Comparison of Multiparameter Flow Cytometry Immunophenotypic Analysis and Quantitative RT-PCR for the Detection of Minimal Residual Disease of Core Binding Factor Acute Myeloid Leukemia

Abstract

Objectives: To examine the value of minimal residual disease (MRD) by multiparameter flow cytometry (MFC) in core binding factor (CBF) acute myeloid leukemia (AML).

Methods: We studied 42 patients with t(8;21)(q22;q22)/RUNX1-RUNX1T1 and 51 with inv(16)(p13.1q22)/CBFB-MYH11. Tandem MRD analyses by MFC and quantitative reverse transcription polymerase chain reaction (qRT-PCR) were performed in 281 bone marrow (BM) samples.

Results: Grouping qRT-PCR levels as ≤0.01, 0.01 to 0.1, 0.1 to 1, 1 to 10, and >10%, and reporting MFC (sensitivity, 0.1%-0.01%) as positive or negative, κ coefficient test showed no agreement between qRT-PCR and MFC in BM samples obtained postinduction (n = 44, κ = 0.041), and only weak agreement during consolidation (n = 108, κ = 0.083), maintenance/follow-up (n = 107, κ = 0.164), and salvage chemotherapy (n = 24, 0.376). In the post induction BM samples, while qRT-PCR <0.1% was associated with lower and ≥10% with higher AML relapse risk (P = .035), qRT-PCR between 0.1% to 1% and 1% to 10% failed to predict relapse. In the latter group with intermediate qRT-PCR results, MFC provided prognostic value for relapse (P = 0.006).

Conclusions:MFC and qRT-PCR are complementary tests in monitoring CBF AML MRD.

Acute myeloid leukemia (AML) with t(8;21)(q22;q22) and inv(16)(p13.1q22) account for about 15% of adult AML and carry an overall favorable prognosis. Using standard therapy, adults with core binding factor (CBF) AML have an improved complete remission (CR) rate, prolonged CR duration, and a better prognosis than patients with AML with a normal karyotype or other chromosomal aberrations.^1,2 Despite these encouraging results, approximately 30% of patients with CBF AML eventually relapse.^1,3,4 Residual leukemic cells present in the bone marrow (BM) of patients in CR are thought to be responsible for the emergence of the relapses.

The development of reverse transcription polymerase chain reaction (RT-PCR) techniques to detect the fusion genes RUNX1-RUNX1T1 and CBFB-MYH11 has enabled the sensitive detection and quantification of residual AML. Minimal residual disease (MRD) monitoring by quantitative (q) RT-PCR at specific time points in CBF AML allows identification of patients at high risk of relapse.^5-8 While most of these studies have shown that a high qRT-PCR level at the end of treatment or a rising qRT-PCR level during follow-up predisposes to relapse, no absolute threshold that predicts outcome has been defined.^7,9,10 On the other hand, it is known that long-term surviving patients with t(8;21) AML may harbor relatively high copy numbers of RUNX1-RUNX1T1 fusion transcripts.^10-12 Furthermore, for an adult with t(8;21)(q22;q22)AML who has achieved morphologic CR with persistence of RUNX1-RUNX1T1 transcripts with one course of induction therapy, no data exists to advocate the immediate use of allogeneic stem cell transplant.¹³

Multiparameter flow cytometry (MFC) is applicable in more than 90% of all patients with AML.^14-16 A positive MRD has been shown to correlate with an inferior leukemic event-free survival and overall survival in patients post induction and consolidation, as well as to predict hematopoietic stem cell transplant (HSCT) outcomes.^17-21 An early study by Perea and colleagues compared the results of AML MRD by a 3-color flow cytometry assay with qRT-PCR in CBF AML, found a concordance rate of 67%; they also noted that the combination of flow cytometry and qRT-PCR improved MRD detection.¹⁰

In our laboratory, we have developed and validated a MFC assay for AML MRD detection that combines the identification of a leukemia-associated immunophenotype (LAIP) and the detection of “deviation from normal” approaches.^22,23 We have applied this assay to MRD assessment in patients with AML associated with t(8;21) or inv(16). In the present study, we examined the value of MFC in CBF AML MRD detection in conjunction with the data of qRT-PCR. The latter is the commonly used method for CBF AML follow-up.

