Banff TCMR Working Group

Banff TCMR Working Group: Issues for Discussion

The TCMR Working Group has been reviewing both the current Banff lesion-scoring rules and the way individual lesion scores are integrated into diagnostic categories.

Ahead of the next Banff meeting, we would like to invite the community to discuss several issues that may warrant clarification or modification. Rather than simply posing questions, we provide below in the attachment some background and the main arguments that have emerged from Working Group discussions.

We welcome your comments, experience from routine practice, and alternative proposals.

Item 1 – Scoring of ti, i and i-IFTA

Background

The current Banff rules do not explicitly state whether inflammatory cells located within peritubular capillaries (PTCs) should be included or excluded when assessing ti, i and i-IFTA, which are intended to reflect interstitial inflammation.

Distinguishing intravascular from interstitial inflammatory cells may be straightforward in mildly inflamed areas, but can become difficult in areas of dense inflammation, particularly when capillary boundaries or basement membranes are poorly preserved.

One possible approach is to score PTC inflammation only in areas away from dense interstitial inflammation. However, this may result in excluding areas where interactions between antibodies, inflammatory cells and the microvascular endothelium are most prominent.

Questions for discussion

• When assessing ti, i and i-IFTA, do you include or exclude inflammatory cells located within PTCs?

• How do you distinguish intravascular from interstitial inflammation in areas of dense inflammation?

• When scoring ptc, do you restrict your assessment to areas away from dense inflammation?

• Would more explicit Banff guidance on compartment-specific inflammatory scoring improve reproducibility?

Item 2 – Definition and scoring of tubular atrophy

·         Denominator for ct scoring

The Working Group supports using the total cortical area as the denominator when grading tubular atrophy as ct1, ct2 or ct3.

The current online Banff Reference Guide specifies the area occupied by cortical tubules as the denominator. Using total cortical area may provide a simpler and more reproducible approach and should therefore be discussed and clarified.

·         Defining tubular atrophy in the setting of acute injury

Tubular size alone may be misleading in biopsies with acute tubular injury, acute rejection or interstitial edema. In these settings, small or compressed tubules should not necessarily be interpreted as atrophic.

The Working Group therefore suggests that tubular atrophy should preferably be supported by structural evidence of chronic injury, particularly thickened and/or wrinkled tubular basement membranes, rather than tubular diameter alone.

The presence of acute injury and edema may also make the assessment of interstitial fibrosis (ci) difficult. One possible approach would be to allow a designation such as ci-ND (not determined) when fibrosis cannot be reliably assessed.

Questions for discussion

• Should the total cortical area formally replace cortical tubular area as the denominator for ct scoring?

• Should Banff provide more explicit morphological criteria for tubular atrophy in the setting of acute injury or rejection?

• Would a category such as ci-ND be useful when acute edema or inflammation prevents reliable assessment of fibrosis?

• 
  

Item 3 – Banff Category 4: Should “acute TCMR” be renamed “active TCMR”?

The Working Group has discussed whether the term acute TCMR should be replaced by active TCMR, analogous to the terminology adopted for ABMR several years ago. A similar approach has recently been adopted in the Banff pancreas classification.

Arguments supporting the term “active TCMR”

The designation acute is currently based primarily on the presence of inflammation and tubulitis in areas without IFTA. However, this histological pattern does not necessarily correspond to an acute clinical presentation.

For example, lesions currently classified as acute TCMR may be detected:

• on protocol biopsies in patients with stable kidney function or in biopsies performed for indications such as proteinuria rather than acute graft dysfunction.

• in biopsies showing concurrent i-IFTA and t-IFTA, with only a slowly progressive increase in serum creatinine;

Moreover, the distinction between acute and chronic active TCMR may not reflect a true biological boundary. Persistent or unresolved active TCMR may progressively lead to chronic active TCMR, whereas the current classification requires discrete thresholds for ti, i-IFTA and t-IFTA. These lesions may develop at different rates, making some biopsies difficult to classify.

Using the broader term active TCMR, while reporting the individual acute and chronic Banff lesion scores, might better reflect a biological continuum rather than impose an artificial temporal distinction.

In the future, this approach could potentially be complemented by TCMR activity and chronicity indices, analogous to approaches being developed for ABMR.

Arguments for retaining “acute TCMR”

The distinction between acute and chronic active TCMR is familiar to clinicians and pathologists and is currently used in clinical decision-making. In some centers, the diagnosis of acute TCMR directly influences whether anti-rejection treatment is initiated.

Removing this terminology could therefore have unintended clinical consequences unless any new framework is clearly defined and validated.

Questions for discussion

• Should acute TCMR be renamed active TCMR?

• Does the distinction between acute and chronic active TCMR reflect biologically distinct entities, or primarily different positions along a continuum of injury?

• Would it be preferable to diagnose active TCMR, report all relevant acute and chronic lesion scores (i, ti, i-IFTA, t, t-IFTA, ci and ct), and complement them with global activity and chronicity indices?

• If so, should diagnostic subcategories  (Borderline, Aute TCMR/ CA YCMR) still be retained alongside these indices?

• What would be the potential impact of such a change on clinical decision-making and treatment?

Item 4 – Banff Category 5: Should the IFTA category be retained?

Arguments for removing Category 5

IFTA is a morphological pattern rather than a specific diagnosis.

The information conveyed by the current IFTA grade is already largely captured by the individual ci and ct scores.

Moreover, the current IFTA grading system may inadequately represent biopsies in which interstitial fibrosis and tubular atrophy are markedly discordant. For example, a biopsy with ci1/ct3 may be assigned IFTA grade 3 despite having only mild interstitial fibrosis.

Eliminating the separate IFTA category could therefore simplify the classification while preserving more precise information through the individual ci and ct scores.

Arguments for retaining a category for non-specific chronic injury

A practical issue remains: where should biopsies in which IFTA is the predominant finding, but no specific disease or rejection process can be identified, be classified?

Several possibilities could be considered.

One option would be to place such biopsies within Category 1, while distinguishing truly normal biopsies from those showing non-specific abnormalities:

• Category 1A: Normal – no significant acute or chronic lesions.

• Category 1B: Non-specific changes – potentially including IFTA without an identifiable specific etiology.

Another possibility would be to place these biopsies in Category 6, particularly when chronic injury may be related to previous or ongoing rejection.

Questions for discussion

• Does IFTA warrant its own Banff diagnostic category?

• Is reporting ci and ct separately sufficient to describe the extent of chronic injury?

• If Category 5 were removed, where should biopsies showing predominantly non-specific IFTA be classified?

• Would separating Category 1 into “Normal” and “Non-specific changes” be useful?

• Should presumed rejection-related chronic scarring be distinguished from IFTA of uncertain or non-alloimmune origin?

Background Information:

DISCUSSION Record of the Banff TCMR Working Group Members:

MARK HAAS

Some excellent points here – not nit-picking at all. Just 2 comments. With regard to active TCMR superimposed on chronic active TCMR, I tend to make this diagnosis when the biopsy meets criteria for both (e.g., i2t2 plus ti2 i-IFTA2 and t2 and/or t-IFTA2), although I do this mainly at the request of our clinicians who are sometimes reluctant to treat chronic active TCMR in light of Vanderlene’s study (KI 2021) that >70% of such lesions do not respond to steroids +/- thymo, especially grade 1B chronic active TCMR.  Yet, there was no evidence in that study (albeit based on a limited number of cases) that patients with a biopsy showing chronic active TCMR with an “active” component responded any better than those without such a component.  It would, however, be helpful to further resolve this moving forward.Parmjeet and others have long stressed that peritubular capillaritis is frequently seen in biopsies showing TCMR that do not meet Banff criteria for AMR or even MVI, DSA- and C4d-negative (i.e., without glomerulitis).   While the argument might be made that the leukocytes within the ptcs in such cases are simply traveling through these capillaries in response to chemotactic factors in order to reach the tubulo-interstitium, molecular studies (e.g., Ivy Rosales’ JASN paper from 2022 or 2023) indicate that outcomes of grafts with histologically “pure” TCMR (i.e., not meeting AMR criteria) are affected by transcripts associated with AMR (or perhaps more accurately MVI) as well as TCMR, suggesting that these cells within the ptcs are simply not just “passing through”.  The linear association of the activity index with outcomes in TCMR cases likewise is consistent with this interpretation, suggesting that biopsies with g0i3t3ptc3 (activity index 9) have a worse prognosis than those with g0i3t3ptc1 (activity index 7).  We are currently doing a study of the impact of activity and chronicity indices on therapeutic responses in cases of TCMR (including borderline, 1a, and 1b active TCMR plus 1a and 1b chronic active TCMR; C4d+, DSA+, and v>0 excluded) and it will be interesting to compare cases with similar (i + t) but different ptc scores.  Hoping to at least have some preliminary findings to include in my talk at Banff.

ILKKA 

As a clinician (and also when reading research papers), I have always found this IFTA-1, IFTA-2, IFTA-3 categorization very confusing and more or less useless. I would much rather just see the ci and ct scores separately, IFTA-2 or IFTA-3 makes me always just wonder what the individual ci and ct scores actually are.. So in my opinion category 5 is unnecessary; a list /table of the individual lesion scores tells the whole picture much more accurately.

