[go: up one dir, main page]

WO2016198749A1 - Biomarqueurs de diagnostic, variables cliniques, et techniques permettant de les sélectionner et de les utiliser - Google Patents

Biomarqueurs de diagnostic, variables cliniques, et techniques permettant de les sélectionner et de les utiliser Download PDF

Info

Publication number
WO2016198749A1
WO2016198749A1 PCT/FI2016/050417 FI2016050417W WO2016198749A1 WO 2016198749 A1 WO2016198749 A1 WO 2016198749A1 FI 2016050417 W FI2016050417 W FI 2016050417W WO 2016198749 A1 WO2016198749 A1 WO 2016198749A1
Authority
WO
WIPO (PCT)
Prior art keywords
mdk
egf
mps
dps
gmcsf
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/FI2016/050417
Other languages
English (en)
Inventor
Matti Poutanen
Taija SALONIEMI
Kaisa Huhtinen
Antti PERHEENTUPA
Pia SUVITIE
Tero AITTOKALLIO
Harri Siitari
Daniel LAAJALA
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Turku
Original Assignee
University of Turku
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by University of Turku filed Critical University of Turku
Publication of WO2016198749A1 publication Critical patent/WO2016198749A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
    • G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
    • G01N33/50—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing
    • G01N33/68—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing involving proteins, peptides or amino acids
    • G01N33/6893—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing involving proteins, peptides or amino acids related to diseases not provided for elsewhere
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B20/00—ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
    • G16B20/20—Allele or variant detection, e.g. single nucleotide polymorphism [SNP] detection
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B20/00—ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
    • G16B40/00—ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
    • G16B40/20—Supervised data analysis
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2800/00—Detection or diagnosis of diseases
    • G01N2800/36—Gynecology or obstetrics
    • G01N2800/364—Endometriosis, i.e. non-malignant disorder in which functioning endometrial tissue is present outside the uterine cavity

Definitions

  • the present disclosure relates to the field of in vitro diagnostics. More specifically the present disclosure relates to non-invasive means and methods for determining a subject's diagnostic state.
  • novel combinations of biomarkers and/or clinical variables, and processes for selecting or identifying them for determining a subject's diagnostic state are disclosed.
  • the presently disclosed techniques for selecting and combining the biomarkers and clinical variables are applicable to determining a subject's diagnostic state of not only endometriosis but also other diseases particularly diseases associated with lower abdominal pain.
  • Endometriosis is a chronic estrogen-dependent gynecological disorder characterized by the presence and growth of endometrial tissue outside of the uterus. Globally, the disease affects up to 10% of women (176 to 181 million worldwide). Patients with endometriosis commonly complain of progressive and debilitating dysmenorrhea that begins already before menstruation. The pain may range from mild to severe, and is characteristically dull and aching. Dyspareunia, dysuria and painful defecation are also commonly associated with endometriosis. The experienced pain correlates poorly with the observed severity of the disease, although advanced endometriosis is strongly associated with infertility.
  • the disease primarily affects women of reproductive age, and the associated pain and infertility symptoms have a significant impact on physical and mental well-being. Consequently, the health burden and social cost associated with endometriosis are very high with a recent estimate of over 9000 € per woman per year in the EU due to loss of productivity and direct health care costs.
  • Serum CA125 is elevated in the serum of -60% of stage III and IV endometriosis patients. However, only 33% of stage l/l I patients have elevated concentrations. Therefore, the assay is not considered to be sufficiently sensitive as a screening test.
  • serum markers have also been studied, while none of the studies have been able to identify a set of serum markers clinically useful for identifying endometriosis. Thus, the diagnosis of endometriosis is currently mostly based on clinical signs and symptoms, physical examination and imaging techniques, such as ultrasound or MR imaging.
  • One aspect of the present disclosure is directed to a computer- implemented in vitro method of determining a subject's diagnostic state, as defined in claim 1.
  • determination of a subject's diagnostic state shall be interpreted in a slightly broader sense compared with unequivocal confirmation or exclusion of a disease.
  • determination of the subject's diagnostic state shall also comprise processes whose outcome is a computed risk for the subject to have or develop a disease, although a conclusive diagnosis is not made.
  • Other aspects of the present disclosure are directed to a processor specially adapted to carry out the method of claim 1 , and to a tangible non-transitory memory, which comprises program code instructions as defined in the independent claim directed to the processor.
