An official website of the United States government
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
National Guideline Alliance (UK). Non-Hodgkin's Lymphoma: Diagnosis and Management. London: National Institute for Health and Care Excellence (NICE); 2016 Jul. (NICE Guideline, No. 52.)
A surgically excised tissue biopsy is widely accepted as the gold standard for the diagnosis of lymphoma based upon the current international guidelines (Lugano 2014 and ESMO 2015). An excision biopsy of a lymph node (or other tissue) allows assessment of micro-architecture, provides adequate material for immunocytochemistry, flow cytometry if received unfixed, FISH studies and extraction of DNA and RNA for molecular diagnostics. Concordance between the results of these investigations provides a high level of confidence in the diagnosis. Where the disease process is focal an excision biopsy is more likely to be diagnostic by virtue of the volume of tissue obtained and excision biopsies, in addition are typically less prone to processing artefacts which can impair morphological interpretation.
The major disadvantages of an excision biopsy are the need for general anaesthesia and the delays that can result from seeking a surgical opinion. These issues can be addressed by using needle core biopsies, but at the expense of a reduction in the range and quality of investigations that can be performed, unless multiple 10-15 mm cores have been taken when the amount of tissue may be similar to some excision biopsies. However, single thin cores of 5mm or less are common and this severely compromises all of the investigations listed above. Inadequate or too few core biopsies reduces the degree of confidence that can be placed in the diagnosis and judging when a needle core biopsy is adequate to support the immediate treatment of the patient is subjective and can be very difficult. This is compounded by routinely cutting step levels through these blocks, which results in a significant amount of the available tissue being discarded; this is common practice in many pathology departments. These problems frequently result in repeat biopsies being required with further delays to diagnosis and treatment.
An additional factor, in the near future, will be the need for a much higher standard of tissue collection and handling to support the diagnostics required for precision medicine. It is likely that unfixed tissue will be required to support sequencing-based techniques and that conditions under which samples are collected, transported and stored will become much more rigorous than is the case at present.
The critical question to be addressed is the circumstances where the loss of information and diagnostic confidence can be justified by logistical benefits and patient convenience. The main determinants will be the site of disease, urgency of treatment, patient preference and fitness.
Clinical question: Is core biopsy an acceptable alternative to excision biopsy for the accurate diagnosis of suspected non-Hodgkin's lymphoma at first presentation?
The review identified no evidence that met the inclusion criteria of the review.
A literature review of published cost-effectiveness analyses did not identify any relevant papers for this topic. Whilst there were potential cost implications of making recommendations in this area, other questions in the guideline were agreed as higher priorities for economic evaluation. Consequently no further economic modelling was undertaken for this question.
| Recommendations | Consider an excision biopsy as the first diagnostic procedure for people with suspected non-Hodgkin's lymphoma at first presentation. In people with suspected non-Hodgkin's lymphoma for whom the risk of a surgical procedure outweighs the potential benefits of an excision biopsy, consider a needle core biopsy procedure. Take the maximum number of cores of the largest possible calibre. For people with suspected non-Hodgkin's lymphoma in whom a diagnosis is not possible after a needle core biopsy procedure, offer an excision biopsy (if surgically feasible) in preference to a second needle core biopsy procedure. Pathology departments should ensure that tissue is conserved when handling needle core biopsies, so that further analysis can be carried out if needed. |
|---|---|
| Relative value placed on the outcomes considered | The GC considered accurate classification of non Hodgkin's lymphoma (NHL) to be the most important outcome when drafting the recommendations because treatment is crucially dependent on this. |
| Quality of the evidence | No published evidence was identified for this topic and so the GC based their recommendations on clinical expertise and experience. |
