Cancer continues to be a menace worldwide. Despite a flurry of scientific innovations, it is still the leading cause of lives lost across the world.
In 2024, in the United States alone, an estimated two million cancer cases will be diagnosed, while the total number of deaths projected will cross the mark of 610,000.
Calculations based on 2017-2019 data show that more than 40% of men and women in the United States will be diagnosed with cancer at some point during their lifetime. Even more alarming is that this year, close to 15,000 children and adolescents will be diagnosed with cancer, and approximately 1,600 will unfortunately succumb to it.
The situation in the United States reflects the global scenario. As one of the leading causes of death worldwide, in 2022, there were almost 20 million new cases and 9.7 million cancer-related deaths worldwide. And there is no relief visible in the near term. By 2040, the number of new cancer cases per year is expected to rise to 29.9 million and the number of cancer-related deaths to 15.3 million.
Therefore, the menace of cancer requires all-around efforts from the scientific, technological, and medical research community. In the coming segment, we will discuss one such effort that makes people around the world optimistic and enthusiastic about its potential for future endeavors.
Click here for a list of top cancer therapy stocks.
AI-Powered Precision Oncology for Cancer Biopsies
A recent scientific journal publication in the Journal of Clinical Oncology mentioned its efforts to train DeepHRD, a deep learning platform for predicting HRD from hematoxylin and eosin (H&E)–stained histopathological slides, using primary breast and ovarian cancers from The Cancer Genome Atlas (TCGA).
This breakthrough is significant because cancers with homologous recombination deficiency (HRD) benefit from platinum salts and poly(ADP-ribose) polymerase inhibitors. However, the standard diagnostic tests for detecting HRD require molecular profiling, which is not universally available.
The research results concluded that DeepHRD could predict HRD in breast and ovarian cancers directly from routine H&E slides across multiple external cohorts, slide scanners, and tissue fixation variables. When its results were compared with the conventional system of molecular testing, the DeepHRD platform could classify 1.8- to 3.1-fold more patients and exhibited better OS in high-grade serous ovarian cancer and platinum-specific PFS in metastatic breast cancer.
While the results show the medical community’s reasons for excitement about the solution, the inclusion of AI meant a big improvement in the space of operational logistics of early cancer diagnosis and pertinent treatment thereof.
Click here to learn how CRISPR can be used to treat cancer.
Why is the Inclusion of AI Protocol Crucial?
According to the UC San Diego researchers at the helm of this experiment, the inclusion of AI in the system allows rapid and low-cost detection of clinically actionable genomic alterations directly from tumor biopsy slides.
The researchers believe that the system could save weeks and thousands of dollars from clinical oncology treatment workflows for breast and ovarian cancer.
Otherwise, not only could the delays prove to be life-threatening, but the high-cost processes could affect patients with resource constraints. In that sense, the inclusion of AI meant a more inclusive cancer diagnostic process.
While explaining the bottlenecks that existed, Erik Bergstrom, PhD, lead author of the study and a postdoctoral researcher in Alexandrov’s lab, said:
“Unfortunately, high costs, tissue requirements, and slow turnaround times have hindered the widespread use of precision oncology, leading to suboptimal — potentially detrimental — treatment for cancer patients.”
AI-powered genomic testing, therefore, comes with relief. It not only saves critical time for patients but also assists doctors by enabling immediate treatment prescriptions after initial tissue diagnosis. This precision oncology method is more reliable than conventional genomic tests, as AI tests have a negligible failure rate. In contrast, conventional genomic tests have a failure rate of between 20 and 30%, which often necessitates retesting or resorting to an invasive biopsy.
The AI identification of HRD empowers this method to become something that allows accurate and instantaneous detection of cancer genomic biomarkers.
Researchers and scientists who’ve been involved in this space for a long time believe the breakthrough to be momentous. Explaining the reasons, Scott Lippman, M.D., UC San Diego distinguished professor of medicine, Center for Engineering and Cancer, and Moores Cancer Center member, said:
“The era of precision oncology took off in the late 90s, but recent U.S. studies show that the vast majority of cancer patients are not getting FDA-approved precision therapy. And the prime reason is because they’re not getting tested. As a clinical oncologist — and I’ve been doing this for nearly 40 years — there is no question that this approach is the future of precision oncology.”
The technology that serves as the foundation for DeepHRD is protected by provisional patents issued through UC San Diego. These patents have been licensed to io9, a company involving Alexandrov, Bergstrom, and Lippman.
In the next segment, we look into this company briefly to understand the kind of capabilities it is built to handle and the solutions it has brought forth so far!
#1. Io9: AI-Driven Digital Pathology
Io9 already boasts an AI platform that efficiently conducts molecular profiling from routine clinical slides. This platform can detect genomic biomarkers from digital slides, making access to treatment possible for all patients worldwide.
Apart from the HRD biomarker-based solution that will work for breast and ovarian cancer now and pancreatic cancer in the future, io9 is also active on a host of biomarkers, including KRAS, EGFR, and BRAF, which will prove useful in the domains of lung, colorectal, and breast cancer and metastatic melanoma in future.
Ludmil B. Alexandrov is the founder and Chief Scientific Officer of io9, while Scott M. Lippman, another among the founders, is on the board of directors.
