IJC International Journal of Cancer

Lymphoid neoplasm incidence by WHO subtype in Australia 1982–2006 Marina T. van Leeuwen1, Jennifer J. Turner2,3, David J. Joske4,5, Michael O. Falster1,6, Preeyaporn Srasuebkul1, Nicola S. Meagher1, Andrew E. Grulich7, Graham G. Giles8,9 and Claire M. Vajdic1 1

Adult Cancer Program, Lowy Cancer Research Centre, Prince of Wales Clinical School, The University of New South Wales, Sydney, NSW, Australia Department of Histopathology, Douglass Hanly Moir Pathology, Sydney, NSW, Australia 3 The Australian School of Advanced Medicine, Macquarie University, Sydney, NSW, Australia 4 Department of Haematology, Sir Charles Gairdner Hospital, Perth, WA, Australia 5 School of Medicine and Pharmacology, The University of Western Australia, Perth, WA, Australia 6 Centre for Health Research, School of Medicine, University of Western Sydney, Sydney, NSW, Australia 7 Kirby Institute, The University of New South Wales, Sydney, NSW, Australia 8 Cancer Epidemiology Centre, Cancer Council Victoria, Melbourne, VIC, Australia 9 Centre for Epidemiology and Biostatistics, School of Population and Global Health, The University of Melbourne, Parkville, VIC, Australia 2

Epidemiology

There are limited data characterizing the subtype-specific incidence of lymphoid neoplasms in the World Health Organization (WHO) Classification era. Data were obtained on all incident lymphoid neoplasms registered in Australia during 1982–2006. Subtypes were grouped using the InterLymph nested hierarchical classification, based on the 2008 WHO Classification. Temporal trends were examined using Joinpoint regression; average annual percentage change in incidence was computed. Multiple Poisson regression was used to compare incidence by sex and age. The incidence of all non-Hodgkin lymphoma (NHL) increased by 2.5%/year during 1982–1996 and was stable thereafter. During 1997–2006, several mature B- and natural killer (NK)-/T-cell NHL subtypes increased in incidence, including diffuse large B-cell (1.3%/year), follicular (2.5%/year), Burkitt (6.8%/year), marginal zone (13.2%/year), mantle cell (4.2%/year), peripheral T-cell lymphoma (4.7%/year) and plasmacytoma (7.1%/year). While chronic lymphocytic leukemia incidence was stable, small lymphocytic lymphoma incidence declined (8.1%/year). Hodgkin lymphoma (HL) incidence increased during 1997–2006 (2.2%/year), both classical (4.3%/year) and nodular lymphocyte predominant (12.1%/year) HL. Diagnostic artifact, evidenced by a sustained decline in the incidence of NHL not otherwise specified (NOS; 5.8%/year) and lymphoid neoplasms NOS (5.6%/year), limits the interpretation of temporal trends for some subtypes. A marked male predominance was observed for almost all subtypes. Incidence of mature B- and NK-/T-cell NHL subtypes increased sharply with age, except for Burkitt lymphoma/leukemia. For HL subtypes, a bimodal age distribution was only evident for nodular sclerosis HL. Variation in incidence patterns over time and by sex and age supports etiological differences between lymphoid neoplasm subtypes.

Key words: lymphoma, non-Hodgkin lymphoma, Hodgkin lymphoma, incidence, trend Additional Supporting Information may be found in the online version of this article. Grant sponsor: Fellowships from the Leukaemia Foundation of Australia (MTvL); Grant sponsor: the Australian National Health and Medical Research Council; Grant numbers: MTvL ID1012141, CMV ID1023159 and AEG ID568819; Grant sponsor: Cancer Institute New South Wales; Grant number: CMV ID10/CDF/2-42 DOI: 10.1002/ijc.28849 History: Received 11 Nov 2013; Accepted 25 Feb 2014; Online 18 Mar 2014 Correspondence to: Claire M. Vajdic, Adult Cancer Program, Lowy Cancer Research Centre, Prince of Wales Clinical School, The University of New South Wales, Sydney, New South Wales 2052, Australia, Tel.: 1612-9385-1424, Fax: 1612-9385-1430, E-mail: [email protected]

C 2014 UICC Int. J. Cancer: 135, 2146–2156 (2014) V

Lymphoid neoplasms, including non-Hodgkin lymphoma (NHL), Hodgkin lymphoma (HL), plasma cell neoplasms and lymphoid leukemias, collectively represent the fifth most common malignancy in Australia.1 Their etiology is largely unknown, with specific infections and immunodeficiency being the only established risk factors. Descriptive studies in the United States and Europe show differences in demographic2–4 and temporal3,5 incidence patterns between lymphoid neoplasm subtypes, which, together with observations from analytical studies,6 suggest differences in etiology. There have been few comprehensive studies of subtypespecific incidence patterns.2–4,7 The existence of multiple classification schemes8 and variable exclusion criteria make comparisons challenging. These schemes include the Working Formulation (WF, 1982), the Revised European-American Lymphoma (REAL) Classification (1994) and the World Health Organization (WHO) Classification of Tumours of

2147

van Leeuwen et al.

What’s new? Lymphoid neoplasms are a very heterogeneous group of cancers, but limited data exist on the incidence of specific lymphoma subtypes. The authors use population-based cancer registry data to systematically describe subtype-specific incidence patterns of lymphoid neoplasms in Australia. While the incidence rate of Non-Hodgkin lymphoma stabilized between 1997 and 2006, increases were observed for diffuse large B-cell, follicular, peripheral T-cell, plasmacytoma, and Hodgkin lymphoma. These observations may help answer questions related to etiology and public health burden associated with the individual lymphoid neoplasm subtypes.

