ABSTRACT
Purpose
To determine classification criteria for toxoplasmic retinitis.
Design
Machine learning of cases with toxoplasmic retinitis and 4 other infectious posterior uveitides / panuveitides.
Methods
Cases of infectious posterior uveitides / panuveitides were collected in an informatics-designed preliminary database, and a final database was constructed of cases achieving supermajority agreement on 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 infectious posterior uveitides / panuveitides. The resulting criteria were evaluated on the validation set.
Results
Eight hundred three cases of infectious posterior uveitides / panuveitides, including 174 cases of toxoplasmic retinitis, were evaluated by machine learning. Key criteria for toxoplasmic retinitis included focal or paucifocal necrotizing retinitis and either positive polymerase chain reaction assay for Toxoplasma gondii from an intraocular specimen or the characteristic clinical picture of a round or oval retinitis lesion proximal to a hyperpigmented and/or atrophic chorioretinal scar. Overall accuracy for infectious posterior uveitides / panuveitides was 92.1% in the training set and 93.3% (95% confidence interval 88.2, 96.3) in the validation set. The misclassification rates for toxoplasmic retinitis were 8.2% in the training set and 10% in the validation set.
Conclusions
The criteria for toxoplasmic retinitis had a low misclassification rate and seemed to perform sufficiently well for use in clinical and translational research.
T oxoplasma gondii is a ubiquitous parasite worldwide and is the most common cause of retinal infection in most populations, resulting in a substantial burden of eye disease and vision loss. , T. gondii reproduces sexually only in the gut of felines, but can reproduce asexually in most other mammals and in birds. Infection occurs via 1 of several routes: through ingestion of materials contaminated with cat feces that contain öocysts; by eating raw or undercooked tissue of infected intermediate hosts; or vertically from mother to an unborn child during pregnancy. With rare exception, vertical transmission occurs only when the mother is first infected during the pregnancy.
In intermediate hosts, including food animals and human beings, öocysts become tachyzoites, the proliferative form of the parasite that can cause clinical disease, but parasites eventually encyst in various tissues, including the retina. These tissue cysts contain the bradyzoite form of parasite that does not induce clinical disease, but remains viable for prolonged periods of time. Tissue cysts reactivate from time to time, releasing bradyzoites, which again convert back to tachyzoites, but the factors that cause reactivation are poorly understood. Proliferation of tachyzoites is self-limited in people with normal immune function. People with postnatally acquired infections may develop a transient illness characterized by lymphadenopathy, fever, and sore throat, but the initial infection often is asymptomatic. Ocular involvement may occur at the time of initial systemic infection or months to years later.
In the United States the age-adjusted seroprevalence of anti– T. gondii antibodies during the period 2011-2014 was approximately 10%. Seroprevalence increases with age and is higher among men, socioeconomically disadvantaged groups, and individuals born outside the United States. It is estimated that, overall, 2% of T. gondii –infected individuals in the United States have ocular involvement. The risk of ocular involvement also is believed to be higher among Hispanic immigrants, presumably because they are infected in their countries of origin by endemic parasites of greater virulence (see below). In 2010, it was estimated that nearly 5,000 people in the United States would develop symptomatic ocular toxoplasmosis each year.
The rates of both infection and ocular involvement are higher in many other parts of the world; for example, in the area of Erechim, in southern Brazil, results of a population-based study showed that 21.3% of individuals over the age of 13 years had ocular toxoplasmosis, and in a prior study from the same area, 98 of 100 children aged 10-15 were infected with T. gondii .
The primary ocular site of T. gondii infection is the retina and eventually may result in full-thickness retinal necrosis. The subjacent choroid can also be destroyed, presumably by accompanying inflammation, which ultimately results in an atrophic scar with white center, owing to exposure of the sclera, and a variably pigmented border as the lesion becomes inactive. Tissue cysts are believed to persist at scar borders after resolution of an active episode. Recurrences arising from these tissue cysts account for the classic appearance of toxoplasmic retinitis: a focus of intense tissue inflammation adjacent to a pre-existing retinochoroidal scar. Not all lesions arise from scars, however. “Primary lesions” (those arising from normal-appearing retina) may occur at the time of an initial infection or may occur later; these late primary lesions are thought to arise from organisms that encyst in the retina at the time of initial infection, but do not immediately cause clinically apparent disease. ,
The clinical appearance of toxoplasmic retinal lesions may vary, based on the duration of parasite proliferation before encystment, and on the severity of associated inflammation , Infections that resolve early, with minimal inflammation, may result in only multiple small outer retinal opacities, a presentation of disease termed “punctate outer retinal toxoplasmosis.” Conversely, persistent infection, as may occur in immunocompromised individuals, may result in large areas of retinal necrosis, possibly mimicking other forms of necrotizing retinitis, such as cytomegalovirus (CMV) retinitis. Occasionally, elderly individuals may develop extensive lesions. , ,
Variation in prevalence of infection and risk for ocular involvement among otherwise healthy individuals appears to reflect parasite strains of different virulence. , Genotypes of parasites endemic to different geographic areas vary considerably; the presence of more virulent strains in food animals of southern Brazil is thought to explain the fact that ocular toxoplasmosis is more prevalent and more severe in that region than in the United States.
