JP2024053899A - CANCER DETECTION DEVICE, CANCER INFORMATION ACQUISITION METHOD, AND PROGRAM - Google Patents

Abstract

[Problem] To provide a cancer diagnosis device, a cancer information acquisition device, and a program that can diagnose early-stage pancreatic cancer and bile duct cancer. [Solution] A cancer diagnosis device (1) includes a determination unit that determines whether or not a subject has bile duct cancer or pancreatic cancer based on bacterial flora data (2) related to bacterial flora in the bile of the subject. [Selected Figure] Figure 1

Description

The present invention relates to a cancer diagnosis device, a cancer information acquisition method, and a program.

Pancreatic and biliary duct cancers have a poor prognosis, with a high proportion of deaths compared to the number of patients. It is known that these cancers have a good prognosis if they are detected early and treatment is started early. For example, the earlier bile duct cancer is detected, the higher the 5-year survival rate. For this reason, it is believed that early diagnosis of pancreatic cancer can contribute to improving survival rates. However, because early pancreatic cancer basically has no symptoms, there are many cases in which pancreatic cancer is already advanced at the time of discovery, except when abnormalities are detected by abdominal echography, endoscopic ultrasound examination, or MR (Magnetic Resonance) cholangiopancreatography, or when it is discovered by computer tomography examination for other diseases, and a method for early diagnosis of pancreatic cancer is needed.

Patent Document 1 discloses a method for testing the risk of cancer by analyzing the intestinal flora, based on the finding that changes in the intestinal flora are associated with the onset of cancer in an analysis of a carcinogenesis model mouse.

Non-patent literature 1 analyzes the relationship between colorectal cancer and intestinal flora, and reports that intestinal flora can be used as a biomarker for determining the prognosis of colorectal cancer.

International Publication No. 2014/126044

Andrew Maltez Thomas, et al., “Metagenomic analysis of colorectal cancer datasets identifies cross-cohort microbial diagnostic signatures and a link with choline degradation”, Nature Medicine, 25, 667-678, 2019

The above-mentioned Patent Document 1 shows that changes in the intestinal flora and an increase in bacteria that produce secondary bile acids are associated with an increase in liver cancer and lung cancer, but does not show an association with pancreatic cancer and bile duct cancer. Furthermore, the subject of research in the above-mentioned Non-Patent Document 1 is colorectal cancer, not pancreatic cancer and bile duct cancer. If there were a biomarker useful for early detection of pancreatic cancer and bile duct cancer, it could contribute to improving the prognosis of pancreatic cancer and bile duct cancer, which are intractable cancers.

The present invention has been made in consideration of the above-mentioned circumstances, and aims to provide a cancer diagnosis device, a cancer information acquisition method, and a program that can diagnose early-stage pancreatic cancer and bile duct cancer.

(Cancer diagnosis device)
The cancer assessment device described in the present specification comprises:
The apparatus includes a determination unit that determines whether or not a subject has bile duct cancer or pancreatic cancer based on information about the bacterial flora in the bile of the subject.

The information about the bacterial flora may include
It may also be the relative abundance of each bacterial genus.

The determination unit is
It may be possible to determine whether or not the subject has bile duct cancer or pancreatic cancer at stage IA or higher.

The determination unit is
When determining whether or not the subject has bile duct cancer, determining whether or not the subject has bile duct cancer using a model generated by supervised learning based on sample data showing a relationship between information on the bacterial flora in the bile of bile duct cancer positive subjects known to have bile duct cancer and bile duct cancer negative subjects known not to have bile duct cancer, and information on whether or not the bile duct cancer positive subjects and bile duct cancer negative subjects have bile duct cancer;
When determining whether or not the subject has pancreatic cancer, whether or not the subject has pancreatic cancer may be determined using a model generated by supervised learning based on sample data showing a relationship between information regarding the bacterial flora in the bile of pancreatic cancer-positive subjects known to have pancreatic cancer and pancreatic cancer-negative subjects known not to have pancreatic cancer, and information regarding whether or not the pancreatic cancer-positive subjects and pancreatic cancer-negative subjects have pancreatic cancer.

