대한안과학회지 213 년제 54 권제 6 호 J Korean Ophthalmol Soc 213;54(6):92-912 pissn: 378-6471 eissn: 292-9374 http://dx.doi.org/1.3341/jkos.213.54.6.92 = 증례보고 = 스펙트럼영역빛간섭단층촬영유의지도에서자동으로구한국소적망막신경섬유층결손면적 서샘 신중원 엄기방 한양대학교의과대학안과학교실 목적 : 자동으로구한국소적망막신경섬유층결손면적의유용성을알아보았다. 대상과방법 : 국소적망막신경섬유층결손이있는 51 명과정상인 53 명을대상으로스펙트럼영역빛간섭단층촬영의유의지도에서결손면적 (p 1%, p 5%) 을 Image J 로수작업으로, Matlab 을이용하여자동으로측정하였다. 사진과지도에서결손면적, 망막신경섬유층두께, 시신경유두측정치, 황반부내망막두께의 AUC (area under the receiver operating characteristic curve) 를구하였다. 결과 : 유의지도에서수기와자동으로구한결손면적간에는상관관계가높았다 (p 1% r=.94, p 5% r=.89). 유의지도에서수기와자동으로구한결손면적 (p 5%) 의 AUC ( 각각.987,.966) 는유의한차이가없었다. 후자의 AUC 는하사분면망막신경섬유층두께 (.936) 보다조금컸으나유의한차이는없었고, 하반부신경절세포 + 내망상층두께 (.894), 수직유두함몰비 (.869) 보다컸으며유의하였다 (p.5). 결론 : 유의지도에서결손면적의자동측정은황반부내망막과시신경유두의측정치보다결손의검출력이우월하였다. < 대한안과학회지 213;54(6):92-912> 망막신경섬유층결손을발견하기위하여망막신경섬유층사진이임상에서널리사용되나안저의색소침착이적은무뉘안저 (tessellated fundus) 이거나광범위하게망막신경섬유층이얇으면망막신경섬유층결손을보기가어렵다. 매체가투명해야잘보이고, 양질의상을얻으려면동공을최대로산동시켜야한다. 또한상의해석은주관적이며전문가간에도평가에있어서상당한차이가있다. 반면에빛간섭단층촬영 (optical coherence tomography, OCT) 은동공의직경이 5 mm 정도면가능하고촬영후바로정량적인망막신경섬유층측정치를알수있다. 스펙트럼영역빛간섭단층촬영기는시간영역빛간섭단층촬영기에비하여스캔속도가빠르고축해상도가높은영상을획득할수있다. 1 스펙트럼영역빛간섭단층촬영기 (3D OCT-2, Topcon, Tokyo, Japan) 는망막신경섬유층두께지도 (thickness map) 와기기에내장된정상기준자료와비교하여 1% 미만은적색, 5% 미만은황색으로결손이있는부위를색상화 Received: 212. 11. 3. Revised: 212. 12. 28. Accepted: 213. 4. 8. Address reprint requests to Ki Bang Uhm, MD, PhD Department of Ophthalmology, Hanyang University Seoul Hospital, #222 Wangsimni-ro, Seongdong-gu, Seoul 133-791, Korea Tel: 82-2-229-857, Fax: 82-2-2291-8517 E-mail: KBUhm@hanyang.ac.kr * 이논문의요지는 213 년대한안과학회제 19 회학술대회에서구연으로발표되었음. 한유의지도 (significance map)(cirrus 스펙트럼영역빛간섭단층촬영기의망막신경섬유층두께 deviation map 과유사함 ) 를제공한다. 이지도들에서망막신경섬유층결손의형태와위치는망막신경섬유층사진과흡사하며, 녹내장진단에유용하다고보고되어있다. 2-5 이전의연구들에서는망막신경섬유층결손을정량적으로분석하기위하여대부분망막신경섬유층결손의폭과위치를사용하여분석하였으나, 6-9 본연구에서는유의지도의망막신경섬유층결손면적을자동으로측정하는프로그램을이용하여결손면적을구하였고, 유용성을알아보기위하여수작업으로측정한결손면적과관계를알아보았으며, 유두주위망막신경섬유층두께, 시신경유두, 황반부내망막측정치들의진단력과비교해보았다. 대상과방법 본연구는전향적연구로 212년 4월부터 212년 9월까지한양대학병원안과를내원한환자들중안과검사에서정상소견을보인정상군 53명 53안과국소적망막신경섬유층결손을보이는녹내장군 51명 51안을대상으로하였다. 모든대상자에서병력조사, 시력측정, 골드만안압측정, 세극등을이용한전안부검사, 앞방각경검사, 초음파를이용한중심각막두께측정 (Sp-3, Tomey Co., Nagoya, Japan), 검안경을이용한시신경유두검사, Humprey field analyzer 92
- 서샘외 : 망막신경섬유층결손면적의자동측정 - (Carl Zeiss Meditec, Dublin, CA, USA) 를이용한시야검사, 주사레이저검안경을이용한 6도망막신경섬유층사진 (F-1, Nidek Inc., Japan), 시신경유두사진, 3D OCT- 2 (software version 7.11) 촬영을하였다. 녹내장안은녹내장성시신경손상과시야검사상녹내장성시야결손을보이고국소적망막신경섬유층결손이있었다. 녹내장성시신경손상은국소적이거나전반적인시신경테의얇음또는패임이있거나양안의수직유두함몰비차이가.2 이상, 망막신경섬유층결손, 시신경유두출혈이있는경우였다. 시야결손전녹내장은국소적망막신경섬유층결손이있고시야결손이없었다. 정상안은안압이 21 mmhg 이하이며안압상승의기왕력이없었고, 위에기술한녹내장성시신경손상및시야검사상녹내장성시야결손이없었다. 구면렌즈대응치가 -6.디옵터이하, +3. 디옵터이상인경우와당뇨망막병증등망막질환, 포도막염, 시야나시신경에영향을줄만한내과또는신경과질환이있는환자는제외하였다. 녹내장성시야결손은 Humphrey 자동시야계의 Swedish interactive threshold algorithm (SITA) 의 central 3-2 검사에서최소한 2회이상 glaucoma hemifield test가 outside normal limit 이거나, pattern standard deviation (PSD) 이 5% 미만이거나, pattern deviation plot에서가장자리를제외한부위에서인접한 3개이상점의역치가정상의 5% 미만으로나타나고그중한개이상은 1% 미만인경우로정하였다. 주시상실 2% 미만, 가양성과가음성반응이 3% 미만으로신뢰할만한결과와재현성있는경우를대상으로하였다. 충분히산동한후동일한검사자가검사를시행하였다. 3D OCT 는광원이 84 nm 파장이며, 축해상도는 5 μm이다. 3D disc protocol로시신경유두를중심으로가로세로 6 mm 사각형부위를 512A 128B 스캔을하였다. 스캔속도가초당 5, A-scan으로시간영역빛간섭단층촬영기보다빠르므로눈의움직임으로인한오류가적다. 소프트웨어가시신경유두중심을결정하고이곳을중심으로한직경 3.4 mm원의망막신경섬유층두께와시신경유두측정치들이자동으로측정된다. 4개의사분면, 12개의시구역, 평균시신경유두주위망막신경섬유층두께와내장된프로그램에서자동으로생성되는망막신경섬유층두께지도와유의지도를분석에사용하였다. 시신경유두측정치로는유두함몰면적, 시신경테면적, 유두함몰면적비 (cup/disc area ratio), linear cup to disc ratio, 수직유두함몰비, 유두함몰용적, 시신경테용적을분석에이용하였다. 또한 3D macular vertical scan으로황반부가로세로 7 mm 사각형부위를 512A 128B 스캔을하여황반부내망막두께를측정하였다. 황반부 6 6 mm 2 에서구한신경섬유층, 신경절세포층, 내망상층의상반부, 하반부, 전체측정치를분석에이용하였다. Quality factor 값이 5 미만이거나중심이이탈되어있는경우는대상에서제외하였다. 