  • the present disclosure is directed to a method of determining a subject's endometriosis state, said method comprising:
  • CA125 CA125
  • MDK Midkine
  • EMILIN1 Elastin microfibril interfacer 1
  • said the biomarkers may further comprise at least one biomarker selected from the group consisting of Granulocyte- macrophage colony-stimulating factor (GM-CSF), phosphatidylcholine acyl- alkyl C38:1 (PC ae C38:1 ), phosphatidylcholine acyl-alkyl C38:2 (PC ae C38:2), and any combination thereof.
  • GM-CSF Granulocyte- macrophage colony-stimulating factor
  • PC ae C38:1 phosphatidylcholine acyl- alkyl C38:1
  • PC ae C38:2 phosphatidylcholine acyl-alkyl C38:2
  • said biomarkers may additionally comprise at least one biomarker selected from the group consisting of epidermal growth factor (EGF), interleukin-1 receptor antagonist (IL-1 Ra), interleukin- 17 (IL-17), and any combination thereof.
  • EGF epidermal growth factor
  • IL-1 Ra interleukin-1 receptor antagonist
  • IL-17 interleukin- 17
  • said biomarkers may further comprise human epididymal secretory protein E4 (HE4).
  • HE4 human epididymal secretory protein E4
  • the method further comprises assessing at least one clinical variable selected from the group consisting of the subject's general information, surgical information, medical history, obstetric history, and pain symptoms.
  • the method comprises assessing at least one pain symptom selected from the group consisting of menstrual pain, intercourse pain, and defecation pain.
  • determination of a subject's endometriosis state includes diagnosing, testing for, prognosing or monitoring endometriosis state, risk of endometriosis, response to treatment of endometriosis, recurrence or relapse of endometriosis, or risk of recurrence or relapse of endometriosis.
  • the presence of endometriosis is determined on the basis of an increased level of said at least one biomarker as compared with a corresponding control level.
  • said determining the presence or absence of endometriosis in step c) includes determining or prognosing any possible re- lapse of the disease or response to treatment.
  • the method is used for excluding the presence or risk of a lower abdominal pain-associated disease other than endometriosis, such as disease selected from the group consisting of endometrial can-cer, ovarian cancer, colon cancer, ulcerative colitis, irritable bowel dis- ease, irritable bowel syndrome and Crohn's disease.
  • a lower abdominal pain-associated disease other than endometriosis such as disease selected from the group consisting of endometrial can-cer, ovarian cancer, colon cancer, ulcerative colitis, irritable bowel dis- ease, irritable bowel syndrome and Crohn's disease.
  • the present disclosure provides an in vitro screening kit comprising reagents for determining a biological sample for the level of biomarkers comprising CA125, MDK, and EMILIN1.
  • the kit may further comprise reagents for de- termining the level of at least one biomarker selected from the group consisting of GM-CSF, PC ae C38:1 , PC ae C38:2, and any combination thereof.
  • the kit may additionally comprise reagents for determining the level of at least one biomarker selected from the group consisting of EGF, IL-1 Ra, IL-17, and any combination thereof.
  • the kit may additionally comprise one or more reagents for determining the level of HE4.
  • the present disclosure is directed to use of a biomarker combination comprising CA125, MDK, and EMILIN1 for determining a subject's endometriosis state.
  • said use may further comprise use of at least one biomarker selected from the group consisting of GM-CSF, PC ae C38:1 , PC ae C38:2, and any combination thereof.
  • said use may further comprise use of at least one biomarker selected from the group consisting of EGF, IL-1 Ra, IL-17, and any combination thereof.
  • said use may additionally comprise use of HE4.
  • Figure 1 shows an overview of a feature selection and classification process in prediction of endometriosis
  • Figure 2 generally illustrates the concept of a decision-making process in which a subject's diagnosis of endometriosis is either confirmed or excluded;
  • FIG. 3 is a more detailed process flow chart
  • Figures 4A and 4B depict sample screens (pages) of a tool view, which is used to obtain clinical variables concerning patients;
  • Figure 4C depicts a sample screen relating to risk assessment
  • Figure 5 shows examples of decision trees with clinical variables and bi- omarkers; typically large numbers of such decision trees generated to fit the patient information provided;
  • Figure 6A depicts ranking of the importance of features (clinical variables and biomarkers) used in the model
  • Figure 6B depicts AUC (area under curve), sensitivity and specificity of the endometriosis diagnosis obtained by using the tool with the three most important clinical variables and three most informative biomarkers included;
  • Figure 7 depicts progress of AUC versus additional features, and maximum and minimum values obtained after 100 different repeats of the algorithm.