The determination unit is
The model may be generated by performing supervised learning using the sample data.

The cancer assessment device described in the present specification comprises:
The apparatus may further include at least one selected from the group consisting of a PCR device that performs PCR, a base sequence analysis unit that analyzes the base sequence of a PCR product, and an analysis unit that analyzes the bacterial flora.

(How to obtain cancer information)
The method for obtaining cancer information described in the present specification includes:
Based on information regarding the bacterial flora in the bile of a subject, information is obtained to assist in determining whether or not the subject has bile duct cancer or pancreatic cancer.

(program)
The program described herein is
Computer,
The device functions as a determination unit that determines whether or not a subject has bile duct cancer or pancreatic cancer based on information about the bacterial flora in the bile of the subject.

The present invention makes it possible to diagnose early-stage pancreatic cancer and bile duct cancer.

1 is a block diagram showing a configuration of a determination device according to an embodiment of the present invention; FIG. 2 is a block diagram showing a hardware configuration of the determination device of FIG. 1 . 2 is a flowchart of a learning process of the determination device of FIG. 1 . 2 is a flowchart of a determination process of the determination device of FIG. 1 . 1 shows the results of a model for pancreatic cancer in an example, in which A shows the distribution of risk scores obtained by a model using a neural network, and B shows a Receiver Operating Characteristic (ROC) curve for the model. 1 shows the results of a model for bile duct cancer in an example, in which A shows the distribution of risk scores obtained with a neural network model, and B shows the ROC curve for the model.

The following describes in detail an embodiment of the present invention with reference to the drawings. In each drawing, the same or equivalent parts are given the same reference numerals.

The cancer assessment device according to this embodiment is a device for assessing whether or not a subject has at least one of bile duct cancer and pancreatic cancer. Throughout this specification, the subject is not particularly limited to an animal having a bile duct and a pancreas, and examples of the subject include mammals such as dogs, cats, pigs, monkeys, and humans, and preferably humans. The subject may be a subject suspected of having bile duct cancer or pancreatic cancer, or a subject not known to have bile duct cancer or pancreatic cancer.

The determination device is constructed on a computer. In other words, the determination device is an information processing device in which information processing by software is specifically realized using hardware resources.

As shown in FIG. 1, the determination device 1 inputs information (bacterial flora data) 2 on the bacterial flora in bile. The bacterial flora data 2 is, for example, the relative abundance of each bacterium contained in the bile of a subject. In classifying bacteria, a phylogenetic classification based on the homology of the base sequence of the gene encoding 16S ribosomal RNA (16S rRNA gene) can be used. Classification levels include phylum, class, order, family, genus, and species. Preferably, the bacterial flora data 2 is the relative abundance of each genus of bacteria contained in bile. The type of bacteria used as the bacterial flora data 2 may be selected from bacteria that are statistically associated with the presence or absence of pancreatic cancer or bile duct cancer, or may be all of the bacteria contained in bile.

The bacterial flora data 2 is obtained by a known method from bile as a specimen obtained from a test subject. Bile may be collected, for example, using a duodenal tube or from a resected gallbladder sample. First, DNA is extracted from the bile, and the 16S rRNA gene contained in the extracted DNA is amplified by polymerase chain reaction (PCR). Next, the base sequence of the obtained amplified product (PCR product) is determined by a next-generation sequencer or the like. Based on the base sequence of the PCR product, the relative abundance according to the classification level of each bacterium can be calculated by clustering or the like. For example, genus-level classifications of bacteria include Veillonella, Rothia, Acidaminococcus, Citrobacter, Prevotella, Pseudomonas, Lactobacillus, Atobium, Streptococcus, Staphylococcus, Pyramidobacter, Acinetobacter, Pseudoxanthomonas, Corynebacterium, Pleomorphomonas, Saccharopolyspora, and Neisseria.

