With the emergence of deep learning techniques based on convolutional neural networks, artificial intelligence (AI) has driven transformative developments in the field of medical image analysis. Recently, large language models (LLMs) such as ChatGPT have also started to achieve distinction in this domain. Increasing research shows the undeniable role of AI in reshaping various aspects of medical image analysis, including processes such as image enhancement, segmentation, detection in image preprocessing, and postprocessing related to medical diagnosis and prognosis in clinical settings. However, despite the significant progress in AI research, studies investigating the recent advances in AI technology in the aforementioned aspects, the changes in research hotspot trajectories, and the performance of studies in addressing key clinical challenges in this field are limited. This article provides an overview of recent advances in AI for medical image analysis and discusses the methodological profiles, advantages, disadvantages, and future trends of AI technologies.
Artificial Intelligence ; Humans ; Image Processing, Computer-Assisted/methods* ; Neural Networks, Computer ; Deep Learning ; Diagnostic Imaging/methods*

Objective To compare the positioning errors,required planning target volume(PTV)outer expansion boundary and three-dimensional displacement of two different positioning fixation devices CIVCO Uni-frame carbon fiber adjustable angle head frame and domestic integrated board in radiotherapy for the pa-tients with head tumor,and to investigate the application value of CIVCO headrest in the positioning of head tumor radiotherpay.Methods The clinical data of 82 patients with head tumors who underwent radiotherapy in this hospital from January to December 2023 were analyzed retrospectively.The patients were divided into the adjustable angle head frame group and integrated plate group,41 cases in each group.Both groups were fixed with U-shaped thermoplastic head film.Radiotherapy was performed by using ELEKTA infinity linear accelerator.The cone beam computed tomography(CBCT)image guidance was used,combined with bone reg-istration and manual calibration to obtain the patient positioning errors in three directions:X axis,Y axis and Z axis.The positioning error data of each CBCT was recorded;based on the PTV outward expansion boundary formula(2.5∑+0.7σ),the PTV outward expansion boundaries required by the two fixation devices in all di-rections for the treatment of head tumors were calculated;the 3-dimensional spatial displacement was calculat-ed based on the patient's single radiotherapy positioning error.Results The absolute positioning errors in the X,Y,and Z directions of the adjustable angle head frame group were 0.13(0.07,0.22),0.16(0.16,0.27),0.10(0.05,0.18)cm respectively,which in the integrated board group were 0.14(0.07,0.24),0.18(0.08,0.32),0.09(0.04,0.10)cm respectively.The absolute error of the Y direction in the adjustable angle head frame group was smaller than that of the integrated board group(P=0.006),and the difference in the relative error between the two groups was not statistically significant(P＞0.05).According to the PTV expansion formula,the calculated expansion boundaries of the X,Y,and Z directions in the adjustable head frame group were 0.24,0.30,0.22 cm,which in the integrated plate group were 0.21,0.36,0.21 cm,respectively.The three-dimensional space comprehensive positioning error value could be calculated by the formula calculation using the positioning error values in three directions.The mean values of the three-dimensional space positio-ning errors of the adjustable angle headstock group and domestic integrated plate group were 0.320 cm and 0.340 cm respectively,there was no statistically significant difference between the two groups.Conclusion The absolute error in Y direction and the median value of three-dimensional displacement of the CIVCO adjustable angle head rest are smaller than those of the domestic integrated plate,which is worthy of clinical promoted applica-tion if conditions permit.

Objective To explore the value of preoperative amide proton transfer(APT)imaging combined with diffusion kurtosis imaging(DKI)in predicting lymphovascular space invasion(LVSI)in cervical cancer.Methods Fifty-one patients with cervical cancer who underwent preoperative MRI examination and had complete postoperative pathology were retrospectively selected.Pelvic MRI scans were performed 1-2 weeks preoperatively,and the corresponding APT values,mean kurtosis(MK)and mean diffusivity(MD)values were obtained respectively,and the occurrence of LVSI was determined based on postoperative pathology results.The predictive effects of APT-and DKI-derived parameters alone or in combination on the LVSI status of cervical cancer were compared.Results Among 51 cases of cervical cancer,36 cases had pathologically confirmed LVSI and 15 cases did not had LVSI.The area under the curve(AUC)of the receiver operating characteristic(ROC)curve for the preoperative APT values,MK and MD values alone and in combination to predict the LVSI status of cervical cancer were 0.820,0.788,0.762,0.894,0.885,and 0.896,respectively.APT values combined with DKI-derived parameters predicted LVSI better than when they were used separately,in which APT values combined with MK and MD values predicted LVSI of cervical cancer with the largest AUC,sensitivity of 91.7％,specificity of 80.0％,and Youden's index of 0.717.Conclusion Preoperative APT imaging and DKI have important value in predicting LVSI of cervical cancer,and the combined application of the two can improve the prediction efficacy,which has certain clinical application value.