Material and Methods Data sources

National data for all incident lymphoid neoplasms (ICD-O-3 9590–9734, 9760–9764, 9820–9837, 9940 and 9948), as notified by statutory obligation to Australian cancer registries, were obtained for 1982–2006. Australian cancer registries implemented ICD-O-2 in 1997 and ICD-O-3 by 2002. Prior to this, a combination of ICD-9 and The Systematized Nomenclature of Medicine Second Edition were used. All data have since been mapped to ICD-O-3. Although the first and second editions of the ICD-O incorporated elements of the WF and REAL Classification, respectively, cancer registry-based studies of cases originally diagnosed and coded prior to the introduction of ICD-O-3 may have suffered from lack of specificity and a large proportion of not otherwise specified (NOS) cases. Grouping codes into subtypes is one method for reducing misclassification for epidemiological studies.8 Therefore, data were grouped using the most recent InterLymph nested hierarchical classification.6,11 This system defines standardized groupings, or hierarchical levels, of lymphoid neoplasms, providing a method for including subtypes originally classified under previous schemes. All subtypes of NHL and HL were included and were grouped according to the highest hierarchical level afforded by the data; at least level five for NHL, and level four for HL. The higher the hierarchical level, the better defined the biological entity. The exceptions were precursor lymphoid neoplasms and NHL NOS, for which data on lineage were not available at the cancer registry level. In accordance with the InterLymph classification, plasma cell neoplasms and lymphoid leukemias were included under the NHL rubric.11 Data were stratified by calendar year of diagnosis, sex and 5-year age group. Population estimates were obtained from the C 2014 UICC Int. J. Cancer: 135, 2146–2156 (2014) V

Australian Bureau of Statistics for the census years 1981, 1986, 1991, 1996, 2001 and 2006 and annual, intercensal population estimates generated using linear interpolation. All data were obtained with institutional ethics committee approval.

Statistical analysis

For each lymphoid neoplasm subtype, directly age- and sexstandardized incidence rates (per 100,000 person-years) were calculated for the years 1982–2006, using the World (2000) population as the standard. For descriptive purposes, average annual rates were also calculated for periods defined by the use of different classification systems in Australia, 1982–1996 (pre-ICD-O-2) and 1997–2006 (post-ICD-O-2). Segmented linear (Joinpoint) regression modeling was used to examine temporal trends in age-standardized rates (Joinpoint Regression Program Version 4.0.1, January 2013, Statistical Research and Applications Branch, National Cancer Institute). Joinpoint modeling assumes that there is no change in trend over time and then selects points at which the trend changes significantly, based on Monte Carlo permutation testing. The annual percentage change (APC) in rates, with 95% confidence intervals (CIs), is estimated for each distinct trend or segment.12 The average APC (AAPC), a weighted average of the segment APCs over a prespecified time period, can also be determined.13 Trends were examined for 1982–1996 (pre-ICD-O-2) and 1997–2006 (post-ICD-O2) for all NHL, HL, NHL NOS, lymphoid neoplasms NOS and subtypes for which diagnostic reliability has remained sufficiently high over time, including diffuse large B-cell lymphoma (DLBCL), follicular lymphoma, plasma cell neoplasms and mycosis fungoides/Sezary syndrome (MF/SS).11 To minimize the potential impact of changes in classification on temporal trends, trends for all other subtypes were examined for 1997–2006 (post-ICD-O-2) only. The AAPC, with 95% CI, was calculated for each period. Trends were examined overall, as well as by sex and age group for selected subtypes. Age groups were selected based on Australian age-specific incidence patterns: 0–14, 15–24, 25–44, 45–64 and 65 years for HL, precursor lymphoid neoplasms and Burkitt lymphoma/leukemia (BL) and 0–44, 45–64 and 65 years for other subtypes. Multiple Poisson regression was used to compare incidence between sexes and age groups as adjusted incidence rate ratios with 95% CI. The models included a term for

Epidemiology

Haematopoietic and Lymphoid Tissues (2001, 2008).9,10 The WHO Classification represents consensus classification for clinical, pathologic and scientific use and wholly incorporates the International Classification of Diseases for Oncology, Third Edition (ICD-O-3, 2001).6,11 Herein we use population-based cancer registry data to systematically describe subtype-specific incidence patterns of lymphoid neoplasms in Australia 1982–2006 according to the InterLymph nested hierarchical classification,11 a widely adopted epidemiological classification based on the 2008 WHO Classification.

2148

calendar year and the log of the population size was used as an offset and a scale parameter was included to adjust for overdispersion. Ordinal variables were modeled as continuous variables to determine linear trend, where indicated. To assess the potential effect of differential trends in incidence by sex or age group, statistical interaction with calendar year was tested; no significant interactions were detected. To provide the most current, clinically relevant estimates, modeling was restricted to 2002–2006, the period during which ICDO-3 was wholly used in Australia, and to subtypes with at least 50 cases during this period. Modeling was performed using STATA Version 11.1 (StataCorp LP, College Station, TX). The level of statistical significance was set at p < 0.05.

Results Of the 128,295 incident lymphoid neoplasms registered in Australia during 1982–2006, 112,747 (88%) were NHL, 9,542 (7%) were HL, 5,999 (5%) were lymphoid neoplasms NOS and 7 were composite HL/NHL. Of the NHL, 18% were DLBCL, 12% follicular lymphoma, 16% chronic lymphocytic leukemia (CLL), 4% small lymphocytic lymphoma (SLL) and 20% were plasma cell neoplasms. Of the HL, 50% were nodular sclerosis and 21% were mixed cellularity/lymphocyte depleted (MC/LD).

Epidemiology

Temporal trends in age-standardized incidence rates

Subtype-specific, age-standardized incidence rates and the AAPC in incidence for 1982–1996 and 1997–2006, as determined by Joinpoint regression, are presented in Table 1 and Figures 1 and 2. Supporting Information Tables 1–4 show the temporal trends by sex and age. The incidence of NHL increased significantly by 2.5%/year (95% CI 5 2.2–2.8%) during 1982–1996, and was stable during 1997–2006 (0.4%, 95% CI 5 20.2–0.9%). For mature B-cell NHL, the incidence of follicular lymphoma increased 4.0%/year (95% CI 5 2.7–5.3%) and 2.5%/ year (95% CI 5 2.1–2.8%) during 1982–1996 and 1997–2006, respectively. Incidence of DLBCL increased 5.5%/year (95% CI 5 4.7–6.4%) during 1982–1996, slowing to 1.3% (95% CI 5 0.4–2.2%) per year during 1997–2006. In the latter period, significantly increasing incidence was observed only for those aged 45–64 and 65 years, having stabilized for those aged 0–44 years. Incidence of plasma cell myeloma increased 1.5%/year (95% CI 5 1.1–1.8%) during 1982–1996, but was stable during 1997–2006. Incidence of plasmacytoma increased 7.1%/year (95% CI 5 5.8–8.4%) during both 1982– 1996 and 1997–2006. The remaining mature B-cell NHL subtypes were analyzed from 1997 to 2006 only. Incidence of marginal zone lymphoma (MZL) increased 13.2%/year (95% CI 5 7.3–19.5%). Increasing incidence was observed for the mucosa-associated lymphoid tissue (MALT), nodal and splenic subtypes. BL incidence increased 6.8%/year (95% CI 5 3.9–9.7%); significantly increasing incidence was observed only for males, and those aged 25–44 and 45–64 years. The incidence of mantle

