Introduction/Background Epithelial ovarian cancer (EOC) is the most lethal cancer among gynecological malignancies worldwide, accounting for 90% of all ovarian cancers. Our study’s primary objective is to discriminate benign from malignant ovarian tumors using tissue samples that underwent formalin fixation using ATR-FTIR spectroscopy.

Methodology This is a retrospective, single center study. Inclusion criteria were of formalin fixed tissue taken from pathology archive samples, resected from females ages 18 and above. Two sets of slides from malignant tumors and 2 sets of slides from benign tumors were used to study the effect of tissue thicknesses on the measured absorption spectra: a set of 4 microns and a set of 12 microns from each tumor type were compared. Spectroscopic measurements were performed on the different slides. The 4 microns tissue thickness group was chosen. Each slide was measured in multiple locations. PCA-LDA Discrimination analysis was performed using the measured spectra. The cross-validation process was repeated five times. The results of these validations were then averaged to produce a single estimation.

Results A total of 74 tissue samples were examined. Absorption spectra of the malignant tumors were consistently different from that of benign tumors at many spectral ranges. Using K-fold cross validation technique, the study showed that the model correctly classified the samples into malignant and benign groups with an accuracy of 94.5%.

• Download figure
• Open in new tab
• Download powerpoint

Abstract #79 Figure 1

Cancer group vs benign group differences

Disclosures Our study exhibits a sensitive method to differentiate between a benign and malignant paraffin block preparation. With further research, this technique could become an alternative to conventional histopathology.