Departments of 1 Anatomical Pathology and 3 Haematology, St Vincents Hospital, Sydney; 2 School of Public Health, University of Sydney, Sydney; 4 National Centre in HIV Epidemiology and Clinical Research, Sydney, Australia
Received 16 October 2003; revised 18 December 2003; accepted 23 December 2003
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Non-Hodgkins lymphoma (NHL) is pathologically diverse. Epidemiological investigations into its increasing incidence and aetiology require accurate subtype classification.
Patients and methods:
Available pathology reports of 717 cases aged from 20 to 74 years in an Australian, population-based epidemiological study of NHL were reviewed by one anatomical pathologist to assign a World Health Organization (WHO) classification category. High or low confidence was assigned to the diagnosis of NHL, cell phenotype and WHO category and reasons given for low confidence.
Results:
The most informative biopsy reports were from open tissue biopsy (79% of cases), tissue core biopsy (8%), cytology (4%) and bone marrow (9%); 8% of cases had inadequate biopsies for diagnostic purposes. Immunohistochemistry or flow cytometry reports were available for 96% of cases, gene rearrangement studies for 6% and cytogenetics for 3%. The reviewer assigned high confidence to the diagnosis of NHL in 93% of cases and also the phenotype in 88%. While a WHO classification could be assigned in 91% of cases, confidence was high in only 57.5%; insufficient immunophenotyping was the commonest reason for low confidence.
Conclusions:
Expert pathology review of a population-based sample of NHL can provide a WHO classification category for most cases. A high level of confidence in the classification, however, would require review of diagnostic material and additional phenotyping.
Key words: epidemiology, NHL subtypes, pathology report review, WHO NHL classification
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Trends in NHL subtypes have been difficult to assess retrospectively because many epidemiological studies of NHL, particularly in the USA, have used cancer registry data and classified cases according to the Working Formulation (WF) [9], a purely morphological classification supplemented after the mid 1980s by limited immunophenotyping. A move to international standardisation came in 1994 with the more broadly based classifications of the Revised EuropeanAmerican Classification of Lymphoid Neoplasms (REAL) [10] and its successor, the World Health Organization (WHO) classification [11], which have become accepted internationally but have yet to be used extensively in epidemiological studies.
The WHO classification defines NHL subtypes according to morphology in combination with cell phenotype, genetic features, clinical features, race, geographic distribution and microbiologic features [11], the relative importance of which depends on availability and subtype. Ideally, a WHO classification category is assigned by reviewing slides from an adequate biopsy, with access to clinical and ancillary test information. Mostly, however, epidemiological studies with large case numbers lack funding for full, centralised expert pathology review, so other strategies are required for uniform and accurate classification.
The New South Wales (NSW) non-Hodgkins Lymphoma Study is a population-based, epidemiological casecontrol study of the role of sun, occupational and infectious exposures in NHL causation. As it is a participant in the International Collaboration of Investigators Working on Lymphoma Epidemiological Studies (InterLymph) based on studies in North America, Europe and Australia [12], standardised and validated pathological diagnoses are needed to share and pool data across studies.
As the first population-based investigation of NHL aetiology in Australia, the NSW casecontrol study offered a unique opportunity to examine subtype classification. Since most lymphoma patients are treated in tertiary referral centres in Australia in which well developed referral networks ensure the availability of ancillary tests and consultant opinions, reasonably accurate classification was expected even though the initial diagnostic biopsy is often done elsewhere.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Pathology report review
Anatomical pathology reports relevant to the diagnosis of lymphoma were obtained from the NSW Cancer Registry or the treating clinician for patients who had given informed consent. Pathology practices were asked for additional reports with potentially relevant results for immunohistochemistry, flow cytometry, cytogenetics, molecular pathology, peripheral blood lymphocyte count, bone marrow or serum or urine protein electrophoresis, for patients who had consented to gaining this additional information (92.5% of patients).
