EIN
(Endometrial Intraepithelial Neoplasia):
Improved Criteria for diagnosing
endometrial precancer
Stanley J. Robboy, MD, FCAP,
FFPath FRCPI (Hon), FRCPath (UK, Hon)
Professor of Pathology, Duke University
Past-President, College American Pathologists
Types: Endometrial Cancers
Feature
Histotype
Behavior
Risk factors
Precursor
Endometrioid
Endometrioid,
Secretory,
Squamous
Indolent
Hormonal
“Hyperplasia”
NonEndometrioid
Serous,
Clear cell,
Carcinosarcoma
Aggressive
None
Serous EIC
Retrospective Studies
(Hertig, 1949)
Time from old biopsies to CA
Interval Finding
>15 yr Normal
> 6 yr Cystic hyperplasia
< 5 yr Adenomatous/Atyp hyperplasia CIS
Endometrial Hyperplasia
Common terms
•
•
•
•
Simple v. Complex v. Atypia
Cystic atrophy v. Hyperplasia
Disordered prolif v Simple hyperplasia
Mild, Mod, Marked (3Ms)
–with/without Atypia
• Adenomatous, Anaplasia & CIS
WHO94 Endometrial
Hyperplasia System
Criteria
Glandular complexity
Nuclear atypicality
WHO94 Endometrial
Hyperplasia System
A
R
C
H
I
T
E
C
T
U
R
E
Simple
→
Complex →
No atypia
With atypia
No atypia
With atypia
WHO94 Endometrial
Hyperplasia System
C
Y
T
O
L
O
G
Y
No atypia→
Atypia
→
Simple
Complex
Simple
Complex
Progression to cancer
Nested case control (2008)
Hyperplasia
Relative Risk
Simple
2.0
Complex
2.8
Atypical
14.0
Lacey 2008
Problems in diagnosis:
GOG experience
Community
cases
submitted as
atypical
hyperplasia
Trimble, Gyn Onc 2004
40% Benign/Hyper
30% Atyp Hyper
30% Cancer
Problems in diagnosis:
GOG experience
Expert
gynecologic
pathologists
disagree
among
themselves
60% disagreement
Zaino, USCAP 2004
EIN: Endometrial
Intraepithelial Neoplasia
Conceptual shift
in thinking
EIN – Conceptual Import
• Genetic changes key, not estrogen
• Computer measurable (Reproducible)
• Weeds out cases that otherwise might
likely be treated
• May identify latent cases
Monoclonal Origin
• Point origin and expansile growth
Select relevant fields
Size changes over time
• Lesion contrasts with normal
Compare internally
• Ratio glands to stroma
– Volume % glands
• Length of basement membrane
– ‘Outer surface density’ of glands
• Nuclear pleomorphism
– Std deviation of shortest nuclear diameter
EIN
• Excessive glands (glands > stroma)
• Abnormal architecture
– Excessive branching, out- or inward
Complexity & papillary snouts
• Cytologic atypia
– Nuclei pleomorphic, dyspolarized,
Irregularly stratified
– Nucleoli uniformly prominent
– Cytoplasm eosinophilic
+ 0.0439 x (Volume % Stroma)
– 0.1592 x (Outer Surface Density glands)
– 3.9934 x Ln (Std Dev Shortest Nuclear Axis)
+ 0.6229
Progression risk:
40-60%
25-30%
~ 0%
-4
+5
0 +1
Frequency:
15-25%
Baak 1988
5%
65%
Contribution to D-Score
§ Volume % Glands
§ Perimeter Basement Membr
65%
25%
§ Standard Deviation
Shortest Nuclear Axis
10%
ARCHITECTURAL FEATURES
MORE IMPORTANT
THAN CYTONUCLEAR FEATURES
Volume Percentage Stroma
OUTER SURFACE DENSITY GLANDS
EIN Gland
Normal Gland
Hyperplasia
Basal Membrane = Outer Surface Density
Increases from Normal à Hyperplasia à EIN
Clinical Outcome of 176 WHO
“Hyperplasias”
100
Followup, Months
80
60
40
Outcome
Cancer
No Cancer
20
0
SH
(3/95)
CH
(2/22)
AH
(18/59)
Mutter, 2002
Clinical Outcome of 176 “Hyperplasias”
