The impact of race/ethnicity on survival in AML and associated cytogenetic subgroups

The inferior survival outcomes observed in young Black and Hispanic patients with AML in comparison with White patients is likely a multifactorial process. One potential factor is greater incidence of high-risk cytogenetics in these minority groups. However, a large-scale analysis of cytogenetics has not been performed. In addition, there is data lacking on survival outcomes within cytogenetic subgroups which would help clarify their clinical impact, and whether it is likely other factors also have influence on the significant survival disparity in young Black and Hispanic patients with AML.

A recent study published by Conneely, et al. in Blood Advances, compiled data from the Therapeutically Applicable Research to Generate Effective Treatments (TARGET) database, containing data from over 900 patients diagnosed with AML between 1996 and 2013. Cytogenetics, survival outcomes, and clinical features were compared based on race/ethnicity.¹

Study design

This was a retrospective cohort study using the TARGET AML database that contains data on young patients (aged <21 years for analysis) diagnosed with de novo AML. Race/ethnicity was characterized by White non-Hispanic, Hispanic, and Black non-Hispanic patients. Patients identifying as both Black and Hispanic were included within the Hispanic group.

The TARGET data is divided into discovery and validation subsets. The majority of patients in the discovery dataset had relapsed, while the validation set was a random set of patients with or without relapse. Both datasets were merged for this study.

Results

Patient characteristics for the analyzed cohort (N = 814) are summarized in table 1 below.

Table 1. Patient characteristics*

CNS, central nervous system; CR1, first complete remission; HSCT, hematopoietic stem cell transplant; IQR, interquartile range; KMT2Ar, KMT2A rearranged; WBC, white blood cell. *Adapted from Conneely, et al.1
Characteristic, % (unless otherwise stated)	White non-Hispanic (n = 552)	Black (n = 108)	Hispanic (n = 154)
Male	54.5	49.1	53.9
Age at diagnosis (range), median	10.1 (3.7–14.9)	11.3 (3.7–13.4)	8.9 (2.6–14.2)
Cytogenetic subtype
KMT2Ar	22.5	23.1	13.6
t(8;21)	10.9	21.3	17.5
Inv(16)	14.3	10.2	15.4
-7/-7q	0.9	1.9	0.6
-5/-5q	4.0	8.3	5.2
Normal	25.2	15.7	18.2
Other	24.1	32.4	30.5
Unknown	4.9	0	7.1
WBC at diagnosis (IQR), median	31.1 (10.9–86.7)	34.2 (15.9–96.9)	33.6 (11.8–90.8)
CNS involvement	6.3	11.1	5.2
HSCT received in CR1	15.9	9.3	14.9

As reported in table 1, both Black (OR, 2.22; 95% confidence interval [CI], 1.28–3.74) and Hispanic (OR, 1.74; 95% CI, 1.05–2.83) patients were more likely to have t(8;21) AML, while Black patients were also more likely to harbor 5q deletion/monosomy or 7q deletion/monosomy (OR, 2.29; 95% C, 1.06–4.68).

KMT2Ar AML was more common in white non-Hispanic patients compared with Hispanic patients (OR, 0.54; 95% CI, 0.32–0.88). Although the KTM2A subtype was not associated with Black race, the specific KMT2A rearrangement t(6;11) (q27;q23) conferring a poor prognosis, was more common in Black patients (OR, 6.12; 95% CI, 1.81–21.59). Additionally, KMT2Ar diagnosis in Black patients was associated with older age (median, 11.7 years [range, 2.3–15.1 years]) compared with other race/ethnicities (median, 3.2 years [range, 1.0–11.8 years])(p = 0.01).

In terms of treatment, fewer Black patients received a transplant following first complete remission compared with non-Hispanic White and Hispanic patients (Table 1), however the odds ratio did not reach statistical significance.

Survival outcomes by race/ethnicity

Survival and treatment outcomes for the entire cohort were analyzed, and demonstrated Black patients had inferior event-free survival (EFS) and increased risk of death compared with non-Hispanic White patients and Hispanic patients (Table 2). When analyzing this risk more specifically, Black patients were also more likely to have early death (≤50 days) (OR, 4.55; 95% CI, 1.44–13.96) or death as the first event (OR, 2.83; 95% CI, 1.46–5.29).

