Answers are grounded in this paper and cite their sources.
We read with interest the recent article by He et al1 comparing conventional karyotyping and fluorescence in situ hybridization (FISH) in acute myeloid leukemia (AML). While the authors clearly show no advantage to FISH when karyotyping is adequate and propose an algorithmic approach to testing, there is no discussion regarding practical matters, such as turnaround time, potential delay in diagnosis, and actual cost vs benefit. Cytogenetics and FISH are often send-out tests at many institutions, and as noted, this is often done concurrently rather than sequentially. Chromosome analysis has long been the gold standard for identification of the common, recurrent chromosome abnormalities in AML. However, this analysis requires dividing cells, takes 5 to 7 days to process, and can miss subtle or cryptic rearrangements (eg, inv[16] and MLL anomalies). We, like others, are also looking at best practices to provide value and quality care in the new health care environment. We find the algorithmic strategy is beneficial in many situations and can prevent unnecessary testing. Being curious of the implications in this scenario, we asked for a turnaround time report for conventional cytogenetics and multiprobe AML FISH panel from our reference laboratory (a large national reference laboratory) for the past 6 months. The average turnaround time for cytogenetics was 7 days vs 2.5 days for FISH. This could potentially result in a significant delay of definitive treatment regimens or increased length of stay. In this regard, time from diagnosis to treatment (TDT) has been reported to predict survival in younger but not older patients with AML.2 In that study, response rate and survival were decreased after a 5-day treatment delay in patients younger than 60 years. On the other hand, Bertoli et al3 recently reported that TDT did not adversely affect outcome of patients with AML. Nonetheless, delay in diagnosis/treatment is often associated with increased hospitalization time and thus overall cost. We would also have to assume adequate specimen would be available 7 days later for further FISH analysis, but this may not always be the case. It has been our experience that clinicians, as well as patients and their families, would rather have the answer sooner, not later. The authors also suggest follow-up FISH analysis on cases that were less than adequate or no growth (30/250 cases; 12%), which would add an additional time to definitive characterization and only with specific probes focused on World Health Organization–defined AML with recurrent genetic abnormalities. While we certainly agree with a focused approach on cases that have morphologically distinct features such as t(8;21), inv(16), and monocytic differentiation, which often harbor 11q23 abnormalities, many leukemias have nonspecific morphologic and immunophenotypic features. Of these cases, a significant proportion may require follow-up FISH testing and further delay if less than adequate cytogenetic studies. In addition, probes for numeric chromosome abnormalities, such as monosomy 7, which has been shown to have significantly poorer prognosis,4‐6 do not appear to be included in the targeted strategy but are generally performed on initial diagnosis. In fact, recent studies suggest karyotypic banding studies may be insufficient, and there may be prognostic differences in only partial monosomy that may be clinically significant.7 Despite the accuracy of cytogenetics in this study, the authors also do not address why targeted FISH is necessary to confirm the cytogenetic abnormality, other than providing a baseline. If a confirmation is necessary, maybe the confirmatory test should have been done first. The authors suggest that the cost for a comprehensive panel “may be quite expensive” and its addition questionable. So how much is this cost difference, and is it potentially “worth it” to have a quicker result? Our institutional costs from our reference testing laboratory for a comprehensive AML FISH panel on initial diagnosis are approximately four times that of conventional cytogenetics. That difference could be equated to less than a day of a hospital admission, another significant cost factor in health care. Note these are not charges to the patient, which are significantly different and can be variable among hospitals and institutions, are often tied to insurance and reimbursement rates from Medicare, and do not reflect true costs of testing. Does the actual cost of the test justify a less costly test with significant delay in turnaround? Or are we ultimately trying to decrease patient and hospital charges while providing the most efficiency and quality in diagnosis and treatment? This is the crux of the question, and clinicians and administrators may have differing opinions. Any delay that affects length of hospital stays, though, is unlikely to be cost beneficial in the long run. While no test is perfect, this study clearly points out that the two methods are very comparable in identifying a chromosomal abnormality. Maybe FISH testing is more valuable than suggested due to its shorter turnaround time, which affects the overall length of hospital stay and should replace cytogenetic testing as a first-line test. While we are all involved in appropriate test utilization, the answer may be more complex than realized. We thank Drs Hammer and Doll for their interest and comments about our article and appreciate their recognition of the