SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Banff Process, Lesions and Classification
  4. The Beginning
  5. Evolution
  6. Strengths
  7. Weaknesses of Banff Classification
  8. Opportunities: The ‘Omics’ Approaches
  9. Threats: Avoid Splitting and Stay Relevant
  10. Acknowledgment
  11. References

The Banff process defined the diagnostic histologic lesions for renal allograft rejection and created a standardized classification system where none had existed. By correcting this deficit the process had universal impact on clinical practice and clinical and basic research. All trials of new drugs since the early 1990s benefited, because the Banff classification of lesions permitted the end point of biopsy-proven rejection. The Banff process has strengths, weaknesses, opportunities and threats (SWOT). The strength is its self-organizing group structure to create consensus. Consensus does not mean correctness: defining consensus is essential if a widely held view is to be proved wrong. The weaknesses of the Banff process are the absence of an independent external standard to test the classification; and its almost exclusive reliance on histopathology, which has inherent limitations in intra- and interobserver reproducibility, particularly at the interface between borderline and rejection, is exactly where clinicians demand precision. The opportunity lies in the new technology such as transcriptomics, which can form an external standard and can be incorporated into a new classification combining the elegance of histopathology and the objectivity of transcriptomics. The threat is the degree to which the renal transplant community will participate in and support this process.


Banff Process, Lesions and Classification

  1. Top of page
  2. Abstract
  3. Banff Process, Lesions and Classification
  4. The Beginning
  5. Evolution
  6. Strengths
  7. Weaknesses of Banff Classification
  8. Opportunities: The ‘Omics’ Approaches
  9. Threats: Avoid Splitting and Stay Relevant
  10. Acknowledgment
  11. References

The Banff consensus process is a moderated self-organizing group to define and standardize organ allograft biopsy interpretation. The goal of the process is to promote international uniformity in the reporting of allograft pathology, which is needed for research, clinical trials and patient management. The consensus process engages in an ongoing dialectic among clinicians, pathologists and basic scientists, highlighted by a meeting every 2 years in which the classification guidelines are reviewed and revised in the light of emerging information. The Banff system provides criteria for classifying various forms of rejection as well as other pathologies such as drug toxicity (1–3).

Starting with a historical overview, this paper presents an informal ‘SWOT’ analysis (Strengths, Weaknesses, Opportunities, and Threats) of the current Banff process and classification. The aim is to improve clinicians' understanding of the Banff classification when using pathology reports to make treatment decisions, conduct clinical trials and perform research. The paper focuses on the criteria for T-cell mediated rejection (TCMR), which remains a major therapeutic challenge, a risk factor for outcomes, and a focus of drug development.

The Beginning

  1. Top of page
  2. Abstract
  3. Banff Process, Lesions and Classification
  4. The Beginning
  5. Evolution
  6. Strengths
  7. Weaknesses of Banff Classification
  8. Opportunities: The ‘Omics’ Approaches
  9. Threats: Avoid Splitting and Stay Relevant
  10. Acknowledgment
  11. References

Inspired by efforts in heart and lung transplantation (4) and the need to standardize renal biopsy interpretation, a small group met in August 1991 in Banff, Alberta. Led by Kim Solez, Lorraine Racusen and Philip Halloran, this meeting established the spirit of the Banff process: it was interdisciplinary (pathologists, nephrologists, surgeons, tissue typing physicians) and international (Canada, USA, Denmark, Finland, Australia). The task was to construct a classification system for allograft pathology (5). The initial working title was ‘Workshop on International Standardization of the Nomenclature and Criteria for the Histologic Diagnosis of Kidney Transplant Rejection’, a title soon simply called ‘Banff’. Since 1991, a Banff meeting has occurred every 2 years, initially in Banff and later in other places, satisfying its international spirit.