Lymphoid neoplasms in Australia 1982–2006

cell lymphoma (MCL) increased 4.2%/year (95% CI 5 0.5– 8.1%). Incidence was stable for CLL, lymphoplasmacytic lymphoma (LPL) and hairy cell leukemia. The incidence of SLL decreased significantly by 8.1%/year (95% CI 5 2.5–13.3%); this pattern was observed for both sexes and all age groups. For mature natural killer (NK)-/T-cell NHL, incidence increased 17.1%/year (95% CI 5 2.0–34.5%) during 1997– 2006 for large granular lymphocytic leukemia, based on small numbers (n 5 36). Incidence increased 4.7%/year (95% CI 5 1.6–8.0%) for peripheral T-cell lymphoma (PTCL). Significantly increasing incidence was observed for both PTCL NOS, and other PTCL, although incidence of primary cutaneous T-cell lymphoma was stable. Following a period of increasing incidence during 1982–1996, incidence of MF/SS was stable during 1997–2002. Incidence was stable for adult T-cell leukemia/lymphoma and extranodal NK-/T-cell lymphoma (ENNKTL). Following a period of stability during 1982–1996, incidence of all HL increased 2.2%/year (95% CI 5 1.3–3.1%) during 1997–2006. During 1997–2006, incidence increased 4.3%/year (95% CI 5 0.0–8.8%) for classical HL NOS, and 1.7%/year (95% CI 5 0.0–3.4%) for nodular sclerosis HL; rates of lymphocyte rich and MC/LD HL were stable. Incidence increased 12.1%/year (95% CI 5 5.4–19.3%) for nodular lymphocyte predominant HL. The temporal trends varied by age; all HL incidence increased for those aged 0–14, 15– 24 and 25–44 but not for the older age groups. For nodular sclerosis HL and nodular lymphocyte predominant the increases were observed for the young and old age groups, while lymphocyte rich HL incidence increased only for the older age groups (Supporting Information Table 4). For other lymphoid neoplasm subtypes, the incidence of precursor lymphoid neoplasms was stable during 1997–2006. Incidence of NHL NOS declined 1.9%/year (95% CI 5 0.4– 3.4) during 1982–1996 and 5.8%/year (95% CI 5 4.9–6.7%) during 1997–2006. Incidence of lymphoid neoplasms NOS declined 3.8%/year (95% CI 5 3.3–10.4%) during 1982–1996 and 5.6%/year (95% CI 5 4.1–7.0%) during 1997–2006. Incidence by sex, 2002–2006

For mature B-cell NHL, male predominance was three- to fourfold for hairy cell leukemia, BL and MCL and up to twofold for CLL/SLL, LPL, plasmacytoma, plasma cell myeloma, DLBCL and follicular lymphoma (Supporting Information Tables 5a and b and 6). There was no sex disparity for MZL overall or for the MALT, nodal or splenic subtypes. For mature NK-/T-cell NHL, incidence was fourfold higher for males for ENNKTL and twofold higher for primary cutaneous anaplastic large cell lymphoma (C-ALCL), MF/SS and PTCL, NOS (Supporting Information Table 7). There was no sex disparity for other types of PTCL. A less than twofold male predominance was seen for the classical HL subtypes of lymphocyte-rich, MC/LD and classical HL, NOS (Supporting Information Table 8). There was no sex disparity for nodular sclerosis HL. Incidence of C 2014 UICC Int. J. Cancer: 135, 2146–2156 (2014) V

Table 1. Trends in the incidence of lymphoid neoplasms by subtype, Australia, 1982–1996 and 1997–2006 1982–19962 ASR3

AAPC

95% CI

No.

ASR3

AAPC

95% CI

All lymphoid neoplasms4

63273

22.4

1.8

1.1–2.4

65022

26.1

0.3

20.1–0.8

Non-Hodgkin lymphoma

54284

19.1

2.5

2.2–2.8

58463

23.0

0.4

20.2–0.9

3950

1.7

3235

1.9

0.2

21.5–1.8

39489

13.6

46209

17.7

1.3

0.6–2.1

9823, 9670

11563

3.9

10479

3.8

21.4

23.5–0.8

CLL

9823

8946

3.0

8546

3.1

0.3

21.3–1.9

SLL

9670

2617

0.9

1933

0.7

28.1

213.3 to 22.5

MCL

9673

386

0.1

1175

0.5

4.2

0.5–8.1

LPL

9671, 9761

840

0.3

1331

0.5

20.6

23.7–2.7

101

0.0

1203

0.5

13.2

7.3–19.5

9699 (excl. C77.0-C77.9), 9764

64

0.0

708

0.3

10.2

3.9–16.9

Nodal

9699 (C77.0-C77.9)

