Multi Parametric Image Analysis and Correlation With Outcomes in Lung Cancer Screening and Early Stage Lung Cancer

Summary

Determine whether CT-based multiparametric analytical models may improve prediction of biopsy and treatment outcome in patients undergoing screening CT scan and/or treatment for early stage lung cancer

Detailed description

The hypothesis is that multiparametric models that incorporate complex image information from screening CT scans will improve prediction of the outcome of subsequent lung biopsy, an invasive diagnostic procedure. In this project, we will construct an image feature-based multiparametric prognostic model for biopsy outcome from screening lung CT scans performed at our institution, and then validate it using theNLST imaging and clinical outcomes dataset. This study involves no treatment or invasive procedures. Investigator will review all charts of patients who were treated for early stage lung cancer with definitive radiation therapy at UTSW or Parkland Memorial hospital, diagnosed with a malignancy from January 1, 2004 to October 31, 2014, to compile demographic, diagnostic, therapeutic, outcome, and toxicity data. Investigator expect that this will include approximately 200 patient charts. This data will be analyzed statistically and used for future directed research. Investigator will also analyze an anonymized dataset of patients from the National Lung Cancer Screening Trial (NLST) provided by the National Cancer Institute (NCI)

Arms & interventions

• OtherRetrospective Study

  The medical charts are the subjects. The institutional charts will be identified by the use of definitive radiation therapy correlating with an early stage lung cancer diagnosis during the above time frame. The data from these charts will be entered into a password protected excel spreadsheet. The charts will be identified by name, medical record number, date of birth, and social security number. These are all patients treated by all hospitals and clinics affiliated with UTSW and Parkland. At the time of study, some of the patients will have expired but some will be alive and in the regional North Texas area. Thus, given the minimal risk nature of this retrospective chart review, we could not reasonably conduct this research with a full waiver of consent. The NLST external dataset is proved by the NCI, with no identifying characteristics.

Outcome measures

Eligibility criteria