While io9 is a spinoff of an entity for educators and researchers to keep working with their innovations in a scaled-up, commercial fashion, there are big companies active in this space as well where AI meets oncology diagnostics and treatment.
#2. GE Healthcare
On the first day of May 2024, GE HealthCare (Nasdaq: GEHC) announced Revolution RT, a new radiation therapy computed tomography (CT) solution with innovative hardware and software that would simplify simulation workflow and make oncology care more personalized and seamless. The Revolution RT was unveiled at the European Society for Therapeutic Radiology and Oncology (ESTRO) 2024 Congress in Glasgow with an updated artificial intelligence (AI)-an enhanced version of the Intelligent Radiation Therapy (IRT) platform.
While speaking about the inclusion of AI in their oncology solutions, Dr. Taha Kass-Hout, chief technology and science officer of GE HealthCare, said the following:
“At GE HealthCare, we are committed to advancing the frontiers of oncology treatment through AI-driven technologies that transform and optimize the care continuum.”
While listing the benefits of incorporating AI, Dr. Taha Kass-Hout said that the enhancement would help view the patient’s anatomy with great accuracy and facilitate precise tumor targeting while aiming to protect nearby healthy tissue. This would result in “a precise, more connected, and efficient care environment that accelerates the delivery of personalized care, with the goal of enabling better patient outcomes.”
For the full year 2023, GE Healthcare reported revenues of US$19.6 billion, registering an increase of 7% year-over-year and 8% on an Organic basis with growth across all segments and regions.
#3. iCAD
Another med-tech entity, which is much smaller in size but wide in its coverage, impact, and ambition, iCAD, is doing specific AI-focused work in the field of breast cancer. The company draws its vision and purpose from real-life threats that exist on the ground. Data shows that 1 in 8 women would get breast cancer in her lifetime, and as high as 20-40% of these cancers are missed in mammogram screenings.
With a woman getting diagnosed with breast cancer every 14 seconds globally, the assessment of cancer risk with traditional models is only 63% accurate.
iCAD’s platform ProFound Breast Health Suite focuses on detection, density, and risk, shining a greater spotlight on cancer and exposing its hiding places.
In the detection phase, ProFound’s solution incorporates the latest developments in deep-learning, 3rd generation artificial intelligence (AI). The accuracy and efficiency are quite unprecedented for 2D and 3D mammography screening with up to 2x enhanced clinical performance compared to other AI platforms.
While elaborating on ProFound AI’s capabilities in detecting breast cancer from the aspect of breast density, Gabriele Pedicalli, MD, Wooster Community Hospital, had the following to say:
“ProFound AI not only helps us to review cases with fatty tissue, but we also see an even greater benefit for those women with dense breasts. Before ProFound AI, I might have read a case with dense breasts and thought, ‘Let’s call them back in 6 months or a year,’ but with ProFound AI, we can more easily tell whether we should be doing another examination or biopsy.”
Finally, in the risk assessment segment, ProFound’s AI-powered solution makes it possible to have the world’s first image-based risk model, which is 2.4x more accurate than traditional models. It serves excellently when it comes to identifying women at high risk of developing cancer before or at their next screening to help find interval cancers earlier.
For the full year 2023, iCAD’s total revenues were US$17.3 million, down 13%.
The Future of AI-Powered Precision Oncology
In April 2024, the globally revered Nature magazine comprehensively examined the role of AI in precision oncology. It duly recognized AI’s transformative capability in this space, as its opening remarks were:
“The fields of cancer research and precision oncology are undergoing a massive transformation due to the application of artificial intelligence (AI).“
To be more specific, the journal article highlighted AI’s role in detecting hidden patterns from multiple sources of information, including molecular profiling, pathology, and medical imaging. It also appreciated AI for integrating omics data to provide a more comprehensive understanding of cancer. Apart from monitoring and diagnosis, AI has also been of great use in developing new assays for characterizing cancer, prognostication, and predicting responses to specific treatments.
However, the summary report did not forget to mention the limitations and challenges that arose when translating these new tools from research settings to clinical practice.
There are many examples of tools from research settings that have been successfully translated into clinical practice, though. For instance, in a developing economy like India, Apollo Cancer Centre, Bengaluru, launched India’s first AI-Precision Oncology Centre (POC) in January 2024.
While elaborating on how such centers could help in introducing a better and more efficient regime for cancer care, Dr. Vishwanath S, Senior Consultant, Medical Oncology, ACC, Bengaluru, said:
“Clinically, it helps in care pathway compliance by monitoring adherence to standard care pathways, also for patient management based on genomic, clinical, and pathological profiles. Clinical alerts and recommendations for diagnostic tests, enrollment for Value-Based Care (VBC) & other patient benefit programs are some of the uses. Patient drop-off and Patient-reported outcome measures (PROMs) alerts for timely interventions are among the other benefits.”
The use of AI in cancer diagnostics is set to witness staggering growth in the future. According to a report published by Towards Healthcare, AI in cancer diagnostics is anticipated to soar at a 9.35% Compound Annual Growth Rate (CAGR) from 2023 to 2032, with the market on a trajectory to achieve an estimated USD 2,084.34 million by the end of 2032 from US$892.23 million in 2022.
Characteristically, it can be said with surety that AI would aid in enhancing diagnostic accuracy and significantly improve the efficiency of treatment planning!
Click here to learn all about investing in artificial intelligence.