Patent Application
20160003830
1. Field of the Invention
The present invention relates to a technique for clinical diagnosis and screening of colon cancer. More specifically, the present invention relates to a multiplex colon cancer marker panel. Even more specifically, the present invention relates to a combination of colon cancer markers having a nigh ability to detect colon cancer.
2. Disclosure of the Related Art
As typical colon cancer markers, Carcinoembryonic antigen-related cell adhesion molecule 5 (CEA) and Carbohydrate Antigen 19-9 (CA19-9) are known. These colon cancer markers are actually used in clinical practice, but it has been demonstrated that they are not suitable for early diagnosis.
JP2008-14937 A (Patent Document 1) and Oncology Reports 2011, Vol. 25, pp 1217-26 (Non-Patent Document 1) report that novel Colon cancer-associated proteins have been identified by proteomic analysis of colon cancer tissues. Further, The Non-Patent Document 1 reports that Galectin-1, Galectin-3, and Galectin-4 are effective as plasma markers for colon cancer.
Each of the colon cancer markers that have been previously reported cannot achieve a satisfactory detection rate of cancer patients (more specifically, sensitivity) when used as a single marker.
On the other hand, there is a case where markers are simply combined for the purpose of improving the reliability of diagnosis. It is true that the combined use of markers improves the detection rate of cancer patients, but specificity (i.e., the percentage of healthy individuals correctly diagnosed as healthy) is reduced. Therefore, it is necessary to minimize a reduction in specificity.
It is therefore an object of the present invention to provide a specific combination of colon cancer markers based on statistical knowledge, which is capable of detecting a larger number of colon cancer patients in an earlier stage while maintaining high specificity.
The present inventors have found that a specific combination of markers can achieve a hiqh detection rate of cancer patients while maintaining high specificity. Such findings have been found for the first time by the present invention.
The present invention includes the following inventions.
Carcinoembryonic antigen-related cell adhesion molecule 5, Carbohydrate antigen 19-9, and APEX nuclease;
Carcinoembryonic antigen-related cell adhesion molecule 5, Carbohydrate antigen 19-9, and Actin-related protein 2;
Carcinoembryonic antigen-related cell adhesion molecule 5, Galectin-4, and APEX nuclease;
Carbohydrate antigen 19-9, Galectin-4, and APEX nuclease;
Carbohydrate antigen 19-9, Galectin-4, and Actin-related protein 2; or
Carbohydrate antigen 19-9, APEX nuclease, and Actin-related protein 2.
acquiring respective measured values of five colon cancer markers of Carcinoembryonic antigen-related cell adhesion molecule 5, Carbohydrate antigen 19-9, Galectin-4, APEX nuclease, and Actin-related protein 2 in a biological sample derived from an individual;
normalizing the respective measured values of the five colon cancer markers to derive respective probability scores of the five colon cancer markers and deriving an average of the probability scores; and
evaluating the average of the probability scores based on whether the average is higher or lower than a criterion value for the five colon cancer markers.
acquiring respective measured values of four colon cancer markers arbitrarily selected from five colon cancer markers of Carcinoembryonic antigen-related cell adhesion molecule 5, Carbohydrate antigen 19-9, Galectin-4, APEX nuclease, and Actin-related protein 2 in a biological sample derived from an individual;
normalizing the respective measured values of the four colon cancer markers to derive respective probability scores of the four colon cancer markers and deriving an average of the probability scores; and
evaluating the average of the probability scores based on whether the average is higher or lower than a criterion value for the four colon cancer markers.
acquiring respective measured values of three colon cancer markers of:
Carcinoembryonic antigen-related cell adhesion molecule 5, Carbohydrate antigen 19-9, and APEX nuclease;
Carcinoembryonic antigen-related cell adhesion molecule 5, Carbohydrate antigen 19-9, and Actin-related protein 2;
Carcinoembryonic antigen-related cell adhesion molecule 5, Galectin-4, and APEX nuclease;
Carbohydrate antigen 19-9, Galectin-4, and APEX nuclease;
Carbohydrate antigen 19-9, Galectin-4, and Actin-related protein 2; or
Carbohydrate antigen 19-9, APEX nuclease, and Actin-related protein 2 in a biological sample derived from an individual;
normalizing the respective measured values of the three colon cancer markers to derive respective probability scores of the three colon cancer markers and deriving an average of the probability scores; and
evaluating the average of the probability scores based on whether the average is higher or lower than a criterion value for the three colon cancer markers.
