Introduction

The Ki-67 labeling index is a well-established marker of cellular proliferation and has been associated with worse disease-free and overall survival in gastrointestinal stromal tumors (GISTs). Recent evidence highlights its potential role in risk stratification, particularly when Ki-67 exceeds certain thresholds. However, current assessment relies on postoperative histopathology. Radiomics, by extracting quantitative features from computer tomography (CT) images, may offer a non-invasive method to predict Ki-67 expression preoperatively. This meta-analysis aimed to evaluate the diagnostic performance of CT-based radiomics for predicting Ki-67 status in GISTs.

Aims & Methods

A systematic review and meta-analysis were conducted according to PRISMA guidelines). Four electronic databases (PubMed, Scopus, Science Direct, Cochrane Library) were searched for relevant articles up to December 2024. Eligible studies assessed CT radiomics in predicting Ki-67 in histologically confirmed GISTs. Pooled sensitivity, specificity, and area under the curve (AUC) were calculated using a random-effects model. Subgroup analyses were performed based on cohort type, CT protocol, number of features, and Ki-67 cutoff values. Study quality was assessed using the QUADAS-2 tool.

Results

Six studies comprising 1,632 patients met the inclusion criteria. Pooled sensitivity and specificity were 0.71 (95% CI: 0.63–0.79) and 0.76 (95% CI: 0.73–0.78), respectively. The pooled AUC was 0.79 (95% CI: 0.74–0.84). Subgroup analyses showed consistent performance across imaging protocols and feature counts, while Ki-67 cutoffs (8% vs. 10%) significantly impacted diagnostic accuracy. Sensitivity analysis confirmed robustness of specificity estimates. Publication bias was detected for specificity (Egger’s test p = 0.006).

Conclusion

CT radiomics demonstrates moderate accuracy for non-invasively predicting Ki-67 expression in GISTs. Given Ki-67's prognostic significance, radiomics could support preoperative risk assessment and personalized treatment planning. Further prospective validation is warranted.

Disclosure

No Conflict of Interest ​

Introduction

Filgotinib, a selective JAK1 inhibitor, has emerged as a promising therapy for moderate-to-severe ulcerative colitis (UC). However, real-world data, especially in multi-refractory patients, remain scarce. The Filguito registry, conducted across 11 Andalusian centers, provides a comprehensive assessment of filgotinib’s effectiveness and safety over 12 months in routine clinical practice.

Aims & Methods

This ambispective, multicenter study included 77 UC patients treated with filgotinib. Clinical and biochemical parameters were collected at baseline, weeks 8 and 16, and months 6 and 12. Clinical remission (CR) was defined as a partial Mayo score <3; combined clinical-biochemical remission (CBR) as partial Mayo <3 plus fecal calprotectin <250 μg/g; and steroid-free remission (SFR) as partial Mayo <3 without corticosteroids since week 8. Effectiveness outcomes were stratified by prior exposure to advanced therapies.

Results

The study population had a mean age of 39.5 years and a mean disease duration of 13.5 years. Most patients (62%) presented with extensive colitis, and the majority had a history of multiple failed advanced therapies (mean 2.1). At baseline, disease activity was high, with a mean partial Mayo score of 5.6 and mean fecal calprotectin of 2,325.5 μg/g. Only a minority were corticosteroid-dependent at initiation.
Filgotinib induced rapid and sustained reductions in disease activity, with the partial Mayo score decreasing from 5.6 at baseline to 1.2 at 12 months (p<0.001). Fecal calprotectin levels fell by 76% (2,325.5 to 548.2 μg/g, p=0.006), and CRP declined significantly by year-end (12.4 to 3.1 mg/L, p=0.022). Clinical remission rates rose sharply to 56.9% by week 8, stabilizing at 53% by 12 months. Composite and steroid-free remission peaked at 32.8% and 45.8%, respectively, during follow-up (Table 1).
Differential Response by Prior Treatment Exposure
Patients with ≥2 prior advanced therapy failures showed markedly lower effectiveness:

• 12-month clinical remission: 32% vs. 58% in those with 1 prior failure (p=0.029).
• Steroid-free remission: 28% vs. 64% (p=0.014).
• Composite remission: <25% across subgroups, with no significant intergroup differences.

By the end of follow-up, 29.8% of patients had discontinued filgotinib, mainly due to lack of primary or secondary response. Only two patients stopped treatment because of adverse events (fever), and one due to pregnancy. No serious adverse events were reported, confirming a favorable safety profile in this real-world cohort.

Table 1: Effectiveness Outcomes

Timepoint	Baseline (n=77)	8 Weeks (n=70)	16 Weeks (n=49)	6 Months (n=46)	12 Months (n=29)	p¹
Partial Mayo Index (mean)	5.6	2.6 (<0.001)²	1.9 (<0.001)²	1.8 (<0.001)²	1.2 (<0.001)²	<0.001
CRP (mg/L, mean)	12.4	6.9 (0.567)²	7.2 (0.567)²	4.8 (0.079)²	3.1 (0.022)²	0.019
Calprotectin (μg/g, mean)	2325.5	1941.0 (0.628)²	1083.2 (0.036)²	813.7 (0.218)²	548.2 (0.005)²	0.006
Clinical Remission (%)	6.5%	56.9% (<0.001)³	54.1% (<0.001)³	54.1% (<0.001)³	53.1% (<0.001)³	-
Clinical-Biochemical Remission (%)	1.3%	25.0% (<0.001)³	16.4% (0.004)³	32.8% (<0.001)³	22.4% (0.009)³	-
Steroid-Free Remission (%)	-	45.8%	37.7% (0.333)³	42.6% (0.014)³	38.8% (0.002)³	-
Treatment Discontinuation	-	11/77 (14.2%)	14/77 (18.1%)	20/77 (25.9%)	23/77 (29.8%)	-

¹ Friedman test for related samples (two-way ANOVA by ranks).
² Pairwise comparisons using the Friedman test.
³ McNemar test for related samples.

Conclusion

In this real-world, multicenter Andalusian cohort of patients with refractory UC, filgotinib demonstrated sustained clinical and biochemical effectiveness over one year, with a favorable safety profile. Although response rates were lower in patients with multiple prior advanced therapy failures, clinically meaningful improvements were observed even in this challenging population. These results support filgotinib as an effective and safe therapeutic option for UC in routine clinical practice