| Trade off between clinical benefits and harms | The GC decided that although no evidence was identified it was still important to make recommendations because accurate initial lymphoma diagnosis can reduce treatment delay and avoid incorrect treatment with serious adverse effects for the patient. Following initial lymphoma diagnosis, the patient typically enters the treatment pathway without further verification, unlike many other cancers where diagnosis is confirmed on material obtained during therapeutic surgery. It is therefore imperative that the correct diagnosis is obtained when a patient initially presents with suspected lymphoma. The GC considered that a correct diagnosis is usually easier to achieve when an excision biopsy has been obtained. The committee acknowledge that diagnosis using excision biopsy takes longer than a core biopsy, and that this delay might harm some patients with suspected lymphoma. However, the GC considered that this would apply to a minority of patients with aggressive disease and centres treating lymphomas should be able to ensure that appropriate services are provided in these cases. The GC also noted that repeat non-diagnostic core biopsies can themselves result in diagnostic delay. Other potential harms of excision biopsy include general anaesthesia and surgical complications. The GC agreed that the benefits of an accurate lymphoma diagnosis outweighed the potential harms because it will ensure the patient enters the correct treatment pathway, but the GC balanced the benefits and harms of the recommendations by allowing for factors specific to an individual patient to guide the choice of diagnostic procedure. The GC noted that inadequate sample quality is a frequently occurring problem and presents a major challenge for diagnostic pathologists in confidently diagnosing suspected lymphoma. The GC therefore made recommendations that collectively serve to ensure that adequate quality tissue samples are obtained for a confident diagnosis to be made. The GC considered the potential benefit of these recommendations will be that a correct diagnosis will be achieved in the highest possible number of patients with suspected lymphoma. |
| Trade off between net health benefits and resource use | No economic evidence was identified for this topic and no model was built. The GC estimated that the recommendations will increase the rate of excision biopsy and the associated costs, but this will be balanced by the decrease in cost associated with fewer diagnostic and non-diagnostic core biopsies and by the reduction in downstream costs associated with more accurate diagnoses. |
| Other considerations | The GC estimated that the change in practice needed to implement the recommendations will be varied: some centres currently carry out this practice, however there will be a significant change in centres where this is not the case. The GC acknowledged that there would be an impact on surgical resources and uptake would be dependent on availability of surgical services. |
Genetic and molecular testing has provided important insights into lymphoma biology. When applied to many lymphoma subtypes they have also demonstrated that the diagnosis and subclassification of lymphomas is more accurate when compared with traditional diagnostic methods such as standard microscopy and immunohistochemistry. Advances in this field may reduce heterogeneity in patients included in clinical trials, allow for greater confidence in the diagnostic process and improve patient outcomes.
Aggressive B-cell lymphoma can be subdivided into six main categories, as well a number of minor or rare subtypes. For the purposes of this question the six main categories are:
At present only the accurate diagnosis of Burkitt Lymphoma impacts on choice of therapy.
In the case of Burkitt lymphoma the presence of a MYC-IGH rearrangement as the sole abnormality identified by FISH in the context of a BCL2 negative germinal centres phenotype is the defining characteristic. The molecular subtypes of DLBCL are determined by gene expression profiling, which is the gold standard for identifying these subtypes, but is not routine practice.
The main problem is that most lymphoma diagnostic technologies are in a phase of rapid change. Data on these newer technologies is limited. Immunocytochemistry is increasingly recognised as being a poorly reproducible method unsuited for biomarker analysis. There is a large body of sequencing data (whole exome, targeted re-sequencing) that is highly relevant particularly to the diagnosis of Burkitt Lymphoma and the differentiation of GCB and ABC types of DLBCL and identification several of the genes within each category that are targets for specific therapy. Combinations of expression profiling and targeted sequencing are likely to become the method of choice over the next few years but experience in routine application is limited at the present time.
Clinical question: What is the most effective genomic/phenotypic testing strategy to diagnose the subtypes of aggressive b-cell non-Hodgkin's lymphoma?
Twenty six studies provided information on diagnostic tests. All were retrospective cross sectional studies using retrospectively collected data.
Four studies (Barrans et al., 2013; Gormley et al., 2005; Soldini et al., 2013 and Iqbal et al., 2015) including 796 patients assessed testing strategies to differentiate between BL and DLBCL. Low quality evidence from one study (Soldini et al, 2013) indicated all patients were accurately classified to their original diagnosis when using FISH. Low quality evidence from two studies (Barrans et al., 2013 and Iqbal et al., 2015) indicated that classic diagnostic methods can accurately diagnose BL and DLBCL compared to gene expression profiling at rates of 93.59-95.4%. Finally, one study (Gormley et al., 2005) provided low quality evidence that immunohistochemistry (IHC) can accurately diagnose patients into BL/DLBCL and GC/ABC subtypes compared to morphology at a rate of 85.5%.