  • biomarker refers to a molecule whose measurement provides information regarding a disease state of a subject and is a distinctive indicator of a pathologic condition or its absence.
  • bi- omarker of endometriosis refers to a biomarker selected from the set of bi- omarkers provided by the present invention and which is indicative of endometriosis.
  • biomarker and marker are used herein interchangeably and they include, but are not limited to, protein markers, nucleic acid bi- omarkers (e.g. mRNA or cDNA) corresponding to or encoding said protein markers, and lipid markers.
  • assessing a biomarker level refers to quantifying a biomarker.
  • level is interchangeable with the terms “amount” and “concentration”, and can refer to an absolute or relative quantity. Assessing a biomarker may also include de- termining the absence or presence of the biomarker. Absolute or relative quantities of one or more biomarkers may also be referred to as “an expression profile” or “an expression pattern”.
  • the level of a biomarker in a biological sample may be determined by a variety of techniques as is readily apparent to a skilled person.
  • suitable methods include mass spectrometry, immunoassays, spectrophotometry, enzymatic assays, ultraviolet assays, kinetic assays, electrochemical assays, colorimetric assays, turbidimetric assays, atomic absorption assays, flow cytometry, liquid chromatography such as high performance/pressure liquid chromatography (HPLC), gas chromatography, nuclear magnetic resonance spectrometry, and any combinations thereof.
  • HPLC high performance/pressure liquid chromatography
  • sample refers to a biological sample, typically a clinical sample which may be, for example, a tissue sample, such as a biopsy or a laparoscopy sample, or a sample of a bodily fluid, such as urine, blood, plasma, or serum, obtained from a subject.
  • tissue sample such as a biopsy or a laparoscopy sample
  • a bodily fluid such as urine, blood, plasma, or serum
  • biological samples may be obtained from the subject at various time points before, during, or after treatment.
  • the expression levels of the present biomarker combinations in the biological sample are then determined and compared with those in a biological sample obtained from the same subject at a different time point, or with a control level obtained, for example, from a reference sample derived from an individual whose endometriosis state is known and/or who has not been exposed to said treatment.
  • predetermined reference values obtained from a pool of apparently healthy individuals may be used as control levels in said comparisons.
  • clinical variables obtained from a subject may be used with or without biomarker data to determine the subject's diagnostic state.
  • the clinical variables may be obtained at various time points be- fore, during, or after treatment and compared with those obtained from the same subject at a different time point.
  • the term "clinical variable” includes various objective or subjective variables concerning a subject's health or general condition.
  • the term includes, but is not limited to, general information such as the subject's age, height, weight, menopausal status, and features of the menstrual cycle.
  • the term also includes the subject's surgical history including status, reason, location, and number of previous abdominal surgery or surgeries, procedures used during previous abdominal surgery or surgeries, possible removal of the ovary and/or the Fallopian tube, and hysterectomy history.
  • the term also includes the subject's obstetric history including gravidity history and parity, ectopic pregnancies and treatments, history of miscarriage and medical abortions, history of infertility treatments, number of in vitro fertilizations (IVF, ICSI), history of insemination and ovulation induction, number of ovulation induction procedures, normality of partner's sperm, success of infertility treat- ments, and current desire for pregnancy (trying for more than a year without success). Furthermore, the term includes other history information such as the subject's medication history (especially concerning hormonal medication), diagnostic history, and history of disorders associated with lower abdominal pain in close family member.
  • the term also includes various pain-associated symp- toms, such as menstrual pain (duration, frequency, and severity of menstrual pain; type of medication required for alleviating menstrual pain); intercourse pain (description of pain, frequency and severity of pain, avoidance of intercourse because of pain), urinary pain i.e. pain associated with urinary bladder function (duration, frequency, and severity of pain, occasions of experienced pain symptoms, association with menstrual cycle, and previously diagnosed bladder related diseases); chronic abdominal pain, i.e.
  • the severity of pain symptoms may be evaluated on a scale such as the one from 0 to 10, wherein 0 denotes no pain and 10 denotes the strongest imaginable pain.
