Development of an automatic surgical planning system for high tibial osteotomy using artificial intelligence

BACKGROUND

This study proposed an automatic surgical planning system for high tibial osteotomy (HTO) using deep learning-based artificial intelligence and validated its accuracy. The system simulates osteotomy and measures lower-limb alignment parameters in pre- and post-osteotomy simulations.

METHODS

A total of 107 whole-leg standing radiographs were obtained from 107 patients who underwent HTO. First, the system detected anatomical landmarks on radiographs. Then, it simulated osteotomy and automatically measured five parameters in pre- and post-osteotomy simulation (hip knee angle [HKA], weight-bearing line ratio [WBL ratio], mechanical lateral distal femoral angle [mLDFA], mechanical medial proximal tibial angle [mMPTA], and mechanical lateral distal tibial angle [mLDTA]). The accuracy of the measured parameters was validated by comparing them with the ground truth (GT) values given by two orthopaedic surgeons.

RESULTS

All absolute errors of the system were within 1.5° or 1.5%. All inter-rater correlation confidence (ICC) values between the system and GT showed good reliability (>0.80). Excellent reliability was observed in the HKA (0.99) and WBL ratios (>0.99) for the pre-osteotomy simulation. The intra-rater difference of the system exhibited excellent reliability with an ICC value of 1.00 for all lower-limb alignment parameters in pre- and post-osteotomy simulations. In addition, the measurement time per radiograph (0.24 s) was considerably shorter than that of an orthopaedic surgeon (118 s).

CONCLUSION

The proposed system is practically applicable because it can measure lower-limb alignment parameters accurately and quickly in pre- and post-osteotomy simulations. The system has potential applications in surgical planning systems.

Automatic software-based 3D-angular measurement for weight-bearing CT (WBCT) is valid

BACKGROUND

The purpose of this study was to compare automatic software-based angular measurement (AM) with validated measurement by hand (MBH) regarding angle values and time spent for Weight-Bearing CT (WBCT) generated datasets.

METHODS

Five-hundred WBCT scans from different pathologies were included in the study. 1st - 2nd intermetatarsal angle, talo-1st metatarsal angle dorsoplantar and lateral, hindfoot angle, calcaneal pitch angle were measured and compared between MBH and AM.

RESULTS

The pathologies were ankle osteoarthritis/instability, n = 147 (29%); Haglund deformity/Achillodynia, n = 41 (8%); forefoot deformity, n = 108 (22%); Hallux rigidus, n = 37 (7%); flatfoot, n = 35 (7%); cavus foot, n = 10 (2%); osteoarthritis except ankle, n = 82 (16%). The angles did not differ between MBH and AM (each p > 0.36). The time spent for MBH / AM was 44.5 / 1 s on average per angle (p < .001).

CONCLUSIONS

AM provided angles which were not different from validated MBH and can be considered as a validated angle measurement method. The time spent was 97% lower for AM than for MBH.

LEVELS OF EVIDENCE

Level III.

Automatic measurement of short-term variability of repolarization to indicate ventricular arrhythmias in a porcine model of cardiac ischaemia

AIMS

An automated method for determination of short-term variability (STV) of repolarization on intracardiac electrograms (STV-ARIauto) has previously been developed for arrhythmic risk monitoring by cardiac implantable devices, and has proved effective in predicting ventricular arrhythmias (VA) and guiding preventive high-rate pacing (HRP) in a canine model. Current study aimed to assess (i) STV-ARIauto in relation to VA occurrence and secondarily (ii-a) to confirm the predictive capacity of STV from the QT interval and (ii-b) explore the effect of HRP on arrhythmic outcomes in a porcine model of acute myocardial infarction (MI).

METHODS AND RESULTS

Myocardial infarction was induced in 15 pigs. In 7/15 pigs, STV-QT was assessed at baseline, occlusion, 1 min before VA, and just before VA. Eight of the 15 pigs were additionally monitored with an electrogram catheter in the right ventricle, underwent echocardiography at baseline and reperfusion, and were randomized to paced or control group. Paced group received atrial pacing at 20 beats per min faster than sinus rhythm 1 min after occlusion. Short-term variability increased prior to VA in both STV modalities. The percentage change in STV from baseline to successive timepoints correlated well between STV-QT and STV-ARIauto. High-rate pacing did not improve arrhythmic outcomes and was accompanied by a stronger decrease in ejection fraction.

CONCLUSION

STV-ARIauto values increase before VA onset, alike STV-QT in a porcine model of MI, indicating imminent arrhythmias. This highlights the potential of automatic monitoring of arrhythmic risk by cardiac devices through STV-ARIauto and subsequently initiates preventive strategies. Continuous HRP during onset of acute MI did not improve arrhythmic outcomes.

