Font Size: a A A

Application Of Chest DR Dual-energy Subtraction Technology And Deep Learning Bone Suppression Imaging Technology In The Diagnosis Of Occupational Pneumoconiosis

Posted on:2023-04-12Degree:MasterType:Thesis
Country:ChinaCandidate:Y L XieFull Text:PDF
GTID:2544306905462594Subject:Medical imaging and nuclear medicine
Abstract/Summary:PDF Full Text Request
Occupational pneumoconiosis is a progressive lung disease characterized by diffuse fibrosis of lung tissue caused by long-term inhalation and retention of industrial mineral dust in the lungs.The higher the stage of pneumoconiosis,complications are more likely to occur,the disease progresses more rapidly,and the disability rate is higher.Therefore,the identification and accurate diagnosis of early pneumoconiosis is the focus of pneumoconiosis research.Imaging examination can not only determine the presence or absence of pneumoconiosis and its grading,but can also detect complications,which is of great significance for the early detection,diagnosis,and treatment of pneumoconiosis.The screening and diagnosis of occupational pneumoconiosis in China is based on chest high kilovolt X-ray photography or digital X-ray photography(DR)as the preferred methods and the main diagnostic basis.Small opacity of 1-3mm are the most common,and DR chest X-ray is a two-dimensional overlapping image.The bilateral lung tissue is easily affected by the structure of the thoracic bone and the mediastinum,and it is easy to miss the diagnosis of small lesions.Therefore,identifying how to fully display the lung field on the DR chest radiograph without being affected by overlapping tissues is the key to the correct diagnosis of pneumoconiosis.With the continuous improvement of X-ray examination technology,a number of new technologies have been derived,such as dual-energy subtraction,bone suppression imaging,etc.These new technologies can basically remove the overlapping influence of the thoracic skeleton and other structures and obtain a single lung tissue image,which significantly improves the clarity and diagnostic performance of small lesions on DR chest radiographs.However,due to many reasons such as case collection and equipment configuration,there are not many reports on the clinical application of such new technologies.This research mainly includes the following three parts:①To explore the clinical application value of chest DR dual energy subtraction(DES)in the diagnosis of occupational pneumoconiosis;②Bone suppression image technology based on deep learning,BSI)on the clinical application of occupational pneumoconiosis diagnosis;③comparative analysis of the two different bone suppression imaging techniques DES and BSI in the diagnosis of occupational pneumoconiosis and their advantages and disadvantages,the results are as follows:Part OneObjective:To analyze the application value of chest DR dual-energy subtraction technique in the diagnosis of occupational pneumoconiosis.Methods:DR chest radiographs(DR group)and dual-energy subtraction soft tissue images(DES group)of 86 patients with suspected pneumoconiosis and 21 dust workers in Sanshui District Disease Control Center of Foshan City were collected.The diagnostic results of chest radiographs of DR group read by 3 pneumoconiosis diagnostic physicians according to the national standard and corresponding standard radiographs were taken as the relative gold standard.The differences in the size and shape of small opacity in pneumoconiosis between DES group and DR combined with DES soft tissue images(combined group),the consistency and overall difference between DES group and relative gold standard in terms of overall profusion and diagnostic stage,as well as the differences in the judgment of small opacity profusion,small opacity aggregation and large opacity in the zone of lung were analyzed.Results:The rate of excellent soft tissue images in DES was lower than that in DR group(56.1%vs 72.9%),and the difference was statistically significant(P<0.05).Compared with the DR group,the DES group and the combined group had no significant difference in the judgment of the size and shape of the small opacity of pneumoconiosis(P>0.05).There was no significant difference(all P values>0.05)and the consistency was strong(k values were 0.860,0.893,all P values<0.05),while the judgment of the overall profusion and diagnostic staging of small opacity in the combined group was the same as that in the DR group.Compared with the overall level,there were statistical differences and strong consistency(k values were 0.939,0.934,P values<0.05);the density of small opacity in each zone of lung was significantly different between the DES group,the combined group and the DR group,respectively.The overall distribution degree of small opacity aggregation and large opacity were statistically