Results Purpose Methods Purpose
Significance of Drusen Regression in Intermediate Age-‐Related Macular CLINICAL & T RANSLATIONAL DegeneraFon (AMD) in Progression to Advanced Disease MEDICINE
Tanya Glaser, E. Lauren Doss, Elvira Agrón, Divya Nigam, Emily Y. Chew, Wai T. Wong NaPonal Eye InsPtute, NaPonal InsPtutes of Health, Bethesda, MD
Results
Purpose
1. IdenFficaFon of Drusen Regression in Study Eyes From Baseline to Year 2: Out of 58 eyes with intermediate To invesFgate the prognosFc implicaFons of drusen regression (DrR) in eyes with intermediate AMD for progression to advanced AMD and development of geographic AMD and large drusen, 31 (53%) were scored as having DrR between baseline and year 2. atrophy. 2. Progression of AMD in Study Eyes from Year 2 to Year 5 Figure 1. ProporFons of eyes in groups with drusen P r o g r e s s io n t o A d v a n c e d A M D f r o m Y e a r 2 to Y e a r 5 regression (with DrR) and without drusen regression Study populaFon (without DrR) progressing to advanced AMD or CNV ParPcipants at the NEI site of AREDS2, age 50+ yrs and with large drusen (≥125 µm) in both eyes or CGA developing new GA lesion from year 2 to year 5. All large drusen in one eye and advanced disease in the fellow eye were prospecPvely followed with N e w N o n -c e n tra l eyes that progressed to CGA had DrR at year 2. Equal G A L e s io n fundus imaging for five years. Eyes without advanced AMD at baseline and not progressing to N o E v e n ts proporPons of eyes with DrR and without DrR advanced disease by year 2 were idenPfied as study eyes and analyzed, with each parPcipant progressed to CNV by year 5 contribuPng at most one eye. Table 2. DescripPon of the Image analysis to idenFfy eyes with drusen regression fates of study. Eyes Annual mulPmodal images were spaPally aligned using i2k RePna Solware (DualAlign LLC, Clilon categorized as having progressed to advanced AMD Park). The occurrence and locaPon of significant DrR (area > circle C1 of diameter 125 µm) Table 2. Progression of Study Eyes to Advanced AMD or new non-‐central GA include eyes with CGA or CNV. between baseline
Methods
P e rc e n t o f E y e s
100
80
60
40
20
0
D rR
Total Eyes w/ # Eyes DrR
Eyes progressing to Advanced AMD DrR
Non-‐ AAMD Eyes developing new Eyes Eyes w/ DrR non-‐central GA Lesion
No DrR
No DrR
DrR
No DrR
45
23/45 (56%)
1/23 (4%)
0/22 (0%)
4/30(13%)
0/27 (0%)
Yr 4 55 28/55 (51%) 9/28 (32%)
4/27 (15%)
42
20/42 (48%)
8/20(40%)
4/22 (18%)
12/28 (42%)
4/27 (15%)
Yr5
5/26 (19%)
40
19/40 (48%)
11/19 (58%)
4/21 (19%)
15/27 (56%)
4/26 (15%)
53 27/53 (51%) 9/27 (33%)
Eyes categorized as having new GA lesion developed new patch of atrophy >175µm during the study. For the atrophy analysis eyes were excluded eyes that developed CNV over the study Pme period. AbbreviaPons : Non-‐ AAMD = non advanced AMD eyes (no CNV or CGA)
3. Progression to advanced AMD associated with Drusen Regression: • Significantly more eyes in the DrR progressed to CGA at Years 3, 4, and 5 compared with those in the groups with No DrR. • Equal proporPons of eyes in the DrR group and the No DrR group progressed to CNV at Years 3, 4, and 5. • Hazard raPos for progression associated with drusen regression (adjusPng for age/sex): To advanced AMD (CGA or CNV) = 1.68 (CI: 0.54 – 5.25) and to CNV = 0.88 (CI: 0.25 – 3.) (DrR relaPve to No DrR). P r o g r e s s io n t o C G A o r D e v e l o p m e n t o f N e w N o n -c e n tra l G A (C N V c e n s o re d ) 100 90 80
L o g - r a n k (M a n te l- C o x ) te s t P v a lu e = 0 .0 0 2 3
70
H a z a r d R a tio (M a n t e l-H a e n s z e l)
60
H R = 4 .1 ( 9 5 % C I 1 .7 - 1 0 .2 )
N o D rR D rR
50 40 30 20
0 0 .0 Atrophy minus DrR ROIs
DrR
2/27 (7%)
10
IdenFfied DrR ROIs
Eyes progressing to CGA or developing new non-‐ central GA lesion
Yr 3 57 30/57 (53%) 7/30 (23%)
E v e n t R a te
and year 2 were detected and outlined as regions of interest (ROIs) using computer-‐ Baseline RF Subtracted Images Year 2 RF assistedimage analysis. Binary Image IdenFfied DrR ROIs Para-‐ROI SpaFal correlaFon of areas of drusen regression with areas of atrophy or CNV For eyes progressing to atrophy or advanced AMD, areas of atrophy on mFC FAF imaging and areas of late fluorescein angiography leakage were delineated by semi-‐automated planimetric techniques. Emerging areas of clinical atrophy and decreased FAF > 175µm in diameter were marked and Baseline FAF Outlined area of Atrophy 5yr FAF delineated as ROIs.