Automated nailfold capillary density measurement method based on improved YOLOv5

Nailfold capillary density is an essential physiological parameter for analyzing nailfold health; however, clinical images of the nailfold are taken in many situations, and most clinicians subjectively analyze nailfold images. Therefore, based on the improved "you only look once v5" (YOLOv5) algorithm, this study proposes an automated method for measuring nailfold capillary density. The improved technique can effectively and rapidly detect distal capillaries by incorporating methods or structures such as 9mosaic, spatial pyramid pooling cross-stage partial construction, bilinear interpolation, and efficient intersection over union. First, the modified YOLOv5 algorithm was used to detect nailfold capillaries. Subsequently, the number of distal capillaries was filtered using the 90° method. Finally, the capillary density was calculated. The results showed that the Average Precision (AP)@0.5 value of the proposed approach reached 85.2 %, which was an improvement of 4.93 %, 5.24 %, and 107 % compared with the original YOLOv5, YOLOv6, and simple-faster rapid-region convolutional network (R-CNN), respectively. For different nailfold images, using the density calculated by nailfold experts as a benchmark, the calculated results of the proposed method were consistent with the manually calculated results and superior to those of the original YOLOv5.

A novel artificial intelligence model for fetal facial profile marker measurement during the first trimester

BACKGROUND

To study the validity of an artificial intelligence (AI) model for measuring fetal facial profile markers, and to evaluate the clinical value of the AI model for identifying fetal abnormalities during the first trimester.

METHODS

This retrospective study used two-dimensional mid-sagittal fetal profile images taken during singleton pregnancies at 11-13+ 6 weeks of gestation. We measured the facial profile markers, including inferior facial angle (IFA), maxilla-nasion-mandible (MNM) angle, facial-maxillary angle (FMA), frontal space (FS) distance, and profile line (PL) distance using AI and manual measurements. Semantic segmentation and landmark localization were used to develop an AI model to measure the selected markers and evaluate the diagnostic value for fetal abnormalities. The consistency between AI and manual measurements was compared using intraclass correlation coefficients (ICC). The diagnostic value of facial markers measured using the AI model during fetal abnormality screening was evaluated using receiver operating characteristic (ROC) curves.

RESULTS

A total of 2372 normal fetuses and 37 with abnormalities were observed, including 18 with trisomy 21, 7 with trisomy 18, and 12 with CLP. Among them, 1872 normal fetuses were used for AI model training and validation, and the remaining 500 normal fetuses and all fetuses with abnormalities were used for clinical testing. The ICCs (95%CI) of the IFA, MNM angle, FMA, FS distance, and PL distance between the AI and manual measurement for the 500 normal fetuses were 0.812 (0.780-0.840), 0.760 (0.720-0.795), 0.766 (0.727-0.800), 0.807 (0.775-0.836), and 0.798 (0.764-0.828), respectively. IFA clinically significantly identified trisomy 21 and trisomy 18, with areas under the ROC curve (AUC) of 0.686 (95%CI, 0.585-0.788) and 0.729 (95%CI, 0.621-0.837), respectively. FMA effectively predicted trisomy 18, with an AUC of 0.904 (95%CI, 0.842-0.966). MNM angle and FS distance exhibited good predictive value in CLP, with AUCs of 0.738 (95%CI, 0.573-0.902) and 0.677 (95%CI, 0.494-0.859), respectively.

CONCLUSIONS

The consistency of fetal facial profile marker measurements between the AI and manual measurement was good during the first trimester. The AI model is a convenient and effective tool for the early screen for fetal trisomy 21, trisomy 18, and CLP, which can be generalized to first-trimester scanning (FTS).

Performance evaluation of a deep learning-based cascaded HRNet model for automatic measurement of X-ray imaging parameters of lumbar sagittal curvature

PURPOSE

To develop a deep learning-based cascaded HRNet model, in order to automatically measure X-ray imaging parameters of lumbar sagittal curvature and to evaluate its prediction performance.

METHODS

A total of 3730 lumbar lateral digital radiography (DR) images were collected from picture archiving and communication system (PACS). Among them, 3150 images were randomly selected as the training dataset and validation dataset, and 580 images as the test dataset. The landmarks of the lumbar curve index (LCI), lumbar lordosis angle (LLA), sacral slope (SS), lumbar lordosis index (LLI), and the posterior edge tangent angle of the vertebral body (PTA) were identified and marked. The measured results of landmarks on the test dataset were compared with the mean values of manual measurement as the reference standard. Percentage of correct key-points (PCK), intra-class correlation coefficient (ICC), Pearson correlation coefficient (r), mean absolute error (MAE), mean square error (MSE), root-mean-square error (RMSE), and Bland-Altman plot were used to evaluate the performance of the cascade HRNet model.