12개의시구역에서망막신경섬유층결손의위치는우안을기준으로표시하였고 9시는우안의이측이다. 망막신경섬유층사진에서나타나는국소적인망막신경섬유층결손의위치, 폭을측정하였고, 면적은 image J를이용하여수작업으로측정하였다. 두께지도는시신경유두를중심으로한 6 6 mm 2 부위에서망막신경섬유층두께를그에해당되는컬러스케일로나타낸다. 두께지도에서결손의정의는시신경유두연에서 1유두직경떨어진지점에서망막의가장큰 A B Figure 1. (A) The cslo RNFL photographs were overlaid with a color disc photograph (5 5 mm 2 ) as a reference image that was obtained by Topcon 3D OCT-2. (B) The retinal blood vessels (arrow) and RNFL defects (arrow head) were matched between the cslo RNFL photograph and the reference image. cslo = confocal scanning laser ophthalmoscope; RNFL = retinal nerve fiber layer; OCT = optical coherence tomography; T = temporal; N = nasal. 93
- 대한안과학회지 213 년제 54 권제 6 호 - RNFL significance map RNFL thickness map Disc photograph cslo RNFL photograph A RNFL significance map RNFL thickness map cslo RNFL photograph Autodetection B Figure 2. (A) Original photograph in the retinal nerve fiber layer (RNFL) significance map, RNFL thickness map, optic disc photograph, and cslo RNFL photograph. The RNFL significance map, RNFL thickness map, and optic disc photograph were obtained by Topcon 3D OCT-2, and the RNFL photograph was captured by a Nidek F-1 cslo. (B) Each boundary of the RNFL defects was manually delineated. The areas of the RNFL defects were measured on the 5 5 mm 2 parapapillary area using Image J software (yellow lines) manually. Superpixels coded in red, red and yellow were automatically calculated by a Matlab computer program according to the red, green, and blue values in the 5 5 mm 2 circumpapillary region. Measurement of the area of localized RNFL defect in the significance map (red area = 1.11 mm 2, red and yellow area = 2.79 mm 2 ), thickness map (1.4 mm 2 ), cslo RNFL photograph (1.77 mm 2 ) manually, and automated measurement using significance map (red area =.84 mm 2, red and yellow area = 3.59 mm 2 ). cslo = confocal scanning laser ophthalmoscope; RNFL = retinal nerve fiber layer; OCT = optical coherence tomography. 혈관직경보다너비가크며쐐기모양을보이고색의급격한변화를보이는부위로하였다. 유의지도에서망막신경섬유층결손은황색으로표시되는정상기준자료의 5% 미만부위또는적색으로표시되는 1% 미만부위가임의로연속적으로 5개이상연결되었을때로정의하였다. 망막신경섬유층사진, 빛간섭단층촬영의두께지도와유의지도의 5 5 mm 2 크기의같은부위를비교하기위해망막신경섬유층사진과시신경유두중심이표시되고 5 5 mm 2 크기인 3D OCT 시신경유두컬러사진을겹쳐놓고, 시신경유두컬러사진을기준으로혈관주행과망막신경섬유층결손에맞춰영상을조정하였다 (Fig. 1). 망막신경섬유층사진, 두께지도, 유의지도에서시신경유두를중심으로한유두주위가로 5 mm, 세로 5 mm의사각형부위내의결손면적을 image J를이용하여수작업으로측정하였다 (Fig. 2). Image J에서의측정은 pixel 단위로이루어지기때문에, 이를 mm 단위 로변환하기위해 image J를통해시신경유두의면적을측정한후, 빛간섭단층촬영에서제공하는시신경유두면적측정값과비교하여변환비율을결정하였다. Matlab R21a (The Math-Works, Inc., Natick, MA, USA) 컴퓨터프로그램을이용하여 5 5 mm 2 의유의지도에서망막신경섬유층이정상기준자료의 5% 미만을나타내는황색, 1% 미만을나타내는적색부위의결손면적을자동으로측정하였다 (Fig. 2). 지도상에서비정상으로표시되는부분만을선택적으로골라내어측정하기위해 Matlab 을이용하여적색, 녹색, 청색의정보를분리하여추출할수있으며, 각각의정보차이를계산했을경우시신경유두와유두주위정상범위의망막신경섬유층인회백색부분을제거할수있다. 선택된구역의 pixel 개수의합이곧면적이다. 측정된유의지도의실제면적은 6 6 mm 2 에해당하며 54 54 pixels의사진으로표현되기때문에, 면적 94
- 서샘외 : 망막신경섬유층결손면적의자동측정 - 계산에있어 6 mm=54 pixels 의변환비율을적용하여길이단위로변환이가능하다. 사진에서의국소적망막신경섬유층결손을기준으로정상군과녹내장군으로분류하였다. 정상군과녹내장군간의수치자료비교는 unpaired t-test 를, 범주자료는 Chi-square test를사용하였다. 유의지도에서수동과자동으로구한망막신경섬유층결손면적은 paired t-test 로비교하였고, 두측정치들의관계를알아보기위하여 Pearson 상관분석을하였다. 유의지도에서수동과자동으로구한망막신경섬유층결손면적은 Bland-Altman plot을사용하여일치도를알아보았다. 정상군과녹내장군에서망막신경섬유층사진과두께지도를이용하여수동으로측정한망막신경섬유층결손면적, 유의지도에서수동과자동으로측정한적색면적, 적색과황색면적, 유두주위망막신경섬유층두께, 시신경유두황반부내망막층측정치들의국소적망막신경섬유층결손검출력을알아보고자 AUC (area under the receiver operating characteristic curve) 와민감도, 특이도를구하였다. AUC 차이의비교는 MedCalc version 9.3.7. (MedCalc Software, Mariakerke, Belgium) 을이용하였다. 통계학적분석은 MedCalc와 SPSS version 13. (SPSS Inc., Chicago, Illinois, USA) 을이용하였고 p값이.5 미만일경우통계적으로유의한것으로간주하였다. 