Smith, Byron H.

Significance Statement Histopathologic assessment of kidney tissue currently relies on manual scoring or traditional image-processing techniques to quantify and classify tissue features, time-consuming approaches that have limited reproducibility. The authors present an alternative approach, featuring a convolutional neural network for multiclass segmentation of kidney tissue in sections stained by periodic acid–Schiff. Their findings demonstrate applicability of convolutional neural networks for tissue from multiple centers, for biopsies and nephrectomy samples, and for the analysis of both healthy and pathologic tissues. In addition, they validated the network’s results with components from the Banff classification system. Their convolutional neural network may have utility for quantitative studies involving kidney histopathology across centers and potential for application in routine diagnostics. Background The development of deep neural networks is facilitating more advanced digital analysis of histopathologic images. We trained a convolutional neural network for multiclass segmentation of digitized kidney tissue sections stained with periodic acid–Schiff (PAS). Methods We trained the network using multiclass annotations from 40 whole-slide images of stained kidney transplant biopsies and applied it to four independent data sets. We assessed multiclass segmentation performance by calculating Dice coefficients for ten tissue classes on ten transplant biopsies from the Radboud University Medical Center in Nijmegen, The Netherlands, and on ten transplant biopsies from an external center for validation. We also fully segmented 15 nephrectomy samples and calculated the network’s glomerular detection rates and compared network-based measures with visually scored histologic components (Banff classification) in 82 kidney transplant biopsies. Results The weighted mean Dice coefficients of all classes were 0.80 and 0.84 in ten kidney transplant biopsies from the Radboud center and the external center, respectively. The best segmented class was “glomeruli” in both data sets (Dice coefficients, 0.95 and 0.94, respectively), followed by “tubuli combined” and “interstitium.” The network detected 92.7% of all glomeruli in nephrectomy samples, with 10.4% false positives. In whole transplant biopsies, the mean intraclass correlation coefficient for glomerular counting performed by pathologists versus the network was 0.94. We found significant correlations between visually scored histologic components and network-based measures. Conclusions This study presents the first convolutional neural network for multiclass segmentation of PAS-stained nephrectomy samples and transplant biopsies. Our network may have utility for quantitative studies involving kidney histopathology across centers and provide opportunities for deep learning applications in routine diagnostics.

Background. Kidneys from deceased donors with acute kidney injury (AKI) are more likely to be discarded because of concerns for poor outcomes after transplantation. The aim of this study was to determine the long-term outcomes of a large cohort of patients transplanted utilizing kidneys from deceased donors with AKI. Methods. All patients receiving a deceased donor kidney transplant during a recent 10-year period were included. Acute Kidney Injury Network (AKIN) criteria were used to classify the donors. Donor kidneys with >10% cortical necrosis or more than mild chronic changes were discarded. The primary outcome is the combined endpoint of death or graft loss. Results. The cohort included 1313 kidneys from 974 donors, AKIN stage 0 (no AKI) in 319 (24.3%), stage 1 in 370 (28.2%), stage 2 in 177 (13.5), and stage 3 in 447 (34.0%). Estimated 5-year graft survival (95% confidence interval) was 78.5% (72.5-84.5), 77.8% (72.8-82.1), 83.8% (76.8-88.9), and 84.6% (79.5-88.7) for AKIN donor stage 0 to 3, respectively (log-rank P = 0.10). After adjusting for baseline differences, the hazard ratio (95% confidence interval) for the combined endpoint for the AKIN stage 3 group (relative to AKIN 0 group) was 0.70 (0.45-1.10). Delayed graft function occurred in 44.6% and 75.4% of AKIN 2 and 3 groups, as compared to 33.9% and 33.5% in AKIN 0 and 1 ( P < 0.001). Conclusion. We conclude that transplanting selected kidneys from deceased donors with AKI with preimplantation biopsy showing <10% cortical necrosis and no more than mild chronic changes have excellent long-term graft survival.