acquiring respective measured values of two colon cancer markers of Carbohydrate antigen 19-9 and APEX nuclease in a biological sample derived from an individual;
normalizing the respective measured values of the two colon cancer markers to derive respective probability scores of the two colon cancer markers and deriving an average of the probability scores; and
evaluating the average of the probability scores based on whether the average is higher or lower than a criterion value for the two colon cancer markers.
According to the present invention, it is possible to provide a specific combination of colon cancer markers based on statistical knowledge, which is capable of detecting a larger number of colon cancer patients in an earlier stage while maintaining high specificity. More specifically, colon cancer patients can be detected at high sensitivity by determining the quantities of expressed specific two or more colon cancer markers in blood samples (plasma samples) from individual subjects. By detecting cancer patients at high sensitivity, it is possible to make an early diagnosis and to select an appropriate cancer treatment, which as a result contributes to an improvement in QOL of patients. Further, the colon cancer marker panel according to the present invention is expected to be applied to colon cancer diagnostic reagents and colon cancer diagnostic equipment.
a) shows a histogram representing the distribution of marker levels of Galectin-4 of healthy individuals (Control) and a histogram representing the distribution of marker levels of Galectin-4 of colon cancer patients (CRC), wherein the histogram of healthy individuals is given by gray bars and the histogram of colon cancer patients is given by black bars, and the horizontal axis represents the marker level and the vertical axis represents the number of samples;
b) shows a curve of probability score obtained by converting the marker levels based on the cumulative distribution function of an extreme-value distribution shown in
The present invention provides a colon cancer marker panel comprising specific two or more colon cancer markers. The colon cancer marker panel according to the present invention is constituted from two to five colon cancer markers, and the expression of each of the markers is increased by colon cancer.
The colon cancer marker panel constituted from five colon cancer markers includes Carcinoembryonic antigen-related cell adhesion molecule 5 (hereinafter, referred to as “CEA”), Carbohydrate antigen 19-9 (hereinafter, referred to as “CA19-9”), Galectin-4, APEX nuclease (DNA-(apurinic or apyrimidinic site) lyase) (hereinafter, referred to as “APEX1”), and Actin-related protein 2 (hereinafter, referred to as “ACTR2”).
The colon cancer marker panel constituted from four colon cancer markers includes four colon cancer markers arbitrarily selected from the above-mentioned five colon cancer markers. Specific combinations of the four colon cancer markers are CEA, CA19-9, Galectin-4, and APEX1; CEA, CA19-9, Galectin-4, and ACTR2; CEA, CA19-9, APEX1, and ACTR2; CEA, Galectin-4, APEX1, and ACTR2; and CA19-9, Galectin-4, APEX1, and ACTR2.
The colon cancer marker panel constituted from three colon cancer markers includes specific three colon cancer markers selected from the above-mentioned five colon cancer markers. Specific combinations of the three colon cancer markers are CEA, CA19-9, and Galectin-4; CEA, CA19-9, and APEX1; CEA, CA19-9, and ACTR2; CEA, Galectin-4 and APEX1; CA19-9, Galectin-4, and APEX1; CA19-9, Galectin-4, and ACTR2; and CA19-9, APEX1, and ACTR2.
The colon cancer marker panel constituted from two colon cancer markers includes specific two colon cancer markers selected from the above-mentioned five colon cancer markers. Specific combinations of the two colon cancer markers are CA19-9 and Galectin-4; and CA19-9 and APEX1.
An object to be analyzed by a method according to the present invention is a biological sample derived from an individual (human individual). The biological sample to be analyzed is preferably a blood sample. However, this is not intended to exclude a body fluid sample or a tissue sample other than a blood sample.