결과 정상군 53명과국소적망막신경섬유결손을보이는녹내장군 51명총 14명 14안을대상으로하였으며, 평균연령은각각 56.3 ± 1.7, 57.7 ± 13.7세였다. 녹내장군은시야결손전녹내장 7안, 시야결손녹내장 44안이있었다. 두군간에연령, 성별, 좌우안, 굴절이상, 중심각막두께는유의한차이가없었다. Humphrey 자동시야검사상정상군과녹내장군에서 mean deviation (MD) 은각각 -1.45 ±.91 Table 1. Demographics and characteristics of subjects with localized RNFL defects and normal controls Normal eyes Eyes with RNFL defects (n = 53) (n = 51) p-value Age (years) 56.3 ± 1.7 57.7 ± 13.7.54 * Gender (male:female) 28:25 29:22.68 Eye (right:left) 24:29 21:3.67 Spherical equivalent (diopters) -.15 ± 1.29 -.65 ± 2.5.15 * Central corneal thickness (μm) 543.78 ± 32.17 539.33 ± 31.98.5 * Mean deviation (db) -1.45 ±.91-6.15 ± 4.33 <.1 * Pattern standard deviation (db) 1.58 ±.69 6.8 ± 4.2 <.1 * Type of glaucoma NA Preperimetric 7 Perimetric 44 Localized RNFL defects NA 64 Superior 11 Inferior 27 Superior and inferior 13 Values are presented as number or mean ± SD. RNFL = retinal nerve fiber layer; NA = not applicable. Unpaired t-test; Chi-square test. Table 2. Area under the receiver operating characteristic curves (AUCs) of the red-free photograph, thickness map, and significance map between normal controls and eyes with RNFL defects Normal (n = 53) Eyes with RNFL defects (n = 51) p-value * AUCs (95% CI) Sn/Sp (Sp 8%) Sn/Sp (Sp 9%) Defect area (mm 2 ) Photographic (manual) 2.41 ± 1.46 <.1 1. (1.-1.) NA 1/1 Thickness map (manual) 2.4 ± 1.35 <.1 1. (1.-1.) NA 1/1 Significance map red (manual) 1.65 ± 1.17 <.1.961 (.924-.998) NA 92/1 Significance map red & yellow (manual).1 ±.1 2.77 ± 1.83 <.1.987 (.967-1.) NA 98/98 Significance map red (automatic).1 ±.4 1.7 ± 1.43 <.1.963 (.925-1.) 94/85 94/92 Significance map red & yellow (automatic).16 ±.32 3.15 ± 2.25 <.1.966 (.932-1.) 98/83 9/94 RNFL = retinal nerve fiber layer; CI = confidence interval; Sn = sensitivity; Sp = specificity; NA = not applicable. * Unpaired t-test. 95
- 대한안과학회지 213 년제 54 권제 6 호 - Table 3. Area under the receiver operating characteristic curves (AUCs) of circumpapillary RNFL thickness between normal controls and eyes with RNFL defects Normal Eyes with RNFL p-value * Sn/Sp Sn/Sp AUCs (95% CI) (n = 53) defects (n = 51) (Sp 8%) (Sp 9%) Circumpapillary RNFL thickness (μm) Average 14.2 ± 8.1 83.3 ± 12.8 <.1.917 (.864-.97) 78/81 76/92 Quadrant Temporal 79.2 ± 1.6 69.3 ± 12.2 <.1.726 (.628-.823) 55/81 35/91 Superior 122.7 ± 12.6 11.7 ± 18.3 <.1.863 (.791-.935) 8/85 67/91 Nasal 84.7 ± 14.4 74.5 ± 15.8.1.71 (.68-.811) 51/81 39/91 Inferior 129.2 ± 1.8 9.3 ± 22.2 <.1.936 (.89-.981) 88/81 82/91 Clock-hour 9 (temporal) 68. ± 1.9 64.3 ± 12.6.12.62 (.492-.713) 35/81 24/91 1 91.2 ± 12.8 77.1 ± 19.4 <.1.719 (.618-.82) 57/81 47/91 11 127.2 ± 17.5 99.6 ± 29.2 <.1.786 (.695-.877) 65/81 61/91 12 (superior) 122.6 ± 18.2 13.3 ± 22.9 <.1.771 (.68-.862) 51/81 39/92 1 122. ± 15.2 12.3 ± 19.3 <.1.795 (.79-.881) 65/81 49/91 2 99.1 ± 16.2 85.3 ± 2.5 <.1.718 (.617-.819) 61/81 33/92 3 (nasal) 74.4 ± 15.5 66.4 ± 15.3.9.667 (.562-.773) 41/81 2/92 4 8.3 ± 18.2 71.6 ± 19.6.2.644 (.537-.751) 41/81 2/91 5 114.1 ± 16.5 95.3 ± 2.5 <.1.767 (.674-.861) 67/83 41/91 6 (inferior) 134.4 ± 17.5 93.4 ± 25.7 <.1.896 (.836-.957) 8/83 75/91 7 138.7 ± 17.5 82. ± 34.6 <.1.97 (.845-.97) 84/85 82/94 8 78.3 ± 12.9 66.6 ± 16.2 <.1.736 (.637-.835) 63/81 49/92 RNFL = retinal nerve fiber layer; CI = confidence interval; Sn = sensitivity; Sp = specificity. * Unpaired t-test. db, -6.15 ± 4.33 db로유의한차이가있었고 (p<.1), PSD는각각 1.58 ±.69 db, 6.8 ± 4.2 db로유의한차이가있었다 (p<.1). 51명의녹내장군에서국소적망막신경섬유층결손은상측이 11안, 하측이 27안, 상하모두에존재하는경우가 13안이었다 (Table 1). 국소적망막신경섬유층결손은 7시방향에위치하는경우가 36개로가장많았으며다음으로는 11시방향이 17개로많이나타났다. 