The determination device 1 performs calculations on the bacterial flora data 2 and outputs information (determination information) 3 on whether or not the subject has at least one of bile duct cancer and pancreatic cancer. The determination information 3 includes information indicating whether or not the subject has bile duct cancer or pancreatic cancer, a numerical value correlating with the risk of the subject having bile duct cancer or pancreatic cancer, and information indicating whether or not the subject has bile duct cancer or pancreatic cancer at a predetermined stage or higher. The stage of the disease is not particularly limited to a classification used in clinical practice, and is, for example, a classification based on the TNM Classification of Malignant Tumors (UICC). In the case of pancreatic cancer, it may be a classification based on the Pancreatic Cancer Treatment Guidelines (edited by the Japan Pancreas Society). Preferably, the determination information 3 is information indicating whether or not the subject has bile duct cancer or pancreatic cancer at stage IA or higher. In the case of bile duct cancer, stage IA means "single tumor with a maximum diameter of 2 cm or less, no vascular invasion or main bile duct invasion." In the case of pancreatic cancer, it means "a tumor limited to the pancreas with a maximum diameter of 2 cm or less."

The configuration of the determination device 1 will be described in more detail using an example in which the determination device 1 determines whether or not a test subject has bile duct cancer. In the following, a test subject that is the subject of the determination is defined as a bile duct cancer positive subject, and a test subject that is determined not to have bile duct cancer is defined as a bile duct cancer negative subject. As shown in FIG. 1, the determination device 1 includes a determination unit 10. The determination unit 10 determines whether or not a test subject has bile duct cancer based on the bacterial flora data 2 of the test subject. In detail, the determination unit 10 determines whether or not a test subject has bile duct cancer using a model generated by supervised learning based on sample data 30 showing a relationship between the bacterial flora data of the bile duct cancer positive subject and the bile duct cancer negative subject and information on whether or not the bile duct cancer positive subject and the bile duct cancer negative subject have bile duct cancer.

Supervised learning is a machine learning technique that uses a set of combinations of explanatory variables and associated objective variables as training data and performs learning by fitting to the training data. Fitting is performed by extracting the features of the explanatory variables included in the training data, selecting features for each objective variable, extracting features of the data belonging to that objective variable, and generating criteria for identifying the objective variable. Fitting generates a model that outputs an objective variable that should correspond to an input explanatory variable from the explanatory variable. The model can output an objective variable that corresponds to a new explanatory variable not included in the training data.

The sample data 30 corresponds to learning data, and the explanatory variables are bacterial flora data obtained from the bile of cholangiocarcinoma-positive subjects and cholangiocarcinoma-negative subjects, and the objective variable is information on whether or not the cholangiocarcinoma-positive subjects and cholangiocarcinoma-negative subjects have cholangiocarcinoma, i.e., in the case of a cholangiocarcinoma-positive subject, information indicating that the subject has cholangiocarcinoma, and in the case of a cholangiocarcinoma-negative subject, information indicating that the subject does not have cholangiocarcinoma. The model used by the determination unit 10 is a model generated as sample data 30, which is a set of combinations of bacterial flora data obtained from the bile of multiple cholangiocarcinoma-positive subjects and cholangiocarcinoma-negative subjects for learning, and information on whether or not each cholangiocarcinoma-positive subject and each cholangiocarcinoma-negative subject has cholangiocarcinoma.

For example, the relative abundances of bacteria M1, M2, M3, ... Mn are r1, r2, r3, ... rn, and r1 to rn are explanatory variables. The objective variable is information indicating whether or not a cholangiocarcinoma-positive subject and a cholangiocarcinoma-negative subject have cancer. When selecting bacteria that are statistically related to the presence or absence of cholangiocarcinoma, for example, the explanatory variables are the relative abundances of bacterial genera G1, G2, and G3 in the bile of a cholangiocarcinoma-positive subject or a cholangiocarcinoma-negative subject, and the objective variable is information indicating whether or not the cholangiocarcinoma-positive subject or the cholangiocarcinoma-negative subject has cancer.