Objective:To explore the current status of presenteeism among ICU nurses and its correlation with resilience-related practice environment.Methods:Convenience sampling was adopted to select 434 ICU nurses from Beijing Tiantan Hospital, Capital Medical University in October 2022 as study subjects. An online questionnaire was conducted using the General Information Questionnaire, Chinese version of the Standford Presenteeism Scale-6 (SPS-6), and Nurse Resilience-related Practice Environment Factor Assessment Scale. Pearson correlation and multiple linear regression analysis were used to explore the relationship between ICU nurses' resilience-related practice environment and presenteeism.Results:Among ICU nurses, the total score on the Nurse Resilience-related Practice Environment Factor Assessment Scale was (122.71±26.14) and the SPS-6 score was (16.11±4.66). Presenteeism was negatively correlated with total scores on the Nurse Resilience-related Practice Environment Factor Assessment Scale and scores on career support and development, practice support and development, and personal support and development ( r=-0.494, -0.471, -0.444, and -0.476; P<0.001). Regression analysis showed that resilience-related practice environment negatively predicted presenteeism in nurses ( P<0.001) . Conclusions:ICU nurses are at a high level of presenteeism, and total scores on the Nurse Resilience-related Practice Environment Factor Assessment Scale are negatively correlated with presenteeism. Nursing administrators should emphasize the psychological state of ICU nurses and improve the resilience-related practice environment in ICUs, thereby reducing the incidence of presenteeism among ICU nurses.

Objective To explore the value of preoperative amide proton transfer(APT)imaging combined with diffusion kurtosis imaging(DKI)in predicting lymphovascular space invasion(LVSI)in cervical cancer.Methods Fifty-one patients with cervical cancer who underwent preoperative MRI examination and had complete postoperative pathology were retrospectively selected.Pelvic MRI scans were performed 1-2 weeks preoperatively,and the corresponding APT values,mean kurtosis(MK)and mean diffusivity(MD)values were obtained respectively,and the occurrence of LVSI was determined based on postoperative pathology results.The predictive effects of APT-and DKI-derived parameters alone or in combination on the LVSI status of cervical cancer were compared.Results Among 51 cases of cervical cancer,36 cases had pathologically confirmed LVSI and 15 cases did not had LVSI.The area under the curve(AUC)of the receiver operating characteristic(ROC)curve for the preoperative APT values,MK and MD values alone and in combination to predict the LVSI status of cervical cancer were 0.820,0.788,0.762,0.894,0.885,and 0.896,respectively.APT values combined with DKI-derived parameters predicted LVSI better than when they were used separately,in which APT values combined with MK and MD values predicted LVSI of cervical cancer with the largest AUC,sensitivity of 91.7％,specificity of 80.0％,and Youden's index of 0.717.Conclusion Preoperative APT imaging and DKI have important value in predicting LVSI of cervical cancer,and the combined application of the two can improve the prediction efficacy,which has certain clinical application value.

Objective:To explore the current status of presenteeism among ICU nurses and its correlation with resilience-related practice environment.Methods:Convenience sampling was adopted to select 434 ICU nurses from Beijing Tiantan Hospital, Capital Medical University in October 2022 as study subjects. An online questionnaire was conducted using the General Information Questionnaire, Chinese version of the Standford Presenteeism Scale-6 (SPS-6), and Nurse Resilience-related Practice Environment Factor Assessment Scale. Pearson correlation and multiple linear regression analysis were used to explore the relationship between ICU nurses' resilience-related practice environment and presenteeism.Results:Among ICU nurses, the total score on the Nurse Resilience-related Practice Environment Factor Assessment Scale was (122.71±26.14) and the SPS-6 score was (16.11±4.66). Presenteeism was negatively correlated with total scores on the Nurse Resilience-related Practice Environment Factor Assessment Scale and scores on career support and development, practice support and development, and personal support and development ( r=-0.494, -0.471, -0.444, and -0.476; P<0.001). Regression analysis showed that resilience-related practice environment negatively predicted presenteeism in nurses ( P<0.001) . Conclusions:ICU nurses are at a high level of presenteeism, and total scores on the Nurse Resilience-related Practice Environment Factor Assessment Scale are negatively correlated with presenteeism. Nursing administrators should emphasize the psychological state of ICU nurses and improve the resilience-related practice environment in ICUs, thereby reducing the incidence of presenteeism among ICU nurses.