Michael MENGEL

I really think TCMR Banff rules need a major overhaul, now with MVI/AMR being potentially suppressed with anti CD38 target treatments.Also the recent paper from Maarten's group highlights that the cellular component of rejection presents in a new light and Banff needs to adjust to that.On top the rules around scoring i, t, ci, ct, i-IFTA, t-IFTA need to be clarified as Parmjeet outlines - but that should happen in a formal data/ AI image analysis driven manner.We should raise this at this Banff meeting and make it a main topic for the 2028 meeting.A white paper coming from the 2026 meeting outlining the problem and the roadmap towards 2028, and then a revised Banff classification after the 2028 meeting = a major TCMR overhaul for the 20th Banff meeting = 40 years after inception......I fully agree that these are the key areas which need to be addressed and you outline a roadmap for specific research to close the existing gaps.I also agree that this should results in a white paper outlining the gaps and making recommendations on what needs to be done to close the knowledge gaps and generate data which can then lead to up-dates of the Banff classification in a data driven approach.I think the timing is right. With targeted treatments towards AMR, we will get a clearer picture of the T cell component of rejection and the Banff rules need to adjust with MVI/AMR being potentially better controlled.Similar to the early 2000's when TCMR was better controlled and the AMR phenotypes emerged in the biopsies. Do we need IFTA as a category with the prospect of adopting the chronicity score?The main reason for the IFTA category was to have something after the term CAN was abandoned. While CAN included chronic rejection, IFTA was supposed to be everything left after you have ruled out specific diagnostic entities as the cause for chronic damage. It was the first step toward precision diagnostics in the Banff classification. It was considered needed since IFTA is a strong correlate with function and prognosis.However, the chronicity and activity indices seem to outperform IFTA scores in that regard.Hence, we can have a discussion about the future of IFTA. We can still describe the extent of ci and ct, but as Candice points out we need a category in the classification for those cases where no specific entity can be diagnosed, but still the biopsy looks not really normal.Having said that, maybe we think in all cases a diagnosis can be rendered and thus IFTA as a category becomes obsolete?

Marion Rabant

Regarding the IFTA category and its potential replacement by a chronicity score, it should definitively be discussed (and even pooled ?) at the upcoming Banff Meeting. I have just the fear that only the total of the scores would be given and it may therefore become difficult to understand how much of a given chronicity score is actually driven by IFTA versus other chronic lesions. I think we should be careful and insist that the individual components of the score should be noted and not only the total score. A useful parallel may be the modified NIH indices used in lupus nephritis, with separate activity and chronicity indices. Although a global score is provided, the individual lesion scores contributing to the total are also important for its interpretation.

SURYA Seshan

I totally agree with all you both have said. It is now time to readjust the TCMR side with added nuances and more exact terms and definitions of descriptive features.

CANDICE Roufosse

To answer Parmjeet's question specifically about IFTA category: I have also always wondered why this was a separate category and when trying to understand why people before us created it, have thought the most likely reason was that it is one of the most common findings in biopsies and its etiology could be just about anything. At least in my practice, biopsies where IFTA-NOS > 10% is the main finding are quite common (often accompanied by ATI category 6). In contrast I hardly ever use category 1 "normal", because there is almost always something wrong in a biopsy (perhaps centres with protocol biopsies use this more often).If we remove category 5, then what category do you attribute cases with just IFTA-NOS and no other specific findings to?We could put them in Category 1 as "non-specific changes" but that would be quite a big change as currently category 1 requires exclusion of IFTA/category 5. Centres with protocol or t0 biopsies might not like that. Category 6 is "Other Changes Not Considered To Be Caused By Acute Or Chronic Rejection" so we can't put it there as who knows if IFTA was caused by rejection or not? 

Agnes FOGO

Thanks to Parmjeet for excellent  summary and leadership with these complex areas. I think a total chronicity score is a good idea- IFTA does not include glomerulosclerosis or  severe vascular sclerosis-I think an additional important point is the integration with clinical situation, as emphasized in the most recent Banff  meeting update  paper. It seems that cutoffs for suggesting  rejection could vary in protocol biopsies with stable creatinine, vs  in indication biopsies- The vast borderline category might be parsed a bit better based on clinical setting - as many of you  know, we still use CCTT criteria in our practice to  indicate aTCMR in  indication biopsies.