RESULTS

The PCK of the cascaded HRNet model was 97.9-100% in the 3 mm distance threshold. The mean differences between the reference standard and the predicted values for LCI, LLA, SS, LLI, and PTA were 0.43 mm, 0.99°, 1.11°, 0.01 mm, and 0.23°, respectively. There were strong correlation and consistency of the five parameters between the cascaded HRNet model and manual measurements (ICC = 0.989-0.999, R = 0.991-0.999, MAE = 0.63-1.65, MSE = 0.61-4.06, RMSE = 0.78-2.01).

CONCLUSION

The cascaded HRNet model based on deep learning algorithm could accurately identify the sagittal curvature-related landmarks on lateral lumbar DR images and automatically measure the relevant parameters, which is of great significance in clinical application.

Performance of systolic pressure index for lower limb peripheral arterial disease in a group of elderly in sub-Saharan Africa

BACKGROUND

The automatic measurement of the ankle-brachial index (ABI) constitutes a reliable, simple, safe, rapid, and inexpensive alternative diagnostic screening test compared with the Doppler method for peripheral arterial disease (PAD). We aimed to compare the diagnostic performance of automatic ABI measurement tests to Doppler ultrasound for PAD in a group of patients aged 65 years and above, in Sub-Saharan Africa.

METHODS

This was an experimental comparative study of the performance of Doppler ultrasound to the automated ABI test in the diagnosis of PAD in patients aged ≥ 65 years followed-up at the Yaoundé Central Hospital, Cameroon between January to June 2018. An ABI threshold < 0.90 is defined as a PAD. We compare the sensitivity, and specificity of the high ankle-brachial index (ABI-HIGH), low ankle-brachial index (ABI-LOW), and the mean ankle-brachial index (ABI-MEAN) for both tests.

RESULTS

We included 137 subjects with an average age of 71.7 ± 6.8 years. In the ABI-HIGH mode, the automatic device had a sensitivity of 55% and a specificity of 98.35% with a difference between the two techniques of d = 0.024 (p = 0.016). In the ABI-MEAN mode, it had a sensitivity of 40.63% and a specificity of 99.15%; d = 0.071 (p < 0.0001). In the ABI-LOW mode, it had a sensitivity of 30.95% and a specificity of 99.11%; d = 0.119 (p < 0.0001).

CONCLUSION

The Automatic measurement of systolic pressure index has a better diagnostic performance in the detection of Peripheral Arterial Disease compared to the reference method by continuous Doppler in sub-Saharan African subjects aged ≥ 65 years.

Deep learning-based automatic sella turcica segmentation and morphology measurement in X-ray images

BACKGROUND

Although the morphological changes of sella turcica have been drawing increasing attention, the acquirement of linear parameters of sella turcica relies on manual measurement. Manual measurement is laborious, time-consuming, and may introduce subjective bias. This paper aims to develop and evaluate a deep learning-based model for automatic segmentation and measurement of sella turcica in cephalometric radiographs.

METHODS

1129 images were used to develop a deep learning-based segmentation network for automatic sella turcica segmentation. Besides, 50 images were used to test the generalization ability of the model. The performance of the segmented network was evaluated by the dice coefficient. Images in the test datasets were segmented by the trained segmentation network, and the segmentation results were saved in binary images. Then the extremum points and corner points were detected by calling the function in the OpenCV library to obtain the coordinates of the four landmarks of the sella turcica. Finally, the length, diameter, and depth of the sella turcica can be obtained by calculating the distance between the two points and the distance from the point to the straight line. Meanwhile, images were measured manually using Digimizer. Intraclass correlation coefficients (ICCs) and Bland-Altman plots were used to analyze the consistency between automatic and manual measurements to evaluate the reliability of the proposed methodology.

RESULTS

The dice coefficient of the segmentation network is 92.84%. For the measurement of sella turcica, there is excellent agreement between the automatic measurement and the manual measurement. In Test1, the ICCs of length, diameter and depth are 0.954, 0.953, and 0.912, respectively. In Test2, ICCs of length, diameter and depth are 0.906, 0.921, and 0.915, respectively. In addition, Bland-Altman plots showed the excellent reliability of the automated measurement method, with the majority measurements differences falling within ± 1.96 SDs intervals around the mean difference and no bias was apparent.

CONCLUSIONS

Our experimental results indicated that the proposed methodology could complete the automatic segmentation of the sella turcica efficiently, and reliably predict the length, diameter, and depth of the sella turcica. Moreover, the proposed method has generalization ability according to its excellent performance on Test2.

