Sergey Avdeychik, Director of Healthcare Technologies at Andersen.
Cancer is one of the main healthcare challenges of the 21st century. According to the International Agency for Research on Cancer (IARC), in 2020, the number of new cases worldwide exceeded 19.3 million, and the total cancer deaths in the same year were almost 10 million. The dominant types of cancer in developed countries are breast (45 per 100,000 of the population), prostate, lung and rectal cancers. IARC predicts that if the current rate of population growth and cancer incidence remains, there will be an increase in cases of up to 28.4 million people per year by 2040.
Cancer care is expensive. According to the forecasts by the U.S. National Cancer Institute (NCI), the U.S. government’s spending on fighting cancer will exceed $246 billion by 2030. This is 34% higher than in 2015; notably, the previous estimated figure was reached twice sooner than expected. The growth in the need for treatment will impact the growth of companies working to apply more technology to this segment of healthcare.
Significance Of Technologies For Detection And Treatment
Every day, we see how the efforts of scientists, clinicians and tech companies in cancer care are combined with the purpose of early diagnosis and accurate prognosis. Lung cancer – the major cause of cancer death worldwide and the deadliest type of cancer out of the four most common (about 75% of patients die within five years). But when lung cancers are found early, then two-thirds of patients survive for at least five years. In my opinion, regular lung X-ray screenings of vulnerable individuals have the potential to reduce the number of deaths.
Digital technologies (AI/ML in particular) are becoming essential and effective assistance in cancer care, and with good reason. In order to achieve the best results in early cancer detection, massive screenings of the population are needed. Additionally, the development of new methods of diagnosis and treatment in oncology requires processing enormous amounts of data acquired from pathology, radiology, laboratory and advanced molecular diagnostics.
How Digital Health Solutions For Cancer Care Work
Radiographic assessment of tumors relies upon visual evaluation by expert clinicians, and these readings may be enhanced by advanced computational analysis. AI helps automate processes of the initial interpretation of images. It is also able to perform advanced scan analysis with volumetric delineation of tumors over time, helping predict their genotype and biological course, estimating the clinical outcomes and the progression of the disease. According to NCI experts’ analysis of more than 60,000 cervical images from a different NCI study, the accuracy of automatic image processing tools for detecting cervical cancer in these images was 91%, which is around 30% higher than the accuracy of examinations by specialists.
I believe that the accuracy of mammography, a traditional method for diagnosing breast cancer, can also be improved. The American Cancer Society estimates that about 12.1 million mammograms are performed in the U.S. annually with almost every second woman being misdiagnosed with breast cancer. The use of AI in one study and evaluation of 500 pathology reports showed a 30% faster data processing speed than a human doctor with 99% accuracy of results. It is intended to eliminate the need for unnecessary biopsies and reduces the stress caused by misdiagnosis.
Due to limitations of human vision, radiologists can overlook tiny malignancies in the early stages when the treatment possibilities are the highest. Up to 35% of lung nodules are not found during initial screening. One deep-learning AI module analyzing CT scans of lungs can detect up to 95% of even the smallest neoplasms (1-3 mm in diameter), which is 30 times more accurate than what a radiologist can provide. In one research paper, during CT diagnostics of lung cancer, ML-based models help reduce the number of false positives by 11% and the number of false negatives by 5%
Interestingly, the sensitivity of a traditional lung X-ray, which delivers about 70 times smaller radiation dose than CT scanning, can also be increased with the help of digital correction tools. These tools help “remove” various artifacts, such as ribs, from the patient’s lung X-ray image, so the provider can get unobstructed views of the smallest pathological nodules hidden behind.
Leading Companies In Cancer Digital Solution Development
Advancements in the field of image analysis for cancer care are achieved by large companies such as IBM and Google, research institutions like NCI and also smaller digital health start-ups. These initiatives come not only from such traditional “nests” of science-saturated startups as the U.S., EU, Japan, South Korea and Israel but also, for example, from Egypt. Rology, which was founded there in 2017, has raised over a million dollars for the development of its teleradiology platform equipped with an AI-enabled DICOM viewer and has aimed to address the shortage of radiologists in the market.
One of the American startups, PathAI, which raised about $250 million in total after round series C for the development of AI-based technologies to augment the work of pathologists, is focusing on the improved detection of abnormalities in tissue sections.
In the digital pathology segment, I would like to mention Google Health’s Augmented Reality Microscope (ARM). The microscope’s utility for detecting metastatic breast cancer and identifying prostate cancer was demonstrated. This shows more potential for improving cancer detection.
I am convinced that cancer care is not only about accurate diagnosis and effective treatment, but also about prevention of cases, increased monitoring of conditions, improved overall quality of life and advanced end-of-life care for cancer patients. In the second part of the article, we shall discover how, in all these areas, digital technologies are finding worthy applications and what that means for business development leaders.