While the classification is in continuous evolution, the fundamental criteria for rejection (now recognized as TCMR after antibody-mediated rejection (ABMR) was defined) developed at the first meeting remain valid. To diagnose TCMR, the Banff classification defines the key lesions in different compartments of kidney and outlines how to semi-quantitatively record these lesions. The lesion scores (g-, i-, t-, v-score) are summarized into a final Banff TCMR grade (1,5) and empirically defined thresholds of scores and grades decide whether a case is designated TCMR. The histologic lesions of TCMR are (5): interstitial inflammation (i-score); tubulitis (t-score = invasion of tubular epithelium by lymphocytes) and intimal arteritis (v-score = invasion of inflammatory cells beneath the endothelium), whereas increasing quantities of each lesion reflects more severe rejection.

However, as early as the late 1970s, it was known that the tubulitis and interstitial inflammation are not specific for TCMR and can be observed in acute tubular necrosis (6), posttransplant lymphoproliferative disorder, in biopsies from stably functioning allografts, and in infectious or drug-induced interstitial nephritis (5). Because of concerns with over-diagnosing and over-treating TCMR, the founders introduced minimum thresholds for the i- and t-score to diagnose TCMR (1,5). In the absence of intimal arteritis, at least 25% of the cortex with interstitial infiltrate and simultaneously at least moderate tubulitis (t2 = 5–10 mononuclear cells/tubular cross section) must be found to diagnose TCMR. Lesser degrees of interstitial infiltrates plus tubulitis qualifies for the so-called borderline category (suspicious for rejection).

In contrast, intimal arteritis has long been regarded as pathogonomic for rejection (1,7), almost as a dogma. Interestingly, a single lymphocyte under the endothelium of an artery is enough to make the diagnosis of rejection independent of interstitial inflammation, but severe tubulitis without significant interstitial infiltrates does not qualify as rejection. This seems biologically implausible and needs reassessment.

In 1991 when Banff criteria for TCMR were first described, immunosuppression was less advanced, and renal biopsies displayed more frequent and severe TCMR than today. Hence, the Banff thresholds were more sensitive for TCMR because the frequency and severity of the disease was higher. With the advent of improved immunosuppressive agents, however, the frequency and severity of TCMR lessened, reducing the sensitivity and probably also the specificity of the criteria, and increasing the frequency of borderline changes, which are still of uncertain clinical relevance. Moreover, data at that time were mostly based on findings in biopsies for cause and not protocol biopsies. Even today, functional deterioration of a graft is a crucial element of the diagnosis of TCMR: in most centers diagnosis of TCMR is based on evaluating the extent of lesions in the clinical context (8). Notably, even the first Banff paper posed the question of whether ‘subclinical’ lesions or borderline changes in biopsies without dysfunction should be regarded as rejection (5), a question that remains open (9).

Evolution

  1. Top of page
  2. Abstract
  3. Banff Process, Lesions and Classification
  4. The Beginning
  5. Evolution
  6. Strengths
  7. Weaknesses of Banff Classification
  8. Opportunities: The ‘Omics’ Approaches
  9. Threats: Avoid Splitting and Stay Relevant
  10. Acknowledgment
  11. References

During the 1993 and 1995 meetings, the first studies applying the new criteria validated the 1991 decisions. It was agreed that no changes should be made to the criteria unless supported by acceptable studies from two independent groups. Meanwhile, an independent, NIH-sponsored group—the Cooperative Clinical Trials in Renal Transplantation (CCTT)—was working under the guidance of Robert Colvin on diagnostic criteria (10). Hence, at the 1997 Banff meeting, a consensus between both groups was worked out and jointly published in 1999 as the revised Banff'97 classification (1). Since the same lesions (i, t, v) were used to diagnose rejection in both Banff and CCTT (7), they were retained and all centers use these same lesions for classification. However, major differences between the two classifications are as follows: the CCTT differentiates three types of rejection (tubulo-interstitial, vascular and humoral) and does not grade rejection. In addition, with CCTT most cases of the Banff borderline category would be classified as CCTT type I rejection, and thus CCTT has a lower threshold for rejection than the Banff classification.

