ACROBATIC Welcomes Two New Projects

ACROBATIC is pleased to welcome two new research projects that have recently become affiliated with the centre: STROMA (WP 2.28) and DaVulvaMob (WP 2.29).

Both projects strengthen ACROBATIC’s collaborative research network and address key challenges in cancer surgery, aiming to improve evidence-based decision-making and patient outcomes.

WP 2.28
WP leader: Charles Vesteghem, Center for Clinical Data Science (CLINDA), Aalborg University & Aalborg University Hospital, and Clinical Cancer Research Centre, Aalborg University Hospital, Denmark
Title: Solid Tumour Resectability and Operability Decision Support Using Multimodal AI (STROMA)

Abstract:
Complete surgical resection (R0) is crucial for improving survival and quality of life in patients with malignant solid tumours. However, preoperative assessment of resectability and operability remains challenging, despite advances in imaging and multidisciplinary team (MDT) evaluation. Incomplete resections significantly increase recurrence risk, while inaccurate operability assessments may deny patients curative treatment or expose them to surgical risk.
This project aims to develop and evaluate artificial intelligence (AI)-based decision-support tools combining imaging and electronic health registry data to improve preoperative prediction of resectability and operability in cancer patients. The primary objective is an AI model for colorectal cancer resectability using retrospective data. Secondary aims include assessing operability in colorectal cancer and extending models to ovarian cancer, sarcoma, breast cancer, and bladder cancer, with the goal of a generic predictive framework for solid tumours as well as conducting a prospective validation.
The study will use multimodal AI methods integrating structured data (e.g., laboratory results, comorbidities), unstructured data (clinical notes, reports), and imaging (CT, MRI, PET, ultrasound). Development will proceed from regional pilot cohorts to national datasets, applying transfer learning for cross-cancer applicability. Data will be processed within secure high-performance computing environments, with prospective validation planned in a randomised clinical design.
Leveraging Denmark’s comprehensive, population-wide health data, the project strives to enhance MDT decision-making, reduce unnecessary surgeries, improve surgical margins, and ultimately optimise patient outcomes. Successful implementation may support regulatory approval and provide a transferable methodology for AI-based surgical decision support across multiple cancer types.

WP 2.29
WP leader: Pernille Tine Jensen, Aarhus University & Aarhus University Hospital, Department of Gynaecology and Obstetrics, Denmark
Title: A National Randomized Controlled Trial Investigating the Impact of Mobilization Restrictions on Wound Complications and Quality of Life Following Vulvar Cancer Surgery – DaVulvaMob

Abstract:
Background: The standard treatment of vulvar cancer (VC) is up-front surgery with or without flap reconstruction and inguinofemoral lymph node dissection. Surgical treatment is associated with a high rate of postoperative morbidity; particularly surgical wound complications (SWCs). Current postoperative recommendations include restricted mobilization. However, immobilization entails risk of SWCs, thromboembolism, infection, and muscle atrophy. There is no evidence supporting restricted mobilization compared to early mobilization; in contrast, early mobilization is in general recommended after surgery according to Enhanced Recovery after Surgery (ERAS) programs.
Aim: This study, DaVulvaMob, evaluates the effect of non-restricted vs restricted mobilization on postoperative SWCs, severe complications, and quality of life following surgery for VC.
Methods: This is a randomised controlled non-inferiority trial. Women undergoing up-front surgery for VC (primary or relapse) are eligible. The women are randomly assigned to either group: unrestricted mobilization (intervention) vs restricted mobilization according to guidelines (standard). The primary endpoint is SWCs as assessed by a validated wound assessment tool. The study is a national multicentre study and includes patients from all regions of Denmark in close collaboration with home care services. A power calculation demands 95 patients in each group to show non-inferiority with a margin of 20%, a level of significance of 5% and power of 80%, and a drop-out rate of 20%.
Perspectives: DaVulvaMob contributes to the development of evidence-based recommendations for postoperative mobilization in VC patients. It contributes to the future transformation of the Danish healthcare system and results are expected to change clinical practice.