The first 2 authors contributed equally to this work.
Incidence trends of mantle cell lymphoma in the United States between 1992 and 2004
Article first published online: 9 JUL 2008
Copyright © 2008 American Cancer Society
Volume 113, Issue 4, pages 791–798, 15 August 2008
How to Cite
Zhou, Y., Wang, H., Fang, W., Romaguer, J. E., Zhang, Y., Delasalle, K. B., Kwak, L., Yi, Q., Du, X. L. and Wang, M. (2008), Incidence trends of mantle cell lymphoma in the United States between 1992 and 2004. Cancer, 113: 791–798. doi: 10.1002/cncr.23608
- Issue published online: 1 AUG 2008
- Article first published online: 9 JUL 2008
- Manuscript Accepted: 7 APR 2008
- Manuscript Revised: 2 APR 2008
- Manuscript Received: 28 FEB 2008
- Agency for Healthcare Research and Quality. Grant Number: R01-HS016743
- mantle cell lymphoma;
- non-Hodgkin's lymphoma;
- geographic variations;
- SEER Program;
- population-based cancer registries
Mantle cell lymphoma (MCL) is a distinct subtype of B-cell non-Hodgkin's lymphoma. To the authors' knowledge, little is known regarding its incidence patterns and associated factors. The purpose of the current study was to examine the incidence of MCL over a period of 13 years and to identify the factors associated with the incidence patterns.
Patients diagnosed with MCL between 1992 and 2004 were identified from the Surveillance, Epidemiology, and End Results (SEER) Tumor registries. SEER*Stat statistical software was used for analysis.
Of the 87,166 patients diagnosed with non-Hodgkin's lymphoma during the 13-year period between 1992 and 2004, 2459 (2.8%) had confirmed MCL. The overall incidence of MCL (per 100,000) was 0.55, which increased with age: 0.07 in patients aged <50 years, 2.97 in patients aged 70 to 79 years, and 2.78 in those aged ≥80 years. The age-adjusted incidence rate increased from 0.27 of 100,000 in 1992 to 0.69 of 100,000 in 2004, and the annual percent change was 5.87% (P < .05). The median age at diagnosis was 68 years. The incidence of MCL was higher in men (0.84 of 100,000) than in women (0.34 of 100,000) (P < .05), and was higher in Caucasians (0.61 of 100,000) than in African Americans (0.32 of 100,000). Late-stage (III-IV) MCL was diagnosed in 74.6% of patients. There were significant geographic variations noted (P < .05).
The incidence of MCL increased from 1992 to 2004, and was significantly higher in men, in Caucasians, and patients aged ≥50 years. Most patients were diagnosed with late-stage MCL, and there also were considerable geographic variations observed in incidence rate. Cancer 2008. © 2008 American Cancer Society.
The incidence of non-Hodgkin's lymphoma (NHL) has been increasing steadily over the past several decades. At present, it is the fifth most common cancer in the US, with an estimated 63,190 new cases and 15,280 deaths reported in 2007.1 Mantle cell lymphoma (MCL) is an uncommon but distinct subtype of NHL, with little known about its incidence in the US. The most commonly cited data, which suggest that MCL accounts for 6% of all NHLs, were based on a re‒evaluation of 1378 cases of NHL diagnosed between January 1, 1988 and December 31, 1990 at 9 study sites around the world.2 The International Lymphoma Study Group performed the analysis after introducing the Revised European-American Lymphoma (REAL) classification system in 1994.3 Other MCL studies conducted outside the US have been limited by small numbers of cases, not being population-based, or covering only a short period of time.4-16 Furthermore, data from these studies have provided inconsistent results on the percentages of patients with MCL among those with NHL, with wide variations across geographic region.4–16 To our knowledge, no comprehensive studies or descriptions of the incidence of MCL in the US have been published to date.
An epidemiologic investigation of MCL requires access to accurate classification. However, several classification systems for lymphoma have evolved during the past several decades. Initially, the classification system used was based primarily on morphology. Specifically, the Rappaport classification system described MCL as a diffuse or vaguely nodular low-grade lymphoma of intermediate differentiation.17 In the 1980s, this entity was termed centrocytic NHL by the Kiel classification system,18, 19 or was called lymphocytic lymphoma of intermediate differentiation in the American nomenclature.20 MCL was then categorized as diffuse small-cleaved cell lymphoma by the Working Formulation system in 1982.21 In the 1990s, an increased understanding of the immune system and of the genetic abnormalities associated with NHL led to the identification of several previously unrecognized types of lymphoma. In 1992, Banks et al22 coined the term mantle cell lymphoma, establishing MCL as a distinct type of lymphoma. In 1994, the International Lymphoma Study Group proposed the REAL classification system,3, 23 followed in 2000 by the World Health Organization (WHO) classification system.24 The REAL and WHO systems made it clear that MCL is a B-cell NHL with distinct clinicopathologic characteristics. MCL is derived from CD5-positive B cells within the mantle zone with t(11;14)(q13;23) translocation involving the bcl-1 oncogene rearrangement and resulting in cyclin D1 overexpression.
