Point-of-Care Ultraviolet Microscopy for Cervical Cancer Screening

This study aims to determine the image features of cervical cells, as measured via ultraviolet microscopy, that would constitute a positive screening and a negative screening result for cervical dysplasia, a precursor to cervical cancer, as measured against liquid-based Papanicolaou testing.

Our study population will consist of patients presenting to one of the three outpatient gynecology clinics within the Banner University Medical Center system for either a routine or follow up cervical cancer screening. Eligible subjects will be informed of the study and asked to consent to participate in accordance with standard informed consent protocol. The subject will undergo a Papanicolaou screening test, using both a spatula and a cytobrush and collected in a Thinprep specimen container according to the manufacturer's instructions. This specimen will be used for standard-of-care cytology. A second specimen will then be obtained in an identical fashion, labeled with a deidentified code unique to the subject, and held under refrigeration pending microscopy. The second specimen will be taken from the clinic for processing in preparation for UV microscopy. A sample of the specimen will be smeared onto a UV slide and images taken. The UV images will be then analyzed to extract the following features for determining the stage of cervical lesion: Cellularity: The density of cells in the sample. A high cellularity indicates an abnormal growth of cells. Nuclear features: The size and shape of the nuclei in the cells. Abnormal nuclear features, such as enlarged nuclei or irregular chromatin distribution, can indicate the presence of a cervical lesion. Cytoplasmic features: The appearance of the cytoplasm in the cells. Abnormal cytoplasmic features, such as vacuolation or hyperchromasia, can indicate the presence of a cervical lesion. Cellular architecture: The organization and arrangement of the cells in the sample. Abnormal cellular architecture, such as loss of polarity or crowding of cells, can indicate the presence of a cervical lesion. A second subject arm will consist of patients presenting to colposcopy clinic for colposcopy due to LSIL or HSIL findings on a screening Pap test. Subjects with test results of ASCUS or ASC-H will be excluded. A cervical cytology specimen will be taken from these subjects using both a spatula and a cytobrush, labeled with a deidentified code unique to the subject, and held under refrigeration pending microscopy in a manner identical to the second specimen processing described above. They will then undergo colposcopy according to the standard of care based upon their prior cytology results.

Data to be collected from subjects at the time of their clinical visit; or upon chart review by the research team; will include demographic information, whether the subject is presenting for a routine screen or for a follow-up, and the subject's most recent cervical cytology result. Data to be collected from the standard-of-care cytology will include the cytologic diagnosis and the presence or absence of human papilloma virus (HPV), if applicable based on the subject's age. The Pap results of each subject will be reviewed and classified as "screen negative" (NIL), "screen positive" (LSIL, HSIL), or "indeterminate" (ASCUS, ASC-H). Subjects with a positive screen, or an indeterminate result, who undergo colposcopy will have their colposcopy results recorded as well. The UV microscopy imaging features will be coded and collected.

The extracted features from the UV microscopic analysis will be compared with the results from the standard-of-care cytology. Gaussian distributions will be plotted for each feature, with a separate distribution plotted for patients who screen positive and who screen negative based on standard-of-care cytology results. Combinations of UV microscopy features compared to the standard-of-care cytology results; and, if applicable, their colposcopy results; will be recorded and plotted as well. Areas of overlap between these distributions will be used to select an appropriate threshold value, with a target false negative rate of 14% or better

Inclusion Criteria: Patients presenting for cervical cancer screening. Patients presenting for follow-up for positive cervical cancer screening. Exclusion Criteria: Patients who do not, or no longer, require cervical cancer screening