|
Lung cancer is the leading cause of cancer death in both men and women in the United States. The disease has been closely associated with smoking since 1964, when the Surgeon General concluded that tobacco smoke was a cause of lung cancer. Today, smoking is thought to cause up to 80-90 percent of lung cancer cases. Many programs have been successful in limiting smoking (e.g. smoking cessation programs, laws and policies that limit exposure to second hand smoke, etc.). However, lung cancer 5 year survival rates have not improved dramatically in the past several decades. The five-year survival rate for lung cancer currently stands at 15.6 percent as compared to an over 90 percent survival rate for breast, colon and prostate cancers.
In October 2010, the National Lung Screening Trial (NLST) announced it had demonstrated a significant reduction in death rates from lung cancer (and death from any cause) when performing screening using low dose CT. More than 2 years later and despite the success of this trial and the endorsement of many agencies recommending the use of CT in screening in high risk populations (The American Cancer Society, The American Lung Association, The American College of Chest Physicians, the American Society of Clinical Oncology, The American Association for Thoracic Surgery and the National Comprehensive Cancer Network), there has not been widespread adoption of CT screening for lung cancer.
While several reasons for this lack of adoption have been identified, one recurrent reason given is the concern over radiation exposure to a large screening population, even with low dose CT. The scans performed in the NLST were considered to be “low dose” because they used an appreciably reduced scanner output (compared to conventional diagnostic CT scans) to carry out these screening studies. Since the NLST was initiated in 2002, there have been tremendous advances in CT scanner technology, specifically with respect to radiation dose reduction. These include the introduction of automatic exposure controls which adapt the scanner output to the size of the patient (reducing it for small patients, etc.) and novel image reconstruction methods that exploit advances in computing power to reduce image noise and allow image quality to be maintained even when very low doses are used. These developments encourage us that lung cancer screening may be accomplished using CT at doses that are significantly lower than the “low dose” scans used in the NLST.
Therefore, there is a critical need to develop Ultra Low Dose CT scanning methods that bring the radiation dose down to a level close to that of Chest X-rays while maintaining the image quality necessary for lung cancer screening. The purpose of this project is to investigate both methods to perform CT exams at reduced dose levels close to those of Chest X-rays and to evaluate the ability to identify and characterize lung nodules in a screening setting under these ultra low dose scanning conditions. This requires an innovative approach which brings together imaging physics, image reconstruction methods and computer vision techniques with specific emphasis on detection of lung lesions and their boundaries.
To accomplish this, we propose two specific aims:
(1) We will create a collection of simulated ultra low dose CT scans of patients using a combination of imaging physics and innovative image reconstruction methods. This collection will be created from data acquired from scans performed for the NLST as well as scans acquired as part of current clinical practice.
(2) We will evaluate the effects of reduced dose images on the relative performance in the both the detection of lung lesions and in the delineation of lesion boundaries to assess lesion size.
The results of this research will allow us to determine the effectiveness of ultra low dose CT scans and may allow lung cancer screening using CT to be deployed on a wide basis with reduced concerns about radiation dose. | 