A fat fraction phantom for establishing new convolutional neural network to determine the pancreatic fat deposition

The determination of fat fraction based on Magnetic Resonance Imaging (MRI) requires extremely accurate data reconstruction for the assessment of pancreatic fat accumulation in medical diagnostics and biological research. In this study, the signal model of the oil and water emulsion was created with a 3.0 T field strength. We examined the quantification of the fat fraction from phantom and the intrapancreatic fat fraction using the techniques of magnetic resonance spectroscopy (MRS) and Iterative Decomposition with Echo Asymmetry and Least-Squares estimate (IDEAL) in magnetic resonance imaging (MRI). Additionally, we contrasted expert manual pancreatic fat assessment with MRS and IDEAL pancreatic fat fraction quantification. There was a strong connection between the true fat volume fraction and the fat fraction from IDEAL and MRS (R² = 0.99 and 0.99, respectively). For both phantom and in vivo measurements, Pearson's correlation and linear regression analysis were used. The findings of the in vivo assessment revealed a variable correlation between the pancreatic fat fraction MRI readings. We also used MR-opsy for manual pancreatic fat fraction segmentation since it read pancreatic fat fractions more accurately than IDEAL and MRS, which aided in the development of machine learning's ability to assess pancreatic fat automatically.

Automatic software-based 3D-angular measurement for Weight-Bearing CT (WBCT) provides different angles than measurement by hand

BACKGROUND

Purpose of this study was to compare automatic software-based angular measurement (AM, Autometrics, Curvebeam, Warrington, PA, USA) with previously validated measurement by hand (MBH) regarding angle values and time spent for the investigator for Weight-Bearing CT (WBCT).

METHODS

Five-hundred bilateral WBCT scans (PedCAT, Curvebeam, Warrington, PA, USA) were included in the study. Five angles (1st - 2nd intermetatarsal angle, talo-metatarsal 1-angle (TMT) dorsoplantar and lateral projection, hindfoot angle, calcaneal pitch angle) were measured with MBH and AM on the foot/ankle (side with pathology). Angles and time spent of MBH and AM were compared (t-test, homoscedatic).

RESULTS

The specific pathologies were ankle osteoarthritis/instability, n = 147 (29%); Haglund deformity/Achillodynia, n = 41 (8%); forefoot deformity, n = 108 (22%); Hallux rigidus, n = 37 (7%); flatfoot, n = 35 (7%); cavus foot, n = 10 (2%); osteoarthritis except ankle, n = 82 (16%). The angles differed between MBH and AM (each p < 0.001) except the calcaneal pitch angle (p = 0.05). The time spent for MBH / AM was 44.5 ± 12 s / 1 ± 0 s on average per angle (p < 0.0011).

CONCLUSIONS

AM provided different angles as MBH and can currently not be considered as validated angle measurement method. The investigator time spent is 97% lower for AM (1 s per angle) than for MBH (44.5 s per angle). Cases with correct angles in combination with almost no time spent showed the real potential of AM. The AM system will have to become reliable (especially in diminishing positive and negative angle values as defined) and valid which has to be proven by planned studies in the future.

LEVEL OF EVIDENCE

Level III.

Reliability of a human pose tracking algorithm for measuring upper limb joints: comparison with photography-based goniometry

Background

Range of motion (ROM) measurements are essential for diagnosing and evaluating upper extremity conditions. Clinical goniometry is the most commonly used methods but it is time-consuming and skill-demanding. Recent advances in human tracking algorithm suggest potential for automatic angle measuring from RGB images. It provides an attractive alternative for at-distance measuring. However, the reliability of this method has not been fully established. The purpose of this study is to evaluate if the results of algorithm are as reliable as human raters in upper limb movements.

Methods

Thirty healthy young adults (20 males, 10 females) participated in this study. Participants were asked to performed a 6-motion task including movement of shoulder, elbow and wrist. Images of movements were captured by commercial digital cameras. Each movement was measured by a pose tracking algorithm (OpenPose) and compared with the surgeon-measurement results. The mean differences between the two measurements were compared. Pearson correlation coefficients were used to determine the relationship. Reliability was investigated by the intra-class correlation coefficients.

Results

Comparing this algorithm-based method with manual measurement, the mean differences were less than 3 degrees in 5 motions (shoulder abduction: 0.51; shoulder elevation: 2.87; elbow flexion:0.38; elbow extension:0.65; wrist extension: 0.78) except wrist flexion. All the intra-class correlation coefficients were larger than 0.60. The Pearson coefficients also showed high correlations between the two measurements (p < 0.001).

Conclusions

Our results indicated that pose estimation is a reliable method to measure the shoulder and elbow angles, supporting RGB images for measuring joint ROM. Our results presented the possibility that patients can assess their ROM by photos taken by a digital camera.