  • the clinical variables are obtained from the subject by a questionnaire comprising any number of relevant questions.
  • the clinical variables to be taken into account, and thus the questions to be comprised in the questionnaire may vary and depend on the disease in question.
  • the herein-disclosed techniques and methods for determining a subject's diagnostic state may be applied to any diseases, disorders, or conditions.
  • Some embodiments concern diseases associated with lower abdominal pain including, but not limited to urinary, disorders, such as bladder or kidney problems, bowel disorders, such as Crohn's disease, colon cancer, ulcerative colitis, irritable bowel disease (IBD) or irritable bowel syndrome (IBS), disorders of the reproductive system, such as endometriosis, endometrial cancer, ovarian cancer, or any other gynecological disease.
  • the term "subject” refers to an animal, preferably to a mammal, more preferably to a human, and in connection with gynecological diseases, most preferably to a woman.
  • said subject may suffer from endometriosis with or without diagnosis, be suspected to suffer from endometriosis, be at risk of endometriosis, or may have already been treated for endometriosis.
  • said subject may also suffer from, be suspected to suffer from, be at risk of, or may have already been treated for a disease other than endometriosis, especially another disease associated with lower abdominal pain.
  • the terms "human subject", "patient” and “individual” are inter-changeable.
  • the term "diagnostic state” refers any distinguishable manifestation of a disease, including non-disease.
  • the term includes, without limitation, information regarding the presence or absence of the disease, the risk of the disease, the stage of the disease, progression or remission, relapse or recurrence of the disease over time, and the severity of the disease.
  • endometriosis state refers any distinguishable manifestation of endometriosis, including non- disease.
  • the term includes, without limitation, information regarding the presence or absence of endometriosis, the risk of endometriosis, the stage of endometriosis, progression or remission, relapse or recurrence of en- dometriosis over time, and the severity of endometriosis.
  • corresponding terminology may be applied to any other disease, disorder, or condition as well, including but not limited to those associated with lower abdominal pain.
  • determining a subject's diagnostic state refers, without limitation, to diagnosing, testing for, prognosing or monitoring state of a disease, progression of said disease, risk of said disease, response to treatment, remission, relapse or recurrence of said disease, or a risk of remission, relapse of recurrence of said disease.
  • determining a subject's endometriosis state refers, without limitation, to diagnosing, testing for, prognosing or monitoring endometriosis state, progression of endometriosis, risk of endometriosis, response to treatment, remission, relapse or recurrence of endometriosis, or a risk of remission, relapse of recurrence of endometriosis.
  • corresponding terminology may be applied to any other disease, disorder, or condition including but not limited to those associated with lower abdominal pain.
  • the term "diagnosing” refers, without limitation, to a process aimed at determining whether or not a subject is afflicted with a disease such as endometriosis or any other disease associated with lower abdominal pain. This is also meant to include instances where the presence or a stage of said is not finally determined but that further diagnostic testing is warranted.
  • the method is not by itself determinative of the presence or absence of said disease, or the stage of said disease in the subject but can indicate that further diagnostic testing is needed or would be beneficial. Therefore, the present method may be combined with one or more other diagnostic methods for the final determination of the presence or absence of said disease , or the stage of said disease in the subject.
  • diagnostic methods are well known to a person skilled in the art, including but not limited to, laparoscopy.
  • the techniques disclosed herein may be used not only for diagnostic purposes but also for prognosis or predicting the outcome of or risk of a disease such as endometriosis or any other disease associated with lower abdominal pain, or monitoring the subject's disease state over time, any possible remission, recurrence or relapse of the disease, or response to treatment.
  • Monitoring of a subject's disease state can be performed by continuously measuring certain parameters and/or performing a medical test repeatedly.
  • a subject's disease state is monitored by obtaining bodily fluid samples repeatedly, assaying the samples using the method disclosed herein and comparing assay results with one another and with a reference value to identify any change in the subject's disease state.
  • the techniques disclosed herein may be used for differential diagnostics, i.e. for distinguishing of a particular disease, disorder, or condition from others that present similar clinical features.
  • the techniques are particularly suitable for differentiating various abdominal pain-associated diseases from each other.
  • the term "indicative of a disease”, when applied to a biomarker, refers to an expression pattern or profile which, using routine statistical methods setting confidence levels at a minimum of 95%, is diagnostic of said disease or a stage of said disease such that the expression pattern is found significantly more often in subjects with said disease or a stage of said disease than in subjects without said disease or another stage of said disease.