Two studies (Dave et al., 2006 and Hummel et al., 2006) including 291 patients assessed testing strategies to differentiate between BL and other NHL subtypes. Low quality evidence from one study (Dave et al., 2006) indictaed that pathological review provides more diagnostic accuracy (87.3%) compared to classic diagnostic methods (73.2%) when diagnosing Burkitt lymphoma. One study (Hummel et al., 2006) provided low quality evidence that morphology can accurately diagnose patients into BL versus other NHL subtypes at a rate of 83.6%.
One study (Votavova et al, 2010) including 82 patients assessed the use of histopathological and clinical review compared to gene expression profiling in the diagnosis of PMBL reporting low quality evidence of a diagnostic accuracy rate of 85.4%.
One study reporting low quality evidence (Deffenbacher et al, 2010) including 17 patients assessed the use of pathological review compared to gene expression profiling in the diagnosis of HIV DLBCL, with a diagnostic accuracy rate of 64.7%.
Five studies (Barrans et al 2012; Malik et al, 2010; Booman et al, 2006; Scott et al, 2013 and Choi et al 2009) including 472 patients provided low quality evidence comparing various immunohistochemistry (IHC) algorithms to gene expression profiling (GEP). The highest rates of diagnostic accuracy (>90%) were reported when using IHC (93.4%; Malik et al. 2010), IHC Hans (91.5%; Scott et al., 2013), IHC Tally (93.6%; Scott et al., 2013) and IHC Choi algorithms (training set: 92.9%, validation set: 93.7%; Choi et al., 2009) and the lowest rate of diagnostic accuracy using IHC reported by Booman et al. (2006; 70%). Rimsza et al. (2009) assessed the use of qNPA at two thresholds (>0.8 and >0.9) compared to GEP reporting low quality accuracy rates of 92.3% (threshold >0.9) and 100% (threshold >0.8). Su et al., (2013) assessed the value of a bivariate mixture model reporting the a diagnostic accuracy rate when using a two-species analysis (human and canine) of 89.7% compared to 89.1% when using a human species alone analysis (89.1%). Finally, Williams et al. (2010) providing low quality evidence on the use of formalin-fixed paraffin embedded tissue when sub-typing DLBCL, reported a 97.7% accuracy rate compared to the use of fresh frozen tissues, and Mareschal et al. (2015) also providing low quality evidence found that GEP using a RT-MLPA assay accurately subtyped patients at a rate of 100% compared to GEP Affymetrix.
Four studies (Poulsen et al, 2005; Gutierrez-Garcia et al, 2011; Haarer et al, 2006 and Visco et al 2012) including 569 patients provided low quality evidence comparing various immunohistochemistry (IHC) algorithms to gene expression profiling (GEP). The highest rates of diagnostic accuracy (>90%) were reported when using IHC (92.7%; Poulsen et al., 2005) and a 3-marker algorithm (92.6%) or 4-marker algorithm (92.8%; Visco et al., 2012) and the lowest rate of diagnostic accuracy was reported when using the IHC Choi algorithm (59.1%; Gutierrez-Garcia et al., 2011). When assessing studies that had reported using the same IHC algorithms (Hans and Choi) there was wide variation between the reported diagnostic accuracy of these algorithms (59.1% compared to 90% for the Choi algorithm and 65.3% and 87.2%).
One study (Chang et al, 2010) assessed the use of FISH compared to polymerase chain reaction in the detection of t(14;18) in 227 patients with NHL reporting low quality evidence of a 70.5% accuracy rate. One study (Dunphy et al, 2008) assessed the use of FISH compared to PCR in the detection of BCL2 in 22 patients with primary mediastinal B-cell lymphoma reporting low quality evidence of a 95.5% accuracy rate. One study (Lynnhtun et al, 2014) assessed the use of FISH compared to immunohistochemistry plus FISH in the detection of MYC in 41 patients with high-grade B-cell lymphomas reporting low quality evidence of accuracy rates of 58.5% with a ≥40% IHC-FISH threshold and 87.8% at ≥70% and ≥80% IHC-FISH threshold. One study (Mationg-Kalaw et al, 2012) reported the use of pathological review compared to immunohistochemistry plus FISH in the detection of Ki67 in 432 patients with diffuse large B-cell lymphoma reporting low quality evidence of a 38.4% accuracy rate at >70% threshold and a 61.6% accuracy rate at >90% threshold. Finally, one study (Zeppa et al, 2012) assessed the use of flow cytometry, immunohistochemistry-FISH and polymerase chain reaction compared to histology and follow-up in the detection of immunoglobulin heavy-chain (IGH) signals in 48 patients with non-Hodgkin's lymphoma, reactive hyperplasia and small lymphocytic lymphoma/chronic lymphocytic leukemia reporting low quality evidence of accuracy rates of 95.8%, 86.4% and 80% (respectively).