The review pathologist (JT), who had a particular interest in haematopathology, read de-identified reports. The original reporting pathologists were classified as anatomical pathologists in tertiary referral centres if they worked in recognised teaching hospitals [13] or served on the Australian Cancer Network Working Party for the Diagnosis of non-Hodgkins Lymphoma. Clinical details, site of lymphoma and diagnostic biopsy specimen type were recorded. The review pathologist assigned a WHO category, where possible, on the basis of all information available including translation from the classification used by the original reporting pathologist (WF, REAL or WHO). One code was used for all three grades of follicular lymphoma, and one code for all subtypes of diffuse large B-cell lymphoma. WHO categories for which the distinction between leukaemia and lymphoma is based on lymphocyte counts in blood and/or bone marrow and which could not reliably be distinguished in this review were given both applicable categories in the form of a combined description, e.g. small lymphocytic lymphoma OR chronic lymphocytic leukaemia. Patients in whom two separate types of lymphoma were clearly described were given two codes, e.g. diffuse large B-cell lymphoma AND follicular lymphoma, except in cases of diffuse large B-cell lymphoma where the marrow was involved by morphologically low-grade lymphoma.
A high or low level of confidence in the diagnosis of NHL, phenotype and WHO classification category was assigned as outlined in Tables 1 and 2, which were based on the published criteria (WHO 2001). To increase certainty, emphasis was placed on ancillary studies, especially flow cytometry, which is standardised, and immunohistochemistry, which is relatively objective. Immunophenotyping was required for confident classification of all subtypes, including follicular lymphoma. Morphological description had to be stated with confidence and consistent with the diagnosis. The essential clinical information depended on the WHO category and, in general, adequate tissue biopsy was required. Fine needle aspiration (FNA) biopsy, bone marrow biopsy or peripheral blood were often considered adequate specimens for eligibility and phenotype but inadequate for WHO classification, with the exception of some categories of lymphoma or leukaemia with primary bone marrow involvement or a diagnostic immunophenotype. Where possible, a differential diagnosis was made for cases with low confidence levels. Reasons for low confidence were recorded.
|
|
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
Immunohistochemistry or flow cytometry reports were available for 96% of cases (immunohistochemistry, 80%; flow cytometry, 55%; both, 40%), and gave the standardised information needed for immunophenotyping. Other ancillary tests were rare (gene rearrangement studies, 6%; cytogenetics, 3.5%). Specialist anatomical pathologists in tertiary referral centres made 48% of the primary diagnoses and in an additional 22% of cases provided a second opinion.
Confidence by the review pathologist in the diagnosis of NHL was high in 667 cases (93% of 717), and high in the phenotypic classification for 635 cases (88% of 717). However, for diagnosis, phenotype and WHO classification, confidence was high in only 412 cases (57% of 717). Low confidence in the diagnosis of NHL was mostly due to the inability to exclude simple reactive changes or Hodgkins lymphoma, or if there was an inadequate specimen (Table 4). Insufficient immunophenotype information was the most common barrier to the review pathologist assigning a WHO category, explaining 129 (58%) ratings of low confidence, while inadequate biopsy material explained another 47 (21%). Insufficient immunophenotyping was the reason for low confidence in 40% of 239 cases classified as follicular lymphoma; immunophenotyping was absent in 6%, non-diagnostic in 2% and insufficient for a WHO classification in 32%.
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
On the basis of expert report review, we could have high confidence in the WHO classification of NHL in only 57% of cases on report review, despite 96% of cases having some immunophenotyping. The commonest reason for low confidence in the WHO classification was insufficient immunophenotyping in 18% of 717 cases. In particular, 40% of cases diagnosed with follicular lymphoma, the most common NHL subtype in the NSW study (33%), had insufficient immunophenotyping for WHO classification. While many pathologists consider that follicular lymphoma can easily be diagnosed using morphological criteria alone i.e. not requiring immunophenotypic confirmation, a follicular or nodular morphologic pattern may also be seen in reactive lymphadenopathy, Hodgkins lymphoma and subtypes of NHL other than follicular lymphoma, including mantle cell lymphoma, marginal zone lymphoma and small lymphocytic lymphoma/chronic lymphocytic leukaemia.
