Objective To observe the value of MRI radiomics model for predicting postoperative prognosis of moderate carpal tunnel syndrome(CTS).Methods A total of 126 patients with moderate CTS who underwent endoscopic release and fat-suppressed proton density weighted imaging(PDWI)before operation were retrospectively enrolled.The patients were divided into good prognosis group(n=80)and poor prognosis group(n=46)based on postoperative functional evaluation,also randomly divided into training set and validation set at a ratio of 7∶3.Volume of interest(VOI)of the median nerve was obtained through delineating ROI of the affected wrist on fat suppressed PDWI.Radiomics features were extracted,and those associated with postoperative prognosis of CTS were screened in training set.Clinical prediction model,radiomics model and combined model of these two were established,and the predictive efficacy of the models were evaluated and compared according to the area under the curve(AUC)of receiver operating characteristic(ROC)curve.Results Patients in poor prognosis group were older than in good prognosis group(P＜0.05).A clinical model was constructed based on age.The radiomics model was constructed based on 6 radiomics features associated with postoperative prognosis of CTS,with predictive efficacy(AUC=0.872)higher than that of clinical model(AUC=0.604,P＜0.05)but not significantly different with that of the combined model(AUC=0.905,P＞0.05).Conclusion MRI radiomics model could be used to effectively predict postoperative prognosis of moderate CTS.

Objective To investigate the value of chest contrast-enhanced CT(C-CECT)combined with artificial intelligence iterative reconstruction(AIIR)for bronchial artery(BA)imaging.Methods Seventy patients who underwent C-CECT were prospectively enrolled.The images were reconstructed with AIIR(AIIR group)and hybrid iterative reconstruction(HIR,HIR group),respectively.The overall image quality,the traceability of BA,the sharpness of BA and the diagnostic confidence of abnormalities of BA were subjectively graded using a 5-point scale by two radiologists,respectively.The subjective scores and inter-observer agreement were compared between groups.The noise(SD)in reconstructed images of thoracic aorta,pulmonary trunk,BA and spinal erectors,the contrast-to-noise ratio(CNR)of the above 3 arteries relative to spinal erectors,and the diameters of BA at the origin,bifurcation and pulmonary hilum were compared between groups.Results The scores of the overall image quality,the traceability of BA,the sharpness of BA and the diagnostic confidence of abnormalities of BA were all significantly higher in AIIR group than those in HIR group(all P＜0.001),all with good inter-observer agreement(Kappa=0.46-0.73).SD of the aorta,pulmonary artery trunk,BA and erector spinal muscle in AIIR group were lower than those in HIR group,while CNR of above 3 arteries were higher than those in HIR group(all P＜0.05).No significant difference of the diameter of BA at each position was found between groups(all P＞0.05),while the consistency of measurement of AIIR group was higher than that of HIR group(intra-class correlation coefficient:0.89-0.94 vs.0.63-0.78).Conclusion C-CECT combined with AIIR could significantly improve imaging quality and diagnostic confidence of BA.

Objective To investigate the diagnostic value of electromyographic(EMG)tremor indicators for Parkinson's disease(PD)using the Logistic regression model.Methods A total of 65 patients with PD(PD group)and 39 patients with essential tremor(ET)(ET group)were enrolled and underwent EMG tremor analysis.General information,disease-related data,and EMG tremor characteristics were compared between the two groups.Multivariate Logistic regression analysis was performed to screen for independent influencing factors of PD,and receiver operating characteristic(ROC)curves were plotted.The area under the curve(AUC)was used to evaluate the diagnostic value of EMG tremor indicators for PD.Results Compared with the ET group,the PD group had a higher proportion of patients with unilateral onset and those with tremor spectrum frequency≥2 times,and a lower proportion of patients with a family history of tremor(P＜0.05).The tremor peak frequencies in the resting,postural,and weight-bearing(1 000 g)states were lower in the PD group than in the ET group(P＜0.05).There were statistically significant differences in the tremor rhythm patterns between the two groups in the resting and weight-bearing states(P＜0.05),with the PD group dominated by alternating contraction patterns and the ET group by synchronous contraction pat-terns.Multivariate Logistic regression analysis revealed that the tremor peak frequency in the weight-bearing state,the tremor rhythm pattern in the resting state,and the frequency of tremor spectrum were independent influencing factors of PD(P＜0.05).The ROC curves showed that the AUCs of the tremor peak frequency in the weight-bearing state,the tremor rhythm pattern in the resting state,and the frequency of tremor spectrum for diagnosing PD were 0.886,0.750,and 0.779,respec-tively.The combination of these three indicators yielded the highest AUC(0.936)for diagnosing PD,with a sensitivity of 81.54％and a specificity of 94.87％.Conclusion The tremor peak fre-quency in the weight-bearing state,the tremor rhythm pattern in the resting state,and the frequency of tremor spectrum provided by EMG tremor analysis can serve as clinical indicators for early diagno-sis of PD,and their combined use offers higher diagnostic value,which can be used to differentiate PD from ET.

