Predictors Of Relapse Research Articles

Magnetic Resonance-Guided Focused Ultrasound Thalamotomy in a Prospective Cohort of 52 Patients with Parkinson's Disease: A Possible Critical Role of Age and Lesion Volume for Predicting Tremor Relapse.

Magnetic resonance-guided focused ultrasound (MRgFUS) thalamotomy of ventral intermediate (Vim) nucleus is useful to treat drug-resistant tremor-dominant Parkinson's disease (TdPD), but tremor relapse may occur. Predictors of relapse have been poorly investigated so far. The aim of this study is to evaluate the role of clinico-demographic, procedural, and neuroradiological variables in determining clinical response, relapse, and adverse events (AEs) in TdPD after MRgFUS Vim-thalamotomy. Fifty-two TdPD patients who consecutively underwent unilateral MRgFUS Vim-thalamotomy were prospectively evaluated at baseline and after 24 hours, 1 month, 6 months, and 12 months using MDS-UPDRS-III in off and on medication conditions. AEs were collected at each evaluation. Lesion volume was calculated at 24-hour magnetic resonance imaging (MRI). Patients with tremor improvement <30% in off medication were considered nonresponders (when detected after 24 hours) or relapsers (if detected from 1-month visit onward). All patients showed tremor improvement >30% at 24 hours. Tremor relapse occurred in 12 patients (23%), exclusively during the first month after thalamotomy. Relapse was associated with younger age (P = 0.030) and smaller lesion volume (P = 0.030). At 1 month, 22 patients (42%) had AEs; at 6 and 12 months, AEs persisted in 19% and 6% of cases. AEs at 6 months were associated with larger lesions (P = 0.018). All AEs were mild. MRgFUS Vim-thalamotomy is effective in treating tremor in TdPD. Relapse is associated with younger age and smaller lesion volume, but larger lesions make AEs more likely to persist. We suggest that a lesion volume between 145 and 220 mm3 on T1-weighted MRI may be the therapeutic window that ensures tremor control without long-lasting AEs. © 2025 The Author(s). Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Establishment and validation a relapse prediction model for bipolar disorder

Establishment and validation a relapse prediction model for bipolar disorder

Relationship Between an Interleukin 6 SNP and Relapse After Allogeneic Bone Marrow Transplantation

Background/Objectives: Unrelated bone marrow transplantation (BMT) is a curative treatment for hematological malignancies. While HLA mismatch is a recognized risk factor in unrelated BMT, the significance of non-HLA single nucleotide polymorphisms (SNPs) remains uncertain. Cytokines play key roles in several aspects of unrelated BMT. Although the relationship between cytokine gene SNPs and BMT outcomes has been examined, the findings obtained have been inconsistent; therefore, further investigations in additional cohorts are warranted. Methods: Four SNPs in the IL2, IL6, IFN-gamma, and TGF-beta1 genes were retrospectively genotyped in 822 malignant patients and their corresponding donors who received unrelated BMT through the Japan Marrow Donor Program with compatibility at minimum HLA-A, -B, and -DRB1. The relationships between these SNP genotypes and BMT outcomes were statistically analyzed. Results: The donor interleukin-6 (IL6) SNP, rs1800796, also known as -572G&gt;C and -634C/G, was associated with the relapse of the original disease in both univariable and multivariable regression analyses (minimum p-value = 0.0013), and the cumulative incidence curve analysis identified CC as a risk genotype (p-value = 0.0012). None of these SNPs correlated with overall survival. Conclusions: The donor IL6 SNP, rs1800796, may serve as a useful predictor of tumor relapses if validated.

The clinical characteristics and treatment outcomes of patients with systemic polyarteritis nodosa (PAN): a single centre study from India.

To describe the clinical profile and compare the long-term outcomes of patients with S-PAN treated with various treatment regimens at our centre in the last 2 decades. Data regarding clinical presentation, treatment allocation, relapses and outcomes of patients fulfilling American College of Rheumatology (ACR) 1990 criteria for PAN in the last 2 decades were recorded from electronic medical records. Relapse-free survival and predictors were analysed using KM survival statistics and regression analysis. Altogether, 53 patients including 2 with hepatitis B infection were included. Cutaneous lesions and peripheral neuropathy were the commonest manifestations. Most patients (64.2%) presented with a five-factor score (FFS) of 0. Disease-attributable hypertension and peripheral gangrene were the most common manifestations of severe disease. During a median follow-up period of 53.5 months in 49 patients, 43 (87.8%) attained complete response while 3(6.1%) had a partial response. Nineteen (40.4%) patients relapsed at a median duration of 82 (IQR 36.3-127.7) months. The relapse-free survival in patients who received induction with mycophenolate (n = 26), was comparable to that with cyclophosphamide (n = 21) [adjusted HR : 0.68]. Smoking history was an independent predictor of relapse (HR = 6.28, p= 0.013) while age was protective (HR = 0.94, p= 0.015). FFS and BVAS at 3 months were among the predictors of mortality (total deaths = 5). In our cohort of S-PAN, relapses were observed in 40.4% of patients. Mycophenolate was similar to cyclophosphamide in maintaining relapse-free survival. Only 10% fatality was recorded. FFS and BVAS at 3 months were predictors of mortality.