Because of the relatively high seroprevalence of T. gondii antibodies in the general population and the relatively low risk of ocular involvement, the presence of IgG antibodies to T. gondii typically is not a useful feature for diagnosing toxoplasmic retinitis; however, a negative serologic test may help to exclude toxoplasmic retinochoroiditis in a patient with a nonspecific focus of retinal inflammation. Conversely, the presence of IgM antibodies may provide information about recently acquired systemic toxoplasmosis. Polymerase chain reaction (PCR) techniques can be used to identify T. gondii DNA in ocular fluids, and are particularly helpful in diagnosing ocular toxoplasmosis in patients with unusual presentations of disease.
Although retinal lesions are self-limited in otherwise healthy individuals, it is believed that treatment with a combination of antimicrobial agents and corticosteroid will reduce tissue damage from associated inflammation. There is no consensus regarding the best antimicrobial agents; most common is use of both a dihydrofolate reductase inhibitor and a sulfonamide, such as pyrimethamine and sulfadiazine or combination trimethoprim-sulfamethoxazole. Despite the absence of class I clinical trials demonstrating the efficacy of antimicrobial treatment of ocular toxoplasmosis, 1 comparative trial in which treatment was assigned by clinical center reported that treatment with pyrimethamine and sulfadiazine resulted in smaller scars than did no treatment, suggesting efficacy in limiting retinal damage. In this trial the recurrence rate was unaffected by the short-term course of treatment. Subsequent small clinical trials suggested efficacy similar to pyrimethamine and sulfadiazine for trimethoprim-sulfamethoxazole, for pyrimethamine and azithromycin, and for intravitreal clindamycin and dexamethasone. Retrospective cohort data suggest that treatment of ocular toxoplasmosis with corticosteroids alone is associated with increased risks of fulminant toxoplasmic retinitis, ocular recurrences, and worse visual outcomes ; therefore, such management generally is discouraged. Severely immunocompromised patients can be treated with an antimicrobial agent alone and are likely to require continued antimicrobial therapy to maintain lesion inactivity. , Treatment with currently available drugs does not eliminate tissue cysts, but continued treatment with an antimicrobial agent, such as trimethoprim-sulfamethoxazole, reduces the risk of recurrences. ,
The Standardization of Uveitis Nomenclature (SUN) Working Group is an international collaboration that has developed classification criteria for 25 of the most common uveitides using a formal approach to development and classification. Among the diseases studied was toxoplasmic retinitis.
Methods
The SUN Developing Classification Criteria for the Uveitides project proceeded in 4 phases, as previously described: (1) informatics, (2) case collection, (3) case selection, and (4) machine learning.
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.
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. , 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. , Because the goal was to develop classification criteria, only cases with a supermajority agreement (>75%) that the case was the disease in question were retained in the final database (ie, were “selected”). ,
Machine Learning
The final database then was randomly separated into a training set (∼85% of the cases) and a validation set (∼15% of the cases) for each disease, as described in the accompanying article. 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 infectious posterior uveitides and panuveitides, the diseases against which toxoplasmic retinitis was evaluated were acute retinal necrosis (ARN), CMV retinitis, syphilitic uveitis, and tubercular uveitis.
Results
Two hundred thirteen cases of toxoplasmic retinitis were collected and 174 (82%) achieved supermajority agreement on the diagnosis during the “selection” phase and were used in the machine learning phase. These cases of toxoplasmic retinitis were compared to cases of infectious posterior uveitides / panuveitides, including 186 cases of ARN, 211 cases of CMV retinitis, 35 cases of syphilitic posterior uveitis, and 197 cases of tubercular uveitis. The details of the machine learning results for these diseases are outlined in the accompanying article. The characteristics of cases with ocular toxoplasmosis are listed in Table 1 , and the classification criteria developed after machine learning are listed in Table 2 . Key features of the criteria include a unifocal or paucifocal (<5 lesions) active retinitis and either (1) evidence of infection with T. gondii , either from PCR of an intraocular fluid specimen or serum IgM antibodies to T. gondii (evidence of acute infection); or (2) classic clinical picture ( Figure 1 ) with hyperpigmented and/or atrophic scar accompanied by either a round or oval area of active retinitis or a recurrent area of active retinitis. The overall accuracy for infectious posterior uveitides / panuveitides was 92.1% in the training set and 93.3% (95% confidence interval 88.2, 96.3%) in the validation set. The misclassification rate for toxoplasmic retinitis in the training set was 8.2% and in the validation set 10%. In the training set the disease with which it was most often confused was CMV retinitis, whereas in the validation set no one disease predominated.