The above various classification systems used since the 1970s have prevented consistent diagnoses. In addition, MCL was not established as a distinct type of lymphoma until 1992. Therefore, our analysis considers the incidence of all MCL diagnosed during the 13-year study period beginning January 1, 1992, when data were first combined from all 13 registries, and ending December 31, 2004, the most recent date for which complete data are available. This analysis used data from the November 2006 Surveillance, Epidemiology, and End Results (SEER) data submission, released in April 2007.25
MATERIALS AND METHODS
Data Sources and Descriptions
The SEER Program has compiled incidence data since 1973 from population-based cancer registries. Each SEER registry includes codes for all cancers, including NHL, entered according to a standard classification scheme based on diagnostic pathology reports in medical records.26 For our analysis, we selected only the first matching record for each SEER MCL case. The International Classification of Diseases for Oncology (ICD-O)-3 code used for MCL is 9673; the codes for diffuse large B-cell lymphoma are 9680 (excluding site C49.9) and 9684 (B); the codes for follicular lymphoma are 9690, 9691, 9695, and 9698.26
Incidence is defined as the number of new cancers of a specific site or type occurring in a specified population during a year, usually expressed as the number of cancers per 100,000 persons at risk. The number of new cancers may include multiple primary cancers occurring in 1 patient. In general, the incidence rate does not include recurrences.25
We examined the data from all patients with MCL diagnosed between 1992 and 2004 in 13 geographic areas: Atlanta, Connecticut, Detroit, Hawaii, Iowa, New Mexico, San Francisco–Oakland, Seattle–Puget Sound, Utah, Los Angeles, San Jose–Monterey, rural Georgia, and among the Alaskan Natives. The SEER 13-Registry database covers 14% of the US population, with meticulous and consistent data collection and standards. The annual incidence rates were age-adjusted for US population in the year 2000 by the direct standardization method. An age-adjusted incidence is a weighted average of the age-specific incidence, in which the weights are the proportions of persons in the corresponding age groups of a standard population. When comparing rates across time (or countries), age-adjusted rates, using the same standard population, are generally used to avoid the potential confounding effect of different age distributions. Annual percent change is used to measure trends or the change in rate over time. It is the average annual rate of change over the time series selected.
Because of the wide variation in incidence rates of MCL resulting from the limited number of cases, we used 2-year age-adjusted cumulative incidence rates to reflect the current incidence and the incidence over time. Cumulative incidence is the number of new cases of disease occurring over a specified period of time in a population at risk at the beginning of the interval. Incidence and survival were calculated using the 2007 SEER*Stat client-server program developed by Information Management Services Inc. (Silver Spring, Md).27
The categories we used for disease stage at the time of diagnosis were stages I, II, III, and IV, and unstaged disease. Stage I included stage I (SEER code 10) and stage IE (code 11). Stage II included stage II (code 20) and stage IIE (code 21). Stage III included stage III (code 30), stage IIIE (code 31), stage IIIS (code 32), and stage IIIES (stage IIIE plus stage IIIS [code 33]). Stage IV was considered to be disseminated (multifocal) involvement of ≥1 extralymphatic organ(s) (code 80). In unstaged disease, the tumor stage was unknown (code 99).28
The cancer mortality rate is the number of deaths wherein cancer is given as the underlying cause of death and occurring in a specified population during 1 year, usually expressed as the number of deaths because of cancer per 100,000 persons. Because the vital status of patients with MCL as of the last date of follow-up (December 31, 2004) is available in the SEER data, we examined the mortality rate, termed incidence-based mortality (IBM).28, 29 IBM is a statistical tool for calculating population-based mortality rates according to tumor characteristics. IBM is the cross-sectional mortality rate in the population as a whole for a given tumor type in a specified calendar period. Consequently, the IBM reflects the combined effect of cancer incidence, diagnosis, and treatment.