Objective:To analyze the clinical characteristics, muscle imaging and pathological features of patients with anti-signal recognition particle positive immune-mediated necrotizing myopathy (SRP-IMNM).Methods:Nine patients with SRP-IMNM were collected in the Neuromuscular Disease Center of Jiaozuo People′s Hospital from May 2018 to May 2023, who were confirmed by skeletal muscle pathology and myositis-specific autoantibodies detection, and their clinical manifestations, muscle imaging and muscle pathology characteristics were systematically summarized.Results:Among the 9 patients with SRP-IMNM, there were 7 females and 2 males. The age of onset ranged from 18 to 59 years. All the patients presented proximal muscle weakness. Seven patients experienced neck weakness, and dysphagia was present in 5 patients. Laboratory examinations showed elevated serum creatine kinase levels in all 9 patients (1 866-6 725 U/L). Eight patients were combined with other antibodies positivity, except for anti-SRP antibody. Among them, 7 patients were combined with anti-Ro-52 antibody positivity, 4 patients combined with anti-Ro-52 antibody positivity alone, and 3 patients combined with 3 or more positive antibodies simultaneously. Those patients who presented with interstitial lung disease and cardiac involvement were all combined with other antibodies positivity. Seven patients completed thigh muscle magnetic resonance imaging (MRI), which showed diffuse skeletal muscle oedema, partial muscle atrophy and fatty replacement, primarily affecting the posterior thigh muscle group. Two patients underwent shank muscle MRI. The soleus involvement was evident, while the tibialis anterior muscle and gastrocnemius muscles were involved in 1 patient. All 9 patients showed varying degrees of scattered muscle fiber necrosis and regeneration on muscle biopsies. In 1 patient, a small amount of inflammatory cell infiltration was observed. Pipestem capillaries were observed in 4 patients. Immunohistochemical staining revealed a small number of CD68-positive lymphocytes in 8 patients. Additionally, 5 patients showed upregulation of major histocompatibility complex Ⅰ expression on the muscle fiber membrane, while 6 patients showed deposition of membrane attack complex (C5b-9) on non-necrotic muscle fibers and capillaries. P62 staining showed homogeneous fine-granular in sarcoplasm in 6 patients.Conclusions:In addition to proximal muscle weakness, patients with SRP-IMNM often experience neck weakness and dysphagia. Those with multiple antibodies are more likely to develop interstitial lung disease and cardiac involvement. SRP-IMNM patients have diffuse oedema in the affected muscles, and the posterior thigh muscles are more prone to atrophy and fatty tissue formation. C5b-9 deposition and pipestem capillaries are significant pathological features of SRP-IMNM, which provide additional evidence for clinical diagnosis.

Objective:To investigate the clinical phenotypes, muscle magnetic resonance imaging (MRI) and pathological changes, and genetic characteristics of myfibrillar myopathies (MFMs) cuased by MYOT gene mutation. Methods:(1) The clinical data of a MFMs family caused by a de novo frameshift mutation in MYOT gene admitted to Department of Neurology, Jiaozuo People's Hospital Affiliated to Xinxiang Medical University in February 2021 were collected. Electromyography, muscle MRI, and pathological examination were used to confirm the changes of the muscle lesions. MYOT gene mutation in the proband and other patients was detected by next generation sequencing (NGS) and Sanger sequencing, respectively. The 3D structure models of myotilin protein before and after gene mutation were predicted by AlphaFold3 and pymol3. (2) Literature on MFMs caused by MYOT gene mutation was searched from Pubmed and China National Knowledge Infrastructure from the establishment of these databases to July 2024; clinical and genetic characteristics of MFMs caused by MYOT gene mutation were summarized. Results:(1) In the 9 patients from this family, 8 had onset in adolescence (16-20 years old). Unilateral or bilateral hand muscle weakness as the first symptoms appeared in most patients, and then, hand muscle atrophy gradually appeared and slowly progressed to the proximal limbs. Electromyography showed myogenic damage. Muscle MRI showed patchy long T1 and long T2 signal intensity in the bilateral anterior tibial muscles. Muscle pathological staining showed typical rimbed vacuoles, cytoplasm, smear-like muscle fibers and desmin abnormal deposition in some muscle fibers; electron microscopy revealed disorganized myofibril structures, focal myofibril lysis, Z-band streaming, and subsarcolemmal or myofibril mitochondrial accumulation. Heterozygous mutation in MYOT gene c.680_683del (p.Val227GlufsTer10) locus was noted in 8 patients and daughter of the proband. Bioinformatics analysis suggested that MYOT gene c.680_683del mutation could cause premature termination of myotilin translation, leading to the production of a truncated protein, thereby disrupting its normal structure and function. (2) Eighty-nine patients with MFMs caused by MYOT gene mutation in previous literature mainly manifested as chronic progressive weakness of the distal or proximal limbs, with some involving the myocardium, respiratory muscles, or peripheral nerves. A total of 12 MYOT gene mutations were identified, with p. Ser60phe being the most common mutation. Except for p.Tyr4_his9del, being an in-frame mutation, the others were missense mutations. Conclusion:MFMs caused by MYOT gene mutation exhibit obvious clinical heterogeneity, characterized by very slow progression of muscle weakness; MYOT gene locus c.680_683del (p.Val227GlufsTer10) is a de novo heterozygous mutation.