Reconsidering remission in recurrent late-life depression: clinical presentation and phenotypic predictors of relapse following successful antidepressant treatment.

Late-life depression (LLD) is characterized by repeated recurrent depressive episodes even with maintenance treatment. It is unclear what clinical and cognitive phenotypic characteristics present during remission predict future recurrence. Participants (135 with remitted LLD and 69 comparison subjects across three institutions) completed baseline phenotyping, including psychiatric, medical, and social history, psychiatric symptom and personality trait assessment, and neuropsychological testing. Participants were clinically assessed every two months for two years while receiving standard antidepressant treatment. Analyses examined group differences in phenotypic measure using general linear models. Concurrent associations between phenotypic measures and diagnostic groups were examined using LASSO logistic regression. Sixty (44%) LLD participants experienced a relapse over the two-year period. Numerous phenotypic measures across all domains differed between remitted LLD and comparison participants. Only residual depressive symptom severity, rumination, medical comorbidity, and executive dysfunction significantly predicted LLD classification. Fewer measures differed between relapsing and sustained remission LLD subgroups, with the relapsing group exhibiting greater antidepressant treatment intensity, greater fatigue, rumination, and disability, higher systolic blood pressure, greater life stress and lower instrumental social support. Relapsing group classification was informed by antidepressant treatment intensity, lower instrumental social support, and greater life stress. A wide range of phenotypic factors differed between remitted LLD and comparison groups. Fewer measures differed between relapsing and sustained remission LLD subgroups, with less social support and greater stress informing vulnerability to subsequent relapse. This research suggests potential targets for relapse prevention and emphasizes the need for clinically translatable relapse biomarkers to inform care.

A vision-language foundation model for precision oncology.

Clinical decision-making is driven by multimodal data, including clinical notes and pathological characteristics. Artificial intelligence approaches that can effectively integrate multimodal data hold significant promise in advancing clinical care1,2. However, the scarcity of well-annotated multimodal datasets in clinical settings has hindered the development of useful models. In this study, we developed the Multimodal transformer with Unified maSKed modeling (MUSK), a vision-language foundation model designed to leverage large-scale, unlabelled, unpaired image and text data. MUSK was pretrained on 50 million pathology images from 11,577 patients and one billion pathology-related text tokens using unified masked modelling. It was further pretrained on one million pathology image-text pairs to efficiently align the vision and language features. With minimal or no further training, MUSK was tested in a wide range of applications and demonstrated superior performance across 23 patch-level and slide-level benchmarks, including image-to-text and text-to-image retrieval, visual question answering, image classification and molecular biomarker prediction. Furthermore, MUSK showed strong performance in outcome prediction, including melanoma relapse prediction, pan-cancer prognosis prediction and immunotherapy response prediction in lung and gastro-oesophageal cancers. MUSK effectively combined complementary information from pathology images and clinical reports and could potentially improve diagnosis and precision in cancer therapy.

Clinical picture, outcomes and predictors of relapse in eosinophilia-associated coronary vasospasm: Data from a European multicentric study

Clinical picture, outcomes and predictors of relapse in eosinophilia-associated coronary vasospasm: Data from a European multicentric study

Circulating tumor DNA detection improves relapse prediction in epithelial ovarian cancer

