Classification criteria for herpes simplex virus anterior uveitis

Abstract

Purpose:

To determine classification criteria for herpes simplex virus (HSV) anterior uveitis

Design:

Machine learning of cases with HSV anterior uveitis and 8 other anterior uveitides.

Methods:

Cases of anterior uveitides were collected in an informatics-designed preliminary database, and a final database was constructed of cases achieving supermajority agreement on the diagnosis, using formal consensus techniques. Cases were split into a training set and a validation set. Machine learning using multinomial logistic regression was used on the training set to determine a parsimonious set of criteria that minimized the misclassification rate among the anterior uveitides. The resulting criteria were evaluated on the validation set.

Results:

One thousand eighty-three of cases of anterior uveitides, including 101 cases of HSV anterior uveitis, were evaluated by machine learning. The overall accuracy for anterior uveitides was 97.5% in the training set and 96.7% in the validation set (95% confidence interval 92.4, 98.6). Key criteria for HSV anterior uveitis included unilateral anterior uveitis with either 1) positive aqueous humor polymerase chain reaction assay for HSV; 2) sectoral iris atrophy in a patient ≤50 years of age; or 3) HSV keratitis. The misclassification rates for HSV anterior uveitis were 8.3% in the training set and 17% in the validation set, respectively.

Conclusions:

The criteria for HSV anterior uveitis had a reasonably low misclassification rate and appeared to perform well enough for use in clinical and translational research.

PRECIS

Using a formalized approach to developing classification criteria, including informatics-based case collection, consensus-technique-based case selection, and machine learning, classification criteria for herpes simplex virus (HSV) anterior uveitis were developed. Key criteria included unilateral anterior uveitis with either 1) positive aqueous humor polymerase chain reaction assay for HSV; 2) sectoral iris atrophy in a patient ≤50 years of age; or 3) HSV keratitis. The resulting criteria had an acceptable misclassification rate.

Herpes simplex virus (HSV) anterior uveitis is an infectious anterior uveitis presumed to be due to replicating virus in the eye, as evidenced by the detection of HSV viral DNA in the aqueous humor of eyes using polymerase chain reaction (PCR) analysis of aqueous humor obtained by paracentesis of the anterior chamber.^1–3 It nearly always is unilateral, may present with elevated intraocular pressure, may be chronic, in 30% to 40% of cases may occur in the context of HSV keratitis (HSV keratouveitis), or may occur as a unilateral uveitis with sectoral iris atrophy without keratitis.^4–6 In case series of patients with uveitis, it accounts for 3 to 10% of all uveitis cases and 5 to 10% of anterior uveitis cases.^7,8 The correct diagnosis affects management as oral antiviral medications typically are used in the treatment of HSV anterior uveitis, with some patients requiring chronic suppressive antiviral medication.^9,10

The Standardization of Uveitis Nomenclature (SUN) Working Group is an international collaboration, which has developed classification criteria for 25 of the most common uveitides using a formal approach to development and classification.^11–17 Among the anterior uveitides studied was HSV anterior uveitis.

Methods

The SUN Developing Classification Criteria for the Uveitides project proceeded in four phases as previously described: 1) informatics, 2) case collection, 3) case selection, and 4) machine learning.¹²,13,^15,16

Informatics.

As previously described, the consensus-based informatics phase permitted the development of a standardized vocabulary and the development of a standardized, menu-driven hierarchical case collection instrument.¹²

Case collection and case selection.

De-identified information was entered into the SUN preliminary database by the 76 contributing investigators for each disease as previously described.^15,16 Cases in the preliminary database were reviewed by committees of 9 investigators for selection into the final database, using formal consensus techniques described in the accompanying article.^15,16 Because the goal was to develop classification criteria, only cases with a supermajority agreement (>75%) that the case was the disease were retained in the final database (i.e. were “selected”).

Machine learning.