  • an expression pattern which is indicative of a disease is found in at least 80% of subjects who have the disease and is found in less than 10% of subjects who do not have the disease.
  • an expression pattern which is indicative of said disease is found in at least 90%, at least 95%, at least 98%, or more in subjects who have the disease and is found in less than 10%, less than 8%, less than 5%, less than 2.5%, or less than 1 % of subjects who do not have the disease.
  • said disease is associated with lower abdominal pain.
  • said disease is endometriosis.
  • expression profile and “expression pattern” refer not only to an expression level of a pro- tein or nucleic acid biomarker or but also to a level of a lipid biomarker, although it is understood that lipids are not expressed through translation but synthetized in a series of defined steps in the cytoplasm.
  • the expression profile of the biomarker in a relevant control has to be determined. Once the control levels are known, the determined marker levels can be compared therewith and the significance of the difference can be assessed using standard statistical methods. In some embodiments of the present disclosure, a statistically significant difference between the deter- mined biomarker level and the control level is indicative of a disease such as endometriosis or any other disease associated with lower abdominal pain. In some further embodiments, before to be compared with the control, the biomarker levels are normalized using standard methods.
  • control may refer to a control sample ob- tained from an apparently healthy individual or a pool of apparently healthy individuals, or it may refer to a predetermined threshold value which is indicative of the presence or absence of a disease in question.
  • control or threshold values may have been de- termined, if necessary, from samples of subjects of the same age, demographic features, and/or disease status, etc.
  • the threshold value may originate from a single individual not affected by a disease in question or be a value pooled from more than one such individual.
  • the term "apparently healthy” refers to an individual or a pool of individuals who show no signs of a disease in question and thus are believed not to be affected by said disease in question and/or who are predicted not to develop said disease in question.
  • the term "increased expression” refers to an increase in the amount of a biomarker in a sample as compared with a corre- sponding control sample. Said increase can be determined qualitatively and/or quantitatively according to standard methods known in the art. The expression is increased if the amount or level of the biomarker in the sample is, for instance, at least about 1.5 times, 1.75 times, 2 times, 3 times, 4 times, 5 times, 6 times, 8 times, 9 times, time times, 10 times, 20 times or 30 times the amount of the same biomarker in the control sample. In some embodiments, the term “increased expression” refers to a statistically significant increase in the level or amount of the biomarker as compared with that of a relevant control.
  • the term “decreased expression” refers to a decrease in the amount of a biomarker in a sample as compared with a corre- sponding control sample. Said decrease can be determined qualitatively and/or quantitatively according to standard methods known in the art. The expression is decreased if the amount of the biomarker in the sample is, for instance, at least about 1.5 times, 1.75 times, 2 times, 3 times, 4 times, 5 times, 6 times, 8 times, 9 times, time times, 10 times, 20 times or 30 times lower than the amount of the same biomarker in the control sample. In some embodiments, the term “decreased expression” refers to a statistically significant decrease in the level or amount of the biomarker as compared with that of a relevant control.
  • sensitivity is a measure of the ability of a marker to detect the disease. In other words, sensitivity represents the probability of a positive test result in subjects with the disease.
  • specificity is a measure of the ability of a marker to detect the absence of the disease. In other words, specificity represents the probability of a negative test result in a subject without the disease.
  • FP false positive
  • FN false negative
  • TP true positive
  • TN true negative
  • uccess rate refers to the percentage-expressed proportion of affected individuals with a positive result
  • FDR false detection rate
  • TP+TN correctly classified subjects
  • TP+TN+FP+FN TP+TN+FP+FN
  • Receiver Operation Characteristic (ROC) curves may be utilized to demonstrate the trade-off between the sensitivity and specificity of a marker, as is well known to skilled persons.
  • the horizontal X-axis of the ROC curve represents 1 -specificity, which increases with the rate of false positives.
  • the vertical Y-axis of the curve represents sensitivity, which increases with the rate of true positives.
  • cut-off i.e. threshold
  • the values of specificity and sensitivity may be determined.
  • data points on the ROC curves represent the proportion of true-positive and false- positive classifications at various decision boundaries. Optimum results are obtained as the true-positive proportion approaches 1.0 and the false-positive proportion approaches 0.0.
  • sensitivity usually is reduced and vice versa.