Objective:To investigate the clinical phenotypes, muscle magnetic resonance imaging (MRI) and pathological changes, and genetic characteristics of myfibrillar myopathies (MFMs) cuased by MYOT gene mutation. Methods:(1) The clinical data of a MFMs family caused by a de novo frameshift mutation in MYOT gene admitted to Department of Neurology, Jiaozuo People's Hospital Affiliated to Xinxiang Medical University in February 2021 were collected. Electromyography, muscle MRI, and pathological examination were used to confirm the changes of the muscle lesions. MYOT gene mutation in the proband and other patients was detected by next generation sequencing (NGS) and Sanger sequencing, respectively. The 3D structure models of myotilin protein before and after gene mutation were predicted by AlphaFold3 and pymol3. (2) Literature on MFMs caused by MYOT gene mutation was searched from Pubmed and China National Knowledge Infrastructure from the establishment of these databases to July 2024; clinical and genetic characteristics of MFMs caused by MYOT gene mutation were summarized. Results:(1) In the 9 patients from this family, 8 had onset in adolescence (16-20 years old). Unilateral or bilateral hand muscle weakness as the first symptoms appeared in most patients, and then, hand muscle atrophy gradually appeared and slowly progressed to the proximal limbs. Electromyography showed myogenic damage. Muscle MRI showed patchy long T1 and long T2 signal intensity in the bilateral anterior tibial muscles. Muscle pathological staining showed typical rimbed vacuoles, cytoplasm, smear-like muscle fibers and desmin abnormal deposition in some muscle fibers; electron microscopy revealed disorganized myofibril structures, focal myofibril lysis, Z-band streaming, and subsarcolemmal or myofibril mitochondrial accumulation. Heterozygous mutation in MYOT gene c.680_683del (p.Val227GlufsTer10) locus was noted in 8 patients and daughter of the proband. Bioinformatics analysis suggested that MYOT gene c.680_683del mutation could cause premature termination of myotilin translation, leading to the production of a truncated protein, thereby disrupting its normal structure and function. (2) Eighty-nine patients with MFMs caused by MYOT gene mutation in previous literature mainly manifested as chronic progressive weakness of the distal or proximal limbs, with some involving the myocardium, respiratory muscles, or peripheral nerves. A total of 12 MYOT gene mutations were identified, with p. Ser60phe being the most common mutation. Except for p.Tyr4_his9del, being an in-frame mutation, the others were missense mutations. Conclusion:MFMs caused by MYOT gene mutation exhibit obvious clinical heterogeneity, characterized by very slow progression of muscle weakness; MYOT gene locus c.680_683del (p.Val227GlufsTer10) is a de novo heterozygous mutation.

Objective:To explore the value of using the mean gamma index (GI) in targer area receiving 50% prescribed dose as reference in VMAT planned dose verification through model-based dose calculation and measurement-based dose reconstruction.Methods:Based on Compass dose verificantion system, the VMAT plans for 70 patients were validated using two method. The mean GI and passing rate in target area receiving 50% of prescribed dose area for each validation plan were obtained to evaluate its application value in dose validation. First, plan information obtained by TPS calculation was input into the Compass system for performing independent dose calculation based on the accelerator data model, and obtain a three-dimensional dose based on the independent model calculation. The planned fluence measured for each patient′s treatment plan on the accelerator was reconstructed through the Compass system to obtain a three-dimensional dose based on measurement reconstruction. The three-dimensional dose obtained by the two method were compared with the three-dimensional dose calculated by TPS.Results:Combined with the gamma criteria of 3%/3 mm in the error setting condition of GI analysis, the mean GI in the area receiving 50% of prescribed dose was evaluated. GI≤0.4 was classified as PASS, 0.4 < GI ≤ 0.6 as being clinically acceptable, and GI > 0.6 as FAIL. The VMAT planned dose verification for 70 patients showed that the model-based independent calculation was in a better agreement with the TPS calculation. The GI values were all < 0.6: GI≤0.4 for 67 patients and 0.4 90% for 88 patients and < 90% for other 2, all meeting the requirements of clinical dose verification. The model-based independent dose verification is better than the measurement-based reconstructed dose verificantion, and the difference is statistically significant ( t=15.20, 10.71, P < 0.05). Conclusions:The mean GI in target area receiving 50% of prescribed dose can be used as a reference to judge the operatability of clinical plan in clinical dose verification. The mean GI value, in combination with the comprehensive result of gamma passing rate, is more convincing to evaluate dose verification. A combination of model-based dose verification, despite time-saving and labor-saving, and the measurement-based dose verification could become a reliable dose verification method for clinical application.