SEER*Stat 6.3.6 is statistical software designed specifically for the analysis of SEER and other cancer-related databases. The statistics calculated using this software included age-adjusted rates, gamma confidence intervals for age-adjusted rates, trends in rates, the annual percent change, and IBM. We calculated the age-adjusted incidence rates for MCL per year from 1992 to 2004. The sex-specific and race-specific, age-adjusted, area-specific incidence rates and their 95% confidence intervals were also reported. As noted, these incidence rates were age-adjusted to the US population in 2000. Age-specific incidence rates were calculated for all patients. We calculated the proportions of patients with MCL who belonged to the following racial or ethnic groups: Caucasian, African American, or other (American Indian/Alaskan Native, Asian/Pacific Islander, other unspecified, and unknown). We also examined the trends in the incidence of MCL in linear Poisson multivariate regression models. Finally, using the SEER*Stat's IBM method, we calculated disease-specific mortality stratified by geographic area.
Incidence of MCL Over 13 Years
Among the 87,166 patients diagnosed with NHL at various ages, 2459 had confirmed MCL, accounting for 2.8% of patients with NHL. As shown in Table 1, the overall 13-year age-adjusted incidence rate was 0.55 for 100,000 persons per year. This rate reflects the number of new cases with onset of disease per 100,000 persons per year. The age-adjusted incidence rate was highest for the population aged 70 to 79 years and lowest for those aged <50 years. The incidence rate in men (0.84) was more than twice that of women (0.34). In addition, the MCL incidence rate of 0.61 of 100,000 in Caucasians was higher than that of other races. The data also demonstrated that the majority of patients were diagnosed at later stages; patients with stage III and stage IV disease accounted for 74.6% of all patients. There were also substantial geographic variations in MCL incidence rates over the 13-year period. For example, Alaskan Natives had the lowest age-adjusted incidence rate, at 0.10 of 100,000, and Seattle had the highest, at 0.67 of 100,000.
|Age-adjusted Incidence Rate per 100,000 (95% CI)*||Relative Risk (95% CI)|
|50-59||0.83 (0.76-0.92)||11.72 (10.00-13.76)|
|60-69||1.96 (1.81-2.12)||27.55 (23.77-32.01)|
|70-79||2.97 (2.76-3.19)||41.70 (36.08-48.34)|
|≥80||2.78 (2.51-3.07)||39.10 (33.29-45.99)|
|Men||0.84 (0.80-0.88)||2.48 (2.28-2.70)|
|African American||0.32 (0.26-0.38)||1.17 (0.90-1.52)|
|Caucasian||0.61 (0.59-0.64)||2.25 (1.89-2.71)|
|Other ethnicities||0.27 (0.23-0.32)||1.00|
|Stage I||0.08 (0.07-0.09)||1.71 (1.43-2.05)|
|Stage II||0.05 (0.04-0.06)||1.00|
|Stage III||0.08 (0.07-0.09)||1.62 (1.35-1.94)|
|Stage IV||0.30 (0.29-0.32)||6.32 (5.42-7.39)|
|Alaska Natives||0.10 (0.00-0.62)||1.00|
|Rural Georgia||0.24 (0.00-0.62)||2.42 (0.22-122.09)|
|Hawaii||0.37 (0.28-0.47)||3.76 (0.57-151.12)|
|San Francisco–Oakland||0.48 (0.42-0.54)||4.88 (0.76-193.40)|
|New Mexico||0.48 (0.40-0.59)||4.97 (0.77-198.11)|
|Atlanta (Metropolitan)||0.48 (0.40-0.58)||4.97 (0.77-197.75)|
|Los Angeles||0.54 (0.49-0.58)||5.49 (0.87-217.38)|
|Connecticut||0.55 (0.49-0.62)||5.67 (0.89-224.90)|
|San Jose–Monterey||0.57 (0.48-0.67)||5.85 (0.91-232.50)|
|Detroit (Metropolitan)||0.59 (0.53-0.67)||6.11 (0.96-241.93)|
|Utah||0.63 (0.53-0.75)||6.49 (1.01-258.01)|
|Iowa||0.64 (0.57-0.73)||6.60 (1.04-261.58)|
|Seattle (Puget Sound)||0.67 (0.59-0.74)||6.83 (1.07-270.76)|
Trend in Incidence of MCL by Factor Analyzed
Overall incidence of MCL and trends in incidence by sex
To find the overall incidence trend, we investigated the age-adjusted incidence rates for MCL between 1992 and 2004 (Fig. 1). The age-adjusted incidence rate for MCL was lowest in 1992 at 0.27 of 100,000, and was highest in 2004 at 0.69 of 100,000, which was nearly 2.5 times than the 1992 rate.
