Breast cancer is most common cancer among women world-wide and can be treated if diagnosed at early-stage. Fluorescence (FL) techniques have a considerable impact in tumor detection as FL is highly sensitive to biochemical and biophysical structure of the tissues, providing novel techniques for early and noninvasive diagnosis of cancer. Due to high sensitivity of FL, it can be used in early-stage breast cancer detection. Fluorescence-guided-surgery (FGS) using exogenous agent in breast cancer is a well-known method, which locates tumor and margins during intra-operative procedure. FL-spectroscopy and FL-imaging has shown the potentiality in FGS of breast cancer independently. Combination of these two modalities in a single system improves the success rate of FGS which can give tumor free tissues. The goal of the current study is to analyze and classify the early-stage breast cancer, locally advanced breast cancer (invasive ductal carcinoma) and normal tissue. For that, we combine FL-imaging and FL-spectroscopy in a single smartphone-based point-of-care devices and recorded data during intra-operative procedure. A total 21 patients of invasive ductal carcinoma and fibroadenoma are included in this study. Total 65 FL-spectra are recorded during intra-operative procedure which are further used in support vector machine (SVM) based classification of fibroadenoma, invasive ductal carcinoma (IDC) and normal tissue. The sensitivity, specificity, positive predictive value, negative predictive value, and overall diagnostic accuracy (total efficiency) is 78.6 %, 90%, 91.6%, 75%, 95%, respectively. Additionally, we observed a red shift in case of IDC and fibroadenoma from normal tissue, which is 5.22±1.77 nm, and 4.96±2.61 nm, respectively.
|