BackgroundEpithelial ovarian cancer (EOC) is a lethal form of gynecological malignancy. Some EOC patients experience relapse after standard primary debulking surgery (PDS) and adjuvant chemotherapy (ACT). Identifying molecular residual disease (MRD) by circulating tumor DNA (ctDNA) detection can timely signal the potential for relapse. However, research on the usage of ctDNA for MRD detection in EOC is limited.MethodsFifty-one EOC patients who received standard PDS and ACT were included. Targeted sequencing based on a panel of 1021 cancer-related genes, along with further validation using Enrich-rare-mutation sequencing, was performed on tumor tissues acquired during PDS and on plasma samples collected before and after PDS/ACT to identify variants reflecting tumor signals.ResultsPost-surgery MRD was associated with relapse (Log-rank p = 0.0006) and was identified as an independent prognostic factor (HR, 3.4; 95% CI, 1.02–11.42; p = 0.047). The negative and positive predictive values were 0.83 and 0.62 respectively. Additionally, post-surgery MRD outperformed CA125 in predicting relapse, and integrating both parameters could provide more accurate risk stratification. Post-ACT MRD detection identified the patients with ctDNA clearance who were still at risk of relapse. Furthermore, baseline ctDNA detection could help determine patients who are not suitable for further tests after surgery.ConclusionsPost-surgery MRD is superior to CA125 in predicting relapse in EOC. Patients exhibiting transient ctDNA clearance, as evaluated by post-ACT MRD, may require longitudinal monitoring. Baseline ctDNA detection could help determine whether post-surgery ctDNA monitoring should be performed.

Poor prognostic factors for relapse of interstitial lung disease in microscopic polyangiitis: the Japanese multicentre REVEAL cohort study

BackgroundThis study investigated poor prognostic factors for the relapse of interstitial lung disease (ILD) in patients with microscopic polyangiitis (MPA) after remission induction therapy.MethodsWe enrolled patients diagnosed with MPA complicated by ILD according to the Chapel Hill Consensus definition from 2001 to 2023 in multiple institutions in the REVEAL cohort. All patients who were treated with immunosuppressive therapy were followed up, and those who relapsed with ILD were extracted in this study. We explored the risk factors for predicting ILD relapse in patients with MPA-ILD by comparing the demographic, clinical, laboratory, and radiological findings and treatments between the relapsed and non-relapsed groups on admission.ResultsOf 243 patients with MPA, 134 (55.1%) with MPA-ILD were enrolled. Among them, 28 (20.9%) relapsed during a mean follow-up of 4.2 years. The initial serum Krebs von den Lungen-6 (KL-6) and surfactant protein-D (SP-D) levels and the prevalence of usual interstitial pneumonia (UIP) pattern were significantly higher in the relapsed group. The biomarkers were also risk factors for relapse in multivariate Cox regression analysis. The best cut-off values of KL-6, SP-D for predicting ILD relapse were 430 U/mL and 89.5 ng/mL, respectively. We created prediction models based on the best cut-off values for KL-6, SP-D, and the presence of the UIP pattern (KSU model). The 10-year relapse rate was significantly different among patients with MPA-ILD stratified by the number of risk factors based on the KSU model. A higher relapse rate was associated with higher all-cause mortality.ConclusionsThe initial serum high KL-6 and SP-D levels and the prevalence of the UIP pattern were associated with ILD relapse in patients with MPA-ILD. Our multicentre cohort study indicated that the KSU model, which consists of KL-6 ≥ 430 U/mL, SP-D ≥ 89.5 ng/mL, and the presence of the UIP pattern, is a useful predictor of ILD relapse in patients with MPA after immunosuppressive therapy.

Factors Influencing Fronto-Orbital Relapse in Patients With Syndromic Craniosynostosis: A 2 Decade Review.

Fronto-orbital retrusion may occur after primary surgical correction of craniosynostosis, particularly in patients with syndromic craniosynostosis. This study investigated reoperation rates and factors contributing to FO relapse among this cohort. A retrospective review evaluated reoperation for FO relapse in patients with syndromic multisuture craniosynostosis who underwent primary fronto-orbital advancement (FOA) + calvarial vault remodeling (CVR) at our institution between 2004 and 2024. Revision surgeries included repeat FOA or monobloc advancement/distraction. FOA advancement distance was measured using postoperative computed tomography and Mimics software. ROC analysis evaluated the accuracy of FOA distance in predicting subsequent FO relapse. Conditional margins identified optimal advancement distances. Logistic regression of predictors of FO relapse adjusted for age at surgery, craniofacial syndrome, posterior vault distraction osteogenesis (PVDO), advancement distance, and postoperative helmet therapy. Fifty-two patients underwent a mean of 2.8±1.9 skeletal craniofacial procedures each. With a mean follow-up time of 9.2±6.5 years, 16 (30.8%) patients required reoperation for FO relapse. Larger advancement distances were the sole significant predictor of relapse, increasing the odds by 49.6% (OR 1.496, 95% CI: 1.085-2.063; P=0.014). Relapse rates were lower with advancements ≤17.2mm (0.0%) than with further advancements (42.0%, P=0.002). Specifically, advancements >18.8mm were associated with significantly higher relapse rates (P<0.05). Almost one-third of patients with multisuture syndromic craniosynostosis underwent FO region readvancement. Advancements <17.2mm during initial FOA appeared to mitigate relapse, while advancing beyond 18.8mm may increase the risk. Investigation of additional protective factors against FO relapse is encouraged to minimize surgical burden.