The final database then was randomly separated into a training set (~85% of cases) and a validation set (~15% of cases) for each disease as described in the accompanying article.¹⁶ Machine learning was used on the training set to determine criteria that minimized misclassification. The criteria then were tested on the validation set; for both the training set and the validation set, the misclassification rate was calculated for each disease. The misclassification rate was the proportion of cases classified incorrectly by the machine learning algorithm when compared to the consensus diagnosis. For HSV anterior uveitis, the diseases against which it was evaluated were: cytomegalovirus (CMV) anterior uveitis, varicella zoster virus (VZV) anterior uveitis, juvenile idiopathic arthritis (JIA)-associated anterior uveitis, spondylitis/HLA-B27-associated anterior uveitis, tubulointerstitial nephritis with uveitis (TINU), Fuchs uveitis syndrome, sarcoid anterior uveitis, and syphilitic anterior uveitis.

The study adhered to the principles of the Declaration of Helsinki. Institutional Review Boards (IRBs) at each participating center reviewed and approved the study; the study typically was considered either minimal risk or exempt by individual IRBs.

Results

Two hundred fifty cases of HSV anterior uveitis were collected, and 101 (40%) achieved supermajority agreement on the diagnosis during the “selection” phase and were used in the machine learning phase. These cases of HSV anterior uveitis were compared to cases of other anterior uveitides, including 89 cases of CMV anterior uveitis, 123 cases of VZV anterior uveitis, 146 cases of Fuchs Uveitis Syndrome, 202 cases of JIA-associated anterior uveitis, 184 cases of spondylitis/HLA-B27-associated anterior uveitis, 94 cases of TINU, 112 cases of sarcoidosis-associated anterior uveitis, and 32 cases of syphilitic anterior uveitis. The characteristics at presentation to a SUN Working Group Investigator of the cases with HSV anterior uveitis are listed in Table 1. The criteria developed after machine learning are listed in Table 2. Key features included evidence of HSV infection, including: 1) positive PCR for HSV in the aqueous; 2) sectoral iris atrophy (Figure 1) in a patient ≤50 years of age; or 3) HSV keratitis, either epithelial or stromal.

Table 1.

Characteristics of Cases with Herpes Simplex Anterior Uveitis

Characteristic	Result
Number cases	101
Demographics
Age, median, years (25th 75th percentile)	44 (2, 87)
Age category, years (%)
≤16	2
17–50	71
51–60	12
>60	14
Gender (%)
Men	42
Women	58
Race/ethnicity (%)
White, non-Hispanic	65
Black, non-Hispanic	3
Hispanic	3
Asian, Pacific Islander	6
Other	7
Missing	16
Uveitis History
Uveitis course (%)
Acute, monophasic	13
Acute, recurrent	31
Chronic	40
Indeterminate	16
Laterality (%)
Unilateral	99
Unilateral, alternating	0
Bilateral	1
Ophthalmic examination
Cornea
No keratitis	77
Keratitis	23
Keratic precipitates (%)
None	26
Fine	24
Round	18
Stellate	5
Mutton Fat	26
Other	2
Anterior chamber cells (%)
Grade ½+	22
1+	31
2+	25
3+	14
4+	2
Hypopyon (%)	0
Anterior chamber flare (%)
Grade 0	38
1+	47
2+	15
3+	0
4+	1
Iris (%)
Normal	36
Posterior synechiae	18
Sectoral iris atrophy	46
Patch iris atrophy	9
Diffuse iris atrophy	9
Heterochromia	1
Intraocular pressure (IOP), involved eyes
Median, mm Hg (25th, 75th percentile)	16 (12, 21)
Proportion patients with IOP>24 mm Hg either eye (%)	34
Vitreous cells (%)
Grade 0	78
1+	12
2+	6
3+	4
4+	0
Laboratory
Aqueous PCR positive for HSV (% all cases)	41
Aqueous PCR positive for HSV* (% cases tested)	100

^*PCR = polymerase chain reaction; HSV =Herpes simplex virus; 41 patients tested & 41 (100%) were positive

Table 2.

Classification Criteria for Herpes Simplex Anterior Uveitis

Criteria

1. Evidence of anterior uveitis

a. anterior chamber cells

b. if anterior vitreous cells are present, severity is less than anterior chamber inflammation

c. no evidence of retinitis

AND

2. Unilateral uveitis (unless there is a positive aqueous PCR* for herpes simplex virus)

AND

3. Evidence of herpes simplex infection in the eye

a. aqueous humor PCR positive for herpes simplex virus OR

b. sectoral iris atrophy in a patient ≤50 years of age OR

c. herpes simplex keratitis

Exclusions

1. Concomitant dermatomal/cutaneous varicella zoster virus (unless aqueous specimen PCR positive for herpes simplex virus)

2. Positive serology for syphilis using a treponemal test

3. Evidence of sarcoidosis (either bilateral hilar adenopathy on chest imaging or tissue biopsy demonstrating non-caseating granulomata)

4. Aqueous specimen PCR positive for cytomegalovirus or varicella zoster virus

^*PCR = polymerase chain reaction

Figure 1.

