Virtual Library
Start Your Search
Yihui Du
Author of
-
+
MA05 - Lung Cancer Screening (ID 174)
- Event: WCLC 2020
- Type: Mini Oral
- Track: Screening and Early Detection
- Presentations: 1
- Moderators:
- Coordinates: 1/30/2021, 11:45 - 12:45, Scientific Program Auditorium
-
+
MA05.11 - Effect of Lowering the Starting Age for Lung Cancer Screening by Low-Dose Computed Tomography Among Women: A Harm-Benefit Analysis (ID 1399)
11:45 - 12:45 | Presenting Author(s): Yihui Du
- Abstract
- Presentation
Introduction
Studies have shown that it is more cost-effective to screen women at a younger age. However, the estimated lifetime risk of developing radiation-induced lung cancers is greater for women than for men, especially when exposure occurs at younger ages. This study aimed to evaluate the effect of lowering the age at which women could undergo annual lung cancer screening by low-dose computed tomography (LDCT) on the rates of spontaneous and induced lung cancers.
Methods
A validated micro-simulation model was used in a simulated cohort of heavy female smokers aged 45–75 years undergoing annual lung cancer screening by LDCT. The tumor induction was developed based on an excess relative risk model. The outcomes were the number of screen-detected and radiation-induced lung cancers per 1000 screened for scenarios with the starting age decreasing from 55 to 45 years. A benefit-harm ratio was calculated by dividing the number of screen-detected lung cancers by the number of radiation-induced lung cancers. An incremental harm-benefit ratio was calculated by dividing the incremental radiation-induced lung cancers by the incremental screen-detected lung cancers relative to a starting age of 55 years.
Results
Decreasing the screening starting age from 55 to 45 years led to increases in the number of screen-detected and radiation-induced lung cancers from 98.86 to 112.92 and from 0.73 to 2.32 per 1000 screened, respectively. Given a lower starting age between 55 and 45 years, the benefit-harm ratio ranged from 136 to 49 and the incremental harm-benefit ratio ranged from 0.046 to 0.113. When applying a lower radiation dose of 0.4 mSv, the benefit-harm ratio of implementing screening at 45 years increased from 49 to 423 and the IHBR relative to age 55 years decreased from 0.113 to 0.017.
Table 1. The effect of lowering the starting age for annual lung cancer screening by LDCT among female heavy smokers
Scenario*
Screen-detected LC cases
(n/1000 screened)
Radiation-induced LC cases
(n/1000 screened)
Benefit-harm ratio
IHBR
A-55-75
98.86
0.73
136
-
A-54-75
101.39
0.84
120
0.046
A-53-75
103.60
0.97
106
0.052
A-52-75
105.45
1.10
96
0.057
A-51-75
107.02
1.24
86
0.063
A-50-75
108.35
1.39
78
0.070
A-49-75
109.66
1.54
71
0.076
A-48-75
110.63
1.72
64
0.084
A-47-75
111.49
1.88
59
0.091
A-46-75
112.30
2.11
53
0.103
A-45-75
112.92
2.32
49
0.113
Abbreviations: IHBR: incremental harm-benefit ratio relative to scenario A-55-75; LC, lung cancer; LDCT, low-dose computed tomography. *: Screening interval (A-annual) – screening start age – screening stop age.
Conclusion
Lowering the starting age by 10 years for women undergoing annual lung cancer screening by LDCT increased the number of spontaneous lung cancers detected by 14 at the expense of 1.6 lung cancers induced per 1000 screened. Applying an ultra-low-dose CT methodology to lung cancer screening may be a viable option in attempts to reduce the increased risk of harm from radiation exposure when lowering the starting age for screening.
Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.
-
+
P42 - Screening and Early Detection - Risk Modelling and Artificial Intelligence (ID 177)
- Event: WCLC 2020
- Type: Posters
- Track: Screening and Early Detection
- Presentations: 1
- Moderators:
- Coordinates: 1/28/2021, 00:00 - 00:00, ePoster Hall
-
+
P42.02 - Evaluating the Feasibility of a Deep Learning-Based Computer-Aided Detection System for Lung Nodule Detection in a Lung Cancer Screening Program (ID 739)
00:00 - 00:00 | Author(s): Yihui Du
- Abstract
Introduction
Deep learning techniques have recently achieved remarkable results in automated lung nodule detection. Previously, we developed a deep learning-based computer-aided detection (DL-CAD) system based on the CT scans from the United States. The purpose of this study was to validate its effectiveness for automatic pulmonary nodule detection in an external validation set from a lung cancer screening program.
