The Banff 2022 Kidney Meeting Report: Re-Appraisal of Microvascular Inflammation and the Role of Biopsy-Based Transcript Analysis

I have read with great interest the Banff 2022 reports. Congratulations and thanks to all the Banff team for their work! I wanted to take advantage of the offer to give my opinion as a clinician user.

- it is not clear to me if the 3 classical criteria will still be use (evidence of tissue injury, evidence of Ab interaction...) as they appear in the text but not in the table. Do you think we should avoid these definitions of criteria following this Banff update? If not, it may be of interest to include these terms in the table?

- The table 3 might be difficult to apprehend at the first glance. However, I think it removes doubts about certain elements.

I am wondering if reading difficulties could be linked to graphical details. For example the bold and underlined typeface: in the first column it refers to the criteria and in the second to the diagnosis, which appear above the the criteria DSA/C4d.

I have several thoughts that could perhaps improve readability:

- in the histomolecular description column, "MVI at or above threshold" appears as a title but this does not match with the molecular option. In the square bellow, "MVI below threshold" appears but bellow is detailed the first criteria. There is also the mention that molecular transcript may be used, which may be redundant with the "AND/OR molecular transcripts" in the square above. To improve readability, maybe the text could be limited to the criterion the user has to search? 

- first column, second square: histological active or chronic criteria appear bellow "MVI below threshold" but it is not really specific to this square and if the user follows the table, after the very first line of the table, we are already in a situation where these criteria are met. Maybe the detail definition of these lesion (e.g "g > 0 in the absence of GN") could appear in the footnotes? Also, to have a step-by-step thinking, these data could be placed in the last column where the classification between active/chronic active/chronic inactive appears?

- in the second column, first line of each square is the diagnosis, then criteria are listed. In the spirit of step-by-step approach evoked by the arrows in the column headings, maybe it would be easier to read with a distinct column "C4d and DSA" yes/no, then a third column with the diagnosis?

One more thought please:

Table S4, last row, causes of C4D negative, DSA negative, MVI:

T-cell and macrophage mediated injury should be mentioned before viral infection and ischemia reperfusion injury (perhaps immeidiately after NK cells). These 2 cell types are ALWAYS demonstrable in these biopsies, whereas the frequency and importance of non-HLA antibodies are a more nebulous concept which needs more studies.

Many astute clinicians consider treating MVI with steroids if there is graft dysfunction. Greater emphasis on the potential role of T-cells in mediating endothelial injury would be consistent with the proposed change.

I commend the Banff Leadership for implementing this policy of opening up Banff Meeting Reports for public comment prior to publication.

I commend the Banff Leadership for implementing this policy of opening up Banff Meeting Reports for public comment prior to publication.

Table S2. Requirement of at least 2 foci of tubulitis should be specified for t-IFTA as it has been for t-Banff

TABLE 5 ON MMDX I HAVE THE FOLLOWING SUGGESTIONS for the column marked MMDX

Current text:

“Strong and validated relation between the histologic picture of AMR and TCMR”

Suggested text:

 “Strong and validated relation with the histologic picture of AMR and a weaker relationship with TCMR”

Rationale:

As per supplementary Table 1: 89% of survey respondents agree that Biopsy based transcript diagnostics developed for TCMR primarily reflect interstitial inflammation. BKV nephropathy cannot be distinguished.  False positives  have been reported with pyelonephritis and granulomatous adverse drug reactions.

Halloran’s own discrepancy analysis paper shows that more than half of histologic TCMR diagnoses can’t be confirmed by MMDX

1.      Discrepancy analysis comparing molecular and histology diagnoses in kidney transplant biopsies.

Madill-Thomsen K, Perkowska-Ptasińska A, Böhmig GA, Eskandary F, Einecke G, Gupta G, Halloran PF; MMDx-Kidney Study Group.

Am J Transplant. 2020 May;20(5):1341-1350. doi: 10.1111/ajt.15752. Epub 2020 Jan 23.

