Epidemiology and survival trend of Adult T-cell Leukemia/Lymphoma in the United States

Abstract

BACKGROUND:

Globally, 5 to 10 million people are infected with human T-cell leukemia virus type 1 (HTLV-1), which causes adult T-cell leukemia/lymphoma (ATLL) in 2–5% of the carriers¹. ATLL is a rare but extremely aggressive malignancy that can be challenging to diagnose. Very little data exist on the incidence patterns of ATLL in the United States (US).

METHODS:

ATLL cases reported to the National Program of Cancer Registries (NPCR), Surveillance, Epidemiology, and End Results (SEER) and New York State Cancer Registry (NYSCR) were used for the study. Age-adjusted incidence rates were calculated by age, race/ethnicity, gender, and year of diagnosis. 5-year survival rate was compared among race/ethnicity groups using the SEER data.

RESULTS:

During 2001–2015, 2148 ATLL cases were diagnosed in the US, 18% of which were in the State of New York (NYS). NYS had the highest incidence rate for ATLL with a rising trend especially among non-Hispanic blacks (NHBs) while the incidence has been stable across the remainder of the US. NHBs were diagnosed at a younger median age (54 years) and had a shorter overall survival (6 months). In NYC, only 22.6% of the ATLL cases diagnosed were born in North America.

CONCLUSIONS:

This is the largest epidemiological study of ATLL in the US showing a rising incidence in NYC. NHBs have a younger age at presentation and poor overall survival. The rising incidence is largely comprised of NHBs originating from the Caribbean.

Precis for use in the Table of Contents:

Very little data exists on the incidence patterns of ATLL in the United States. This is the largest comprehensive epidemiological study of ATLL in the US using the NPCR, SEER and NYSCR databases and shows a rising incidence with younger age at presentation and poor overall survival in NHBs.

Introduction

Adult T-cell leukemia/lymphoma (ATLL) is a rare and extremely aggressive T cell neoplasm with a dismal prognosis. The disease is caused by a retrovirus human T cell leukemia virus (HTLV-1) that is endemic in southwestern Japan, sub-Saharan Africa, South America, the Caribbean region, and certain areas in the middle east and Australia². HTLV-1 is also prevalent at low levels in the United States (US)^3–5. Within the US, New York, and Florida contain the majority of ATLL cases due to the higher number of Caribbean immigrants in these areas^6–9.

Previously, two studies estimated the incidence of ATLL in the US based on the Surveillance, Epidemiology, and End Results (SEER) registries. In one study, 362 cases of ATLL were reported in approximately 28% of the US population between 2000 to 2012¹⁰; in the other,140 cases of ATLL were reported between 1993 to 2008 in approximately 9.5% of the US population¹¹. Because SEER does not include states like New York and Florida where most ATLL cases are seen, a true estimate of the disease burden in this country remains unknown^10,11. (Of note, New York State Cancer Registry (NYSCR) became a part of the SEER program in 2018)

The current study aims to examine the epidemiology and clinical outcomes of ATLL in the US using population-based cancer registry data with a focus on NYS. This is the largest epidemiological study of ATLL in the US. Our results demonstrate a rising incidence with younger presentation and poor overall survival in Non-Hispanic Blacks (NHBs).

Methods

Data Sources

ATLL (HTLV-1 positive) (includes all variants) cases diagnosed between 2001 to 2015 were identified using the International Classification of Diseases for Oncology, Third Edition codes (ICD-O-3) as 9827/3 of patients ≥ 15 years of age. Three data sources were used in the study.

The first data source we used was the combined dataset obtained from the Centers for Disease Control and Prevention’s (CDC) National Program of Cancer Registries (NPCR). This dataset contains cancer data compiled from population-based registries that participate in the NPCR and/or the NCI’s SEER Program and meet high-quality data criteria. These registries cover approximately 99% of the U.S. population. However, since the NPCR dataset does not include survival information, we used data from SEER18 registries for detailed survival analysis, which covers 27.8% of the entire US population. In addition, since the NPCR dataset contains only limited data fields and the SEER dataset did not include NYS until 2018, where a large fraction of ATLL cases are diagnosed, we also obtained data from the NYSCR.

