The Image of Image Analyzers

 

Jonathan Myers, MD

Wills Eye Hospital

Philadelphia, Pennsylvania

 

Summary

Optic nerve imaging technologies continue to improve, both in their technological abilities to measure, and in their software to evaluate the optic nerve.  The current generation of instruments remains limited by the variability in the optic nerve appearance and in the overlap between glaucoma and normal structure.  Researchers are improving assessment algorithms, but many of these are not yet available to most clinicians.  In their current, commercially available forms, each of these instruments can yield information to supplement the examination and aid in the diagnosis and monitoring of glaucoma.  None of these instruments is sufficient in and of itself to replace the traditional methods of examination.  Physician judgment remains crucial in evaluating the output of these devices and in incorporating that into the diagnosis and management of individual patients.

 

Among the contributions to patient care that the optic nerve imagers have made is to increase interest in the clinical evaluation of the optic nerve.  In their current form, these instruments may not surpass the ultimate clinician’s ability to diagnose and monitor glaucoma, but they certainly can improve the ability of a hurried clinician who may not be able to provide the ultimate clinical examination for every glaucoma suspect.

 

The recent enthusiasm for disc imaging demonstrates that clinician interest may be influenced not just by science, but also by marketing and reimbursements.  Physicians need to remain focused on the clinical utility of these machines to ensure that more than market forces determine what tools are available in the future to care for our patients.  This is especially true as the business community moves quickly to decide fates, but it is the long term performance of these instruments in monitoring this chronic disease that will matter most to our patients.

 

Optic Nerve Imaging Issues

Optic Nerve Evaluation Goals:

 

Optic Nerve Assessment Technologies Potential

 

Challenges for the clinician and the machine

 

Old Technologies

                        • Both affordable and effective, but efficacy varies with clinician skills

                        • Not truly objective

 

Newer Technologies

 

Heidelberg Retina Tomograph (HRT II)

 

HRT to screen for glaucoma

neither sufficiently sensitive nor specific for population screening

not highly correlated with clinical diagnosis

                     Results may be affected by size of optic disc

 

HRT for diagnosis:     

     –Sensitivity: 80%, Specificity 65%

    –Sensitivity: 31-53%, Specificity: 90-92%

                -Sensitivity 74%, Specificity 94% (Miglior et al, AJO 2003)

                -Sensitivity 78%, Specificity 81% (Ford et al, Ophtho 2003)

 

 

HRT to monitor for progression:

            limited by patient and technician- 30-100m

            reduced in sloping vs flat areas- may limit early detection

 

HRT to monitor progression:

Jatuthong, Hoffman, Garway-Heath, Coleman, Zeyen, Caprioli, ARVO 2000

17/67 eyes with progression

Neural Network:

Sensitivity 71%            Specificity 89%

Rim Volume and Rim Height greatest concordance

 

Chauhan et al, Arch Ophthalmol 2001

77 patients prospectively followed avg. 5.5 years

27% no progression

40% HRT progression only

29% HRT and VF progression

4% VF progression only

16 patients photos: 81% agreement with HRT

Conclusion: HRT sensitive for progressive VF loss, but how much of other 40% HRT positive only are “false positives”

Of the 29% with progression by both VF and HRT: nearly equally split HRT detection first vs. VF detection first

            If HRT truly more sensitive, would expect mostly HRT progression first

 

Ervin JC, Lemij HG, Mills RP, Quigley HA, Thompson HW, Burgoyne CF

Ophthalmology 2002;109(3):467-481

Clinician change detection viewing longitudinal stereophotographs compared to confocal scanning laser tomography in the LSU Experimental Glaucoma (LEG) Study.

