The diagnostic performance and clinical value of deep learning-based nodule detection system concerning influence of location of pulmonary nodule
Abstract Background The deep learning-based nodule detection (DLD) system improves nodule detection performance of observers on chest radiographs (CXRs). However, its performance in different pulmonary nodule (PN) locations remains unknown. Methods We divided the CXR intrathoracic region into non-da...
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2023-09-01
|
| Series: | Insights into Imaging |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s13244-023-01497-4 |
| _version_ | 1827710614563192832 |
|---|---|
| author | Seulgi You Ji Hyun Park Bumhee Park Han-Bit Shin Taeyang Ha Jae Sung Yun Kyoung Joo Park Yongjun Jung You Na Kim Minji Kim Joo Sung Sun |
| author_facet | Seulgi You Ji Hyun Park Bumhee Park Han-Bit Shin Taeyang Ha Jae Sung Yun Kyoung Joo Park Yongjun Jung You Na Kim Minji Kim Joo Sung Sun |
| author_sort | Seulgi You |
| collection | DOAJ |
| description | Abstract Background The deep learning-based nodule detection (DLD) system improves nodule detection performance of observers on chest radiographs (CXRs). However, its performance in different pulmonary nodule (PN) locations remains unknown. Methods We divided the CXR intrathoracic region into non-danger zone (NDZ) and danger zone (DZ). The DZ included the lung apices, paramediastinal areas, and retrodiaphragmatic areas, where nodules could be missed. We used a dataset of 300 CXRs (100 normal and 200 abnormal images with 216 PNs [107 NDZ and 109 DZ nodules]). Eight observers (two thoracic radiologists [TRs], two non-thoracic radiologists [NTRs], and four radiology residents [RRs]) interpreted each radiograph with and without the DLD system. The metric of lesion localization fraction (LLF; the number of correctly localized lesions divided by the total number of true lesions) was used to evaluate the diagnostic performance according to the nodule location. Results The DLD system demonstrated a lower LLF for the detection of DZ nodules (64.2) than that of NDZ nodules (83.2, p = 0.008). For DZ nodule detection, the LLF of the DLD system (64.2) was lower than that of TRs (81.7, p < 0.001), which was comparable to that of NTRs (56.4, p = 0.531) and RRs (56.7, p = 0.459). Nonetheless, the LLF of RRs significantly improved from 56.7 to 65.6 using the DLD system (p = 0.021) for DZ nodule detection. Conclusion The performance of the DLD system was lower in the detection of DZ nodules compared to that of NDZ nodules. Nonetheless, RR performance in detecting DZ nodules improved upon using the DLD system. Critical relevance statement Despite the deep learning-based nodule detection system’s limitations in detecting danger zone nodules, it proves beneficial for less-experienced observers by providing valuable assistance in identifying these nodules, thereby advancing nodule detection in clinical practice. Key points • The deep learning-based nodule detection (DLD) system can improve the diagnostic performance of observers in nodule detection. • The DLD system shows poor diagnostic performance in detecting danger zone nodules. • For less-experienced observers, the DLD system is helpful in detecting danger zone nodules. Graphical Abstract |
| first_indexed | 2024-03-10T17:42:26Z |
| format | Article |
| id | doaj.art-0a8b8d1cbf5f437386e690f974b5517f |
| institution | Directory Open Access Journal |
| issn | 1869-4101 |
| language | English |
| last_indexed | 2024-03-10T17:42:26Z |
| publishDate | 2023-09-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | Insights into Imaging |
| spelling | doaj.art-0a8b8d1cbf5f437386e690f974b5517f2023-11-20T09:39:01ZengSpringerOpenInsights into Imaging1869-41012023-09-0114111110.1186/s13244-023-01497-4The diagnostic performance and clinical value of deep learning-based nodule detection system concerning influence of location of pulmonary noduleSeulgi You0Ji Hyun Park1Bumhee Park2Han-Bit Shin3Taeyang Ha4Jae Sung Yun5Kyoung Joo Park6Yongjun Jung7You Na Kim8Minji Kim9Joo Sung Sun10Department of Radiology, Ajou University School of MedicineOffice of Biostatistics, Ajou Research Institute for Innovative Medicine, Ajou University Medical CenterOffice of Biostatistics, Ajou Research Institute for Innovative Medicine, Ajou University Medical CenterOffice of Biostatistics, Ajou Research Institute for Innovative Medicine, Ajou University Medical CenterDepartment of Radiology, Ajou University School of MedicineDepartment of Radiology, Ajou University School of MedicineDepartment of Radiology, Ajou University School of MedicineDepartment of Radiology, Ajou University School of MedicineDepartment of Radiology, Ajou University School of MedicineDepartment of Radiology, Ajou University School of MedicineDepartment of Radiology, Ajou University School of MedicineAbstract Background The deep learning-based nodule detection (DLD) system improves nodule detection performance of observers on chest radiographs (CXRs). However, its performance in different pulmonary nodule (PN) locations remains unknown. Methods We divided the CXR intrathoracic region into non-danger zone (NDZ) and