Poster Presentations: Wednesday, October 26, 2011 |

A Novel Shape-Based Diagnostic Approach for Early Diagnosis of Lung Nodules FREE TO VIEW

Muralidhar Kondapaneni, MBBS; Matthew Nitzken, MS; Eric Bogaert, MD; Georgy Gimel’farb, PhD; Robert Falk, MD; Mohamed Abou El-Ghar, MD; Ayman El-Baz, PhD
Chest. 2011;140(4_MeetingAbstracts):655A. doi:10.1378/chest.1119506
Text Size: A A A
Published online


PURPOSE: Pulmonary nodules are frequent incidental findings during chest computed tomography (CT). The goal of radiological evaluations for small pulmonary nodules is to differentiate benign from malignant lesions non-invasively and with the hope that early detection of lung cancer may improve mortality. However, diagnostic options for small malignant nodules are limited due to difficulties in their accessibility, especially if they are located deep in the tissue or away from large airways. The popular direction of detecting small cancerous nodules is to analyze their growth rate over time. However, this method has not been proven cost-effective or beneficial. Malignant nodules are known to have different shapes than benign nodules. Based on this well known fact, we propose a new novel shape based software analysis approach to differentiate malignant from benign nodules.

METHODS: The proposed analysis begins using lung nodules that are segmented by applying our previously described stochastic deformable model. After constructing a mesh of the nodule the shape can be described in terms of a linear combination of spherical harmonics (SHs). A shape index that is based on the number of SHs required to reconstruct a nodule quantifies the shape complexity of the detected nodule. We pilot tested this analysis framework on a database of clinical multislice 3D chest LDCT scans of 109 lung nodules with diameters ranging from 3mm to 30mm.

RESULTS: Based on biopsy these nodules were categorized as 51 malignant and 58 benign. Our approach resulted in correct classification of 47 out of 51 (92.1% accuracy) malignant and 55 out of 58 (94.8% accuracy) benign nodules. Traditional growth rate approach over one year classified only 29 out of 51 malignant (56.9% accuracy) and 51 out of 58 (84.4% accuracy) benign nodules.

CONCLUSIONS: The overall accuracy using the proposed 3D shape based analysis with a 95% confidence level is 94.4%.

CLINICAL IMPLICATIONS: The preliminary results justify further elaboration of the shape based analysis for diagnosing malignant nodules. The proposed nodule shape analysis could lead to more accurate and faster diagnoses of pulmonary nodules without the need of successive LDCT images.

DISCLOSURE: The following authors have nothing to disclose: Muralidhar Kondapaneni, Matthew Nitzken, Eric Bogaert, Georgy Gimel’farb, Robert Falk, Mohamed Abou El-Ghar, Ayman El-Baz

We will be discussing about a new shape based software analysis for diagnosing malignant lung nodules. This is a research project in process and not presently approved by FDA for clinical diagnostic purposes.

09:00 AM - 10:00 AM




Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

CHEST Journal Articles
PubMed Articles
Radiomics of pulmonary nodules and lung cancer. Transl Lung Cancer Res 2017;6(1):86-91.
Computerized Detection of Lung Nodules through Radiomics. Med Phys Published online May 11, 2017;
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543