Materials and Methods

Patients

From June 2012 to June 2014, assays to detect MRD by MFC and qRT-PCR were performed as part of the routine workup of patients with CBF AML. These BM samples were collected post induction or salvage chemotherapy, during consolidation and maintenance therapy, surveillance and post HSCT. For patients with a diagnosis made at the referring hospital, BM pathology, along with cytogenetic and molecular studies, was reviewed at our hospital, and patients without sufficient clinical and laboratory data were excluded from this study. All cases included in this study had a morphological negative BM (BM blast count <5%).The clinicopathologic information was obtained by review of the medical charts. This study was approved by the Institutional Review Board of MD Anderson Cancer Center.

Patients, either with t(8;21) or inv(16) AML, received one or two courses of induction therapy with FLAG-Ida (fludarabine 30 mg/m² intravenously, days 1-5; cytarabine 2 g/m² intravenously, days 1-5; idarubicin 6 mg/m² intravenously, days 3-4; and filgrastim 5 μg/kg, day 1). Consolidation therapy consisted of 6 cycles of FLAG, except that fludarabine and cytarabine were given for 3 rather than 5 days. Idarubicin was administered during induction and again on one more occasion between consolidation cycles 3 to 6. For patients older than 60 years, patients with other comorbidities, or patients deemed unsuitable for further FLAG-based consolidation, treatment could be changed to decitabine 20 mg/m² IV daily for 5 days every 4 to 6 weeks for up to 12 cycles. For relapsed patients, salvage therapies included high-dose cytarabine-based regimens in most patients. In some patients, hypomethylating agent-based salvage regimens were used. Maintenance included patients with no therapy, and patients who received DNA methyltransferase (DNMT)-inhibitor (decitabine or azacytidine) for maintenance therapy.

Morphologic Assessment

Morphologic evaluation was performed independently without knowledge of the MFC or qRT-PCR findings. Each case was reviewed at least by one hematopathologist at the time of diagnosis, and cases that fulfilled the initial inclusion criteria for this study were confirmed by another pathologist. For each case, routine H&E-stained histologic sections of BM biopsy and aspirate clot, and Wright-Giemsa-stained BM aspirate smears were evaluated. A 500 cell count was performed based on examination of multiple fields of BM aspirate smears. Cases with inadequate quality of BM smears were excluded from this study.

Flow Cytometric Immunophenotyping (FCI) of MRD

Eight-color flow cytometry analysis was performed as previously described.^22,23 The panel included four tubes as follows: (1) CD7FITC, CD33 PE, CD19 PerCP-Cy5.5, CD34 PE-Cy7, CD13 APC, CD38 BV421, CD45 V500; (2) HLADR-FITC, CD117 PE, CD4 PerCP-Cy5.5, CD34 PE-Cy7, CD123 APC, CD19-eF780, CD38 BV421, CD45 V500; (3) HLA-DR-FITC, CD36 PE, CD56 PerCP-Cy5.5, CD34 PE-Cy7, CD64 APC, CD19-eF780, CD14 V450, CD45 V500; (4) CD5-FITC, CD2 PE, CD22 PerCP-Cy5.5, CD34 PE-Cy7, CD38 APC, CD19-eF780, CD15 V450, CD45 V500. All antibodies were obtained from Becton Dickinson (San Jose, CA) or eBioscience (San Diego, CA). Samples were acquired on FACSCanto II instruments (BD Biosciences, San Diego, CA) that were standardized daily using CS&T beads. A minimum of 200,000 live events were acquired to achieve a potential sensitivity of at least 10⁻⁴ (0.01%). Minimal residual disease was defined as a neoplastic blast population with an abnormal pattern of antigen expression deviating from normal regenerating myeloid progenitors. The alterations in blasts were categorized by (1) altered expression levels of antigens normally expressed, either decreased (CD38, HLA-DR) or increased (CD13, CD34, CD117, CD123), as well as altered CD45/side scatter pattern; (2) asynchronous expression of myelomonocytic antigens on myeloblasts including CD4, CD64, CD36, and CD15; and (3) aberrant expression of lymphoid antigens, eg, CD2, CD5, CD7, CD19, CD22, CD56. To increase sensitivity and consistency, the levels of expression were measured by mean fluorescence intensity (MFI). The abnormal blast population was qualified as a percentage of total events. Any level of an abnormal blast population (≥0.01%) detected by FCI was considered MRD positive.