AIM: To explore the factors affecting the whole-eye astigmatism after pterygium excision combined with autologous limbal stem cell transplantation.METHODS: A retrospective analysis was conducted on the medical records of 42 patients(42 eyes)with primary pterygium admitted in the ophthalmology department of Xijing Hospital from January 2023 to October 2023. They underwent pterygium excision combined with autologous limbal stem cell transplantation. The maximum invasion depth of pterygium into the cornea was measured with anterior segment optical coherence tomography(AS-OCT)before operation, the length of the pterygium invading cornea, the width of the limbus and the area of the invading cornea were measured during the operation, and three-dimensional values of corneal astigmatism of anterior segment, index of surface variance(ISV), index of vertical asymmetry(IVA), best corrected visual acuity(BCVA)and whole-eye astigmatism were collected before and at 1 mo after surgery. Patients with astigmatism ≤0.50 D or >0.50 D of the whole eye at 1 mo after surgery were assigned to group A and B, respectively. The differences of clinical data before and at 1 mo after surgery between the two groups, and the correlation between pre-operative clinical indicators and whole-eye astigmatism were analyzed. The decision tree algorithm was performed to explore the influencing factors of whole-eye astigmatism at 1 mo postoperatively.RESULTS: The maximum invasion depth of pterygium in the group A was significantly less than that in the group B [80.00(40.00, 180.00)μm vs 175.00(123.00, 190.00)μm, P=0.002]. Preoperative BCVA(LogMAR), whole-eye astigmatism, cornea astigmatism, ISV, IVA and maximum invasion depth of pterygium were positively correlated with whole-eye astigmatism at 1 mo after surgery(rs=0.317, P=0.041; rs=0.545, P<0.001; rs=0.448, P=0.003; rs=0.389, P=0.011; rs=0.382, P=0.013; rs=0.391, P=0.010). The decision tree algorithm screened out two influential factors: the maximum invasion depth of pterygium into the cornea and preoperative whole-eye astigmatism. The risk of whole-eye astigmatism >0.50 D at 1 mo after operation was higher with maximum invasion depth of pterygium into the cornea >95 μm than that with ≤95 μm. Among the patients with whole-eye astigmatism >2.63 D before operation, the probability of residual whole-eye astigmatism >0.50 D was 88.9%, and the predictive model AUC was 0.804.CONCLUSION: The whole-eye astigmatism after pterygium resection is mainly affected by the maximum invasion depth of pterygium into the cornea and preoperative whole-eye astigmatism. When the maximum invasion depth of pterygium into the corneal is >95 μm and the whole-eye stigmatism is >2.63 D before surgery, the patient should receive surgical treatment as soon as possible in order to obtain good clinical benefits.

With the rapid development of artificial intelligence （AI） technology， the ethical governance of AI has gained increasing attention. The Recommendation on the Ethics of Artificial Intelligence was issued by the United Nations Educational， Scientific and Cultural Organization in 2021， which clarified several principles for the ethical governance of AI. In the field of rehabilitation medicine， the research and application of AI technology have significantly improved patients’ quality of life and survival. However， due to the specificity of the service population in rehabilitation medicine， which is mostly for the sick， injured， disabled， and elderly， a series of complex ethical issues have also arisen. This paper analyzed in detail the ethical issues and challenges encountered in the research and application of AI technology in the field of rehabilitation medicine from various aspects， such as informed consent， security of privacy and data， patients’ physical and mental rehabilitation， compliance regulation， protection of specific groups， and promotion of equity. According to the principles of the Recommendation on the Ethics of Artificial Intelligence and others， response strategies were proposed， including multi-party collaboration and interdisciplinary cooperation， improving and refining relevant laws and regulations， strengthening ethical education across society， establishing accountability mechanisms， increasing investment， promoting equity， and other measures， to promote the healthy development of research and application of AI technology in the field of rehabilitation， as well as benefit humanity.

Objective To explore the diagnostic value of MRI radiomics analysis in mild carpal tunnel syndrome(CTS).Methods Seventy patients with mild CTS and 86 healthy volunteers who underwent wrist MRI examination were retrospectively selected.MRI fat-suppressed proton density weighted imaging(PDWI)were imported into 3D Slicer software,and the region of interest(ROI)delineation was performed by two radiologists independently.The 830 radiomics parameters were extracted,including first-order fea-tures,shape features,texture features,and wavelet-transform features.Radiomics parameter selection was performed through observer intraclass correlation coefficient(ICC),correlation analysis,and multivariate logistic regression.Five diagnostic models were estab-lished,including logistic regression,support vector machine,naive Bayes,decision tree,and random forest.Receiver operating charac-teristic(ROC)curve was used to analyze the diagnostic efficiency of the models.Results Seven radiomics features were selected for inclusion in the diagnostic models.The logistic regression model demonstrated the best performance,with an area under the curve(AUC)of 0.91[95%confidence interval(CI)0.86-0.96],a sensitivity of 88.63%,and a specificity of 89.00%in the training group.In the test group,the AUC was 0.92(95%CI 0.85-0.97),with a sensitivity of 90.48%and a specificity of 84.62%.Conclusion MRI radiomics analysis can be used to diagnose mild CTS,and the logistic regression model demonstrates superior diagnostic per-formance.