importance of appropriate test utilization. We concur that the turnaround time (TAT) of conventional cytogenetics is longer than that of fluorescence in situ hybridization (FISH). However, our conventional cytogenetic testing in diagnostic acute myeloid leukemia (AML) cases is usually rush-ordered by the reviewing hematopathologist, and the TAT is reduced to 3 days. We think their argument that our approach of performing conventional cytogenetics upfront with FISH reflex could delay AML diagnosis/treatment and increase hospitalization time deserves further discussion and clarification. First, AML is a myeloid neoplasm defined and diagnosed by the presence of at least 20% myeloid lineage blasts in bone marrow and/or peripheral blood. The recognition and percent determination of blasts are rendered via morphologic evaluation, and myeloid lineage assignment is accomplished by flow cytometric immunophenotyping and cytochemical stains. Conventional cytogenetics and FISH play no role in the diagnosis of AML. The only exceptions are cases harboring t(15;17);PML-RARA, t(8;21);RUNX1-RUNX1T1, and inv(16)/t(16;16);CBFB-MYH1, which do not require the 20% blast count for AML diagnosis. In this regard, our testing algorithm explicitly excludes acute promyelocytic leukemia (APL) with t(15;17);PML-RARA since the APL workup deviates from other AMLs and always requires rapid FISH or reverse transcription–polymerase chain reaction to confirm a PML-RARA fusion. The other two scenarios, t(8;21);RUNX1-RUNX1T1 and inv(16)/t(16;16);CBFB-MYH1, are adequately covered in our algorithm since both present typically with characteristic morphologic features. For cases with morphologic features suspicious for t(8;21) or inv(16)/t(16;16), the respective targeted FISH testing for RUNX1-RUNX1T1 or CBFB-MYH11 is ordered upfront along with the conventional cytogenetics to confirm the morphologic impression. Second, the major role of cytogenetics in AML is for risk stratification, complemented by key molecular genetic findings, such as FLT3, NPM1, and CEBPA in normal-karyotype AML (which also requires a 3- to 5-day TAT). In patients younger than 60 years, risk stratification is used to guide postremission therapy and allows ample time for the return of both conventional cytogenetics and reflex FISH test results. Importantly, the standard induction therapy in this age group is a 3-day anthracycline and 7-day cytarabine (3 + 7) regimen, and induction therapy is routinely not held for conventional cytogenetics or FISH results, except in APL, which requires rapid confirmation and was excluded from our algorithm. In older patients (≥60 years), the medical intervention choice at time of diagnosis is determined individually, depending largely on the patient’s performance status and comorbidities. Cytogenetics does play a role here, and patients with high-risk findings may be offered a less intense induction therapy even with good performance status and no comorbidities. However, in this age group, patient outcome is generally poorer, and multiple studies have shown no detrimental effect from a slight diagnosis to treatment delay. Therefore, the argument that our algorithm would cause delay of AML diagnosis/treatment and prolongation of hospital delay is not substantiated. Hammer and Doll also question whether an adequate specimen would be available for further FISH testing after conventional cytogenetics. This should not be an issue since it is routine laboratory practice to store methanol/acetic acid–fixed cell pellets for further FISH testing in many cytogenetic laboratories, including ours. This approach was also endorsed and recommended by the International Expert Panel of the European LeukemiaNet. Furthermore, for cases with insufficient conventional cytogenetics, we propose testing for the World Health Organization–defined recurrent genetic abnormalities but do not exclude any additional appropriate targeted FISH testing. For cases with ambiguous chromosome findings, the choice of additional FISH probes is left to the discretion of the cytogeneticists for clarification with integration of the karyotypic, morphologic, and clinical findings. This approach includes use of chromosome enumeration FISH probes as indicated. We want to further reiterate that conventional cytogenetics provides a global assessment of the full complement of chromosomes and is the cornerstone of risk stratification in AML. No matter how extensive an AML FISH panel can be, it is impossible to cover all the findings of conventional cytogenetics or to replace it. In fact, it is not uncommon to see conventional cytogenetic abnormalities that are not targeted by concurrent FISH testing and provide important additional information for risk stratification. We are glad to have the opportunity to further our discussion on the topic of appropriate cytogenetic test utilization in diagnostic adult AML workup. We envision that our reply helps to clarify some confusion regarding the roles of cytogenetic findings in AML diagnosis and treatment, as well as some specific approaches proposed in our algorithm. Last and just as important, one should realize that algorithms, in and of themselves, are not dogmatic. They are, by definition, applicable to most scenarios. Algorithms are by no means exclusionary, and for the occasional exceptions, sound professional and scientific judgment should be followed.
No takes yet. Share an insight, caveat, or question.
Hammer et al. (2016) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: