Faster cancer diagnoses with Artificial Intelligence
Genomics | QIAseq NGS Solutions

Faster cancer diagnoses using Artificial Intelligence

1 February 2021

All local recommended safety guidelines followed at the time of interview.

The future is here: an e-commerce and gaming giant teamed up with an Artificial Intelligence (AI) startup in Japan to create an innovative, promising solution to identify – and fight – cancer using QIAGEN's QIAseq Kits for next-generation sequencing.
Cancer is the world’s second leading cause of death and accounted for some 9.6 million deaths in 2018, according to the World Health Organization (WHO). Researchers have long tried to find new biomarkers for cancer diagnostics; however, biological differences in patients mean that no single biomarker is reliable enough for diagnostics. 

One of the companies making groundbreaking developments in this field is the Japanese company PFDeNA Inc, established in 2016 as a partnership between DeNA and Preferred Networks Inc., a Tokyo-based artificial intelligence (AI) company. PFDeNA Inc.’s legacy began some years earlier when Tomoko Namba announced in 2011 that she was stepping down as CEO of DeNA Co., one of Japan’s most successful IT startups, to care for her cancer-stricken husband. When he passed away, Namba’s commitment to fighting cancer was an inspiration for the innovative startup whose mission was to pioneer a new front in the global battle against the disease.
PFDeNA Inc
Established in 2016 as a partnership between DeNA and Preferred Networks Inc., PFDeNA is a Tokyo-based artificial intelligence (AI) company founded in 2014. The joint research project is developing a deep learning-based system that creates a blood test system to detect 14 types of cancers based on a small amount of blood. It aims to extend healthy life expectancy through early cancer detection, particularly in developed countries which are continually faced with increasing cancer rates.
icon-cta-blockquote2
“Just knowing you have cancer is not enough for treatment. You need to know where. We, therefore, want to develop a pan-cancer screening assay.”
Dr. Kiyo Ishikura, Associate Director of healthcare, PFDeNA

From scanning biomarkers to pattern recognition 

As developed countries such as Japan struggle with aging populations and increased incidences of cancer, research shows that AI can detect cancers more quickly and accurately, helping patients get the care they need. The team at PFDeNA and its founding companies are using deep learning to identify common features of miRNA in samples from cancer patients. With anonymized samples from Japan’s National Cancer Center, PFDeNA is working to develop assays that can quickly screen for 14 types of cancer, such as prostate, stomach, colon and esophageal cancer. To do this, the total expression patterns for each extracellular RNA (ExRNA), including miRNA, are examined. 

“Since only one or a few such molecules is not enough to differentiate cancer from healthy cells, we’re targeting hundreds of different kinds of ExRNA for cancer screening,” says Dr. Kiyohide Ishikura, associate director of PFDeNA’s healthcare business. “Just knowing you have cancer is not enough for treatment. You need to know where. We, therefore, want to develop a pan-cancer screening assay. Through a single, conventional blood sample, you will know if you have a likelihood of developing cancer as well as the specific cancer type.” 

Dr. Kiyohide Ishikura
Dr. Kiyo Ishikura is the Associate Director of healthcare at PFDeNA. Beginning his career at Celera Genomics, where he worked on decoding the human genome, he has spent the last 20 years focusing on genomic medicine. After working in global scientific affairs and managing the Danaher’s local marketing for the company's diagnostic business, he introduced NGS into hospitals across the country and has worked with the local oncology community, the government, and the authority on the topic.
icon-cta-blockquote2
“We believe machine learning and deep learning brings much higher sensitivity and specificity than conventional assays for cancer screening.”
Dr. Kiyo Ishikura, Associate Director of healthcare, PFDeNA

Harnessing the power of AI

There are high hopes that patterns of miRNA expression, which are found in easily tested bodily fluids such as blood, can reliably indicate the presence of cancer in different organs. The joint venture harnesses the power of artificial intelligence (AI) to develop a diagnostic system that can identify multiple types of cancer from liquid biopsy samples. Its dynamic AI technique, where algorithms learn from massive volumes of data, is one of the most promising new applications of deep learning. 

PFDeNA analyzes samples with next-generation sequencers, looking at global expression patterns of small ribonucleic acids, mainly micro RNAs (miRNAs). “We believe machine learning and deep learning brings much higher sensitivity and specificity than conventional assays for cancer screening,” says Ishikura, referring to the modern, high-throughput genetic sequencing techniques.

Tatsuya Yamaguchi
Tatsuya Yamaguchi has devoted his career to developing approved diagnostics systems and currently heads Lab Operations in the area of healthcare at PFDeNA. His most recent challenge is to improve the accuracy as well as the reproducibility of miRNA measurements for commercial tests.
icon-cta-blockquote2
“QIAGEN is a vital, reliable partner in our work and has provided us with high-quality, cutting-edge reagents and ensured a stable supply.”
Tatsuya Yamaguchi, head of lab operations, PFDeNA

Next-generation sequencing is here

An essential tool that the teams at PFDeNA are using to build their new screening system is QIAGEN’s QIAseq Kits for next-generation sequencing. These enable researchers to perform differential expression analysis and generate the data that Preferred Networks engineers can use to create deep learning algorithms for pattern recognition.

“QIAGEN is a vital, reliable partner in our work and has provided us with high-quality, cutting-edge reagents and ensured a stable supply,” says Tatsuya Yamaguchi, head of lab operations at PFDeNA. “This is very important because it has allowed us to generate the data necessary to bring deep learning and machine learning in to help tackle this challenge.”

Ishikura believes PFDeNA has what it takes to succeed, with Preferred Networks’ expertise in developing cutting-edge AI solutions, the state-of-the-art Harumi Lab generating quality data, and DeNA’s agile decision-making from its long experience in mobile services. After all, in 2014, the mobile giant launched a direct-to-consumer genetic testing service called MYCODE that has seen about 90% of customers make lifestyle modifications to protect their health.

“We will need to challenge not only regulations in the current medical system, but how it fundamentally works – from a ‘sickcare’ system in which people get sick and then go to hospital to a ‘healthcare’ system based on preventive diagnosis,” says Ishikura. “We believe people will be more driven to maintain good health when much better tools are available to them. Detecting cancer early is an important key to achieving this goal and we believe we can contribute to this.”

PFDeNA Inc.
Different types of cancer show up as different patterns: AI can spot these subtle differences that are too subtle for the human eye to detect and let them be classified by the computer.
Explore innovative QIAseq solutions for your NGS workflows
Achieve effective rRNA removal from mammalian RNA-seq samples for maximum gene expression reads