A literature review of published cost-effectiveness analyses did not identify any relevant papers for this topic. Whilst there were potential cost implications of making recommendations in this area, other questions in the guideline were agreed as higher priorities for economic evaluation. Consequently no further economic modelling was undertaken for this question.
| Recommendations | Consider using FISH (fluorescence in situ hybridisation) to identify a MYC rearrangement in all people newly presenting with histologically high-grade B-cell lymphoma. If a MYC rearrangement is found, use FISH to identify the immunoglobulin partner and the presence of BCL2 and BCL6 rearrangements. |
|---|---|
| Relative value placed on the outcomes considered | Diagnostic accuracy including sensitivity, specificity, positive predictive value and negative predictive value were considered to be the critical outcomes for this topic. The GC considered sensitivity to be important in avoiding incorrect treatment as a result of disease misclassification. Test reproducibility and turnaround time were also important but no evidence was found for these outcomes. |
| Quality of the evidence | The quality of the evidence was low as assessed using QUADAS2. The reason for this was because the primary focus of the studies was not diagnostic accuracy so the publications lacked information about the index and reference standard tests. Additionally, studies provided limited information on selection of participants/samples and tended to use small sized hospital samples or databases without explanation for inclusion and exclusion of participants resulting in a large amount of uncertainty. Low quality evidence about testing strategies for primary mediastinal B-cell lymphoma came from a single study so the GC decided not to formulate a recommenation for this subgroup. The recommendation to undertake further studies when a MYC translocation is identified is based on the experience of the GC, as no evidence was identified that compared the outcomes of patients with MYC translocations to those of patients with both MYC, BCL2 and/or BCL6 translocations. |
| Trade off between clinical benefits and harms | The GC considered that these recommendations would lead to more accurate diagnosis and as a result, treatment could be more appropriately directed. The recommendations will also facilitate informed decision making with the patient. No harms were indentified. The GC recommended investigating cases of diffuse large B-cell lymphoma for the presence of a MYC rearrangement, and where a rearrangement is detected, to undertake further studies to identify rearrangements of BCL2 and BCL6. Distinguishing between cases with MYC as a sole abnormality and those with additional abnormalities is important in the differential diagnosis of Burkitt lymphoma and poor prognosis DLBCL. There was evidence that the presence of 2 or 3 of these abnormalities in DLBCL portends an adverse clinical outcome and although other factors (for example age) might modify this the GC thought that patients and clinicians would want to know this information. The GC also noted that there is an important clinical issue about misdiagnosis of Burkitt lymphoma as DLBCL and these recommendations will assist with this problem |
| Trade off between net health benefits and resource use | No health economic evidence was identified and no health economic model was built for this topic. The recommended tests are already being used. The GC considered the recommendations would be cost neutral due to a greater number of FISH tests but fewer immunohistochemistry tests. |
| Other considerations | The GC noted that FISH is currently the only method that can be used on a formalin fixed sample. It is also well documented that looking for evidence of the gene abnormality could be more useful in this context than looking for protein expression because the latter is unreliable. There was a lack of consensus on the methodology of gene expression profiling (GEP); although the various systems worked in research settings they were not yet robust enough to be used in routine practice. Research is moving towards newer practices so efforts are now being made to establish the best GEP platforms. |
Advanced molecular diagnostics will have a major impact on the diagnosis and stratification of all patients with lymphoma. Although the technologies are the same across lymphoma subtypes the data supporting its routine clinical application is greatest in high grade B-cell lymphomas.
In high grade B-cell lymphoma the application of molecular diagnostics is important in two areas:
Clinical question: What is the most effective genomic/phenotypic testing strategy to determine therapeutic stratification and prognostic subtypes of aggressive b-cell non-Hodgkin's lymphoma?
