McMaster HealthTech

Author: Ashley Smith, McMaster University

Cancer impacts millions globally, presenting challenges in diagnosis, treatment, and prognosis. Physicians and healthcare providers often require significant time to diagnose cancers according to their specific type, stage, and progression, as well as to develop a treatment plan for the patient. What if there were a tool that could diagnose cancers quickly, accurately, and even recommend treatment plans? That is where machine learning offers a novel and transformative role.

Machine learning is a branch of artificial intelligence that enables computers to learn from data without being explicitly programmed. By developing algorithms capable of identifying patterns and relationships in large datasets, machine learning enables computers to analyze data, make predictions, and assist in decision making.¹ In the context of cancer detection, these algorithms can identify abnormalities, predict outcomes, and recommend treatment outcomes with high accuracy, which has the potential to rapidly translate complex findings into actionable insights for healthcare providers.¹ 

The algorithms in machine learning can detect biomarkers from genetic information and patient samples within datasets, identifying patterns and relationships that help differentiate between cancer types and classify them accurately.^1,2 This is a critically important feature, as human analyses can overlook subtle associations within patient profiles that, with proper training, machine learning can detect.³ More importantly, machine learning continuously improves by learning from predictive models, enhancing its ability to predict cancer types and disease progression.^3,4 Additionally, algorithms can use ensemble methods, which combine multiple approaches to analyzing clinical data, generating more precise predictions and further improving the accuracy and robustness of machine models in classifying various cancers.¹ Overall, the key advantage of machine learning lies in its ability to efficiently process vast, complex data sets while leveraging predictive models and ensemble methods to achieve highly accurate and reliable results.¹

There are various types of machine learning, including supervised learning, unsupervised learning, and reinforcement learning.¹ In supervised learning, machines use labeled datasets to make predictions, whereas unsupervised learning identifies patterns in unlabeled data.^1,5 Furthermore, reinforcement learning involves training a machine to improve its findings and output by maximizing a “reward” signal tied to achieving a specific goal.¹ This process enables the machine enhance its accuracy and precision, increasing its reliability in decision-making.¹

Furthermore, machine learning enhances cancer care by improving diagnosis, prognosis estimation, and treatment planning.^1-5 It can accurately identify differences in cancer types, subtypes, and stages, while also detecting biomarkers and drug targets for effective therapies.^1-5 For example, in skin cancers, machine learning can identify the type and stage of skin cancer from photographs, often with accuracy and efficiency comparable to or exceeding a dermatologist.¹ Similarly, in lung cancers, machine learning can identify cancer nodules on a computed tomography scans to determine the type and stage of the cancer, while also differentiating between benign and malignant tumours.¹ Currently, machine learning has shown great promise in its insights for a variety of additional cancer types, such as breast cancer, prostate cancer, liver cancer, and brain cancers. These applications enable therapies to be tailored according to genetic and tumor profiles, facilitating more patient-centered interventions.^1-5

Overall, machine learning can enhance patient outcomes through timely and accurate diagnosis, prognosis, and patient-centered treatment plans.¹ Machine learning has the potential to become a powerful tool to assist healthcare providers in making informed decisions about a patient’s care plan, particularly in oncology, where diagnoses can be life-changing and devastating for patients.¹ By improving treatment accuracy, machine learning also fosters trust between patients and healthcare providers, helping to build stronger therapeutic relationships during challenging times in patients’ lives.¹

Works Cited:

1. Yaqoob A, Aziz RM, Verma NK. Applications and techniques of machine learning in cancer classification: a systematic review. Hum Cent Intell Syst. 2023 Sep 11. https://doi.org/10.1007/s44230-023-00041-3.

2. Mishra KN, Mishra A, Ray S, Kumari A, Waris SM. Enhancing cancer detection and prevention mechanisms using advanced machine learning approaches. Informatics Med Unlocked. 2024 Aug;101579:1-9. https://doi.org/10.1016/j.imu.2024.101579.

3. Chennupati AK. Artificial intelligence and machine learning for early cancer prediction and response. World J Adv Eng Technol Sci. 2024 May 30;12(1):35-40. https://doi.org/10.30574/wjaets.2024.12.1.0178.

4. Khanam N, Kumar R. Recent applications of artificial intelligence in early cancer detection. Curr Med Chem. 2022 Feb 22;29. https://doi.org/10.2174/0929867329666220222154733.

5. Kourou K, Exarchos TP, Exarchos KP, Karamouzis MV, Fotiadis DI. Machine learning applications in cancer prognosis and prediction. Comput Struct Biotechnol J. 2015;13:8-17. https://doi.org/10.1016/j.csbj.2014.11.005.