A consensus between Banff and CCTT was elaborated with aspects of the CCTT system being incorporated (1). The Banff borderline category survived the merger. The revised Banff'97 classification also includes detailed technical guidelines and minimal adequacy of a specimen to be sufficient for reliable review to minimize the influence of sampling errors. Because tubulitis in atrophic tubules is a frequent finding, and regarded as nonspecific, the Banff'97 paper introduced the rule that tubulitis can only be scored in nonatrophic or mildly atrophic tubules (1). Also, inflammation in the immediate subcapsular cortex, fibrotic areas and the adventitia of large vessels was considered to be nonspecific and was therefore excluded from the i-score (1).

Features of ABMR were added to the schema in 2001 and 2005, introducing immunohistochemistry (i.e. C4d) to the previous light microscopic classification (3,11). Thus, what had previously been called ‘rejection’ was now designated TCMR. More recently, the concept of ‘interstitial fibrosis and tubular atrophy not otherwise specified’ (IFTANOS) was introduced to permit discrimination of nonspecific scarring and atrophy from specific entities of allograft damage (3), like transplant glomerulopathy (TG) and other consequences of slow ABMR. This eliminated the imprecise term chronic allograft nephropathy (CAN), which had unfortunately lumped IFTANOS and all other specific entities of chronic allograft damage (e.g. TG) together.

Application of the classification to protocol biopsies of ‘stable’ kidneys is an ongoing and controversial process (12,13). The main finding in protocol biopsies is that the lesions of TCMR can be observed in allografts with no significant change in function, and thus do not necessarily have a fixed relationship to the immunologic mechanisms of TCMR. How the criteria designed for grafts with dysfunction should be applied to protocol biopsies remains to be clarified, and is a topic at the 2007 Banff meeting. This illustrates the Banff process, addressing unsolved issues to create new consensus in a classical dialectic of thesis, antithesis and synthesis (Figure 1).

image

Figure 1. The Banff process. A self-organizing consensus communication in renal transplantation.

Download figure to PowerPoint

Strengths

  1. Top of page
  2. Abstract
  3. Banff Process, Lesions and Classification
  4. The Beginning
  5. Evolution
  6. Strengths
  7. Weaknesses of Banff Classification
  8. Opportunities: The ‘Omics’ Approaches
  9. Threats: Avoid Splitting and Stay Relevant
  10. Acknowledgment
  11. References

The Banff process is one of the key engines of progress in renal transplantation, applied to basic and clinical research, virtually all drug trials, and in routine clinical care. The clinical relevance of the Banff classification has been shown by independent groups (14–16). Banff lesions respond to treatment, and pure tubulointerstitial TCMR responds better to treatment than cases with intimal arteritis, which has a worse prognosis and often requires more aggressive treatment. Severe tubulitis (t3) has a worse prognosis than mild tubulitis (t1) (14). In addition, the internationally standardized grading of the lesions has provided objectivity for publication, data sharing and statistical analysis—all of which are essential prerequisites for clinical trials. Most Banff meeting participants report that they alter clinical practice as a direct consequence of the meeting (2).

The Banff lesions have been confirmed in animal models of rejection, allowing analysis of basic mechanisms of TCMR (17). Thus, in a mouse kidney transplant model, the lesions of TCMR have been shown to be T-cell dependent but independent of granzymes and perforin, as well as being independent of B cells and antibody (17,18). It seems that the tubulitis reflects epithelial deterioration secondary to a delayed-type hypersensitivity reaction in the interstitium. Changes in the epithelium such as loss of cadherins develop before apparent direct contact with effector T cells (19). The changes in the epithelium are part of a massive alteration in gene expression, including reexpression of embryonic genes and loss of many solute carriers and metabolism genes (20), that is a stereotyped injury response evoked by the interstitial inflammation.