Objective·To evaluate the feasibility and diagnostic performance of diffusion-relaxation correlation spectroscopic imaging(DR-CSI)in assessing the heterogeneity of benign and malignant head and neck tumors,as well as in identifying occult lymph node metastasis(OLNM).Methods·A prospective study was conducted from January to December 2024 at Shanghai Ninth People's Hospital,Shanghai Jiao Tong University School of Medicine,enrolling patients with suspected head and neck tumors who were scheduled for surgery and had a confirmed pathological diagnosis.All patients underwent preoperative routine head and neck plain and contrast-enhanced magnetic resonance imaging(MRI)examinations,including DR-CSI sequence.Conventional imaging parameters,including maximal diameter(MD),depth of invasion(DOI)for tumors,and MD and short diameter(SD)for lymph nodes,were obtained.Post-processing was performed to obtain apparent diffusion coefficient(ADC),T2 value,and D-T2 spectra for all lesions.The compartment segmentation strategy was optimized based on the spectral peak distribution characteristics of different diseases,and the volume fraction(Vi)of each compartment was obtained.Independent sample t-tests,Mann-Whitney U tests,chi-square tests,or Fisher's exact tests were used to compare intergroup differences in clinical data and imaging metrics.Principal component analysis(PCA)and Adonis analysis were employed to evaluate the discriminative ability of imaging metrics among different subtypes of benign tumors.Receiver operating characteristic(ROC)analysis was used to evaluate the ability of univariate and multivariable models to characterize the malignancy of head and neck squamous cell carcinoma(HNSCC)and identify OLNM.Results·A total of 97 cases were collected,including 28 benign tumors and 69 HNSCCs.Fifteen pathologically confirmed OLNMs and 20 benign lymph nodes(BLNs)were also enrolled.Among the 28 benign tumors,there were 6 cases of pleomorphic adenoma stroma-rich(PA stroma-rich),9 cases of pleomorphic adenoma stroma-poor(PA stroma-poor),9 cases of Warthin's tumor(WT),and 4 cases of basal cell adenoma(BCA).Statistically significant differences were observed in certain imaging parameters(ADC,T2,and DR-CSI Vi)among benign tumor subtypes.PCA analysis demonstrated a strong discriminative ability of imaging parameters in distinguishing pathological subtypes of benign tumors(R2=0.64,P<0.001).Among the 69 HNSCCs,47 were classified as Grade 1(well/moderately well-differentiated)and 22 as Grade 2(moderately/poorly differentiated).Compared to Grade 1,Grade 2 showed lower ADC and higher T2 values,though differences were not statistically significant.As HNSCC malignancy increased,VA4 decreased and VB4 increased significantly.OLNM showed a significant increase in SD and VA4 compared to BLNs.The combination of SD and VA4 for preoperative OLNM identification achieved a diagnostic efficiency of 0.843.Conclusion·DR-CSI can analyze diffusion and relaxation characteristics at the sub-voxel level,offering valuable insights for characterizing benign head and neck tumor subtypes,assessing HNSCC malignancy,and identifying OLNMs.Compared to traditional parameters like ADC or T2,DR-CSI provides enhanced tissue microstructure analysis.

Objective To explore the diagnostic value of MRI radiomics analysis in mild carpal tunnel syndrome(CTS).Methods Seventy patients with mild CTS and 86 healthy volunteers who underwent wrist MRI examination were retrospectively selected.MRI fat-suppressed proton density weighted imaging(PDWI)were imported into 3D Slicer software,and the region of interest(ROI)delineation was performed by two radiologists independently.The 830 radiomics parameters were extracted,including first-order fea-tures,shape features,texture features,and wavelet-transform features.Radiomics parameter selection was performed through observer intraclass correlation coefficient(ICC),correlation analysis,and multivariate logistic regression.Five diagnostic models were estab-lished,including logistic regression,support vector machine,naive Bayes,decision tree,and random forest.Receiver operating charac-teristic(ROC)curve was used to analyze the diagnostic efficiency of the models.Results Seven radiomics features were selected for inclusion in the diagnostic models.The logistic regression model demonstrated the best performance,with an area under the curve(AUC)of 0.91[95%confidence interval(CI)0.86-0.96],a sensitivity of 88.63%,and a specificity of 89.00%in the training group.In the test group,the AUC was 0.92(95%CI 0.85-0.97),with a sensitivity of 90.48%and a specificity of 84.62%.Conclusion MRI radiomics analysis can be used to diagnose mild CTS,and the logistic regression model demonstrates superior diagnostic per-formance.

