Cervical cancer is a preventable disease. Early detection and treatment strategies have reduced the rate of cervical cancer by 65% in high-income countries. However, cervical cancer remains the first or second leading cause of cancer in low- and middle-income countries (LMICs) with more than 85% of cervical cancer death occuring in LMICs. A lack of resources in these settings has made it difficult to implement effective cervical cancer screening, diagnosis, and preventative treatment. The goal of my thesis work was to develop low-cost technologies that can improve early detection and prevention of cervical cancer in low-resource settings. Currently, the standard method of cervical cancer screening and diagnosis in many LMICs is visual inspection with acetic acid (VIA). During VIA, 2-5% acetic acid is applied on the cervix, which causes precancerous and cancerous lesions to turn white. While VIA has become a low-cost method of cervical cancer detection, it is associated with a low specificity (high false-positive rate), leading to the overtreatment of women and a misallocation of limited health care resources. At the same time colposcopy, which is standard for detecting high-grade cervical precancer and cancer in high-income countries, also has a low specificity. The first part of this thesis work focused on the evaluation of high-resolution microendoscopy (HRME) as a tool to increase the specificity for the detection of cervical precancer at the point-of-care. These evaluations were completed in medically-underserved areas of Texas as well as in El Salvador, where rates of cervical cancer are elevated. In Texas, I conducted a study to compare the performance of colposcopy alone to that of colposcopy with HRME imaging in a population of screen-positive women referred for follow-up testing. In El Salvador, I conducted a study to compare two screening strategies (VIA and HPV DNA testing) and three follow-up strategies (VIA alone, colposcopy alone and VIA/colposcopy with HRME imaging). In both studies, I found that including HRME with VIA or colposcopy significantly increased the specificity for detecting high-grade precancer and cancer. In the Texas study, adding HRME had a significantly higher specificity (56%) than traditional colposcopy (45%) on a per-site basis, p=0.01. Similarly, in El Salvador the specificity when adding HRME (62%, 69%) was significantly higher than VIA (51%, 58%) alone and colposcopy alone (53%, 59%) on a per-patient and per-site basis, p<0.001. I also worked to develop a new HRME system that was more portable and less expensive than the original system. This was accomplished by replacing the computer tablet and CCD camera sensor with a Raspberry Pi single-board computer and Raspberry Pi CMOS camera, reducing the overall cost from $2,400 to $1,200. Similar technology was also used to develop a low-cost lateral flow assay reader that can automatically read diagnostic tests at the point-of-care. In many underserved areas, there are not enough trained clinicians to screen and provide preventative treatment for cervical cancer, in part due to a lack of education and training resources. To address this need, I led a team of students to create the Low-cost Universal Cervical cancer Instructional Apparatus (LUCIA), a low-cost cervical cancer training model ($47) that can be used in low-resource areas to train medical providers how to perform a variety of skills for cervical cancer prevention, including VIA, Pap and HPV sample collection, colposcopy, endocervical curettage, cervical biopsy, cryotherapy, and loop electrosurgical excision procedure. This model has since been used to provide hands-on training in courses held in Texas, El Salvador, and Mozambique, with average feedback being 4.12/5 for usefulness, 4.46/5 for skill improvement, and 4.43/5 for ease of skill evaluation.
Overall, the development of these technologies has begun to build a more comprehensive framework to better provide cervical cancer prevention services in low-resource settings.