The prevalent renal transplant population presents an opportunity to observe the adaptive changes in the alloimmune response over time, but such studies have been limited by uncertainties in the conventional biopsy diagnosis of T cell-mediated rejection (TCMR) and antibody-mediated rejection (ABMR). To circumvent these limitations, we used microarrays and conventional methods to investigate rejection in 703 unselected biopsies taken 3 days to 35 years post-transplant from North American and European centers. Using conventional methods, we diagnosed rejection in 205 biopsy specimens (28%): 67 pure TCMR, 110 pure ABMR, and 28 mixed (89 designated borderline). Using microarrays, we diagnosed rejection in 228 biopsy specimens (32%): 76 pure TCMR, 124 pure ABMR, and 28 mixed (no borderline). Molecular assessment confirmed most conventional diagnoses (agreement was 90% for TCMR and 83% for ABMR) but revealed some errors, particularly in mixed rejection, and improved prediction of failure. ABMR was strongly associated with increased graft loss, but TCMR was not. ABMR became common in biopsy specimens obtained .1 year post-transplant and continued to appear in all subsequent intervals. TCMR was common early but progressively disappeared over time. In 108 biopsy specimens obtained 10.2-35 years post-transplant, TCMR defined by molecular and conventional features was never observed. We conclude that the main cause of kidney transplant failure is ABMR, which can present even decades after transplantation. In contrast, TCMR disappears by 10 years post-transplant, implying that a state of partial adaptive tolerance emerges over time in the kidney transplant population.

Intimal arteritis (the presence of v-lesions) in kidney transplant biopsy specimens is believed to have major prognostic and diagnostic significance. We assessed the relationship of v-lesions to prognosis in 703 indication biopsy specimens and used microarray-based molecular tests to re-examine the relationship of v-lesions to rejection. v-Lesions were noted in 49 specimens (7%) and were usually mild (v1). The presence of v-lesions had no effect on graft survival compared with the absence of v-lesions. Pathologists using current conventions almost always interpreted v-lesions as reflecting T cell-mediated rejection (TCMR), either pure or mixed with antibody-mediated rejection (ABMR). The molecular scores questioned the conventional diagnoses in 29 of 49 specimens (59%), including ten that were conventional TCMR with no molecular rejection and nine that were conventional TCMR mixed with pure ABMR molecularly. The presence of tubulointerstitial inflammation (i-t) meeting TCMR criteria allowed subclassification of v-lesion specimens into 21 i-t-v-lesion specimens and 28 isolated v-lesion specimens. Molecular TCMR scores were positive in 95% of i-t-v-lesion specimens but only 21% of isolated v-lesion specimens. Molecular ABMR scores were often positive in isolated v-lesion biopsies (46%). Time of biopsy after transplantation was critical for understanding isolated v-lesions: most early isolated v-lesion specimens had no molecular rejection and were DSA negative, whereas most isolated .1 year after transplantation had positive DSA and ABMR scores. Therefore, v-lesions in indication biopsy specimens do not affect prognosis and can reflect TCMR, ABMR, or no rejection. Time after transplantation, DSA, and accompanying inflammation provide probabilistic basis for interpreting v-lesions.

The key lesions in antibody-mediated kidney transplant rejection (ABMR) are microcirculation inflammation (peritubular capillaritis and/or glomerulitis lesions, abbreviated ''pg'') and glomerular double contours (cg lesions). We used these features to explore subphenotypes in 164 indication biopsies with ABMR-related diagnoses: 137 ABMR (109 pure and 28 mixed with T cellmediated rejection [TCMR]) and 27 transplant glomerulopathy (TG), identified from prospective multicenter studies. The lesions indicated three ABMR subphenotypes: pgABMR, cgABMR, and pgcgABMR. Principal component analysis confirmed these subphenotypes and showed that TG can be reclassified as pgcgABMR (n ¼ 17) or cgABMR (n ¼ 10). ABMR-related biopsies included 45 pgABMR, 90 pgcgABMR, and 25 cgABMR, with four unclassifiable. Dominating all time intervals was the subphenotype pgcgABMR. The pgABMR subphenotype presented earliest (median <2 years), frequently mixed with TCMR, and was most associated with nonadherence. The cgABMR subphenotype presented late (median 9 years). Subphenotypes differed in their molecular changes, with pgABMR having the most histologic-molecular discrepancies (i.e. potential errors). Donor-specific antibody (DSA) was not identified in 29% of pgcgABMR and 46% of cgABMR, but failure rates and molecular findings were similar to cases where DSA was known to be positive. Thus, ABMR presents distinct subphenotypes, early pg-dominant, late cg-dominant, and combined pgcg phenotype, differing in time, molecular features, accompanying TCMR, HLA antibody, and probability of nonadherence.

Recognition that some lesions typical of T cellmediated rejection (TCMR) also occur in antibodymediated rejection requires revision of the histologic TCMR definition. To guide this process, we assessed the relative importance of various lesions and the performance of new histology diagnostic algorithms, using molecular TCMR scores as histology-independent estimates of true TCMR. In 703 indication biopsies, random forest analysis and logistic regression indicated that interstitial infiltrate (i-lesions) and tubulitis (t-lesions) were the key histologic predictors of molecular TCMR, with arteritis (v-lesions) having less importance. Histology predicted molecular TCMR more accurately when diagnoses were assigned by strictly applying the Banff rules to the lesion scores and redefining isolated v-lesion TCMR. This improved prediction from area under the curve (AUC) 0.70 with existing rules to AUC 0.80. Further improvements were achieved by introducing more categories to reflect inflammation (AUC 0.84), by summing the lesion scores (AUC 0.85) and by logistic regression (AUC 0.90). We concluded that histologic assessment of TCMR can be improved by placing more emphasis on i-and t-lesions and incorporating new algorithms for diagnosis. Nevertheless, some discrepancies between histologic and molecular diagnoses persist, partially due to the inherent nonspecificity of i-and t-lesions, and molecular methods will be required to help resolve these cases.