Discrepancy analysis comparing molecular and histology diagnoses in kidney transplant biopsies - PubMed (nih.gov)

( Table 2: 139 Histologic TCMRs, 55 called on MMDX)

Current text:

“Technical validity has been demonstrated ---”

Suggested text:

 “Limited technical validity has been demonstrated for categorizing histologic phenotypes, but discrepancy rates with histologic labels used to train the classifiers discrepancy  vary between 27.9 to 46.9%.  Another caveat is that statistical measures of variability of principal component scores, archetype scores and classifier scores is not presented in any of the landmark publications in this field.”

Rationale:

Consider referring to data in the following papers

 1. In-silico performance, validation, and modeling of the Nanostring Banff Human Organ transplant gene panel using archival data from human kidney transplants.

Smith RN.

BMC Med Genomics. 2021 Mar 19;14(1):86. doi: 10.1186/s12920-021-00891-5.

PMID: 33740956

https://bmcmedgenomics.biomedcentral.com/articles/10.1186/s12920-021-00891-5

2.   The MMDx® diagnostic system: A critical re-appraisal of its knowledge gaps and a call for rigorous validation studies.

Randhawa P.Clin Transplant. 2022 Nov;36(11):e14747. doi: 10.1111/ctr.14747. Epub 2022 Jun 16.PMID: 35678044

https://pubmed.ncbi.nlm.nih.gov/?term=randhawa+knowledge+gaps&sort=date&size=50#:~:text=MMDx%C2%AE%20diagnostic,PMID%3A%2035678044

Thank you very much for this manuscript and nice summary of discussions.

I have very few comments, mostly on the Supplementary Table 2.

For the tubulitis, would it be clearer to add "in the most affected tubule" in the definition of score? For example, t2: 2 or more foci with 5 to 10 mononuclear cells/tubular cross section (or 10 tubular cells) in the most affected tubule. Otherwise, we could think that we need at least 2 tubules with at least 5 to 10 cells.

For the tubulitis score, when talking about severely atrophic tubules, we could add: see definition of severely atrophic tubules below.

In the main text, in the paragraph "Arterial intimal fibrosis" the number of the Table is missing.

Thank you very much.

Marion Rabant

Congratulations for this great work. I have 3 remarks:

- As non-HLA DSA are not routinely tested and the STAR Working Group does not recommend testing for non-HLA antibodies for clinical decision making, the « DSA-negative » statement in « MVI, DSA-negative and C4d-negative » seems a bit vague. Couldn’t “without detectable DSA” or “HLA-DSA negative” be more proper terms for now?

- In table 3: “cases with “Probable AMR” with only histological chronic lesions cg or ptcml can be labelled as “chronic inactive AMR””. This is a bit confusing as per definition, “Probable AMR” does not have “only histological chronic lesions” as they meet criterion 1 for ABMR (minor active lesions).

- Missing closing parenthesis in Supplemental Table 1, 2nd round, first item “(meeting all 3 Banff criteria)”

Thank you for a comprehensive summary of the discussions at the Banff 2022 meeting.

I think it would be most desirable for the transplant community to keep the table format for the updated Banff classification as close as possible to that of the previous versions, which provided a streamlined view of the whole classification, and to include in such format the changes proposed. Table 3 is very difficult to digest.

In response to some of the comments below on Table 3:

It was decided at the Banff 2022 meeting, and also reached strong consensus in the post-meeting survey, that we update “Category 2: Antibody-mediated rejection (AMR) and microvascular inflammation (MVI)”. This update defines criteria for inclusion in the Banff Classification of

1. MVI, DSAneg and C4dneg = cases with MVI that are DSA-negative and C4d-negative (1st and 2nd AMR criteria fulfilled, but 3rd criterion negative), and for

2. Probable AMR = cases fulfilling the 1st and 3rd but not the 2nd AMR criterion

both of which according to Banff 2019 falled in the “No rejection (Category 1)” Diagnostic Category (before 2017 they were denominated “suspicious for AMR” within Category 2).