An essential strength of the Banff process is its consensus approach (Figure 1). The willingness and capacity of the process to continuously adapt in response to new research ensures that it will remain relevant, incorporating new sciences as they emerge. Nobody owns Banff: it has been a moderated self-organizing system from the beginning. It is not the classification of ‘Dr. Banff’, nor of one center, but of the community. The moderator, Kim Solez, supervises the process but does not direct or drive the outcome. Refinement of the classification has always been an open source, like today's ‘Linux’ or ‘Wikipedia’. Banff reflects the state of the art, which should be in motion forever, and no method (e.g. histology) is more important than other emerging approaches (e.g. transcriptomics) if they are superior. If well-conducted studies indicate that the current criteria should be changed, or that the new technology can add to established procedures, a process follows (Figure 1) (21): (1) the respective groups are invited to the next Banff meeting to present their results (only the speakers are invited, everybody else can present his data as poster presentations and participate at the meeting and discussion), (2) the results are discussed by all participants moderated by a session chair, and (3) outcome of this consensus discussion might be acceptance of changes by the group, request for further data and adjourn until the next meeting, or a decision not to change(usually revealed by voting). Following every meeting, a report describing its outcome should be published (2,3,21). This time-consuming procedure is not only needed to describe the rationale underlying the recent changes on the schema but also useful as being a continued platform to discuss controversial issues.

Weaknesses of Banff Classification

  1. Top of page
  2. Abstract
  3. Banff Process, Lesions and Classification
  4. The Beginning
  5. Evolution
  6. Strengths
  7. Weaknesses of Banff Classification
  8. Opportunities: The ‘Omics’ Approaches
  9. Threats: Avoid Splitting and Stay Relevant
  10. Acknowledgment
  11. References

The Banff classification of rejection possesses two main weaknesses (Table 1): poor reproducibility, and lack of independent validation (discussed in the next section). The continuing problem of borderline changes and the lack of mechanistic understanding are also noteworthy.

Table 1.  Banff process and classification of rejection issues in progress, still warranting further discussion
1. Reproducibility
2. Independent validation of thresholds and diagnostic features
 of rejection
3. Robust quantification and measurement of rejection/
 inflammation
4. Borderline category
5. Subclinical rejection/application to protocol biopsies
6. Integration of features with sole prognostic impact

Reproducibility is a major issue, as it is for all histopathological classification systems (22). Two types of reproducibility have to be considered: intra- and interobserver reproducibility. Both are assessed mathematically by Kappa statistics, which measure the percent agreement between two votes (15). Studies on Banff scoring (15,22–25) (Table 2) showed reproducibility as ‘moderate’ (Kappa: 0.41–0.60) and more often ‘fair’ (Kappa: 0.21–0.40). Discriminating between rejection/no rejection without scoring was better. Furness meticulously analyzed numerous parameters potentially influencing reproducibility of Banff classification (15,23). He circulated slides between European transplant pathologists and found alarmingly low Kappa values. Grading tubulitis was essentially not reproducible (Kappa value 0.17) (15). Earlier studies based on the 1991 Banff classification reporting better Kappa values were done by pathologists who were trained together and were involved in developing the classification (22,24). Pathologists participating in the Furness trial were never trained together and were truly testing the operation of the classification in practice. Lesions where the definition was based on ‘percent area involved’ (i-score) had worse reproducibility. When different pathologists review the same slide, much variation resulted from examining different areas of the slide and different interpretation of the rules.

Table 2.  Reproducibility of Banff process and classification scores measured by Kappa values
ReferenceSolez 1995 (24)Marcussen 1995 (22)Furness 2001 (15)Gough 2002 (25)Furness 2003 (23)Veronese 2005 (26)
  1. Kappa values: <0.20 = poor, 0.21–0.40 = fair, 0.41–0.60 = moderate, 0.61–0.80 = good, >0.80 = excellent.

  2. NA = not available.

  3. *Three observers were compared to one another.

Kappa value for 
 i-score0.330.330.340.420.35NA
 t-score0.350.350.170.480.21NA
 v-score0.470.470.350.500.38NA
 Rejection (yes/no)0.620.56NA0.77NA0.46–0.72*

However, in the Furness trials, the pathologists showed consistency in individual scores and diagnosis (excellent intraobserver reproducibility) (15). If all participants are aware of the Banff ‘wobble’, harm to patients should be minimal. In contrast, in the setting of multicentric trials, the Banff variability is a problem, and may account for different centers reporting different rejection rates despite comparable immunosuppression (15). Central pathological reevaluation of biopsies should be obligatory in multicenter drug trials, but this does not solve the problem: it is still just one opinion.

