Abstract Background Emergency department (ED) crowding is a global challenge with adverse effects on patient outcomes and staff well-being. Traditional crowding scores, such as the National ED Overcrowding Study (NEDOCS) score, have limited ability to predict imminent crowding. The aim of this study is to develop and validate a real-time, artificial intelligence (AI)-driven prediction model using a long short-term memory (LSTM) neural network to predict ED crowding, using ambulance diversion status as the reference standard and comparing performance with the (modified) NEDOCS score. Methods In this single-centre, retrospective cohort study, we extracted parameters related to crowding from electronic health records and workforce data. Ambulance diversion status served as the reference standard for crowding. The LSTM model was trained and evaluated with walk-forward validation and performance was assessed using the area under the receiver operating characteristic curve (AUC). Model calibration was assessed using Brier score and expected calibration error (ECE). Results The optimal LSTM model used the previous hour of data to predict one hour ahead, yielding an AUC of 0.84 (95% CI 0.81–0.86). Key features contributing to the predictive quality of our model included the number of pre-notified arrivals, patients in the yellow and orange triage categories, and the number of ED supervisors. Compared with the modified NEDOCS score (AUC of 0.73 (95% CI 0.70–0.76)), the LSTM model showed higher discriminative performance in internal validation within this single-centre study. Conclusions Our AI-driven prediction model demonstrated good predictive performance for ED crowding in internal validation. These findings highlight the potential of real-time prediction models to support ED clinical workflow.
The R2 (lenalidomide-rituximab) and R2 with bendamustine (R2B) regimens are both feasible in relapsed/refractory follicular lymphoma (R/R FL), but prospective phase II data on R2B are lacking. In this multinational, prospective, randomised, non-comparative trial, patients with R/R FL were randomised 1:1 to R2 (arm A) or R2B (arm B). Patients achieving CT-based partial or complete remission (PR/CR) received rituximab maintenance every 3 months for 2 years. Co-primary endpoints were investigator-assessed CR rates at end-of-induction (EOI) and severe toxicity rates. Between December 2014 and July 2019, 92 patients were randomized. The trial was stopped prematurely due to slow accrual. CR rates at EOI were 11.4% (95% CI [3.8, 24.6%]) for R2 and 15.2% (CI [6.3, 28.9%]) for R2B. With a median follow-up of 74 months, predefined severe toxicities occurred in 3 (6.8%; CI [1.4, 18.7%]) and 6 pts (13.0%, CI [4.9, 26.3%]) in Arm A and Arm B respectively, with 2 pneumonia-related deaths in arm A and no treatment-related deaths in arm B; 43% and 66% experienced any grade 3-4 AE (2%/7% grade 4). At 60 months, in arms A and B event free survival (EFS) was 39.5% (CI [25.1, 53.6%]) and 56.4% (CI [40.1, 69.3%]). Overall survival (OS) was excellent at 72.1% and 86.3%, respectively. R2 and R2B are both effective treatment regimens for R/R FL. While CT-based CR rates were low, treatment with R2B resulted in numerically longer median EFS and OS, but this was associated with higher toxicity.
LBA3506 Background: Recurrence after local treatment for colorectal liver metastases (CLM) occurs in up to 70% of patients, frequently confined to the liver. Dutch guidelines do not recommend adjuvant chemotherapy, because three randomized controlled trials (RCTs) have shown no overall survival (OS) benefit of perioperative systemic chemotherapy in resectable CLM. Hepatic arterial infusion pump (HAIP) chemotherapy delivers high doses of floxuridine directly to the liver. This trial evaluated the effectiveness of adjuvant HAIP chemotherapy with floxuridine compared to resection alone in patients with resectable CLM and a low clinical risk score (CRS). Methods: This is an open-label, investigator-initiated, multicenter, randomized phase III trial. Adult patients with resectable CLM, no extrahepatic disease, and CRS 0–2 were randomized 1:1 to resection plus adjuvant HAIP chemotherapy with floxuridine versus resection alone, both without adjuvant systemic chemotherapy. Preoperative systemic chemotherapy prior to randomization was allowed. Patients in both groups who signed informed consent but did not fulfill inclusion criteria at the time of surgery were excluded and replaced. Patients were scheduled for 6 cycles of HAIP chemotherapy with floxuridine (0.12 mg/kg/day) that was initiated 4–12 weeks after placement of a Tricumed constant flow pump. The primary endpoint was progression-free survival (PFS) calculated from the date of surgery to the date of a recurrence or death. Secondary endpoints included hepatic PFS (hPFS) and ninety-day mortality. Survival was estimated using Kaplan-Meier method and compared using a log-rank test. Results: Between August 2018 and March 2026, 243 patients were randomized to resection followed by adjuvant HAIP (n=120) or resection alone (n=123). At time of surgery, 25 patients were excluded due to presence of extrahepatic disease, unresectable CLM or histopathological confirmation of benign disease. In this analyses, 110 patients were included in the resection and adjuvant HAIP group and 108 patients in the resection alone group. In the HAIP group, 100 patients (91%) initiated HAIP chemotherapy, and the median number of administered cycles was 5 [IQR 3-6]. Treatment was discontinued in 8 patients (8%) due to recurrence and in 32 patients (32%) due to toxicity. The median PFS was 15.0 months in HAIP group vs 16.0 months in resection alone group (HR 0.88; 95% CI 0.62–1.25; p=0.48). The median hPFS was 37.6 months in HAIP group vs 21.8 months in resection alone group (HR 0.80; 95% CI 0.54–1.17; p=0.25). Ninety-day postoperative mortality was observed in 4 patients (3.3%) in the HAIP group and in 1 patient (0.8%) after resection alone. No mortality was attributed to pump placement or HAIP chemotherapy. Conclusion: In patients with resectable CLM and a low CRS, no improvement in PFS after adjuvant HAIP chemotherapy with floxuridine compared to resection alone could be demonstrated. Mature results for overall survival are expected in 2029. EudraCT number: 2018-001696-21. Clinical trial information: 2018-001696-21.
BACKGROUND:Over recent decades, there has been a more pronounced shift towards earlier therapeutic intervention in luminal Crohn's disease (CD). However, it remains unclear if similar strategies apply to active perianal fistulizing CD (pCD). This study explores the patient journey by assessing pCD management across diagnostic time periods in a large national cohort. METHODS:Patients with active pCD were included in a multicentre cohort across 41 Dutch academic and non-academic hospitals. Diagnostic modalities, medical and surgical treatments were evaluated over the pCD disease course, particularly within the first year after pCD diagnosis, and compared across four time periods: < 2010, 2010-2015, 2015-2020 and > 2020. RESULTS:Four hundred and forty patients were included (53.0% female; median pCD disease course 4.1 years). 12% were diagnosed before 2010, 11% in 2010-2015, 31% in 2015-2020, 46% after 2020. Over time, imaging (MRI or EUS) within the first year after pCD diagnosis was used in 37% of cases diagnosed before 2010 to 66%, 83% and 95% in subsequent periods. Endoscopic evaluation for proctitis was performed in 23%, 40%, 57% and 60%. First-line anti-TNF use was initiated in 27% followed by 58%, 68% and 69%. A seton was placed in 27% initially and in 42%, 51% and 52% in later years. Fistula surgery aimed at closure was performed in 14%, 4%, 8% and 19% over time. CONCLUSIONS:This national cohort illustrates substantial progress in early pCD management over time. These shifts likely reflect growing clinical awareness of the importance of early and multidisciplinary interventions. However, surgical procedures aimed at fistula closure remain infrequently performed in the early disease course, which may be related to the inclusion of a predominantly therapy-refractory patient group.
