A deep learning model based on fusion images of chest radiography and X-ray sponge images supports human visual characteristics of retained surgical items detection

Kawakubo, Masateru; Waki, Hiroto; Shirasaka, Takashi; Kojima, Tohru; Mikayama, Ryoji; Hamasaki, Hiroshi; Akamine, Hiroshi; Kato, Tomoko S.; Baba, Shingo; Ushiro, Shin; Ishigami, Kousei

doi:10.1007/s11548-022-02816-8

Cited by 5 publications

(7 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is also critical in medical image segmentation, where one needs to outline and define objects, such as RSIs. Faster region-based convolutional neural networks (R-CNN) gained wide appreciation as they were the first to reliably detect objects due to the use of region proposal networks (RPN) and the Fast R-CNN component [ 30 ]. This model has the potential to achieve higher degrees of sensitivity and specificity, accurately identifying surgical items in images, which is paramount for medical diagnosis.…”

Section: Deep Learning-based Methodsmentioning

confidence: 99%

“…Despite proposing a novel DL software that uses post-processed images created by combining X-ray images of typical post-operative radiography and surgical sponges, researchers have not fully investigated the relationship between the detectability of RSIs and human visual assessment. In [ 30 ], a study examined the relationship between the detectability of RSIs and human subjectivity using DL. A DL model was created using 2987 training shots and 1298 validation images that were post-processed by integrating X-ray images of typical post-operative radiography and surgical sponges.…”

Section: Deep Learning-based Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Minimization of occurrence of retained surgical items using machine learning and deep learning techniques: a review

Abo-Zahhad,

El-Malek,

Sayed

et al. 2024

BioData Mining

View full text Add to dashboard Cite

Retained surgical items (RSIs) pose significant risks to patients and healthcare professionals, prompting extensive efforts to reduce their incidence. RSIs are objects inadvertently left within patients’ bodies after surgery, which can lead to severe consequences such as infections and death. The repercussions highlight the critical need to address this issue. Machine learning (ML) and deep learning (DL) have displayed considerable potential for enhancing the prevention of RSIs through heightened precision and decreased reliance on human involvement. ML techniques are finding an expanding number of applications in medicine, ranging from automated imaging analysis to diagnosis. DL has enabled substantial advances in the prediction capabilities of computers by combining the availability of massive volumes of data with extremely effective learning algorithms. This paper reviews and evaluates recently published articles on the application of ML and DL in RSIs prevention and diagnosis, stressing the need for a multi-layered approach that leverages each method’s strengths to mitigate RSI risks. It highlights the key findings, advantages, and limitations of the different techniques used. Extensive datasets for training ML and DL models could enhance RSI detection systems. This paper also discusses the various datasets used by researchers for training the models. In addition, future directions for improving these technologies for RSI diagnosis and prevention are considered. By merging ML and DL with current procedures, it is conceivable to substantially minimize RSIs, enhance patient safety, and elevate surgical care standards.

show abstract

Section: Deep Learning-based Methodsmentioning

confidence: 99%

Section: Deep Learning-based Methodsmentioning

confidence: 99%

Minimization of occurrence of retained surgical items using machine learning and deep learning techniques: a review

Abo-Zahhad,

El-Malek,

Sayed

et al. 2024

BioData Mining

View full text Add to dashboard Cite

show abstract

“…Yet, these authors note that a limitation of the study was that the software only identified specific surgical sponges and could not recognize other retained surgical objects [ 18 ]. Kawakubo et al also developed a DL model to detect retained surgical items by post-processing fused images of surgical sponges and unremarkable postoperative X-rays [ 19 ]. The authors subsequently compared the model to two experienced radiologists identifying retained surgical sponges [ 19 ].…”

Section: Retained Surgical Bodiesmentioning

confidence: 99%

“…Kawakubo et al also developed a DL model to detect retained surgical items by post-processing fused images of surgical sponges and unremarkable postoperative X-rays [ 19 ]. The authors subsequently compared the model to two experienced radiologists identifying retained surgical sponges [ 19 ]. The deep learning model had higher sensitivity and lower specificity for sponge detection compared to both human observers, suggesting its potential to support diagnostic ability by reducing the rate of missed RSBs.…”

Section: Retained Surgical Bodiesmentioning

confidence: 99%

“…Thus, the model will not be able to identify and classify foreign objects it has not previously encountered. This was particularly evident in the model developed by Kawakubo et al, as the software only identified specific surgical sponges for which the model was trained and could not recognize other retained surgical objects [ 19 ]. Moreover, the breadth and variety of training datasets, encompassing objects in diverse orientations and forms, are crucial for AI's ability to detect foreign bodies.…”

Section: Limitations Of Ai Imagingmentioning

confidence: 99%

See 1 more Smart Citation

Artificial intelligence in the detection of non-biological materials

Eibschutz,

Lu,

Abbassi

et al. 2024

Emerg Radiol

View full text Add to dashboard Cite

Artificial Intelligence (AI) has emerged as a transformative force within medical imaging, making significant strides within emergency radiology. Presently, there is a strong reliance on radiologists to accurately diagnose and characterize foreign bodies in a timely fashion, a task that can be readily augmented with AI tools. This article will first explore the most common clinical scenarios involving foreign bodies, such as retained surgical instruments, open and penetrating injuries, catheter and tube malposition, and foreign body ingestion and aspiration. By initially exploring the existing imaging techniques employed for diagnosing these conditions, the potential role of AI in detecting non-biological materials can be better elucidated. Yet, the heterogeneous nature of foreign bodies and limited data availability complicates the development of computer-aided detection models. Despite these challenges, integrating AI can potentially decrease radiologist workload, enhance diagnostic accuracy, and improve patient outcomes.

show abstract

Patient Identification Based on Deep Metric Learning for Preventing Human Errors in Follow-up X-Ray Examinations

Ueda

Morishita

2023

J Digit Imaging

View full text Add to dashboard Cite

Biological fingerprints extracted from clinical images can be used for patient identity verification to determine misfiled clinical images in picture archiving and communication systems. However, such methods have not been incorporated into clinical use, and their performance can degrade with variability in the clinical images. Deep learning can be used to improve the performance of these methods. A novel method is proposed to automatically identify individuals among examined patients using posteroanterior (PA) and anteroposterior (AP) chest X-ray images. The proposed method uses deep metric learning based on a deep convolutional neural network (DCNN) to overcome the extreme classification requirements for patient validation and identification. It was trained on the NIH chest X-ray dataset (ChestX-ray8) in three steps: preprocessing, DCNN feature extraction with an EfficientNetV2-S backbone, and classification with deep metric learning. The proposed method was evaluated using two public datasets and two clinical chest X-ray image datasets containing data from patients undergoing screening and hospital care. A 1280-dimensional feature extractor pretrained for 300 epochs performed the best with an area under the receiver operating characteristic curve of 0.9894, an equal error rate of 0.0269, and a top-1 accuracy of 0.839 on the PadChest dataset containing both PA and AP view positions. The findings of this study provide considerable insights into the development of automated patient identification to reduce the possibility of medical malpractice due to human errors.

show abstract

A deep learning model based on fusion images of chest radiography and X-ray sponge images supports human visual characteristics of retained surgical items detection

Cited by 5 publications

References 19 publications

Minimization of occurrence of retained surgical items using machine learning and deep learning techniques: a review

Minimization of occurrence of retained surgical items using machine learning and deep learning techniques: a review

Artificial intelligence in the detection of non-biological materials

Patient Identification Based on Deep Metric Learning for Preventing Human Errors in Follow-up X-Ray Examinations

Contact Info

Product

Resources

About