Trial registration

This study was registered in the Clinical Trials Center of The First Affiliated Hospital, Sun Yat-sen University (2021–387).

Supplementary Information

The online version contains supplementary material available at 10.1186/s12891-022-05826-4.

Automated measurement of alpha angle on 3D-magnetic resonance imaging in femoroacetabular impingement hips: a pilot study

BACKGROUND

Femoroacetabular impingement (FAI) syndrome is an established pre-osteoarthritic condition. Diagnosis is based on both clinical and radiographic parameters. An abnormal manually calculated alpha angle in magnetic resonance imaging (MRI) is traditionally utilized to diagnose abnormal femoral head-neck offset. This pilot study aimed to assess the feasibility of automated alpha angle measurements in patients with FAI syndrome, and to compare automated with manual measurements data with regard to the time and effort needed in each method.

METHODS

Alpha angles were measured with manual and automated techniques, using postprocessing software in nineteen hip MRIs of FAI syndrome patients. Two observers conducted manual measurements. Intra- and inter-observer reproducibility and correlation of manual and automated alpha angle measurements were calculated using intra-class correlation (ICC) analysis. Both techniques were compared regarding the time taken (in minutes) and effort required, measured as the amount of mouse button presses performed.

RESULTS

The first observer's intra-observer reproducibility was good (ICC 0.77; p < 0.001) while the second observer's was good-to-excellent (ICC 0.93; p < 0.001). Inter-observer reproducibility between both observers in the first (ICC 0.45; p < 0.001) and second (ICC 0.56; p < 0.001) manual alpha angle assessment was moderate. The intra-class correlation coefficients between manual and automated alpha angle measurements were ICC = 0.24 (p = 0.052; observer 1, 1st measurement), ICC = 0.32 (p = 0.015; observer 1, 2nd measurement), ICC = 0.50 (p < 0.001; observer 2, 1st measurement), and ICC = 0.45 (p < 0.001; observer 2, 2nd measurement). Average runtime for automatic processing of the image data for the automated assessment was 16.6 ± 1.9 min. Automatic alpha angle measurements took longer (time difference: 14.6 ± 3.9 min; p < 0.001) but required less effort (difference in button presses: 231 ± 23; p < 0.001). While the automatic processing is running, the user can perform other tasks.

CONCLUSIONS

This pilot study demonstrates that objective and reliable automated alpha angle measurement of MRIs in FAI syndrome hips is feasible. Trial registration The Ethics Committee of the University of Düsseldorf approved our study (Registry-ID: 2017084398).

Feasibility study of hallux valgus measurement with a deep convolutional neural network based on landmark detection

OBJECTIVE

To develop a deep learning algorithm based on automatic detection of landmarks that can be used to automatically calculate forefoot imaging parameters from radiographs and test its performance.

MATERIALS AND METHODS

A total of 1023 weight-bearing dorsoplantar (DP) radiographs were included. A total of 776 radiographs were used for training and verification of the model, and 247 radiographs were used for testing the performance of the model. The radiologists manually marked 18 landmarks on each image. By training our model to automatically label these landmarks, 4 imaging parameters commonly used for the diagnosis of hallux valgus could be measured, including the first-second intermetatarsal angle (IMA), hallux valgus angle (HVA), hallux interphalangeal angle (HIA), and distal metatarsal articular angle (DMAA). The reference standard was determined by the radiologists' measurements. The percentage of correct key points (PCK), intragroup correlation coefficient (ICC), Pearson correlation coefficient (r), root mean square error (RMSE), and mean absolute error (MAE) between the predicted value of the model and the reference standard were calculated. The Bland-Altman plot shows the mean difference and 95% LoA.

RESULTS

The PCK was 84-99% at the 3-mm threshold. The correlation between the observed and predicted values of the four angles was high (ICC: 0.89-0.96, r: 0.81-0.97, RMSE: 3.76-6.77, MAE: 3.22-5.52). However, there was a systematic error between the model predicted value and the reference standard (the mean difference ranged from - 3.00 to - 5.08°, and the standard deviation ranged from 2.25 to 4.47°).

CONCLUSION

Our model can accurately identify landmarks, but there is a certain amount of error in the angle measurement, which needs further improvement.