Ian Gibson

Agree that the chronicity score is progress beyond IFTA only for prognosis. But for chronic glomerular pathology, chronicity score formula only takes account of Banff cg score. The lupus chronicity score does factor in global GS and FSGS. We need to factor in chronic glomerular pathology of all types for a prognostically meaningful allograft chronicity score. It needs to incorporate global GS, FSGS, chronic ischemic glomerulopathy and atubular glomeruli / glomerular cysts. A few of my slides below from a presentation at an USCAP / RPS meeting a few years ago made this point.

Mark Haas

Ian raises a good point here regarding the chronicity score.  In this regard, it is worth noting that this score was initially developed in the specific context of AMR (Kidney Int 2023; 103: 187-195) and thus its focus on cg to the exclusion of other chronic glomerular changes (global and segmental glomerulosclerosis), although the data from the study of Vaulet, Naesens and coworkers (Nature Communications 2025; 16: 10417) did nicely demonstrate its applicability to other rejection and rejection-like (e.g., MVI, DSA- and C4d-negative) lesions.  Still, to extend the applicability of the chronicity score to lesions not related to rejection it may be beneficial to test whether inclusion of additional lesions such as those noted by Ian improve on its overall predictability .  Also, others have used the Mayo chronicity score (Kidney Int 2017; 91: 787–789) or the closely related Boston chronicity score ( J Am Soc Nephrol 2018; 29: 2213–2224) for scoring chronic changes, albeit much more in native kidney biopsies than transplant biopsies.This might be an interesting topic for further discussion at Banff.

Lynn Cornell

Thank you, Parmjeet and others, for the summary of many of the issues we are facing now in transplant pathology. I could comment on everything, but for now, I’d like to comment on the Banff diagnostic categories, at the risk of expanding this conversation way beyond the scope of the TCMR working group. Rather than considering only the existing category 5, I think this is an opportunity to rethink the overall structure and purpose of the Banff diagnostic categories.These Banff categories over the years have incorporated our evolving understanding of diseases affecting the kidney transplant, but also are a bit of a “fossil” in that they preserve the older categories that, in my opinion, we should move on from (e.g., CAN evolving into IFTA NOS). The existing categories are also not all "diagnostic” categories, in the sense that in renal pathology we aim to assign an etiologic or attributive diagnosis. Instead, the current Banff categories are a mishmash of different types of concepts: etiologic diagnoses (TCMR, ABMR), a diagnostic threshold pattern (borderline), a chronic injury pattern (CAN/IFTA NOS), and a catch-all (category 6).I think diagnostic categories are useful to have for teaching and conceptual purposes, while recognizing limitations such as that histologic lesions are not necessarily specific (e.g., transplant arteriopathy) and that some patterns we do not always know how to categorize etiologically (e.g., DSA- C4d- MVI pattern).If I were designing Banff diagnostic categories now in 2026, I would suggest:1.      Normal biopsy or no diagnostically significant changes2.      Antibody-associated injuryo       Including AMR as currently defined as well as related histologic injury patterns like DSA- C4d- MVI3.      T cell-mediated rejection•       Including borderline changes, active (or acute) TCMR, and chronic TCMR4.      Infection (e.g., BK polyomavirus)5.      Other specific diseases and attributed injuries, not transplant-specifico       Including recurrent, de novo, or donor-derived glomerulonephritis and tubulointerstitial diseases (not due to rejection or infection); drug toxicity; acute tubular injury; vascular disease (not due to rejection); and PTLD6.      Morphologic patterns of allograft injury of uncertain or unattributed etiologyo       Including IFTA-NOS, i-IFTA of uncertain etiology (e.g., differential diagnosis may include chronic rejection, infection, and residual inflammation following an infection), and sub-threshold lesions not able to be placed in another categoryI would prefer a term like “antibody-associated injury” rather than antibody-mediated rejection, as I think calling some of these manifestations antibody-mediated overstates our understanding of the mechanism of several of the manifestations of this type of rejection.Category 5 above is essentially “native kidney diseases” where there may be existing classifications for the native kidney, but we need to recognize that in the transplant these diseases may have different histologic and immunophenotypic appearances, and may have different clinical presentations and outcomes.I would separate out infection as a separate category (above, category 4), rather than including it in category 5 above, since several infections are relatively transplant-specific and because infection can mimic rejection.SeshanI agree with all, that the total number of glomeruli available, the absolute or percentage that are viable/preserved or sclerosed (GS) is an important prognostic indicator.This point was discussed at the AI-CI WG meetings. One other caveat that was discussed at the time was whether a portion or all of the GS was located in a subcapsular scar or another larger cortical scar. This needs to be mentioned. As in native kidney biopsies, I have always included these numbers in the diagnostic line for the clinician. As Mark mentioned, other resources where global sclerosis has been used as a chronicity indicator, in addition to the Lupus AI-CI that Ian mentioned.