Figure 2 provides a simplified diagram to indicate this clear rationale behind the criteria for defining these two phenotypes" (MVI DSAneg C4dneg; probable AMR), in relation to the AMR definitions within Category 2. Since the details are as important as the general rationale, the details of these definitions are in Table 3; the writing committee of all coauthors suggested keeping Table 3 in the main manuscript to avoid overlooking the details.

@ Maike & Kerstin.

You draw attention to important points of terminology. Although this might sound very detailed, definitions need to be crystal clear so as to eliminate a source of inter-observer variability, even before looking down the microscope.

"Non-sclerotic" and "non-sclerosed" have been used in definitions inconsistently, to refer to globally sclerosed glomeruli, whereas those terms might also be interpreted by some to include segmental sclerosis. In response to your comment, minor changes will be made to the website, using the terminology "global sclerosis" as in Haas et al. Kidney International (2020) 98, 1120–1134 (consensus glomerular definitions).

- Glomerulitis is assessed in all non-globally sclerosed glomeruli; the denominator is all non-globally sclerosed glomeruli. Ischaemia is not relevant. Concerning your comment on the 2019 Banff table and footnotes, please be aware that the website is now the sole repository of the current, up-to-date classification. The Banff 2022 report does not contain any Classification table to avoid minor errors and changes accidentally being introduced in Banff manuscript and leading to problems like this.

- Adequacy remains 10 glomeruli, by convention, whether they are obsolete or not. If you think the biopsy does not contain enough viable glomeruli to adequately assess for glomerulitis because most are obsolete, you could comment on that as a limitation in your report . This same problem applies to most classifications, like lupus nephritis ISN/RPS or MESTC score too.

- Tubulitis: if only one focus is noted in the biopsy, this is t0. If there are at least 2 foci, then there is a Banff t lesion score, and it is determined by the most affected, so in your example (2 foci one with moderate and one with severe tubulitis) this is t3

- A "focus" of tubulitis is at least 1 tubule with tubulitis. As far as I know there is no definition for the min distance between 2 foci - a point for the TCMR working group to address.

- t-IFTA: this lesion was introduced in Banff 2019, and it was not mentioned whether there was a requirement for 2 foci. Another point for the TCMR working group to address

- ct. This score does include endocrine type atrophy. Please see Glossary on website for definition of Tubular Atrophy. A minor change will be made to the ct score paragraph to refer readers on to the Glossary for the definition. Thyroidisation type has not been mentioned to date, perhaps also a point for the TCMR working group to address.

Thank you for your careful reading! Candice

After reading the report again, I think we need to be consistent in terminology about what MVI+/DSA and C4d negative cases are: in the current draft we call it entity, phenotype, and category - but to me those are different things. To me it is not an entity and not a new Banff category. I think it is a definition of a phenotype we need to study more to understand the underlying pathomechanisms.

I agree Table 3 is very complex and can be read in different ways - The corresponding figure is way more intuitive and clear.

Thank you for the very thought-out and helpful refinements of the Kidney Banff classification, that have been proposed. The kidney working group surely managed to pick up major issues which we are frequently confronted with in our daily routine as transplant pathologists. In particular the term MVI, C4d negative, DSA negative is very important to learn a better understanding of this not so infrequent phenomenon.

With regard to MVI it would be very helpful to have a clearer statement on how to score glomerulitis. The morphology of glomerulitis has so far been well defined and is comprehensive, however, the denominator is not clear. In the Banff homepage it is indicated that non-sclerosed glomeruli are the denominator, although in the description of the score it says „% of glomeruli“ as opposed to mm-score where it is indicated „% of non-sclerotic glomeruli“ which appears a little contradictory. Moreover, in the 2019 Banff report it is indicated in a footnote that globally sclerosed and ischemic glomeruli should not be scored, but it is not indicated what the denominator is and whether both should be excluded. The adequacy of a sample is defined by 10 glomeruli in total, but would that also make sense for glomerulitis if the sclerosed glomeruli were not included in the denominator? As the MVI threshold is of paramount importance for the diagnosis of AMR it would be great if this could be clarified. In an example of a biopsy with 24 glomeruli, 8 sclerostic and 10 ischemic and glomerulitis in 5 of the remaining 6 this would mean g1 if all glomeruli were in the denominator, g2 if sclerotic glomeruli were excluded and g3 if sclerotic and ischemic glomeruli were excluded. This is a very artificial example for sure, but it might illustrate our problem.