28

0.0

363

0.1

15.6

9.8–21.8

Splenic

9689

9

0.0

132

0.1

21.9

7.2–38.7

9940

602

0.2

645

0.3

21.9

25.9–2.2

10638

3.6

1.6

1.3–1.9

11440

4.2

0.0

21.0–1.0 5.8–8.4

Precursor lymphoid neoplasms

ICD-O-3 code

9727-29, 9835-37

Mature B-cell neoplasms5 CLL/SLL

MZL MALT

Hairy cell leukemia Plasma cell neoplasms Plasmacytoma

9731, 9734

236

0.1

7.1

5.8–8.4

450

0.2

7.1

Plasma cell myeloma

9732, 9733

10402

3.5

1.5

1.1–1.8

10990

4.0

20.3

21.4–0.8

9690-91, 9695, 9698

6213

2.2

4.0

2.7–5.3

7566

3.1

2.5

2.1–2.8

5.5

4.7–6.4

Follicular lymphoma Diffuse large B-cell lymphoma

9678-80

8850

3.1

11742

4.5

1.3

0.4–2.2

Burkitt lymphoma/leukemia

9687, 9826

294

0.1

573

0.3

6.8

3.9–9.7

1202

0.4

2483

1.0

3.6

1.4–5.9 2.0–34.5

Mature T/NK-cell neoplasms6 Large granular lymphocytic leukemia

9831

0

0.0

36

0.0

17.1

Adult T-cell leukemia/lymphoma

9827

35

0.0

45

0.0

3.3

29.6–18.0

ENNKTL

9719

19

0.0

92

0.0

5.2

24.7–16.0

543

0.2

1542

0.6

4.7

1.6–8.0

Peripheral T-cell lymphoma, NOS

9702

389

0.1

675

0.3

4.1

1.4–6.8

PCTCL

9709

74

0.0

359

0.1

23.0

29.8–4.3

Peripheral T-cell lymphoma, other

9705, 9708, 9714, 9716-17

80

0.0

508

0.2

11.2

5.8–16.8

9700-01

596

0.2

3.4

1.9–5.0

672

0.3

22.2

26.0–1.8

9591, 9675, 9684, 9832

9643

3.4

1.9

0.4–3.4

6536

2.4

25.8

26.7 to 24.9

5081

2.0

20.1

20.9–0.7

5020

2.0

Peripheral T-cell lymphoma

MF/SS Non-Hodgkin lymphoma, NOS Hodgkin lymphoma Classical

4461

2.3

2.2

1.3–3.1

4287

2.2

1.6

0.4–2.9

Lymphocyte rich

9651

394

0.2

246

0.1

1.5

214.5–20.4

Mixed cellularity/lymphocyte depleted

9652–9655

1276

0.5

754

0.4

22.4

25.5–0.8

Nodular sclerosis

9663–9667

2226

0.9

2579

1.4

1.7

0.0–3.4

Classical, NOS

9650, 9661–9662

1124

0.4

708

0.3

4.3

0.0–8.8

9659

61

0.0

174

0.1

12.1

5.4–19.3

9590, 9820

3908

1.3

2091

0.7

25.6

27.0 to 24.1

Nodular lymphocyte predominant Lymphoid neoplasm, NOS

23.8

210.4–3.3

Abbreviations: CLL/SLL: chronic lymphocytic leukemia/small cell lymphocytic lymphoma; MALT lymphoma: marginal zone lymphoma of mucosaassociated lymphoid tissue lymphoma; DLBCL: diffuse large B-cell lymphoma; NK-cell: natural killer cell; ENNKTL: extranodal NK-/T-cell lymphoma; zary syndrome; MC/LD: mixed cellularity/ NOS: not otherwise specified; PCTCL: primary cutaneous T-cell lymphoma; MF/SS: mycosis fungoides/Se lymphocyte-depleted; ICD-O-3: International Classification of Diseases for Oncology, Third Edition; ASR: age-standardised rate; AAPC: average annual percent change; CI: confidence interval. 1 AAPC could not be computed for some subtypes due to insufficient cases in some strata including B-cell prolymphocytic leukemia (9833; 1982– 1996 n 5 2, 1997–2006 n 5 21); heavy chain disease (9762 1982–1996 n 5 0, 1997–2006 n 5 5); immunoproliferative disease (9760; 1982–1996 n 5 0, 1997–2006 n 5 29); T-cell prolymphocytic leukemia (9834;1982–1996 n 5 7, 1997–2006 n 5 25); aggressive NK-cell leukemia (9948; 1982– 1996 n 5 2, 1997–2006 n 5 6); primary cutaneous anaplastic large-cell lymphoma (9718; 1982–1996 n 5 0, 1997–2006 n 5 65); composite Hodgkin/non-Hodgkin lymphoma (9596; 1982–1996 n 5 0, 1997–2006 n 5 7) and HL subtypes 1982–1996. 2 AAPC computed only for subtypes with sufficient diagnostic reliability during 1982–1996. 3 Age-standardised to World (2000) population and expressed per 100,000 person-years. 4 All lymphoid neoplasms include non-Hodgkin lymphoma, Hodgkin lymphoma, lymphoid neoplasm NOS and composite Hodgkin/non-Hodgkin lymphoma. 5 Mature B-cell neoplasms include B-cell prolymphocytic leukemia, heavy chain disease and immunoproliferative disease. 6 Mature T/NK-cell neoplasms include T-cell prolymphocytic leukemia, aggressive NK-cell leukemia and primary cutaneous anaplastic large-cell lymphoma.

Epidemiology

1997–2006

No.

Lymphoid neoplasm subtype

1

2150

Lymphoid neoplasms in Australia 1982–2006

Epidemiology

Figure 1. Trends in the incidence of lymphoid neoplasms by subtype and sex, Australia, 1982–2006. *All incidence rates are standardised to the World (2000) population. Squares denote yearly age-standardized incidence rates (ASR), with trends defined by Joinpoint analysis. Abbreviations are defined in Table 1.

nodular lymphocyte predominant HL was threefold higher for males. For other lymphoid neoplasm subtypes, incidence was modestly higher for males for precursor lymphoid neoplasms, NHL NOS and lymphoid neoplasms NOS (Supporting Information Table 9). Incidence by age, 2002–2006

Incidence increased sharply with age for men and women for all subtypes of mature B- and NK-/T-cell NHL, with the exception of BL, where an increase with age was only observed for women (Figs. 3a–3k). This was also reflected in multiple Poisson regression modeling (Supporting Information Tables 5a and b, 6 and 7); increasing incidence was observed with increasing age for all subtypes (p 5 0.014 for BL; p < 0.001 for other subtypes), although there was variation in the magnitude of effect. The greatest increases in incidence with age were seen for CLL/SLL, MCL, LPL and plasma cell myeloma, for which incidence was typically 100fold higher for those aged 65 years or more compared with those aged 0–44 years; there were almost no cases recorded

in children for these subtypes. More modest increases in incidence with age, generally 10–40-fold for those aged 65 years compared with those aged 0–44 years, were observed for MZL, hairy cell leukemia, follicular lymphoma, DLBCL, plasmacytoma, MF/SS, PTCL and C-ALCL. For BL, only a twofold increase was seen for those aged 65 years compared with those aged 0–14 years. For HL subtypes, we observed a significant trend with increasing age (p < 0.001) for lymphocyte-rich, MC/LD and classical HL NOS (Figs. 3o23s; Supporting Information Table 8), although the change in magnitude was slight in comparison with that seen for mature B- and NK-/T-cell NHL. Nodular sclerosis HL showed a peak in incidence for those aged 15–24 years. There was little variation in age for nodular lymphocyte predominant HL. For other lymphoid neoplasm subtypes, incidence of precursor lymphoid neoplasms was highest for children (Fig. 3l; Supporting Information Table 7). Incidence of NHL NOS and lymphoid neoplasms NOS (Figs. 3m23n; Supporting Information Table 9) was 30- to 50-fold higher for those aged 65 years compared with those aged 0–44 years. C 2014 UICC Int. J. Cancer: 135, 2146–2156 (2014) V

