1
|
Michelin B, Labani A, Bilbault P, Roy C, Ohana M. Potential added value of an AI software with prediction of malignancy for the management of incidental lung nodules. RESEARCH IN DIAGNOSTIC AND INTERVENTIONAL IMAGING 2023; 8:100031. [PMID: 39076687 PMCID: PMC11265191 DOI: 10.1016/j.redii.2023.100031] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Subscribe] [Scholar Register] [Received: 08/04/2022] [Accepted: 08/09/2023] [Indexed: 07/31/2024]
Abstract
Purpose To determine the impact of an artificial intelligence software predicting malignancy in the management of incidentally discovered lung nodules. Materials and methods In this retrospective study, all lung nodules ≥ 6 mm and ≤ 30 mm incidentally discovered on emergency CT scans performed between June 1, 2017 and December 31, 2017 were assessed. Artificial intelligence software using deep learning algorithms was applied to determine their likelihood of malignancy: most likely benign (AI score < 50%), undetermined (AI score 50-75%) or probably malignant (AI score > 75%). Predictions were compared to two-year follow-up and Brock's model. Results Ninety incidental pulmonary nodules in 83 patients were retrospectively included. 36 nodules were benign, 13 were malignant and 41 remained indeterminate at 2 years follow-up.AI analysis was possible for 81/90 nodules. The 34 benign nodules had an AI score between 0.02% and 96.73% (mean = 48.05 ± 37.32), while the 11 malignant nodules had an AI score between 82.89% and 100% (mean = 93.9 ± 2.3). The diagnostic performance of the AI software for positive diagnosis of malignant nodules using a 75% malignancy threshold was: sensitivity = 100% [95% CI 72%-100%]; specificity = 55.8% [38-73]; PPV = 42.3% [23-63]; NPV = 100% [82-100]. With its apparent high NPV, the addition of an AI score to the initial CT could have avoided a guidelines-recommended follow-up in 50% of the benign pulmonary nodules (6/12 nodules). Conclusion Artificial intelligence software using deep learning algorithms presents a strong NPV (100%, with a 95% CI 82-100), suggesting potential use for reducing the need for follow-up of nodules categorized as benign.
Collapse
Affiliation(s)
- Bastien Michelin
- Department of Diagnostic Imaging (Radio B), Hôpitaux universitaires de Strasbourg, Strasbourg 67000, France
| | - Aïssam Labani
- Department of Diagnostic Imaging (Radio B), Hôpitaux universitaires de Strasbourg, Strasbourg 67000, France
| | - Pascal Bilbault
- Emergency Department, Hpitaux universitaires de Strasbourg, Strasbourg 67000, France
| | - Catherine Roy
- Department of Diagnostic Imaging (Radio B), Hôpitaux universitaires de Strasbourg, Strasbourg 67000, France
| | - Mickaël Ohana
- Department of Diagnostic Imaging (Radio B), Hôpitaux universitaires de Strasbourg, Strasbourg 67000, France
| |
Collapse
|
2
|
Bintein F, Yannoutsos A, Chatellier G, Fontaine M, Damotte D, Paterlini-Bréchot P, Meyer G, Duchatelle V, Marini V, Schwering KL, Labrousse C, Beaussier H, Zins M, Salmeron S, Lajonchère JP, Priollet P, Emmerich J, Trédaniel J. Patients with atherosclerotic peripheral arterial disease have a high risk of lung cancer: Systematic review and meta-analysis of literature. JOURNAL DE MÉDECINE VASCULAIRE 2021; 46:53-65. [PMID: 33752847 DOI: 10.1016/j.jdmv.2020.12.005] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 12/23/2020] [Accepted: 12/23/2020] [Indexed: 12/24/2022]
Abstract
PURPOSE Lung cancer and atherosclerosis share common risk factors. Literature data suggest that the prevalence of lung malignancy in patients with peripheral arterial disease (PAD) is higher than in the general population. Our goal was to determine, through a systematic literature review, the prevalence of lung cancer in patients with PAD. METHODS We consulted available publications in the Cochrane library, MEDLINE, PUBMED, EMBASE, and ClinicalTrials.gov. We included all articles, written in English or French, published between 1990 and 2020 reporting the prevalence of lung cancer in patients with PAD (atherosclerotic aortic aneurysm or peripheral occlusive diseases). Patients with coronary artery disease, cardiac valvulopathy or carotid stenosis were not included. We did not include case reports. We performed a critical analysis of each article. Data were collected from two independent readers. A fixed effect model meta-analysis allowed to estimate a summary prevalence rate. RESULTS We identified 303 articles, and selected 19 articles according to selection criteria. A total of 16849 patients were included (mean age 68.3 years, 75.1% of males). Aortic aneurysms were found in 29% of patients and atherosclerotic occlusive disease in 66% of patients. Lung cancer was identified in 538 patients, representing a prevalence of 3%. DISCUSSION Lung cancer is found in 3% of patients with atherosclerotic PAD. This prevalence is higher than that found in lung cancer screening programs performed in the general population of smokers and former smokers. These patients should be screened for lung cancer. Their selection may dramatically increase the benefit of lung cancer screening.
