Chronic Allograft Nephropathy Grades in Sequential Protocol Biopsies

1
Chronic Allograft Nephropathy Grades in
Sequential Protocol Biopsies
Patricia E. Birk, MD, FRCP(C)1 and Ian W.
Gibson, MBChB, MRCPath(UK)2 Departments of
1Pediatrics and 2Pathology University of Manitoba
Disclosures P. Birk Astellas (Investigator,
speaker), Wyeth (Investigator), Hoffman LaRoche
(Investigator)I. Gibson None
2
Overview

• Review the Pediatric experience with protocol
  biopsies for chronic allograft nephropathy (CAN)
  surveillance.
• Importance of protocol biopsies for pediatric
  recipients
• 2. Discuss the methodological problems
  associated with the use of histological
  end-points in clinical studies.
• Limitations of Banff chronic interstitial (ci)
  scores
• 3. Develop research design strategies for
  clinical trials using CAN as an end-point.
• Computerized image analysis of routine
  histological stains

3
Overview

• Review the Pediatric experience with protocol
  biopsies for chronic allograft nephropathy (CAN)
  surveillance.
• Importance of protocol biopsies for pediatric
  recipients
• 2. Discuss the methodological problems
  associated with the use of histological
  end-points in clinical studies.
• Limitations of Banff chronic interstitial (ci)
  scores
• 3. Develop research design strategies for
  clinical trials using CAN as an end-point.
• Computerized image analysis of routine
  histological stains

4
Overview

• Review the Pediatric experience with protocol
  biopsies for chronic allograft nephropathy (CAN)
  surveillance.
• Importance of protocol biopsies for pediatric
  recipients
• 2. Discuss the methodological problems
  associated with the use of histological
  end-points in clinical studies.
• Limitations of Banff chronic interstitial (ci)
  scores
• 3. Develop research design strategies for
  clinical trials using CAN as an end-point.
• Computerized image analysis of routine
  histological stains

5
Rationale for CAN Surveillance in Pediatric Renal
Allograft Recipients
6
13 yr old female 5 yrs post-transplant Cr 78
?mol/L (0.9 mg/dl)
7
13 yr old female 5 yrs post-transplant Cr 78
?mol/L (0.9 mg/dl) Grade III CAN
8
18 yr old female 10 yrs post-transplant Cr 94
?mol/L (1.1 mg/dl)
9
18 yr old female 10 yrs post-transplant Cr 94
?mol/L (1.1 mg/dl) Grade II CAN with sclerosing
glomerulopathy
10
Incidence and Progression of CAN(N21)

• 18/21 (86) pts
• gt Grade I CAN
• 13/21 (62) Grade I
• CAN
• 5/21 (24) Grade II
• CAN

Birk and Gibson, Pediatr Transplantation, 2004
11
CAN Function vs. Histology (CAN vs. No CAN,
pNS)
Birk and Gibson, Pediatr Transplantation, 2004
12
Banff Chronic Pathological ScoringMethodological
Issues
13

Limitations of Banff ci Scoring

• Nonlinear increases
• Reported as ranges
• Large ranges mask incremental (and potentially
  clinically significant) increases in interstitial
  fibrosis

14
ci00-5
ci226-50
ci2 ? ci1 x2
ci16-25
ci3gt50
15

Limitations of Banff ci Scoring

• Nonlinear increases
• eg. ci2 ? ci1 x2
• Reported as ranges
• What is a mean ci score?
• Large ranges mask incremental (and potentially
  clinically significant) increases in interstitial
  fibrosis

16
ci00-5
ci226-50
11 Mean ci Score1.7 42
ci16-25
ci3gt50
17

Limitations of Banff ci Scoring

• Nonlinear increases
• eg. ci2 ? ci1 x2
• Reported as ranges
• What is a mean ci score?
• Large ranges mask incremental (and potentially
  clinically significant) increases in interstitial
  fibrosis

18
ci16-25
6 12 18
19
Computerized Image Analysis of Renal Allograft
Biopsies

• Seron et al., 1993
• 43 biopsies in 43 patients
• Nicholson et al., 1996
• 107 biopsies in 130 patients
• Nicholson et al., 1999
• 156 biopsies in 52 patients
• Seron et al., 2001
• 107 biopsies in 40 patients
• Grimm et al., 2003
• 190 biopsies
• Pape et al., 2003
• 56 biopsies in 56 patients
• Diaz Encarnacion et al., 2004
• 49 biopsies

20
Computerized Image Analysis of Massons Trichrome
Stain
21
Aniline blue counterstain
22
Allograft Interstitial Fibrosis A Surrogate
for CAN

• Well-visualized
• IF gtgtarterial intimal hyperplasia
• Associated with ? allograft function and survival
• Isoniemi et al., 1994
• Nickerson et al., 1998
• Moreso et al., 2001
• Nankivell et al., 2001, 2003, 2004