Table 2. Survival outcomes for entire cohort stratified by race/ethnicity*

CI, confidence interval; EFS, event-free survival; HR, hazard ratio; OS, overall survival. *Adapted from Conneely, et al.1
Race/ethnicity	EFS, %	HR (95% CI)	OS, %	HR (95% CI)
White non-Hispanic	50	Ref	71	Ref
Black	35	1.57 (1.21–2.04)	52	1.97 (1.47–2.65)
Hispanic	50	1.04 (0.81–1.34)	63	1.22 (0.91–1.63)

Survival outcomes by cytogenetic class

Survival outcomes by race/ethnicity were stratified for KTM2Ar and core binding factor (CBF) AML subtypes, with KTM2Ar inferring worse prognosis and CBF AML associated with more favorable prognosis.

For KMT2Ar AML, both Black and Hispanic patients showed inferior overall survival (OS) and EFS when harboring KTM2Ar. Death as the first event was also more likely in Black patients (OR, 10.08; 95% CI, 2.30–52.36). Finally, Hispanic patients with KMT2Ar were more likely to experience induction failure (OR, 9.52; 95% CI, 2.58–36.88).

Table 3. Survival outcomes for cytogenetic subgroups of AML*

CBF, core-binding factor; NS, not significant. *Adapted from Conneely, et al.1
Race/ethnicity	EFS, %	HR (95% CI)	OS, %	HR (95% CI)
KMT2Ar AML
White non-Hispanic	47	Ref	69	Ref
Black	14	2.31 (1.41–3.79)	47	2.54 (1.43–4.51)
Hispanic	29	2.20 (1.27–3.80)	43	2.07 (1.09–3.93)
CBF AML
White non-Hispanic	64	Ref	87	Ref
Black	44	1.93 (1.14–3.28)	61	3.24 (1.60–6.57)
Hispanic	67	0.99 (0.57–1.71)	77	2.03 (1.00–4.10)
CBF AML, inv(16)
White non-Hispanic	52	—	85	—
Black	36	NS	73	NS
Hispanic	73	NS	86	NS
CBF AML, t(8;21)
White non-Hispanic	80	ref	90	ref
Black	48	3.15 (1.44–6.92)	56	4.43 (1.69–11.66)
Hispanic	63	—	70	2.82 (1.02–7.80)

Black patients showed inferior OS and EFS within the CBF AML subtype and had increased risk of death as the first event (OR, 4.48; 95% CI, 1.18–17.11). Early death was also more common among Black patients (OR, 12.87; 95% CI, 1.59–265.05).

When analyzing further within two subgroups of CBF AML; inv(16) and t(8;21), EFS for Black patients was the lowest, however did not reach statistical significance. No significant differences were shown in OS for inv(16) AML, however both Black and Hispanic patients had inferior OS compared with non-Hispanic White patients when diagnosed with t(8;21) AML.

Survival outcomes of patients treated with gemtuzumab

EFS appeared to increase overall across AML patients and for those with KMT2Ar when receiving gemtuzumab, however this did not reach statistical significance. OS was not significantly improved with gemtuzumab use.

Gemtuzumab use had a significant impact on EFS only within Black patients when considering AML overall (HR, 0.84; 95% CI, 0.69–1.03). For KM2TAr AML, when analyzed as a whole cohort, a significant improvement in EFS with gemtuzumab was observed (HR, 0.64; 95% CI, 0.43–0.95). Stratifying by race revealed only White non-Hispanic patients had significant improvement in EFS (HR, 0.60; 95% CI, 0.36–0.98). Finally, when analyzing CBF AML, gemtuzumab again showed the greatest effect on EFS among Black patients (HR, 0.30; 95% CI, 0.10–0.92) with no other racial-ethnic group showing significant improvements.

Black patients were also the only group with significant improvement in OS within AML when receiving gemtuzumab (HR, 0.57; 95% CI, 0.33–0.99). In contrast, no effect was seen by race/ethnicity for KMT2Ar AML with gemtuzumab, and again no difference in improvement was observed for CBF AML. CD33 expression in myeloblasts, which is associated with response to gemtuzumab, did not differ significantly by race/ethnicity.

Conclusion

Overall, TARGET data provides evidence for increased incidence of high-risk cytogenetics in Black patients, which may account for some of the survival variation. However, survival differences were still observed within each cytogenetic subtype, with significantly inferior EFS and OS in black and Hispanic patients with KM2TAr AML compared with non-Hispanic White patients, despite its high prevalence in the white population. Survival differences were particularly notable in CBF AML, which is generally thought to confer a favorable prognosis; however, for Black patients with CBF AML, OS was more aligned with that seen in intermediate-high risk AML than favorable-risk AML.

Limitations of this study were firstly the inclusion of the discovery dataset which was biased toward patients experiencing relapse and therefore inferior outcomes. A small number of patients were observed within certain cytogenetic subgroups, making multivariate analysis impossible and increasing the size of confidence intervals. Finally, using self-reported race-ethnicity may not accurately categorize patients by genetic and non-genetic risk factors.