Objective·To evaluate the feasibility and diagnostic performance of diffusion-relaxation correlation spectroscopic imaging(DR-CSI)in assessing the heterogeneity of benign and malignant head and neck tumors,as well as in identifying occult lymph node metastasis(OLNM).Methods·A prospective study was conducted from January to December 2024 at Shanghai Ninth People's Hospital,Shanghai Jiao Tong University School of Medicine,enrolling patients with suspected head and neck tumors who were scheduled for surgery and had a confirmed pathological diagnosis.All patients underwent preoperative routine head and neck plain and contrast-enhanced magnetic resonance imaging(MRI)examinations,including DR-CSI sequence.Conventional imaging parameters,including maximal diameter(MD),depth of invasion(DOI)for tumors,and MD and short diameter(SD)for lymph nodes,were obtained.Post-processing was performed to obtain apparent diffusion coefficient(ADC),T2 value,and D-T2 spectra for all lesions.The compartment segmentation strategy was optimized based on the spectral peak distribution characteristics of different diseases,and the volume fraction(Vi)of each compartment was obtained.Independent sample t-tests,Mann-Whitney U tests,chi-square tests,or Fisher's exact tests were used to compare intergroup differences in clinical data and imaging metrics.Principal component analysis(PCA)and Adonis analysis were employed to evaluate the discriminative ability of imaging metrics among different subtypes of benign tumors.Receiver operating characteristic(ROC)analysis was used to evaluate the ability of univariate and multivariable models to characterize the malignancy of head and neck squamous cell carcinoma(HNSCC)and identify OLNM.Results·A total of 97 cases were collected,including 28 benign tumors and 69 HNSCCs.Fifteen pathologically confirmed OLNMs and 20 benign lymph nodes(BLNs)were also enrolled.Among the 28 benign tumors,there were 6 cases of pleomorphic adenoma stroma-rich(PA stroma-rich),9 cases of pleomorphic adenoma stroma-poor(PA stroma-poor),9 cases of Warthin's tumor(WT),and 4 cases of basal cell adenoma(BCA).Statistically significant differences were observed in certain imaging parameters(ADC,T2,and DR-CSI Vi)among benign tumor subtypes.PCA analysis demonstrated a strong discriminative ability of imaging parameters in distinguishing pathological subtypes of benign tumors(R2=0.64,P<0.001).Among the 69 HNSCCs,47 were classified as Grade 1(well/moderately well-differentiated)and 22 as Grade 2(moderately/poorly differentiated).Compared to Grade 1,Grade 2 showed lower ADC and higher T2 values,though differences were not statistically significant.As HNSCC malignancy increased,VA4 decreased and VB4 increased significantly.OLNM showed a significant increase in SD and VA4 compared to BLNs.The combination of SD and VA4 for preoperative OLNM identification achieved a diagnostic efficiency of 0.843.Conclusion·DR-CSI can analyze diffusion and relaxation characteristics at the sub-voxel level,offering valuable insights for characterizing benign head and neck tumor subtypes,assessing HNSCC malignancy,and identifying OLNMs.Compared to traditional parameters like ADC or T2,DR-CSI provides enhanced tissue microstructure analysis.

Objective To observe the value of MRI radiomics model for predicting postoperative prognosis of moderate carpal tunnel syndrome(CTS).Methods A total of 126 patients with moderate CTS who underwent endoscopic release and fat-suppressed proton density weighted imaging(PDWI)before operation were retrospectively enrolled.The patients were divided into good prognosis group(n=80)and poor prognosis group(n=46)based on postoperative functional evaluation,also randomly divided into training set and validation set at a ratio of 7∶3.Volume of interest(VOI)of the median nerve was obtained through delineating ROI of the affected wrist on fat suppressed PDWI.Radiomics features were extracted,and those associated with postoperative prognosis of CTS were screened in training set.Clinical prediction model,radiomics model and combined model of these two were established,and the predictive efficacy of the models were evaluated and compared according to the area under the curve(AUC)of receiver operating characteristic(ROC)curve.Results Patients in poor prognosis group were older than in good prognosis group(P＜0.05).A clinical model was constructed based on age.The radiomics model was constructed based on 6 radiomics features associated with postoperative prognosis of CTS,with predictive efficacy(AUC=0.872)higher than that of clinical model(AUC=0.604,P＜0.05)but not significantly different with that of the combined model(AUC=0.905,P＞0.05).Conclusion MRI radiomics model could be used to effectively predict postoperative prognosis of moderate CTS.