Borderline lesions remain an unsolved problem. Some studies conclude that the most borderline cases are indeed TCMR (27), but discrepant percentages (30–80%) of borderline cases have been reported to progress to TCMR without treatment (28,29). Such discrepancy may be explained by the ‘single cell cut off concept’ of a histologic classification: having sufficient interstitial inflammation detecting four or five lymphocytes in a tubule, discriminates between borderline and TCMR. Acknowledging such concerns, the classification insists that the therapeutic decisions in borderline cases must be made in the clinical context (1,5).

Opportunities: The ‘Omics’ Approaches

  1. Top of page
  2. Abstract
  3. Banff Process, Lesions and Classification
  4. The Beginning
  5. Evolution
  6. Strengths
  7. Weaknesses of Banff Classification
  8. Opportunities: The ‘Omics’ Approaches
  9. Threats: Avoid Splitting and Stay Relevant
  10. Acknowledgment
  11. References

The emerging ‘omic’ sciences—transcriptomics and later metabolomics and ultimately proteomics—offer opportunities for objective, quantitative, biology-based measurements that will improve on histopathology-based diagnoses. After new diagnostic criteria are accepted in principle, the choice of the platform will follow, and the technology must be validated in independent laboratories. Incorporation of omics technologies into diagnostics will require standardization of many issues, that is sample storage, RNA purification, RNA quality, normalization methods, etc.

Transcriptome measurements can be highly reproducible, because the assays can be internally controlled and normalized. We have data in our own laboratory showing high intralaboratory reproducibility (correlation coefficient of 0.98) of microarrays run on different days. Some variation between multiple cores is inevitable, but our experience to date is that this is much less than morphology variations. Such efforts will benefit from progress in the use of transcriptomics and microarray technology for other diagnostic applications such as cancer.

At that point, the results of the new technology as an external validation of the existing criteria, especially in problem areas, will be exciting and transforming, infusing new dynamism into clinical research and inviting basic researchers to examine mechanisms. The long process of understanding and interpreting these new insights will then follow, integrating the new methods into the existing histopathology criteria to create a true gold standard that is robustly measured and biologically rational.

Ultimately transcript measurements will be incorporated into the existing histology-based Banff classification. Thus, TCMR would be classified by a combination of morphology and objective measurement by transcripts representing infiltrating cells and parenchymal changes (30,31). Such numbers should be standardized so that they mean the same to clinicians in Prague or Pittsburg.

At the 2005 Banff meeting, the development of a ‘genomics’ supported Banff classification was already an important objective (3), and the participants concluded that gene expression analysis would be complementary rather than being competitive with conventional histopathology. This suggests greater interdisciplinary participation: clinicians, pathologists, ‘omics-specialists,’ bioinformaticians, with the development of mathematically based algorithms integrating clinical, histological and molecular features. However, a principle of the Banff process is that the classification should be applicable worldwide, including developing countries. Therefore, optimization of conventional histology will continue, guided by insights arising from molecular studies.

Threats: Avoid Splitting and Stay Relevant

  1. Top of page
  2. Abstract
  3. Banff Process, Lesions and Classification
  4. The Beginning
  5. Evolution
  6. Strengths
  7. Weaknesses of Banff Classification
  8. Opportunities: The ‘Omics’ Approaches
  9. Threats: Avoid Splitting and Stay Relevant
  10. Acknowledgment
  11. References

The Banff process can adapt indefinitely to evolving challenges and technologies, incorporating new findings with the best of the old, but only if the participants support the principle of bringing all opinions together in one process. By dialogue, merger and consensus, the Banff process has largely avoided emergence of competitive and contradictory systems, as it occurred with the European and North-American lymphoma classifications (32).