Transplant in ALL: who, when, and how?

Allogeneic hematopoietic stem cell transplantation (allo-HCT) remains a cornerstone in the treatment of high-risk acute lymphoblastic leukemia (ALL), yet optimal patient selection is challenging in the era of rapidly changing modern therapy. Refined molecular characterization allows for better risk assessment, sparing low-risk patients from allo-HCT toxicity while identifying those who may benefit from intensified approaches. Measurable residual disease (MRD) has emerged as a powerful predictor of relapse irrespective of treatment strategy, challenging the necessity of transplant in MRD-negative patients. Further, expanded donor options, particularly haploidentical transplantation coupled with reduced intensity conditioning, have extended the applicability of allo-HCT to a broader range of patients. Finally, immunotherapies and targeted treatments are increasingly integrated into both initial and relapsed treatment protocols yielding deep remission and allowing for successful transplant in patients with a history of advanced disease. In this review, we provide an overview of the contemporary role of transplant in adult patients with ALL, focusing on indications for allo-HCT in first remission, optimal sequencing of transplant with novel therapies, and advancements in donor selection and conditioning regimens.

Mutations and MRD: clinical implications of clonal ontogeny.

Measurable residual disease (MRD) is a strong but imprecise predictor of relapse in acute myeloid leukemia. Many patients fall into the outlier categories of MRD positivity without relapse or MRD negativity with relapse. Why? We will discuss these states in the context of "clonal ontogeny" examining how mutations, clonal structure, and Darwinian rules impact response, resistance, and relapse.

Lack of Tissue Virological Response as a Predictor of Relapse in Chronic Hepatitis D after Completion Bulevirtide Therapy

Aim: to evaluate the significance of a positive polymerase chain reaction result for hepatitis D virus RNA (HDV RNA) in liver biopsy specimens of patients with chronic hepatitis D (CHD) after completion of antiviral therapy (AVT) as a predictor of infection relapse.Materials and methods. The study included 21 patients with CHD who received combined AVT with peginterferon alpha and bulevirtide for 48 weeks, followed by bulevirtide monotherapy for 48–96 weeks, making the total duration of antiviral therapy 96–144 weeks. In all patients HDV RNA became undetectable in serum 24–96 weeks from the start of treatment, with aviremia maintained for at least 48 weeks until the end of AVT. At the end of treatment, all patients underwent liver biopsy to detect HDV RNA in liver tissue.Results. Out of 21 patients with sustained complete virological response (negative polymerase chain reaction result for HDV RNA in serum), 8 (38 %) had HDV RNA detected in liver tissue, indicating that a tissue virological response was not achieved. All 8 patients experienced a relapse of CHD within 24 weeks after discontinuing AVT.Conclusions. In patients with chronic hepatitis D who have achieved a complete virological response in serum, the absence of a virological response in liver tissue (detection of HDV RNA in liver biopsy) is a predictor of relapse, providing a rationale for the continuation of antiviral therapy.

Predictors of early recurrence of atrial tachyarrhythmias after catheter ablation of atrial fibrillation

Aim. Search for predictors of early recurrence of atrial tachyarrhythmias after radiofrequency ablation (RFA) of atrial fibrillation (AF).Methods. The study included 57 subjects with persistent (n = 17; 30%) and paroxysmal (n = 40; 70%) forms of AF, admitted for the RFA. All patients underwent transthoracic echocardiography, assessment of deformation of both atria using 2D Strain, computed tomography (CT) with 3D reconstruction of the left atrium (LA). Intraoperatively, high-density voltage mapping of LA was performed before RF pulmonary vein isolation. All patients underwent follow-up after 3 months.Results. Recurrence of atrial tachyarrhythmia after 3 months was recorded in 17.5% of patients. High prevalence of low-amplitude activity zones in the LA and persistent AF were the strongest predictors. The LA reservoir function below 21.7%, the conduction function below 15.7%, the LA stiffness index above 0.314 relative units, the LA volume with the appendage above 121.7 ml, and the LA vertical size according to CT data above 65.5 mm statistically significantly predicted early recurrences of atrial tachyarrhythmias with high sensitivity and specificity.Conclusion. The decreased LA deformation in the reservoir and conductor phase, increased LA stiffness index, the prevalence of low-amplitude activity zones, vertical size and volume of the LA with an auricle according to CT data and persistent AF are significant predictors of early relapses after interventional treatment of AF.

