Virtual Library
Start Your Search
Dawei Yang
Author of
-
+
P3.04 - Clinical Design, Statistics and Clinical Trials (ID 720)
- Event: WCLC 2017
- Type: Poster Session with Presenters Present
- Track: Clinical Design, Statistics and Clinical Trials
- Presentations: 1
- Moderators:
- Coordinates: 10/18/2017, 09:30 - 16:00, Exhibit Hall (Hall B + C)
-
+
P3.04-001 - Evaluate the Utility of the ProLung China Test in the Diagnosis of Lung Cancer (ID 9541)
09:30 - 09:30 | Presenting Author(s): Dawei Yang
- Abstract
Background:
This Study will assess the stability of the ProLung China Test classification algorithm when used as an adjunct to CT scan. Also, we will assess whether there are any potential safety concerns of the ProLung China Test when used to evaluate patients with a positive CT scan for lung cancer.
Method:
This study is a multicentre, prospective, open, self-control study which aims to evaluate the utility of the ProLung China Test in diagnosis of lung cancer (ClinicalTrials.gov ID: NCT02726633 ). The subject whose age is between 18 and 80 years old and CT result shows a 4 ~ 50 mm nodule within 30 days is our object. In these objects, we will exclude those people who has TB, pulmonary edema, chronic lung infection, abnormal anatomy, skin disease effecting bioconductance and other tumors.The expected sensitivity and specificity of Prolung China Test are 70 % and 61%, and non-inferiority margin is 10 %. Based on these statistical information, four clinical trial centers, in the study, will enroll at least 452 samples with 20 % dropout rate. These samples must contain at least 182 effective malignant sample and 194 benign samples.
Result:
According to the inclusion and exclusion criteria, excision biopsy or follow-up examination will perform on enrolled subjects. Before these examination, a Prolung China test will be operated on these subjects. The subject in follow-up will be followed at least 24 months. The pathology result and follow-up result will be the gold standard in this study. The diagnosis result and adverse event will be recorded during the experiment.
Conclusion:
To demonstrate safety and efficacy of the ProLung China Test in the risk stratification of patients with pulmonary lesions identified by CT that are suspicious for lung cancer.
-
+
P3.13 - Radiology/Staging/Screening (ID 729)
- Event: WCLC 2017
- Type: Poster Session with Presenters Present
- Track: Radiology/Staging/Screening
- Presentations: 1
- Moderators:
- Coordinates: 10/18/2017, 09:30 - 16:00, Exhibit Hall (Hall B + C)
-
+
P3.13-037 - Deep Learning System for Lung Nodule Detection (ID 9503)
09:30 - 09:30 | Presenting Author(s): Dawei Yang
- Abstract
Background:
Ever since its first success in large scale image recognition problem, deep convolutional neural networks (DCNNs) have shown their capabilities in solving many challenging visual perceptual tasks, such as image classification, segmentation, object detection. In some cases, DCNNs have already achieved near-human performance. In medical image analysis, DCNNs have also been successfully applied to lesion detection, segmentation, and diagnosis. The power of DCNN lies in its ability to learn a hierarchical representation of raw input data, without hand-crafted features.
Method:
The focus of this study was to improve the performance of DCNN in automatic detection of pulmonary nodule on CT scan. In particular, nodules whose size is less than 4mm were considered. A total of 171 scans were collected at Zhongshan Hospital Fudan University, which was first process by the proposed DCNN system (12Sigma). The detection results were then carefully reviewed by an experienced physician, all false positives and missed nodules were manually labeled. To refine the system, all false positives and true nodules of sizes 4mm and under were selected.
Result:
The data set was randomly split into training and test sets, where the training set consists of 90% (154 scans) and the test set consists of 10% of data (17 scans). Figure 1 shows the FROC curves on the test set before and after re-training. Overall, the detection sensitivity were improved at all false positive levels, but the improvement was most significant at low FP rate region. For example, when FP is 0.5 per scan, the detection sensitivity increased from 0.36 to 0.49. Figure 1
Conclusion:
This improvement suggests that, even with limited data, the deep neural network can learn from its mistakes and be easily tuned to be more sensitive to small nodules. We believe that when more data is colllected, more significant improve in high FP rate region will be observed.