“This study provides the first direct evidence that an existing CSLT may reasonably meet or exceed the ONH surface change detection performance of fellowship-trained glaucoma specialists in at least those eyes with good CSLT reproducibility.” (Note: This study was of monkey eyes with experimental glaucoma- it’s unclear how much glaucoma fellowships emphasize monkey glaucoma)

 

HRT in clinical practice

 

 

Nerve Fiber Analyzer GDx

 

Choplin and Lundy, Ophthalmology 2001

 

GDx in clinical practice

 

NFA Sensitivity and Specificity

• Weinreb         82%     62%     The Number

                               74%     92%     Best Algorithm

            83 normals, 84glaucoma subjects

• Kwon  78%     60%     The Number

            18 normals, 20 glaucoma subjects

• Sinai   94%     91%     Best Algorithm

            34 normals, 34 suspects, 34 glaucoma subjects     

• Costa 84%     79%     The Number <32

            93%     91%     Best Algorithm

            94 normals, 91 glaucoma subjects

 

 

Optical Coherence Tomography (OCT II)

Assesses RNFL thickness by analyzing the temporal delay of backscattered light from tissue structures via low-coherence interferometry.

In laboratory setting, very accurate and reproducible- best of these four instruments in these regards (in the laboratory)

In clinical setting, there exists a definite “learning curve” to optimally utilize the instrument- this is much improved in the OCT III vs. the OCT II

 

Appealing to measure neuronal thickness/volume to assess for current or progressive glaucoma damage

Simple graphical print-outs easier to compare than disc photos

Challenged by same issues of overlap between normal, suspect and glaucoma patients’ parameters according to many studies and clinical experience

RNFL Measurements correlated to HRT measurements (Mistlberger)

 

Comparing HRT, GDx, and OCT

HRT, GDx, OCT: Best Parameters

Zangwill et al, Arch Ophthalmol 2001

50 normals, 41 early to moderate glaucoma

Similar area under ROC among instruments

At specificity>90%, GDx less sensitive

Stereophotographs, HRT, OCT equal

 

HRT, GDx, OCT: Summary Data

Zangwill et al, Arch Ophthalmol 2001

50 normals, 39 early to moderate glaucoma

Two MDs, One PhD evaluate summary data

Sensitivity        Specificity

HRT:    64-75%            68-80%

GDx:    72-82%            56-82%

OCT:   76-79%            68-81%

 

DP (disc photo), HRT, GDx, OCT vs. Diagnosis

• 39 normal and 50 age matched glaucoma

• MD     normal 0.0+0.9 dB      glaucoma -3.9+2.3 dB

                        Sens.               Spec.               AROC

• DP                  0.94                 0.87                 0.93

• CSLO             0.84                 0.90                 0.92

• SLP                0.89                 0.87                 0.94

• OCT               0.82                 0.84                 0.88

Greaney MJ, Hoffman DC, Garway-Heath DF, Nakla M, Coleman AL, Caprioli J

IOVS January 2002;43(1):140-145

 

 

Discam

            Digitized near simultaneous stereo disc photography

            Excellent images may be achieved through dilated pupils with clear media

            Computer display facilitates stereo viewing and comparison

            Computer analysis of disc parameters possible

            Easy storage, retrieval, and comparison of images

            Future software will allow more sophisticated analysis of disc and rim,

e.g. sectoral evaluation of rim thickness

Current Use

            Can replace standard disc photography for evaluation, documentation,

serial comparisons

            Lowest purchase price of these four instruments

            Stereo disc photography without a photographer

 

Discam limitations

            Very few published studies in peer reviewed journals (Three on Medline, 8/03)

            Clinician or technician draws disc and cup margins

            Does not correct measurements for refractive error

            Simplistic data evaluation tools (for the time-being)

 

In a comparison with scanning laser ophthalmoscopy and disc photos, Discam had similar performance overall, although the techniques showed variation in the results on individual patients. Correnti AJ, Wollstein G, Price LL, Schuman JS. Ophthalmology. 2003

 

 

 

Selected References

 

Heidelberg Retinal Tomograph

 “Optical coherence tomography and scanning laser polarimetry were capable of differentiating glaucomatous from nonglaucomatous populations in this cohort; however considerable measurement overlap was observed among normal, ocular hypertensive, and glaucomatous eyes.”

