Aim: Proper identification in real time of different types of tissues during intraoperative procedures represents a vital and challenging task. This paper addresses tissue segmentation in two different medical applications using hyperspectral imaging (HSI) and machine learning in two main steps.
Methods: The first step consists of data preprocessing designed to overcome the most common problems linked with HSI, involving inter- and intra-patient variability of the tissue spectra and the high dimensionality of the spectra. The preprocessing step involves outlier removal, spectral smoothing, normalization, and dimensionality reduction using principal component analysis applied in the spectral domain of HSI data. In the spatial domain, multiple levels of analysis are performed using Gaussian filters. The second step consists of tissue segmentation using an optimized machine learning model. The most suitable model was selected under statistical comparison of seven machine learning models involving three different levels of spatial analysis.
Results: According to the experimental results, the U-Net achieves the highest precision (0.908) for detection of liver, bile duct, artery, and portal vein tissues in a set of 18 HSI data, while the logistic regression with the elasticnet regularization combined with multiscale spatial analysis obtains the highest F1-score (0.673) and segmentation accuracy (0.803) for thyroid and parathyroid glands segmentation in a set of 21 HSI data.
Conclusion: In addition to the computational experiments, combining machine learning with HSI represents a promising approach to perform image-guided surgery.
Aim: Some surgeons have been using some form of handheld robotics (HR) since liver resections began being done minimally invasively (MI); however, with the development of the complete robotic surgical systems (CRSS), they have lived in limbo neither being truly laparoscopic nor robotic. While doing the Study: International and Multi-centered on Minimally Invasive Liver Resections for colorectal metastases (CRLM), we decided to evaluate these two different degrees of robotics, specifically the HR group and the group undergoing a completely robotic (CR) approach.
Methods: Four international centers (one in France, one in Germany, one in Taiwan, and one in the United States of America) were asked to join a retrospective review of cases to compare short- and long-term outcomes after open, laparoscopic, and robotic liver resection for CRLMs. For this study, only patients who had either HR or CR liver resections were included. HR was defined as cases done with a robotically controlled camera holder and a powered stapling device. Only patients with ≤ 3 tumors that were ≤ 5 cm were included so that the preoperative characteristics of the two groups would be similar.
Results: In total, three centers did CR for CRLM (28 patients) and one center used HR (49 patients). MI resections were possible in 92.5% of patients when HR was used compared to 34.2% (22.6% laparoscopic, 11.5% CR) when centers used CRSS. Mean operating room times were significantly longer after CR compared to HR resections,
Conclusion: HR and CR liver resections have similar short- and long-term outcomes; however, when HR is used, over 90% of cases can be done MI compared to under 35% with CR. The added benefit of haptics and the ability for the operating surgeon to use hand assistance may account for this discrepancy. CRSS do not use haptics and surgeons must rely on visual cues; as robots become more autonomous, it may make more sense for computer engineers to work on the robot perceiving feedback and not the surgeon. HR may be the safest way to develop more autonomous actions in surgery and may yield the most benefits for patients by keeping the surgeon in the loop.