qRT-PCR

Nanofluidics-based qualitative multiparametric RT-PCR was performed for detection of recurrent fusion transcripts in all newly diagnosed acute leukemia. The panel tests included t(8;21)(q22;q22); RUNX1-RUNX1T1, inv(16)(p13.1q22) or t(16;16)(p13.1q22); CBFB-MYH11 variant A and CBFB-MYH11 variant D. In cases positive for any of these fusion products, qRT-PCR was performed for monitoring treatment response. In brief, extracted RNA was analyzed by real-time PCR for the RUNX1/RUNX1T1 and CBFB/MYH11 fusion transcripts on an ABI HT 7900 platform from Applied Biosystems (Foster City, CA). Quantitative values were expressed as a percentage ratio of fusion transcripts to ABL1 transcript levels (%), and the sensitivity is 0.01%.

Cytogenetics Analysis

G-banded metaphase cells were prepared from unstimulated 24- and 48-hour BM aspirate cultures at the time of diagnosis for conventional chromosomal analysis. Interphase fluorescence in situ hybridization (FISH) analysis was performed at the time of diagnosis and at follow-up whenever possible. FISH for CBFB (dual color, breakapart probe, Abbott Molecular/Vysis, Des Plaines, IL) and RUNX1T1/RUNX1 (dual color, dual fusion probe, Abbott Molecular/Vysis, Des Plaines, IL) was performed on freshly harvested BM cells (metaphase or interphase) or BM aspirate smear (interphase). Two hundred cells were counted. The positive cut-off established in our lab was 0.4% for RUNX1T1/RUNX1 and 4.2% for CBFB rearrangement. Two hundred nuclei were counted, and the percentage of abnormal cells was calculated.

Statistical Analysis

Relapse-free survival was estimated by the Kaplan-Meier method and the log-rank test. The agreement (κ coefficient) of MFC with qRT-PCR for detecting MRD was calculated using categorized values (MFC: positive and negative, qRT-PCR: <0.01%, 0.01% to 0.1%, 0.1% to 1%, 1% to 10%, ≥10%). Weighted κ values were used to categorize results as very good (0.81-1.0), good (0.61-0.8), moderate (0.41-0.6), fair (0.21-0.4) or slight (0-0.2).²⁴ The statistical analyses were performed using SPSS software (IBM Corporation, Armonk, NY). Results were considered statistically significant if P values were less than .05 in a two-tailed test.

Results

Patients and Samples

A total of 93 patients with CBF AML were included in this study. There were 42 patients with t(8;21) and 51 patients inv(16). Of these patients, 76 patients received treatment for first-time diagnosis of AML, and 17 had relapsed AML Figure 1. Of the first group of 76 patients, at the time of post induction, 54 patients had MFC performed; 59 patients had qRT-PCR data available. All included patients had a morphological remission BM (<5% blasts). The demographic and laboratory data are shown in Table 1.

Figure 1.