정상군과녹내장군간에수기로구한두께지도망막신경섬유층결손면적, 수기와자동으로구한유의지도의결손면적, 유두주위망막신경섬유층두께, 황반부내망막두께, 시신경유두측정치들은 9시망막신경섬유층두께, 시신경테용적을제외하고모두유의한차이를보였다 (p<.5) (Table 2-4). 수기와자동으로구한유의지도적색면적은유의한차이가없었고상관계수는.94, 수기와자동으로구한적색과황색면적은유의한차이가있었고 (p=.1) 상관계수는.89 이었다 (Fig. 3) (Table 5). Bland-Altman plot에서적색면적, 적색과황색면적모두 proportional bias가없었고, 적색과황색면적은자동이수기로구한면적보다컸다 (Fig. 4). 수기로구한망막신경섬유층사진과수기로구한유의지도의결손면적들과의상관계수는.92 이상, 수기로구한망막신경섬유층사진과자동으로구한유의지도의결손면적들과의상관계수는.798 이상으로밀접한연관성이있었다. 또한자동으로구한유의지도의결손면적과시야지표와의 상관계수 (MD<-.717, PSD>.78) 도높았다 (Table 6). 국소적망막신경섬유층검출에유용한측정치들의 AUC, 특이도, 특이도 9% 이상에서의민감도는다음과같다. 수기로구한두께지도결손면적의진단력 (1., 1%, 1%) 이가장높았고, 유의지도의적색과황색면적을수기로구한면적 (.987, 98%, 98%) 과자동으로구한면적 (.966, 94%, 9%), 유의지도의적색면적을자동으로구한면적 (.963, 92%, 94%) 과수기로구한면적 (.961, 1%, 92%) 순이었으며이들간에 AUC는차이가없었다 (p>.21) (Fig. 5, 6) (Table 2, 7). 유두주위망막신경섬유층두께들중하사분면 (.936, 91%, 82%), 황반부내망막측정치들중하반부신경절세포 + 내망상층두께 (.894, 91%, 71%), 시신경유두측정치들중수직유두함몰비 (.869, 91%, 73%) 가국소적망막신경섬유층결손의검출에가장유용하였다 (Table 3, 4). 자동으로구한유의지도적색과황색면적의 AUC (.966) 는유두주위하사분면망막신경섬유층두께의 AUC (.936) 보다컸으나유의성은없었고 (p=.22), 수동으로구한유의지도적색과황색면적의 AUC (.987) 는유두주위하사분면망막신경섬유층두께의 AUC (.936) 보다컸으며유의한차이가있었다 (p=.4). 자동으로구한유의지도적색과황색면적의 AUC (.966) 는하반부신경절세포 + 내망상층두께 (.894), 수직유두함몰비 (.869) 보다컸 96
- 서샘외 : 망막신경섬유층결손면적의자동측정 - Table 4. Area under the receiver operating characteristic curves (AUCs) of macular inner retina thickness and optic disc parameter between normal controls and eyes with RNFL defects Normal Eyes with RNFL p-value * Sn/Sp Sn/Sp AUCs (95% CI) (n = 53) defects (n = 51) (Sp 8%) (Sp 9%) Macular inner retina thickness (μm) NFL Superior 34.1 ± 4.9 29.6 ± 7.5 <.1.719 (.616-.822) 59/81 39/92 Inferior 37.6 ± 4.2 24.3 ± 1.5 <.1.859 (.778-.941) 75/83 73/92 Total 35.8 ± 4. 26.9 ± 7.6 <.1.846 (.762-.93) 78/83 73/94 GCL + IPL Superior 72.1 ± 5.7 65.9 ± 5.5 <.1.789 (.72-.876) 51/85 39/94 Inferior 7.5 ± 5.9 6.4 ± 6.2 <.1.894 (.832-.956) 75/85 71/91 Total 71.3 ± 4.8 63.2 ± 5. <.1.879 (.815-.943) 65/83 59/94 GCC Superior 16.2 ± 7.8 95.5 ± 12. <.1.77 (.678-.863) 61/83 39/91 Inferior 18. ± 8.3 84.6 ± 15.6 <.1.893 (.829-.958) 82/81 73/91 Total 17.2 ± 7.8 9.1 ± 11.9 <.1.875 (.86-.943) 82/81 67/92 Optic nerve head Cup area (mm 2 ) 1.6 ±.5 1.61 ±.68 <.1.732 (.636-.828) 47/83 43/91 Rim area (mm 2 ) 2.1 ±.47 1.24 ±.52 <.1.869 (.796-.941) 84/85 65/91 C/D area ratio.34 ±.14.56 ±.17 <.1.827 (.747-.97) 69/81 63/91 Linear CDR.57 ±.14.74 ±.13 <.1.824 (.744-.95) 69/81 61/91 Vertical CDR.58 ±.13.79 ±.14 <.1.869 (.798-.94) NA 73/91 Cup volume (mm 3 ).24 ±.18.35 ±.25.1.628 (.52-.736) 43/81 18/91 Rim volume (mm 3 ).17 ±.19.15 ±.17.56.562 (.448-.675) 37/83 22/91 RNFL = retinal nerve fiber layer; CI = confidence interval; Sn = sensitivity; Sp = specificity; NFL = nerve fiber layer; GCL = ganglion cell layer; IPL = inner plexiform layer; GCC = ganglion cell complex; CDR = cup to disc ratio; NA = not applicable. * Unpaired t-test. Table 5. Comparison of significance map area analyzed by manual and automatic method Manual Automatic Difference p-value * r p-value Significance map red area (mm 2 ).81 ± 1.16.84 ± 1.31 -.3 ±.56.54.94 <.1 Significance map red and yellow area (mm 2 ) 1.38 ± 1.89 1.64 ± 2.19 -.26 ± 1..9.89 <.1 * Paired t-test; Correlation. A Significance map red area (automatic) (sq mm) 8 7 6 5 4 3 2 1 1 2 3 4 5 1 2 3 4 5 6 7 Significance map red area (manual) (sq mm) Significance map red and yellow (manual) (sq mm) Figure 3. (A) Scatter plots showing the relationship of manually measured red area (p < 1%) of significance map against automatically measured red area (p < 1%) of significance map (r =.94, p <.1). (B) Relationship of manually measured red and yellow area (p < 5%) of significance map against automatically measured red and yellow area (p < 5%) of significance map (r =.89, p <.1). B Significance map red & yellow area (automatic) (sq mm) 12 1 8 6 4 2 97