  • the area under the ROC curve is a measure of the utility of a marker in the correct identification of disease subjects, i.e. subjects who are affected by a disease.
  • the AUC values can be used to determine the effectiveness of the test.
  • An area of 1.0 represents a perfect test; an area of 0.5 represents a worthless test.
  • a traditional rough guide for classifying the accuracy of a diagnostic test is the following: AUC val- ues 0.9 to 1.0 represent a test with excellent diagnostic power, AUC values 0.80 to 0.90 represent a test with good diagnostic power, AUC values 0.70 to 0.80 represent a test with fair diagnostic power, AUC values 0.60 to 0.70 represent a test with poor diagnostic power, and AUC values 0.50 to 0.60 represent a test with failed diagnostic power.
  • AUC val- ues 0.9 to 1.0 represent a test with excellent diagnostic power
  • AUC values 0.80 to 0.90 represent a test with good diagnostic power
  • AUC values 0.70 to 0.80 represent a test with fair diagnostic power
  • AUC values 0.60 to 0.70 represent a test with poor diagnostic power
  • AUC values 0.50 to 0.60 represent a test with failed diagnostic power.
  • Figure 1 shows an overview of a feature selection and classification process in prediction of a subject's classification into one of diagnostic states, such as patient or control group.
  • the following description and the associate drawings relate to an environment used for developing an algorithm for diagnosing endometriosis.
  • the algorithmic classification approach is applicable to other diseases or conditions for which the classification is based on combinations of clinical variables and biomarkers.
  • wrapper feature subset selection methodology was used to select the best discriminating features (clinical variables and biomarkers) for the prediction problem, namely disease versus control classification.
  • "best" discriminating features means those that maximize the ratio of correct classifications to all classifications.
  • sensitivity vs. (1 -specificity) is plotted as a normalized curve, ie, a curve whose x- and y- axes are normalized to the range of ⁇ 0, 1 ⁇ , and the area under the curve (AuC) is used as the metrics.
  • Sensitivity can be expressed as the ratio true positives /(true positives + false negatives) and specificity as the ratio true nega- tives/(true negative + false positives) of prediction.
  • a heuristic function guides the search to find feature subsets with the highest-scoring evaluator functions.
  • “wrapper" selection methods the actual accuracy is not known, but an accuracy estimate is often used as both the heuristic function (which guides the search) and the evaluator function (which evaluates the goodness of a candidate subset).
  • the inventors used repeated cross-validation as the evaluator function.
  • wrapper feature subset selection methodology is one of several filtering approaches described by Kohavi et al.
  • the inventors found wrapper feature subset selection attractive because it forces the accuracy estimation to execute cross-validation more times on small datasets than on large datasets.
  • the overall accuracy estimation time which is the sum of the induction algorithm running time and the cross-validation time, does not grow too fast. This means that small datasets will be cross-validated many times to overcome the high variance resulting from small amounts of data.
  • the inventors experimented with the wrapper subset selection strategy in a repeat- ed cross-validation setting and assembled the results across different repeats of the cross-validation.
  • mtry parameter which is the number of variables randomly sampled as candidates at each split
  • some others were default parameters of the Random forest package (e.g. ntree which determines the number of trees to grow).
  • Repeated cross-validation was used to search for the best parameters of random forest and to quantify the predictive power of the machine learning model. The final results reported are averaged over all the repeats of the cross-validation. In final classification, 100 repeats of cross- validation were executed. Each repeat was randomly seeded at each iteration in both feature selection and final classification for variation in different random parameters (e.g. division of data into test and training set in cross-validation). The reported results are averages of different repeats.
  • biomarker discovery can be formulated as:
  • X ⁇ X(i ), X(2), X(D) ⁇ be a set of predictors (features, i.e., biological measurements) and T be the target variable (i.e., the disease status we want to predict).
  • decision trees are excellent modelling tools which have found usage in many application domains.
  • a major drawback of decision trees is that they are prone to overfitting to their training sets.
  • Bagging with a random feature subset is thus a technique which address the tendency of decision-tree-based decision algorithms to overfit to their training data sets. It uses a collection of decision trees, each of which is trained on different random permutations of the data as well as the predictive features. Finally, the results from these different decision trees are combined (typically, averaged) to get the final unbiased prediction, which ameliorates the disadvantages of a single decision tree and provides better and more robust generalization capabilities.