Priorities change, and the process must address these changes. The incidence and severity of TCMR has decreased with the advances in immunosuppression (most achieved by trials using the Banff classification). But long-term prognosis has improved less, underscoring the need to define mechanisms and pathogenesis of late graft deterioration and develop criteria that can guide clinical trials. For example, new therapeutic strategies for patients with mild or ‘nonspecific’ inflammation below the current Banff thresholds might be developed (33). The continuing definition of ABMR phenotypes also opens prospects for intervention studies. The basis for any new therapeutic strategies has to be reliably refined classification criteria.

As in the past, future refinement will be driven by the dynamic, self-organized Banff process (Figure 1). This may require that more biologists and bioinformaticians and ‘omics’ specialists join the Banff process alongside clinicians and pathologists. In 1991, the ultimate hope of the initiators of the Banff process was that this classification would lead to major improvements in patient care and management (5)—it has achieved this goal and continues to do so!

References

  1. Top of page
  2. Abstract
  3. Banff Process, Lesions and Classification
  4. The Beginning
  5. Evolution
  6. Strengths
  7. Weaknesses of Banff Classification
  8. Opportunities: The ‘Omics’ Approaches
  9. Threats: Avoid Splitting and Stay Relevant
  10. Acknowledgment
  11. References
  • 1
    Racusen LC, Solez K, Colvin RB et al. The Banff 97 working classification of renal allograft pathology. Kidney Int 1999; 55: 713723.
  • 2
    Racusen LC, Halloran PF, Solez K. Banff 2003 meeting report: New diagnostic insights and standards. Am J Transplant 2004; 4: 15621566.
  • 3
    Solez K, Colvin RB, Racusen LC et al. Banff '05 meeting report: Differential diagnosis of chronic allograft injury and elimination of chronic allograft nephropathy (‘CAN’). Am J Transplant 2007; 7: 518526.
  • 4
    Billingham ME, Cary NRB, Hammond ME et al. A working formulation for the standardization of nomenclature in the diagnosis of heart and lung rejection: Heart rejection study group. J Heart Transplant 1990; 9: 587593.
  • 5
    Solez K, Axelsen RA, Benediktsson H et al. International standardization of criteria for the histologic diagnosis of renal allograft rejection: The Banff working classification of kidney transplant pathology. Kidney Int 1993; 44: 411422.
  • 6
    Marcussen N, Lai R, Olsen TS, Solez K. Morphometric and immunohistochemical investigation of renal biopsies from patients with transplant ATN, native ATN, or acute graft rejection. Transplant Proc 1996; 28: 470476.
  • 7
    Dammin GJ. The kidney as a homograft and its host. Med Bull (Ann Arbor) 1960; 26: 278283.
  • 8
    Solez K, Racusen L, Rayner D, Olsen S, Halloran P. The Banff schema four years later. Transplant Proc 1996; 28: 450452.
  • 9
    Rush D, Nickerson P, Gough J et al. Beneficial effects of treatment of early subclinical rejection: A randomized study. J Am Soc Nephrol 1998; 9: 21292134.
  • 10
    Colvin RB, Cohen AH, Saiontz C et al. Evaluation of pathologic criteria for acute renal allograft rejection: Reproducibility, sensitivity, and clinical correlation. J Am Soc Nephrol 1997; 8: 19301941.
  • 11
    Racusen LC, Colvin RB, Solez K et al. Antibody-mediated rejection criteria – an addition to the Banff 97 classification of renal allograft rejection. Am J Transplant 2003; 3: 708714.
  • 12
    Salomon DR. Protocol biopsies should be part of the routine management of kidney transplant recipients. Con. Am J Kidney Dis 2002; 40: 674677.
  • 13
    Rush D. Protocol biopsies should be part of the routine management of kidney transplant recipients. Pro. Am J Kidney Dis 2002; 40: 671673.
  • 14