- 대한안과학회지 213 년제 54 권제 6 호 - 4 3 RNFL defect No Yes A 7 6 5 RNFL defect No Yes B 2 1 +1.96 SD 1.1 4 3 2 +1.96 SD 2.2 Mean. 1 Mean.3-1 -1.96 SD -1.1-1 -2-1.96 SD -1.7-2 -3 1 2 3 4 5 6 7 2 4 6 8 1 Figure 4. Bland-Altman plots of agreement between manually and automatically measured (A) red area (p < 1%) and (B) red and yellow area (p < 5%) in the significance map. Differences between automatically and manually measured areas were plotted against the means. Dash lines represent the 95% confidence intervals. Table 6. Relationships (correlation coefficients) between retinal nerve fiber layer defect areas and structural-functional parameters RNFL defects area Red-free Thickness p < 1% p < 1 & 5% p < 1% p < 1 & 5% MD PSD Photograph Map (manual) (manual) (automatic) (automatic) (manual) (manual) RNFL defects area Photograph (manual).962.95.92.798.817 -.821.816 Thickness map (manual).962.929.94.818.831 -.822.836 p < 1% (manual).95.929.957.94.95 -.88.819 p < 1 & 5% (manual).92.94.957.842.89 -.788.792 p < 1% (automatic).798.818.94.842.97 -.717.78 p < 1 & 5% (automatic).817.831.95.89.97 -.723.721 VF indices MD -.821 -.822 -.88 -.788 -.717 -.723 -.862 PSD.816.836.819.792.78.721 -.862 All correlations are significant at the p <.1 level (two-tailed). RNFL = retinal nerve fiber layer; VF = visual field; MD = mean deviation; PSD = pattern standard deviation. 12 1 8 6 Significance map red & yellow area (automatic) (sq mm) 으며유의하였다 ( 각각 p=.18, p=.8). 유두주위하사분면망막신경섬유층두께의 AUC는하반부신경절세포 + 내망상층두께, 수직유두함몰비의 AUC와유의한차이는없었다 ( 각각 p=.15, p=.7). 하반부신경절세포 + 내망상층두께와수직유두함몰비의 AUC도유의한차이는없었다 (p=.45) (Fig. 6) (Table 7). 4 2 고 찰 No no RNFL defect Yes yes >.5464 Sens: 9.2 Spec: 94.3 Figure 5. The cut-off point (>.546 mm 2 ), sensitivity (9.2%), and specificity (94.3%) of automated measurements of significance map red and yellow area (p < 5%). 이전의대부분의연구들에서는망막신경섬유층결손을정량적으로분석하기위하여망막신경섬유층결손의폭과위치를사용하여분석하였으나, 6-9 스펙트럼영역빛간섭단층촬영기의두께지도와유의지도에서의결손은망막신경섬유층사진의망막신경섬유층결손형태와유사하게보이므로 98
- 서샘외 : 망막신경섬유층결손면적의자동측정 - Table 7. Probability values for the differences (upper triangle of numbers) between area under the receiver operating characteristic curves (AUC) Significance map red manual (.961) Significance map red & yellow manual (.987) Significance map red auto (.963) Significance map red & yellow auto (.966) RNFLT inferior quadrant (.936) Macular inner retina thickness Inferior GCL + IPL (.894) Vertical CDR (.869) Significance map red manual (.961) -.21.92.79.37.4.4 Significance map red & yellow manual (.987) -.26.31.4.5.2 Significance map red auto (.963) -.77.34.4.7 Significance map red & yellow auto (.966) -.22.2.8 RNFLT inferior quadrant (.936) -.15.7 Macular inner retina thickness Inferior GCL+IPL (.894) -.45 Vertical CDR (.869) - AUC = area under the receiver operating characteristic curve; RNFLT = retinal nerve fiber layer thickness; GCL = ganglion cell layer; IPL = inner plexiform layer; CDR = cup to disc ratio. Sensitivity Sigmap red & yellow area (p < 5%)(manual)(.987) Sigmap red & yellow area (p < 5%)(auto)(.966) Inferior quadrant RNFL thickness (.936) Inferior GCL + IPL thickness (.894) Vertical CDR (.869) 1 8 6 4 2 2 4 6 8 1 1-Specificity Figure 6. Receiver operating characteristics curves of the manually measured significance map red and yellow area (p < 5%), automatically measured significance map red and yellow area (p < 5%), circumpapillary inferior quadrant retinal nerve fiber layer (RNFL) thickness, inferior ganglion cell layer + inner plexiform layer thickness (GCL + IPL thickness), and vertical cup to disc ratio (CDR). 