Follicular lymphoma (Working Formulation categories B, C and D) in the European Organisation for Research and Treatment of Cancer (EORTC) trials was confirmed at slide review according to the REAL classification in only 223 (69%) of 323 patients [16]. Three cases were reclassified as chronic lymphocytic leukaemia, 23 as mantle cell lymphoma, eight as marginal zone lymphoma and six as diffuse large B-cell lymphoma, while an additional 17 cases were identified among other Working Formulation categories of the total group of 670 cases [16]. The five expert haematopathologists in the Non-Hodgkins Lymphoma Classification Project [17] could not accurately diagnose follicular lymphoma on morphology alone in 7% of cases, although immunophenotyping was considered to improve accuracy by very little. Data from these two studies suggest that morphologic diagnosis of follicular lymphoma is not accurate in a proportion of cases. In our study, the review pathologist required immunophenotyping to assign a confident diagnosis of follicular NHL.
Inadequate biopsy material (tissue biopsy reported as inadequate, bone marrow alone or FNA alone), prevented confident WHO classification in 8% of 717 cases. Minimally invasive biopsy techniques, which can produce suboptimal material, are increasingly used although open surgical tissue biopsy for diagnosis and classification of lymphoma is still generally recommended [18]. Now that FNA cytology combined with flow cytometry is an accepted triage procedure for initial lymphoma diagnosis, some centres promote FNA cytology as the definitive biopsy for lymphoma diagnosis and subtyping if ancillary studies are available [19]. However, FNA has technical and sampling limitations and does not allow the vital architectural assessment [20] needed in diagnosing follicular lymphoma, one of the commonest lymphomas in Western countries. NHL therapeutic trials do not accept FNA as an adequate biopsy, and the InterLymph collaboration has made the same decision.
Clinical information is essential in some categories of the WHO classification. Our report review relied on clinical information in anatomical pathology reports, often no more than the patients age and sex and the biopsy site. Since classification of T-cell lymphomas and leukaemias is strongly dependent on clinical information and ancillary studies, confidence in the WHO classification was, as a consequence, low in many of these cases and in those that required separating marginal zone lymphoma from lymphoplasmacytic lymphomas. On the other hand, recording of biopsy site was useful in applying the WHO classification, especially in marginal zone lymphoma.
Error in assigning cases to subtypes may have substantial effects on the direction and shape of observed doseresponse relationships in epidemiological studies [21]. We intend to measure the error in our report review assignment of WHO classification categories by comparing these assignments with expert review of pathology slides in a proportion of cases. Additional immunochemistry will be done when needed and possible. This will increase confidence in the WHO classification in many cases, although not in those with inadequate material.
The problems of inadequate specimens and insufficient immunophenotyping for use of the WHO classification are being addressed in a number of countries by the preparation of clinical practice guidelines recommending that biopsy material should be adequate and have careful triage for ancillary studies, including flow cytometry, when lymphoma is a possibility [18, 22]. Guidelines in Australia are currently being developed by a Working Party of the Australian Cancer Network. When adopted and implemented, the accuracy of the WHO classification of lymphoma recorded in cancer registries and obtained by pathology report review in large population-based lymphoma studies will be greatly enhanced.
![]() |
Acknowledgements |
---|
![]() |
FOOTNOTES |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2. Cartwright RA. Changes in the descriptive epidemiology of non-Hodgkins lymphoma in Great Britain? Cancer Res 1992; 52 (19 Suppl): 5441s5442s.
3. Grulich AE, Li Y, McDonald AM et al. Decreasing rates of Kaposis sarcoma and non-Hodgkins lymphoma in the era of potent combination anti-retroviral therapy. AIDS 2001; 15: 629633.[CrossRef][ISI][Medline]
4. Eltom MA, Jemal A, Mbulaiteye SM et al. Trends in Kaposis sarcoma and non-Hodgkins lymphoma incidence in the United States from 1973 through 1998. J Natl Cancer Inst 2002; 94: 12041210.
5. Carli PM, Boutron MC, Maynadie M et al. Increase in the incidence of non-Hodgkins lymphomas: evidence for a recent sharp increase in France independent of AIDS. Br J Cancer 1994; 70: 713715.[ISI][Medline]
6. Cartwright RA, Brincker H, Carli PM et al. The rise in incidence of lymphomas in Europe 19851992. Eur J Cancer 1999; 35: 627633.[CrossRef][ISI][Medline]
7. Groves FD, Linet MS, Travis LB, Devesa SS. Cancer surveillance series: non-Hodgkins lymphoma incidence by histologic subtype in the United States from 1978 through 1995. J Natl Cancer Inst 2000; 92: 12401251.
