Page 19 - 2021_07-Haematologica-web

Genetic-based personalized treatment of MPN
scripts had deeper and faster responses compared to those with e13a2 transcripts.18 Subsequently, Baccarani and col- leagues observed lower response rates and inferior out- comes following nilotinib treatment in patients with e13a2 transcripts.19 These investigators also noted a relationship of transcript type with age and gender. More recently, a French study, led by Genthon and colleagues, also documented deeper and faster responses, in terms of achieving MR3 and MR4, in patients expressing e14a2 compared with those expressing e13a2 transcripts.20 The differences in clinical outcomes based on transcript subtype has so far only been investigated in small studies and this issue needs to be assessed prospectively in larger cohorts receiving any of the five currently licensed TKI treatments for CML. A study of 20 patients who were in TFR for longer than 1 year, having achieved sustained MR4.5 following treatment with TKI, suggested the potential importance of the lineage of meas- urable residual disease as a potential predictive biomarker of TFR outcome.21 The investigators, Pagani and colleagues, used fluorescence-activated cell sorting of CML cells known to express BCR-ABL1 mRNA, granulocytes, monocytes, B cells T cells and natural killer cells; BCR-ABL1 DNA PCR was used to investigate the lineage of these residual CML cells. The observation of BCR-ABL1 DNA being present only in B- and T-lymphocytes begs the question of the CML cell lineage contributing to the success of TFR. The study could not, however, establish the absence of CML stem cells (CD34+38–26+) in a cohort of patients in TFR with undetectable BCR-ABL1 mRNA transcripts, which are con- sidered to account for the loss of TFR in such patients.22 Regardless, larger studies assessing the lineage of measura- ble residual disease in TFR patients are warranted.
In patients with SM, the identification and quantification of the KIT D816V mutation in hematopoietic stem and pro- genitor cells, as well as mature mast cells, by highly sensi- tive (sensitivity <0.01%) and specific assays, such as allele- specific oligonucleotide quantitative PCR and digital droplet PCR, allows an accurate diagnosis.23-26 This informa- tion is also useful for risk stratification, complementing con- ventional biomarkers such as serum tryptase level and per- centage of bone marrow infiltration by mast cells, and may be used for monitoring patients receiving treatment. More recently, the use of targeted NGS panels for the characteri-
Figure 2. The International Scale for quantitative reverse transcrip- tase polymerase chain reaction analysis of BCR-ABL1 transcripts. IRIS: International Randomized Study of Interferon and STI571; MMR: major molecular response; MR4, MR4.5 and MR5: 4-, 4.5- and 5-log reductions, respectively, from the IRIS standardized base- line; Ph: Philadelphia chromo- some; RT-qPCR: quantitative reverse transcriptase polymerase chain reaction.
zation of associated gene mutations, such as SRSF2, ASXL1 and RUNX1, has improved prognostication and has led to the development of novel prognostic scoring systems to optimize clinical management.27
Testing for BCR ABL kinase domain mutations
Mutation testing for BCR-ABL1 kinase domain mutations and the co-existence of subclones in general can be assessed by a variety of technologies. In general, the more sensitive the test, the more complex and expensive it is. Sanger sequencing has a low error rate but a poor sensitivity of only 10-20%. In contrast, NGS has a better sensitivity of roughly 1% and is useful for the identification of com- pound mutations.28 However, the error rate associated with the library amplification and preparation can be up to one in a 1,000 base pairs, particularly when sequencing from mRNA, which relies on the error-prone reverse transcrip- tase. Barcode correction techniques may improve this, but the best technique may be so-called duplex sequencing, which is novel in that it sequences both DNA strands, dra- matically reducing the error rate since mutations are only called if the complementary base change is seen on the other strand. Indeed, recent reports suggest that low fre- quency mutations that are detected by NGS but have unpredictable clinical courses (e.g., disappear spontaneous- ly in some patients) may represent artifacts from the error rate inherent in NGS.29 Soverini and colleagues recently con- ducted the first prospective, ‘real-world’ assessment of NGS-based BCR-ABL1 knockdown mutation testing com- pared with Sanger sequencing in a large cohort of consecu- tively studied CML patients in whom TKI had failed or who were in the ‘warning’ category described by the European Leukemia Net (ELN) guidelines.30 The researchers demonstrated the importance of low-level mutations, defined as mutations with a variant allele frequency of 3- 20%, and their clinical relevance (Figure 3). These observa- tions are the basis for guiding genomic-based personalized therapies for CML further, in particular through the identi- fication of high-risk individuals prior to the ELN ‘warning’ stage. These tests should also improve the efficiency and safety of clinical trials designed to reduce the risk of blast transformation in patients who respond suboptimally to TKI. Important challenges now are to improve sensitivity
haematologica | 2021; 106(7)
1789