Using big data from long-form recordings to study development and optimize societal impact

Big data are everywhere. In this chapter, we focus on one source: long-form, child-centered recordings collected using wearable technologies. Because these recordings are simultaneously unobtrusive and encompassing, they may be a breakthrough technology for clinicians and researchers from several diverse fields. We demonstrate this possibility by outlining three applications for the recordings-clinical treatment, large-scale interventions, and language documentation-where we see the greatest potential. We argue that incorporating these recordings into basic and applied research will result in more equitable treatment of patients, more reliable measurements of the effects of interventions on real-world behavior, and deeper scientific insights with less observational bias. We conclude by outlining a proposal for a semistructured online platform where vast numbers of long-form recordings could be hosted and more representative, less biased algorithms could be trained.

An automatic measurement system of distal femur morphological parameters using 3D slicer software

In the field of joint surgery, the computer-aided design of knee prostheses suitable for the Chinese population requires a large quantity of anatomical knee data. In this study, we propose a new method that uses 3D Slicer software to automatically measure the morphological parameters of the distal femur. First, 141 femur samples were segmented from CT data to establish the femoral shape library. Next, balanced iterative reducing and clustering using hierarchies (BIRCH) combined with iterative closest point (ICP) and generalised procrustes analysis (GPA) were used to achieve fast registration of the femur samples. The statistical model was automatically calculated from the registered femur samples, and an orthopaedic surgeon marked the points on the statistical model. Finally, we developed an automatic measurement system using 3D Slicer software, and a deformable model matching method was applied to establish the point correspondence between the statistical model and the other samples. By matching points on the statistical model to corresponding points in other samples, we measured all other samples. We marked six points and measured eight parameters. We evaluated the performance of automatic matching by comparing the points marked manually with those matched automatically and verified the accuracy of the system by comparing the manual and automatic measurement results. The results indicated that the average error of the automatic matching points was 1.03 mm, and the average length error and average angle error measured automatically by the system were 0.37 mm and 0.63°, respectively. These errors were smaller than the intra-rater and inter-rater errors measured manually by two different surgeons, which showed that the accuracy of our automatic method was high. Taken together, this study established an accurate and automatic measurement system for the distal femur based on the secondary development of 3D Slicer software to assist orthopaedic surgeons in completing the measurements of big data and further promote the improved design of Chinese-specific knee prostheses.

Artificial intelligence X-ray measurement technology of anatomical parameters related to lumbosacral stability

PURPOSE

To develop a deep learning-based model for measuring automatic lumbosacral anatomical parameters from lateral lumbar radiographs and compare its performance to that of attending-level radiologists.

METHODS

A total of 1791 lateral lumbar radiographs were collected through the PACS system and used to develop the deep learning-based model. Landmarks for the four used parameters, including the lumbosacral lordosis angle (LSLA), lumbosacral angle (LSA), sacral horizontal angle (SHA), and sacral inclination angle (SIA), were identified and automatically labeled by the model. At the same time, the measurement results were obtained through landmarks on the test set compared to manual measurements as the reference standard. Statistical analyses of the Percentage of Correct Key Points (PCK), intra-class correlation coefficient (ICC), Pearson correlation coefficient, mean absolute error (MAE), root mean square error (RMSE), and Bland-Altman plots were performed to evaluate the performance of the model.

RESULTS

The mean differences between the reference standard and the model for LSLA, LSA, SHA, and SIA, were 0.39°, 0.09°, 0.13°, and 0.12°, respectively. A strong correlation and consistency between the four parameters were found between the model and reference standard (ICC = 0.92-0.98, r = 0.92-0.97, MAE = 1.35-1.84, RMSE = 1.82-2.51), while with statistically significant difference for LSLA (p = 0.02).

CONCLUSIONS

The presented model revealed clinically equivalent measurements in terms of accuracy, while superior measurements were obtained in terms of cost-effectiveness, reliability, and reproducibility. The model may help clinicians improve their understanding and evaluation of lumbar diseases and LBP from a quantitative perspective in practical work. (ChiCTR2100048250).

A high-throughput and low-cost maize ear traits scorer

In this study, based on automatic control and image processing, a high-throughput and low-cost maize ear traits scorer (METS) was developed for the automatic measurement of 34 maize ear traits. In total, 813 maize ears were measured using METS, and the results showed that the square of the correlation coefficient (R2) of the manual measurements versus the automatic measurements for ear length, ear diameter, and kernel thickness were 0.96, 0.79, and 0.85, respectively. These maize ear traits could be used to classify the type, and the results showed that the classification accuracy of the support vector machine (SVM) model for the test set was better than that of the random forest (RF) model. In addition, the general applicability of the image analysis pipeline was also demonstrated on other independent maize ear phenotyping platforms. In conclusion, equipped with image processing and automatic control technologies, we have developed a high-throughput method for maize ear scoring, which could be popularized in maize functional genetics, genomics, and breeding applications.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11032-021-01205-4.

