Can artificial intelligence help decision-making in arthroscopy? Part 2: The IA-RTRHO model - a decision-making aid for long head of the biceps diagnoses in small rotator cuff tears.

INTRODUCTION: The possible applications of artificial intelligence (AI) in orthopedic surgery are promising. Deep learning can be utilized in arthroscopic surgery due to the video signal used by computer vision. The intraoperative management of the long head of biceps (LHB) tendon is the subject of a long-standing controversy. The main objective of this study was to model a diagnostic AI capable of determining the healthy or pathological state of the LHB on arthroscopic images. The secondary objective was to create a second diagnostic AI model based on arthroscopic images and the medical, clinical and imaging data of each patient, to determine the healthy or pathological state of the LHB. HYPOTHESIS: The hypothesis of this study was that it was possible to construct an AI model from operative arthroscopic images to aid in the diagnosis of the healthy or pathological state of the LHB, and its analysis would be superior to a human analysis. MATERIALS AND METHODS: Prospective clinical and imaging data from 199 patients were collected and associated with images from a validated protocoled arthroscopic video analysis, called "ground truth", made by the operating surgeon. A model based on a convolutional neural network (CNN) modeled via transfer learning on the Inception V3 model was built for the analysis of arthroscopic images. This model was then coupled to MultiLayer Perceptron (MLP), integrating clinical and imaging data. Each model was trained and tested using supervised learning. RESULTS: The accuracy of the CNN in diagnosing the healthy or pathological state of the LHB was 93.7% in learning and 80.66% in generalization. Coupled with the clinical data of each patient, the accuracy of the model assembling the CNN and MLP were respectively 77% and 58% in learning and in generalization. CONCLUSION: The AI model built from a CNN manages to determine the healthy or pathological state of the LHB with an accuracy rate of 80.66%. An increase in input data to limit overfitting, and the automation of the detection phase by a Mask-R-CNN are ways of improving the model. This study is the first to assess the ability of an AI to analyze arthroscopic images, and its results need to be confirmed by further studies on this subject. LEVEL OF EVIDENCE: III Diagnostic study.

Can artificial intelligence help decision-making in arthroscopy? Part 1: Use of a standardized analysis protocol improves inter-observer agreement of arthroscopic diagnostic assessments of the long head of biceps tendon in small rotator cuff tears.

INTRODUCTION: Injuries of the long head of biceps (LHB) tendon are common but difficult to diagnose clinically or using imaging. Arthroscopy is the preferred means of diagnostic assessment of the LHB, but it often proves challenging. Its reliability and reproducibility have not yet been assessed. Artificial intelligence (AI) could assist in the arthroscopic analysis of the LHB. The main objective of this study was to evaluate the inter-observer agreement for the specific LHB assessment, according to an analysis protocol based on images of interest. The secondary objective was to define a video database, called "ground truth", intended to create and train AI for the LHB assessment. HYPOTHESIS: The hypothesis was that the inter-observer agreement analysis, on standardized images, was strong enough to allow the "ground truth" videos to be used as an input database for an AI solution to be used in making arthroscopic LHB diagnoses. MATERIALS AND METHOD: One hundred and ninety-nine sets of standardized arthroscopic images of LHB exploration were evaluated by 3 independent observers. Each had to characterize the healthy or pathological state of the tendon, specifying the type of lesion: partial tear, hourglass hypertrophy, instability, fissure, superior labral anterior posterior lesion (SLAP 2), chondral print and pathological pulley without instability. Inter-observer agreement levels were measured using Cohen's Kappa (K) coefficient and Kappa Accuracy. RESULTS: The strength of agreement was moderate to strong according to the observers (Kappa 0.54 to 0.7 and KappaAcc from 86 to 92%), when determining the healthy or pathological state of the LHB. When the tendon was pathological, the strength of agreement was moderate to strong when it came to a partial tear (Kappa 0.49 to 0.71 and KappaAcc from 85 to 92%), fissure (Kappa -0.5 to 0.7 and KappaAcc from 36 to 93%) or a SLAP tear (0.54 to 0.88 and KappaAcc from 90 to 97%). It was low for unstable lesion (Kappa 0.04 to 0.25 and KappaAcc from 36 to 88%). CONCLUSION: The analysis of the LHB, from arthroscopic images, had a high level of agreement for the diagnosis of its healthy or pathological nature. However, the agreement rate decreased for the diagnosis of rare or dynamic tendon lesions. Thus, AI engineered from human analysis would have the same difficulties if it was limited only to an arthroscopic analysis. The integration of clinical and paraclinical data is necessary to improve the arthroscopic diagnosis of LHB injuries. It also seems to be an essential prerequisite for making a so-called "ground truth" database for building a high-performance AI solution. LEVEL OF EVIDENCE: III; inter-observer prospective series.

Evaluation of algorithms for Multi-Modality Whole Heart Segmentation: An open-access grand challenge.

Knowledge of whole heart anatomy is a prerequisite for many clinical applications. Whole heart segmentation (WHS), which delineates substructures of the heart, can be very valuable for modeling and analysis of the anatomy and functions of the heart. However, automating this segmentation can be challenging due to the large variation of the heart shape, and different image qualities of the clinical data. To achieve this goal, an initial set of training data is generally needed for constructing priors or for training. Furthermore, it is difficult to perform comparisons between different methods, largely due to differences in the datasets and evaluation metrics used. This manuscript presents the methodologies and evaluation results for the WHS algorithms selected from the submissions to the Multi-Modality Whole Heart Segmentation (MM-WHS) challenge, in conjunction with MICCAI 2017. The challenge provided 120 three-dimensional cardiac images covering the whole heart, including 60 CT and 60 MRI volumes, all acquired in clinical environments with manual delineation. Ten algorithms for CT data and eleven algorithms for MRI data, submitted from twelve groups, have been evaluated. The results showed that the performance of CT WHS was generally better than that of MRI WHS. The segmentation of the substructures for different categories of patients could present different levels of challenge due to the difference in imaging and variations of heart shapes. The deep learning (DL)-based methods demonstrated great potential, though several of them reported poor results in the blinded evaluation. Their performance could vary greatly across different network structures and training strategies. The conventional algorithms, mainly based on multi-atlas segmentation, demonstrated good performance, though the accuracy and computational efficiency could be limited. The challenge, including provision of the annotated training data and the blinded evaluation for submitted algorithms on the test data, continues as an ongoing benchmarking resource via its homepage (www.sdspeople.fudan.edu.cn/zhuangxiahai/0/mmwhs/).