Prediction of sustained biologic and targeted synthetic DMARD-free remission in rheumatoid arthritis patients

Explainable deep learning for disease activity prediction in chronic inflammatory joint diseases

Analysing complex diseases such as chronic inflammatory joint diseases (CIJDs), where many factors influence the disease evolution over time, is a challenging task. CIJDs are rheumatic diseases that cause the immune system to attack healthy organs, mainly the joints. Different environmental, genetic and demographic factors affect disease development and progression. The Swiss Clinical Quality Management in Rheumatic Diseases (SCQM) Foundation maintains a national database of CIJDs documenting the disease management over time for 19'267 patients. We propose the Disease Activity Score Network (DAS-Net), an explainable multi-task learning model trained on patients' data with different arthritis subtypes, transforming longitudinal patient journeys into comparable representations and predicting multiple disease activity scores. First, we built a modular model composed of feed-forward neural networks, long short-term memory networks and attention layers to process the heterogeneous patient histories and predict future disease activity. Second, we investigated the utility of the model's computed patient representations (latent embeddings) to identify patients with similar disease progression. Third, we enhanced the explainability of our model by analysing the impact of different patient characteristics on disease progression and contrasted our model outcomes with medical expert knowledge. To this end, we explored multiple feature attribution methods including SHAP, attention attribution and feature weighting using case-based similarity. Our model outperforms temporal and non-temporal neural network, tree-based, and naive static baselines in predicting future disease activity scores. To identify similar patients, a k-nearest neighbours regression algorithm applied to the model's computed latent representations outperforms baseline strategies that use raw input features representation.

Prevalence and predictors of sustained remission/low disease activity after discontinuation of induction or maintenance treatment with tumor necrosis factor inhibitors in rheumatoid arthritis: a systematic and scoping review

BACKGROUND

To determine the prevalence of sustained remission/low disease activity (LDA) in patients with rheumatoid arthritis (RA) after discontinuation of tumor necrosis factor inhibitors (TNFi), separately in induction treatment and maintenance treatment studies, and to identify predictors of successful discontinuation.

METHODS

We performed a systematic literature review of studies published from 2005 to May 2022 that reported outcomes after TNFi discontinuation among patients in remission/LDA. We computed prevalences of successful discontinuation by induction or maintenance treatment, remission criterion, and follow-up time. We performed a scoping review of predictors of successful discontinuation.

RESULTS

Twenty-two induction-withdrawal studies were identified. In pooled analyses, 58% (95% confidence interval (CI) 45, 70) had DAS28 < 3.2 (9 studies), 52% (95% CI 35, 69) had DAS28 < 2.6 (9 studies), and 40% (95% CI 18, 64) had SDAI ≤ 3.3 (4 studies) at 37-52 weeks after discontinuation. Among patients who continued TNFi, 62 to 85% maintained remission. Twenty-two studies of maintenance treatment discontinuation were also identified. At 37-52 weeks after TNFi discontinuation, 48% (95% CI 38, 59) had DAS28 < 3.2 (10 studies), and 47% (95% CI 33, 62) had DAS28 < 2.6 (6 studies). Heterogeneity among studies was high. Data on predictors in induction-withdrawal studies were limited. In both treatment scenarios, longer duration of RA was most consistently associated with less successful discontinuation.

CONCLUSIONS

Approximately one-half of patients with RA remain in remission/LDA for up to 1 year after TNFi discontinuation, with slightly higher proportions in induction-withdrawal settings than with maintenance treatment discontinuation.

Factors associated with discontinuation of biologics in patients with inflammatory arthritis in remission: data from the BIOBADASER registry

BACKGROUND

The objectives of this study were to assess the discontinuation of biologic therapy in patients who achieve remission and identify predictors of discontinuation of biologics in patients with inflammatory arthritis in remission.

METHODS

An observational retrospective study from the BIOBADASER registry comprising adult patients diagnosed with rheumatoid arthritis (RA), ankylosing spondylitis (AS), or psoriatic arthritis (PsA) and receiving 1 or 2 biological disease-modifying drugs (bDMARDs) between October 1999 and April 2021. Patients were followed yearly after initiation of therapy or until discontinuation of treatment. Reasons for discontinuation were collected. Patients who discontinued bDMARDs because of remission as defined by the attending clinician were studied. Predictors of discontinuation were explored using multivariable regression models.

RESULTS

The study population comprised 3,366 patients taking 1 or 2 bDMARDs. Biologics were discontinued owing to remission by 80 patients (2.4%): 30 with RA (1.7%), 18 with AS (2.4%), and 32 with PsA (3.9%). The factors associated with a higher probability of discontinuation on remission were shorter disease duration (OR: 0.95; 95% CI: 0.91-0.99), no concomitant use of classic DMARDs (OR: 0.56; 95% CI: 0.34-0.92), and shorter usage of the previous bDMARD (before the decision to discontinue biological therapy) (OR: 1.01; 95% CI: 1.01-1.02); in contrast, smoking status (OR: 2.48; 95% CI: 1.21-5.08) was associated with a lower probability. In patients with RA, positive ACPA was associated with a lower probability of discontinuation (OR: 0.11; 95% CI: 0.02-0.53).

CONCLUSIONS

Discontinuation of bDMARDs in patients who achieve remission is uncommon in routine clinical care. Smoking and positive ACPA in RA patients were associated with a lower probability of treatment discontinuation because of clinical remission.

Tapering biologic DMARDs in rheumatoid arthritis

With the arrival of biologics and the shift toward treat-to-target therapy, the possibility of a sustained clinical response has become an achievable goal for many patients with rheumatoid arthritis (RA). Although biologics have revolutionized the treatment of RA, they are costly, potentially inconvenient, and carry risks of side effects. Whether they can or should be tapered in patients with tight disease control is a matter of clinical uncertainty. The major international rheumatology professional societies have all issued guidelines on this question, but across recommendations, consensus is lacking on how and when to consider therapy de-escalation. Recent evidence suggests that sustained remission or low disease activity is more attainable with dose reduction as opposed to outright discontinuation of biologic therapy, and certain predictors of successful taper have begun to be described. This article will (1) summarize the current evidence base for biologic tapering in RA, (2) outline real-world outcomes findings, (3) review important contextual factors relevant to therapy de-escalation, such as cost-effectiveness considerations and patient perspectives, and (4) conclude by summarizing current guidelines.