    Mueller A, Schnuelle P, Waldherr R, Van Der Woude FJ. Impact of the Banff '97 classification for histological diagnosis of rejection on clinical outcome and renal function parameters after kidney transplantation. Transplant 2000; 69: 11231127.
  • 15
    Furness PN, Taub N. International variation in the interpretation of renal transplant biopsies: Report of the CERTPAP Project. Kidney Int 2001; 60: 19982012.
  • 16
    Haas M, Kraus ES, Samaniego-Picota M, Racusen LC, Ni W, Eustace JA. Acute renal allograft rejection with intimal arteritis: Histologic predictors of response to therapy and graft survival. Kidney Int 2002; 61: 15161526.
  • 17
    Jabs WJ, Sedlmeyer A, Ramassar V et al. Heterogeneity in the evolution and mechanisms of the lesions of kidney allograft rejection in mice. Am J Transplant 2003; 3: 15011509.
  • 18
    Halloran PF, Urmson J, Ramassar V et al. Lesions of T-cell-mediated kidney allograft rejection in mice do not require perforin or granzymes A and B. Am J Transplant 2004; 4: 705712.
  • 19
    Einecke G, Fairhead T, Hidalgo LG et al. Tubulitis and epithelial cell alterations in mouse kidney transplant rejection are independent of CD103, perforin or granzymes A/B. Am J Transplant 2006; 6: 21092120.
  • 20
    Einecke G, Broderick G, Sis B, Halloran PF. Early loss of renal transcripts in kidney allografts: Relationship to the development of histologic lesions and alloimmune effector mechanisms. Am J Transplant 2007; 7: 11211130.
  • 21
    Solez K, Benediktsson H, Cavallo T et al. Report of the third Banff conference on allograft pathology (July 20–24, 1995) on classification and lesion scoring in renal allograft pathology. Transplant Proc 1996; 28: 441444.
  • 22
    Marcussen N, Olsen TS, Benediktsson H, Racusen L, Solez K. Reproducibility of the Banff classification of renal allograft pathology. Inter- and intraobserver variation. Transplant 1995; 60: 10831089.
  • 23
    Furness PN, Taub N, Assmann KJ et al. International variation in histologic grading is large, and persistent feedback does not improve reproducibility. Am J Surg Pathol 2003; 27: 805810.
  • 24
    Solez K, Hansen HE, Kornerup HJ et al. Clinical validation and reproducibility of the Banff schema for renal allograft pathology. Transplant Proc 1995; 27: 10091011.
  • 25
    Gough J, Rush D, Jeffery J et al. Reproducibility of the Banff schema in reporting protocol biopsies of stable renal allografts. Nephrol Dial Transplant 2002; 17: 10811084.
  • 26
    Veronese FV, Manfro RC, Roman FR et al. Reproducibility of the Banff classification in subclinical kidney transplant rejection. Clin Transplant 2005; 19: 518521.
  • 27
    Saad R, Gritsch HA, Shapiro R et al. Clinical significance of renal allograft biopsies with “borderline changes,” as defined in the Banff Schema. Transplant 1997; 64: 992995.
  • 28
    Meehan SM, Siegel CT, Aronson AJ et al. The relationship of untreated borderline infiltrates by the Banff criteria to acute rejection in renal allograft biopsies. J Am Soc Nephrol 1999; 10: 18061814.
  • 29
    Schweitzer EJ, Drachenberg CB, Anderson L et al. Significance of the Banff borderline biopsy. Am J Kidney Dis 1996; 28: 585588.
  • 30
    Einecke G, Melk A, Ramassar V et al. Expression of CTL associated transcripts precedes the development of tubulitis in T-cell mediated kidney graft rejection. Am J Transplant 2005; 5: 18271836.
  • 31
    Racusen LC. Molecular techniques in transplantation. Transplant Proc 2004; 36: 731732.
  • 32
    Harris NL, Jaffe ES, Diebold J, Flandrin G, Muller-Hermelink HK, Vardiman J. Lymphoma classification–from controversy to consensus: The R.E.A.L. and WHO classification of lymphoid neoplasms. Ann Oncol 2000; 11(Suppl 1): 310.
  • 33
    Mengel M, Gwinner W, Schwarz A et al. Infiltrates in protocol biopsies from renal allografts. Am J Transplant 2007; 7: 356365.