  • Bagging bootsstrap aggregation of a decision tree with a random subset of features involves an ensemble of multiple decision trees that are useful for classification, regression, and also missing value imputation.
  • bagging is used to arrive at a classification setting which models a binary response (Patient or Control) as a function of predictor variables (biomarkers with or without clinical variables concerning patients).
  • decision tree algorithms create a binary tree by repeatedly splitting the features in the dataset into two groups. At each branching point (a node in the tree), any feature can be divided into two groups (children) which results in the best separation between the categorical response. The splitting is performed until a criterion for stopping the split- ting is completed. Typically the splitting is stopped when a minimum number of nodes nmin is reached of the tree needs no further splitting because its leaves contain observations of a single class
  • Gini impurity index (G) of the parents and its prospective child is often used as the splitting criterion. If the current node consists of n observations in P classes, let n p be the number of observations in the p th class. We can now mathematically formulate the Gini impurity index as:
  • the Gini index G increases when observations from different classes are accumulated in the same node. Similar to the Gini index, other potential candidates for split- ting criterion are: information gain, variance reduction, and other statistical tests.
  • the process of constructing of the tree is begun by computing the
  • Gini index in the parent node is assigned to different unique binary partitions, first splitting between the first and second lowest values and continuing with the second and third lowest values and so on.
  • samples with a value lower than or equal to the threshold of split is assigned to the one of the two children of the node (e.g. left child of the binary tree) and the samples with values larger than the threshold of split is assigned to the other child of the node (e.g. right child of the binary tree).
  • the two-way arrows in Figure 1 indicate that feature selection and/or classification may be performed iteratively, to reduce the influence of random artefacts on final results.
  • Figure 2 generally illustrates the concept of an in vitro decisionmaking process in which a subject's risk of having or developing endometriosis is determined.
  • features from one or two classes are obtained from the subject, namely biomarkers and/or clinical variables.
  • Bi- omarkers which are measurable indicators of some biological state or condi- tion, can be obtained via well-known methods, which are beyond the scope of the present disclosure. It should be noted that the decision-making process described herein does not use physical samples obtained from subjects but data elements describing the physical samples.
  • Clinical variables are information elements which objectively and/or subjectively describe the subject.
  • objective descriptors include the subject's age, weight, menopausal status, or the like.
  • subjective descriptors include the subject's own evaluation of various pains.
  • the features obtained from the subject are inputted to a computer-implemented model, which is based on the above-described machine-learning process. In that pro- cess, candidate sets of decisions trees are filtered to yield an optimized set of decisions trees, which is used to process the features obtained from the subject.
  • FIG. 2 The result of the in vitro decision-making process shown in Figure 2 may be a qualitative diagnosis (eg endometriosis or not) or a quantitative risk assessment (eg x% likelihood of having endometriosis, with y% confidence and z% accuracy).
  • Figure 3 is a more detailed flow chart of an overall process, which comprises model creation (shown as phase A) and diagnostic decision-making (shown as phase B).
  • the model-creation phase comprises step 3-2, in which clinical variables describing large numbers of samples are obtained.
  • large numbers of samples means sample sizes large enough to draw scientifically solid conclusions, such as correlations between feature sets and diagnostic status.
  • the clinical variables are obtained via computerized user interfaces, which present one or more questionnaires. Each questionnaire comprises multiple questions, whose answers will form the set of clinical variables concerning one sample.
  • the set of clinical variables is optimized in the Feature selection phase described in connection with Figure 1.
  • biomarkers describing the same samples are also available, which is why the test 3-4 is answered positively and the flow proceeds to step 3-6, in which the computer-implemented model (see Figure 2) is trained on both clinical variables and biomarkers obtained from the same samples.
  • This model-training step 3-6 includes the feature-selection phase described in connection with Figure 1.
  • phase B the model trained in step 3-6 will be used in an in vitro diagnostic phase (or risk-assessment phase) concerning an individual subject (prospective patient).
  • Phase B begins in step 3-12, in which biomarkers and clinical variables of the subject are obtained. This step is analogous with steps 3-2 and 3-6 described above, apart from the fact that the steps of phase A were performed with respect to a large number of samples.
  • step 3-14 the model trained in step 3-6 (namely the filtered sets of decision trees) is used to assess the subject's diagnostic status.