Rediagnosed by EIN Criteria
100
Followup, Months
80
60
40
20
Outcome
Cancer
No Cancer
No EIN
(1/111)
EIN
(22/65)
Mutter 2002
Timing - a critical flaw
Concurrent
Appears in 1st year
Progression
Appears in 2nd or later years
Concept of Progression
• Concurrent:
Tumor appears in 1st year, i.e., < 1 yr follow-up
197 women
• Progression:
Appears > 1 year
477 women (median 48 mo, max 22 yrs)
Progression – WHO 94
<1 yr
SH
CH
SAH
CAH
%
>1 yr
%
4
9
20
53
0.7
9
7
20
Progression – EIN D-score
<1 yr
Score
%
>1 yr
%
0
41
0.6
29
>1
<1
Cancer Outcomes
in 477 “Hyperplasias” restratified by EIN
Probability of Remaining
Without Progression
1.0
D-Score>1
.8
D-Score 0-1
.6
D-Score <0
.4
HR D-Score >1 vs. 0-1= 28
.2
.0
0
HR D-Score >1 vs <0 = 72
60
120
180
Follow-up Time (> 12 mo)
From Baak, Mutter, Robboy et al, Cancer June 2005
Proliferative
EIN
Complex
Atypical
Simple
Atypical
N=56
Complex
Non-Atypical
N=67
52%
N=65
43%
25%
25%
Simple
Non-Atypical
N=188
36%
19%
31%
Endometrial Intraepithelial
Neoplasia
20%
Baak et al 2005
EIN
No EIN
2 of every 3 “hyperplasia” cases
are benign
EIN: ICD-9
As of January 1, 2010
621.34
Benign endometrial hyperplasia
621.35
Endometrial intraepithelial neoplasia [EIN]
EIN Reproducibility
Usubutum A et al
Modern Pathol 25: 877-884, 2012
Questionaire, 20 reviewers
Terminology preferred
WHO 80%
Read Robboy’s PFRT
Yes
90%
Visit EM.org Website
Yes
90%
EIN system easy to learn Yes
85%
Easy to apply
70%
Yes
Community: "Expert"
EIN Diagnostic
Reproducibility
Case
Expert
Consensus
kappa=0.74
Community-Expert:
20 pathologists
79% agree w expert.
Community to expert
kappas= 0.72 (.45-.84)
Pathologist
Style Group
20
28
36
29
40
37
64
18
27
30
09
21
04
31
60
48
63
17
08
16
32
38
19
14
15
26
11
39
54
62
03
23
61
44
25
06
51
45
07
22
46
42
24
58
56
52
49
43
33
12
05
01
10
13
41
47
50
53
57
59
55
35
Diagnosis
Cancer
EIN
Benign, non-EIN
No Data
T S R N H J D K B I C G O P Q M E FAL Expert
Concensus
GREEN
YELLOW
RED
Usubutun et al, 2012
Discordant Cases
Defines Pathologist Style
Ref Dx Red Green
Explanation
OverDx UnderDx
EIN
EIN
B9
Small focus, EIN in anov
Polyp; Loose, Subtle
B9
EIN
B9
Fragment, Shattered, Thick
PTEN & mutations
PAX2 (10q24)
• Transcription factor
• Embryonic expression required for:
Kidney
Mesonephric structures
Paramesonephric ducts
• 5-fold reduced in endometrial Ca
PAX2 knockout leads to
Mullerian and renal atresia
Dressler, 199
EIN:
EIN:
08-111
H&E
PTEN
EIN: PAX2
Coinactivated PAX2 & PTEN in EIN
H&E
PTEN
PAX2
normal
background
+
+
*
+
*
*
*
*
+
+
*
+
EIN
08-111
Mutter, 2010
PAX2 & PTEN null rates
by dx
PE
(normal)
PAX2 null
36%
PTEN null
49%
Both express
36%
EIN
Cancer
71%
44%
15%
77%
68%
10%
Mutter, 2010
True
Precancer = EIN
kappa=0.54-0.62
Target
Interobserver
Reproducibility
Improved
Atypical
Hyperplasia
kappa=0.34-0.47
WHO-2014 (New)
Endometrial Hyperplasia
System
Benign
Hyperplasia without atypica
Precancer
Atypical hyperplasia /
Endometrioid Intraepithelial Hyperplasia
Acknowledgements
• Jan Baak, MD, Professor of Pathology
Stavanger University Hospital, Norway
University of Munich, Germany
• George Mutter, MD, Professor of Pathology
Brigham & Womens’ Hospital
Harvard University Medical School
Also see www.endometrium.org