Patterns of breast cancer locoregional relapse, metastasis, and subtypes in Ghana

BackgroundSignificant advances have been made in targeted therapeutics and systemic therapy regimens for breast cancer (BC) treatment over the past decade. Tumour cells can however remain in the body, leading to locoregional relapse and/or metastasis. Subtypes of BC have distinct prognostic effects and have been linked to varying risks of early locoregional relapse and metastases, response to treatment, and overall survival. Most Low- and middle-income countries (LMICs) have no registries of BC locoregional relapse and metastasis.MethodsThis study comprehensively reviewed, a 3-year retrospective single-centre data of female BC visiting the Komfo Anokye Teaching Hospital (KATH), Ghana to determine the prevalence of locoregional relapse and metastasis across our patient population. Prevalence of metastasis among the various BC subtypes was also determined.ResultsPrevalence of BC locoregional relapse and metastasis were 3.4% and 47.6% respectively. For BC patients with documented locoregional relapse (N = 36), 27.8% (CI = 15.8 − 44.0%) had relapse to the contralateral breast, 41.7% (CI = 27.1 – 57.8%) had relapse to the ipsilateral breast, and 30.6% (CI = 18.0 − 46.9%) had relapse to regional lymph nodes. For BC patients with documented metastasis (N = 503), 151 (30%) had multiple organs involvement, 141 (28%) had lung metastases, 80 (16%) had bone metastases, 45 (9%) had liver metastases, 16 (3%) had brain metastases and 70 (14%) had other metastases (ovary, uterus, spleen, peritoneum, or distant lymph nodes). Basal subtype was the most common subtype (n = 82, 41%), followed by Luminal A (n = 69, 34.5%), HER2+ (n = 37, 18.5%) and Luminal B (n = 12, 6%). Basal subtypes had the most metastasis (35%), with multiple metastasis being the most prevalent (13%).ConclusionClose to half of the patients (46%) presented with metastatic BC. BC subtypes could influence the specific metastatic site. The most common BC subtype was the Basal subtype and had the most metastases (35%), with multiple metastasis being the most prevalent (13%). These findings should serve as a guide in the management of patients to enhance early prediction and detection of locoregional relapse and metastasis for improved overall treatment outcomes.

The journey of MASLD: Tracking resolution, relapse, and predictive factors after sleeve gastrectomy and one-anastomosis gastric bypass, a propensity score-matched cohort study.

To assess the rates and predictors of resolution and relapse of metabolic-dysfunction associated steatotic liver disease (MASLD) in individuals undergoing sleeve gastrectomy (SG) or one-anastomosis gastric bypass (OAGB). This observational prospective cohort study involved 1618 propensity score-matched participants (81.5% female) with concurrent MASLD and obesity who underwent SG or OAGB between 2013 and 2023. In the context of a maximum follow-up of four years with a median follow-up of 2.2 years (IQR: 1.0-3.3), the overall rates of MASLD resolution and relapse were 71.1 per 1000 person-month and 8.7 per 1000 person-month, respectively. These rates were comparable between the SG and OAGB groups. Significant resolution predictors were a lower percentage of hepatic steatosis, a higher percentage of 12-month excess weight loss (EWL%), and younger age. In contrast, an increased percentage of liver steatosis, a higher pre-operative (Pre-Op) fat mass percentage (FM%), and older age were significant predictors of relapse. This study found no significant differences in MASLD resolution and relapse rates between SG and OAGB. Key factors influencing MASLD outcomes included the percentage of hepatic steatosis, 12-month EWL%, Pre-Op FM%, and age.

Identifying Key Prognostic Indicators for Relapse and Chronic Epilepsy in Autoimmune Encephalitis: Insights from a Multicenter Retrospective Study.

