High prevalence and inferior long‐term outcomes for TP53 mutations in therapy‐related acute lymphoblastic leukemia

The entity of therapy-related ALL has been used to describe acute lymphoblastic leukemia (ALL) diagnosed following exposure to cytotoxic therapies including chemotherapy and/or radiation for unrelated preceding malignancies.1-3 This nomenclature was adapted in parallel with that of therapy-related-myeloid neoplasm (t-MN) which is a well-established entity and has been included in 2016 WHO classification of myeloid neoplasms and acute leukemia.4 In contrast, the existence of t-ALL remains debatable as it is less well-described in the literature and due to the paucity of wide genomic studies. We and others have previously described the entity of therapy-related ALL to be relatively common when defined in its clinical context of prior cytotoxic exposure.1, 3 There are additionally specific observed patterns amongst t-ALL patients including older age at diagnosis, a female predominance when compared to de novo ALL,1, 3 and a predilection to carry high-risk cytogenetics,1-3, 5 similar to therapy-related MN. Nevertheless, molecular studies illustrating the underlying genetic alterations beyond cytogenetics in t-ALL remain scarce.2 A recently reported study of a 21 patient cohort with t-ALL illustrated that TP53 alterations were more prevalent in t-ALL as compared to de novo B-cell ALL (42% vs. 23%).2 This is a crucial finding given that TP53 mutations (TP53m) are considered one of the key features for t-MN and contribute largely to their inherently dismal prognosis.6 Here, we retrospectively reviewed adults with a diagnosis of therapy-related ALL treated at our center. We defined t-ALL as ALL diagnosed after exposure to chemotherapy or/and radiation for unrelated malignancy. We only included t-ALL cases in which their leukemia was evaluated by molecular analysis with our in-house targeted next-generation sequencing (NGS) panel. Table S1 illustrates targeted genes. The study was approved by our local IRB. Descriptive statistics were used to describe patient baseline demographic, disease, and prior treatment information. The differences in baseline characteristics and TP53 mutation status were examined using a two-sample Wilcoxon test for continuous variables, and Fisher's exact test for categorical variables. Leukemia-free survival (LFS) and overall survival (OS) were defined as the period from the date of diagnosis to the first observed date of relapse or death, and date of death, respectively. If patients remained disease free and alive, LFS was censored at the time of the last disease assessment. OS was censored at the time of the last follow-up if they were last known to be alive. The differences in LFS and OS by TP53 mutation were evaluated using Kaplan–Meier curves and log-rank test. All p values were two-sided at the .05 significance level. Statistical analysis was performed using SAS version 9.4 (SAS Institute, Cary, NC). Amongst 127 adult patients with ALL treated at City of Hope between 2000 and 2021 who met the clinical definition of t-ALL, 31 (24%) patients had available molecular analysis by targeted NGS and were included in this analysis. The median age was 63 years (range: 32–89), and the majority were females (n = 20; 64.5%) and Whites (n = 19; 61.3%). Most cases (n = 27; 87.1%) of t-ALL had B-cell lineage with only four cases having T cell lineage. The median white blood cell count (WBC) at presentation was 4.8 K/μl (range: 1–440). Three patients had a prior or concurrent diagnosis of myelodysplasia (MDS) at the time of ALL diagnosis. Table S2 summarizes the patient's characteristics. The median latency interval between prior malignancy diagnosis and ALL onset was 8 years (range: 1.5–30). The most common prior malignancies were breast cancer and multiple myeloma (each 9; 29%), and this was followed by prostate cancer (n = 4; 12.9%). Prior cytotoxic therapies included chemotherapy in 13 (41.9%) patients, radiation therapy in 9 (29%) patients, and a combination of chemotherapy and radiation in 8 (25.8%) patients. There were 9 (29%) patients, all with multiple myeloma, who received prior therapy with immunomodulatory drugs. Conventional cytogenetics and FISH studies showed complex karyotype (defined as ≥5 abnormalities) in 10 (32.3%) patients, chromosome 5 or/and 7 abnormalities in 8 (25.8%) patients, KMT2Ar in 5 (16.1%) patients, normal karyotype in 3 (9.7%) patients, and Philadelphia (Ph)-chromosome in 3 (9.7%) patients. Molecular analysis showed that TP53 mutations were the most common alterations in this cohort and were observed in 12 (38.7%) patients. None of the Ph + ALL or ALL with T-cell lineage cases carried TP53 mutations. The median number of TP53 mutations per patient was 1 (range: 1–3) and the median variant allele frequency (VAF) was 59% (range: 15%–96%). When limiting analysis to only cases with Ph-negative B cell ALL, the incidence of TP53m was 50%. Other detected mutations in this cohort included K- or N-RAS (n = 5; 16.1%), FLT3 (n = 2; 6.5%), NOTCH (n = 3; 9.7%), RUNX1 (n = 1; 3.2%), ETV6 (n = 1; 3.2%), IDH2 (n = 1;3.2%), and PTPN11 (n = 1; 3.2%), DNMT3A (n = 1; 3.2%), and SF3B1 (n = 1; 3.2%). No somatic mutations were observed in t-ALL cases with Ph chromosome. Figure 1A depicts mutations frequency for this cohort. For 29 patients who received induction, 23 (74.2%) achieved complete remission (CR)/CR with incomplete count recovery (CRi) following initial therapy. There were 10 patients who underwent consolidation with allogeneic hematopoietic cell transplantation (HCT) in first CR (n = 9; 29%) or second CR (n = 1; 3.2%). The conditioning regimen used was myeloablative in four patients and reduced intensity in the remaining six patients. No statistically significant differences were noted in patient characteristics or treatment choice between TP53m patients and those without TP53m t-ALL. Nonetheless, there were inherent observed patterns amongst the TP53m subgroup including older age at the time of diagnosis (median age: 69 vs. 58 years), less female predominance (50% vs. 74%), mainly B-cell phenotype (100% vs. 78%), the tendency for antecedent MDS (17% vs. 5%), lower median WBC at presentation (3.1 vs. 11.7), the likelihood of complex cytogenetics (42% vs. 26%), shorter median latency for ALL onset (5.5 vs. 8 years), and unexpectedly, propensity to achieve CR/CRi with initial induction as compared to patients with no TP53 mutations. Table S1. With a median follow up of 39.8 months (range: 0.7–59.4), there was a trend toward inferior median LFS for t-ALL patients with TP53m (7.1 months; 95% CI: 4.3–16.6) as compared to t-ALL patients without TP53m (15.4 months; 95% CI: 5.5-not reached) (p = .066). However, median OS was not different between patients with (12.0; 95% CI: 5.4–19.4]) and without TP53m (17.1; 95% CI: 5.5–45.5) (p = .18). Refer to Figure 1B,C. Herein, we have demonstrated an in-depth evaluation of the molecular profile of t-ALL in a cohort of 31 patients utilizing a targeted NGS panel. We demonstrated that despite differing lineage phenotypes, there are evident overlaps in molecular alterations observed in t-ALL cases compared with t-MN. Our findings are both illustrative and supportive of a t-ALL entity existing in the myriad of leukemic categories. On molecular review of our therapy-related ALL cases, the high prevalence of TP53 mutation is of key interest given its accordance with these observations in t-MN. Additionally, we noted that the incidence of TP53 mutations is particularly overrepresented in cases with Ph-negative B-cell ALL and constitutes half of the observed cases. There is also significant overlap in other genetic aberrations seen frequently in t-MN including mutations in RAS, DNMT3A, IDH and FLT3. This parallel molecular finding has also been described in previous studies focused more exclusively on therapy-related B-ALL.5 Although we could not demonstrate statistical differences between patients and disease characteristics for TP53m and non-TP53m cases of t-ALL as a result of the small study cohort, there were observed trends toward unique features for TP53m patients with t-ALL including older age, lower WBC at presentation, exclusively observed B-cell lineage, antecedent MDS, complex cytogenetics, and shorter latency for ALL onset. While t-ALL patients with TP53m have a high chance of attaining remission, there was a trend toward inferior long-term outcomes consistent with t-MN harboring TP53.6 Through our institutional cohort of 31 patients with t-ALL is relatively large when considering the molecular profiling of these cases, it is limited in that this is not a cross-institutional analysis and is of course a retrospective review. We have some level of standardization in that all patients were evaluated with the same next-generation sequencing analysis, but it will be prudent to eventually gain larger multi-institution pooled analyses to further validate these molecular characteristics and specifically their effect on prognosis and outcomes. Our study successfully depicts the molecular similarities between t-MN and t-ALL and observed mutational profiles including the propensity for TP53m to represent a significant proportion of these cases and to possibly drive poorer outcomes. We will need larger studies to further validate these findings and to build on understanding the genetic interactions and milieu of t-ALL. We will additionally need to evaluate the outcomes with TP53m directed therapies in these patients harboring TP53m allowing for more tailored treatment choices beyond chemotherapy to improve disease and survival outcomes. Future studies would perhaps evaluate the association between clonal hematopoiesis after exposure to cytotoxic therapy and the emergence of t-ALL, given their shared genomic alterations. This may lead to understanding the predilection of disease and correlations with these specific molecular subtypes of t-ALL. The authors report no relevant conflict of interest. The data that supports the findings of this study are available in the supplementary material of this article. Table S1. Targeted Genes Table S2. Patient Characteristics Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.