The determination unit 10 inputs the bacterial flora data 2 obtained from the bile of the subject into a model generated in advance, and obtains as output determination information 3 indicating whether or not the subject has bile duct cancer. In this way, the determination unit 10 determines whether or not the subject has bile duct cancer.

Any known model may be used as the model generated by supervised learning. Examples of models for supervised learning include linear regression, linear classification, logistic regression, support vector machine, random forest, decision tree, neural network, convolutional neural network, perceptron, deep learning, and k-nearest neighbor method. Various frameworks such as R, Jubatus, Theano, and TensorFlow can be used for supervised learning. Preferably, the model is a neural network. When binary data is used as information indicating whether or not a subject has cholangiocarcinoma, the model serves as a classifier that outputs whether or not the subject has cholangiocarcinoma.

The determination unit 10 generates a model by performing supervised learning using sample data 30. In practice, the sample data 30 is obtained from a plurality of bile duct cancer positive subjects and bile duct cancer negative subjects. When determining bile duct cancer, the accuracy of the determination can be improved by using a large number of sample data 30 obtained from bile duct cancer negative subjects (e.g., patients with benign diseases accompanied by inflammation in the bile duct, such as cholelithiasis, sclerosing cholangitis, and chronic pancreatitis) and bile duct cancer positive subjects having bile duct cancer at each stage. Note that, when determining pancreatic cancer, the accuracy of the determination can be improved by using a large number of sample data 30 obtained from liver cancer positive subjects and pancreatic cancer positive subjects having pancreatic cancer at each stage.

The determination device 1 shown in Fig. 1 is realized, for example, by a computer having the hardware configuration shown in Fig. 2 implementing a software program. Specifically, the determination device 1 is composed of a CPU (Central Processing Unit) 21 that controls the entire device, a main memory unit 22 that operates as a work area for the CPU 21, an external memory unit 23 that stores the operating program of the CPU 21, an operation unit 24, a display unit 25, a communication interface 26, and an internal bus 27 that connects these.

The main memory unit 22 is composed of RAM (Random Access Memory) and the like. The CPU 21's judgment program 28, machine learning program 29, sample data 30, and bacterial flora data 2 are loaded into the main memory unit 22 from the external memory unit 23. The main memory unit 22 is also used as a working area (temporary data storage area) for the CPU 21.

The external storage unit 23 is composed of a non-volatile memory such as a flash memory and a hard disk. A judgment program 28 to be executed by the CPU 21 and a machine learning program 29 are stored in advance in the external storage unit 23. The external storage unit 23 also stores sample data 30 and bacterial flora data 2 obtained from the bile of the subject.

The operation unit 24 is composed of devices such as a keyboard and a mouse, and an interface device that connects these devices to the internal bus 27.

The display unit 25 is composed of display devices such as a CRT (Cathode Ray Tube) and an LCD monitor. The display unit 25 displays the judgment information 3 and the like.

The communication interface 26 is an interface for transmitting and receiving data to and from external devices. Sample data 30 and bacterial flora data 2 are transmitted and received via the communication interface 26.

Next, a cancer information acquisition method realized by the operation of the determination device 1 according to this embodiment, i.e., the processing executed by the determination device 1, will be described. The processing executed by the determination device 1 is divided into a learning processing shown in FIG. 3 and a determination processing shown in FIG. 4.

First, the learning process will be described. In the learning stage, the machine learning program 29 is loaded from the external storage unit 23 into the main storage unit 22, and the CPU 21 executes the machine learning program 29 to perform the process.

As shown in FIG. 3, first, the determination device 1 reads the sample data 30 (step S1). Here, the sample data 30 is read from the external storage unit 23 into the main storage unit 22.

Then, the judgment unit 10 performs supervised learning using the sample data 30 (step S2). The model generated by the supervised learning is stored in the main memory unit 22 and the external memory unit 23 for use in the judgment program 28 described later. When step S2 is completed, the judgment device 1 ends the learning process.

Next, the judgment process will be described. In the judgment process, the judgment program 28 is loaded from the external storage unit 23 into the main storage unit 22, and the CPU 21 executes the judgment program 28 to perform the process.

