3.1.1. Breast Cancer

Based on the World Health Organization (WHO), breast cancer is the most frequent cancer among women, affecting over 1.5 million women each year, and is responsible for the most significant cancer-related deaths among women. In 2015, 570,000 women died from breast cancer [60]. This highlights the potential of biosensors for the early detection of cancer cells. Biosensors are promising and selective detection devices which hold immense potential as point-of-care (POC) tools [61]. Several studies have shown the application of optical-based biosensors to detect breast cancer cells, demonstrating the promising potential of biosensors for early-stage detection of breast cancer cells.

The surface plasmon resonance (SPR) sensor is an optical sensor employing a unique mode of electromagnetic field called the surface plasmon, which propagates at the interface of a metal and a dielectric. The SPR sensor utilizes the evanescent field generated by the surface plasmon to detect alterations in the refractive index of the dielectric material near the interface [62]. Numerous studies have suggested the effectiveness of SPR sensors in the early detection of cancers [63]. Kumar et al. [64] described a photonic crystal fiberbased surface plasmon resonance (SPR) sensor for detecting breast cancer cells based on their refractive index (Figure 2A). They used simulations and numerical analysis to measure the wavelength sensitivity and resolution of the sensors for normal and cancerous cells, achieving a high sensitivity and low resolution. The refractive index of normal and cancerous cells was estimated using a multi-layer perceptron, and the machine learning algorithm was used to optimize the structural parameters. The proposed sensor shows promising results and could be a potential alternative sensing device for early-stage breast cancer diagnosis. In another study, Verma et al. [65] developed a machine learning approach for breast cancer cell detection using a surface plasmon resonance (SPR) based on a photonic crystal fiber sensor, which is shown in Figure 2B. The sensor operates by detecting changes in the refractive index of the fiber when breast cancer cells are present. The machine learning algorithm is trained on a dataset of SPR spectra obtained from both breast cancer and noncancerous cells and is used to classify new samples as either cancerous or non-cancerous based on their spectral patterns.

Another type of optical sensor is the fluorescence sensor, widely used to identify and measure biomolecules or metal ions. The advantages of this type of sensor include its sensitivity, specificity, resistance to light scattering, and ease of use [66]. In a study reported by Jin et al. [67], they developed a breast cancer liquid biopsy system that integrates a fluorescence sensor array with a deep learning model. The sensor array uses fluorescent probes to gather diverse information about cells and exosomes. The deep learning model employs a CNN-based architecture to distinguish between normal and cancerous cells. The system has demonstrated successful discrimination between normal and different cancerous cells and achieved a 100% accurate classification of different breast cancer cells. In addition, Pala et al. [68] constructed and tested a digital in-line holographic microscope for imaging breast cancer cells using holography, which is shown in Figure 2C. The microscope was constructed using a white LED for illumination, a pinhole to make the light semicoherent, and a CMOS sensor to record images of the plane above it. Holograms were captured and numerically reconstructed, and the amplitude of individual cells was collected. Using machine learning, these images were transformed into a fractal dimension and rotated to calculate the identifying features of each cell. Upon testing the accuracy of this system, the team achieved an accuracy of 99.65%.