Methods
The proprietary DL-CAD system was pre-trained on the public dataset, LIDC-IDRI, which was collected from seven academic centers resulting in the AI prototype software MIPNOD 1.0. We retrospectively collected 2,127 low-dose CT scans from a Chinese lung cancer screening project between July and December 2017. Three hundred sixty scans were used in this study and they were evaluated independently by radiologists in a double reading fashion and the DL-CAD system. An extra senior radiologist checked all the results and made the consensus as to the reference standard. Free-response Receiver operating characteristic analysis was applied to assess the detection performance of the DL-CAD system.
Results
After making the consensus, there were 260 nodules in 196 participants and 164 participants without nodules. The DL-CAD system achieved a sensitivity of 90.0% with one false positive per scan, while radiologists had a sensitivity of 76.5% for detection during double reading. The performance comparison between the DL-CAD system and radiologists in nodule types and Lung-RADS categories was as follows. (1) solid nodules: 90.2% vs 77.2%, P = 0.007; part-solid nodules: 95.2% vs 81.0%, P = 0.810; non-solid nodules: 87.3% vs 72.7%, P = 0.134. (2) Lung-RADS 2: 86.0% vs 66.2%, P < 0.001; Lung-RADS 3: 94.3% vs 90.0%, P = 0.549; Lung-RADS 4: 100.0% vs 97.0%, P = 1.000.
Conclusion
The deep learning-based CAD system showed good performance for automatic pulmonary nodule detection on an external dataset and could provide assistance for radiologists in lung cancer screening programs.
-
+
P45 - Screening and Early Detection - Radiological Risk Stratification (ID 182)
- Event: WCLC 2020
- Type: Posters
- Track: Screening and Early Detection
- Presentations: 1
- Moderators:
- Coordinates: 1/28/2021, 00:00 - 00:00, ePoster Hall
-
+
P45.03 - Lung Nodule Management Based on Diameter and Volume in Lung Cancer Screening with Low-Dose Computed Tomography (ID 3242)
00:00 - 00:00 | Presenting Author(s): Yihui Du
- Abstract
Introduction
The Netherlands and China Big 3 diseases (NELCIN-B3) project is ongoing. One of the objectives is to evaluate the performance of the diameter-based and volume-based management of lung nodules in Chinese lung cancer screening setting. The aim of this study was to report the interim findings regarding lung nodule management using diameter-based and volume-based protocols in the first screening round.
Methods
Lung cancer screening with low-dose computed tomography (LDCT) was conducted in a Chinese asymptomatic population aged 40-74 years. In total, 1000 consecutive baseline LDCT scans were included and were independently read twice. The diameter of lung nodules was manually measured in the first reading and both diameter and volume were measured semi-automatically in the second reading. The percentage of lung nodules required short-term LDCT scan or further work-up according to the recommendations of diameter-based protocol (diameter ≥ 6 mm) and volume-based protocol (volume ≥ 100 mm3) was compared. The disagreement in nodules classification between the two protocols was evaluated using McNemar's test.
Results
In total, 96 lung nodules in 79 participants were measured in both readings. The median diameter of the nodules in manual and semi-automatic measurements was 5.0 (4.0-7.0) mm and 6.0 (5.0-8.0) mm, respectively. The median volume of the nodules was 100.0 (61.0-184.0) mm3. The percentage of lung nodules required short-term LDCT scan and further work-up using the diameter-based protocol with manual and semi-automatic measurements, and using the volume-based protocol was 37.5%, 65.6% and 46.9%, respectively. The disagreement in nodules classification between diameter-based protocol using manual measurement and volume-based protocol was observed for 15 (15.6%) lung nodules (McNemar's χ2 = 4.27, P = 0.039). The disagreement in nodules classification between diameter-based protocol using semi-automatic measurement and volume-based protocol was observed for 20 (20.8%) lung nodules (McNemar's χ2 = 14.45, P < 0.001) (Table 1).
Table 1 Agreement in lung nodules classification for the diameter-based management using manual measurement and volume-based management (n=96)
Nodule management based on diameter
Nodule management based on volume
Annual LDCT screening
Short-term LDCT scan or further work-up
Total
Based on manual diameter measurement
Annual LDCT screening
48
12
60
Short-term LDCT scan or further work-up
3
33
36
Total
51
45
96
Based on semi-automatic diameter measurement
Annual LDCT screening
32
1
33
Short-term LDCT scan or further work-up
19
44
63
Total
51
45
96
Conclusion
Compared to the diameter-based protocol using manual measurements, the volume-based protocol for lung nodule management is associated with a higher referral rate to short-term LDCT scan and further work-up. However, compared to the diameter-based protocol using semi-automatic measurement, the volume-based protocol is associated with a lower referral rate.