As shown in FIG. 4, first, the determination device 1 reads the bacterial flora data 2 (step S11). Here, the bacterial flora data 2 is read from the external memory unit 23 to the main memory unit 22. Next, the determination unit 10 determines whether or not the subject has bile duct cancer from the bacterial flora data 2 according to the model generated by supervised learning, and the determination information 3 is output (step S12). When step S12 is completed, the determination device 1 ends the determination process.

The determination device 1 according to this embodiment determines whether or not a subject has bile duct cancer based on bacterial flora data 2 acquired from bile. As shown in the following examples, bile duct cancer can be determined with high accuracy based on information about the bacterial flora in bile.

In this embodiment, the bacterial flora data 2 may be the relative abundance of each bacterial genus. The relative abundance of bacteria can be relatively easily determined using a commercially available kit, etc., so it is possible to easily determine whether or not the subject has bile duct cancer. The determination unit 10 may also determine whether or not the subject has pancreatic cancer or bile duct cancer of stage IA or higher. Stage IA pancreatic cancer and bile duct cancer have a maximum diameter of 2 cm or less. The determination device 1 can determine intractable pancreatic cancer and bile duct cancer even at such an early stage.

In the present embodiment, the determination device 1 generates a model in the determination unit 10, but the determination unit 10 may determine whether or not the subject has bile duct cancer using a model that is generated outside the determination device 1 and pre-stored in the external storage unit 23, rather than using the generated model.

The determination information 3 is information for assisting in the determination of whether or not the subject has bile duct cancer. Therefore, in another embodiment, a cancer information acquisition method is provided, which includes an acquisition step of acquiring information for assisting in the determination of whether or not the subject has bile duct cancer based on the bacterial flora data 2 of the subject.

The above describes the configuration of the determination device 1 when determining whether or not a subject has bile duct cancer. However, for the configuration of the determination device 1 when determining whether or not a subject has pancreatic cancer, the above description of the configuration of the determination device 1 when determining whether or not a subject has bile duct cancer can be referred to by replacing bile duct cancer with pancreatic cancer. When determining whether or not a subject has pancreatic cancer, the determination unit 10 determines whether or not the subject has pancreatic cancer based on the bacterial flora data 2 of the subject. In detail, the determination unit 10 determines whether or not the subject has pancreatic cancer based on a model generated by supervised learning based on sample data 30 indicating the relationship between the bacterial flora data of pancreatic cancer positive subjects and pancreatic cancer negative subjects and information on whether or not pancreatic cancer positive subjects and pancreatic cancer negative subjects have pancreatic cancer.

The hardware and software configurations of the determination device 1 described above are merely examples, and can be changed or modified as desired. For example, the determination device 1 may further include at least one selected from the group consisting of a PCR device that performs PCR, a base sequence analysis unit that analyzes the base sequence of the PCR product, and an analysis unit that analyzes the bacterial flora. A known thermal cycler or the like may be used as the PCR device. The base sequence analysis unit may employ a known method that applies the Sanger method, for example. It is preferable to use a next-generation sequencer for analyzing the base sequences of many DNA fragments such as PCR products. The analysis unit statistically analyzes the base sequences output from the base sequence analysis unit using a known method, and calculates the relative abundance of each bacterium according to the classification level. The PCR device, the base sequence analysis unit, and the analysis unit may operate under the control of the CPU 21, or may be controlled by another CPU.

In another embodiment, a method for early diagnosis of cholangiocarcinoma is provided. The method for early diagnosis of cholangiocarcinoma determines whether or not a subject has cholangiocarcinoma based on information about the bacterial flora in the bile of a subject, using a model generated by supervised learning based on sample data showing a relationship between information about the bacterial flora in the bile of cholangiocarcinoma-positive subjects and cholangiocarcinoma-negative subjects and information about whether or not cholangiocarcinoma-positive subjects and cholangiocarcinoma-negative subjects have cholangiocarcinoma, and diagnoses subjects determined to have cholangiocarcinoma as having cholangiocarcinoma.

