Current Applications of Machine Learning in Spine: From Clinical View

Ren, GuanRui; Yu, Kun; Xie, Zhi‐Yang; Wang, Peiyang; Zhang, Wei; Huang, Yong; Wang, Yuntao

doi:10.1177/21925682211035363

Cited by 24 publications

(23 citation statements)

References 170 publications

(226 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The seven systematic reviews focus on the most important ML and computer models to realize phase recognition in general surgery. Secondary aims include presenting commonly used concepts of AI and ML and discussions of data types and limitations of currently used model systems [3,4,7,10] , variance analysis for predictive ability [8] , and overall accuracy of those models [7,10] . The most recently published systematic review is a comprehensive review on surgical data science, providing evidence [9] of progress in this field since 2004.…”

Section: Resultsmentioning

confidence: 99%

“…The performance of ML is ranked according to its level of discrimination (probability of predicting outcomes accurately) and calibration (the degree of over-or underestimating the predicted versus true outcome) [7] .…”

Section: Types Of Machine Learningmentioning

confidence: 99%

“…Data labeling could be expensive, while gathering poorly labeled data will produce poor results and waste money [3,13,19] . It could also have a high inter-operator variability; for example, the total surface area of a burn injury could present a barrier to effective ML due to varying degrees of accuracy of labeled data [7] .…”

Section: Data Annotationmentioning

confidence: 99%

“…In addition to ANNs, convolutional neural networks (CNNs) are increasingly being used in image processing as they are designed to process pixel data [6] and find efficacy in the field of computer vision (CV). Interestingly, CNNs structurally resemble the neural architecture used by the human visual cortex, which is situated in the occipital lobe [7] .…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

The challenges of deep learning in artificial intelligence and autonomous actions in surgery: a literature review

Taher¹,

Grasso²,

Tawfik³

et al. 2022

Art Int Surg

View full text Add to dashboard Cite

Aim: Artificial intelligence (AI) is rapidly evolving in healthcare worldwide, especially in surgery. This article reviews important terms used in machine learning and the challenges of deep learning in surgery. Methods: A review of the English literature was carried out focused on the terms “challenges of deep learning” and “surgery” using Medline and PubMed between 2018 and 2022. Results: In total, 54 articles discussed the challenges of deep learning in general. We include 25 articles from various surgical specialties discussing challenges corresponding to their respective specialties. Conclusion: The increased utilization of AI in surgery is faced with a wide variety of technical, ethical, clinical, and business-related challenges. The best way to expedite its expansion in surgery in the safest and most cost-efficient manner is by ensuring that as many surgeons as possible have a clear understanding of basic AI concepts and how they can be applied to the preoperative, intraoperative, postoperative, and long-term follow-up phases of the surgical patient care.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Types Of Machine Learningmentioning

confidence: 99%

Section: Data Annotationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

The challenges of deep learning in artificial intelligence and autonomous actions in surgery: a literature review

Taher¹,

Grasso²,

Tawfik³

et al. 2022

Art Int Surg

View full text Add to dashboard Cite

show abstract

“…The rapid advancement of artificial intelligence/machine learning technology in the field of spine care will make near instant quantitative measurements of spondylolisthesis, disc heights, angles and other metrics possible 1–3 . That technology will require validated guidelines for how to optimally interpret these measurements.…”

