JPH0451943A - Cancer diagnostic device - Google Patents

Abstract

PURPOSE:To decide the cancer with high accuracy by executing a decision of high/low values for executing a cancer decision, and a discrimination analysis for processing a result of measurement in accordance with a prescribed discrimination algorithm and executing a decision based on combination of the result of measurement. CONSTITUTION:As a result of measurement by A reagent, a measured value of each tumor marker is recorded as data in a floppy disk 1 together with body data of a person to be examined. The floppy disk 1 is sent to a cancer deciding organization of a university, etc., and read by a reader 2 provided on the cander deciding organization. The read data is subjected to decision processing in accordance with a prescribed algorithm by a decision processor 3 consisting of a personal computer, etc., and an inspection report is printed out by a printer 4. The printed-out report is fed back to a medical practitioner, etc., from the cancer deciding organization, and transmitted to the person to be examined. According to this method, the person to be examined whose possibility of a cancer is small can be excluded from in a tumor marker high value group.

Description

DETAILED DESCRIPTION OF THE INVENTION <Industrial Application Field> The present invention relates to a cancer diagnostic device that measures a plurality of tumor markers using reagents and determines cancer based on the measured values.

<Prior Art and Problems to be Solved by the Invention> Conventionally, when examining cancer, a plurality of tumor markers possessed by a subject are measured using predetermined reagents. In processing the measurement results, in order to detect cancer patients early, if even one tumor marker is positive, it is indicated as a possibility of malignancy, and a detailed examination is performed each time.

However, this method has a high probability of being diagnosed as positive even if it is not actually cancer, that is, the false positive rate is high.
Even if all tumor markers are low (negative), a false negative result may occur in which the patient actually has cancer.

Therefore, there has been a desire for a cancer diagnostic device that can reduce the false positive rate and false negative rate and reliably detect only true cancer patients. To this end, it is considered essential to perform multivariate analysis of the measurement results of each tumor marker to improve diagnostic accuracy.However, how to actually select and measure tumor markers and the measurement results is unclear. How to combine and process them is only being studied through trial and error.

The purpose of the present invention is to focus on several specific types of tumor markers and to perform better analysis based on the measured values of these tumor markers, in order to detect cancer with higher certainty than in the past. It is an object of the present invention to provide a cancer diagnostic device with higher accuracy that employs this method.

Means for Solving the Problems> The cancer diagnostic device of the present invention for achieving the above objects has the following features:
CA125. TPAf group 1j lipbutide anti-j
X), FER (ferritin), CE A
(CEA-5), AFP (a-1
a recording means for respectively recording the measurement results for six types of tumor markers, 5LX (allele le''-i antigen); a reading means for reading each measurement result from the recording means; A means for comparing the measurement data read from the means with reference values corresponding to the age category of the subject and dividing the measurement results into groups for each tumor marker, and a means for making a cancer diagnosis based on the grouping of the measurement results. a value group determination means; a discriminant analysis means for processing the read data in accordance with a predetermined discrimination algorithm based on statistical data of actual cancer patients to determine cancer based on a combination of tumor markers; /Low value group determination means 1 Comprehensive determination means for performing comprehensive cancer determination from the results of the discriminant analysis means.

Effects> The above device performs cancer judgment using the direct/low value group judgment means, and by combining the discriminant analysis that analyzes trends in actual cancer patient data, it is possible to identify cancer markers from among the high tumor marker group. Subjects with a low possibility of cancer can be excluded. By integrating the analysis, it is possible to perform cancer determination with a low incidence of false positives and false negatives.

<Example> Hereinafter, the case of diagnosing ovarian cancer will be described in detail with reference to the accompanying drawings. However, it should be noted at the outset that the present invention is applicable not only to ovarian cancer but also to other cancer diagnoses.

FIG. 3 is a block diagram showing an example of the configuration of a cancer diagnostic device aimed at diagnosing ovarian cancer.

The test subject's blood is sent to the testing institution through a medical practitioner, etc.
CA125. TPA, FER, CEA.

Tumor markers AFP and SLX are measured. Examples of reagents include CA125 two cent core CA125RIA kit (normal value 35U/m
1 or less) (TFB Toray Fuji Bionics Co., Ltd. TPA (Tissue polypeptide
e antigen) Niflorifigen TPA Kit “Daiichi ■Tissue Voribetide Antigen (normal value 100U/l or less) (Daiichi Radioisotope Research Institute) FER (Ferrittn): Ferritin Kit “Daiichi” (normal value 15-220ng
/-1 male; 10-80ng/m1 female) (Daiichi Radioisotope Research Institute) CE A (Carcinoevbryor+tea
ntigen): CEA rear beads R (normal value 2.
5 ng71 or less) (Guina Hot Co., Ltd.) AFP (Alpha retoprotein)
: α-Funito Lia beads (normal value 10ng/11
Below) (Dynabot Co., Ltd.) S L X (Sialyl Le” -4): 5
LXr Otsuka" (normal value 38/1 or less) (Otsuka Pharmaceutical Co., Ltd.).