| Characteristic | Result |
|---|---|
| Number of cases | 174 |
| Demographics | |
| Age, median, years (25th, 75th percentile) | 28 (21, 43) |
| Sex (%) | |
| Male | 50 |
| Female | 50 |
| Race/ethnicity (%) | |
| White, non-Hispanic | 55 |
| Black, non-Hispanic | 8 |
| Hispanic | 9 |
| Asian, Pacific Islander | 11 |
| Other | 13 |
| Missing | 4 |
| Uveitis history | |
| Uveitis course (%) | |
| Acute, monophasic | 43 |
| Acute, recurrent | 24 |
| Chronic | 26 |
| Indeterminate | 7 |
| Laterality (%) | |
| Unilateral | 88 |
| Unilateral, alternating | 0 |
| Bilateral | 12 |
| Ophthalmic examination | |
| Keratic precipitates (%) | |
| None | 61 |
| Fine | 18 |
| Round | 10 |
| Stellate | 1 |
| Mutton fat | 9 |
| Other | 1 |
| Anterior chamber cells, grade (%) | |
| 0 | 45 |
| ½+ | 14 |
| 1+ | 16 |
| 2+ | 12 |
| 3+ | 10 |
| 4+ | 3 |
| Anterior chamber flare, grade (%) | |
| 0 | 63 |
| 1+ | 25 |
| 2+ | 9 |
| 3+ | 2 |
| 4+ | 1 |
| Iris (%) | |
| Normal | 97 |
| Posterior synechiae | 3 |
| Iris nodules | 0 |
| Iris atrophy (sectoral, patchy, or diffuse) | 0 |
| Heterochromia | 0 |
| IOP, involved eyes | |
| Median, mm Hg (25th, 75th percentile) | 16 (13, 18) |
| Proportion of patients with IOP > 24 mm Hg either eye (%) | 7 |
| Vitreous cells, grade (%) | |
| 0 | 21 |
| ½+ | 13 |
| 1+ | 30 |
| 2+ | 27 |
| 3+ | 7 |
| 4+ | 2 |
| Vitreous haze, grade (%) | |
| 0 | 30 |
| ½+ | 19 |
| 1+ | 27 |
| 2+ | 14 |
| 3+ | 9 |
| 4+ | 1 |
| Retinitis characteristics | |
| Number of lesions per eye, including active lesions & scars (%) a | |
| Unifocal (1) | 5 |
| Paucifocal (2-4) | 82 |
| Multifocal (≥5) | 8 |
| Indeterminate (lesion not photographed or dense vitritis) | 5 |
| Number of active lesions per eye (%) a | |
| Unifocal (1) | 78 |
| Paucifocal (2-4) | 2 |
| Multifocal (≥5) | 0 |
| Indeterminate (lesion not photographed or dense vitritis) | 20 |
| Proximate/adjacent hyperpigmented/atrophic scars (%) a | |
| Present | 82 |
| Absent (active lesion only) | 16 |
| Indeterminate (dense vitritis) | 2 |
| Lesion shape (%) | |
| Round or ovoid | 59 |
| Placoid | 16 |
| Ameboid | 9 |
| Wedge-shaped | 3 |
| Punctate | 2 |
| Missing | 11 |
| Lesion character (%) | |
| Circumferential | 1 |
| Confluent | 7 |
| Granular | 1 |
| Lesion location (%) | |
| Posterior pole involved | 52 |
| Midperiphery and/or periphery only | 48 |
| Lesion size (%) | |
| <250 µm | 6 |
| 250-500 µm | 10 |
| >500 µm | 84 |
| Other features (%) | |
| Retinal vascular sheathing or leakage or occlusion | 17 |
| Retinal hemorrhage | 6 |
| Systemic disease | |
| Immunocompromised patients (%) | 6 |
| Human immunodeficiency virus infection | 5 |
| Organ transplant | 0 |
| Chemotherapy or other immunosuppression | 1 |
| Laboratory data (%) | |
| Aqueous or vitreous specimen PCR b positive for Toxoplasma gondii | 10 |
| Positive serology for antibodies to Toxoplasma gondii c | 75 |
| Positive IgM antibodies to Toxoplasma gondii | 21 |
| Positive IgG antibodies to Toxoplasma gondii | 74 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
