A comparative evaluation of diffuse reflectance and Raman spectroscopy in the detection of cervical cancer

Rubina Shaikh, Vasumathi G. Prabitha, Tapas Kumar Dora, Supriya Chopra, Amita Maheshwari, Kedar Deodhar, Bharat Rekhi, Nita Sukumar, C. Murali Krishna, Narayanan Subhash

Research output: Contribution to journalArticlepeer-review

20 Citations (Scopus)

Abstract

Optical spectroscopic techniques show improved diagnostic accuracy for non-invasive detection of cervical cancers. In this study, sensitivity and specificity of two in vivo modalities, i.e diffuse reflectance spectroscopy (DRS) and Raman spectroscopy (RS), were compared by utilizing spectra recorded from the same sites (67 tumor (T), 22 normal cervix (C), and 57 normal vagina (V)). Data was analysed using principal component – linear discriminant analysis (PC-LDA), and validated using leave-one-out-cross-validation (LOOCV). Sensitivity, specificity, positive predictive value and negative predictive value for classification between normal (N) and tumor (T) sites were 91%, 96%, 95% and 93%, respectively for RS and 85%, 95%, 93% and 88%, respectively for DRS. Even though DRS revealed slightly lower diagnostic accuracies, owing to its lower cost and portability, it was found to be more suited for cervical cancer screening in low resource settings. On the other hand, RS based devices could be ideal for screening patients with centralised facilities in developing countries. (Figure presented.).

Original languageEnglish
Pages (from-to)242-252
Number of pages11
JournalJournal of Biophotonics
Volume10
Issue number2
DOIs
Publication statusPublished - 1 Feb 2017
Externally publishedYes

Keywords

  • cancer screening
  • cervical intraepithelial neoplasia
  • diffuse reflectance spectroscopy
  • in vivo
  • principal component-linear discriminant analysis
  • Raman spectroscopy

Fingerprint

Dive into the research topics of 'A comparative evaluation of diffuse reflectance and Raman spectroscopy in the detection of cervical cancer'. Together they form a unique fingerprint.

Cite this