The inferior survival outcomes observed in young Black and Hispanic patients with AML in comparison with White patients is likely a multifactorial process. One potential factor is greater incidence of high-risk cytogenetics in these minority groups. However, a large-scale analysis of cytogenetics has not been performed. In addition, there is data lacking on survival outcomes within cytogenetic subgroups which would help clarify their clinical impact, and whether it is likely other factors also have influence on the significant survival disparity in young Black and Hispanic patients with AML.

A recent study published by Conneely, et al. in Blood Advances, compiled data from the Therapeutically Applicable Research to Generate Effective Treatments (TARGET) database, containing data from over 900 patients diagnosed with AML between 1996 and 2013. Cytogenetics, survival outcomes, and clinical features were compared based on race/ethnicity.¹

Study design

This was a retrospective cohort study using the TARGET AML database that contains data on young patients (aged <21 years for analysis) diagnosed with de novo AML. Race/ethnicity was characterized by White non-Hispanic, Hispanic, and Black non-Hispanic patients. Patients identifying as both Black and Hispanic were included within the Hispanic group.

The TARGET data is divided into discovery and validation subsets. The majority of patients in the discovery dataset had relapsed, while the validation set was a random set of patients with or without relapse. Both datasets were merged for this study.

Results

Patient characteristics for the analyzed cohort (N = 814) are summarized in table 1 below.

Table 1. Patient characteristics*

CNS, central nervous system; CR1, first complete remission; HSCT, hematopoietic stem cell transplant; IQR, interquartile range; KMT2Ar, KMT2A rearranged; WBC, white blood cell. *Adapted from Conneely, et al.1
Characteristic, % (unless otherwise stated)	White non-Hispanic (n = 552)	Black (n = 108)	Hispanic (n = 154)
Male	54.5	49.1	53.9
Age at diagnosis (range), median	10.1 (3.7–14.9)	11.3 (3.7–13.4)	8.9 (2.6–14.2)
Cytogenetic subtype
KMT2Ar	22.5	23.1	13.6
t(8;21)	10.9	21.3	17.5
Inv(16)	14.3	10.2	15.4
-7/-7q	0.9	1.9	0.6
-5/-5q	4.0	8.3	5.2
Normal	25.2	15.7	18.2
Other	24.1	32.4	30.5
Unknown	4.9	0	7.1
WBC at diagnosis (IQR), median	31.1 (10.9–86.7)	34.2 (15.9–96.9)	33.6 (11.8–90.8)
CNS involvement	6.3	11.1	5.2
HSCT received in CR1	15.9	9.3	14.9

As reported in table 1, both Black (OR, 2.22; 95% confidence interval [CI], 1.28–3.74) and Hispanic (OR, 1.74; 95% CI, 1.05–2.83) patients were more likely to have t(8;21) AML, while Black patients were also more likely to harbor 5q deletion/monosomy or 7q deletion/monosomy (OR, 2.29; 95% C, 1.06–4.68).

KMT2Ar AML was more common in white non-Hispanic patients compared with Hispanic patients (OR, 0.54; 95% CI, 0.32–0.88). Although the KTM2A subtype was not associated with Black race, the specific KMT2A rearrangement t(6;11) (q27;q23) conferring a poor prognosis, was more common in Black patients (OR, 6.12; 95% CI, 1.81–21.59). Additionally, KMT2Ar diagnosis in Black patients was associated with older age (median, 11.7 years [range, 2.3–15.1 years]) compared with other race/ethnicities (median, 3.2 years [range, 1.0–11.8 years])(p = 0.01).

In terms of treatment, fewer Black patients received a transplant following first complete remission compared with non-Hispanic White and Hispanic patients (Table 1), however the odds ratio did not reach statistical significance.

Survival outcomes by race/ethnicity

Survival and treatment outcomes for the entire cohort were analyzed, and demonstrated Black patients had inferior event-free survival (EFS) and increased risk of death compared with non-Hispanic White patients and Hispanic patients (Table 2). When analyzing this risk more specifically, Black patients were also more likely to have early death (≤50 days) (OR, 4.55; 95% CI, 1.44–13.96) or death as the first event (OR, 2.83; 95% CI, 1.46–5.29).

Table 2. Survival outcomes for entire cohort stratified by race/ethnicity*

CI, confidence interval; EFS, event-free survival; HR, hazard ratio; OS, overall survival. *Adapted from Conneely, et al.1
Race/ethnicity	EFS, %	HR (95% CI)	OS, %	HR (95% CI)
White non-Hispanic	50	Ref	71	Ref
Black	35	1.57 (1.21–2.04)	52	1.97 (1.47–2.65)
Hispanic	50	1.04 (0.81–1.34)	63	1.22 (0.91–1.63)

Survival outcomes by cytogenetic class

Survival outcomes by race/ethnicity were stratified for KTM2Ar and core binding factor (CBF) AML subtypes, with KTM2Ar inferring worse prognosis and CBF AML associated with more favorable prognosis.