As a result of measurement using these reagents, #I of each tumor marker
The fixed value is used as data, and the third
The information is recorded on the floppy disk 1 shown in the figure.

Floppy disk 1 will be sent to a cancer assessment institution such as a university.
It is read by a reading device 2 provided at a cancer diagnosis institution.

The read data is subjected to judgment processing according to a predetermined algorithm by a judgment/processing device 31 consisting of a personal computer or the like, and an inspection report is outputted to the printer 4. The generated report is fed back from the cancer assessment organization to a medical practitioner, etc., and then communicated to the patient.

The determination processing algorithm will be explained in detail below.

First, the measured values of each tumor marker are divided into groups according to the age category of the subject (Tables 1 to 6). This is because the normal value range of tumor markers fluctuates. (Margins below) Table 2 (Margins below) Table 1 (Margins below) Table 3 (Margins below) Table 4 Table 5 (Margins below) (Margins below) Table 6 Figure 1 shows the overall process. It is a figure showing an outline. If only the above grouping is possible, the subject will have groups ■ to ■ for each tumor marker, but if one of the five tumor markers excluding SLX is in group ■, then the tumor marker The high value group judgment algorithm is followed as all the values are not normal values. In the high value group determination algorithm, discriminant analyzes A and B are executed depending on the case.

All groups of 5 tumor markers except SLX
If it is ■, it is assumed that all the tumor marker values are normal values independently, and the determination algorithm for the low value group is followed. In the low value group determination algorithm, discriminant analyzes B, C, and D of discriminant analysis are executed depending on the case. Finally, a comprehensive judgment is made for inclusion in the inspection report. The comprehensive judgment includes any of judgments 1 to 9, and comments may be added depending on the case.

In the high value group determination algorithm (see Figure 2 (A)),
First, check whether it is CA125 or 300 or higher. CA1
25 is investigated first because the factor analysis results indicate that it is most closely related to CA125 or cancer. If it is 300 or more, determination 7 is performed. The content of judgment 7 is, "There is a possibility that you have ovarian cancer. Please undergo a detailed examination." Additionally, comment A is added: ``Please perform ovarian cancer screening frequently and carefully monitor the patient's progress, keeping in mind the possibility of malignant disease.''

If it is less than 300, then TPA, which is an effective tumor marker for diagnosing cancer, is examined after CA125 as a result of factor analysis. i.e. TPA or under 50 years old, 50-59 years old, 6
Standard values set according to the age classification of 0 to 69 years old and 70 years old and above (110,133,2°120.9.144.5
) or more, and if it is, cancer determination 7 is performed. If it is below the standard value, CA125. TPA,
Mahalanobis general distance and its posterior probability from five tumor markers: FER, CEA, and AFP (Haruo Yanai, Yoshio Takane,
``Multivariate Analysis Method,'' Modern Human Statistics, Breakfast Shoten, 1977, Kawaguchi Osamube: ``Introduction to Multivariate Analysis,'' Morikita Publishing, 1973)
is calculated, and the result is used to make a determination. That is, the general distance between the test subject and the non-cancer group and the general distance between the test subject and the cancer group are calculated, and the post-hoc probability PR that the test subject has cancer is determined by discriminant analysis A and B (described later), respectively. demand. Then, if at least one PRc is 0.5 or more, determination 7 of cancer is performed, and if both are less than 0.5, determination 6 of non-cancer is performed. This judgment 6 states, ``There is a tumor marker that showed a high value, so please re-examine in 1 to 2 months after your menstrual period ends.'' and ``Ovarian cancer screening will be conducted in 1 to 2 months.'' Please add comment B.

In the low value group determination algorithm (see Figure 2 (B)),
Based on the questionnaire, the doctor checks whether the pelvic examination is normal, and if both the uterus and ovaries are normal, the PRc is 0 based on discriminant analysis B.
Check if it is 5 or more. If it is less than that, there is almost no possibility of cancer, and the statement says, ``All the tumor marker values are normal, and the results of the tumor marker combination test have also been determined to be non-cancerous, so please follow the comprehensive evaluation.'' Add comment E.

If PRc according to discriminant analysis B is 0.5 or more, it indicates that there is a possibility of ovarian cancer, and comment C is added saying "Please perform ovarian cancer examination in 3 months."