Examples of the blood sample include whole blood, blood plasma, and blood serum, and the like. The blood sample can be prepared by appropriately treating whole blood collected from an individual. When collected whole blood is treated to prepare a blood sample, treatment performed on the whole blood is not particularly limited as long as it is clinically acceptable. For example, anticoagulation treatment and centrifugal separation may be performed. The blood sample directly subjected to measurement of marker levels may be one that has been appropriately stored at low temperatures, for example, in a frozen state, in the course of or after its preparation. It is to be noted that the blood sample used in the present invention is discarded without being returned to an individual as its source.
The method according to the present invention absolutely includes the step of measuring each of colon cancer markers constituting the above-described colon cancer marker panel.
A high detection rate can be expected by performing the step of measuring two or more other colon cancer markers in addition to the step of measuring each of colon cancer markers constituting the above-described colon cancer marker panel. On the other hand, even when a high detection rate is achieved by increasing the number of markers used in combination, it is clinically meaningless if specificity is low. Therefore, the optimum number of markers can be determined using, as an index, a median AUC (Area Under the Curve) obtained by repeating analysis 100 times (which will be described later) per number of markers.
The analysis of cancer marker levels according to the present invention is performed by acquiring respective measured values of colon cancer markers constituting the colon cancer marker panel, and using the measured values which are sigmoidally normalized based on an extreme-value distribution.
The parameters of the extreme-value distribution are determined using only the marker levels of samples of healthy individuals. The measured value of each of the markers is converted to a “probability score” by the cumulative distribution function of the extreme-value distribution. The probability score (hereinafter, sometimes simply referred to as a “score”) refers to the probability that a patient has colon cancer at a certain marker level, and is a normalized value between 0 and 1.
An average value is derived from the respective scores derived from the respective measured marker values. Based on the score average determined in such a manner as described above, a diagnosis of colon cancer is made. When the score average of a sample is larger than a criterion value, the sample is regarded as positive, and when the score average of a sample is smaller than the criterion value, the sample is regarded as negative. When a sample is regarded as positive based on its score average, a human individual as a source of the sample can be diagnosed as having colon cancer.
A specific example of the criterion value to be compared with the score average value is a threshold value of average of colon cancer marker scores. The threshold value used in the present invention can be previously set depending on race, age, etc. The threshold value can be set by reference to averages of scores of the colon cancer markers converted by the above-described normalization of the measured quantity values of the respective colon cancer markers present in samples derived from individuals belonging to a healthy individual group and from individuals belonging to a colon cancer patient group.
As the threshold value, a cutoff value yielding high diagnostic accuracy is selected. The threshold value can be appropriately selected by those skilled in the art from cutoff values preferably yielding a specificity of 80% or higher, e.g., 95%. The upper limit of the specificity is not particularly limited, but may be, for example, 98%.
A method for setting the threshold value is appropriately selected by those skilled in the art. One example of the method is ROC Curve (Receiver Operating Characteristic Curve) analysis.
In the method according to the present invention, the colon cancer markers are preferably measured by an assay based on biospecific affinity. The assay based on biospecific affinity is well-known to those skilled in the art and is not particularly limited. However, an immunoassay is preferred. Specific examples of the immunoassay include competitive and non-competitive immunoassays such as western blotting, radioimmunoassay, Enzyme-Linked ImmunoSorbent Assay (ELISA; including all sandwich, competitive, and direct immunoassays), immunoprecipitation, precipitation reaction, immunodiffusion, immunoagglutination, complement-binding reaction, immunoradiometric assay, fluoroimmunoassay, and protein A immunoassay. In the immunoassay, antibodies that bind to the colon cancer markers in a blood sample are detected. At this time, a colon cancer detection chip may be used, in which antibodies that bind to all the proteins constituting the colon cancer marker panel are immobilized onto the surface of one substrate.
The colon cancer markers are measured by bringing a sample into contact with antibodies under conditions where colon cancer marker proteins to be measured can form immunocomplexes with antibodies against the colon cancer marker proteins.
A specific protocol of the immunoassay can be easily selected by those skilled in the art.
Alternatively, the colon cancer markers may be measured based on mass spectrometry. A method of mass spectrometry is not particularly limited as long as it can perform quantitative analysis, and can be appropriately selected by those skilled in the art.
