Ruchi Deshpande MS, Presenter: Nothing to Disclose

John J. Demarco PhD, Abstract Co-Author: Nothing to Disclose

Brent Julius Liu PhD, Abstract Co-Author: Nothing to Disclose

We have developed a decision support system for Radiation Therapy that utilizes a comprehensive DICOM RT specific database of retrospective treatment planning data to perform data mining. These data mining results may corroborate a clinician’s radiation therapy treatment plan, thus increasing confidence or they may highlight potential ways to improve the treatment plan by lowering dose to sensitive organs surrounding the tumor. Since the success of such systems depends heavily on the size and composition of the training database, we have also developed an infrastructure to facilitate data collection for cloud-based research collaborations.

The infrastructure and decision support algorithm have both been tested and evaluated with 51 sets of retrospective treatment planning data of head and neck cancer patients. An expert has tested and validated the integrity of our client-side JavaScript DICOM parser and anonymizer. An expert has also verified HIPAA compliance of our data collection mechanism.

An efficient data collection mechanism is essential to ensure a constantly growing and self-updating big data repository for good system performance. DICOM compatibility ensures vendor neutrality, and a HIPAA-compliant data sharing protocol encourages research participation. This provides the potential for a large-scale cancer registry containing vital anatomical, dosimetric and treatment planning information, which is harnessed by a data mining decision support algorithm in a cloud-based environment. The cloud-based infrastructure promotes big data analytics in the field of Radiation Oncology, by paving the way for building data warehouses, dealing with HIPAA regulations and providing a platform for testing data mining algorithms.

We will present an infrastructure for facilitating large-scale collection of radiation therapy treatment planning data, and demonstrate the benefits of a data mining decision support algorithm that utilizes this infrastructure. The various components of the system, including a client-side DICOM parser, a HIPAA-compliant data sharing protocol, a cloud-based data mining and analytics engine, and a DICOM specific database will be presented.

Deshpande, R, Demarco, J, Liu, B, Moving Towards Big Data Analytics in Radiation Therapy: Dynamic Decision Support through Data Mining.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14011416.html Accessed March 14, 2025