Hoh ST; Greenfield DS; Mistlberger A; Liebmann JM, Ishikawa H; Ritch R

Optical coherence tomography and scanning laser polarimetry in normal, ocular hypertensive, and glaucomatous eyes.

Am J Ophthalmol 2000 Feb;129(2):129-35

 

“In a broad clinical setting including normal, OHT and glaucoma patients, the HRT and visual field tests have fair to poor agreement in detecting glaucoma...These values did not change when normal or OHT patients were excluded from the analysis.  In the clinical setting, caution should be used when interpreting HRT results on the basis of multivariate discriminant analysis.”

Miglior S, Casual M, Aguareschi M, Marchetti I, Iester M, Orzalesi N.  Clinical ability of Heidelberg retinal Tomograph examination to detect glaucomatous visual field changes.  Ophthalmology 2001;108:1621-1627

 

“Most patients with field changes also had disc changes; however, less than half of those with disc changes had field changes.”

Chauhan BC, McCormick TA, Nicolela MT, LeBlanc RP.  Optic disc and visual field changes in a prospective longitudinal study of patients with glaucoma.  Comparison of scanning laser tomography with conventional perimetry and optic disc photography.  Arch Ophthalmol 2001;119:1492-1499

 

“This study provides the first direct evidence that an existing CSLT may reasonably meet or exceed the ONH surface change detection performance of fellowship-trained glaucoma specialists in at least those eyes with good CSLT reproducibility.” Note: Monkey Eyes

Ervin JC, Lemij HG, Mills RP, Quigley HA, Thompson HW, Burgoyne CF

Clinician change detection viewing longitudinal stereophotographs compared to confocal scanning laser tomography in the LSU Experimental Glaucoma (LEG) Study.

Ophthalmology 2002;109(3):467-481

Miglior S, Guareschi M, Albe' E, Gomarasca S, Vavassori M, Orzalesi N.  Detection of glaucomatous visual field changes using the Moorfields regression analysis of the Heidelberg retina tomograph.  Am J Ophthalmol. 2003 Jul;136(1):26-33.

Ford BA, Artes PH, McCormick TA, Nicolela MT, LeBlanc RP, Chauhan BC.  Comparison of data analysis tools for detection of glaucoma with the Heidelberg Retina Tomograph.  Ophthalmology. 2003 Jun;110(6):1145-50

Miglior S, Guareschi M, Albe' E, Gomarasca S, Vavassori M, Orzalesi N.  Detection of glaucomatous visual field changes using the Moorfields regression analysis of the Heidelberg retina tomograph.  Am J Ophthalmol. 2003 Jul;136(1):26-33.

“Rim area is reproducible and potentially useful as a marker of progression. Pattern of variability and the influence of different reference planes, disease, operators, and visits should be considered when evaluating progression.”
Tan JC, Garway-Heath DF, Fitzke FW, Hitchings RA   Reasons for rim area variability in scanning laser tomography.  Invest Ophthalmol Vis Sci. 2003 Mar;44(3):1126-31

Bowd C, Chan K, Zangwill LM, Goldbaum MH, Lee TW, Sejnowski TJ, Weinreb RN.   Comparing neural networks and linear discriminant functions for glaucoma detection using confocal scanning laser ophthalmoscopy of the optic disc.  Invest Ophthalmol Vis Sci. 2002 Nov;43(11):3444-54.

“This study demonstrated quantitatively that the progression of optic disk damage in glaucoma is faster in eyes with disk hemorrhage than in eyes without disk hemorrhage.”
Ahn JK, Park KH.  Morphometric change analysis of the optic nerve head in unilateral disk hemorrhage cases.  Am J Ophthalmol. 2002 Dec;134(6):920-2.

 

Wollstein G; Garway-Heath DF; Hitchings RA

Identification of early glaucoma cases with the scanning laser ophthalmoscope.

Ophthalmology 1998 Aug;105(8):1557-1563

 

Zangwill LM, Van Horn S, De Souza M, Sample PA, Weinreb RN.

Optic nerve head Topography in ocular hypertensive eyes using confocal scanning laser ophthalmoscopy.