Collapse
Affiliation(s)
- F Bintein
- Groupe hospitalier Paris Saint-Joseph, 75014 Paris, France.
| | - A Yannoutsos
- Groupe hospitalier Paris Saint-Joseph, 75014 Paris, France; Inserm UMR 1153 Center of Research in Epidemiology and Statistics, Université Paris Descartes, Sorbonne Paris Cité, Paris, France
| | - G Chatellier
- Hôpital européen Georges-Pompidou, AP-HP, 75015 Paris, France; Université Paris Descartes, Sorbonne Paris Cité, Paris, France
| | | | - D Damotte
- Hôpital Cochin, AP-HP, 75014 Paris, France; Université Paris Descartes, Sorbonne Paris Cité, Paris, France; Unité Inserm U1138, centre de recherche des Cordeliers, Paris, France
| | | | - G Meyer
- Hôpital européen Georges-Pompidou, AP-HP, 75015 Paris, France; Université Paris Descartes, Sorbonne Paris Cité, Paris, France
| | - V Duchatelle
- Groupe hospitalier Paris Saint-Joseph, 75014 Paris, France
| | - V Marini
- Groupe hospitalier Paris Saint-Joseph, 75014 Paris, France
| | | | - C Labrousse
- Groupe hospitalier Paris Saint-Joseph, 75014 Paris, France
| | - H Beaussier
- Groupe hospitalier Paris Saint-Joseph, 75014 Paris, France
| | - M Zins
- Groupe hospitalier Paris Saint-Joseph, 75014 Paris, France
| | - S Salmeron
- Groupe hospitalier Paris Saint-Joseph, 75014 Paris, France
| | - J-P Lajonchère
- Groupe hospitalier Paris Saint-Joseph, 75014 Paris, France
| | - P Priollet
- Groupe hospitalier Paris Saint-Joseph, 75014 Paris, France
| | - J Emmerich
- Groupe hospitalier Paris Saint-Joseph, 75014 Paris, France; Inserm UMR 1153 Center of Research in Epidemiology and Statistics, Université Paris Descartes, Sorbonne Paris Cité, Paris, France; Université Paris Descartes, Sorbonne Paris Cité, Paris, France
| | - J Trédaniel
- Hôpital européen Georges-Pompidou, AP-HP, 75015 Paris, France; Université Paris Descartes, Sorbonne Paris Cité, Paris, France; Unité Inserm UMR-S 1124, toxicologie, pharmacologie et signalisation cellulaire, Paris, France
| |
Collapse
|
3
|
Chen G, Xue Y, Wei J, Duan Q. The undiagnosed potential clinically significant incidental findings of neck CTA: A large retrospective single-center study. Medicine (Baltimore) 2020; 99:e22440. [PMID: 33120738 PMCID: PMC7581090 DOI: 10.1097/md.0000000000022440] [Citation(s) in RCA: 5] [Impact Index Per Article: 1.3] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 12/21/2022] Open
Abstract
To assess the prevalence and missed reporting rate of potential clinically-significant incidental findings (IFs) in the neck CTA scans.All consecutive patients undergoing neck CTA imaging, from January 1, 2017 to December 31, 2018, were retrospectively evaluated by a radiologist for the presence of incidental findings in the upper chest, lower head and neck regions. These incidental findings were subsequently classified into 3 categories in terms of clinical significance: Type I, highly significant, Type II, moderately significant; and Type III, mildly or not significant. Type I and Type II IFs were determined as potential clinically significant ones and were retrospectively analyzed by another 2 radiologists in consensus. The undiagnosed findings were designated as those that were not reported by the initial radiologists. The differences in the rate of unreported potential clinically significant IFs were compared between the chest group and head or neck group.A total of 376 potential clinically significant IFs were detected in 1,698 (91.19%) patients, of which 175 IFs were classified as highly significant findings (Type I), and 201 (53.46%) as moderately significant findings (Type II). The most common potential clinically significant findings included thyroid nodules (n = 88, 23.40%), pulmonary nodules (n = 56, 14.89%), sinus disease (n = 39, 10.37%), intracranial or cervical artery aneurysms (n = 30, 7.98%), enlarged lymph nodes (n = 24, 6.38%), and pulmonary embolism (n = 19, 5.05%). In addition, 184 (48.94%) of them were not mentioned in the initial report. The highest incidence of missed potential clinical findings were pulmonary embolism and pathologic fractures and erosions (100% for both). The unreported rate of the chest group was significantly higher than that of the head or neck one, regardless of Type I, Type II or all potential clinically significant IFs (χ = 32.151, χ = 31.211, χ = 65.286, respectively; P < .001 for all).Important clinically significant incidental findings are commonly found in a proportion of patients undergoing neck CTA, in which nearly half of these patients have had potential clinically significant IFs not diagnosed in the initial report. Therefore, radiologists should beware of the importance of and the necessity to identify incidental findings in neck CTA scans.
