Year: 2018
Type: letter
Abstract: To the Editor: We welcome this opportunity to elaborate on our study findings that showed no impact of MYD88L265P mutation status on overall survival (OS) of patients with Waldenström macroglobulinemia (WM).1 First, we concur with Castillo and coworkers that the diagnostic criteria for WM need to be uniform for cross-study comparisons. To that end, our study had excluded all the cases of IgM monoclonal gammopathy of undetermined significance (MGUS) as defined by the World Health Organization2 in 2017 and supported by recent studies, as well as smoldering WM because these patients do not require treatment.3-6 Concerns were raised about the sensitivity of the assay we employed: allele-specific polymerase chain reaction (AS-PCR), a validated test now routinely used in our as well as other laboratories in the United States and other countries to detect the presence of MYD88L265P mutation. The sensitivity of the assay used is but one part of the equation, and the readers are likely to misinterpret the findings without taking into context the lymphoplasmacytic burden of the bone marrow samples utilized in the study. Expensive and laborious tests with higher sensitivity, including more sensitive AS-PCR testing, Sanger or next-generation sequencing have been employed in some studies, including by members of our group.7 However, the utility of such assays largely comes into play when the tumor-burden is very low. As exemplified in our article, this was a non-issue for the cohorts studied because the disease burden in the samples tested was substantially higher than the lower limit of detection of the assay employed. Importantly, our assay is a clinically validated assay, the sensitivity of which is in accordance with the vast majority of assays utilized for routine clinical purposes in the United States. The MYD88L265P mutation has been variably identified in 70% to over 95% of patients with WM.6, 8 In our study, the prevalence of MYD88L265P mutation was 79% (95% CI: 73–84) using an assay designed to detect a minimum of 1% mutant allele in a wild-type background. Although there were 10 patients with lymphoplasmacytic infiltrate of less than 10% in the bone marrow samples used, all samples tested had at least 5% lymphoplasmacytic lymphoma cells, and thus we are confident in this assay's ability to detect the MYD88L265P in each of our patients. To minimize the possibility of false-negative results affecting our data, we have performed a subgroup analysis by excluding these 10 patients, and found that the results remain unchanged from that originally reported (median OS 10.2 [95% CI:8.4–16.5] years in MYD88L265P mutant cohort and 13.9 [95% CI: 8.3–29.4] years for patients with MYD88WT genotype, P = .98).1 As discussed in the paper, the inability of our assay to detect the non-L265P mutations could lead to false-negative results1; however due to the rarity of such mutations,9 the outcome of patients in our wild-type cohort, all of whom received non-BTK inhibitor-based therapies, are unlikely to be affected. It is well established that alterations in the MYD88 gene can favorably impact response to BTK inhibitors such as ibrutinib.6 However, as indicated in the paper, the Mayo clinicians were blinded to their patients’MYD88 mutation status, and ibrutinib was used in only 2% of patients in MYD88L265P group and none in the wild-type group.1 Therefore, BTK inhibitor-based regimens had minimal potential to affect the outcomes. Furthermore, studies have recently shown that non-BTK inhibitor-based regimens are effective in WM, irrespective of the presence of MYD88 mutation.10-12 The retrospective data presented by Castillo and coworkers reported a poorer prognosis in patients with wild-type disease compared to patients with mutations in the MYD88 gene appear incomplete, without the crucial information regarding the therapies administered.9 Our rationale for assessing OS in patients with WM from the time of treatment initiation i.e. active disease rather than from the time of diagnosis of WM, as explained in the paper, was to reduce the lead-time bias for patients who have smoldering (asymptomatic) disease which does not require treatment.1 In other words, the prognostic implications of the mutations on survival outcomes would be more accurately assessed if their impact is analyzed from the time of symptomatic disease when treatment is begun. The baseline disease characteristics and International Prognostic Scoring System for Waldenström's Macroglobulinemia (IPSS-WM) were similar at the time of active disease in the MYD88L265P mutant and MYD88WT cohorts of Mayo Clinic patients. Therefore, the argument by Castillo et al. that the patients with wild-type disease may have had aggressive disease at diagnosis is not in accordance with what was observed in our study. Regardless, to further allay this concern raised, we reanalyzed OS from the time of diagnosis, irrespective of the disease state (active or smoldering WM) of our patients. The median follow-up was similar for MYD88L265P and MYD88WT WM patients (7.4 vs. 8.1 years respectively, P = .25). The overlapping survival curves in this repeat analysis (Figure 1) speak for themselves, once again confirming our original findings that MYD88L265P mutation status does not predict survival outcome in patients with WM who are primarily treated with therapies outside of BTK inhibitors. Overall survival of patients with MYD88L265P versus MYD88WT cohorts from the time of diagnosis of WM. The median OS was 11.6 (95% CI: 10.1–17.5) years for patients with MYD88L265P and 15.7 (8.3–29.4) years, for patients with MYD88WT genotype, P = 0.67. Our preliminary findings demonstrating that the proportion of patients who transformed to a higher grade lymphoma was greater in the WT cohort (14% vs. 2% in MYD88L265P cohort, P = .002) were indeed subsequently confirmed by Castillo and coworkers.9 All transformations that occurred in our study were metachronous to the diagnosis of WM, and it remains unclear as to what extent the increased use of alkylating agents in the WT cohort contributed to this observation. At 10 years, the risk of transformation was 1.6% and 24.5% in the MYD88L265P and MYD88WT cohorts, respectively (P = .002). However, unlike Castillo's study,9 the higher rate of transformation did not negatively impact the OS of the entire MYD88WT cohort. We have recognized the limitations associated with a retrospective study involving a heterogeneously treated cohort in our paper.1 Additional studies involving homogeneously treated populations that could potentially reduce the impact of confounding factors are needed to establish a causal relationship, if any, between the MYD88WT genotype and disease transformation. Of 557 patients seen at our institution between January 1, 1995 and July 30, 2016, 219 had archived samples for assessment of MYD88 status, and were the focus of this study. To reduce the associated biases, we compared the OS of patients whose MYD88L265P mutation status was known (n = 219; 39%) to the remainder of (n = 338; 61%) the cohort; the lack of OS difference between these two cohorts indirectly suggests that biases that were introduced by focusing only on a subset of patients were likely minimal. MYD88L265P mutation is an early oncogenic event in WM. To date, there are no convincing data to support that the MYD88 mutation status changes with prior therapy. Therefore, it is unlikely that the prevalence of MYD88L265P genotype in our study was affected by the number and types of prior therapies that the patients had received. In summary, without taking the context into account, the compelling findings of our study may be misinterpreted as a mere comparison of “apples and oranges.” Nothing to report.
Source: American Journal of Hematology
DOI: 10.1002/ajh.25029
Language: English
Subfield: Genetics
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W2783851435