C 2014 UICC Int. J. Cancer: 135, 2146–2156 (2014) V

Epidemiology

Figure 2. Trends in the incidence of lymphoid neoplasms by subtype and sex, Australia, 1997–2006. *All incidence rates are standardised to the World (2000) population. Squares denote yearly age-standardized incidence rates (ASR), with trends defined by Joinpoint analysis. Abbreviations are defined in Table 1.

van Leeuwen et al.

2151

Figure 3. Incidence of lymphoid neoplasms by subtype, sex and age, Australia, 2002–2006. *All incidence rates are standardized within age groups to the World (2000) population. Abbreviations are defined in Table 1.

Epidemiology

2152 Lymphoid neoplasms in Australia 1982–2006

C 2014 UICC Int. J. Cancer: 135, 2146–2156 (2014) V

Discussion In this national, population-based study of lymphoid neoplasms registered in Australia during 1982–2006, marked variation was observed in subtype-specific incidence patterns over time and by sex and age. Despite stabilization in the incidence of NHL overall, the incidence of many subtypes of mature B-cell and NK-/T-cell NHL increased in the last 10 years, including DLBCL, follicular, marginal zone, mantle cell, Burkitt and peripheral T-cell lymphoma and plasmacytoma. The incidence of HL also increased during this period. Decreasing incidence of the NOS subtypes supports improved diagnostic specificity over time. The introduction and increasing use of ancillary diagnostic tests may have contributed to the temporal increases in incidence for some subtypes. Most lymphoid neoplasms show a strong male predominance and sharp increases in incidence with increasing age. Comparison of incidence patterns reinforces growing epidemiological and biological evidence for subtype-specific variation in etiology, and informs clinicians, patients and agencies providing health services. The trends in overall NHL incidence observed in Australia since the 1980s are broadly consistent with those reported for other Western countries. Incidence increased by 2.5% annually during 1982–1996, compared with 3–4% for Europe7,14,15 and the United States16,17 for an NHL rubric that excluded plasma cell neoplasms and lymphoid leukemias. Reasons for this broad-based NHL epidemic remain unknown. While improvements in diagnosis, disease classification and cancer registration, together with increasing numbers of people with human immunodeficiency virus (HIV) infection, may have been contributing factors,18,19 they are not wholly responsible.5,19,20 The subsequent stabilization in incidence in Australia during 1997– 2006 was also seen in parts of Europe from the mid-1990s.15,21 Significant declines in incidence were observed in the United States in the late 1990s for males aged 25–54, although incidence continued to increase at 1–2%/year for middle-aged and older adults.16 When HIV-related neoplasms are excluded from the United States data, NHL incidence increased throughout 1992–2009 for 15–39 year olds and stabilized in the mid-2000s for older age groups.5 In Australia the stabilization in NHL incidence since 1997 was observed for men and women and for all age groups. Much of the stabilization appeared to be driven by plateaus in the incidence of CLL/SLL, plasma cell myeloma and a slowing in the increase in incidence of DLBCL. Together, these neoplasms comprised more than half of all NHL cases registered in Australia during this period. Reasons for the downtrend in CLL/SLL incidence, also seen in the United States,3 are not known, but our data appear to support a progressive reclassification of SLL as other types of B-cell lymphoma, especially MZL and MCL, over the period as pathologists became more familiar with newer subtypes. In Australia the incidence of DLBCL increased 1.3%/year over the last 10 years. The increase was greatly attenuated compared to that seen in 1982–1996 (5.5%/year) and was

C 2014 UICC Int. J. Cancer: 135, 2146–2156 (2014) V

2153

observed in men, women and those aged at least 45 years. In the United States during 1992–2001, overall rates of DLBCL declined over time, predominantly in males 25–54 years of age, but increased in the elderly.3 When HIV infection is taken into account, overall rates of DLBCL in the United States appear to have stabilized since 2003.5 In Australia the incidence of NHL in those with HIV halved following the introduction of highly active antiretroviral therapy (HAART) in the mid-1990s.22 Nevertheless, the contribution by HIVrelated DLBCL to temporal trends in Australia is expected to be inconsequential on account of the low incidence. Indeed, a prior study estimated that 165 cases of HIV-related DLBCL were diagnosed during 1996–2004,22 which is less than 1.5% of the total DLBCL. Diagnostic reliability for DLBCL has remained high over time,11 but there has probably been increased diagnostic accuracy of “gray zone” cases between DLBCL and BL23 and increasing classification of “double-hit” lymphomas as DLBCL instead of BL.24 Reasons for the sustained increase in follicular lymphoma incidence in Australia are also unclear. Follicular lymphoma incidence increased in the United States until the mid-1990s,2 but stabilized thereafter.3,5 Diagnostic reliability for follicular lymphoma has remained consistently high, even before the widespread use of immunophenotyping.11 The etiology of follicular lymphoma is not well understood, but it is unrelated to immune deficiency.5 Increased risk has been associated with tobacco smoking,25 family history,26 and genetic susceptibility.27 Consistent with incidence trends in the United States,3 the increasing incidence of MZL likely reflects the implementation of disease-specific ICD-O-2 codes, and diagnosis as lymphoma of cases that would previously have been diagnosed as lymphoid hyperplasia, making true increases in incidence difficult to establish. MALT lymphoma, comprising more than half of all MZL cases in our series, is causally associated with chronic inflammation arising from Helicobacter pylori infection and possibly other infectious agents.28 However, the seroprevalence of H. pylori infection has decreased in Western populations over the last several decades.29,30 Furthermore, in Australia the seroprevalence of H. pylori infection is higher for males than females,31 but a male excess was not observed for MZLs. MALT lymphoma is associated with history of autoimmune disease, particularly Sj€ogren’s syndrome, Hashimoto thyroiditis and systemic lupus erythematosus,10,32,33 thus the gender balance may be related to the higher prevalence of certain autoimmune diseases, including lupus, among females.34 An increasing incidence of MCL, also seen in the United States,3,35,36 likely reflects its recognition as a distinct disease entity in 1992 and the introduction of immunohistochemical staining for cyclin D1, overexpressed in almost all cases, and cytogenetic testing. Potential true increases in incidence are therefore difficult to determine, especially given that the etiology of this malignancy is unknown.37 Reasons for the increasing incidence of BL, seen both in Australia and the United States,3 are uncertain. Increasing