Data Analysis:

Age-adjusted incidence rate with a 95% confidence interval for ATLL was calculated using three different datasets, by state (NPCR dataset only), gender, race/ethnicity (non-Hispanic white (NHW), NHB, non-Hispanic Asian Pacific Islander (NHAPI), and Hispanic), age group (15–39, 40–49, 50–59, 60–69, 70–79, 80+), year of diagnosis (2001–2005, 2006–2010, 2011–2015), and region (NYSCR dataset only). All rates were age-adjusted to the 2000 standard US population and expressed as per 100,000 population. Case counts less than 16 were suppressed in the NPCR dataset by default. The SEER*Stat software (version 8.3.5; Surveillance Research Program, National Cancer Institute, Bethesda, MD) was used for rate calculation. Incidence trends from 2001 to 2015 were also evaluated using Joinpoint regression model. Using the SEER18 data, the 5-year observed survival rate was calculated using the Kaplan Meier method for different race/ethnicity groups. The Log-rank test was used to test statistical significance. Since NYSCR data contains the place of birth, the frequency distribution by patients’ birthplace was tabulated.

Results:

NYS had the highest age-adjusted incidence rate for ATLL in the US

2,148 ATLL cases were reported from 2001 to 2015 in the US, representing an overall rate of 0.06 per 100,000 population. New York had the highest rate (0.16 per 100,000) among all states in the nation, followed by Hawaii (0.15 per 100,000) and Florida (0.11 per 100,000) (Figure 1A, B). SEER18 reported 457 cases across the 15 year period, showing a similar ATLL incidence rate as the entire U.S population (0.05 per 100,000), however, it did not report on some of the most populous states such as New York and Florida, which are known to have a higher incidence of ATLL. In terms of the accumulative case count during this 15-year period, the top 5 states are New York (383; 17.8%), Florida (285; 13.3%), California (137; 6.4%), Texas (136; 6.3%), and New Jersey (89; 4.1%).

Figure 1.

Geographic and racial characteristics of ATLL in the US

(a) Age-Adjusted Incidence Rate of ATLL in the US (2001–2015) (NPCR)

(b) Racial and ethnic distribution of ATLL by state (NPCR)

ATLL predominantly affected non-Hispanic blacks with a younger median age at diagnosis and had no gender difference

The age-adjusted ATLL incidence rates by sex, race/ethnicity, age group, and time period of diagnosis are shown in Table 1. NHB had the highest ATLL incidence rate among all race/ethnicity groups, and the rate difference was particularly significant in NYS, with the rate about 20 times higher in NHB compared to NHW. Hispanics also had higher incidence rates than NHW, but these differences were only shown in NPCR and NYSCR data, but not in SEER data. As for most cancers, the ATLL incidence rate increased with age. NHB and Hispanic were diagnosed at a relatively younger age compared to NHW (Figure 2). Across all race/ethnic groups, males showed a slightly higher ATLL incidence rate than females, but the difference was not statistically significant (not shown).

Table 1.

ATLL incidence rate by demographic characteristics based on data from National Program of Cancer Registries (NPCR), Surveillance, Epidemiology, and End Results (SEER), and New York State Cancer Registry (NYSCR) respectively, 2001–2015

		Source of data
		NYSCR			SEER18			NPCR
		Count	Percent	Rate (95% CI)	Count	Percent	Rate (95% CI)	Count	Percent	Rate (95% CI)
Total		385	100	0.16 (0.14–0.17)	457	100	0.05 (0.04–0.05)	2148	100	0.06 (0.06–0.06)
Sex	Male	185	48	0.17 (0.14–0.19)	227	50	0.05 (0.04–0.06)	1106	51	0.07 (0.06–0.07)
	Female	200	52	0.15 (0.13–0.17)	230	50	0.04 (0.04–0.05)	1042	49	0.05 (0.05–0.06)
Race/ Ethnicity	Non-Hispanic White	63	16	0.04 (0.03–0.05)	212	46	0.03 (0.03–0.04)	1009	47	0.04 (0.03–0.04)
	Non-Hispanic Black	252	65	0.75 (0.66–0.85)	143	31	0.14 (0.12–0.17)	817	38	0.20 (0.19–0.22)
	Non-Hispanic Asian-Pacific Islander	NA	NA	NA	47	10	0.06 (0.04–0.07)	82	4	0.05 (0.04–0.07)
	Hispanic	60	16	0.22 (0.16–0.28)	47	10	0.03 (0.02–0.04)	205	10	0.06 (0.05–0.06)
Age Group	15–39	46	12	0.05 (0.04–0.06)	64	14	0.02 (0.01–0.02)	328	15	0.02 (0.02–0.02)
	40–49	68	18	0.16 (0.12–0.20)	51	11	0.03 (0.02–0.04)	272	13	0.04 (0.04–0.05)
	50–59	85	22	0.22 (0.18–0.28)	86	19	0.05 (0.04–0.07)	390	18	0.07 (0.06–0.07)
	60–69	90	23	0.34 (0.28–0.42)	83	18	0.08 (0.06–0.10)	443	21	0.11 (0.10–0.12)
	70–79	64	17	0.39 (0.30–0.50)	94	21	0.15 (0.12–0.18)	424	20	0.17 (0.15–0.18)
	80+	32	8	0.28 (0.19–0.40)	79	17	0.19 (0.15–0.24)	291	14	0.18 (0.16–0.20)
Year of Diagnosis	2001–2005	79	21	0.10 (0.08–0.13)	160	35	0.053 (0.05–0.06)	643	30	0.06 (0.05–0.06)
	2006–2010	136	35	0.17 (0.14–0.20)	141	31	0.04 (0.04–0.05)	730	34	0.06 (0.05–0.06)
	2011–2015	170	44	0.2 (0.17–0.23)	156	34	0.04 (0.04–0.05)	775	36	0.06 (0.05–0.06)
NYS Region	NYC	291	76	0.30 (0.26–0.33)	NA	NA	NA	NA	NA	NA
	Upstate NY	94	24	0.06 (0.05–0.08)	NA	NA	NA	NA	NA	NA

