A case control study of 153 pancreatic cancer patients and 309 healthy people, published November 12, 2024 in eClinicalMedicine, described by Craif as a sister journal of The Lancet. Specificity was 92.9 percent and sensitivity across all stages was 88.2 percent.

Craif Inc., together with researchers at Keio University School of Medicine, Hokuto Hospital and Nagoya University, reported that microRNA recovered from a single urine sample and read by a machine learning model detected early stage pancreatic cancer more accurately than CA19-9, the blood marker in routine clinical use. The study was published on November 12, 2024 in eClinicalMedicine, which the Japanese release describes as a sister journal of The Lancet.

Pancreatic cancer progresses quickly and produces few symptoms while the tumor is small. It is often already too advanced for surgical removal by the time it is found, and the five year survival rate is around 10 percent, one of the worst of any cancer. Early detection is the key to treatment, but current methods make early diagnosis difficult, which is why the Japanese release calls the development of a noninvasive, high accuracy method urgent.

What the study found

At the cutoff that best balanced sensitivity and specificity, the urinary microRNA assay showed:

• Specificity of 92.9 percent.
• Sensitivity of 92.9 percent in Stage I and IIA disease.
• Sensitivity of 88.2 percent across all stages.

CA19-9 had been measured in 140 of the 153 pancreatic cancer patients. Its sensitivity was 37.5 percent in Stage I and IIA and 63.6 percent across all stages. The Japanese release states that CA19-9's predictive ability tended to be lower the earlier the stage, and concludes that urinary microRNA is the better performer for detecting pancreatic cancer at an early stage.

The release also reports that the urinary microRNA patterns reflected not only the tumor tissue itself but changes in the tumor microenvironment. At an early stage, when the tumor is small, the signal coming from the tumor itself is expected to be weak; the authors suggest the assay may be picking up early disease because it uses that broader information as well.

How the assay works

The team focused on extracellular vesicles and the microRNA they carry, which cells use to send information to distant cells and which cancer cells release actively to make their surroundings favorable to growth and to metastasis. The method concentrates extracellular vesicles in urine, which lets a large quantity of microRNA be extracted and analyzed comprehensively. A machine learning model was then built on that urinary microRNA data.

Because the sample is urine, the Japanese release says the method is simple, places little burden on the patient, and allows collection at home. On that basis it says the approach may contribute to broad population screening and to early detection of pancreatic cancer in areas where access to hospitals is limited.

Where the samples came from

The Japanese release describes a joint clinical study run across six medical institutions, and the paper's own title calls the design a case control study.

• Pancreatic cancer patients, 153 in total, enrolled at Hokuto Hospital, Kawasaki Medical School Hospital, the National Cancer Center Hospital, Kagoshima University Hospital and Kumagaya General Hospital.
• Healthy participants, 309 in total, enrolled at Hokuto Hospital and Omiya City Clinic.

Urine was collected from both groups and microRNA and CA19-9 were measured.

What this release does not establish

This is a study result, not a product announcement. The Japanese page makes no regulatory claim, does not describe the assay as a diagnostic, and says nothing about availability outside Japan. Craif's commercial urine test in Japan is a cancer risk test, a Japanese consumer testing category that is not the same regulatory thing as a cleared or approved diagnostic in the United States.

In a case control design, specificity measured against 309 healthy participants is not the false positive rate that would be seen screening an average risk population, and the reported sensitivity is measured in patients already known to have pancreatic cancer. The Japanese page does not state whether the reported figures come from cross validation, from a held out test set, or from the data the model was fitted on.

Terms as the Japanese release defines them

• microRNA: small RNA that regulates biological function. Many kinds exist inside cells and they regulate a range of biological functions.
• Extracellular vesicles: vesicles secreted by cells, 40 to 5,000 nm in diameter.
• CA19-9: a tumor marker in blood used for diagnosis of pancreatic and biliary tract cancers and for monitoring treatment response. Its blood concentration is known to rise as pancreatic cancer progresses, but its sensitivity is low at an early stage, and it also rises in some healthy people and in other diseases, so it has limits for early diagnosis of pancreatic cancer.
• Sensitivity and specificity: sensitivity is the proportion of people with the disease who test positive; specificity is the proportion of people without the disease who test negative.
• Tumor microenvironment: the normal cells, blood vessels, immune cells and extracellular matrix around a cancer cell, which influence the cancer's growth and metastasis and play an important role in progression, treatment resistance and the control of immune response. 

Publication information

Journal: eClinicalMedicine

Title: A noninvasive urinary microRNA-based assay for the detection of early-stage pancreatic cancer: a case control study

Authors: Shogo Baba, MS; Tadatoshi Kawasaki, MD; Satoshi Hirano, MD; Toru Nakamura, MD; Toshimichi Asano, MD; Ryo Okazaki, MD; Koji Yoshida, MD; Tomoya Kawase, MD; Hiroshi Kurahara, MD; Hideyuki Oi, MD; Masaya Yokoyama, MD; Junji Kita, MD; Johji Imura, MD; Kazuya Kinoshita, MD; Shunsuke Kondo, MD; Mao Okada, MD; Tomoyuki Satake, MD; Yukiko Shimoda Igawa, MD; Tatsuya Yoshida, MD; Hiroki Yamaguchi, PhD; Yoriko Ando, PhD; Mika Mizunuma, PhD; Yuki Ichikawa, PhD; Kyoko Hida, MD; Hiroshi Nishihara, MD; Yasutaka Kato, MD.

DOI: https://doi.org/10.1016/j.eclinm.2024.102936

Published: November 12, 2024

Funding

Japan Agency for Medical Research and Development, Advanced Medical Devices and Systems Development Project, grant number JP24he2302007, Development of a simple machine analysis system for urinary microRNA to achieve ultra high precision, noninvasive early cancer detection.

About Craif

Craif is a venture company founded in 2018 that came out of Nagoya University. It holds a core technology, NANO IP (NANO Intelligence Platform), that detects disease related biological substances, including microRNA, with high accuracy from body fluids that are easy to collect, such as urine. Craif uses NANO IP to develop tests aimed at early detection of cancer and at care matched to the individual.

Company details

• Company name: Craif Inc.
• Representative: Ryuichi Onose, Representative Director
• Founded: May 2018
• Business: research and development of next generation tests toward early detection of disease, centered on the cancer field, and toward individualized medicine; provision of the miSignal series of urine based cancer risk tests
• Head office: ITP Hongo Office 5F, 2-25-7 Yushima, Bunkyo City, Tokyo

URL: https://craif.com/