Keyframe extraction from laparoscopic videos based on visual saliency detection

Artificial Intelligence in the Management of Women with Endometriosis and Adenomyosis: Can Machines Ever Be Worse Than Humans?

Artificial intelligence (AI) is experiencing advances and integration in all medical specializations, and this creates excitement but also concerns. This narrative review aims to critically assess the state of the art of AI in the field of endometriosis and adenomyosis. By enabling automation, AI may speed up some routine tasks, decreasing gynecologists' risk of burnout, as well as enabling them to spend more time interacting with their patients, increasing their efficiency and patients' perception of being taken care of. Surgery may also benefit from AI, especially through its integration with robotic surgery systems. This may improve the detection of anatomical structures and enhance surgical outcomes by combining intra-operative findings with pre-operative imaging. Not only that, but AI promises to improve the quality of care by facilitating clinical research. Through the introduction of decision-support tools, it can enhance diagnostic assessment; it can also predict treatment effectiveness and side effects, as well as reproductive prognosis and cancer risk. However, concerns exist regarding the fact that good quality data used in tool development and compliance with data sharing guidelines are crucial. Also, professionals are worried AI may render certain specialists obsolete. This said, AI is more likely to become a well-liked team member rather than a usurper.

It's time for a minimum synoptic operation template in patients undergoing laparoscopic cholecystectomy: a systematic review

BACKGROUND

Despite the call to enhance accuracy and value of operation records few international recommended minimal standards for operative notes documentation have been described. This study undertook a systematic review of existing operative reporting systems for laparoscopic cholecystectomy (LC) to fashion a comprehensive, synoptic operative reporting template for the future.

METHODS

A search for all relevant articles was conducted using PubMed version of Medline, Scopus and Web of Science databases in June 2021, for publications from January 1st 2011 to October 25th 2021, using the keywords: laparoscopic cholecystectomy AND operation notes OR operative notes OR proforma OR documentation OR report OR narrative OR audio-visual OR synoptic OR digital. Two reviewers (NOC, GMC) independently assessed each published study using a MINORS score of ≥ 16 for comparative and ≥ 10 for non-comparative for inclusion. This systematic review followed PRISMA guidelines and was registered with PROSPERO. Synoptic operative templates from published data were assimilated into one "ideal" laparoscopic operative report template following international input from the World Society of Emergency Surgery board.

RESULTS

A total of 3567 articles were reviewed. Following MINORS grading 25 studies were selected spanning 14 countries and 4 continents. Twenty-two studies were prospective. A holistic overview of the operative procedure documentation was reported in 6/25 studies and a further 19 papers dealt with selective surgical aspects of LC. A unique synoptic LC operative reporting template was developed and translated into Chinese/Mandarin, French and Arabic.

CONCLUSION

This systematic review identified a paucity of publications dealing with operative reporting of LC. The proposed new template may be integrated digitally with hospitals' medical systems and include additional narrative text and audio-visual data. The template may help define new OR (operating room) recording standards and impact on care for patients undergoing LC.

Keyframe Extraction From Laparoscopic Videos via Diverse and Weighted Dictionary Selection

Laparoscopic videos have been increasingly acquired for various purposes including surgical training and quality assurance, due to the wide adoption of laparoscopy in minimally invasive surgeries. However, it is very time consuming to view a large amount of laparoscopic videos, which prevents the values of laparoscopic video archives from being well exploited. In this paper, a dictionary selection based video summarization method is proposed to effectively extract keyframes for fast access of laparoscopic videos. Firstly, unlike the low-level feature used in most existing summarization methods, deep features are extracted from a convolutional neural network to effectively represent video frames. Secondly, based on such a deep representation, laparoscopic video summarization is formulated as a diverse and weighted dictionary selection model, in which image quality is taken into account to select high quality keyframes, and a diversity regularization term is added to reduce redundancy among the selected keyframes. Finally, an iterative algorithm with a rapid convergence rate is designed for model optimization, and the convergence of the proposed method is also analyzed. Experimental results on a recently released laparoscopic dataset demonstrate the clear superiority of the proposed methods. The proposed method can facilitate the access of key information in surgeries, training of junior clinicians, explanations to patients, and archive of case files.

