**1. Introduction**

Cancer is a major cause of death worldwide, and the incidence of cancer continues to rise, making early detection and diagnosis essential. In 2020, nearly 10 million people died of cancer, according to the World Health Organization [1]. Cancer is a genetic disease caused by changes in the DNA sequence of cells due to DNA damage, harmful substances, or errors in cell division [2]. Lung and breast cancers are prevalent types of cancer that affect men and women. Carcinoma, lymphoma, leukemia, sarcoma, and melanoma are frequently encountered cancer classifications that can emerge in various organs of the human body. Aging, prolonged exposure to sunlight, smoking, radiation exposure, viral infections, the use of hormonal medications, and exposure to certain chemicals are among the known factors contributing to the development of cancer [3].

Cancer involves the abnormal growth of cells that are clones of each other, and these cancerous cells can divide and spread beyond the neighboring cells, generating tumors. Tumors can be benign or malignant [4]. Benign tumors stay in their primary stage and may grow in size without spreading into neighboring tissues or organs. In contrast, malignant tumors consist of cancer cells that grow uncontrollably and spread into neighboring cells, tissues, or organs. Cancer cells can spread to other parts of the body via the lymphatic node system or through the bloodstream. From benign tumors to malignant, cancer passes through four stages, where the fourth stage is called metastatic, and the chances of survival

**Citation:** Kokabi, M.; Tahir, M.N.; Singh, D.; Javanmard, M. Advancing Healthcare: Synergizing Biosensors and Machine Learning for Early Cancer Diagnosis. *Biosensors* **2023**, *13*, 884. https://doi.org/10.3390/ bios13090884

Received: 21 August 2023 Revised: 8 September 2023 Accepted: 9 September 2023 Published: 13 September 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

are very low [5]. Unfortunately, there are no advanced methods to diagnose cancer at the earliest stages.

Typical cancer detection processes involve physical exams, lab tests that use expensive and harmful equipment based on electromagnetic radiation, or a surgical diagnosis called a biopsy [6]. These lab tests use biological samples, such as blood, urine, saliva, or sweat, to detect specific biomarkers. Imaging-based tests include magnetic resonance imaging (MRI), computed tomography (CT), positron emission tomography (PET), ultrasound, and X-rays [7–10]. Although these imaging tests are non-invasive, they require expensive equipment and experienced pathologists. Biopsies are useful for determining the type of tumor, but they are invasive, painful, time-consuming, and expensive [11–13]. These issues with the current cancer diagnostic and detection methods make them unsuitable for point-of-care testing (POCT) and early cancer diagnosis [14–19]. Therefore, there is an urgent need to develop biosensing platforms for early cancer diagnosis that are low-cost, non-invasive, accurate, less time-consuming, and can be used at the point of care.

Biosensors can fill the void of point-of-care testing by providing low-cost, non-invasive, high-speed, and on-site testing of bio samples. A biosensor is a specialized type of sensor designed to detect and capture biological signals or molecules within a given sample or environment. These biological signals can include various entities, such as specific proteins, enzymes, DNA sequences, antibodies, or even whole cells. Biosensors are particularly significant in the field of medical and healthcare applications due to their ability to provide valuable information about biological processes, disease diagnosis, and patient monitoring [20]. Various point-of-care devices and lab-on-chip technologies have been employed to detect and diagnose various biomarkers from bio samples, such as pregnancy tests, blood glucose, or tuberculosis. Biosensors search for a specific biomarker in the bio sample and use either electrical or optical sensing modalities to detect biomarkers [19,21–26] associated with a cancer type. Additionally, researchers have conducted significant research on biosensors that use sweat or exhaled air as a potential sensing mode and have tried to associate the concentrations of volatile organic compounds with the presence of cancer or some other disease [25,27].

Biosensors can generate large amounts of data at the time of diagnosis and may require complex processing algorithms to generate the results. To effectively analyze the vast amounts of data obtained from biosensors and extract valuable information, researchers have harnessed the power of machine learning (ML) techniques. These methods include but are not limited to support vector machines (SVM), k-nearest neighbor (KNN), decision trees (DT), artificial neural networks (ANN), and convolutional neural networks (CNN) [28,29]. This paper is a review of the research projects that develop biosensors for early cancer detection and diagnosis and employ ML algorithms to enhance data analysis. In this regard, a review of ML algorithms is presented, along with the fundamental working principles of different biosensing techniques. The papers in this review study are divided into two main categories: electrical detection and optical detection biosensors. To facilitate a reader's understanding of how ML models are advancing biosensors in diagnosing different types of cancer, sub-categories have been created based on the type of cancer detected using specific biomarkers or cells.