Hereinbelow, the present invention will be described more specifically with reference to the following examples, but is not limited to the examples.
In the following Reference Example 2 and Example 1, plasma samples were prepared in the following manner. About 15 mL of blood was collected from each individual into a BD Vacutainer CPTTM tube. After the collection of blood, the collected blood was immediately centrifuged (1,700×g, 4° C., 20 min) to obtain a supernatant as a plasma component (about 5 mL). The obtained plasma sample was stored at −80° C.
The plasma sample was thawed before measurement and diluted 5,000 to 20,000-fold to obtain a blood sample used to measure the concentrations of the colon cancer markers according to the present invention.
Out of the proteins identified by proteomic analysis using cancer tissues in JP2008-14937A, 40 proteins whose ELISA measurement systems have been established were selected as candidates for colon cancer markers. These 40 proteins are shown in Tables 1 to 8. In Tables 1 to 8, the protein names and gene names of the 40 proteins, existing ELISA kits, standard proteins, capture antibodies, detection antibodies, conjugated enzymes, secondary antibodies, and substrates are shown. It is to be noted that the protein names and the gene names in Tables are names registered in the UniProt database, “CA19-9” in Table 1 refers to carbohydrate antigen 19-9, recombinant proteins marked with 3) in Tables 2 and 3 were synthesized using “Transdirect insect cell” (manufactured by Shimadzu Corporation), and antibodies marked with 4) in Tables 2, 3 and 4 were prepared by immunizing synthetic peptides.
Plasma samples of patients whose informed consent had been obtained in accordance with the ethical guidelines of Faculty of Medicine of Osaka University were analyzed in the following manner. In the following analysis, “sensitivity” refers to the percentage of colon cancer patients who are correctly diagnosed as having colon cancer, and “specificity” refers to the percentage of healthy individuals as healthy who are correctly diagnosed, and “false-positive rate” refers to the percentage of healthy individuals who are diagnosed as having colon cancer.
The plasma samples were prepared according to the method described in Reference Example 1 from blood collected from 105 colon cancer patients and 100 healthy individuals. The concentrations of the 40 proteins in each of the plasma samples of the colon cancer patients and the healthy individuals were measured using the ELISA measurement systems shown in Tables 1 to 8.
Out of the 40 markers, 13 proteins showed statistically-significant differences (p<0.05) between the colon cancer patients and the healthy individuals. The analysis results of the 13 proteins showing statistically-significant differences are more specifically shown in Table 9. In Table 9, the area under ROC curve (AUC) and the P value; the cutoff values expressed as concentration and probability score; and the sensitivities for different stages (Stage 0, Stage I, Stage II, Stage III, Stage IV) and all stages (All Stages) of colon cancer determined using the cutoff value of each of the markers in healthy individuals vs colon cancer patients (namely, Control vs CRC) are shown. It is to be noted that the cutoff values and the sensitivities are values when the specificity is 95% (i.e., when an allowable false-positive rate is 5%). Further, the significant difference in concentration was based on verification using Mann-Whitney test. The stages of colon cancer are based on TMN classification, and primary cancer is represented as Stage 0 (in-situ cancer), Stage I, and Stage II, and lymph node metastatic cancer is represented as Stage III and Stage IV (the same shall apply hereinafter).
As shown in Table 9, when these 13 proteins were used alone as markers, sensitivity for all stages was about 40% at a maximum. That is, it can be said that these proteins are poor in sensitivity when used as single markers.
The effectiveness of a combined use of colon cancer markers for increasing the detection rate of colon cancer patients was verified.
