Academic Background

Tubercular Uveitis (TBU) is an ocular condition caused by Mycobacterium tuberculosis infection or immune-mediated inflammation. While tuberculosis (TB) remains a major global public health issue, the diagnosis and treatment of TBU continue to pose significant challenges. TBU presents a wide range of clinical manifestations, with a lack of specific diagnostic criteria often leading to misdiagnosis or delayed treatment. Traditional diagnostic methods, such as the Tuberculin Skin Test (TST) and Interferon Gamma Release Assay (IGRA), can only indicate TB infection but do not assess whether the infection is active or directly related to ocular inflammation. Moreover, the sensitivity of Polymerase Chain Reaction (PCR) testing of ocular fluids is low, and the invasive nature of sample collection adds further complexity to the diagnosis.

To address these challenges, the Collaborative Ocular Tuberculosis Study (COTS) group developed an online tool — the COTS calculator. This tool is designed to assist clinicians in determining whether to initiate antitubercular therapy (ATT) in the absence of conclusive diagnostic evidence. It integrates a patient’s clinical phenotype, regional TB endemicity, TST/IGRA results, and chest imaging findings to generate a score that guides ATT initiation.

Paper Source

This study was authored by an international team of experts, including Ludi Zhang, William Rojas-Carabali, Shannon Sheriel Choo, and others. The researchers represent institutions such as Nanyang Technological University (Singapore), Stanford Medicine (USA), and Sankara Nethralaya (India). The paper was published online on October 31, 2024, in the journal JAMA Ophthalmology.

Study Design and Methods

Study Design

This diagnostic study aimed to evaluate the accuracy of the COTS calculator in guiding ATT initiation. The data was sourced from the COTS-1 study, a retrospective observational study that included 962 suspected TBU patients from 25 international ophthalmology centers between January 2004 and December 2014. A total of 492 patients met the study’s inclusion criteria and underwent a 12-month follow-up after treatment.

Study Tool

The COTS calculator generates a score from 1 to 5 based on five clinical parameters: 1. Clinical phenotype (e.g., anterior uveitis, intermediate uveitis, pan uveitis, retinal vasculitis, or choroiditis); 2. TB endemicity of the region (endemic or non-endemic); 3. TST result; 4. IGRA result; 5. Chest imaging findings (e.g., chest X-ray or CT scan).

A score of 5 indicates an 81%-100% likelihood of initiating ATT, while a score of 1 suggests a 0%-20% likelihood.

Data Analysis

The study compared the diagnostic accuracy of three tests: 1. Clinician judgment (Test 1); 2. COTS calculator scores of 4 or 5 (Test 2); 3. COTS calculator score of 5 alone (Test 3).

A 2 × 2 table was constructed to calculate sensitivity, specificity, positive predictive value (PPV), precision, recall, and F1 score. Additionally, subgroup analyses stratified results by TB endemicity (endemic vs. non-endemic regions).

Study Results

Key Findings

Among the 492 participants, the COTS calculator identified 225 patients (45.7%) with a high or very high likelihood of ATT initiation (scores of 4 or 5), of which 111 (22.5%) had a very high likelihood (score of 5). COTS-5 demonstrated the highest specificity (88.7%), while clinician judgment showed the highest sensitivity (95.5%). COTS-4 and COTS-5 together achieved a balance between specificity (64.3%) and sensitivity (48.8%).

Subgroup Analysis

In endemic regions, the PPV and sensitivity of COTS-5 were higher than in non-endemic regions, indicating that the COTS calculator holds greater value in reducing misdiagnosis in high-prevalence settings.

Conclusions and Implications

This study demonstrated that the COTS calculator (score ≥4) provides greater specificity compared to clinician judgment alone in guiding ATT initiation. While clinician judgment is effective in identifying potential true positives with high sensitivity, combining it with the high PPV of COTS-5 could reduce false positives. This tool is especially suitable for high-prevalence, low-resource settings, enabling more selective ATT recommendations and minimizing unnecessary treatments.

Research Highlights

1. Innovative Tool: The COTS calculator is the first consensus-based online tool designed to assist clinicians in deciding ATT initiation when diagnostic evidence is inconclusive.
2. High Specificity: COTS-5 achieved the highest specificity, effectively reducing the rate of misdiagnosis.
3. Broad Applicability: The tool is particularly valuable in TB-endemic regions, offering diagnostic support in resource-limited settings.

Limitations of the Study

1. Retrospective Design: The retrospective nature of the study made it challenging to determine true disease status, especially given the lack of definitive diagnostic standards.
2. Selection Bias: The COTS-1 database comprised patients with a high index of suspicion for TBU, possibly leading the COTS calculator to favor higher scores.
3. Variability Across Centers: Differences in follow-up practices and treatment approaches across centers may have introduced bias into the results.

Future Research Directions

Future studies should explore the predictive accuracy of the COTS calculator using larger and more heterogeneous patient populations from different regions. Additionally, prospective studies incorporating genomic or microbiological confirmation would enhance diagnostic accuracy.

Summary

This study validated the accuracy of the COTS calculator in guiding ATT initiation, showing its value in reducing misdiagnosis and optimizing treatment decisions. The tool’s selective approach to ATT initiation minimizes unnecessary treatments and potential drug-related adverse effects, demonstrating particular utility in TB-endemic regions. Clinician judgment should serve as an initial guide, followed by consultation with the COTS calculator to achieve a more judicious use of ATT.