43 studies provided evidence about the prognostic value of molecular diagnostics in people with high grade B-cell lymphoma.
Moderate quality evidence from 22 studies (n=5065 patientes) reported overall survival does not differ between patients with GCB and non-GCB DLBCL subtype, although two additional studies suggest that overall survival may be inferior in patients with non-GCB (Molina, 2012, 2013; Mitrovic , 2013; n = 776; reported HRs ranged from 1.9-2; low quality). Progression-free survival (17 studies; n = 3177; moderate quality) does not differ between patients with GCB and non-GCB DLBCL subtype, although one additional study suggest that progression-free survival may be inferior in patients with non-GCB (Molina, 2012, 2013; n = 640; HR = 1.9; low quality).
Moderate quality evidence from 12 studies (n=1804 patients) reported overall survival does not differ between patients with GCB and non-GCB DLBCL subtype, although low quality evidence from one additional study suggest that overall survival may be inferior in patients with non-GCB (Perry, 2014 validation set; n = 215; reported HRs ranged from 2.07-2.14).
Moderate quality evidence from 9 studies (n=1396 patients) reported similar progression/event-free survival is either similar between patients with GCB and non-GCB/ABC DLBCL while low to moderate quality evidence from 3 studies (n=592 patients) reported lower progression/event free survival in patients with the non-GCB/ABC DLBCL subtype (HRs ranged from 2-2.27).
Five studies (n=1127 patients) provided low quality evidence that overall survival is either similar between patients with GCB and non-GCB DLBCL (4 studies; n = 652) or inferior in patients with the non-GCB DLBCL subtype (1 study; n = 475; HR = 0.56). Four studies (n=1187 patients) provided low quality evidence that progression-free survival is either similar between patients with GCB and non-GCB DLBCL (3 studies; n = 475) or inferior in patients with the non-GCB DLBCL subtype (1 study; n = 712; HRs ranged from 0.59-0.63).
Twelve studies (n=2051 patients) provided low-moderate quality evidence that overall survival does not differ between patients with GCB and non-GCB/ABC DLBCL.
Eight studies (n=1173 patients) provided low to moderate quality evidence that progression-free survival does not differ between patients with GCB and non-GCB/ABC DLBCL.
Low to moderate quality evidence from 6 studies (n=1573 patients) reported that overall survival is similar between patients with GCB and non-GCB/ABC DLBCL while five studies (n=1768 patients) provided low to moderate quality evidence that overall survival was inferior in patients with the non-GCB/ABC DLBCL subtype (reported HRs ranged from 0.53-2.1 [these span 0 as different reference groups are used]). There was large patient overlap between these studies. Progression-free survival is either similar between patients with GCB and non-GCB/ABC DLBCL (4 studies; n = 1488; low-moderate quality) or inferior in patients with the ABC DLBCL subtype (4 studies; n = 1577; HRs ranged from 0.63-2.6 [these span 0 as different reference groups are used]; low-moderate quality).
Seven studies (n=1821 patients) provided low to moderate quality evidence that overall survival is either similar between patients with and without MYC translocation while 4 studies (n=1066) provided low to moderate quality evidence that overall survival was inferior in patients with MYC translocation (reported HRs ranged from 1.68-4.87). Progression-free survival (9 studies; n = 1967; low-moderate quality) does not differ between patients with and without MYC translocation (as assessed by FISH), although one additional study found inferior progression-free survival in patients with MYC translocation (Kojima, 2013; n = 100; HR = 2.717; unclear quality).
No evidence was found for the following comparisons:
Low to moderate quality evidence from nine studies (n=2139 patients) reported no difference in overall survival and from eight studies (n=1771 patients) reported no difference in progression-free survival between patients with and without BCL2 translocation (as assessed by FISH), although one additional study may have found inferior overall survival in patients with BCL2 translocation (Horn, Ziepert, Bart et al., 2013; n = 112; unclear quality).
Low to moderate quality evidence from seven studies (n=1982 patients) showed no difference in overall survival while low to moderate quality evidence from four studies (n=1247 patients) showed no difference in progression-free survival between patients with and without BCL6 translocation (as assessed by FISH).
