Developing Automated tools for Cervical Cancer Radiotherapy Treatment for Zambian Women living with HIV (Biospecimen/Cohort)

University Of Tx Md Anderson Can Ctr, Houston TX

Investigators

Linked publications, trials & patents

Abstract

Project Summary The long-term goal of this research is to improve the management of cervical cancer across the lifespan for women living with HIV (WLH). The purpose of this proposal is to determine how cervical cancer treatment planning techniques can be adapted and automated for WLH to include bone marrow sparing and treatment of pelvic nodes—both which are valuable considerations for clinics with limited resources where the incidence of cervical cancer and HIV are high. The hypothesis is that at least 90% automated treatment plans created for WLH that include nodal boosting and bone marrow sparring are clinically acceptable. The first specific aim will be a retrospective planning study where automated methods for bone marrow sparing are validated on an existing biocohort dataset from Zambia. Ionizing radiation dose to marrow containing structures and adjacent organs will be studied and associated with number and location of enlarged nodes that are boosted to a high radiation dose using the adapted technique. The second specific aim will develop and validate deep learning models to autocontour enlarged nodes (from HIV-reactive adenopathy and malignancy). Differences in the sensitivity and accuracy of autocontouring between cohorts of WLH and women without HIV will be reported. The outcomes from this research will be a comprehensive understanding of how treatment planning techniques can be adapted for WLH that receive cervical radiotherapy. In addition, a suite of tools specifically designed for these scenarios described above will be integrated into a deployable FDA approved platform to increase access to this treatment adaptation for WLH and allow for future clinical studies of hematological toxicity and outcome for WLH.