Epidemiology

van Leeuwen et al.

Epidemiology

2154

incidence in Australia was confined to males, and to those aged 25–44 and 45–64 years. This profile largely mirrors the Australian population living with HIV, and acquired immunodeficiency syndrome (AIDS)-related BL has not declined in incidence in Australia since the introduction of HAART.22 A contribution by AIDS-related BLs to temporal incidence trends is likely,5 but the magnitude is unknown. The impact of changes in the pathological diagnosis of BL in the 2001 WHO classification, with broader morphological criteria but narrower immunophenotypic and cytogenetic criteria, is difficult to assess. Wider availability of FISH probes for c-myc translocations may have increased the identification of BL cases with DLBCL morphology.23,24 We are therefore unable to distinguish a true increase from diagnostic artifact. As noted previously, the age profile for BL is younger than that for all other lymphoid neoplasms. Recent data from the United States show peaks in rates around 10 and 75 years, and a third peak for males near age 40 years,38 suggesting differing etiologies. This pattern was evident to some extent in our data, although could not be confirmed in multiple Poisson regression modeling due to the comparatively small number of cases (n 5 342). The increasing incidence of mature NK-/T-cell NHL and PTCL in Australia is consistent with United States’ data, and the cause is unknown.3,5 Diagnostic artifact cannot be excluded, particularly for large granular lymphocytic leukemia, which can now be readily diagnosed using peripheral blood flow cytometry. In agreement with recent trends in the United States,39,40 the incidence of cutaneous T-cell lymphoma was stable, although the incidence of other specified peripheral T-cell lymphoma increased markedly over time. In contrast to the past3,5 but not most recent40 United States experience, we observed stabilization in incidence of MF/SS, although increases were observed in females and those aged 65 years or more. Classification changes resulting in shifts in the distribution of cutaneous T-cell lymphoma subtypes may have contributed to this pattern.41 The incidence of HL in Australia was stable during 1982– 1996, a finding broadly consistent with stable or declining HL incidence observed in the United Kingdom (1984– 1993),42 Nordic countries (1978–97),43 Europe (1983– 1997;1985–1992),7,14 and the United States (1969–1984; 1992–2001).3,44 In the subsequent 10-year period, 1997–2006, HL incidence increased in Australia. Increases were observed for all HL (2%/year), and the classical (2%/year), nodular sclerosis (4%/year) and nodular lymphocyte predominant (12%/year) subtypes. Nodular sclerosis HL increased in incidence for young adults, as observed during prior eras in Nordic countries and the United States,43,45 but also for those aged 65 years or more. Other novel findings were increases in rates of lymphocyte-rich classical HL for those aged 45 years or more, and nodular lymphocyte predominant HL for young adults and those aged 45–64 years. While the temporal trend for all HL is likely to be robust to improvements in diagnosis over time, this cannot be assured for the sub-

Lymphoid neoplasms in Australia 1982–2006

types,44 which could not be analyzed for 1982–1996 because of zero cells for some strata. The only HL subtype to exhibit a bimodal pattern, with peaks in young adulthood and older age, was nodular sclerosis HL. This neoplasm was also the only HL subtype not to exhibit a male predominance. HL diagnosed in childhood and older age is more likely to be positive for infection with Epstein–Barr virus (EBV).46 Evidence is emerging of the role of genetic variants in the major histocompatibility complex region in EBV-positive and EBV-negative classical HL and its subtypes.47 In contrast to 1982–1996, we observed large declines in the incidence of lymphoid neoplasms that were NOS during 1997–2006. During this period, NHL NOS and lymphoid neoplasms NOS together comprised 13% of all lymphoid neoplasms registered in Australia, a figure similar to that seen in the United States during 1992–2001.3 The concentration of cases in those aged 65 years (66% in the last 5years) potentially highlights age-related disparities in diagnostic practices or reluctance to pursue invasive procedures necessary for complete classification in patients in whom illness limits treatment options. The striking male preponderance seen for most, but not all, lymphoid neoplasm subtypes is consistent with other published registry-based data from the United States2,3 and Europe.4 The greatest male-to-female ratios, in excess of twofold, were observed for BL, hairy cell leukemia and MCL. Differential exposure to causative agents may account for some of the sex imbalance; as yet there are no established hormonal or reproductive risk factors.48 The increasing incidence of mature B- and NK-/T-cell NHL with age is well documented.2–4 The greatest age-related increases were observed for CLL/SLL, MCL, LPL and plasma cell myeloma, with virtually no cases in children. This large, national, population-based study systematically examined temporal trends in the subtype-specific incidence of lymphoid neoplasms in Australia, as well as incidence patterns by sex and age. This has not been performed previously for Australia, and complements analyses published for the United States2,3 and Europe.4 As recommended for epidemiological research, subtypes were defined according to the InterLymph nested hierarchical classification, based on the 2008 WHO Classification. Concordance between ICD-O-3, which incorporates the WHO Classification, and cases originally diagnosed and coded under previous systems is generally high, particularly when grouped into subtypes.8,11 However, errors of coding translation to ICD-O-3 are possible, particularly during the earlier time period (1982–1996), as evidenced by 394 cases of lymphocyte rich HL during this period. Nevertheless, improved diagnostic specificity, as evidenced by the ongoing decline in incidence of unspecified lymphoid neoplasms, will have contributed to the temporal incidence trends observed for some subtypes. Additionally, misclassification is inevitable, particularly in the absence of central pathology review, with which the National Cancer C 2014 UICC Int. J. Cancer: 135, 2146–2156 (2014) V

2155

van Leeuwen et al.