We would also be happy for some more detailed defintion with respect to histological criteria of TCMR.

It is stated that 2 foci are needed for the assignment to a t-score. But what exactly is a focus? Is a single tubule with tubulitis sufficient and how far do the tubules have to be apart not to belong to the same focus.

In the definition of t1-t3 it reads as if 2 foci with each mild, moderate or severe tubulitis need to be present to reach the threshold for the resp. score. How about 2 foci one with moderate and one with severe tubulitis? Is this t2 or t3? We would include this in the t3 group, because the most severly affected tubule shows severe tubulitis, however, this is true for only one focus.

Is the definition of tIFTA also restricted to the presence of 2 foci?

ct-score: shouldn´t neuroendocrine like tubules and thyroidization be included in the definition?

Thank you again for this great work,

Maike Büttner and Kerstin Amann

Congratulations on the great work, & thanks for all you have done on this! I understand the need to integrate the understanding of microvascular inflammation, antibody-mediated rejection, transcriptomics, etc., more; & I applaud you for your efforts in doing that! However, I also agree that Table 3 is quite complicated & that it should be made clear what is being changed in the current diagnostic criteria (if anything). If Table 3 isn't really a change to the current diagnostic criteria, then the suggestion to place that in a Supplement is a nice suggestion.

It seems that one of the main conclusions from recent research & the meeting was that "suspicious" or "possible" antibody-mediated rejection should be recognized in cases that have been ignored or neglected as the microvascular inflammation & other thresholds & criteria have waxed & waned.

Again, thanks so much for all you have done; & best wishes in bringing it to completion!

Congrats for this work. The report is a nice summary of the scientific discussions that took place at Banff and is very informative.

I have only one comment: it is not clear if they are changes in the classification. Some paragraphs seem to say yes, others no, mostly relying on the fact there is probably not yet sufficient scientific supporting evidence to change the classification considering all the clinical implications that this could have. To avoid any confusion, and state that there is still a need for additional data and scientific discussions to confirm these changes, it should probably be necessary to change and/or move in the supplementary material the Table 3 (which I believe is by the way not so intuitive, and quite difficult to understand).

Warn congratulations to the fantastic job in summarizing the Banff meeting.

I only have three comments;

1) It is my understanding that 2 new entities were delineated by the working group. Does this mean that Banff classification is changing?

2) If so, please provide a clear Table displaying the new classification .

3) A Table summarizing the new Banff classification would be more informative and readable than Table 3 which represents a proposed "clinical reasoning framework for intuitive classification of cases to come to a final disease classification".

Fantastic job summarized here,

I just would like to share to remarks:

_ The sentence "if the testing methods are sufficiently standardized and clinically validated', used throughout the text and in Table 3. I perfectly understand challenges raised here but it seems very difficult to appy for clinical practice... A companion guide could be added to help clinician to decide which is considered by the group as (close to be) validated...

_ Table 1 and 2, 4 and 5 are perfectly clear. However, although Figure 2 is clear, Table 3 is a bit difficult to understand (especially the 3 columns).

Table 3 is overly complicated and gives the wrong impression that there is a change in the classification as well as a change in the diagnostic criteria. This will be really confusing for both pathologists and clinicians in their decision making process to treat patients. It also distracts from the main take home messages of the paper. Table 3 has to go in the supplementary appendix so that the manuscript focuses on its main goals.