NA: Not available

Figure 2:

Median age at diagnosis (NYSCR and SEER18).

As shown in Fig 1B, the proportion of NHB ATLL cases is lower in some states such as California and Alabama. However, when looking at the number of ATLL cases per 100,000 in each race, the incidence rate remains highest in NHB in all states except Hawaii (Supplemental Fig 2 and Supplemental Table 1).

ATLL incidence rate was stable in the US but was rising in NYS

We also examined changes in ATLL incidence rate over time. As shown in Figure 3, the incidence of ATLL in the U.S. remained largely stable from 2001 to 2015 (Annual Percentage Change (APC) = 0.28%, p>0.05), but showed an increasing trend in the state of New York (APC = 6.9%, p<0.0001). Nationwide, there was an increasing incidence in NHB (APC = 3.7%, p<0.05) mirrored by a decreasing incidence in NHW (APC = −2.7% p<0.05).

Figure 3:

Age-Adjusted Incidence Rate of ATLL in the US (NPCR and SEER) and NY State (NYSCR) over time (2001–2015)

Non-Hispanics blacks had the worst observed survival in ATLL among all race/ethnicity groups

Since neither NPCR nor NYSCR datasets have survival information, SEER18 data was used for survival analysis. The observed 5-year survival rate for ATLL patients was considerably low (23.4%) with a median survival time of 11 months. Non-Hispanic blacks had the 5-year worst survival rate (13.1%) compared to other race/ethnicity groups (17.1% for NHAPI, 26.5% for Hispanic, and 30.1% for NHW). The log-rank test detected a significant survival difference among the race/ethnicity groups (p < 0.0001, Figure 4).

Figure 4:

SEER ATLL survival estimates stratified by race/ethnicity (2001–2015)

The majority of ATLL patients in NYS live in New York City (NYC) and were born in the Caribbean region

Among 385 ATLL cases diagnosed in NYS, 76% of them lived in NYC and 24% lived in the remainder of the state. The incidence rate was also much higher in NYC (0.30 per 100,000) compared to the rest of the state (0.06 per 100,000). 56% of the patients were born in the Caribbean, 22.6% in North America and 4.9% in South America (Figure 5).

Figure 5:

Distribution of ATLL cases (n) in NY by place of birth (total cases n = 385)

Discussion:

This is the first comprehensive study that examines the incidence of ATLL in the US using NPCR, SEER and NYSCR datasets. In addition to a much wider population coverage, this study also provides more recent data compared to the two previous reports^10,11. Due to the rarity and heterogeneity in the distribution of this disease, prior studies in the US that included single institutional retrospective analyses^6–8 or SEER data that excluded Florida and New York do not provide accurate measures of disease burden^10,11. One study used the SEER9 database that consists of 9.5% of the US population and identified a rising incidence rate from 1993 to 2008 with an annual percent change of +6.2%¹¹. The other study analyzed SEER18 data from 2000 to 2012 and identified 362 cases with a higher incidence in NHBs and worse survival as well¹⁰.

In the US, the age-adjusted incidence rate of ATLL remained largely stable at 0.06 per 100,000 during the 15-year period of 2001–2015 (Figure 3); however, there has been a significant increase in the number of NHB cases (Figure S1B). Contrasting the stable nationwide incidence rate, the incidence of ATLL in NYS has more than doubled and the annual increase is statistically significant (Figure 3). This can be partially attributed to the fact that NYC has remained the mecca of immigration to the US, and this is particularly true for immigrants from the Caribbean region where HTLV-1 is endemic¹². Consistent with this interpretation, 56.1% of ATLL patients in NYS were born in Caribbean countries. This confirms previously reported single institution findings of ATLL in a predominantly Caribbean immigrant population^6,7. However, it is important to note that 22.6% of ATLL patients in NYC are born in North America because ATLL is often missed in American-born patients as it is not expected and this has potential implications for treatment. This low-level endemicity of HTLV-1 is likely accounted for by vertical transmission from infected mothers in the US or horizontal transmission through sexual intercourse to people born in the US^13,14.

