Spectroscopy of blood samples for the diagnosis of endometrial cancer and classification of its different subtypes

Paraskevaidi, Maria, Medeiros-De-morais, Camilo De lelis orcid iconORCID: 0000-0003-2573-787X, Raglan, Olivia, Lima, Kassio M. G., Martin-Hirsch, Pierre L., Paraskevaidis, Evangelos, Kyrgiou, Maria and Martin, Francis L orcid iconORCID: 0000-0001-8562-4944 (2017) Spectroscopy of blood samples for the diagnosis of endometrial cancer and classification of its different subtypes. Journal of Clinical Oncology, 35 (S15). p. 5596. ISSN 0732-183X

Full text not available from this repository.

Official URL: https://doi.org/10.1200/JCO.2017.35.15_suppl.5596


Background: Symptoms of endometrial cancer often appear in early stages, thus a diagnosis, based on microscopic histological examination of endometrial tissue, can be given relatively on time. However, this procedure interferes subjective interpretation allowing human error, while screening of the asymptomatic population is not widely performed because of the high cost of the available tests (e.g. transvaginal ultrasound) and the relative invasiveness [biopsy or dilation and curettage (D+C)]. Consequently, there is a widespread need to develop inexpensive, non-invasive techniques that would accurately diagnose endometrial cancer, as well as classify the different subtypes. Spectrochemical methods generate a signature fingerprint of biological material in the form of spectra. Unlike immunological methods, which detect only one molecule at a time, the spectra obtained from a clinical sample represent all the molecular constituents within that sample, including proteins, lipids and carbohydrates; this provides a holistic picture of the sample. Previous studies have confirmed spectroscopy’s ability to diagnose gynecologic cancers in blood. Methods: Attenuated total reflection-Fourier transform infrared (ATR-FTIR) spectroscopy was used to analyse blood plasma and serum from 71 women with endometrial cancer and 18 age-matched healthy controls; classification algorithms were then applied to extract the underlying biological information. Results: Principal component analysis followed by support vector machine (PCA-SVM) diagnosed endometrial cancer with 100% accuracy in plasma and 95% in serum. Discrimination between the different subtypes [endometrioid adenocarcinoma (n = 43) vs carcinosarcoma (n = 14)] was achieved with 98.33% accuracy in both plasma and serum. The spectral regions responsible for discrimination were attributed to protein and lipid alterations. Conclusions: Our preliminary results suggest an accurate and objective diagnostic tool for endometrial cancer with blood testing, allowing therefore thoughts for a potential screening test in high risk populations. Future work will include higher number of normal cases and different subtypes and grades.

Repository Staff Only: item control page