Background Nonadherence to immunosuppression after kidney transplant is an important contributor to graft failure. Little is known about how nonadherence changes 3 years posttransplant when Medicare coverage of immunosuppression ends and how that nonadherence impacts allograft histology. The goal of this study was to compare rates of nonadherence during posttransplant years 1 to 3 to years 3 to 5 and examine the relationship between nonadherence during years 3 to 5 and 5-year allograft histology. Methods We retrospectively analyzed 552 conventional kidney allografts in patients transplanted at our center between January 1, 1999, and June 1, 2010, who used the Mayo Clinic Specialty Pharmacy for the first 5 years posttransplant. Nonadherence was defined as less than 80% proportion of days covered. Overall adherence to immunosuppression appeared to be higher during years 3 and 5 compared to between years 1 and 3 (89.4% vs 82.9%, respectively; P < 0.0001 [paired t test]). Results Overall nonadherence during posttransplant years 3 to 5 appeared to be associated with fibrosis and inflammation on 5-year allograft biopsy but not with transplant glomerulopathy (16.9% vs 5.9%, P = 0.004; 10.4% vs 8.5%, P = 0.61, respectively). After adjusting for nonadherence to calcineurin inhibitor and prednisone therapy, only nonadherence to antimetabolite therapy remained significantly associated with 5-year fibrosis and inflammation (odds ratio, 10.6; 95% confidence interval, 1.5-76.1; P = 0.02). Conclusions Efforts to improve long-term adherence, possibly through the use of specialty pharmacies and increased adherence to antimetabolite therapy, may improve long-term allograft histology and survival, although further studies are needed to confirm these findings.

Background The concepts that obesity is merely a consequence of overeating, and that metabolic health then reflects obesity, may be insufficient and potentially flawed. The role of fructose intake and metabolic endotoxemia has gained attention recently, but data in kidney transplantation are lacking. This study evaluated the risk factors for metabolic syndrome (MS), its components, and other associated markers in kidney transplant recipients (KTRs), focusing particularly on fructose intake and systemic endotoxemia. Methods This cross-sectional observational study enrolled 128 KTRs longer than 1 year posttransplantation. Clinical, biochemical, anthropometric, and questionnaire assessments were undertaken. Results Obesity (body mass index, ≥30 kg/m2) and MS (International Diabetes Federation Definition) were found in 36.7% and 50% of KTRs, respectively. Both increased fructose intake (P = 0.01) and endotoxin level (P = 0.02) were independently associated with MS; and higher fructose intake was independently associated with obesity (P < 0.001). Specifically, increased fructose intake was associated with the central obesity (P = 0.01) and hyperglycemia (P < 0.001) criteria of MS, whereas higher endotoxin level was associated with the hypertriglyceridemia (P = 0.003) and low HDL cholesterol concentration (P = 0.002) criteria of MS. Neither saturated fat nor total caloric intakes were independently associated with obesity and MS; and neither obesity nor central obesity were independently associated with the dyslipidemia and hyperglycemia criteria of MS. Principal component analysis demonstrated relationships between higher levels of endotoxin, soluble endothelial selectin, triglycerides, and insulin resistance (r > 0.6), as well as relationships between increased fructose intake, inflammation, and blood glucose (r > 0.6). Conclusions Dietary modifications through decreasing fructose intake and addressing systemic endotoxemia are plausible targets for improving metabolic health of KTRs.

Background and objectives Delayed graft function is a form of AKI resulting from ischemia-reperfusion injury. Our aim was to study the effect of delayed graft function on the progression of interstitial fibrosis after deceased donor kidney transplantation. Design, setting, participants, & measurements Our study is a retrospective study of all patients transplanted at our center between July of 2003 and September of 2014 using a kidney from a deceased donor. The primary outcome was the progression of interstitial fibrosis on serial protocol biopsies done during the first year post-transplant. We analyzed the distribution of the change in the Banff interstitial fibrosis (ci) score between the delayed graft function and nondelayed graft function groups for all of the paired biopsies done at time 0 and 12 months post-transplant ( Δ fibrosis). We also performed a linear mixed model analyzing the difference in the slopes for the progression of mean Banff ci score for all of the biopsies done at time 0 and 1, 4, and 12 months post-transplant. Results There were 343 (36.7%) in the delayed graft function group and 591 in the control group. The biopsy rates for the delayed graft function and nondelayed graft function groups at time 0 were 65.3% ( n =224) and 67.0% ( n =396), respectively, and at 12 months, they were 64.4% ( n =221) and 68.4% ( n =404), respectively. Paired biopsies were available for 155 in the delayed graft function group and 283 in the control group. In a risk-adjusted model, Banff ci score >0 on the time 0 biopsy had a higher odds of delayed graft function (odds ratio, 1.70; 95% confidence interval, 1.03 to 2.82). The distribution of the Δ fibrosis between 0 and 12 months was similar in delayed graft function and control groups ( P =0.91). The slopes representing the progression of fibrosis were also similar between the groups ( P =0.66). Conclusions Donor-derived fibrosis may increase the odds of delayed graft function; however, delayed graft function does not seem to increase the progression of fibrosis during the first year after transplantation.