본연구에서는유의지도의망막신경섬유층결손면적을자동으로측정하는컴퓨터프로그램을이용하여측정한면적이국소적망막신경섬유층결손의검출에유용한지알아보았다. 본연구에서유의지도에나타난적색 (p<1%), 적색과황색면적 (p<5%) 을자동으로구한측정치는수기로구한측정치보다컸으나 p<5% 만유의한차이가있었다. 이는 Fig. 2와같이유의지도망막신경섬유층결손부위이외의황색부분에서더측정되어자동으로구한황색면적은수기로구한황색면적보다컸다고본다. 유의지도에서자동으로구한더넓은망막신경섬유층결손부위는눈에띄지않는망막신경섬유층이얇은부위라생각한다. Ye et al 1 도망막신경섬유층사진의결손면적은 Cirrus 스펙트럼영역빛단층촬영기의망막신경섬유층두께 deviation map 의내장된기기정상기준자료의 5% 미만인황색면적보다적다고하였으며, 이는망막신경섬유층사진에서보이는것보다스펙트럼영역빛간섭단층촬영기에서추가의망막신경섬유층결손을찾을수있음을의미한다고하였다. 또한자동으로구한유의지도에서결손면적이수동으로구한것보다큰것은위양성의가능성도있다고본다. 망막신경섬유층사진과시간영역빛간섭단층촬영기의망막신경섬유층결손의폭과위치는상관관계가높다고하였다. 6-9 스펙트럼영역빛간섭단층촬영기로측정한본연구에서도수기로구한유의지도적색면적과자동으로구한유의지도적색면적과의상관계수는.94, 수기로구한적색과황색면적과자동으로구한적색과황색면적의상관계수는.89으로연관성이높았다. 또한수기로구한망막신경섬유층사진과수기로구한유의지도의결손면적들과의상관계수는.92 이상, 수기로구한망막신경섬유층사진과자동으로구한유의지도의결손면적들과의상관계수는.798 이상으로밀접한연관성이있었다. 또한자동으로구한유의지도의결손면적과기능과의상관계수 (MD<-.717, PSD>.78) 도높아구조와기능은밀접한관련이있음을알수있었다. 이전에는망막신경섬유층사진이망막신경섬유층결손의평가에널리사용되었으나, 스펙트럼영역빛간섭단층촬영기로생성된유의지도의자동측정은망막신경섬유층결손을정량화하고시각화하여녹내장평가에이용이가능하다고본다. 99
- 대한안과학회지 213 년제 54 권제 6 호 - 시간영역빛간섭단층촬영기에서수기로구한망막신경섬유층결손의폭으로녹내장의진단능력을평가한논문들에서는빛간섭단층촬영의 12개시구역중사진의국소적망막신경섬유층결손에해당하는위치에 1시간이상비정상인 p<5% 를보이는경우결손을탐지한것으로간주하였고, 1개시구역이상에서 5% 이하의이상소견이있는경우 78.7-85.9% 의민감도와 94.7-97.4% 의특이도로가장적정한지표라고하였다. 6-8 스펙트럼영역빛간섭단층촬영은 6 6 mm 2 범위의망막신경섬유층두께지도와함께, 각지점에서의망막신경섬유층두께를빛간섭단층촬영기에내장된정상기준자료와비교하여색상화한유의지도를제공한다. 최근에 3D OCT 의유의지도와유사한 Cirrus 스펙트럼영역빛간섭단층촬영의 deviation map을이용하여정량화한망막신경섬유층결손은다른진단지표에비해높은진단력을보인다는보고들이있다. 2-5 Kim et al 2 은평균 MD 가 -4.66 db인국소적망막신경섬유층결손이있는녹내장에서 Cirrus 빛간섭단층촬영의망막신경섬유층두께 deviation map에서수기로구한결손폭의 AUC가.91, 민감도 92.4%, 특이도 83.3% 로유두주위망막신경섬유층두께보다국소적망막신경섬유층결손검출력이높다고하였다. Leung et al 3 은평균 MD가 -8.99 db인녹내장에서 Cirrus 빛간섭단층촬영의 deviation map에서망막신경섬유층결손의크기, 형태, 두께, 시신경유두에서의거리를 -5로점수화하여점수 3-5의민감도는 92.1-98.3% 이고, 이방법은망막신경섬유층결손의위치와형태를제공하며기존의유두주위망막신경섬유층두께보다민감도가높다고하였다. Leung et al 4 은평균 MD가 -11.38 db인녹내장을대상으로한다른연구에서 Cirrus 빛간섭단층촬영의 deviation map에서 Matlab 컴퓨터프로그램을이용하여정상범위의망막신경섬유층면적의비율인 RNFL area index(%) 를자동으로구하였고, 특이도 9.% 에서민감도는 95.7% 라고보고하였다. Kang et al 5 은평균 MD가 -5.12 db인녹내장을대상으로한연구에서 Cirrus 빛간섭단층촬영기의 deviation map 에서수기로 Image J를이용하여유의지도의비정상부위를측정하였고, 비정상부위에서적색 (p<1%) 과, 황색 (p<5%) 에임의로가중치를부여하여구한 deviation score의 AUC가.958로시신경유두주위망막신경섬유층두께의 AUC.892 보다유의하게높다고하였다. 본연구에서녹내장군의평균 MD가 -6.15 db였으며유의지도결손면적의국소적망막신경섬유층결손의검출력은이미보고된결과와비교할만하였다. 수기로구한두께지도결손면적의 AUC는 1., 민감도 1%, 특이도 1% 로가장높았고, 수기로구한유의지도의적색과황색면적의 AUC는.987, 민감도 98%, 특이도 98%, 자동으로구한 유의지도의적색과황색면적의 AUC는.966, 민감도 9%, 특이도 94% 였으며, 이들간에 AUC는유의한차이가없어서진단력은비슷하였다. 이전의많은연구들에서시신경유두주위망막신경섬유층두께는시신경유두측정치, 망막내층두께등과의구조적연관성, 시기능과의연관성이입증된녹내장진단지표이나 3.4 mm 원형스캔선을벗어난망막신경섬유층결손은검출이어렵다는한계가있다. 11-17 또한시간영역빛간섭단층촬영기는검사시마다원형스캔을수동으로시신경유두중심에정확히위치시키기어렵다는단점이있어경과관찰시변화를분석하는데어려움이있다. 스펙트럼영역빛간섭단층촬영기를이용한망막신경섬유층두께지도는이러한한계극복이가능하고, 지형도로 6 6 mm 2 시신경유두주위전체의두께와결손위치정보를제공하며, 자동으로시신경유두중심의위치가설정되는이점을보인다. 본연구에서자동으로구한유의지도적색과황색면적의 AUC (.966) 는유두주위하사분면망막신경섬유층두께의 AUC (.936) 보다컸으나유의성은없었고, 수동으로구한유의지도적색과황색면적의 AUC (.987) 는유두주위하사분면망막신경섬유층두께의 AUC (.936) 보다컸으며유의한차이가있었다. 자동과수동으로구한측정치의 AUC가차이가나는이유는불분명하나, 유의지도망막신경섬유층결손부위의측정에서수기로구할때에는망막신경섬유층사진에서보이는결손이외의황색부분을무시하였고, 자동으로구할때에는결손부위이외의황색부분이포함되었기때문으로생각한다. 본연구의중요한결과는자동으로구한유의지도적색과황색면적의진단력이하반부신경절세포 + 내망상층두께, 수직유두함몰비보다우월하다는것이다. 또한자동과수동으로구한유의지도결손면적의진단력은차이가없었고, 시신경유두를중심으로한직경 3.4 mm원으로스캔하여구한유두주위망막신경섬유층두께는결손의깊이는알수있지만결손의형태를알수없는반면에, 유의지도는망막신경섬유층결손형태와위치가망막신경섬유층사진과비슷하므로자동으로망막신경섬유층결손을정량화하는본연구의방법은녹내장의진단및경과관찰에있어유용할것으로생각한다. 또한수기로망막신경섬유층결손을정량화하는경우결손부위의경계부를그려야하고, 면적을계산하는데시간이많이소요되며, 특히경계가모호한미만성망막신경섬유층결손은측정이어렵다는단점이있는데자동으로구하는저자들의방법은이러한단점을극복할수있다. 녹내장에서망막외층은거의영향을미치지않으나, 신경섬유층, 신경절세포층, 내망상층은영향을미치며, 이세층을신경절세포복합체 (ganglion cell complex) 라한다. 14,18 시간영역빛간섭단층의전체황반부망막두께에비하여스펙 91