The aims of this study were to investigate clinical factors associated with encephalitis relapse and chronic epilepsy development, and to evaluate the effectiveness of immunotherapy on encephalitis relapse. Patients with autoimmune encephalitis diagnosed as positive for neuronal surface antibodies in five general hospitals were included. A minimum 12-month follow-up period was conducted, and binary logistic regression analysis was used to identify predictors of encephalitis relapse and chronic epilepsy development. Additionally, decision curve analysis (DCA) was employed to assess the clinical net benefit of predicting encephalitis relapse and chronic epilepsy. The study encompassed 65 patients with autoimmune encephalitis. The one-year relapse rate for encephalitis was 13.9%. The CASE score (P=0.045) was associated with encephalitis relapse, with subsequent immunotherapy proving beneficial in enhancing outcomes. Chronic epilepsy prevalence at one year was 26.2%, particularly higher among patients with positive LGI1 antibodies. Although adjustments in antiseizure medications were partially effective, 41.2% of patients developed drug-resistant epilepsy (DRE). DCA confirmed that the predictive models provided significant net clinical benefit in assessing the risk of encephalitis relapse and chronic epilepsy. Notably, the presence of diffuse cortical atrophy, medial temporal lobe atrophy, or cerebellar hemisphere atrophy was linked to relapsing encephalitis and chronic epilepsy. Most cases of autoimmune encephalitis are effectively managed, however, a minority of patients experience relapse or chronic epilepsy. The CASE score and LGI1 antibodies are independent risk factors for encephalitis relapse and chronic epilepsy development, respectively. Immunotherapy remains beneficial for relapsing patients, yet a portion may progress to DRE. Individuals with relapses and chronic epilepsy are predisposed to the development of cortical, temporal lobe, and cerebellar atrophy.

102 Transcriptomic prediction of psoriasis relapse based on resolved or non-lesional skin sites with or without ex vivo T cell activation

102 Transcriptomic prediction of psoriasis relapse based on resolved or non-lesional skin sites with or without ex vivo T cell activation

Effect of past extensive ulcers on fecal calprotectin in ulcerative colitis

BackgroundUlcerative colitis (UC) causes extensive ulceration attributable to intestinal inflammation. This study investigated the effect of past extensive ulcers (PEUs) on fecal calprotectin (FC).MethodsThis retrospective, single-center, observational study included patients with UC with a Mayo endoscopic subscore of 0. UC with scarring or pseudopolyposis was defined as PEU and FC and fecal immunochemical occult blood test (FIT) values were compared. The marker levels of patients in the PEU and non-PEU groups were examined to assess clinical relapse within 12 months.ResultsOf the 61 included patients, 27 had UC with PEUs and 34 had UC without PEUs. Albumin, hemoglobin, and FIT values between groups were not significantly different; however, the C-reactive protein and FC values of the PEU group were significantly higher than those of the non-PEU group. The FC values of the clinical relapse and remission groups within 12 months differed significantly. The cutoff values for the prediction of relapse within 12 months for all patients (area under the curve [AUC], 0.709; 95% confidence interval [CI], 0.512–0.907) and the non-PEU group (AUC, 0.893; 95% CI, 0.724–1.000) were both 118 mg/kg.ConclusionsUC patients with PEUs have higher FC values than those without PEUs, even in remission. Additionally, FC values do not predict relapse in PEU patients. Therefore, FC may not be a sufficiently accurate biomarker for patients with PEUs.

Development of Time-Aggregated Machine Learning Model for Relapse Prediction in Pediatric Crohn's Disease.

Pediatric Crohn's disease (CD) easily progresses to an active disease compared with adult CD, making it important to predict and minimize CD relapses. However, prediction of relapse at various time points (TPs) during pediatric CD remains understudied. We aimed to develop a real-time aggregated model to predict pediatric CD relapse in different TPs and time windows (TWs). This retrospective study was conducted on children diagnosed with CD between 2015 and 2022 at Severance Hospital. Laboratory test results and demographic data were collected starting at 3 months after diagnosis, and cohorts were formed using data from 6 different TPs at 1-month intervals. Relapse-defined as a pediatric CD activity index ≥ 30 points-was predicted, and TWs were 3-7 months with 1-month intervals. The feature importance of the variables in each setting was determined. Data from 180 patients were used to construct cohorts corresponding to the TPs. We identified the optimal TP and TW to reliably predict pediatric CD relapse with an area under the receiver operating characteristic curve score of 0.89 when predicting with a 3-month TW at a 3-month TP. Variables such as C-reactive protein levels and lymphocyte fraction were found to be important factors. We developed a time-aggregated model to predict pediatric CD relapse in multiple TPs and TWs. This model identified important variables that predicted relapse in pediatric CD to support real-time clinical decision making.