8. Anderson JR, Armitage JO, Weisenburger DD. Epidemiology of the non-Hodgkins lymphomas: distributions of the major subtypes differ by geographic locations. Non-Hodgkins Lymphoma Classification Project. Ann Oncol 1998; 9: 717720.[Abstract]
9. The Non-Hodgkins Lymphoma Pathologic Classification Project. National Cancer Institute sponsored study of classifications of non-Hodgkins lymphomas: summary and description of a working formulation for clinical usage. Cancer 1982; 49: 21122135.[ISI][Medline]
10. Harris NL, Jaffe ES, Stein H et al. A revised EuropeanAmerican classification of lymphoid neoplasms: a proposal from the International Lymphoma Study Group. Blood 1994; 84: 13611392.
11. Jaffe ES, Harris NL, Stein H, Vardiman JW (eds). Pathology and Genetics of Tumours of Haematopoietic and Lymphoid Tissues. World Health Organization Classification of Tumours. Lyon, France: IARC Press 2001.
12. Boffetta P, Linet M, Armstrong B. The InterLymph Collaboration; a consortium of molecular epidemiological studies of non-Hodgkins lymphoma. Proc Am Assoc Cancer Res 2003; 44: 308 (Abstr 1579).
13. Royal Australasian College of Physicians. Requirements for Physician Training in Australia and New Zealand. [online] http://www.racp.edu.au.
14. Glaser SL, Dorfman RF, Clarke CA. Expert review of the diagnosis and histologic classification of Hodgkin disease in a population-based cancer registry: interobserver reliability and impact on incidence and survival rates. Cancer 2001; 92: 218224.[CrossRef][ISI][Medline]
15. Siebert JD, Harvey LA, Fishkin PA et al. Comparison of lymphoid neoplasm classification. A blinded study between a community and an academic setting. Am J Clin Pathol 2001; 115: 650655.[CrossRef][ISI][Medline]
16. Pittaluga S, Bijnens L, Teodorovic I et al. Clinical 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 EuropeanAmerican Classification of Lymphoid Neoplasms: a comparison with the Working Formulation. Blood 1996; 87: 43584367.
17. The Non-Hodgkins Lymphoma Classification Project. A clinical evaluation of the International Lymphoma Study Group classification of non-Hodgkins lymphoma. Blood 1997; 89: 39093918.
18. ESMO minimum clinical recommendations for diagnosis, treatment and follow-up of newly diagnosed large cell non-Hodgkins lymphoma. Ann Oncol 2001; 12: 12091210.[CrossRef][ISI][Medline]
19. Meda BA, Buss DH, Woodruff RD et al. Diagnosis and subclassification of primary and recurrent lymphoma. The usefulness and limitations of combined fine-needle aspiration cytomorphology and flow cytometry. Am J Clin Pathol 2000; 113: 688699.[CrossRef][ISI][Medline]
20. Sandhaus LM. Fine-needle aspiration cytology in the diagnosis of lymphoma. The next step. Am J Clin Pathol 2000; 113: 623627.[CrossRef][ISI][Medline]
21. Armstrong B, Boffetta P. Measurement of the exposure and outcome in epidemiological studies used for quantitative estimation and prediction of risk. In Moolgavkar S, Krewski D, Zeise L et al. (eds): Quantitative Estimation and Prediction of Human Cancer Risks. Lyon, France: IARC 1999; 75102.
22. Zelentz AD, Appelbaum FR, Cabanillas F et al. National Comprehensive Cancer Network Practice Guidelines in Oncology, version 1: Non-Hodgkins Lymphoma 2003. [online] http://www.nccn.org.
23. Fritz A, Percy C, Jack A et al. International Classification of Diseases for Oncology (ICD-O), 3rd edition. Geneva, Switzerland: World Health Organisation 2000.