Dynamic surgical anatomy using 3D reconstruction technology in complex hepato-biliary surgery with vascular involvement. Results from an international multicentric survey.

INTRODUCTION: Three-dimensional liver modeling can lead to substantial changes in choosing the type and extension of liver resection. This study aimed to explore whether 3D reconstruction helps to better understand the relationship between liver tumors and neighboring vascular structures compared to standard 2D CT scan images. METHODS: Contrast-enhanced CT scan images of 11 patients suffering from primary and secondary hepatic tumors were selected. Twenty-three experienced HBP surgeons participated to the survey. A standardized questionnaire outlining 16 different vascular structures (items) having a potential relationship with the tumor was provided. Intraoperative and histopathological findings were used as the reference standard. The proper hypothesis was that 3D accuracy is greater than 2D. As a secondary endpoint, inter-raters' agreement was explored. RESULTS: The mean difference between 3D and 2D, was 2.6 points (SE: 0.40; 95 % CI: 1.7-3.5; p < 0.0001). After sensitivity analysis, the results favored 3D visualization as well (mean difference 1.7 points; SE: 0.32; 95 % CI: 1.0-2.5; p = 0.0004). The inter-raters' agreement was moderate for both methods (2D: W = 0.45; 3D: W = 0.44). CONCLUSION: 3D reconstruction may give a significant contribution to better understanding liver vascular anatomy and the precise relationship between the tumor and the neighboring structures.

Compliance to Adjuvant Chemotherapy of Patients Who Underwent Surgery for Rectal Cancer: Report from a Multi-institutional Research Network.

INTRODUCTION: Adjuvant chemotherapy for locally advanced rectal cancer is associated with improved overall survival. However, recent evidence from randomized trials showed a compliance rate of 43 to 73%, which may affect efficacy. The aim of this multicenter retrospective analysis was to investigate the compliance rate to adjuvant treatment for patients who underwent rectal surgery for cancer. METHODS: Patients who underwent surgery with curative intent for rectal cancer in six Italian colorectal centers between January 2013 and December 2017 were retrospectively reviewed. Exclusion criteria were age less than 18 years, palliative or emergency surgery, and stage IV disease. Parameters of interest were patients' characteristics, preoperative tumor stage, neo-adjuvant chemoradiation therapy, intra-operative and postoperative outcomes. Although the participating centers referred to the same treatment guidelines for treatment, the chemotherapy regiment was not standardized across the institutions. Reasons for not starting adjuvant chemotherapy when indicated, interruption, and modification of drug regimen were collected to investigate compliance. RESULTS: A total of 572 patients were included in the analysis. Two hundred and fifty-two (44.1%) patients received neo-adjuvant chemoradiation therapy. All patients underwent high anterior rectal resection, low anterior rectal resection, or Miles' procedure. Of 399 patients with an indication to adjuvant chemotherapy, 176 (44.1%) completed the treatment as planned. Compliance for patients who started chemotherapy was 56% (95% CI 50.4-61.6%). Sixty-six patients interrupted the treatment, 76 patients significantly reduced the drug dose, and 41 patients had to switch to other therapeutic regimens. CONCLUSIONS: The present multicenter investigation reports a low compliance rate to adjuvant chemotherapy after rectal resection for cancer. Multidisciplinary teams should focus on future effort to improve compliance for these patients.

Main challenges on the curation of large scale datasets for pancreas segmentation using deep learning in multi-phase CT scans: Focus on cardinality, manual refinement, and annotation quality.

Accurate segmentation of the pancreas in computed tomography (CT) holds paramount importance in diagnostics, surgical planning, and interventions. Recent studies have proposed supervised deep-learning models for segmentation, but their efficacy relies on the quality and quantity of the training data. Most of such works employed small-scale public datasets, without proving the efficacy of generalization to external datasets. This study explored the optimization of pancreas segmentation accuracy by pinpointing the ideal dataset size, understanding resource implications, examining manual refinement impact, and assessing the influence of anatomical subregions. We present the AIMS-1300 dataset encompassing 1,300 CT scans. Its manual annotation by medical experts required 938 h. A 2.5D UNet was implemented to assess the impact of training sample size on segmentation accuracy by partitioning the original AIMS-1300 dataset into 11 smaller subsets of progressively increasing numerosity. The findings revealed that training sets exceeding 440 CTs did not lead to better segmentation performance. In contrast, nnU-Net and UNet with Attention Gate reached a plateau for 585 CTs. Tests on generalization on the publicly available AMOS-CT dataset confirmed this outcome. As the size of the partition of the AIMS-1300 training set increases, the number of error slices decreases, reaching a minimum with 730 and 440 CTs, for AIMS-1300 and AMOS-CT datasets, respectively. Segmentation metrics on the AIMS-1300 and AMOS-CT datasets improved more on the head than the body and tail of the pancreas as the dataset size increased. By carefully considering the task and the characteristics of the available data, researchers can develop deep learning models without sacrificing performance even with limited data. This could accelerate developing and deploying artificial intelligence tools for pancreas surgery and other surgical data science applications.

Creation of a rectal cancer registry in Italy by the Advanced International Mini-Invasive Surgery (AIMS) academy clinical research network.

Background: The management of rectal cancer is multimodal and involves a multidisciplinary team of cancer specialists with expertise in medical oncology, surgical oncology, radiation oncology and radiology. It is crucial for highly specialized centers to collaborate via networks that aim to maintain uniformity in every aspect of treatment and rigorously gather patients' data, from the first clinical evaluation to the last follow-up visit. The Advanced International Mini-Invasive Surgery (AIMS) academy clinical research network aims to create a rectal cancer registry. This will prospectively collect the data of patients operated on for non-metastatic rectal cancer in high volume colorectal surgical units through a well design pre-fashioned database for non-metastatic rectal cancer, in order to take all multidisciplinary aspects into consideration. Methods/Design: The protocol describes a multicenter prospective observational cohort study, investigating demographics, frailty, cancer-related features, surgical and radiological parameters, and oncological outcomes among patients with non-metastatic rectal cancer who are candidates for surgery with curative intent. Patients enrolled in the present registry will be followed up for 5 years after surgery. Discussion: Standardization and centralization of data collection for neoplastic diseases is a virtuous process for patient care. The creation of a register will allow the control of the quality of treatments provided and permit prospective and retrospective studies to be carried out on complete and reliable high quality data. Establishing data collection in a prospective and systematic fashion is the only possibility to preserve the enormous resource that each patient represents.