2 areas AI can help improve patient outcomes in cancer care

3 min read

Photo: Pexels-pavel-danilyuk

The British Journal of Cancer recently reported that going forward around 50% of the UK population is likely to receive a cancer diagnosis at some point during their lifetime. Similar statistics have also been observed in the US. These findings broadly suggest in the coming years, we should expect a significant increase in the number of people affected by cancer, worldwide.

The rise in patient numbers and the prevailing inconsistency in cancer care are cause for concern. Inconsistent care often results in unwelcome complications and increases the risk premature death. Ironically the problem isn't always the cancer itself, but rather its incorrect diagnosis and/or treatment.

Throughout a patient's medical journey, there are three key phases: diagnosis, treatment, and either survival or palliative care. Diagnosis and treatment phases play a crucial role in determining patient outcomes, and could potentially benefit from the implementation of AI. By leveraging AI during these phases, we can reduce the risk of a negative domino effect that could result in unnecessary harm to patients.

First Phase - Diagnostic

The first important step towards effective cancer treatment is an accurate and timely diagnosis. Failure to do so can lead to negative consequences. According to the Institute of Medicine*, "Most people will experience at least one diagnostic error in their lifetime, sometimes with devastating consequences".

Misdiagnosing cancer happens and does so frequently enough that it's problematic. Misdiagnosis can happen in many ways. For example, a patient is correctly diagnosed with cancer but the extent of the cancer is misjudged. Consequently, this can affect their treatment and outcome. There are times when patients do not receive the requisite diagnostic tests or when their results are inaccurately interpreted. For example, a lump in the breast can be misdiagnosed as being cancerous, resulting in an unnecessary life-changing surgical procedure.

Imaging as a diagnostic test is crucial in diagnosing many types of cancer. Scans such as MRI/CT/PET can help determine the extent of the disease, which is essential in developing a treatment plan. For instance, in rectal cancer, a scan revealing only a small and confined tumour may not require surgery. Therefore, ensuring scans are accurately interpreted is critical towards effective treatment.

AI can assist in detecting tumours in scans. A recent study in Sweden** demonstrated that the combination of AI and breast cancer radiologists improved the overall diagnostic accuracy. Notably, without the AI, almost 2000 patients would have been wrongly diagnosed as not having cancer.

Second Phase - Treatment

Despite advances in cancer medicine, a major challenge remains - a significant number of patients fail to respond positively to standard cancer treatments, putting their lives at risk. For instance, only around 35% of patients with liver cancer who receive standard care show a positive response.

The field also lacks real-time feedback and analysis, based on combined clinical data, on how patients respond to their treatment and progress thereafter.

How can AI make a meaningful difference?

Patients with certain types of cancer, such as lung cancer, are not only expected to undergo routine tests but also genetic tests. Such tests not only help patients receive personalised treatment but also determine if they are suitable for certain types of treatments.

Cancer testing generates huge amounts of data from blood, urine, scans, genomics, and biopsies. Analyzing these datasets can provide valuable insights for cancer care. AI algorithms can efficiently analyze complex clinical data to predict how the patient's cancer may progress and provide feedback to adjust treatment in a timely manner.

AI-powered insights can revolutionize clinical practice and personalise treatment to change patients' lives.

Conclusion

Making sure patients receive effective cancer treatment is crucial, and AI can play a key role in achieving it. With its assistance, there are two stages where AI can make a significant impact: identifying the disease and developing personalised treatment plans. By leveraging AI's capabilities, we can improve cancer diagnosis and treatment outcomes.

* National Academies of Sciences, Engineering, and Medicine; Institute of Medicine; Board on Health Care Services; Committee on Diagnostic Error in Health Care; Erin P. Balogh, Bryan T. Miller, and John R. Ball, Editors.

**Artificial intelligence for breast cancer detection in screening mammography in Sweden: a prospective, population-based, paired-reader, non-inferiority study; by Karin Dembrower, and others.