One study reported that the turnaround time of the GEP testing strategy employed was less than 1 day and repeated testing of up to 40 patients in parallel was possible (Rumimy, 2013; n = 141; unclear quality).
No studies were identified that reported health-related quality of life.
A literature review of published cost-effectiveness analyses did not identify any relevant papers for this topic. Whilst there were potential cost implications of making recommendations in this area, other questions in the guideline were agreed as higher priorities for economic evaluation. Consequently no further economic modelling was undertaken for this question.
| Recommendations | Do not use immunohistochemistry to assess the prognostic value associated with cell of origin in people with diffuse large B-cell lymphoma. Interpret FISH results (MYC, BCL2 and BCL6 rearrangements) in the context of other prognostic factors (particularly the person's age and International Prognostic Index [IPI]). Explain FISH results and their potential prognostic value to people with B-cell lymphoma. |
|---|---|
| Relative value placed on the outcomes considered | The GC considered overall survival (OS) to be the critical outcome when drafting the recommendations as OS and progression free survival (PFS) are closely aligned in diffuse large B cell lymphoma (DLBCL) with only a small number of relapsing patients being cured by salvage therapy. Health related quality of life and turnaround time for the test were also important but no comparative evidence was identified for these outcomes. |
| Quality of the evidence | The quality of the evidence about overall and progression free survival, assessed using the NICE checklist for prognostic studies, varied from low to moderate quality. There was a high degree of overlap in the populations used by the included studies resulting in an over-estimation of population sizes for each comparison. As a result the GC decided to treat the gene expression profiling (GEP) evidence with caution. The GC noted that the evidence suggested the adverse prognostic impact of MYC translocations may be modified by age and IPI, and could be difficult to interpret. The GC were aware of studies looking at outcomes of people with double hit lymphomas treated with modern as well as experimental chemotherapy arms which suggested the negative impact of these abnormalities is reduced by patient age, such that younger patients may not experience the same adverse outcomes when these genetic abnormalities are present. Approximately half of the included studies did not control separately for the effect of age on test results, which may have confounded the results. |
| Trade off between clinical benefits and harms | A strong recommendation was made not to use immunohistochemistry to assess the prognostic value associated with cell of origin in people with DLBCL on the basis of a large body of moderate quality evidence showing overall and progression free survival did not consistently differ between GCB and non-GCB or ABC subtypes identified using immunohistochemistry. The GC concluded that the survival difference seen in GEP-based studies between ABC and GCB groups is not consistently replicated in immunohistochemistry studies. The GC also considered, based on their clinical experience, that immunohistochemical tests are associated with insufficient reliability and reproducibility, which limits its use as a biomarker. The GC were unable to make any recommendations for GEP because although this is a highly effective technique with consistent results across the major studies the technology and analytical methods are rapidly changing at the present time. The GC noted that although the various systems worked in research settings they were not yet robust enough to be used in routine practice. Based on their clinical experience, the GC decided to recommend that results from FISH should be an integral part of patient information and should be interpreted in the context of other prognostic factors specifically age and IPI because the adverse prognostic effect of MYC rearrangements is much smaller in younger patients with lower IPI compared to older patients and patients with high IPI. Overall the GC considered the benefits to these recommendations are improved diagnostic accuracy and prognostic stratification, which will, in turn, improve the patient outcomes and experience. The GC identified no additional associated harms because the recommendations refer to further analyses conducted on samples already taken from the patient. |
| Trade off between net health benefits and resource use | No health economic evidence was identified and no health economic model was built for this topic. The recommendation will increase the number of FISH tests performed. Although this will be counteracted by some reduction in the use of immunocytochemistry there is likely to be a net increase in cost overall. However, the GC thought that this increased cost would be justified by an improvement in diagnostic accuracy and prognostic stratification, which should also lead to improvements in patient outcomes making the recommendations potentially cost effective. |
| Other considerations | The GC acknowledged these recommendations will require a more systematic approach to the investigation of DLBCL to be implemented in all centres. The extent of change will vary according to current practice. |
All rights reserved. NICE copyright material can be downloaded for private research and study, and may be reproduced for educational and not-for-profit purposes. No reproduction by or for commercial organisations, or for commercial purposes, is allowed without the written permission of NICE.
Your browsing activity is empty.
Activity recording is turned off.
See more...