Collapse
Affiliation(s)
- Guangliang Chen
- Department of Radiology, Fujian Medical University Union Hospital
- School of Medical Technology and Engineering, Fujian Medical University, University Town, Fuzhou, China
| | - Yunjing Xue
- Department of Radiology, Fujian Medical University Union Hospital
- School of Medical Technology and Engineering, Fujian Medical University, University Town, Fuzhou, China
| | - Jin Wei
- Department of Radiology, Fujian Medical University Union Hospital
- School of Medical Technology and Engineering, Fujian Medical University, University Town, Fuzhou, China
| | - Qing Duan
- Department of Radiology, Fujian Medical University Union Hospital
- School of Medical Technology and Engineering, Fujian Medical University, University Town, Fuzhou, China
| |
Collapse
|
4
|
Mazzei MA, Guerrini S, Gentili F, Galzerano G, Setacci F, Benevento D, Mazzei FG, Volterrani L, Setacci C. Incidental extravascular findings in computed tomographic angiography for planning or monitoring endovascular aortic aneurysm repair: Smoker patients, increased lung cancer prevalence? World J Radiol 2017; 9:304-311. [PMID: 28794826 PMCID: PMC5529319 DOI: 10.4329/wjr.v9.i7.304] [Citation(s) in RCA: 3] [Impact Index Per Article: 0.4] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 12/27/2016] [Revised: 04/16/2017] [Accepted: 05/31/2017] [Indexed: 02/06/2023] Open
Abstract
AIM To validate the feasibility of high resolution computed tomography (HRCT) of the lung prior to computed tomography angiography (CTA) in assessing incidental thoracic findings during endovascular aortic aneurysm repair (EVAR) planning or follow-up.
METHODS We conducted a retrospective study among 181 patients (143 men, mean age 71 years, range 50-94) referred to our centre for CTA EVAR planning or follow-up. HRCT and CTA were performed before or after 1 or 12 mo respectively to EVAR in all patients. All HRCT examinations were reviewed by two radiologists with 15 and 8 years’ experience in thoracic imaging. The results were compared with histology, bronchoscopy or follow-up HRCT in 12, 8 and 82 nodules respectively.
RESULTS There were a total of 102 suspected nodules in 92 HRCT examinations, with a mean of 1.79 nodules per patient and an average diameter of 9.2 mm (range 4-56 mm). Eighty-nine out of 181 HRCTs resulted negative for the presence of suspected nodules with a mean smoking history of 10 pack-years (p-y, range 5-18 p-y). Eighty-two out of 102 (76.4%) of the nodules met criteria for computed tomography follow-up, to exclude the malignant evolution. Of the remaining 20 nodules, 10 out of 20 (50%) nodules, suspected for malignancy, underwent biopsy and then surgical intervention that confirmed the neoplastic nature: 4 (20%) adenocarcinomas, 4 (20%) squamous cell carcinomas, 1 (5%) small cell lung cancer and 1 (5%) breast cancer metastasis); 8 out of 20 (40%) underwent bronchoscopy (8 pneumonia) and 2 out of 20 (10%) underwent biopsy with the diagnosis of sarcoidosis.
CONCLUSION HRCT in EVAR planning and follow-up allows to correctly identify patients requiring additional treatments, especially in case of lung cancer.