If the uterus or ovaries are not normal as a result of the internal examination, it is determined whether PRo is 0.5 or more using discriminant analysis C and discriminant analysis D (described later). If both are above, comment C is added as there is a possibility of cancer.

If either of these values is less than 0.5, there is a slight possibility that you have cancer, so please conduct an ovarian cancer screening in 4-5 months.

” is added with comment D. If both values are less than 0.5, comment E is added because there is almost no possibility of cancer.

Next, the contents of the discriminant analyzes A, B, C, and D will be explained. However, each measurement value of the tumor marker of the subject was
, TPA-X2. Fr-X3, CEA-X4. A
FP-X5, Xl, X2. X3. X4. Let X be a five-dimensional vector having X5 as a component.

(1) Discriminant analysis A is a discriminant analysis for distinguishing between a non-cancer group and an ovarian cancer group in the high tumor marker group.

The discriminant is determined under the condition that 5false positive<15%.

According to Y-X-Kl (Kl is a five-dimensional constant vector), Yl, Y2 . Y3. Y4. Find the vector Y whose component is Y5, and calculate the Mahalanobis general distance D with the non-cancer group.
1 to D 1- Y'A-'Y-2Ln(0,04)
Find it with A is a 5-by-5 matrix consisting of constant components.

Next, we set the formula Z-X- to 2 (K2 is a five-dimensional constant vector), calculate the vector Z whose components are 21, 22.23, and 24.25, and calculate the Mahalanobis generalized distance D2 with respect to the ovarian cancer group. D2-Z'A-IZ-2Ln(1-0,
04).

The post-establishment PRo is is required.

(2) Discriminant analysis B is a discriminant analysis for distinguishing between the non-cancer group and the ovarian cancer group in both the low tumor marker value group and the high tumor marker value group. tatashi, false positive is 13
Approximately %.

First, Yl, Y2. Y3. Y4. Find the vector Y with Y5 as a component, and calculate the Mahalanobis generalized distance D1 to the non-cancer group by D 1- Y'B -'Y-2Ln(0,0(10
03) Find it. B is a 5-by-5 matrix consisting of a fixed number of components.

Next, the formula Z-X- is changed to 4 (K4 is a 5-dimensional constant vector), Takatte, Zl, Z2. Z3. Z4. Vector 2 with Z5 as a component is calculated, and the Mahalanobis general distance D2 with respect to the ovarian cancer group is calculated as D 2- Z 'B -' Z -2Ln(1-0,
00003).

The post-establishment PRC is Find it with

The post-establishment PRC is is required.

(3) Discriminant analysis C is a discriminant analysis for distinguishing between the uterine benign tumor group and the ovarian cancer group in both the low tumor marker value group and the high tumor marker value group. Tatami, ralse positive
Find the discriminant under the condition that e<16%.

First, Yl, Y2 . Y3. Y4. A vector Y having Y5 as a component is determined, and the Mal\Lanobis generalized distance D1 to the non-cancer group is determined as D1-Y'C-'Y-2Ln (0,005). C is a 5-by-5 matrix consisting of a fixed number of components.

Next, the formula Z-X-KG (K6 is a five-dimensional constant vector), Zl, 22. z3. Z4. Find the vector Z with Z5 as a component, and calculate the mano1lanobis generalized distance D2 to the ovarian cancer group as D2-Z'C-'Z-2Ln(1-0,0
05).

(3) Discriminant analysis is a discriminant analysis to distinguish between the uterine benign tumor group and the ovarian cancer group in the group with low tumor marker values. demand.

First, we want to set Y−X− to 7 (K7 is a 5-dimensional constant t\tukuru), and then Yl, Y2. Y3. Y4. A vector Y having Y5 as a component is obtained, and the malX Lanobis general distance MD1 with the non-cancer group is calculated as D1. -Y'D-'Y-2Ln(0,38
). p is a 5-by-5 matrix consisting of a fixed number of components.

Next, in the formula Z-X-8 (K8 is a 5-dimensional constant vector) Nishi Takate, Zl, Z2. Z3. Z4. Find the vector Z with Z5 as a component, and calculate the Mal\Lanobis general distance D2 to the ovarian cancer group as D2-Z'D-'Z-2Ln(1-0,
38).

The post-establishment PRc is is required.

As described above, there is a single judgment in which the measured values of each tumor marker are evaluated and determined by each group ■ to By combining the combination determination made from a certain post-establishment PRc, it becomes possible to perform cancer determination with higher probability.

In addition, when CEA is Group ■, we add the comment ``Please also undergo a test for uterine cancer,'' and when TPA is Group ■, or FER is under 50 years old and Group ■, we add the comment ``Please undergo a test for pancreatic cancer.'' If the TPA is in group ■, a comment such as "Please also undergo a breast cancer examination" is added.