Comprehensive Network reclassified 6% of common B-cell NHLs diagnosed in the WHO Classification era.49 A further limitation is that neoplasms could not be examined by HIV infection or tumor-EBV status, characteristics that are not routinely tested nor recorded by Australian cancer registries. We were also unable to examine diversity within subtypes, such as for DLBCL, nor new entities recently incorporated into the WHO Classification, including immunodeficiency-associated lymphoproliferative disorders, in situ lesions or tumors with composite subtypes.50 Finally, race, strongly associated with both NHL and HL incidence, is not uniformly collected by Australian cancer registries and could not be included in our analyses. While the Australian population is predominantly Caucasian, and a minority (2%) Aboriginal, the proportion of individuals with Asian ancestry has increased over time to 6% in 2006. While this makes comparisons with rates stratified by race challenging, it is not expected to have a large impact on the overall figures. Overall, the incidence of NHL has stabilized in Australia, although the incidence of some subtypes may be increasing. The incidence of HL, particularly classical HL, also appears to be increasing for young adults. Changes in diagnostic

practice and classification will continue to limit the interpretation of incidence trends for some subtypes. This should decrease in future as pathologists become more familiar with the WHO classification, but the lack of reimbursement for out-of-pocket expenses such as FISH probes for example will continue to be an issue. Use of the WHO classification, for all subtypes of lymphoid neoplasms, is encouraged. Reasons for the sustained increase in incidence of follicular lymphoma, DLBCL and HL, less likely to have been influenced by diagnostic artifact, are unclear. Marked subtype-specific differences in incidence patterns by sex and age were also observed. These descriptive observations provide further population-level evidence for risk factor similarities and differences among lymphoid neoplasm subtypes. These data can also be used to forecast the future subtype-specific public health burden.

Acknowledgements We thank the Australian state and territory cancer registries for the use of their data, and the Australian Institute of Health and Welfare for cancer data extraction. We thank the Australian Bureau of Statistics for the provision of population census data.

1.

2.

3.

4.

5.

6.

7.

8.

9.

Australian Institute of Health and Welfare (AIHW) and Australasian Association of Cancer Registries. Cancer in Australia: an overview, 2010. Canberra: AIHW, 2010. Groves FD, Linet MS, Travis LB, et al. Cancer Surveillance Series: non-Hodgkin’s lymphoma incidence by histologic subtype in the United States from 1978 through 1995. J Natl Cancer Inst 2000;92:1240–51. Morton LM, Wang SS, Devesa SS, et al. Lymphoma incidence patterns by WHO subtype in the United States, 1992–2001. Blood 2006;107: 265–76. Sant M, Allemani C, Tereanu C, et al. Incidence of hematologic malignancies in Europe by morphologic subtype: results of the HAEMACARE project. Blood 2010;116:3724–34. Shiels MS, Engels EA, Linet MS, et al. The epidemic of non-Hodgkin lymphoma in the United States: disentangling the effect of HIV, 1992– 2009. Cancer Epidemiol Biomarkers Prev 2013;22: 1069–78. Morton LM, Turner JJ, Cerhan JR, et al. Proposed classification of lymphoid neoplasms for epidemiologic research from the Pathology Working Group of the International Lymphoma Epidemiology Consortium (InterLymph). Blood 2007;110:695–708. Cartwright R, Brincker H, Carli PM, et al. The rise in incidence of lymphomas in Europe 1985– 1992. Eur J Cancer 1999;35:627–33. Clarke CA, Undurraga DM, Harasty PJ, et al. Changes in cancer registry coding for lymphoma subtypes: reliability over time and relevance for surveillance and study. Cancer Epidemiol Biomarkers Prev 2006;15:630–8. Jaffe ES, Harris NL, Stein H, eds. WHO classification of tumours. Pathology and genetics of tumours of haematopoietic and lymphoid tissues. Lyon, France: IARC, 2001.

10. Swerdlow SH, Campo E, Harris NL, et al., eds. WHO classification of tumours of haematopoietic and lymphoid tissues. Lyon, France: IARC, 2008. 11. Turner JJ, Morton LM, Linet MS, et al. InterLymph hierarchical classification of lymphoid neoplasms for epidemiologic research based on the WHO classification (2008): update and future directions. Blood 2010;116:e90–8. 12. Kim HJ, Fay MP, Feuer EJ, et al. Permutation tests for joinpoint regression with applications to cancer rates. Stat Med 2000;19:335–51. 13. Clegg LX, Hankey BF, Tiwari R, et al. Estimating average annual per cent change in trend analysis. Stat Med 2009;28:3670–82. 14. Adamson P, Bray F, Costantini AS, et al. Time trends in the registration of Hodgkin and nonHodgkin lymphomas in Europe. Eur J Cancer 2007;43: 391–401. 15. Sandin S, Hjalgrim H, Glimelius B, et al. Incidence of non-Hodgkin’s lymphoma in Sweden, Denmark, and Finland from 1960 through 2003: an epidemic that was. Cancer Epidemiol Biomarkers Prev 2006;15:1295–300. 16. Clarke CA, Glaser SL. Changing incidence of non-Hodgkin lymphomas in the United States. Cancer 2002;94:2015–23. 17. Howe HL, Wingo PA, Thun MJ, et al. Annual report to the nation on the status of cancer (1973 through 1998), featuring cancers with recent increasing trends. J Natl Cancer Inst 2001;93: 824– 42. 18. Zheng T, Mayne ST, Boyle P, et al. Epidemiology of non-Hodgkin lymphoma in connecticut 1935– 1988. Cancer 1992;70:840–9. 19. Hartge P, Devesa SS. Quantification of the impact of known risk factors on time trends in nonHodgkin’s lymphoma incidence. Cancer Res 1992; 52:5566s–9s. 20. Holford TR, Zheng T, Mayne ST, et al. Time trends of non-Hodgkin’s lymphoma: are they

C 2014 UICC Int. J. Cancer: 135, 2146–2156 (2014) V

21.

22.

23.

24.

25.

26.

27.

28.