Collapse
|
5
|
Karkhanis S, Patil S, Hoey ET, Ganeshan A. The prevalence and significance of thoracic findings in patients undergoing extended coverage computed tomography for assessment of abdominal aortic aneurysms. Br J Radiol 2016; 89:20150723. [PMID: 26987373 DOI: 10.1259/bjr.20150723] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/21/2022] Open
Abstract
OBJECTIVE Many major vascular centres, including ours, incorporate coverage of the thoracic aorta in CT scans evaluating abdominal aortic aneurysms (AAAs) (extended coverage contrast-enhanced CT (EC-CECT) scan]. We sought to determine the prevalence of thoracic pathology in a large consecutive series of patients with AAA undergoing EC-CECT. METHODS All patients who underwent EC-CECT for AAA between April 2013 and 2014 were identified from our radiology information system. Reports were retrospectively reviewed and for each study, sex, age and reported thoracic aortic and other non-vascular findings were extracted. Findings were initially categorized into "major" or "minor" depending on if they were mentioned in the report impression/conclusion. Any major thoracic pathology was termed "significant" if there was a new diagnosis/patient intervention/investigation and a change in patient management. RESULTS Of the 150 scans included in the study, 97 (65%) had at least one thoracic finding. These findings included 24 scans (16%) with major findings and 73 scans (48%) with minor findings. In 13 scans (9%), the findings were significant and resulted in a delay (n = 11) or cancellation (n = 2) of AAA repair. CONCLUSION Pre-procedural EC-CECT helps to identify significant intrathoracic findings in patients with AAA, which can have a major impact on AAA repair. This study supports the routine use of pre-procedural EC-CECT in planning AAA repair. ADVANCES IN KNOWLEDGE This study describes the prevalence of significant thoracic pathology, which can impact on AAA repair. This information could potentially change the pre-procedure imaging protocol for patients with AAA.
Collapse
Affiliation(s)
- Salil Karkhanis
- 1 Radiology Department, New Queen Elizabeth Hospital, Birmingham, UK
| | - Shahbaz Patil
- 1 Radiology Department, New Queen Elizabeth Hospital, Birmingham, UK
| | - Edward Td Hoey
- 2 Radiology Department, Birmingham Heartland Hospital, Birmingham, UK
| | - Arul Ganeshan
- 2 Radiology Department, Birmingham Heartland Hospital, Birmingham, UK
| |
Collapse
|
6
|
CT before transcatheter aortic valve replacement: Value of venous phase imaging for detection and interpretation of findings with impact on the TAVR procedure. J Cardiovasc Comput Tomogr 2015; 9:422-7. [DOI: 10.1016/j.jcct.2015.03.007] [Citation(s) in RCA: 6] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 02/28/2014] [Revised: 03/08/2015] [Accepted: 03/17/2015] [Indexed: 12/21/2022]
|
7
|
Callister MEJ, Baldwin DR, Akram AR, Barnard S, Cane P, Draffan J, Franks K, Gleeson F, Graham R, Malhotra P, Prokop M, Rodger K, Subesinghe M, Waller D, Woolhouse I. British Thoracic Society guidelines for the investigation and management of pulmonary nodules. Thorax 2015; 70 Suppl 2:ii1-ii54. [PMID: 26082159 DOI: 10.1136/thoraxjnl-2015-207168] [Citation(s) in RCA: 560] [Impact Index Per Article: 62.2] [Reference Citation Analysis] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/21/2022]
Affiliation(s)
- M E J Callister
- Department of Respiratory Medicine, Leeds Teaching Hospitals, Leeds, UK
| | - D R Baldwin
- Nottingham University Hospitals, Nottingham, UK
| | - A R Akram
- Royal Infirmary of Edinburgh, Edinburgh, UK
| | - S Barnard
- Department of Cardiothoracic Surgery, Freeman Hospital, Newcastle, UK
| | - P Cane
- Department of Histopathology, St Thomas' Hospital, London, UK
| | - J Draffan
- University Hospital of North Tees, Stockton on Tees, UK
| | - K Franks
- Clinical Oncology, St James's Institute of Oncology, Leeds, UK
| | - F Gleeson
- Department of Radiology, Oxford University Hospitals NHS Trust, Oxford, UK
| | | | - P Malhotra
- St Helens and Knowsley Teaching Hospitals NHS Trust, UK
| | - M Prokop
- Department of Radiology and Nuclear Medicine, Radboud University Medical Center, Nijmegen, Netherlands
| | - K Rodger
- Respiratory Medicine, St James's University Hospital, Leeds, UK
| | - M Subesinghe
- Department of Radiology, Churchill Hospital, Oxford, UK
| | - D Waller
- Department of Thoracic Surgery, Glenfield Hospital, Leicester, UK
| | - I Woolhouse
- Department of Respiratory Medicine, University Hospitals of Birmingham, Birmingham, UK
| | | | | |
Collapse
|
8
|
Systematic Approach to the Management of the Newly Found Nodule on Screening Computed Tomography. Thorac Surg Clin 2013; 23:141-52. [DOI: 10.1016/j.thorsurg.2013.01.007] [Citation(s) in RCA: 8] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/21/2022]
|