Development and User Evaluation of a Smartphone-Based System to Assess Range of Motion of Wrist Joint

Purpose

Measuring range of motion (ROM) in the wrist joint is an essential part of hand and wrist functional evaluations, especially before and after surgery. However, accurate measurements require experience and time. To reduce patient and surgeon burdens related to ROM measurement, a smartphone-based system, which enables participants to measure the ROM of the wrist joint semiautomatically using self-taken pictures on a smartphone, was developed and evaluated in this study.

Methods

In the developed system, participants were asked to take a picture of their wrist by using the other hand to position the joint first into full flexion and then into full extension. The hand and arm regions were automatically extracted in the program, and the ROM was estimated after the area of the hand and forearm was cropped. To verify the accuracy of ROM measurements in this system, the proposed method was tested on 66 images of hands from 33 participants; measurements were compared with those taken by hand surgeons. A limit of agreement and an intraclass correlation coefficient (ICC) were used for evaluation.

Results

The smallest averages (95% limits of agreement) of flexion and extension were 11.32° (95% confidence interval [CI], 8.88° to 13.76°) and 11.01° (95% CI, 8.64° to 13.39°), respectively. The ICC (1,1) for 3 measurements taken by one assessor was 0.99 (95% CI, 0.986–0.992), and the ICC (2,1) for 2 measurements taken by both assessors was 0.97 (95% CI, 0.947–0.977).

Conclusions

In this study, we developed a system to measure the semiautomatic ROM of the wrist joint using a smartphone image. Its accuracy was within a clinically usable error range that was comparable with that of a hand surgeon.

Clinical relevance

This system can reduce the burden of ROM measurement for both patients and doctors.

Feasibility of automatic measurements of hip joints based on pelvic radiography and a deep learning algorithm

PURPOSE

To develop and evaluate an automatic measurement model for hip joints based on anteroposterior (AP) pelvic radiography and a deep learning algorithm.

METHODS

A total of 1260 AP pelvic radiographs were included. 1060 radiographs were randomly sampled for training and validation and 200 radiographs were used as the test set. Landmarks for four commonly used parameters, such as the center-edge (CE) angle of Wiberg, Tönnis angle, sharp angle, and femoral head extrusion index (FHEI), were identified and labeled. An encoder-decoder convolutional neural network was developed to output a multi-channel heat map. Measurements were obtained through landmarks on the test set. Right and left hips were analyzed respectively. The mean of each parameter obtained by three radiologists was used as the reference standard. The Percentage of Correct Key points (PCK), intraclass correlation coefficient (ICC), Pearson correlation coefficient (r), root mean square error (RMSE), mean absolute error (MAE), and Bland-Altman plots were used to determine the performance of deep learning algorithm.

RESULTS

PCK of the model at 3 mm distance threshold range was from 87 % to 100 %. The CE angle, Tönnis angle, Sharp angle and FHEI of the left hip generated by the model were 29.8°±6.1°, 5.6°±4.2°, 39.0°±3.5° and 19 %±5 %, respectively. The parameters of the right hip were 30.4°±6.1°, 7.1°±4.4°, 38.9°±3.7° and 18 %±5 %. There were good correlation and consistency of the four parameters between the model and the reference standard (ICC 0.83-0.93, r 0.83-0.93, RMSE 0.02-3.27, MAE 0.02-1.79).

CONCLUSIONS

The new developed model based on deep learning algorithm can accurately identify landmarks on AP pelvic radiography and automatically generate parameters of hip joint. It will provide convenience for clinical practice of measurement.

Radiographic scoliosis angle estimation: spline-based measurement reveals superior reliability compared to traditional COBB method

INTRODUCTION AND OBJECTIVE

Although being standard for scoliosis curve size estimation, COBB angle measurement is well known to be inaccurate, due to a high interobserver variance in end vertebra selection and end plate contour delineation. We propose a stepwise improvement by using a spline constructed from vertebra centroids to resemble spinal curve characteristics more closely. To enhance precision even further, a neural net was trained to detect the centroids automatically.

MATERIALS & METHODS

Vertebra centroids in AP spinal X-ray images of varying quality from 551 scoliosis patients were manually labeled by 4 investigators. With these inputs, splines were generated and the computed curve sizes were compared to the manually measured COBB angles and to the curve estimation obtained from the neural net.

RESULTS

Splines achieved a higher interobserver correlation of 0.92-0.95 compared to manual COBB measurements (0.83-0.92) and showed 1.5-2 times less variance, depending on the anatomic region. This translates into an average of 1° of interobserver measurement deviation for spline-based curve estimation compared to 3°-8° for COBB measurements. The neural net was even more precise and achieved mean deviations below 0.5°.

CONCLUSION

In conclusion, our data suggest an advantage of spline-based automated measuring systems, so further investigations are warranted to abandon manual COBB measurements.