As shown in Figure 1, the age-adjusted incidence rates for MCL increased progressively for both sexes. The annual percent change for both sexes was 5.87% (P < .05, which is significantly different from zero). However, this rate was much greater for men than for women. In 1992, the age-adjusted incidence rate was 0.34 of 100,000 for men and 0.21 of 100,000 for women. By 2004, however, the age-adjusted incidence rate for men was more than 2 times that of women: 1.01 of 100,000 in contrast to 0.44 of 100,000. Therefore, the steady increase in the overall age-adjusted incidence rate was mainly because of the increase noted in men. The gap in age-adjusted incidence rate between men and women has progressively increased in the past 13 years. It is reflected in the difference in annual percent change between sexes. The annual percent change was 7.79% (P < .05) for men and was 2.53% (P > .05) for women.
Trends in incidence of MCL by age
As shown in Figure 2, the age-adjusted incidence rates increased with age. Patients aged <50 years had the lowest incidence rates. The highest incidence rate of MCL occurred in patients aged >80 years. The age-adjusted incidence rates for patients aged ≥50 years have increased progressively with time. The annual percent change in patients ages 50 to 59 years, 60 to 69 years, 70 to 79 years, and ≥80 years were 4.73%, 7.65%, 6.41%, and 7.98%, respectively, which were all statistically significant (P < .05). In contrast, the trend in age-adjusted incidence rates remained flat for patients aged <50 years; the annual percent change was −2.85% (P > .05). We found that the median age at the time of MCL diagnosis was 68 years (67 years for men and 70 years for women), and all the incidence rate increases occurred in the older populations.
Trends in incidence of MCL by race
As shown in Figure 3, the age-adjusted incidence rate in Caucasians increased progressively with time, more so than it did in African Americans or other races. The annual percent change from 1992 to 2004 for Caucasians was 6.31% (P < .05), and that for African Americans was 1.28% (P > .05). The incidence rates of MCL in other races were generally lower than in Caucasians and African Americans, except between the years 1999 and 2002. However, the annual percent change for other races (5.34%) was higher than that of African Americans (P > .05).
Trends in incidence of MCL by tumor stage
As shown in Figure 4, the age-adjusted incidence rate for the early stages (I-III) of MCL remained steady over a period of 13 years. The annual percentage changes were −3.06% (P > .05), 2.31% (P < .05), and 4.88% (P < .05) for stages I, II, and III disease, respectively. In contrast, the age-adjusted incidence rate for stage IV increased dramatically with time, with an annual percent change of 10.22%.
Variation in incidence of MCL by geographic areas
As shown in Table 1, the incidence rates varied substantially in different registry areas over a period of 13 years. Alaskan Natives had the lowest incidence rate at 0.10 of 100,000, and Seattle had the highest rate at 0.67 of 100,000. The results of incidence trend analysis from 1992 to 2004 showed that Los Angeles, San Francisco–Oakland, Iowa, New Mexico, and Seattle experienced a significant increase in the incidence of MCL. The annual percent changes for these 5 areas were 4.72%, 5.70%, 6.07%, 6.75%, and 8.01%, respectively. The annual percentage changes for other areas were 3.09% in Utah, 3.74% in Hawaii, 4.89% in San Jose–Monterey, 5.22% in Connecticut, 5.36% in Detroit, and 5.88% in Atlanta. The annual percentage changes in incidence for Alaskan Natives and rural Georgians were not available.
Trends in Incidence over 13 Years: Comparison of Mantle Cell Lymphoma, Diffuse Large B-cell Lymphoma, Follicular Lymphoma, and a Subtype of Non-Hodgkin Lymphoma
The overall trend for all 4 curves was an increased incidence rate over the 13 years (Fig. 5). The annual percent changes for NHL, diffuse large B-cell lymphoma, and follicular lymphoma were 0.2% (P > .05), 0.17% (P > .05), and 1.23% (P < .05), respectively. Their annual percent changes were all much smaller than that of MCL, which had an annual percent change of 5.87%. The incidence rates for all types of NHL and diffuse large B-cell lymphoma were relatively stable. For follicular lymphoma, the incidence rate increased significantly. The trend in the incidence of diffuse large B-cell lymphoma was similar to that of all types of NHL. For MCL, the trend differed from that of the 3 other disease categories, demonstrating a sharp increase in incidence between 1992 and 2004.