Patients and bone marrow samples tested. Patients inside the dash-box were used for survival comparison (see Figure 2). Thirty-three of the 44 patients with tandem multiparameter flow cytometry (MFC) and quantitative reverse transcription polymerase chain reaction (qRT-PCR) performed were eligible for 18-month survival comparison (Table 3). AML, acute myeloid leukemia; MRD, minimal residual disease.

Table1.

Clinicopathologic Characteristics of Patients With Core Binding Factor Acute Myeloid Leukemia

	RUNX1/RUNX1T1 (n = 42)	CBFB/MYH11 (n = 51)
Demographics
Median age (range), y	52 (11-78)	45 (7-84)
No. (%) of patients >60 y	7 (17)	9 (18)
No. (%) of men/women	23 (55)/19 (45)	29 (57)/22 (43)
Clinical No. (%) with diagnosed AML/t-AML
Newly diagnosed AML	33 (79)/10	43 (84)/13
Relapsed AML/t-AML	9 (21)/1	8 (16)/2
Peripheral blood
Median (range) WBC, ×109	10.1 (0.8-61)	17.2 (1.4-255)
Median (range) hemoglobin, g/L	8.9 (2.33-13.0)	9 (6.1-15.1)
Median (range) platelets, ×1012	37 (13-273)	35 (7-315)
Bone marrow
Median (range) cellularity, %	85 (20-100)	90 (10-100)
Median (range) blasts, %	33 (10-95)	42 (7-93)
Median (range) follow-up, mo	28 (11-78)	28 (8-91)

AML, acute myeloid leukemia; t-AML, therapy-related acute myeloid leukemia.

All patients with newly diagnosed AML were treated with FLAG-IDA based induction chemotherapy. Among the patients with relapsed AML, 15 were treated with cytarabine-based high-dose salvage chemotherapy, one with decitabine, and another with DNMT inhibitor SGI-110. A total of 25 patients underwent HSCT, including 22 patients for relapsed AML and three patients following the first CR.

The median follow-up of these patients, from the time of diagnosis (including the initial diagnosis made at an outside hospital) to last follow-up, was 28 months (range, 6-91). Leukemia-free survival (LFS) was only calculated for patients postinduction who achieved CR or CR with incomplete recovery of blood (CRi). In brief, LFS was assessed from the date of CR/CRi until the date of relapse or death from any cause; patients not known to have relapsed or died at last follow-up were censored on the date they were last examined. Patients who received HSCT were censored at the date of the procedure.

MFC Analysis

The common blast immunophenotype of AML with t(8;21)/RUNX1-RUNX1T1 was CD34+, CD117+, CD13+, CD33+, CD19+, and CD56+, whereas AML with inv(16)/CBFB-MYH11 was often myelomonocytic, with a subset of blasts expressing a typical myeloid immunophenotype and another subset of blasts expressing a monocytic immunophenotype. The sensitivity, after considering factors of sample quality and admixing normal regenerating myeloid precursors, was between 0.01% and 0.1%.

At the time of post induction for first-time diagnosis AML, a total of 54 patients had a MFC study performed by us on a morphologic CR BM. Of these, six (11%) patients were positive for MRD, with a median of 1.05% blasts (range 0.05%-2.38%); and 48 patients were negative for MRD. By the end of follow-up, 14 patients relapsed, including five of the six patients with a positive MFC MRD and nine of the 48 patients with a negative MFC MRD. A positive MFC MRD study in CR was a strong predictor for AML relapse (Kaplan-Meier, log rank, P < .001) Figure 2.

Figure 2.