Section: Introductionmentioning

confidence: 99%

Factors in measuring lumbar spondylolisthesis with reference data from NHANES-II

Hipp¹,

Grieco²,

Newman³

et al. 2022

Preprint

View full text Add to dashboard Cite

Background Although spondylolisthesis, disc height loss, and abnormal disc angles are commonly assessed for in clinical practice and research studies, the factors that influence measurements are incompletely understood, and reference data representing a very large and diverse population are not available. Over 7000 lumbar spine x-rays were collected as part of the 2nd National Health and Nutrition Examination Survey (NHANES-II). This nationwide probability sample may facilitate development of robust reference data to objectively classify spondylolisthesis, disc heights, or disc angles as normal vs abnormal. Analysis of lumbar spine x-rays may also help to document whether disc heights and disc angles must be considered when assessing for spondylolisthesis. Methods Anatomic landmarks were obtained using neural networks and coded logic for L1 to S1 in the NHANES-II lumbar spine radiographs. Nine disc and sagittal plane offset (SPO) measurements were calculated from the landmarks. These data were trimmed to exclude abnormal discs and SPO. The factors that affect SPO were explored along with associations between the metrics and with age, sex, race, nation of origin and BMI. The prevalence of abnormalities was tabulated. Metrics were also calculated for previously analyzed lumbar flexion-extension x-rays to better understand the dependence of SPO on disc angles and disc heights. The errors that occur due to variability in radiographic projection were also assessed. Results SPO and disc metrics were obtained for 35,490 levels from 7,358 NHANES-II subjects who were 25 to 74 years old. Descriptive statistics for nine SPO and disc metrics were tabulated before and after trimming the data. Age, BMI, and sex were statistically significant but explained little of the variance in the metrics. SPO was significantly dependent on disc angle and height, though less so in the NHANES radiographs than in lumbar flexion-extension studies. Errors in the metrics due to out-of-plane imaging are generally small but can be large with extensive out-of-plane x-rays. Discussion The NHANES-II collection of lumbar x-rays allows for establishing reference data for SPO and disc metrics. These reference data allow for easily interpreted standardized reporting in units of std dev from average normal. SPO was significantly dependent on disc angle and disc height although the effect is small when there is limited flexion or extension. If SPO is being assessed from flexion or extension, a simple correction can be made. Caution is needed when interpreting measurements when the x-ray beam passes obliquely through the endplates or posterior wall (out-of-plane imaging). Conclusions The NHANES-II data may facilitate standardized assessments of SPO, disc heights and disc angles. Adjustments should be applied to SPO measurements if made from x-rays with the patient in flexed or extended positions.

show abstract

Developing a new tool for scoliosis screening in a tertiary specialistic setting using artificial intelligence: a retrospective study on 10,813 patients: 2023 SOSORT award winner

Negrini,

Cina,

Ferrario

et al. 2023

Eur Spine J

View full text Add to dashboard Cite

Purpose The study aims to assess if the angle of trunk rotation (ATR) in combination with other readily measurable clinical parameters allows for effective non-invasive scoliosis screening. Methods We analysed 10,813 patients (4–18 years old) who underwent clinical and radiological evaluation for scoliosis in a tertiary clinic specialised in spinal deformities. We considered as predictors ATR, Prominence (mm), visible asymmetry of the waist, scapulae and shoulders, familiarity, sex, BMI, age, menarche, and localisation of the curve. We implemented a Logistic Regression model to classify the Cobb angle of the major curve according to thresholds of 15, 20, 25, 30, and 40 degrees, by randomly splitting the dataset into 80–20% for training and testing, respectively. Results The model showed accuracies of 74, 81, 79, 79, and 84% for 15-, 20-, 25-, 30- and 40-degrees thresholds, respectively. For all the thresholds ATR, Prominence, and visible asymmetry of the waist were the top five most important variables for the prediction. Samples that were wrongly classified as negatives had always statistically significant (p ≪ 0.01) lower values of ATR and Prominence. This confirmed that these two parameters were very important for the correct classification of the Cobb angle. The model showed better performances than using the 5 and 7 degrees ATR thresholds to prescribe a radiological examination. Conclusions Machine-learning-based classification models have the potential to effectively improve the non-invasive screening for AIS. The results of the study constitute the basis for the development of easy-to-use tools enabling physicians to decide whether to prescribe radiographic imaging.

show abstract

Current Applications of Machine Learning in Spine: From Clinical View

Cited by 24 publications

References 170 publications

The challenges of deep learning in artificial intelligence and autonomous actions in surgery: a literature review

The challenges of deep learning in artificial intelligence and autonomous actions in surgery: a literature review

Factors in measuring lumbar spondylolisthesis with reference data from NHANES-II

Developing a new tool for scoliosis screening in a tertiary specialistic setting using artificial intelligence: a retrospective study on 10,813 patients: 2023 SOSORT award winner

Contact Info

Product

Resources

About