- 서샘외 : 망막신경섬유층결손면적의자동측정 - 트럼영역빛간섭단층촬영기로측정한황반부내망막두께인황반부신경절세포복합체의측정치는녹내장진단력을높였고, 스펙트럼영역빛간섭단층촬영기로측정한유두주위망막신경섬유층두께는황반부신경절세포복합체와녹내장진단력이비슷하다고하였다. 13,19-22 본연구에서도유두주위하사분면망막신경섬유층두께와하반부신경절세포 + 내망상층두께의녹내장진단력은비슷하였다. Cirrus 스펙트럼영역빛간섭단층촬영연구에서시신경유두측정치와유두주위망막신경섬유층두께측정치의진단력은차이가없다고하였다. 본연구에서도유두주위하사분면망막신경섬유층두께와수직유두함몰비의녹내장진단력은비슷하였다. 11,22 결론적으로스펙트럼영역빛간섭단층촬영기의망막신경섬유층유의지도는망막신경섬유층사진과비슷하고기기에내장된정상치자료보다 5% 미만으로비정상인국소적망막신경섬유층결손면적을자동으로구할수있으며, 이는황반부내망막측정치와시신경유두측정치에비하여국소적망막신경섬유층결손을검출하는데더높은진단력을보여녹내장진단에유용하였다. 본연구에서는국소적망막신경섬유층결손이있는경우만을대상으로유의지도를분석하였으나미만성망막신경섬유층결손이있는경우도유의지도결손면적의자동측정이가능하므로, 향후미만성망막신경섬유층결손이있는녹내장진단과경과관찰에유용한지에대한연구가필요하다고본다. REFERENCES 1) Schuman JS. Spectral domain optical coherence tomography for glaucoma (an AOS thesis). Trans Am Ophthalmol Soc 28;16: 426-58. 2) Kim NR, Lee ES, Seong GJ, et al. Spectral-domain optical coherence tomography for detection of localized retinal nerve fiber layer defects in patients with open-angle glaucoma. Arch Ophthalmol 21;128:1121-8. 3) Leung CK, Lam S, Weinreb RN, et al. Retinal nerve fiber layer imaging with spectral-domain optical coherence tomography: analysis of the retinal nerve fiber layer map for glaucoma detection. Ophthalmology 21;117:1684-91. 4) Leung CK, Choi N, Weinreb RN, et al. Retinal nerve fiber layer imaging with spectral-domain optical coherence tomography: pattern of RNFL defects in glaucoma. Ophthalmology 21;117: 2337-44. 5) Kang SY, Sung KR, Na JH, et al. Comparison between deviation map algorithm and peripapillary retinal nerve fiber layer measurements using Cirrus HD-OCT in the detection of localized glaucomatous visual field defects. J Glaucoma 212;21:372-8. 6) Jeoung JW, Park KH, Kim TW, et al. Diagnostic ability of optical coherence tomography with a normative database to detect localized retinal nerve fiber layer defects. Ophthalmology 25; 112:2157-63. 7) Yoo YC, Park KH. Comparison of optical coherence tomography and scanning laser polarimetry for detection of localized retinal nerve fiber layer defects. J Glaucoma 21;19:229-36. 8) Yoo YC, Park KH. Influence of angular width and peripapillary position of localized retinal nerve fiber layer defects on their detection by time-domain optical coherence tomography. Jpn J Ophthalmol 211;55:115-22. 9) Kang SM, Uhm KB. False negative findings of optical coherence tomography in eyes with localized nerve fiber layer defects. J Korean Ophthalmol Soc 211;52:454-61. 1) Ye C, To E, Weinreb RN, et al. Comparison of retinal nerve fiber layer imaging by spectral domain optical coherence tomography and scanning laser ophthalmoscopy. Ophthalmology 211;118: 2196-22. 11) Mwanza JC, Oakley JD, Budenz DL, Anderson DR; Cirrus Optical Coherence Tomography Normative Database Study Group. Ability of cirrus HD-OCT optic nerve head parameters to discriminate normal from glaucomatous eyes. Ophthalmology 211;118:241-8. e1. 12) Schulze A, Lamparter J, Pfeiffer N, et al. Diagnostic ability of retinal ganglion cell complex, retinal nerve fiber layer, and optic nerve head measurements by Fourier-domain optical coherence tomography. Graefes Arch Clin Exp Ophthalmol 211;249:139-45. 13) Seong M, Sung KR, Choi EH, et al. Macular and peripapillary retinal nerve fiber layer measurements by spectral domain optical coherence tomography in normal-tension glaucoma. Invest Ophthalmol Vis Sci 21;51:1446-52. 14) Tan O, Li G, Lu AT, et al. Mapping of macular substructures with optical coherence tomography for glaucoma diagnosis. Ophthalmology 28;115:949-56. 15) Taliantzis S, Papaconstantinou D, Koutsandrea C, et al. Comparative studies of RNFL thickness measured by OCT with global index of visual fields in patients with ocular hypertension and early open angle glaucoma. Clin Ophthalmol 29;3:373-9. 