different(P<0.05),and the consistency was strong(k values were 0.918 and 0.934,P values were all<0.05).The detection rate of small opacity profusion,overall density and pneumoconiosis of grade 1 and above in the DES group was lower than that in the DR group and the combined group.Conclusions:1.The image quality of DES soft tissue images is affected by respiratory motion artifacts,and its overall excellent rate is worse than that of DR chest radiographs.2.The DES group and the combined group have strong consistency with the DR group in the size and shape of the pneumoconiosis small opacity,the density of the lung area,the overall profusion,and the diagnosis stage.3.The detection rate and diagnosis rate of pneumoconiosis lesions in the DES group were lower than those in the DR group,while the combined group could improve the detection rate of pneumoconiosis lesions and the diagnosis rate of pneumoconiosis.Part TwoObjective:To analyze the application value of deep learning-based chest DR bone suppression imaging technology in the diagnosis of occupational pneumoconiosis.Methods:A total of 351 cases of suspected occupational pneumoconiosis DR chest X-rays(DR group)found in outpatient and occupational health examinations from January 2014 to December 2019 in Sanshui District Center for Disease Control and Prevention of Foshan City were collected,and bone suppression software was used for bone suppression.After treatment,a DR bone-suppressed soft tissue image group(BSI group)was obtained.Using the diagnosis results of three pneumoconiosis diagnosticians who read DR chest radiographs jointly according to the national diagnostic standard radiographs as the relative gold standard,the differences in the size and shape of small shadows in the BSI group and DR combined with bone depression soft tissue imaging group(combined group)were analyzed,as well as the consistency and overall difference between BSI group and DR combined with bone depression soft tissue imaging group and relative gold standard.And the difference in judgment of small opacity profusion,small opacity aggregation and large opacity in the zone of lung.Results:Compared with the DR group,the BSI group and the combined group had no significant difference in the judgment of the size and shape of pneumoconiosis small opacity,small opacity aggregation and the large opacity(all P>0.05).The judgments of the overall profusion were all highly consistent(k values were 0.749,0.750,0.768,and 0.778;P values were all<0.05);the diagnosis of pneumoconiosis staging in the BSI group was more consistent with the DR group(k value).Compared with the DR group,the diagnosis of pneumoconiosis staging in the combined group was more consistent(k value was 0.819,P<0.05);both groups could significantly reduce the overall profusion of each the zone of lung by 1.At the same time,it can significantly reduce the diagnosis rate of pneumoconiosis in stage I and improve the diagnosis rate of stage II.The overall opacity and pneumoconiosis stage of the BSI group and the combined group were significantly different from those of the DR group(all P<0.05).Conclusions:1.Pure BSI soft tissue images and BSI combined with DR images have no significant difference with DR chest X-ray in judging the shape and size of pneumoconiosis small opacity,small opacity aggregation and large opacity.2.Compared with DR chest X-ray,the diagnosis of pneumoconiosis small shadow density,overall profusion and pneumoconiosis staging in BSI group and combined group were stronger or more consistent.The results show that DR chest radiography can improve the identification of small opacity and the profusion of small opacity with the assistance of BSI technology,thereby improving the overall profusion of small opacity and the diagnosis and staging of pneumoconiosis.3.The difference analysis of the overall degree of pneumoconiosis between the BSI group and the combined group and the DR group in the overall profusion and diagnostic stages of the pneumoconiosis showed that the diagnostic performance of the BSI group and the BSI combined with the DR group was equivalent.Part ThreeObjective:To compare the diagnostic performance of DES technology and BSI technology for pneumoconiosis.Methods:A total of 123 suspected cases of occupational pneumoconiosis found in the Department of Occupational Disease Prevention and Control of Sanshui District Center for Disease Control and Prevention in Foshan City from July 2019 to June 2020 were collected.DES images(DES group)were obtained by dual-energy subtraction technology.The processing software performed bone suppression processing on the DR original images obtained in the DES exposure to obtain a bone suppression soft tissue image group(BSI group).Three diagnostic physicians of pneumoconiosis were used to read and diagnose the soft tissue images obtained by