The levels of almost all the 13 markers selected in Reference Example 2 in the plasma samples of the healthy individuals were relatively low. On the other hand, the plasma samples of the colon cancer patients had relatively high marker levels, and some of them had very high marker levels (i.e., outliers). A histogram representing the concentration (marker level) of Galectin-4 in the plasma samples of the healthy individuals and a histogram representing the concentration of Galectin-4 in the plasma samples of the colon cancer patients are shown in
In view of the above, the marker levels were sigmoidally normalized based on an extreme-value distribution. The parameters of the extreme-value distribution were determined using only the marker levels of the samples of the healthy individuals, and the marker level of each of the samples was converted to a “probability score” (hereinafter, sometimes simply referred to as a “score”) by the cumulative distribution function of the extreme-value distribution. The thus obtained curve of probability score is shown in
After each of the marker levels was normalized in the same manner as described above, the average of normalized scores was used as an index to discriminate between colon cancer patients and healthy individuals. Combinations of markers effective in discriminating between colon cancer patients and healthy individuals were determined in the following manner using a Monte-Carlo method.
For example, in the case of analysis of a combination of two markers, the step of estimating parameters for normalization using 50 samples selected from the 100 samples of the healthy individuals and performing discrimination between colon cancer patients and healthy individuals using the remaining 50 samples of the healthy individuals and 53 samples selected from the 105 samples of the colon cancer patients was repeated 100 times (100-times analysis). In this way, the averages of scores at the time when two markers randomly selected from the 40 markers shown in Tables 1 to 8 were used in combination were calculated.
Further, in the case of analysis of a combination of three markers, the step of estimating parameters for normalization using 50 samples selected from the 100 samples of the healthy individuals and performing discrimination between colon cancer patients and healthy individuals using the remaining 50 samples of the healthy individuals and 53 samples selected from the 105 samples of the colon cancer patients was repeated 100 times (100-times analysis). In this way, the averages of scores at the time when three markers randomly selected from the 40 markers shown in Tables 1 to 8 were used in combination were calculated.
The above-described 100-times analysis was performed in the same manner as described above on combinations of 4, 5, 6, . . . and 40 markers, and the averages of scores were calculated.
The obtained results were reevaluated based on a receiver operating characteristic (ROC) curve. The relationship between the number of markers used in combination (Number of markers) and the average of the areas under the ROC curve (AUC) is shown by a box plot in
Combinations of two to five of the above-described top five markers frequently selected in Reference Example 3 (CA19-9, Galectin-4, APEX1, CEA, and ACTR2) were used in cancer marker panels to determine the ability of each of the cancer marker panels to detect colon cancer. The results are shown in Table 10. In Table 10, the area of under the ROC curve (AUC); the cutoff value represented as probability score; and the sensitivities for different stages (Stage 0, Stage I, Stage II, Stage III, and Stage IV) and all stages (All Stages) of colon cancer determined using the cutoff value of each of the combinations in healthy individuals vs colon cancer patients (Control vs CRC) are shown. It is to be noted that the cutoff value and the sensitivities are values when the specificity is 95% (i.e., when an allowable false-positive rate is 5%). Further, the cancer marker panels according to the present invention are marked with an asterisk.
In Example 1, samples whose average of probability scores of the markers exceeded the cutoff value shown in Table 10 were regarded as positive to discriminate between healthy individuals and colon cancer patients.
The AUC value was highest when the 5 markers (CA19-9, CEA, Galectin-4, APEX1, and ACTR2) were used in combination, and it has been confirmed that the combination of the 5 markers has the highest ability to discriminate between cancer patients and healthy individuals. Further, it has been also confirmed that all the other combinations of markers according to the present invention marked with an asterisk have a higher AUC value than the combination of conventional markers CEA and CA19-9.
Further, when sensitivity was compared, sensitivity was most improved particularly when the above-mentioned 5 markers were used in combination.
Further, a comparison was made using the positive plasma samples between when the each of the above-mentioned 5 markers was used alone and when the above-mentioned 5 markers were used in combination. The results are shown in
As can be seen from
| Number | Date | Country | Kind |
|---|---|---|---|
| 2011-280149 | Dec 2011 | JP | national |
This application is a Divisional Application of patent application Ser. No. 13/723,133 filed on Dec. 20, 2012, which is based on Japanese Patent Application No. JP2011-280149 filed on Dec. 21, 2011, the entire contents of which are hereby incorporated by reference.
| Number | Date | Country | |
|---|---|---|---|
| Parent | 13723133 | Dec 2012 | US |
| Child | 14852308 | US |
