Background: Performing minimally invasive surgery can be technically challenging. In addition to its inherent difficulty, other factors can contribute into making cases particularly difficult. For instance, patient characteristics such as a narrow pelvis, a high BMI and a low tumour can pose an additional challenge in low anterior resections. As technical difficulty is associated with immediate oncological results and patient outcomes, it is important to explore novel methods to prepare for challenging cases, taking into account the individual patient and disease characteristics. The aim of the current project is to develop and test case specific rehearsal methods, establishing the feasibility of their application in a real clinical environment.
Methods: Patient specific virtual and physical (i.e. synthetic) anatomical models were developed using 3D reconstruction and modelling, based on MRI and CT images of patients. These were then combined with mental practice and tested in a simulated (two studies) and a clinical environment (one study). The first study compared MP to MP with virtual 3D models and to a control group; the second study compared MP to MP with 3D visual aids after a significant degree of anatomical variation was introduced; and the clinical trial compared MP with the use of three different aids (Virtual, physical models – including simulation and MRI) to routine clinical practice (control group).
Results: The first study showed performance differences across groups, with the control group performing worse (time to complete LC (F(2,17) = 8.77, p = .002, ηp² = .51), Control group: Median (M) = 1447sec, SD = 341sec) 3D & MP group (M = 670sec, SD = 326sec) (p = .002)). The second study showed equal performance when the anatomy was “normal” [MP vs. MP and 3D Model Total CAT score – NA: 23.63 vs. 26.69 p=0.2 – SCD: 20.5 vs. 26.31 p=0.02 2=0.32 – DA: 24.75 vs. 30.5 p=0.03 2=0.28] but superior performance for the MP and 3D model group for complex anatomy. Although the clinical trial showed no difference in overall performance (Median control: 30.5, MRI: 34.25, virtual: 31.75, physical: 34, p = 0.75, 2 <0.01), the time spent not performing dissection (“nothing” time) was significantly shorter for the SMR with MRI group compared to the control group (57.5 vs. 42min, p < 0.001, 2 = 0.212).