DES alone,DR images obtained by DES combined with DES exposure,BSI soft tissue images obtained by BSI combined with DES exposure respectively according to the national diagnostic standard tablets.The image quality of SOFT tissue images of DES and BSI was compared.Analysis of DES and BSI group,DES joint DR group and BSI DR each zone of the lung area small opacity profusion of pneumoconiosis,overall profusion and staging diagnosis efficiency.According to the image quality of soft tissue image and BSI,the diagnostic performance of DES group and BSI group,DES combined with DR group and BSI combined with DR group on the small opacity profusion,overall profusion and staging of pneumoconiosis in each zone of the lung was analyzed.Results:BSI soft tissue image quality was better than DES soft tissue image(all P values<0.01).The DES group and BSI group had strong consistency in judging the overall profusion of the small opacity and pneumoconiosis diagnosis and staging(k values were 0.717 and 0.666,respectively,P values were all<0.05);The overall profusion of small opacity and the diagnosis of pneumoconiosis staging showed strong consistency(k values were 0.924,0.925,P values<0.05,respectively).DES combined with DR group and BSI combined with DR group showed strong consistency in the diagnosis of small opacity opacity and pneumoconiosis staging(K value 0.924,0.925,P<0.05).All groups showed strong consistency in the diagnosis of the overall profusion in the six upper,middle and lower lung areas of left,right(K value 0.836,0.903,P<0.05),and the consistency in the upper lung area>middle lung area>lower lung area,and the consistency in the right lung area>left lung area.In addition,between the DES group and the BSI group,only the small opacity profusion in the bilateral upper lung region showed strong consistency,the double middle lung region and the right lower lung region showed strong consistency,and the left lower lung region showed moderate consistency(P values were all<0.05),while the DES combined and BSI combined groups showed strong consistency in the diagnosis of the small opacity profusion in the upper and middle lung areas,and strong consistency in the lower lung areas(both P values<0.05).In terms of overall degree of difference,there are differences between DES and BSI group,DES combined and BSI combined group.The diagnostic performance of the BSI combined group is better than DES combined,the BSI is better than DES,and BSI is completely consistent with BSI combined group for the diagnosis and staging of pneumoconiosis.There was no significant difference in the small opacity aggregation and large opacity between the two groups(P>0.05).Conclusion:1.Because the quality of DES soft tissue images is affected by respiratory motion,the image quality of DES in pneumoconiosis cases is not as good as that of BSI soft tissue images.2.The BSI combined group and the DES combined group are better than the DES and BSI groups in the diagnosis of pneumoconiosis small opacity profusion,overall profusion and staging.The BSI combined group is better than the DES combined group,and the BSI group is better than DES.Lesion identification and diagnosis are superior to DES technology.3.The diagnosis of pneumoconiosis lesions by BSI soft tissue image is equivalent to that of BSI combined with DR image.Full text summary:1.The image quality of DES soft tissue images is significantly affected by respiratory motion artifacts,and the overall excellent rate is worse than that of DR chest radiographs and BSI soft tissue images.2.Pure DES soft tissue imaging has certain limitations in the identification of pneumoconiosis lesions,but DES combined with DR images can improve the detection rate of pneumoconiosis lesions and the diagnosis rate of pneumoconiosis stages.3.Pure BSI soft tissue images and BSI combined with DR images have the same ability to identify and diagnose pneumoconiosis.With the assistance of BSI technology,DR chest X-ray can improve the identification of small opacity and the profusion of small opacity,thereby improving the overall profusion of small opacity and the diagnosis and staging of pneumoconiosis.4.BSI combined with DR image and BSI image alone are better than DES combined with DR image and DES image respectively in the diagnosis of pneumoconiosis.The research shows that BSI technology is better than DES technology in the identification and diagnosis of pneumoconiosis lesions.5.Unlike DES,BSI,as a post-processing software,does not require specific equipment and does not generate additional radiation dose,and is more suitable for widespread application in the screening and diagnosis of occupational pneumoconiosis.
Keywords/Search Tags:pneumoconiosis, chest, digital radiography, dual energy subtraction, deep learning, bone suppression, diagnosis
PDF Full Text Request
Related items