The above-mentioned method for early diagnosis of bile duct cancer may include obtaining information about the bacterial flora in the bile of the subject. Obtaining information about the bacterial flora in the bile of the subject may include obtaining bile from the subject, amplifying the 16S rRNA gene of bacteria in the bile of the subject by PCR, and determining the base sequence of the PCR amplified product of the 16S rRNA gene of the bacteria in the bile by a sequencer. In this specification, the information about the bacterial flora in the bile may be base sequence information of the PCR amplified product of the 16S rRNA gene of bacteria in the bile of the subject.

The core part of the processing of the determination device 1, which is composed of the CPU 21, main memory 22, external memory 23, operation unit 24, display unit 25, communication interface 26, internal bus 27, etc., can be realized by using a normal computer system, not a dedicated system. For example, the computer program for executing the above operations may be stored and distributed on a computer-readable recording medium (such as a flexible disk, a CD-ROM (compact disc read only memory), or a DVD-ROM (digital versatile disk read only memory)), and the determination device 1 that executes the above processing may be configured by installing the computer program on a computer. Also, the determination device 1 may be configured by storing the computer program in a storage device of a server device on a communication network such as the Internet, and a normal computer system downloading the computer program.

When the functions of the determination device 1 are realized by sharing between an OS (operating system) and an application program, or by cooperation between an OS and an application program, only the application program portion may be stored in a recording medium or storage device.

It is also possible to superimpose a computer program on a carrier wave and distribute it over a communications network. For example, the computer program may be posted on a bulletin board system (BBS) on the communications network and distributed over the network. The computer program may then be started and executed under the control of the OS in the same way as other application programs, thereby enabling the above-mentioned processing to be performed.

The bile duct cancer and pancreatic cancer determined by the above-mentioned determination device 1 and cancer information acquisition method may be early-stage cancer. In the case of bile duct cancer, early-stage cancer is, for example, T=1 cancer in the TNM classification (single tumor with a maximum diameter of 2 cm or less, no vascular invasion or main bile duct invasion). In the case of pancreatic cancer, early-stage cancer is, for example, T=1 cancer in the TNM classification (tumor limited to the pancreas with a maximum diameter of 2 cm or less).

The present invention will be explained in more detail with reference to the following examples, but the present invention is not limited to these examples.

[DNA extraction]
Using the QIAamp PowerFecal DNA Kit (Qiagen), DNA was extracted from bile samples from subjects with a definitive diagnosis of pancreatic cancer and bile duct cancer. The yield of DNA after extraction was confirmed using the Qubit 3 system (Invitrogen). Samples with a DNA concentration of 10 ng/ml or less were excluded.

[Creation of next-generation sequencer analysis library]
The bacterial 16S rRNA gene was amplified by PCR using KAPA HiFi HotStart ReadyMix DNA polymerase (Roche Diagnostics) and Pro341F/Pro805R (V3-V4) primers (SEQ ID NOs: 1 and 2) designed with Nextera overlap adapters (Illumina). Next, a second PCR was performed using KAPA HiFi HotStart ReadyMix DNA polymerase (Roche Diagnostics) to attach dual indexes and Illumina sequencing adapters with Nextera XT index primers (Integrated DNA Technologies). The PCR products obtained in the second PCR were purified using AMPure XP beads (Beckman Coulter).

[Sequence decision]
The PCR products were quantified using the Qubit 3 system, and equimolar ratios were pooled from each sample, and the base sequences were determined using the MiSeq System (Illumina).

[Data Mining]
Raw reads from the sequencer were filtered and trimmed according to barcode and primer sequences using the MiSeq system, and then quality assessed, filtered, and chimeras detected using the DADA2 pipeline in QIIME2 [34] v2019.4.

[Bacterial flora analysis]
The PCR product variants were classified using 99% clustering of SILVA 132 (https://www.arb-silva.de/download/archive/) in QIIME2 [34] v2019.4. The relative abundance (per million) of all sequences as bile bacterial flora information was then calculated using qiime feature-table rarefy in QIIME2.