Collateral motion saliency-based model for Trypanosoma cruzi detection in dye-free blood microscopy

The motion performed by some protozoa is a crucial visual stimulus in microscopy analysis, especially when they have almost imperceptible morphological characteristics. Microorganisms can be distinguished through the interactions of their locomotion with neighboring elements, as observed in some parasitological analysis of Trypanosoma cruzi. In dye-free blood microscopy, the low contrast of this parasite makes it difficult to detect them. Thus, the parasite's interaction with the neighborhood, such as collisions with blood cells and shocks during the escape of confinements in cell clumps, generates collateral motions that assist its detection. Assuming that the collateral motion of the parasite can be sufficiently noticeable to overcome the dynamic contexts of inspection, we propose a novel computational approach that is based on motion saliency. We estimate motion in microscopy videos using dense optical flow and we investigate vestiges in saliency maps that could characterize the collateral motion of parasites. Our biological-inspired method shows that the parasite's collateral motion is a relevant feature for T. cruzi detection. Therefore, our computational model is a promising aid in the research and medical diagnosis of Chagas disease.

Patch-based classification of gallbladder wall vascularity from laparoscopic images using deep learning

PURPOSE

In this study, we propose a deep learning approach for assessment of gallbladder (GB) wall vascularity from images of laparoscopic cholecystectomy (LC). Difficulty in the visualization of GB wall vessels may be the result of fatty infiltration or increased thickening of the GB wall, potentially as a result of cholecystitis or other diseases.

METHODS

The dataset included 800 patches and 181 region outlines of the GB wall extracted from 53 operations of the Cholec80 video collection. The GB regions and patches were annotated by two expert surgeons using two labeling schemes: 3 classes (low, medium and high vascularity) and 2 classes (low vs. high). Two convolutional neural network (CNN) architectures were investigated. Preprocessing (vessel enhancement) and post-processing (late fusion of CNN output) techniques were applied.

RESULTS

The best model yielded accuracy 94.48% and 83.77% for patch classification into 2 and 3 classes, respectively. For the GB wall regions, the best model yielded accuracy 91.16% (2 classes) and 80.66% (3 classes). The inter-observer agreement was 91.71% (2 classes) and 78.45% (3 classes). Late fusion analysis allowed the computation of spatial probability maps, which provided a visual representation of the probability for each vascularity class across the GB wall region.

CONCLUSIONS

This study is the first significant step forward to assess the vascularity of the GB wall from intraoperative images based on computer vision and deep learning techniques. The classification performance of the CNNs was comparable to the agreement of two expert surgeons. The approach may be used for various applications such as for classification of LC operations and context-aware assistance in surgical education and practice.

Saliency-driven system models for cell analysis with deep learning

BACKGROUND AND OBJECTIVES

Saliency refers to the visual perception quality that makes objects in a scene to stand out from others and attract attention. While computational saliency models can simulate the expert's visual attention, there is little evidence about how these models perform when used to predict the cytopathologist's eye fixations. Saliency models may be the key to instrumenting fast object detection on large Pap smear slides under real noisy conditions, artifacts, and cell occlusions. This paper describes how our computational schemes retrieve regions of interest (ROI) of clinical relevance using visual attention models. We also compare the performance of different computed saliency models as part of cell screening tasks, aiming to design a computer-aided diagnosis systems that supports cytopathologists.

METHOD

We record eye fixation maps from cytopathologists at work, and compare with 13 different saliency prediction algorithms, including deep learning. We develop cell-specific convolutional neural networks (CNN) to investigate the impact of bottom-up and top-down factors on saliency prediction from real routine exams. By combining the eye tracking data from pathologists with computed saliency models, we assess algorithms reliability in identifying clinically relevant cells.

RESULTS

The proposed cell-specific CNN model outperforms all other saliency prediction methods, particularly regarding the number of false positives. Our algorithm also detects the most clinically relevant cells, which are among the three top salient regions, with accuracy above 98% for all diseases, except carcinoma (87%). Bottom-up methods performed satisfactorily, with saliency maps that enabled ROI detection above 75% for carcinoma and 86% for other pathologies.

CONCLUSIONS

ROIs extraction using our saliency prediction methods enabled ranking the most relevant clinical areas within the image, a viable data reduction strategy to guide automatic analyses of Pap smear slides. Top-down factors for saliency prediction on cell images increases the accuracy of the estimated maps while bottom-up algorithms proved to be useful for predicting the cytopathologist's eye fixations depending on parameters, such as the number of false positive and negative. Our contributions are: comparison among 13 state-of-the-art saliency models to cytopathologists' visual attention and deliver a method that the associate the most conspicuous regions to clinically relevant cells.