Probability of acute myeloid leukemia relapse by minimal residual disease (MRD) status as detected by quantitative reverse transcription polymerase chain reaction (qRT-PCR) (A) (P = .035) and multiparameter flow cytometry (B) (P < .001).

qRT-PCR Analysis

The qRT-PCR levels were divided into five categories: ≤0.01%, >0.01% but ≤0.1%, >0.1% to 1%, >1% to 10%, and >10%. qRT-PCR data were available in 59 patients who were treated for first-time AML and had a morphologic CR BM at the time of post induction. Of these patients, 15 (25%) showed a BM qRT-PCR level ≤0.1%, including eight patients with ≤0.01% and seven patients 0.01% to 0.1%; four patients (7%) had ≥10%, and 40 (68%) patients had a level between 0.1% and 10%, including 25 with 0.1%-1% and 15 patients with 1% to 10%. While none of the patients in CR with ≤0.1% fusion transcripts relapsed, two of four patients with >10% qRT-PCR levels relapsed. The levels between 0.1% to 1% and 1% to 10% showed no significant difference in predicting AML relapse (P =  .47) (Figure 2).

Agreement of MFC and qRT-PCR Detection of MRD

Tandem MFC and qRT-PCR tests were performed on 281 morphologically CR BM samples, including 44 post induction samples, 108 samples collected in the course of consolidation therapy, 105 samples collected in the phase of maintenance/follow-up and 24 samples undergoing salvage chemotherapy for relapse AML. MFC was expressed as positive or negative, and qRT-PCR levels were divided into 5 categories. No agreement was observed between these two methods (MFC results and the levels of qRT-PCR) in the post induction BM samples, only a slight agreement in the samples obtained in the consolidation and the follow-up phase, and a fair agreement in samples in patients with treated relapse (κ = 0.376) Table 2. For all samples, the agreement was weak (κ = 0.152).

Table 2.

Correlation of Minimal Residual Disease Detected by MFC and RT-PCR in Paired Bone Marrow Samples and the Results of Interphase FISH^a

N (%)	Induction (n = 44)		Consolidation (n = 108)		Follow-up (n = 105)			Salvage Treatment (n = 24)		Total (n = 281)		Interphase FISH (n = 159)
	MFC –	MFC +	MFC –	MFC +	MFC –	MFC +		MFC –	MFC +	MFC –	MFC +	MFC –	MFC +
RT-PCR ≤0.01	7 (100)	0	43 (100)	0	78 (100)	0b		6 (100)	0	134 (100)	0a	0/69	0a
RT-PCR >0.01 to ≤0.1	8 (100)	0	29 (100)	0	17 (100)	0b		5 (100)	0	59 (100)	0a	0/31	0a
RT-PCR >0.1 to ≤1	13 (72)	5 (28)	25 (92)	2 (8)	5 (83)	1 (17)		4 (80)	1 (20)	47 (84)	9 (16)	1/31 (2.0)	0/5
RT-PCR >1 to ≤10	8 (100)	0	4 (67)	2 (34)	1(50)	1 (50)		0	1 (100)	13 (76)	4 (24)	0/7	0/2
RT-PCR >10	2 (67)c	1 (33)	0	3 (100)	0	2 (100)		0	7 (100)	2 (13)	13 (87)	1/2 (1.5)	8/12 (5.5, 0.5-22)
Total	38	6	101	7	101	4		15	9	255	26	2/140	8/19
κ value	0.041		0.083		0.164			0.376		0.152
P value	.147		.000		.000			.000		.000

FISH, fluorescence in situ hybridization; MFC, multiparameter flow cytometry; RT-PCR. reverse transcription polymerase chain reaction.

^aData are given as No. (%).

^bThese two samples were obtained from one patient who developed therapy-related myelodysplastic syndrome, FISH negative for RUNX1-RUNX1T1; chromosomal analysis revealed a complex karyotype without t(8;21). FISH was positive for -5 and -7.

^cBoth of the negative MFC samples were acute myeloid leukemia with RUNX1-RUNX1T1 postinduction (see text for details).

MFC Utilities in Post Induction and Consolidation BM Assessment

We investigated the clinical significance of MRD detected by MFC, particularly in the post induction samples with qRT-PCR levels between 0.1% to 1% and 1% to 10%, since qRT-PCR levels between these amounts failed to differentiate risk of AML relapse. For this comparison, patients who received HSCT were censored at the time of transplant, and a minimum of 18 months follow-up was required for patients who were in sustained remission.