Am J Ophthalmology 1996; 122: 520-5.

 

Ister M, Mikelberg FS, Swindale NV, Drance SM. ROC analysis of  Heidelberg Retina Tomograph optic disc shape measures in glaucoma. Can J Ophthalmol 1997; 32:382-388.

 

Ister M, Mikelberg FS, Drance SM.

The effect of optic disc size on diagnostic precision with the Heidelberg Retina Tomograph.

Ophthalmology 1997; 104:545-548.

 

Brigatti L; Weitzman M, Caprioli J.

Regional test-retest variability of confocal scanning laser tomography. 

Am J Opthalmol 1995 Oct;120(4):433-440

 

IOVS Vol. 41:4:Supplement pg 90

Abstracts

471 Detection of Progressive Glaucomatous Damage with Conofocal Scanning Laser Ophthalmoscopy, Jatuthong et al

 

473 Six Year Prospective Cohort Study of Early Glaucoma Using the Heidelberg REtinal Tomograph. Hatch et al

 

GDx Nerve Fiber Analyzer
”Both observers discriminated better than the Number. At a critical value of 23, the specificity of the Number was 81.5%, which matched the lowest specificity of the 2 observers: 82.5% and 92.0% for observers 1 and 2, respectively. At these specificities, the sensitivity of the 2 observers and of the Number were 92.0%, 89.5%, and 85.5%, respectively. The sensitivity increased with the severity of glaucoma. The Kappa values for intraobserver agreement were 0.80 and 1.0. CONCLUSIONS: The Number yielded acceptable sensitivity and specificity values at a critical value of 23 in this test population. However, the clinical judgments of the printouts by both expert observers resulted in a better separation between normal and glaucomatous eyes, particularly in the group with mild glaucoma.”

Colen TP, Lemij HG.  Sensitivity and specificity of the GDx: clinical judgment of standard printouts versus the number.   J Glaucoma. 2003 Apr;12(2):129-33

Mojon DS.  Low specificity of scanning laser polarimetry.  Ophthalmologica. 2003 Jan-Feb;217(1):17-9.

” Correction for CPA significantly increases the correlation between retinal nerve fiber layer structural damage and visual function and significantly improves the discriminating power of SLP for detection of mild-to-moderate glaucoma.”
Greenfield DS, Knighton RW, Feuer WJ, Schiffman JC, Zangwill L, Weinreb RN.  Correction for corneal polarization axis improves the discriminating power of scanning laser polarimetry.  Am J Ophthalmol. 2002 Jul;134(1):27-33.
 

“Localized RNFL defects were visible in GDx retardation maps obtained with IC. The defects closely matched those observed in red-free fundus photographs. With FC, however, the GDx retardation images did not correlate well with red-free fundus photography. …An individualized anterior segment compensation in the GDx improves the visualization of localized glaucomatous loss.”
Reus NJ, Colen TP, Lemij HG.  Visualization of localized retinal nerve fiber layer defects with the GDx with individualized and with fixed compensation of anterior segment birefringence.  Ophthalmology. 2003 Aug;110(8):1512-6.

“In general, the correlation of the two [visual field and GDx] is better when there is significant visual field defect than when the visual field is close to normal.”

Kwon YH, Hong S, Honkanen RA, Alward WL.  Correlation of automated visual field parameters and peripapillary nerve fiber layer thickness as measured by scanning laser polarimetry.

J Glaucoma 2000;Aug;9(4):281-8

 

“Scanning laser polarimetry may have good sensitivity and specificity for separating normal from abnormal eyes, but it is not as good for classifying unknown subjects when glaucoma suspects are included.”

Choplin NT, Lundy DC.  The sensitivity and specificity of scanning laser polarimetry in the detection of glaucoma in a clinical setting.  Ophthalmology 2001;108:899-904

 

“Compared with fixed compensation, mean-based SLP parameters generated with SLP-VCC have greater correlation with visual function and RNFL thickness assessments obtained with OCT.”
Bagga H, Greenfield DS, Feuer W, Knighton RW.  Scanning laser polarimetry with variable corneal compensation and optical coherence tomography in normal and glaucomatous eyes.  Am J Ophthalmol. 2003 Apr;135(4):521-9.