16) Cvenkel B, Kontestabile AS. Correlation between nerve fibre layer thickness measured with spectral domain OCT and visual field in patients with different stages of glaucoma. Graefes Arch Clin Exp Ophthalmol 211;249:575-84. 17) Lopez-Peña MJ, Ferreras A, Larrosa JM, et al. Relationship between standard automated perimetry and retinal nerve fiber layer parameters obtained with optical coherence tomography. J Glaucoma 211;2:422-32. 18) Harwerth RS, Charles F. Prentice Award Lecture 26: a neuron doctrine for glaucoma. Optom Vis Sci 28;85:436-44. 19) Tan O, Chopra V, Lu AT, et al. Detection of macular ganglion cell loss in glaucoma by Fourier-domain optical coherence tomography. Ophthalmology 29;116:235-14. e1-2. 2) Mori S, Hangai M, Sakamoto A, Yoshimura N. Spectral-domain optical coherence tomography measurement of macular volume for diagnosing glaucoma. J Glaucoma 21;19:528-34. 21) Nakatani Y, Higashide T, Ohkubo S, et al. Evaluation of macular thickness and peripapillary retinal nerve fiber layer thickness for detection of early glaucoma using spectral domain optical coherence tomography. J Glaucoma 211;2:252-9. 22) Mwanza JC, Durbin MK, Budenz DL, et al. Glaucoma diagnostic accuracy of ganglion cell-inner plexiform layer thickness: comparison with nerve fiber layer and optic nerve head. Ophthalmology 212;119:1151-8. 911
- 대한안과학회지 213 년제 54 권제 6 호 - =ABSTRACT= Usefulness of Automated Measurements of Localized Retinal Nerve Fiber Layer Defects Area Using Significance Map Sam Seo, MD, Joong Won Shin, MD, Ki Bang Uhm, MD, PhD Department of Ophthalmology, Hanyang University College of Medicine, Seoul, Korea Purpose: To evaluate the usefulness of automated measurements of the localized retinal nerve fiber layer (RNFL) defects area in patients with glaucoma. Methods: Fifty one patients with localized RNFL defects in RNFL red-free photographs and 53 healthy subjects were included. All participants were imaged with 3D spectral-domain optical coherence tomography (OCT). The area of defects was measured with the RNFL significance map (red = p < 1% and yellow = p < 5%) using Image J manually and Matlab software automatically. The area under the receiver operating characteristic curve (AUC) was calculated for the RNFL defect area of the RNFL photograph and RNFL maps, circumpapillary RNFL thickness, optic disc parameter, and macular inner retina thickness. Results: High correlation was observed between manually and automatically measured defect areas in the significance map (red area r =.94, red and yellow area r =.89). The AUC for manually and automatically measured defects area (.987,.966; p < 5%, p =.31, respectively) in the significance map was comparable. The latter demonstrated slightly higher but insignificant difference in AUC for inferior quadrant circumpapillary RNFL thickness (.936; p =.22) and was significantly higher than the inferior ganglion cell layer plus inner plexiform layer thickness (.894) and vertical cup to disc ratio (.869) (p =.18, p =.8, respectively). Conclusions: The automated measurements of the RNFL defect area in the significance map performed adequately in detecting localized RNFL defects and had a better performance than macular inner retina and optic nerve parameters. J Korean Ophthalmol Soc 213;54(6):92-912 Key Words: Glaucoma, Optic nerve, Retinal nerve fiber layer, Significance map, Spectral-domain optical coherence tomography Address reprint requests to Ki Bang Uhm, MD, PhD Department of Ophthalmology, Hanyang University Seoul Hospital #222 Wangsimni-ro, Seongdong-gu, Seoul 133-791, Korea Tel: 82-2-229-857, Fax: 82-2-2291-8517, E-mail: KBUhm@hanyang.ac.kr 912