23
Quantitative Image Analysis
Computerized Image Analysis
24
Quantitative Image Analysis
Computerized Image Analysis
25
Digitalized Image, ci2
26
Digitalized Image, ci2
27
Digitalized Image, ci2
Class 1 White (0-150 U) Class 2 Blue (151-170
U) Class 3 Purple, red (171-255 U)
28
Hue Saturation Intensity (HSI) Analysis
Hue ? light Saturation Purity of
color Intensity Strength of signal
29
Digitalized Image, ci2
30
Segmentation Step
31
Segmentation Step (IF26.75)
32
Percentage of Interstitial Fibrosis
Class 2 Class 1 Class 2 Class 3
33
Mean Area Percentage of Interstitial Fibrosis
(IF)
IF ?Area IF / Total Area??1 ?Area IF /
Total Area??2 ?3... n
34

Statistical Analysis of Protocol Biopsy Data

• Protocol biopsies Repeated measurements ?
  Autocorrelation
• Must account for autocorrelation in statistical
  analyses
• eg. General estimating equations (GEE) modeling
• (Hanley et al., Am J Epidemiol, 2003)

35
Mean Area Interstitial Fibrosis (IF) According
to Banff ci Scores in Pediatric Protocol Biopsies
(N105)
20.53 8.74
10.39 5.23
6.83 3.94
0-5
6-25
26-50
36
Mean Area Interstitial Fibrosis (IF) According
to Banff ci Scores in Pediatric Protocol Biopsies
(N105)
20.53 8.74
10.39 5.23
6.83 3.94
0-5
6-25
26-50
37
Relationship Between IF and Banff ci Scores
Using GEE Modeling (N105)
Birk, Gill and Gibson, CST, ATC, 2005
38
Relationship Between IF and Banff ci Scores
Using GEE Modeling (N105)
Birk, Gill and Gibson, CST, ATC, 2005
39
ci1 with Inflammatory Infiltrate, i3
40
Segmentation Step (IF11.47)
41
Relationship Between IF and Banff ci Scores
Using GEE Modeling (N105)
Birk, Gill and Gibson, CST, ATC, 2005
42
Relationship Between IF and Estimated GFR in
Pediatric Protocol Biopsies Without AR (N83)
(plt0.004)
Birk, Gill and Gibson, CST, ATC, 2005
43

Advantages of HSI Technology

• Applicability to routine histological stains
• Rapid (lt30 min/slide)
• Simple
• Technician-operated
• Requires basic knowledge of histology
• (Relatively) inexpensive
• Microscope with digital camera
• Desk-top computer
• Software

44
Protocol Biopsies The Problem of Sampling Error
45
3 yrs post-transplant, Grade III CAN
5 yrs post-transplant, Grade II CAN
46
CAN Grade Variability in Progressors (N13)
Birk and Gibson, CST, CSN, ATC, 2004
47
1 mo, ci1
12 mo, ci2
Progression
Regression
Regression
24 mo, ci1
3 mo, ci0
Progression
48
CAN Grade Variability in a Pediatric Protocol
Biopsy Patient
2
1
0
Birk and Gibson, CST, CSN, ATC, 2004
49
CAN Grade Variability in a Pediatric Protocol
Biopsy Patient
2
10
1
5
Interstitial Fibrosis
0
0
Birk and Gibson, CST, CSN, ATC, 2004
50
Magnitude of Slope (IF) vs. DTPA GFR
Birk and Gibson, CST, CSN, ATC, 2004
51
Summary

• HSI image analysis of trichrome stain can be used
  to quantify renal allograft interstitial
  fibrosis.
• Objective
• Precise (unaffected by AR)
• Generates continuous variables for clinical
  studies
• HSI methodology is a valuable tool for the
  evaluation of interstitial fibrosis progression
  in longitudinal biopsies confounded by sampling
  error.

52
Summary

• HSI image analysis of trichrome stain can be used
  to quantify renal allograft interstitial
  fibrosis.
• Objective
• Precise (unaffected by AR)
• Generates continuous variables for clinical
  studies
• HSI methodology is a valuable tool for the
  evaluation of interstitial fibrosis progression
  in longitudinal biopsies confounded by sampling
  error.

53
Iceberg Analogy of Renal Allograft Injury
Clinical AR ?Cr
Subclinical AR ? Cr
54
Iceberg Analogy of Renal Allograft Injury
(Children)
Clinical AR/ Established CAN ?Cr
Subclinical AR/ Early CAN ? Cr
55
Collaborators and Support
Pediatric Nephrology M. Ogborn T. Blydt-Hansen K.
Pederson Pathology J. Gartner I.
Gibson Research Laboratory H. Dostmohammed H.
Konrad E. Saftiuk
University of British Columbia J.
Gill Funding Canadian Institutes for Health
Research Kidney Foundation of Canada Manitoba
Institute of Child Health Astellas Pharma Canada
Inc.