Sectoral iris atrophy in a patient with herpes simplex virus anterior uveitis.

As a check on the clinical criteria for patients without PCR being done, we evaluated the 14 cases ≥51 years of age with positive aqueous humor PCR for HSV for clinical features. Of the 4 cases aged 51 to 60 years, 3 of 4 had HSV keratitis in addition to the uveitis (and would have been diagnosed as HSV uveitis on clinical grounds) and 1 had no distinguishing features on clinical grounds (and hence needed PCR of the aqueous humor to make the diagnosis). Of the 10 cases aged >60 years, 5 had HSV keratitis in addition to the uveitis (and would have been diagnosed as HSV uveitis on clinical grounds), 3 had atypical iris atrophy (not sectoral, and hence needed PCR of the aqueous to make a diagnosis), and 1 had no distinguishing clinical features (and hence needed PCR off the aqueous to make the diagnosis). Therefore, none of these cases would have been diagnosed as a different infectious uveitis (e.g. VZV anterior uveitis) by the clinical criteria, supporting the utility of the HSV criteria.

The overall accuracy for anterior uveitides was 97.5% in the training set and 96.7% in the validation set (95% confidence interval 92.4, 98.6).¹⁶ The misclassification rate for HSV anterior uveitis in the training set was 8.3% and in the validation set 17%. The disease with which it most often was confused was VZV anterior uveitis.

Discussion

The classification criteria outlined in Table 2 appear to perform reasonably well with reasonably low misclassification rates. Because of the sample size of the validation set, it’s estimate is more subject to one or two cases having a disproportionate effect, and the misclassification rate in the training set was 8.3%.

Herpes simplex anterior uveitis has an appearance often distinct from non-infectious anterior uveitides but has ocular features that may overlap with other anterior uveitides caused by an active viral infection, particularly VZV anterior uveitis.^{1,3,5,6,18,19} Herpes simplex anterior uveitis occurs in a younger age group than VZV anterior uveitis, but the age distributions overlap.^3,5,18 Concomitant or recent dermatomal zoster is present in the majority of patients with VZV anterior uveitis and is useful in its diagnosis.¹⁶ Keratitis is seen in 30 to 40% of eyes with HSV anterior uveitis but <3% of eyes with VZV anterior uveitis.⁵ Endotheliitis can be seen with CMV anterior uveitis, but the coin-shaped nodular keratic precipitates seen with CMV anterior uveitis are not present in HSV uveitis, and the iris atrophy, when present in eyes with CMV anterior uveitis, tends to be patchy rather than sectoral as in HSV and VZV anterior uveitis.²⁰

The Herpetic Eye Disease Study (HEDS) used the following criteria for diagnosis of HSV anterior uveitis: anterior uveitis plus either 1) confirmed prior HSV ocular disease, 2) HSV stromal keratitis, or 3) positive serum antibodies to HSV in the absence of other identifiable uveitides.⁹ However, the HEDS enrolled patients prior to the use of PCR of the aqueous for diagnosis of viral uveitis, and positive serum antibodies to HSV have a poor positive predictive value due to their widespread presence in the general population. Studies using PCR on eyes with anterior uveitis and sectoral iris atrophy have demonstrated the presence of either HSV or VZV in over 95% of eyes.^3,6 Patients under 50 years of age typically had HSV, whereas patients over 60 year of age had VZV.^3,6 Nevertheless, the overlapping age distributions make it impossible to diagnose which virus is causative in patients aged 50 to 60 years without a positive PCR or either HSV keratitis (for HSV) or concurrent dermatomal herpes zoster (for VZV). As sectoral iris atrophy is a later manifestation of HSV anterior uveitis,¹⁸ HSV anterior uveitis can be diagnosed in its absence with either HSV keratitis or a PCR of the aqueous positive for HSV.