Incidence and Mortality Rates in MCL
To investigate the association between incidence and mortality, we compared the age-adjusted incidence rate with the mortality rate (Fig. 6). The annual percent change in the incidence of MCL and annual percentage of mortality of MCL were 5.87% and 8.27%, respectively. Unfortunately, the mortality rate of MCL in the SEER data also increased with time and was in parallel (correlation coefficient of 0.89). This illustrates the finding that despite the availability of many therapies, outcomes had not yet substantially improved, indicating that MCL remains incurable.
NHL is a group of complex malignancies with many subtypes.30, 31 Different subtypes of NHL often need different therapies with different prognosis. Therefore, it is important to study each subtype of NHL separately.31 In the case of MCL, there is lack of data from population-based studies. In this study, we have reported incidence trends of MCL by age, sex, race, and geographic areas within the US over a 13-year period from 1992 to 2004. To our knowledge, the current study is the first report that such demographic data and geographic variations have been described.
The SEER Program reevaluated all NHL cases registered between 1975 and 1991. There has been concern regarding the accuracy of these case reclassifications. Since MCL became a distinct subtype of NHL with specific diagnostic criteria in 1992, the diagnosis for cases from 1992 onward that were included in the current study was considered to be accurate overall, although certain subtypes of lymphoma might have relatively lower reliability.30–32 NHL incidence rates have been reported to increase with time.33-36 We found the trend in MCL similar to the trends in NHL, diffuse large B-cell lymphoma, and follicular lymphoma, with the MCL curve being steeper than that of NHL and the increase in the incidence of MCL greater. The overall increase in age-adjusted incidence rate was mainly due to the increased age-adjusted incidence rate for patients with stage IV MCL. This is consistent with the natural history of MCL, since MCL is an insidious disease that is not associated with clinical symptoms until the late stage. We found the increase in incidence to be significantly higher in Caucasian men aged >50 years. This finding is consistent with other studies that MCL is more common in elderly Caucasian men. Moreover, the highest increase in incidence was seen in patients with stage IV disease, indicating that early diagnosis continues to be a major challenge to the medical community.
The cause of this remarkable increase in the age-adjusted incidence rate in MCL over a 13-year period is unknown. This increase in MCL incidence may not be due to the improvement in detection tools or in the pathologic recognition, because the incidence increase was largely limited to stage IV disease, rather than all stages. It was unlikely due to a change in the diagnostic criteria, since between 1992 and 2004, the criteria remained relatively unchanged. However, it might be interesting to investigate whether the increase in MCL incidence is associated with changes in environment and geographic locations over time. The results of the current study demonstrated a significant geographic variation within the 13 SEER regions, but the reasons are unclear to us. Again, the increase in MCL incidence was unlikely to be due to variation in diagnosis in these 13 SEER areas, because the SEER registries implement uniform diagnosis criteria and perform the highest standard of data quality control. However, the possibility of applying these diagnosis criteria differently in actual diagnosis and coding by various registry staff cannot be ruled out.
The current study demonstrated incidence and mortality curves ran parallel, indicating that increased incidence rate is accompanied by higher mortality. However, this finding does not reflect the newer, highly effective therapies that have been tested and approved since 2004. During this time period, novel efficacious therapies were approved and used for patients with MCL. For example, rituximab (a chimeric monoclonal anti-CD20 antibody) revolutionized the treatment of NHL and improved survival responses.37, 38 A regimen of rituximab plus hyper-CVAD (fractionated cyclophosphamide, vincristine, doxorubicin[ adriamycin], and dexamethasone) alternating with rituximab plus high-dose methotrexate and cytarabine was proven to be effective in previously untreated, aggressive MCL by Romaguera et al.39 In that study, 97% of 97 assessable patients responded, and 87% achieved a complete response or unconfirmed complete response. With a median follow-up time of 40 months, the 3-year failure-free survival and overall survival rates were 64% and 82%, respectively. Bortezomib, approved by the US Food and Drug Administration (FDA) on December 8, 2007, has also been effective in the treatment of relapsed MCL.40, 41 Autologous stem cell transplants have helped improve survival rates as well.42 All such advances in the treatment of MCL occurred after the SEER data were compiled in 2004. The SEER data do not reflect changes in survival rates brought about by more recent treatment improvements. Therefore, further studies may be needed to examine the incidence and mortality trend in patients with MCL after the new data from 2004 through 2007 become available.