Post induction, a total of 33 patients were eligible for assessment. Nine patients with RUNX1-RUNX1T1 AML had a qRT-PCR level falling in this range (0.1%-10%), but only one patient was positive by MFC. By the end of follow-up, three patients relapsed including the patient with a positive MFC; the other two patients had biopsy-proven myeloid sarcoma at presentation. There were 12 patients with inv(16) and a qRT-PCR level in this range post induction; three of these patients were positive by MFC, and all three relapsed, whereas among the nine who were negative by MFC, only one patient relapsed Table 3. Overall, MFC positivity emerged as a strong predictor for AML relapse in this group of patients (P = .006).

Table 3.

Cumulative Incidence of Relapse According to MFC and qRT-PCR Minimal Residual Disease Identification in Bone Marrow Remission Samples at 18 Months

	qRT-PCR <0.1%	qRT-PCR 0.1%-10%	qRT-PCR ≥10%	Total	P Valuea
Post induction	0/9	8/21	1/3	9/33
Relapse/MFC+	0	4/4	1/1	5/5
Relapse/MFC–	0/9	3/17	0/2	3/28	.006
End of consolidation	3/25	3/11	2/2	6/39
Relapse/MFC+	0	1/1	2/2	3/3
Relapse/MFC–	3/25	2/10	0	5/35	NS
Salvage therapy	2/10	2/2	5/5	9/17
Relapse/MFC+	0	2/2	5/5	7/7	NS
Relapse/MFC–	2/10	0	0	2/10

MFC, multiparameter flow cytometry; NS, not significant; qRT-PCR. quantitative reverse transcription polymerase chain reaction.

^aMRD by MFC and relapse incidences in patients with a qRT-PCR between 0.1% and 10% patients.

The number of paired study sets after course four consolidations was small, and a total of 39 patients were eligible for assessment, including 23 with inv(16) and 16 patients with t(8;21). Five patients with inv(16) who had a median qRT-PCR level of 0.08% (range negative to 2.46%) relapsed by the end of this study. Only the patient with a qRT-PCR level of 2.46% was positive by MFC. In patients with t(8;21), MFC was positive in two patients (both with qRT-PCR ≥10%) and both relapsed. In contrast, only one of 13 patients with a negative MFC (median qRT-PCR 0.23%) relapsed. Overall, a negative MFC was not able to predict sustained remission in patients with a qRT-PCR level <10% (P = .227).

Salvage Therapy

A total of 17 patients were treated for relapsed disease, including six with t(8;21) and 11 with inv(16). Eleven patients were treated for first relapse, and 6 patients were treated for second relapse or beyond. As stated above, 15 patients received high-dose chemotherapy and two patients received DNMT-inhibitor therapy with decitabine and SGI-110, respectively. Seven patients with inv(16) and five patients with t(8;21) received HSCT after CR or CRi regardless of MRD status. Excluding one patient who was lost to follow-up, a total of six patients relapsed, including three patients who relapsed post HSCT. The MFC results and qRT-PCR levels had the strongest agreement in this group of patients. Although patients with a low qRT-PCR level (<0.1%) and a negative MFC were less likely to relapse, the assessment was difficult due to the confounding factor of HSCT and small sample size.

Correlation of MFC With qRT-PCR and Interphase FISH

Interphase FISH was performed on a total 159 samples, including 24 post induction, 65 during consolidation therapy, 54 during maintenance therapy, and 16 during salvage chemotherapy for relapsed AML. The FISH results in relation with MFC and qRT-PCR are shown in Table 2.