No D rR
0 .5
1 .0
1 .5
2 .0
2 .5
3 .0
T im e ( Y e a r s ) F o llo w in g Y e a r 2 v is it
3 .5
4. Development of new non-‐central GA associated with Drusen Regression: • Significantly more eyes with DrR developed new areas of non-‐central GA at Years 3, 4, and 5 compared with those in the groups No DrR. • Hazard raFo (non-‐adjusted) for developing either CGA or new non-‐central GA lesion: 4.1 (CI: 1.7 – 10.2, p = 0.0023). Figure 2. Kaplan-‐Meier analysis for the development of new atrophy (defined as progression to CGA or development of a new non-‐central GA lesion) between year 2 and year 5. PaFents censored at end of follow-‐up or aGer development of CNV.
Results
SpaFal CorrelaFon Between Regions of Drusen Regression and Regions of Atrophy Atrophy ROIs Overlap DrR ROIs Methods for ColocalizaFon calculaFon : 1
1. Regions of drusen regression and rePnal atrophy were delineated within a field comprising of a 6mm diameter circle centered on the fovea and their total areas computed. 2. p(DrR) was calculated as the proporPon of total DrR area within the 6mm field 3. p(Atrophy) was calculated as the proporPon of total atrophy area within the 6mm field 4. The colocalizaPon parameter, C, of Atrophy to DrR was defined as the area of overlap between areas of DrR and atrophy, divided by the total area of atrophy: 5. If DrR and Atrophy occur independently, the probability that any pixel is located within both DrR and Atrophy is 6. If DrR and Atrophy occur independently then the probability of overlap is 1 (no greater than chance): 7. Simplified as:
NCC = the normalized colocalizaFon coefficient. NCC > 1 indicates that colocalizaFon of DrR and atrophy is greater than chance.
• The mean percentage of DrR areas that developed subsequent atrophy (n=15) was 26±13% and the mean probability of colocalizaFon of DrR and atrophy relaFve to chance was 6.5 (95% CI: 1.2 – 11.7, range 0 – 41.7). • The mean percentage of DrR areas that developed CGA (n=4) was 55±27% and the mean probability of colocalizaPon of DrR and CGA relaPve to chance was 13.0 (range 1.0 – 41.7).
Conclusion § Drusen regression occurs commonly in eyes with intermediate AMD. § The occurrence of drusen regression does not appear to be associated with an increased short-‐term risk of progressing to CNV. § The occurrence of drusen regression is associated with an increased short-‐term risk of progressing to CGA or of developing new areas of non-‐central GA. § The high probability of overlap between areas of CGA and DrR suggests that areas of DrR may be locally predisposed to the formaFon of CGA, and that drusen regression may consFtutes an early step in local events leading to AMD advancement.
References: 1. Smith RT, Chan JK, Busuoic M, Sivagnanavel V, Bird AC, Chong NV. Autofluorescence characterisPcs of early, atrophic, and high-‐risk fellow eyes in age-‐related macular degeneraPon. Invest Ophthalmol Vis Sci 2006;47:5495-‐504.