We did not observe any gender difference in ATLL incidence which is similar to the report from Japan¹⁵ (Table 1). As expected, our analysis reveals a distinct geographic distribution with 30% of ATLL cases diagnosed in the states of New York and Florida (Figure 1A). Interestingly, the age-adjusted incidence rate is similar in New York and Hawaii but there is a difference in the race-ethnicity distribution of the ATLL patients for the two states (Figure 1B). This is in line with the high percentage of Japanese Americans that live on the Hawaii islands. Except for Hawaii, all other states had a higher age-adjusted incidence rate of ATLL in NHBs. This would mean that even though the absolute number of NHB cases may be low in some states such as Alabama or California, the incidence of ATLL cases per 100,000 remains high in this subgroup as compared to other race-ethnicities. This is consistent with the demographic differences in these states and previously reported studies^16,17.

The incidence of ATLL increases with age with the highest incidence rate found in the older age categories (age group of 70–79 in NYSCR, age group of 80+ in NPCR and SEER, Figure S1B). NHBs were more likely to be diagnosed at a younger age group as compared to other race-ethnic groups. The median age at diagnosis in the US was 62 in the SEER18 data and 59 in NYSCR. In Japan, the median age of diagnosis is 68 years, which is almost a decade later than in the US¹⁵. The lower median age at diagnosis in the US, especially in NHBs, compared to the Japanese needs further study. In NYS, the proportion of NHBs was higher as compared to the rest of the country and might account for the lower median age at diagnosis as compared to SEER. We also analyzed the survival of patients with ATLL by race-ethnicity based on SEER18 data. Survival variation among race-ethnicity groups was observed with the poorest survival for NHBs identified (Figure 4). This is consistent with the survival reported by others and us in a predominantly African American cohort^6–8. The worse outcomes in NHBs may be related to a delay in diagnosis or differences in genetic characteristics⁹.

One of the challenges in estimating ATLL prevalence is the paucity of data that directly estimates HTLV-1 prevalence in the US^3–5. The closest estimation we have is by screening US blood donors for HTLV-1 and blood donors enrich for populations in good health with low-risk behaviors, thus underestimating the prevalence by 2 to 5-fold. A study reported the seroprevalence and demographics of HTLV-1 in first time blood donors in the US between 2002 and 2009 in the southeast, southwest, west and north central regions but did not include the northeast regions such as New York, Maryland and some southeast regions such as Florida where a majority of ATLL cases are located. Overall, the estimates included 5.1 per 100,000 donors screened for HTLV-1 in these regions and the distribution was not uniform. High HTLV-1 incidence was limited to certain small regions in a particular state rather than a more uniform distribution. Even though some of these high-density regions corresponded to states where ATLL cases or incidence was high such as Texas, California, and Alabama, not all regions of higher prevalence were associated with a higher incidence of ATLL³. Hence, it is difficult to associate regions of seroprevalence of HTLV-1 with ATLL in the US and the exact reasons for this differing racial makeup by state are not known.

Our study has several limitations. The datasets available for this study lack information on disease subtype, treatment, mode of transmission of disease, detailed clinical features of the patient, presence or absence of other comorbidities. Subtype-specific information would be valuable given the differences in survival based on ATLL subtype¹⁸. As such, our analysis was limited to demographic pattern and survival trend¹⁹. Furthermore, race is self-reported which can lead to a misclassification bias. NHWs could include not only Caucasians but other individuals that identify themselves as white²⁰. Because ATLL is not a disease seen in Caucasians, the numbers of NHW cases in the SEER18 and NPCR datasets could be the result of potential misclassification. Another shortcoming of this study is that we were unable to confirm HTLV-1 seropositivity information for the ATLL cases although the 9827/3 coding requires this. Since HTLV-1 positivity is a prerequisite for ATLL diagnosis, it would be useful for cancer registries to capture this information moving forward.

Future directions include developing a national HTLV-1 surveillance program with mandatory reporting (like HIV) in the US to accurately identify cases as well as requiring documentation of HTLV-1 status in all cancer databases. Until then measures to raise awareness and prevent transmission such as fetal-maternal transmission through breast feeding in high-risk groups should be undertaken.

In summary, this is the first report that provides comprehensive analysis on the magnitude, epidemiology, and prognosis of ATLL in the US. The secondary focus of this study is on NYS which has the highest incidence of ATLL and is not included in SEER18. The geographic and racial/ethnic characteristics demonstrated in this study are expected to facilitate future clinical and laboratory research of this rare and fatal disease.