[Construction of an early diagnosis model]
For bile duct cancer, bile bacterial flora information at the Genus level and clinical information (TNM classification: handling guidelines, 6th edition) were used as learning data, and machine learning was performed using a neural network to determine whether T was 0 or 1 (single lesion, 2 cm or less in maximum diameter, no vascular invasion or main bile duct invasion (Stage IA)).

For pancreatic cancer, bile bacterial flora information at the Genus level and clinical information (TNM classification: handling guidelines, 6th edition) were used as learning data, and machine learning was performed using a neural network to determine whether T is 0 or 1 (the tumor is confined to the pancreas and the maximum diameter is 2 cm or less (stage IA)).

[result]
The risk scores for pancreatic cancer at T=0 and T≧1 are shown in FIG 5A. FIG 5B shows the ROC curve of leave-one-out cross-validation (LOOCV) of the neural network model for pancreatic cancer. The sensitivity, specificity and predictive value for determining pancreatic cancer were 76%, 85% and 82%, respectively.

Figure 6A shows the risk scores for T=0 and T≧1 for bile duct cancer. Figure 6B shows the ROC curve of the LOOCV of the neural network model for bile duct cancer. The sensitivity, specificity, and predictive value for diagnosing bile duct cancer were 84%, 88%, and 86%, respectively.

This invention allows for various embodiments and modifications without departing from the broad spirit and scope of the invention. Furthermore, the above-described embodiment is intended to explain the invention and does not limit the scope of the invention. In other words, the scope of the invention is indicated by the claims, not the embodiments. Furthermore, various modifications made within the scope of the claims and within the scope of the meaning of the invention equivalent thereto are considered to be within the scope of the invention.

1 Determination device, 2 Bacterial flora data, 3 Determination information, 10 Determination unit, 21 CPU, 22 Main memory unit, 23 External memory unit, 24 Operation unit, 25 Display unit, 26 Communication interface, 27 Internal bus, 28 Determination program, 29 Machine learning program, 30 Sample data

Claims (8)

A determination unit that determines whether or not a subject has bile duct cancer or pancreatic cancer based on information about a bacterial flora in the bile of the subject.
Cancer diagnosis device. The information about the bacterial flora may include
The relative abundance of each bacterial genus,
The cancer diagnosis device according to claim 1 . The determination unit is
Determine whether the subject has stage IA or higher bile duct cancer or pancreatic cancer;
The cancer diagnosis device according to claim 1 or 2. The determination unit is
When determining whether or not the subject has bile duct cancer, determining whether or not the subject has bile duct cancer using a model generated by supervised learning based on sample data showing a relationship between information on the bacterial flora in the bile of bile duct cancer positive subjects known to have bile duct cancer and bile duct cancer negative subjects known not to have bile duct cancer, and information on whether or not the bile duct cancer positive subjects and bile duct cancer negative subjects have bile duct cancer;
When determining whether or not the subject has pancreatic cancer, the method determines whether or not the subject has pancreatic cancer using a model generated by supervised learning based on sample data showing a relationship between information on bacterial flora in bile of pancreatic cancer-positive subjects known to have pancreatic cancer and pancreatic cancer-negative subjects known not to have pancreatic cancer, and information on whether or not the pancreatic cancer-positive subjects and pancreatic cancer-negative subjects have pancreatic cancer.
The cancer diagnosis device according to claim 1 or 2. The determination unit is
generating the model by performing supervised learning using the sample data;
The cancer diagnosis device according to claim 4. The apparatus further includes at least one selected from the group consisting of a PCR device for performing PCR, a base sequence analysis unit for analyzing the base sequence of a PCR product, and an analysis unit for analyzing the bacterial flora.
The cancer diagnosis device according to claim 1 or 2. obtaining information for assisting in determining whether or not the subject has bile duct cancer or pancreatic cancer based on information regarding the bacterial flora in the bile of the subject;
How to get information about cancer. Computer,
and functioning as a determination unit that determines whether or not a subject has bile duct cancer or pancreatic cancer based on information about the bacterial flora in the bile of the subject.
program.