A total of 15 BM samples showed a qRT-PCR level ≥10%; interphase FISH was performed on 14 samples, and nine (64%) were positive showing a median 5.5% (range, 0.5%-22%) fusion signals. In this subset, MFC was positive in 13 of 15 (87%) samples. The two negative MFC samples were AML with RUNX1-RUNX1T1 postinduction. FISH detected 2.0% and 1.5% fusion signals, but became negative in a repeat BM specimen 1 week later. One patient completed induction and consolidation, and remained in CR with a qRT-PCR level of 0.28%, 18 months after diagnosis. The other patient completed treatment, achieved CR and underwent HSCT (pre-HSCT qRT-PCR 0.02%). However, post HSCT, the patient developed therapy-related (t)-MDS/AML with a complex karyotype including chromosome 3, 5, and 17 abnormalities. Of the two samples collected post HSCT when the patient developed t-MDS, RUNX1-RUNX1T1 qRT-PCR was 0.02% and negative, respectively, whereas MFC detected aberrant blasts differing from original AML on both occasions. The patient died of t-MDS with no recurrence of the t(8;21) AML.

A total of 195 samples had a qRT-PCR level ≤0.1%. Except for the above-mentioned patient who developed t-MDS, none of the samples were positive by MFC. Interphase FISH was performed in one.

Discussion

In this study, we compared the results of MFC with qRT-PCR and FISH in CBF AML. We found that a very low (<0.1%) or a high level (≥10%) of fusion transcripts detected by qRT-PCR correlated well with MFC results. However, in patients with a qRT-PCR level between 0.1% and 10%, MFC provides important information on residual disease status in the post induction samples and a positive MFC result is a strong predictor for disease recurrence. Overall, FISH is insensitive, but may provide additional confirmation in some cases.

We first looked into the datasets collected at the time of postinduction. We showed that a qRT-PCR level <0.1% was significantly associated with a reduced risk for AML relapse, whereas a qRT-PCR level ≥10% was highly predictive of future AML relapse. However, there was no difference in AML relapse rates among patients with qRT-PCR levels of 0.1% to 1% and 1% to 10%. These results are similar to what has been published by others.^5,7,11,25,26 In the United Kingdom Medical Research Council AML-15 trial,⁵ a >3 log reduction vs a <1 log reduction in RUNX1-RUNX1T1 or <10 copies vs >500 copies of CBFB-MYH11 transcripts in the post induction remission BM, were the most useful predictors of relapse. On the other hand, a log reduction between 1 to 2 and 2 to 3 in patients with AML associated with RUNX1-RUNX1T1, or a copy number between 10 and 500 of CBFB-MYH11, failed to predict AML relapse. In our study, MFC for MRD was performed in 54 post induction BM samples, and a positive MFC finding was highly predictive of AML relapse. Similar results have been reported by Perea et al.¹⁰ In 33 patients in this study, parallel MFC and qRT-PCR testing was performed on post induction CR BM. The results by MFC were negative in all patients with a qRT-PCR level <0.1% and none of these patients relapsed. In patients with a qRT-PCR level between 0.1% to 1% and 1% to 10%, a positive MFC was a strong discriminator for future AML relapse. A positive MFC result also predicted future relapse among patients with a qRT-PCR level >10% in RUNX1-RUNX1T1 AML post induction in whom a negative MFC study predicted a low risk for leukemia relapse. An exception to this rule included patients with concomitant myeloid sarcoma or extramedullary involvement in whom AML blasts were eradicated from BM but not from an extramedullary site. In an earlier study focused on patients with a negative MFC result in BM samples with ≥5% blasts by morphology, two patients had persistent AML with t(8;21)/RUNX1-RUNX1T1, and three had persistent AML with inv(16)/CBFB-MYH11 post induction. In all five patients the BM became negative on follow-up examination.²³ Molecular fusion transcripts may exist in nondividing cells or dying cells immediately after high-dose chemotherapy. Conceivably, low levels of MRD by PCR could be controlled by subsequent chemotherapy or immune reconstitution and hence fail to predict relapse. Interestingly, in the study by Zhu et al,²⁷ the predictive value of qRT-PCR MRD was not found at earlier checkpoints (after induction and first consolidation) but was most apparent after the second consolidation, corroborating the above hypotheses. Nevertheless, our data indicate that MRD assays by MFC may be especially helpful in patients with a qRT-PCR between 0.1% and 10%, to avoid unnecessary additional induction chemotherapy.²⁸