 

“In eyes with visual field progression after optic disc hemorrhage, a significant change in the SLP image was not seen. Fluctuation of SLP results in patients with glaucoma necessitates confirmation of progression seen on SLP images.”
Boehm MD, Nedrud C, Greenfield DS, Chen PP.  Scanning laser polarimetry and detection of progression after optic disc hemorrhage in patients with glaucoma.  Arch Ophthalmol. 2003 Feb;121(2):189-94.

 

Invest Ophthalmol Vis Sci. 2002 May;43(5):1465-74.   Correction for the erroneous compensation of anterior segment birefringence with the scanning laser polarimeter for glaucoma diagnosis. Garway-Heath DF, Greaney MJ, Caprioli J.

 

Yamada N; Chen PP; Mills RP; Leen MM; Stamper RL; Lieverman MF; Xu L; Stanford DC.  Glaucoma screening using the scanning laser polarimeter.

J Glaucoma 2000 Jun;9(3):254-61

 

Weinreb RN; Zangwill L; Berry CC; Bathija R; Sample PA

Detection of glaucoma with scanning laser polarimetry.

Arch Ophthalmol 1998 Dec;116(12):1583-9

 

Choplin NT; Lundy DC; Dreher AW

Differentiating patients with glaucoma from glaucoma suspects and normal subjects by nerve fiber layer assessment with scanning laser polarimetry.

Ophthalmology 1998 Nov;105(11):2068-2076

 

Ocular Coherence Tomagraphy (OCT)
Carpineto P, Ciancaglini M, Zuppardi E, Falconio G, Doronzo E, Mastropasqua L.  Reliability of nerve fiber layer thickness measurements using optical coherence tomography in normal and glaucomatous eyes.  Ophthalmology. 2003 Jan;110(1):190-5.

 

Greenfield DS, Bagga H, Knighton RW.  Macular thickness changes in glaucomatous optic neuropathy detected using optical coherence tomography.  Arch Ophthalmol. 2003 Jan;121(1):41-6.

Mok KH, Lee VW, So KF.  Retinal nerve fiber layer measurement by optical coherence tomography in glaucoma suspects with short-wavelength perimetry abnormalities.  J Glaucoma. 2003 Feb;12(1):45-9.

Williams ZY, Schuman JS, Gamell L, Nemi A, Hertzmark E, Fujimoto JG, Mattox C, Simpson J, Wollstein G.  Optical coherence tomography measurement of nerve fiber layer thickness and the likelihood of a visual field defect.   Am J Ophthalmol. 2002 Oct;134(4):538-46

Schuman JS, Wollstein G, Farra T, Hertzmark E, Aydin A, Fujimoto JG, Paunescu LA.  Comparison of optic nerve head measurements obtained by optical coherence tomography and confocal scanning laser ophthalmoscopy.  Am J Ophthalmol. 2003 Apr;135(4):504-12.

Bowd C; Weinreb RN; Williams JM; Zangwill LM.

The retinal nerve fiber layer thickness in ocular hypertensive, normal, and glaucomatous eyes with optical coherence tomography.

Arch Ophthalmol. 2000 Jan;118(1):22-26

 

Mistlberger A; Liebmann JM; Greenfield DS; Pons ME; Hoh ST; Ishikawa H; Ritch R

Heidelberg retina tomography and optical coherence tomography in normal, ocular-hypertensive, and glaucomatous eyes.

Ophthalmology 1999 Oct;106(10):2027-32

 

Cioffi GA, RObin AL, Eastman RD, Perell JF, Sarafarazi FA, Kelman SE.  Confocal Laser Scanning Ophthalmoscope-Reproducibility of Optic Nerve Head Topographic Measurements with the Confocal Laser Scanning Ophthalmoscope.  Ophthalmology 1993;100:57-62

 

Brigatti L; Weitzman M, Caprioli J.

