REPRODUCIBILITY OF THE MORPHOMETRIC QUANTIFICATION OF VASCULAR LESIONS IN RENAL ALLOGRAFT BIOPSIES.

 

JL Bosmans1, F Moreso², M Mengel³, D Seron², H Haller4, ME De Broe1 for the European Study Group of Renal Allograft Protocol Biopsies.

 

1 Department of Nephrology-Hypertension, University Hospital Antwerpen, Antwerpen,

  Belgium.

² Department of Nephrology, Hospital de Bellvitge, Barcelona, Spain.

³ Department of Pathology, Medizinische Hochschule, Hannover, Germany.

4  Department of Nephrology, Medizinische Hochschule, Hannover, Germany.

 

 

Vascular lesions in renal allograft biopsies, either at implantation or in protocol biopsies, have been consistently shown to adversely affect the outcome of the graft. The development of a reproducible quantification method of these vascular alterations could therefore have important applications in both diagnostic and therapeutical interventional studies.

The present study investigated the reproducibility of the morphometric quantification of fibrous intimal thickening and arterial wall thickness in protocol biopsies of renal allografts, as assessed by two observers from two different centres, using two different image analysis programs.

Three transplant centres each provided fifteen renal allograft protocol biopsies. The biopsies were performed at a median time of 180 days after transplantation, and were either fixed in Duboscq-Brazil or in formalin, and cut at 4 mm. These 45 biopsies were collected and separately analysed in two centres, A and B. All biopsies were stained with Masson’s trichrome and silver methenamine. The serumcreatinine at the time of biopsy was 142 ± 61 mmol /L.

Morphometric analysis was performed without knowledge of the clinical data. All arterial sections in a single biopsy were analysed. Arteries were defined as vessels containing at least two muscular layers. In centre A, the 45 protocol biopsies were analysed by observer A, on silver methenamine sections at 400x magnification, by means of the point counting method.

In centre B, the same 45 protocol biopsies were analysed by observer B, on Masson’ trichrome stained sections at 400x magnification, by means of manual delineation of the lumen, intima and media, using a KS 400 (Kontron Elektronik GmbH, Release 2.0) imaging system, connected to a Leica-DMR microscope.

Following variables were analysed by both observers: mean time required to analyse a biopsy, number of arteries, ratio of intimal area / medial area, ratio of arterial wall area / total arterial area. Results are expressed as the mean ± standard deviation. Paired T-test statistics were used to compare the results. The Pearson correlation coefficient was used to study the correlation between both morphometric methods. Bland-Altman plots were used to compare agreement between both centres. The statistical analysis was performed with Medcalc 7.0.

The mean number of arteries per biopsy was 2 ± 2. Six out of fourty-five biopsies (13%) had no arteries. There was an excellent agreement between the two observers (R= 0.89, p<0.0001), with however a larger number of arteries per biopsy in centre A (mean difference: 0.4; p=0.02). There was a very good agreement between both observers with regard to the quantification of fibrous intimal thickening (R= 0.72; p<0.0001), with however a significant difference (mean difference: 0.19; p<0.0001). In addition, the agreement was almost optimal at the lower and middle values, but showed symmetric divergence at the highest values. There was a good agreement between the two observers with regard to the quantification of arterial wall thickness (R=0.67; p<0.0001), without any significant difference between the two measurements. Almost maximal agreement was observed in the higher range of values, while symmetrical disagreement was more prominent at the lower values. The mean time, required to analyse one biopsy, was comparable in both centres, and was 10 ± 5 min in centre A, while it was 11 ± 4 min in centre B, allowing the analysis of 5 to 6 biopsies per hour.

Our study clearly demonstrates that the morphometric analysis of arterial lesions in protocol biopsies of renal allografts, performed by two different observers, from two different centres, and using two distinct image analysis programs, yields a reproducible quantification of these lesions. Approximately five to six biopsies can be analysed per hour with both methods. Morphometric analysis of arteries in renal allograft biopsies may have important applications in future studies.

 

REPRODUCIBILITY OF THE MORPHOMETRIC QUANTIFICATION OF VASCULAR LESIONS IN RENAL ALLOGRAFT BIOPSIES.

 

JL Bosmans1, F Moreso², M Mengel³, D Seron², H Haller4, ME De Broe1 for the European Study Group of Renal Allograft Protocol Biopsies.

 

1 Department of Nephrology-Hypertension, University Hospital Antwerpen, Antwerpen,

  Belgium.

² Department of Nephrology, Hospital de Bellvitge, Barcelona, Spain.

³ Department of Pathology, Medizinische Hochschule, Hannover, Germany.

4  Department of Nephrology, Medizinische Hochschule, Hannover, Germany.

 

 

Corresponding Author:

 

Dr. Jean-Louis Bosmans,

University Hospital Antwerpen,

Department of Nephrology-Hypertension,

Wilrijkstraat, 10

B 2650 Edegem

Belgium

Phone: 00-32-3-821.37.92

Fax: 00-32-3-821.90.00

e-mail: bosmans@uia.ua.ac.be