danger zone (DZ). The DZ included the lung apices, paramediastinal areas, and retrodiaphragmatic areas, where nodules could be missed. We used a dataset of 300 CXRs (100 normal and 200 abnormal images with 216 PNs [107 NDZ and 109 DZ nodules]). Eight observers (two thoracic radiologists [TRs], two non-thoracic radiologists [NTRs], and four radiology residents [RRs]) interpreted each radiograph with and without the DLD system. The metric of lesion localization fraction (LLF; the number of correctly localized lesions divided by the total number of true lesions) was used to evaluate the diagnostic performance according to the nodule location. Results The DLD system demonstrated a lower LLF for the detection of DZ nodules (64.2) than that of NDZ nodules (83.2, p = 0.008). For DZ nodule detection, the LLF of the DLD system (64.2) was lower than that of TRs (81.7, p < 0.001), which was comparable to that of NTRs (56.4, p = 0.531) and RRs (56.7, p = 0.459). Nonetheless, the LLF of RRs significantly improved from 56.7 to 65.6 using the DLD system (p = 0.021) for DZ nodule detection. Conclusion The performance of the DLD system was lower in the detection of DZ nodules compared to that of NDZ nodules. Nonetheless, RR performance in detecting DZ nodules improved upon using the DLD system. Critical relevance statement Despite the deep learning-based nodule detection system’s limitations in detecting danger zone nodules, it proves beneficial for less-experienced observers by providing valuable assistance in identifying these nodules, thereby advancing nodule detection in clinical practice. Key points • The deep learning-based nodule detection (DLD) system can improve the diagnostic performance of observers in nodule detection. • The DLD system shows poor diagnostic performance in detecting danger zone nodules. • For less-experienced observers, the DLD system is helpful in detecting danger zone nodules. Graphical Abstracthttps://doi.org/10.1186/s13244-023-01497-4Deep learningChest radiographySolitary pulmonary nodule |
| spellingShingle | Seulgi You Ji Hyun Park Bumhee Park Han-Bit Shin Taeyang Ha Jae Sung Yun Kyoung Joo Park Yongjun Jung You Na Kim Minji Kim Joo Sung Sun The diagnostic performance and clinical value of deep learning-based nodule detection system concerning influence of location of pulmonary nodule Insights into Imaging Deep learning Chest radiography Solitary pulmonary nodule |
| title | The diagnostic performance and clinical value of deep learning-based nodule detection system concerning influence of location of pulmonary nodule |
| title_full | The diagnostic performance and clinical value of deep learning-based nodule detection system concerning influence of location of pulmonary nodule |
| title_fullStr | The diagnostic performance and clinical value of deep learning-based nodule detection system concerning influence of location of pulmonary nodule |
| title_full_unstemmed | The diagnostic performance and clinical value of deep learning-based nodule detection system concerning influence of location of pulmonary nodule |
| title_short | The diagnostic performance and clinical value of deep learning-based nodule detection system concerning influence of location of pulmonary nodule |
| title_sort | diagnostic performance and clinical value of deep learning based nodule detection system concerning influence of location of pulmonary nodule |
| topic | Deep learning Chest radiography Solitary pulmonary nodule |
| url | https://doi.org/10.1186/s13244-023-01497-4 |
| work_keys_str_mv | AT seulgiyou thediagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT jihyunpark thediagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT bumheepark thediagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT hanbitshin thediagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT taeyangha thediagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT jaesungyun thediagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT kyoungjoopark thediagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT yongjunjung thediagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT younakim thediagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT minjikim thediagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT joosungsun thediagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT seulgiyou diagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT jihyunpark diagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT bumheepark diagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT hanbitshin diagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT taeyangha diagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT jaesungyun diagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT kyoungjoopark diagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT yongjunjung diagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT younakim diagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT minjikim diagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule AT joosungsun diagnosticperformanceandclinicalvalueofdeeplearningbasednoduledetectionsystemconcerninginfluenceoflocationofpulmonarynodule |