29.

real? What do they mean? Cancer Res 1992;52: 5443s–6s. Bosetti C, Levi F, Ferlay J, et al. Incidence and mortality from non-Hodgkin lymphoma in Europe: the end of an epidemic? Int J Cancer 2008;123:1917–23. van Leeuwen MT, Vajdic CM, Middleton MG, et al. Continuing declines in some but not all HIV-associated cancers in Australia after widespread use of antiretroviral therapy. AIDS 2009; 23:2183–90. Dave SS, Fu K, Wright GW, et al. Molecular diagnosis of Burkitt’s lymphoma. N Engl J Med 2006;354:2431–42. Aukema SM, Siebert R, Schuuring E, et al. Double-hit B-cell lymphomas. Blood 2011;117: 2319–31. Morton LM, Hartge P, Holford TR, et al. Cigarette smoking and risk of non-Hodgkin lymphoma: a pooled analysis from the International Lymphoma Epidemiology Consortium (InterLymph). Cancer Epidemiol Biomarkers Prev 2005; 14:925–33. Wang SS, Slager SL, Brennan P, et al. Family history of hematopoietic malignancies and risk of non-Hodgkin lymphoma (NHL): a pooled analysis of 10 211 cases and 11 905 controls from the International Lymphoma Epidemiology Consortium (InterLymph). Blood 2007;109: 3479–88. Conde L, Halperin E, Akers NK, et al. Genomewide association study of follicular lymphoma identifies a risk locus at 6p21.32. Nat Genet 2010; 42:661–4. Bouvard V, Baan R, Straif K, et al. A review of human carcinogens—part B: biological agents. Lancet Oncol 2009;10:321–2. Roosendaal R, Kuipers EJ, Buitenwerf J, et al. Helicobacter pylori and the birth cohort effect: evidence of a continuous decrease of infection

Epidemiology

References

2156

30.

31.

32.

33.

34.

35.

37. Smedby KE, Hjalgrim H. Epidemiology and etiology of mantle cell lymphoma and other nonHodgkin lymphoma subtypes. Semin Cancer Biol 2011;21:293–8. 38. Mbulaiteye SM, Anderson WF, Ferlay J, et al. Pediatric, elderly, and emerging adult-onset peaks in Burkitt’s lymphoma incidence diagnosed in four continents, excluding Africa. Am J Hematol 2012;87:573–8. 39. Bradford PT, Devesa SS, Anderson WF, et al. Cutaneous lymphoma incidence patterns in the United States: a population-based study of 3884 cases. Blood 2009;113:5064–73. 40. Korgavkar K, Xiong M, Weinstock M. Changing incidence trends of cutaneous T-cell lymphoma. JAMA Dermatol 2013;149:1295–9. 41. Criscione VD, Weinstock MA. Incidence of cutaneous T-cell lymphoma in the United States, 1973–2002. Arch Dermatol 2007;143:854–9. 42. Cartwright R, McNally R, Roman E, et al. Incidence and time trends in Hodgkin’s disease: from parts of the United Kingdom (1984–1993). Leuk Lymphoma 1998;31:367–77. 43. Hjalgrim H, Askling J, Pukkala E, et al. Incidence of Hodgkin’s disease in Nordic countries. Lancet 2001;358:297–8. 44. Glaser SL, Dorfman RF, Clarke CA. Expert review of the diagnosis and histologic classification of Hodgkin disease in a population-based cancer registry. Cancer 2001;92:218–24.

45. Glaser SL, Swartz WG. Time trends in Hodgkin’s disease incidence. The role of diagnostic accuracy. Cancer 1990;66:2196–204. 46. Glaser SL, Lin RJ, Stewart SL, et al. Epstein–Barr virus-associated Hodgkin’s disease: epidemiologic characteristics in international data. Int J Cancer 1997;70:375–82. 47. Urayama KY, Jarrett RF, Hjalgrim H, et al. Genome-wide association study of classical Hodgkin lymphoma and Epstein–Barr virus statusdefined subgroups. J Natl Cancer Inst 2012;104: 240–53. 48. Kane EV, Roman E, Becker N, et al. Menstrual and reproductive factors, and hormonal contraception use: associations with non-Hodgkin lymphoma in a pooled analysis of InterLymph case– control studies. Ann Oncol 2012;23:2362–74. 49. LaCasce AS, Kho ME, Friedberg JW, et al. Comparison of referring and final pathology for patients with non-Hodgkin’s lymphoma in the National Comprehensive Cancer Network. J Clin Oncol 2008;26:5107–12. 50. Campo E, Swerdlow SH, Harris NL, et al. The 2008 WHO classification of lymphoid neoplasms and beyond: evolving concepts and practical applications. Blood 2011;117:5019–32.

Epidemiology

36.

rates in childhood. Am J Gastroenterol 1997;92: 1480–2. Banatvala N, Mayo K, Megraud F, et al. The cohort effect and Helicobacter pylori. J Inf Dis 1993;168:219–21. Lin SK, Lambert JR, Nicholson L, et al. Prevalence of Helicobacter pylori in a representative Anglo-Celtic population of urban Melbourne. J Gastroenterol Hepatol 1998;13:505–10. Smedby KE, Vajdic CM, Falster M, et al. Autoimmune disorders and risk of non-Hodgkin lymphoma subtypes: a pooled analysis within the InterLymph Consortium. Blood 2008;111: 4029–38. Troch M, Woehrer S, Streubel B, et al. Chronic autoimmune thyroiditis (Hashimoto’s thyroiditis) in patients with MALT lymphoma. Ann Oncol 2008;19:1336–9. Jacobson DL, Gange SJ, Rose NR, et al. Epidemiology and estimated population burden of selected autoimmune diseases in the United States. Clin Immunol Immunopathol 1997;84:223– 43. Zhou Y, Wang H, Fang W, et al. Incidence trends of mantle cell lymphoma in the United States between 1992 and 2004. Cancer 2008;113:791–8. Aschebrook-Kilfoy B, Caces DB, Ollberding NJ, et al. An upward trend in the age-specific incidence patterns for mantle cell lymphoma in the USA. Leuk Lymphoma 2013;54:1677–83.

Lymphoid neoplasms in Australia 1982–2006

C 2014 UICC Int. J. Cancer: 135, 2146–2156 (2014) V