For KMT2Ar AML, both Black and Hispanic patients showed inferior overall survival (OS) and EFS when harboring KTM2Ar. Death as the first event was also more likely in Black patients (OR, 10.08; 95% CI, 2.30–52.36). Finally, Hispanic patients with KMT2Ar were more likely to experience induction failure (OR, 9.52; 95% CI, 2.58–36.88).

Table 3. Survival outcomes for cytogenetic subgroups of AML*

CBF, core-binding factor; NS, not significant. *Adapted from Conneely, et al.1
Race/ethnicity	EFS, %	HR (95% CI)	OS, %	HR (95% CI)
KMT2Ar AML
White non-Hispanic	47	Ref	69	Ref
Black	14	2.31 (1.41–3.79)	47	2.54 (1.43–4.51)
Hispanic	29	2.20 (1.27–3.80)	43	2.07 (1.09–3.93)
CBF AML
White non-Hispanic	64	Ref	87	Ref
Black	44	1.93 (1.14–3.28)	61	3.24 (1.60–6.57)
Hispanic	67	0.99 (0.57–1.71)	77	2.03 (1.00–4.10)
CBF AML, inv(16)
White non-Hispanic	52	—	85	—
Black	36	NS	73	NS
Hispanic	73	NS	86	NS
CBF AML, t(8;21)
White non-Hispanic	80	ref	90	ref
Black	48	3.15 (1.44–6.92)	56	4.43 (1.69–11.66)
Hispanic	63	—	70	2.82 (1.02–7.80)

Black patients showed inferior OS and EFS within the CBF AML subtype and had increased risk of death as the first event (OR, 4.48; 95% CI, 1.18–17.11). Early death was also more common among Black patients (OR, 12.87; 95% CI, 1.59–265.05).

When analyzing further within two subgroups of CBF AML; inv(16) and t(8;21), EFS for Black patients was the lowest, however did not reach statistical significance. No significant differences were shown in OS for inv(16) AML, however both Black and Hispanic patients had inferior OS compared with non-Hispanic White patients when diagnosed with t(8;21) AML.

Survival outcomes of patients treated with gemtuzumab

EFS appeared to increase overall across AML patients and for those with KMT2Ar when receiving gemtuzumab, however this did not reach statistical significance. OS was not significantly improved with gemtuzumab use.

Gemtuzumab use had a significant impact on EFS only within Black patients when considering AML overall (HR, 0.84; 95% CI, 0.69–1.03). For KM2TAr AML, when analyzed as a whole cohort, a significant improvement in EFS with gemtuzumab was observed (HR, 0.64; 95% CI, 0.43–0.95). Stratifying by race revealed only White non-Hispanic patients had significant improvement in EFS (HR, 0.60; 95% CI, 0.36–0.98). Finally, when analyzing CBF AML, gemtuzumab again showed the greatest effect on EFS among Black patients (HR, 0.30; 95% CI, 0.10–0.92) with no other racial-ethnic group showing significant improvements.

Black patients were also the only group with significant improvement in OS within AML when receiving gemtuzumab (HR, 0.57; 95% CI, 0.33–0.99). In contrast, no effect was seen by race/ethnicity for KMT2Ar AML with gemtuzumab, and again no difference in improvement was observed for CBF AML. CD33 expression in myeloblasts, which is associated with response to gemtuzumab, did not differ significantly by race/ethnicity.

Conclusion

Overall, TARGET data provides evidence for increased incidence of high-risk cytogenetics in Black patients, which may account for some of the survival variation. However, survival differences were still observed within each cytogenetic subtype, with significantly inferior EFS and OS in black and Hispanic patients with KM2TAr AML compared with non-Hispanic White patients, despite its high prevalence in the white population. Survival differences were particularly notable in CBF AML, which is generally thought to confer a favorable prognosis; however, for Black patients with CBF AML, OS was more aligned with that seen in intermediate-high risk AML than favorable-risk AML.

Limitations of this study were firstly the inclusion of the discovery dataset which was biased toward patients experiencing relapse and therefore inferior outcomes. A small number of patients were observed within certain cytogenetic subgroups, making multivariate analysis impossible and increasing the size of confidence intervals. Finally, using self-reported race-ethnicity may not accurately categorize patients by genetic and non-genetic risk factors.