ACKNOWLEDGEMENTS Support in part by grants from Medical Research Scholars Program (MRSP) at the NaPonal InsPtutes of Health
Tanya Glaser, E. Lauren Doss, Elvira Agrón, Divya Nigam, Emily Y. Chew, Wai T. Wong NaPonal Eye InsPtute, NaPonal InsPtutes of Health, Bethesda, MD
Results
Purpose
1. IdenFficaFon of Drusen Regression in Study Eyes From Baseline to Year 2: Out of 58 eyes with intermediate To invesFgate the prognosFc implicaFons of drusen regression (DrR) in eyes with intermediate AMD for progression to advanced AMD and development of geographic AMD and large drusen, 31 (53%) were scored as having DrR between baseline and year 2. atrophy. 2. Progression of AMD in Study Eyes from Year 2 to Year 5 Figure 1. ProporFons of eyes in groups with drusen P r o g r e s s io n t o A d v a n c e d A M D f r o m Y e a r 2 to Y e a r 5 regression (with DrR) and without drusen regression Study populaFon (without DrR) progressing to advanced AMD or CNV ParPcipants at the NEI site of AREDS2, age 50+ yrs and with large drusen (≥125 µm) in both eyes or CGA developing new GA lesion from year 2 to year 5. All large drusen in one eye and advanced disease in the fellow eye were prospecPvely followed with N e w N o n -c e n tra l eyes that progressed to CGA had DrR at year 2. Equal G A L e s io n fundus imaging for five years. Eyes without advanced AMD at baseline and not progressing to N o E v e n ts proporPons of eyes with DrR and without DrR advanced disease by year 2 were idenPfied as study eyes and analyzed, with each parPcipant progressed to CNV by year 5 contribuPng at most one eye. Table 2. DescripPon of the Image analysis to idenFfy eyes with drusen regression fates of study. Eyes Annual mulPmodal images were spaPally aligned using i2k RePna Solware (DualAlign LLC, Clilon categorized as having progressed to advanced AMD Park). The occurrence and locaPon of significant DrR (area > circle C1 of diameter 125 µm) Table 2. Progression of Study Eyes to Advanced AMD or new non-‐central GA include eyes with CGA or CNV. between baseline
Methods
P e rc e n t o f E y e s
100
80
60
40
20
0
D rR
Total Eyes w/ # Eyes DrR
Eyes progressing to Advanced AMD DrR
Non-‐ AAMD Eyes developing new Eyes Eyes w/ DrR non-‐central GA Lesion
No DrR
No DrR
DrR
No DrR
45
23/45 (56%)
1/23 (4%)
0/22 (0%)
4/30(13%)
0/27 (0%)
Yr 4 55 28/55 (51%) 9/28 (32%)
4/27 (15%)
42
20/42 (48%)
8/20(40%)
4/22 (18%)
12/28 (42%)
4/27 (15%)
Yr5
5/26 (19%)
40
19/40 (48%)
11/19 (58%)
4/21 (19%)
15/27 (56%)
4/26 (15%)
53 27/53 (51%) 9/27 (33%)
Eyes categorized as having new GA lesion developed new patch of atrophy >175µm during the study. For the atrophy analysis eyes were excluded eyes that developed CNV over the study Pme period. AbbreviaPons : Non-‐ AAMD = non advanced AMD eyes (no CNV or CGA)
3. Progression to advanced AMD associated with Drusen Regression: • Significantly more eyes in the DrR progressed to CGA at Years 3, 4, and 5 compared with those in the groups with No DrR. • Equal proporPons of eyes in the DrR group and the No DrR group progressed to CNV at Years 3, 4, and 5. • Hazard raPos for progression associated with drusen regression (adjusPng for age/sex): To advanced AMD (CGA or CNV) = 1.68 (CI: 0.54 – 5.25) and to CNV = 0.88 (CI: 0.25 – 3.) (DrR relaPve to No DrR). P r o g r e s s io n t o C G A o r D e v e l o p m e n t o f N e w N o n -c e n tra l G A (C N V c e n s o re d ) 100 90 80
L o g - r a n k (M a n te l- C o x ) te s t P v a lu e = 0 .0 0 2 3
70
H a z a r d R a tio (M a n t e l-H a e n s z e l)
60
H R = 4 .1 ( 9 5 % C I 1 .7 - 1 0 .2 )
N o D rR D rR
50 40 30 20
0 0 .0 Atrophy minus DrR ROIs
DrR
2/27 (7%)
10
IdenFfied DrR ROIs
Eyes progressing to CGA or developing new non-‐ central GA lesion
Yr 3 57 30/57 (53%) 7/30 (23%)
E v e n t R a te
and year 2 were detected and outlined as regions of interest (ROIs) using computer-‐ Baseline RF Subtracted Images Year 2 RF assistedimage analysis. Binary Image IdenFfied DrR ROIs Para-‐ROI SpaFal correlaFon of areas of drusen regression with areas of atrophy or CNV For eyes progressing to atrophy or advanced AMD, areas of atrophy on mFC FAF imaging and areas of late fluorescein angiography leakage were delineated by semi-‐automated planimetric techniques. Emerging areas of clinical atrophy and decreased FAF > 175µm in diameter were marked and Baseline FAF Outlined area of Atrophy 5yr FAF delineated as ROIs.