Regional test-retest variability of confocal scanning laser tomography.

Am J Opthalmol 1995 Oct;120(4):433-440

 

Kamal DS, Viswanathan AC, Garway-Heath DF, Hitchings RA, Poinoosawmy D, Bunce C.  Detection of optic disc change with the Heidelberg retina tomograph before confirmed visual field change in ocular hypertensives converting to early glaucoma.  Br J Ophthalmol 1999;83:290-294

 

Discam

Shuttleworth GN; Khong CH; Diamond JP.

A new digital optic disc stereo camera: intraobserver and interobserver repeatability of optic disc measurements.

 

Br J Ophthalmol 2000 Apr;84(4):403-7
Sung VC, Bhan A, Vernon SA.  Agreement in assessing optic discs with a digital stereoscopic optic disc camera (Discam) and Heidelberg retina tomograph.  Br J Ophthalmol. 2002 Feb;86(2):196-202.

Comparison Articles

“The quantitative methods CSLO [computerized scanning laser ophthalmoscopy], SLP [scanning laser polarimetry], and OCT [ocular coherence tomography] were no better than qualitative assessment of disc ONHPs [optic nerve head photography] by experienced observers at distinguishing normal eyes from those with early to moderate glaucoma. A combination of the imaging methods significantly improves this capability.”

Greaney MJ, Hoffman DC, Garway-Heath DF, Nakla M, Coleman AL, Caprioli J. Comparison of optic nerve imaging methods to distinguish normal eyes from those with glaucoma. Invest Ophthalmol Vis Sci 2002 Jan;43(1):140-5

”Among the topographic parameters generated by the HRT, rim area had the best correlation with visual field indices. The "number," maximum modulation, and ellipse modulation generated by the GDx also had correlations with visual field indices. The correlations were better for the sectoral parameters than the global parameters. However, great interindividual variation was found in the association.”

Lan YW, Henson DB, Kwartz AJ.  The correlation between optic nerve head topographic measurements, peripapillary nerve fibre layer thickness, and visual field indices in glaucoma.  Br J Ophthalmol. 2003 Sep;87(9):1135-41 

 

“Nerve fiber loss as detected on GDx correlates well with topographic optic nerve head changes as measured with the HRT-II. However, automated diagnosis on the two machines showed poor agreement.”
Sihota R, Gulati V, Saxena R, Agarwal HC, Sharma AK.Correlation between confocal scanning laser ophthalmoscopy and scanning laser polarimetry in open angle glaucoma.   Eur J Ophthalmol. 2003 Apr;13(3):266-75

 

“Although the area under the ROC curves was similar among the best parameters from each instrument, qualitative assessment of stereophotographs and measurements from the OCT ad HRT generally had higher sensitivities than the measurements from the GDx.”

Zangwill LM, Bowd C, Berry CC, Williams J, Blumenthal EZ, Sanchez-Galeana CA, Vasile C, Weinreb R.  Discriminating between normal and glaucomatous eyes using the Heidelberg retina Tomograph, GDx nerve fiber analyzer, and optical coherence tomograph.  Arch Ophthalmol 2001;119:985-993

 

“When used alone, HRT, GDx, and OCT summary data reports can differentiate between normal and glaucomatous eyes with mild to moderate visual field loss.  However, none of the instruments provided sensitivity and specificity that justify summary data reports being used as a screening tool for early to moderate glaucoma.”

Sanchez-Galeana C, Bowd C, Zblumenthal EZ, Gokhale PA, Zangwill LM, Weinreb RN.  Using optical imaging summary data to detect glaucoma.  Ophthalmology 2001;108:1812-1818

 

“AROC curve values among all three techniques imply that the Discam, CSLO, and stereography perform equally for the determination of glaucoma status.”
Correnti AJ, Wollstein G, Price LL, Schuman JS. Comparison of optic nerve head assessment with a digital stereoscopic camera (discam), scanning laser ophthalmoscopy, and stereophotography.
Ophthalmology. 2003 Aug;110(8):1499-505