Estimation of sex based on patella measurements in a contemporary Chinese population using multidetector computed tomography: An automatic measurement method

Sex estimation is an important part of creating a biological profile, and ultimately assisting in creating a presumptive identification of unidentified skeletal remains. However, manual methods of anthropometric are time-consuming and prone to observer variability. The present study is an attempt to estimation of sex from automatic measurement of patella by multidetector computed tomography (MDCT) in a contemporary Chinese population. Four measurements for every patella, including maximum height (MAXH), maximum breadth (MAXB), maximum thickness (MAXT) and patellar volume (PV), were automatically provided by the software from CT image of 300 Chinese. The sample is composed of 156 males and 144 females with an average age of 41.44 and 45.68 years, respectively. The statistical analyses showed that all variables were sexually dimorphic. Receiver operating characteristic (ROC) analysis was performed to estimate sex from patella. The univariate analysis of each patellar parameter yielded a sex classification accuracy rate of 73.1% to 85.7%. The classification accuracy rates of sex estimation using the combination of the patellar parameters are 81.9% to 91.6%. This paper provides indications that the patella is important bone for sex estimation and it may be used as an alternative in forensic cases when the skull and pelvis are unavailable.

[Adult femur CT modeling and 3D automatic measurement of anatomical parameters]

Objective: To reconstruct a 3D model from adult femur CT scan data and automate a measurement of femoral anatomical parameters to study the characteristics of Chinese femur anatomical parameters. Methods: Using Mimics17.0, models from the CT data of 148 adult patients were established. The completed model was imported into Geomagic Studio and the anatomical landmarks of the femur were extracted to establish 3D coordinate system and unified coordinate system. Programmed with Matlab, using the nearest point iterative ICP algorithm and the 3D automatic extraction algorithm of anatomical landmarks to provide precision positioning, the femoral anatomical parameters were automatically measured and analyzed. The data were compared by using independent sample t test. Results: In this group, the diameter of the male femoral condyle was (84.1±3.6) mm, and it was (74.8±3.3) mm in the female; the anteroposterior diameter of the male femoral condyle was (66.5±3.7) mm, and it was (61.2±3.5) mm in the female; the diameter of the male ball head was (48.8±2.1) mm and it was (43.4±2.2) mm in the female; the differences between the two genders were all statistically significant (t=16.21, 8.84, 15.20, all P<0.05). The male femoral moment radius was (112.5±24.5) mm, and it was (124.7±19.2) mm in the female (t=3.30, P=0.002). The neck angle in male participants was 124.9°±4.0°, and it was 126.1°±5.5° in the female (t=1.40, P>0.05). As the height growed, most of the anatomical parameters increased accordingly. Conclusions: The 3D automatic measurement of femoral anatomical parameters is more reproducible and accurate than manual measurement. It is necessary to establish and enrich the femoral anatomical database to design and develop internal fixation products that meet the needs of Chinese people.

Automated online measurement of N2, N2O, NO, CO2, and CH4 emissions based on a gas-flow-soil-core technique

The gas-flow-soil-core (GFSC) technique allows to directly measure emission rates of denitrification gases of incubated soil cores. However, the technique was still suffering some drawbacks such as inadequate accuracy due to asynchronous detection of dinitrogen (N2) and other gases and low measurement frequency. Furthermore, its application was limited due to intensive manual operation. To overcome these drawbacks, we updated the GFSC system as described by Wang et al. (2011) by (a) using both a chemiluminescent detector and a gas chromatograph detector to measure nitric oxide (NO), (b) synchronizing the measurements of N2, NO, nitrous oxide (N2O), carbon dioxide (CO2) and methane (CH4), and (c) fully automating the sampling/analysis of all the gases. These technical modifications significantly reduced labor demands by at least a factor of two, increased the measurement frequency from 3 to 6 times per day and resulted in remarkable improvements in measurement accuracy (with detection limits of 0.5, 0.01, 0.05, 2.3 and 0.2μgN or Ch(-1)kg(-1)ds, or 17, 0.3, 1.8, 82, and 6μgN or Cm(-2)h(-1), for N2, N2O, NO, CO2, and CH4, respectively). In some circumstances, the modified system measured significantly more N2 and CO2 and less N2O and NO because of the enhanced measurement frequency. The modified system distinguished the differences in emissions of the denitrification gases and CO2 due to a 20% change in initial carbon supplies. It also remarkably recovered approximately 90% of consumed nitrate during incubation. These performances validate the technical improvement, and indicate that the improved GFSC system may provide a powerful research tool for obtaining deeper insights into the processes of soil carbon and nitrogen transformation during denitrification.