We also looked into the paired datasets collected at the end of consolidation. We found that a positive MFC remained a strong predictor for AML relapse in patients with t(8;21). RUNX1-RUNX1T1 was detectable at a low level in the most patients who were in clinical remission. This result is in keeping with what is commonly accepted in patients with AML associated with t(8;21); low levels of MRD in both BM and peripheral blood are compatible with durable remission, and molecular MRD negativity is not a prerequisite for long-term remission.⁵ In contrast, of the five AML patients with inv(16) who relapsed, only one patient with a qRT-PCR level of 2.46% had a positive MFC. It is known in patients with AML associated with CBFB-MYH11 that some long-term survivors fail to achieve complete PCR negativity, whereas about 20% to 30% of qRT-PCR-negative patients eventually relapse.^11,26,29 It is known that the sensitivity of MFC is lower than qRT-PCR, and it may consequently fail to detect a small residual tumor burden, leading to false-negative results. In keeping with this theory, we found that among patients who had a negative MFC result but experienced AML relapse, the qRT-PCR levels were low and FISH studies were all negative. Another reason that might have affected MFC sensitivity could be due to a frequent myelomonocytic immunophenotype of leukemic blasts of AML with CBFB-MYH11. While the myeloblast component has a very distinct immunophenotype, the monocytic blasts exhibit significant overlap with normal/regenerating monocytes.³⁰ The latter may make detection of MRD by MFC challenging if a low number of persistent/relapse AML blasts show predominantly monocytic differentiation after chemotherapy. Based on these findings, we suggest that during the follow-up interval, if patients have molecular relapse or rising qRT-PCR levels on serial monitoring, preemptive treatment should start regardless of MFC study findings. On the other hand, a positive MFC with a negative qRT-PCR in the follow-up BM may help to detect an emerging secondary malignancy. In one patient with t(8;21)/RUNX1-RUNX1T1 AML, qRT-PCR became negative post HSCT, but MFC detected small numbers of aberrant myeloblasts with an immunophenotype differing from the original AML. The positive MFC finding facilitated a diagnosis of t-MDS in this patient.

Approximately 30% of patients with CBF-AML eventually relapse. High-dose chemotherapy,³¹ followed by allogeneic HSCT in second complete remission is recommended for patients who relapse.³² In this study 17 patients were treated for relapsed AML, including six patients who had multiple relapses. The MFC results and qRT-PCR levels had better concordance in this group of patients. It is uncertain whether relapsed AML leukemic blasts exhibit more immunophenotypic aberrancies allowing for increased detection by MFC. Although patients with a low qRT-PCR level (< 0.1%) and a negative MFC appeared less likely to experience relapse, the assessment was difficult due to the confounding factor of HSCT and small sample size. A larger cohort of relapsed CBF AML patients is needed to assess the utility of MRD detected by qRT-PCR and MFC in this group of patients.

In summary, we showed that agreement between MFC and qRT-PCR is weak overall, and MFC and qRT-PCR provide complementary information for assessment of MRD in patients with CBF AML. Post induction, BM MFC study is useful in predicting AML relapse, particularly in patients with a discordant MFC and qRT-PCR result. During clinical follow-up, MFC may not be adequately sensitive to detect a minimal number of leukemic cells to predict early relapse; therefore, patients who have a molecular relapse or rising qRT-PCR levels on serial monitoring should be considered for preemptive treatment, including investigational agents and potentially HSCT, regardless of MFC findings. In patients with relapsed AML, the utility of MRD evaluation by either qRT-PCR or MFC needs evaluation in a larger cohort of patients.