomethylating agent-based regimen [HMA] vs none), and receipt of HCT. We modeled HCT as both a landmark and a time-dependent covariate.

Results:  We identified 388 patients with TP53-altered AML or MDS/AML. The median age was 66 yrs (range 18 - 100); 253 (66%) had AML, 133 (34%) had MDS/AML, 59% had ECOG 0-1, and median treatment related mortality score (TRM) = 7.9. Patients were classified as follows: TP53^mut without del(17p) (n = 190, 49%), isolated del(17p) (n = 45, 12%), and both TP53^mut/del(17p) (n = 151, 39%). Baseline sex, ECOG, diagnosis, WBC, platelet count, and initial treatment intensity (including receipt of no therapy) did not differ among groups. The median OS was 8.5 months. Three- and five-year OS were 10.7% and 7.1%, respectively.

In univariate analysis, compared to isolated TP53^mut, TP53^mut/del(17p) had worse OS (HR 1.37, p = 0.0065) and EFS (HR 1.28, p = 0.032), whereas there was no significant difference in OS or EFS comparing del(17p)/TP53^wt vs TP53^mut w/o del(17p) (HR 1.05, p = 0.80 and HR 1.08, p = 0.67). Additional gene mutations were not associated with EFS or OS. Univariate analyses also found an association with TP53 VAF and OS (HR 1.08, p = 0.006). When controlling for TP53 VAF, there was no association for OS and CK, del(17p) or a second TP53^mut (i.e. biallelic TP53^mut, HR 1.12, p = 0.57). Univariate analyses identified significant associations between OS and age (HR 1.01, p = 0.001), ECOG (HR 1.72, p <0.001), diagnosis (MDS/AML vs AML, HR = 0.68, p < 0.001), HCT (HR 0.45, p < 0.001).  Initial HMA therapy was associated with improved OS compared with IC (HR 0.77, p = 0.027).

In multivariable analyses, treatment intensity significantly modified the association between HCT and OS. Patients receiving IC without HCT had the poorest outcomes (HR 4.23, p<0.001), whereas patients receiving HMA without HCT also had inferior survival (HR 2.04, p=0.003), both compared with patients receiving IC followed by HCT. Among patients who underwent HCT, OS did not differ significantly according to initial treatment intensity (HR 1.08, p=0.81). Concurrent TP53mut/del(17p) remained independently associated with inferior OS (HR 1.43, p=0.01).

Conclusions: Among patients with newly diagnosed TP53-altered myeloid neoplasms, isolated TP53 mutation and isolated chromosome 17p deletion conferred similar risk, whereas patients with concurrent TP53 mutation and chromosome 17p deletion had inferior survival. We observed a significant interaction between treatment intensity and HCT with respect to OS. Patients initially treated with IC who did not undergo HCT experienced particularly poor outcomes, whereas among patients who underwent HCT, OS did not differ according to whether initial therapy was IC or HMA. Given the retrospective design, these findings should be interpreted cautiously because residual confounding and selection bias—including differences in transplant eligibility and patients who died or became ineligible before HCT—may contribute to the observed associations. Prospective studies are needed to define the optimal integration of IC, HMA-based therapy, and HCT in TP53-altered AML and MDS/AML.