Furthermore, if the values of all six types of tumor markers are croup ■ or ■, it will say, ``There is almost no possibility of ovarian cancer. The latter is fine.'', and if it is FER or Croup■,
Judgment 2 is issued, which says, ``Please check for anemia.'' If the examinee's questionnaire answers, ``I don't know if you are pregnant,'' and the test result is CA125 or croup, Please check” Judgment 3
issue. Furthermore, if you are in CA125 or group ■, please check again to see if you have endometriosis or adenomyosis.If not, please re-examine after your menstrual period ends in 1-2 months. ” is issued as a judgment 4. If not all of the six types of tumor markers belong to group ■, but even one of them belongs to group ■, then the doctor will say, ``There is almost no possibility of ovarian cancer, but the next test will be done in about 6 months, after menstruation ends.'' Please accept the request.'' Judgment 5 is issued. However, when determination 1 and determination 5 overlap, priority is given to determination 5. Judgment 8 is “Possibility of tumor other than ovarian cancer (
Liver cancer, intestinal cancer, lung cancer) may also be suspected, so please undergo a thorough examination. ”, and AFP is a group■,
Liver cancer when TPA is group ■ or under 50 years old and FER is group ■; intestinal cancer when CEA is group ■; C
Lung cancer is displayed when EA is Group ■, TPA is Group ■, or FER is under 50 years old or Group ■. However, when determination 6 overlaps determination 1, determination 4, and determination 5, priority is given to determination 6. When judgment 6 and judgment 7 overlap, judgment 7 is given priority. When Judgment 6, Judgment 7, and Judgment 8 overlap, Judgment 7 and Judgment 8 are given priority.

Next, a determination example will be shown. When 10,341 cases were diagnosed using this system in Shizuoka Prefecture, the results were as shown in Table 7. (Margins below) From the table above in Table 7, the correlation between the judgments made by this system and the results of detailed medical examinations is quite high, supporting the high reliability of this system.

In the above example, 53-year-old subject (1) is CA125-
230. TPA-35,2, FER-13゜2, CEA
-1,0, AFP-5,0, and SLX-22,3, and only CA125 was abnormally high, but as a result of the combination determination, it was determined to be non-cancerous. When I underwent a re-examination, it was revealed that it was just a uterine fibroid. Subject (2), 53 years old, received CAl25-12, TPA-80,4, FE.
R-82,7, CEA-1,0, AFP-5,0,5L
The measurement result was X-23.9, and all were normal by independent judgment. However, as a result of the combination determination, it was determined that the patient had cancer. A re-examination revealed that she had ovarian cancer.

The test report displays the normal values of each tumor marker and also displays the measured values in a graph, with A and 82 reference values attached to the graph. Standard values A and B are set to different values depending on the age category. Figure 4 shows CA125~SLX
An example of a graph displaying the measurement results is shown below. In addition, the inspection report includes the judgment results 1 to 8 along with the comments mentioned above.

<Effects of the Invention> As described above, according to the cancer diagnostic device of the present invention, CA12
5. TPA, FER, CEA, AFP.

Six types of tumor markers of SLX are measured, and the measurement results of tumor markers are evaluated according to age category, high/low values are determined for cancer diagnosis, and the measurement results are processed according to a predetermined discrimination algorithm. , discriminant analysis that makes a judgment based on a combination of measurement results, and comprehensive judgment based on these judgment results, it is possible to make a more accurate cancer judgment than in the past.

[Brief explanation of drawings]

Figures 1 and 2 are flowcharts showing an algorithm for determining cancer based on each tumor marker and their combination. In particular, Figure 1 shows an overview of the overall process, and Figure 2 (A) High value group determination algorithm. FIG. 2(B) shows a low value group determination algorithm. FIG. 3 is a schematic configuration diagram of the cancer diagnostic device, and FIG. 4 is a diagram showing an example of a graph display used for a determination report. 1... Floppy 2... Reading device, 3... Judgment processing device

Claims (1)

[Claims] 1. CA125, TPA, FER, CEA, AFP, S
A recording means for recording the measurement results for the six types of tumor markers of LX, a reading means for reading each measurement result from the recording means, and a reading means for reading out the measurement data from the reading means according to the age category of the subject. A means for grouping the measurement results for each tumor marker by comparing them with reference values determined by the tumor marker, a high/low value group determination means for determining cancer based on the grouping of the measurement results, and a means for comparing the read data with the actual cancer A discriminant analysis means that processes according to a predetermined discrimination algorithm based on patient statistical data and makes a cancer determination based on a combination of tumor markers, a high/low value group determination means, and a comprehensive cancer diagnosis based on the results of the discriminant analysis means. 1. A cancer diagnostic device comprising: comprehensive determination means for making a determination.