No D rR
0 .5
1 .0
1 .5
2 .0
2 .5
3 .0
T im e ( Y e a r s ) F o llo w in g Y e a r 2 v is it
3 .5
4. Development of new non-‐central GA associated with Drusen Regression: • Significantly more eyes with DrR developed new areas of non-‐central GA at Years 3, 4, and 5 compared with those in the groups No DrR. • Hazard raFo (non-‐adjusted) for developing either CGA or new non-‐central GA lesion: 4.1 (CI: 1.7 – 10.2, p = 0.0023). Figure 2. Kaplan-‐Meier analysis for the development of new atrophy (defined as progression to CGA or development of a new non-‐central GA lesion) between year 2 and year 5. PaFents censored at end of follow-‐up or aGer development of CNV.
Results
SpaFal CorrelaFon Between Regions of Drusen Regression and Regions of Atrophy Atrophy ROIs Overlap DrR ROIs Methods for ColocalizaFon calculaFon : 1
1. Regions of drusen regression and rePnal atrophy were delineated within a field comprising of a 6mm diameter circle centered on the fovea and their total areas computed. 2. p(DrR) was calculated as the proporPon of total DrR area within the 6mm field 3. p(Atrophy) was calculated as the proporPon of total atrophy area within the 6mm field 4. The colocalizaPon parameter, C, of Atrophy to DrR was defined as the area of overlap between areas of DrR and atrophy, divided by the total area of atrophy: 5. If DrR and Atrophy occur independently, the probability that any pixel is located within both DrR and Atrophy is 6. If DrR and Atrophy occur independently then the probability of overlap is 1 (no greater than chance): 7. Simplified as:
NCC = the normalized colocalizaFon coefficient. NCC > 1 indicates that colocalizaFon of DrR and atrophy is greater than chance.
• The mean percentage of DrR areas that developed subsequent atrophy (n=15) was 26±13% and the mean probability of colocalizaFon of DrR and atrophy relaFve to chance was 6.5 (95% CI: 1.2 – 11.7, range 0 – 41.7). • The mean percentage of DrR areas that developed CGA (n=4) was 55±27% and the mean probability of colocalizaPon of DrR and CGA relaPve to chance was 13.0 (range 1.0 – 41.7).
Conclusion § Drusen regression occurs commonly in eyes with intermediate AMD. § The occurrence of drusen regression does not appear to be associated with an increased short-‐term risk of progressing to CNV. § The occurrence of drusen regression is associated with an increased short-‐term risk of progressing to CGA or of developing new areas of non-‐central GA. § The high probability of overlap between areas of CGA and DrR suggests that areas of DrR may be locally predisposed to the formaFon of CGA, and that drusen regression may consFtutes an early step in local events leading to AMD advancement.
References: 1. Smith RT, Chan JK, Busuoic M, Sivagnanavel V, Bird AC, Chong NV. Autofluorescence characterisPcs of early, atrophic, and high-‐risk fellow eyes in age-‐related macular degeneraPon. Invest Ophthalmol Vis Sci 2006;47:5495-‐504.
ACKNOWLEDGEMENTS Support in part by grants from Medical Research Scholars Program (MRSP) at the NaPonal InsPtutes of Health