MCL has the poorest prognosis of all lymphomas, but important advances have been made in recent years. New biologic agents such as monoclonal antibodies (rituximab), proteasome inhibitors (bortezomib), mammalian target of rapamycin inhibitors (CCI-779), and immunomodulatory agents (thalidomide and lenalidomide) have been demonstrated to be promising clinically. Therefore, a better understanding of the epidemiology of MCL, the development of novel agents, more research funding, and increased public awareness are all needed to change the natural history of MCL.
The current study has several weaknesses. This research has been hindered by the many coexisting classification schemes. Patients diagnosed before 2001 may have diagnosis codes from earlier ICD-O versions that need to be converted to the ICD-O-3, which may have resulted in higher proportions of unclassified (eg, lymphoma and not otherwise specified) cases. Clarke et al studied the agreement of computer-converted ICD-O-3 codes to ICD-O-3 codes generated directly from diagnostic pathology reports vis-à-vis the reproducibility of unclassified status, and found that classification of MCL had a relatively low reliability.30, 31 Another weakness lies in the fact that the cases after 2004 were not available in SEER and could not be included in our study. Thus, the impact of newly approved treatments on the incidence and survival in patients with MCL cannot be examined in our analysis. This limitation can be resolved by using new data in the next few years.
In conclusion, the incidence of MCL increased progressively between 1992 and 2004 and was significantly higher in men, in Caucasians, and in patients aged ≥50 years. The majority of patients were diagnosed in later stages, and there were considerable geographic variations in incidence rate.
We acknowledge the contributions of the National Cancer Institute Information Management Services Inc., the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries, and SEER*Stat software to the creation of this database. The interpretation and reporting of these data are the sole responsibilities of the authors.
- 1American Cancer Society. Cancer facts & figures 2007. Available at: http://www.cancer.org/downloads/STT/CAFF2007PWSecured.pdf Accessed October 2, 2007.
- 2The Non-Hodgkin's Lymphoma Classification Project. A clinical evaluation of the International Lymphoma Study Group classification of non-Hodgkin's lymphoma. Blood. 1997; 89: 3909–3918.
- 6Clinical analysis of 670 cases in two trials of the European Organization for the Research and Treatment of Cancer Lymphoma Cooperative Group subtyped according to the Revised European-American Classification of Lymphoid Neoplasms: a comparison with the Working Formulation. Blood. 1996; 87: 4358–4367., , , et al.
- 8REAL classification of malignant lymphomas in the Republic of Korea: incidence of recently recognized entities and changes in clinicopathologic features. Hematolymphoreticular Study Group of the Korean Society of Pathologists. Revised European-American lymphoma. Cancer. 1998; 83: 806–812., , , et al.
- 18Histopathology of Non-Hodgkin's Lymphomas: Based on the Kiel Classification. New York, NY: Springer-Verlag; 1981., .
- 24World Health Organization Classification of Tumours. Pathology and Genetics of Tumours of Heamatopoietic and Lymphoid Tissues. Lyon: IARC Press; 2001., , , et al.
- 25Surveillance, Epidemiology, and End Results (SEER) Program. SEER Stat Database: Incidence–Seer 13 Regs Limited-Use, Nov 2006 Sub (1992–2004)–Linked to County Attributes–Total U.S., 1969–2004 Counties, National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2007, based on the November 2006 submission. Available at URL: http://www.seer.cancer.gov.
- 26SEER Extent of Disease 1998, Codes and Coding Instructions (No. T-899; Pub. No. 98-4354). 3rd ed. Bethesda, MD: Cancer Statistics Branch, Surveillance Program, Division of Cancer Control and Population Sciences, National Cancer Institute, US Department of Health and Human Services, Public Health Service, National Institutes of Health, 1998;1: 7., , eds.
- 27The Surveillance Research Program of the Division of Cancer Control and Population Sciences, National Cancer Institute. Surveillance Research Program. SEER Stat Software, version 6.3.6. Bethesda, MD: National Cancer Institute. Available at URL: http://www.seer.cancer.gov/seerstat.
- 29National Cancer Institute. SEER*Stat statistical methods for incidence and mortality. Available at http://srab.cancer.gov/software/seerstat.html. Accessed June 11, 2007.
- 35Increasing incidence of non-Hodgkin's lymphoma in Karachi, 1995-2002. Asian Pac J Cancer Prev. 2005; 6: 364–369., , , et al.
- 42The role of hematopoietic stem cell transplantation in non-Hodgkin lymphoma. Clin Adv Hematol Oncol. 2006; 4: 521–530., .