Pomen kazalnikov za glavkom v presejalnem programu za diabetično retinopatijo

Podnar, Barbara (author), Albreht, Tit (mentor), Cvenkel, Barbara (comentor)

Ozadje in namen dela: Zgodnje odkrivanje glavkoma lahko izboljšamo z oportunističnim presejanjem v presejalnem programu za diabetično retinopatijo (PPDR). S študijo smo želeli ugotoviti, kakšna je napovedna vrednost kazalnikov s fotografij očesnega ozadja kot testa za glavkom v populaciji bolnikov, ki se spremljajo v PPDR, kakšna je relativna pomembnost kazalnikov pri ugotavljanju glavkoma v PPDR in kakšna njihova diagnostična vrednost. Metodologija: Pregledali smo fotografije očesnega ozadja bolnikov, ki so bili obravnavani v PPDR na Očesni kliniki Univerzitetnega kliničnega centra Ljubljana od novembra 2019 do januarja 2020 in od maja 2020 do avgusta 2020. Bolnike z vsaj enim kazalnikom za glavkom sumljivega videza papile vidnega živca in nekaj naključno izbranih bolnikov brez kazalnikov smo povabili na očesni pregled. Za bolnike z glavkomom in sumom na glavkom smo smatrali, da so bili na pregled napoteni upravičeno. S pomočjo logistične regresije in posplošene ocenjevalne enačbe z logistično regresijo smo določili, kolikšna je napovedna vrednost kazalnikov za ugotavljanje, kdo potrebuje pregled zaradi suma na glavkom na podlagi kazalnikov s fotografij očesnega ozadja. Z ROC analizo smo določili diagnostično vrednost kazalnikov in napovednega modela. Rezultati: Pregledali smo fotografije 2230 bolnikov iz PPDR. 209 bolnikov (10,1 %) je imelo na fotografijah vsaj 1 kazalnik. Skupno se je na vabilo na pregled odzvalo 149 bolnikov (129 z vsaj 1 kazalnikom, 20 brez kazalnikov), izmed katerih pri 79 (53,0 %) bolnikih nismo ugotavljali glavkoma, 54 bolnikom (36,2 %) smo postavili diagnozo sum na glavkom in 16 bolnikom (10,7 %) diagnozo glavkom. Sedem bolnikov z glavkomom za diagnozo predhodno ni vedelo. Vsi bolniki s kazalnikom zažetje nevroretinskega obroča so imeli glavkom. Kazalnik C/D razmerje je bil najpomembnejši za ugotavljanje glavkoma/suma na glavkom (razmerje obetov 7,59 (95 % interval zaupanja 3,98–14,47; p < 0,001)) in je ostal statistično značilno pomemben v multivariatnem modelu. Vpliv na ugotavljanje glavkoma/suma na glavkom je imela tudi pozitivna družinska anamneza za glavkom, vendar je ostala statistično značilno pomembna le v multivariatnem modelu. Ploščina pod ROC krivuljo za končni LOGIT model (oz. za verjetnosti za izid, ki jih model napove), v katerem sta bila vsebovana družinska anamneza za glavkom in C/D razmerje, je bila 0,967 (95 % interval zaupanja 0,944-0,990; p<0,001). Zaključek: PPDR predstavlja priložnost za odkrivanje novih primerov bolezni, saj je glavkom v približno polovici primerov nediagnosticiran. Študija je pokazala, da sta za napotovanje iz PPDR na očesni pregled zaradi suma na glavkom/glavkoma najpomembnejša kazalnik C/D razmerje in zažetje NRO. Glede na pomembnost pozitivne družinske anamneze za glavkom bi bilo smiselno bolnike iz PPDR povprašati o glavkomu v družini. Potrebna bi bila validacija modela na novem vzorcu PPDR.

• Language: Slovenian
• Keywords: glavkom, presejanje, presejalni test, presejalni program za diabetično retinopatijo, fotografija očesnega ozadja, regresijska analiza
• Work type: Doctoral dissertation
• Organization: MF - Faculty of Medicine
• Year: 2024
• PID: 20.500.12556/RUL-158532
• Publication date in RUL: 15.06.2024

Secondary language

• Language: English
• Title: ​Importance of indicators for glaucoma in diabetic retinopathy screening programme

Abstract

Background and objectives: Retinal images from diabetic retinopathy screening programmes (DRSP) provide an opportunity to improve early glaucoma detection. The aim of our study was to find out the predictive value of the retinal image indicators as a test for glaucoma in a population of patients from a DRSP, the relative importance of retinal image indicators for glaucoma detection in a DRSP and their diagnostic value. Methods: We reviewed retinal images of patients, who attended the DRSP at the University Medical Centre Ljubljana from November 2019 to January 2020 and from May to August 2020. We invited patients with at least one indicator and some randomly selected patients without indicators for an ophthalmic examination. Patients with glaucoma and suspect glaucoma were considered accurately referred. Logistic regression (LOGIT) and generalised estimating equation with logistic regression (GEE) were used to find out the referral accuracy of retinal image indicators. ROC analysis was used to determine the diagnostic value of the indicators and the predictive model. Results: Of the 2230 patients reviewed, 209 patients (10.1 %) had at least one indicator on their retinal images. 129 patients with at least one indicator and 20 patients without indicators (n = 149) attended the eye exam. Seventy-nine patients (53.0 %) were diagnosed glaucoma negative, 54 (36.2 %) suspect glaucoma, and 16 (10.7 %) glaucoma positive. Seven glaucoma patients were newly detected. All patients with a neuroretinal rim notch had glaucoma. The cup-to-disc ratio was the most important indicator for accurate referral (odds ratio 7.59 (95 % confidence interval 3.98–14.47; p < 0.001) and stayed statistically significant in a multivariable analysis. Positive family history of glaucoma also affected referral accuracy but was showed to be statistically significant only in the LOGIT multivariable model. Other indicators and confounders lost their statistical significance in multivariable analyses. The area under the ROC curve for a final LOGIT model (i.e. for its predicted outcome probabilities), consisting of the family history of glaucoma and the cup-to-disc ratio, was 0,967 (95 % confidence interval 0,944-0,990; p<0,001). Conclusions: With approximately half of the patients undiagnosed, DRSP has the potential for glaucoma detection. Our study showed that the cup-to-disc ratio and the neuroretinal rim notch are the most important indicators for accurate glaucoma referral from retinal images in DRSP. It would be reasonal to ask patients from the DRSP about their glaucoma family history since this it was showed to be important in predicting suspect glaucoma referrals. Validation on a new DRSP sample is needed.

• Keywords: glaucoma, screening, screening test, diabetic retinopathy screening programme, retinal image, regression analysis