Anterior uveitis due to HSV is nearly always unilateral, 99% in our series, but bilateral cases do occur, often in the context of immune compromise.^19,20 Therefore, for a patient with bilateral anterior uveitis to be diagnosed as HSV anterior uveitis, a positive PCR should be obtained. Because aqueous paracentesis for PCR for viruses is not always performed in patients with unilateral anterior uveitis,²² the criteria need to provide methods for diagnosis in its absence. Hence the criteria outlined in Table 2 were developed, and they have an acceptable misclassification rate.

The presence of any of the exclusions in Table 2 suggests an alternate diagnosis, and the diagnosis of HSV anterior uveitis should not be made in their presence. In prospective studies many of these tests will be performed routinely, and the alternative diagnoses excluded. However, in retrospective studies based on clinical care, not all of these tests may have been performed. Hence the presence of an exclusionary criterion excludes HSV anterior uveitis, but the absence of such testing does not exclude the diagnosis of HSV anterior uveitis if the criteria for the diagnosis are met.

Classification criteria are employed to diagnose individual diseases for research purposes.¹⁷ Classification criteria differ from clinical diagnostic criteria, in that although both seek to minimize misclassification, when a trade-off is needed, diagnostic criteria typically emphasize sensitivity, whereas classification criteria emphasize specificity,¹⁷ in order to define a homogeneous group of patients for inclusion in research studies and limit the inclusion of patients without the disease in question that might confound the data. The machine learning process employed did not explicitly use sensitivity and specificity; instead it minimized the misclassification rate. Because we were developing classification criteria and because the typical agreement between two uveitis experts on diagnosis is moderate at best,¹⁵ the selection of cases for the final database (“case selection”) included only cases which achieved supermajority agreement on the diagnosis. It is likely that there will be some cases which clinicians believe have HSV anterior uveitis which will not meet the criteria outlined in Table 2.

In sum, the criteria outlined in Table 2 appear to perform reasonably well enough to be used as classification criteria for research.¹⁶

The Standardization of Uveitis Nomenclature (SUN) Working Group

Douglas A. Jabs, MD, MBA,^2,3 Nisha R. Acharya, MD,⁴ Laure Caspers, MD,⁵ Soon-Phaik Chee, FRCOphth, FRCS (G), FRCS (Ed), MMed (Singapore),⁶ Debra Goldstein, MD,⁷ Peter McCluskey, MD,⁸ Philip I. Murray, PhD, FRCP, FRCS, FRCOphth,⁹ Neal Oden, PhD,¹⁰ Alan G. Palestine, MD,¹¹ James T. Rosenbaum, MD,^12,13 Jennifer E. Thorne, MD, PhD,^2,3 and Brett E. Trusko, PhD, MBA.¹⁴

²Department of Epidemiology, the Johns Hopkins University Bloomberg School of Public Health, Baltimore, MD, USA

³Wilmer Eye Institute, the Department of Ophthalmology, the Johns Hopkins University School of Medicine, Baltimore, MD, USA

⁴Francis I. Proctor Foundation, the University of California, San Francisco School of Medicine, San Francisco, CA, USA

⁵Department of Ophthalmology, CHU St. Pierre, Universite Libre de Bruxelles, Brussels, Belgium

⁶Singapore National Eye Centre, Singapore Eye Research Institute, Department of Ophthalmology, Yong Loo Lin School of Medicine, National University of Singapore, Duke-NUS Medical School, Singapore

⁷Department of Ophthalmology, the Northwestern Feinberg School of Medicine, Chicago, IL, USA

⁸Save Sight Institute, Department of Ophthalmology, University of Sydney School of Medicine, Sydney, NSW, Australia

⁹Academic Unit of Ophthalmology, University of Birmingham, Birmingham, UK

¹⁰Emmes Company, LLC, Rockville, MD, USA

¹¹Department of Ophthalmology, University of Colorado School of Medicine, Aurora, Co, USA

¹²Departments of Medicine and Ophthalmology, Oregon Health and Science University, Portland, OR, USA

¹³Legacy Devers Eye Institute, Portland, OR, USA

¹⁴Department of Medicine, Texas A&M University, College Station, TX, USA.