  • step 3-16 the result of step 3-12 are outputted from the computerized model. For instance, the results can be shown to a physician and/or the subject, but what the physician does with the results is beyond the scope of the present disclosure.
  • Figures 4A and 4B depict sample screens (pages) of a tool view, which is used to obtain clinical variables concerning patients.
  • the screen shown in Figure 4A is used to obtain objective information, but similar screens can be used to obtain subjective information, such as the subject's own assessment of various pains.
  • the screen shown in Figure 4B may be used to combine biomarker data with clinical variables obtained by screens analogous with the one shown in Figure 4A.
  • Figure 4C depicts a sample screen relating to risk assessment.
  • the exemplary screen indicates a 94.06% risk for a disease (eg endometriosis), and its inverse value, namely 5.94% likelihood of not having that disease.
  • a risk for having that disease in a future period of time may be calculated and shown.
  • the two rightmost columns indicate an estimated confidence (97.30%) and estimated accuracy (87,78%).
  • Figure 5 shows examples of decision trees with clinical variables and biomarkers; typically large numbers of such decision trees generated to fit the patient information provided.
  • Reference number 5-10 is one exemplary decision tree.
  • the decision tree 5-10 has a root node 5-11 , which is associated with a feature (clinical variable or biomarker) 5-12, herein menstrual pain strength, which may be entered via a user interface analogous with the one shown in Figure 4A.
  • the feature 5-12 has a threshold value 5-13 (here: a value of 4 for subjective assessment of menstrual pain strength).
  • Still further parameters may be associated with the feature 5-12, such as a confidence value 5-14.
  • Use of the decision tree 5-10 begins at the root node 5-11.
  • Its associated feature and threshold are used to retrieve the corresponding feature obtained from the subject, which is compared with the threshold. If, say, the feature ob- tained from the subject (menstrual pain) is less than or equivalent to the threshold, traversal of the decision tree 5-10 proceeds to intermediate node 5-15, which is associated with feature (biomarker) MDK and threshold 0.316. If the corresponding feature (level of MDK) obtained from the subject is less than or equivalent to the threshold 0.316, the threshold, traversal of the decision tree proceeds to 5-21 , which is one of the leaf nodes 5-21 ... 5-25 of the decision tree 5-10. Leaf node 5-21 indicates an (approximately) 43% likelihood for classification as patient. Other decision trees, of which two exemplary ones are depicted by reference numbers 5-30 and 5-40 will be used to improve the confidence and accuracy of classification (prediction).
  • the nodes of decision trees which are not leaf nodes, have associated features and threshold values, which the features are to be compared with.
  • threshold values Several efficient features and feature sets will be given later in this disclosure, particularly for endometriosis.
  • the same threshold values that were used to create the decision trees in the machine-learning phase can also be used in the diagnostic or risk assessment phase.
  • Figure 6A depicts ranking of the importance of features (clinical variables and biomarkers) used in the model
  • Figure 6B depicts AUC (area under curve), sensitivity and specificity of the endometriosis diagnosis obtained by using the tool with the three most important clinical variables and three most informative biomarkers included.
  • Figure 7 depicts progress of AUC versus additional features, and maximum and minimum values obtained after 100 different repeats of the algorithm. More information will be given in the Tables section at the end of this description.
  • any available clinical variables may be sub- jected to the algorithm disclosed herein for determining a subject's diagnostic state.
  • the number of the clinical variables to be utilized may vary from an embodiment to embodiment, and may include at least two, at least three, at least four, at least five, at least six, at least seven, at least eight, at least nine, at least ten, etc. clinical variables selected independently from each other among various variable categories such as the subject's general information, surgical information, medical history, obstetric history, and various pain symptoms including but not limited to, menstrual pain, intercourse pain, and pain associated with defecation or urination.
  • menstrual pain, intercourse pain, and defecation pain are the most relevant clinical variables for determining a subject's endometriosis state.
  • CA125 human cancer antigen 125
  • Serum concentrations of human cancer antigen 125 are known to be elevated in patients with ovarian cancer, advanced endometriosis with peritoneal or deep lesions, or ovarian endometriomas.
  • lack of sensitivity and specificity of CA125 has significantly hampered its use as a diagnostic tool.
  • the classical cut-off value of CA125 (35 U/ml) used for diagnostics of ovarian cancer is far too high for diagnosing endometriosis.
  • CA125 may be used as a biomarker for determining a subject's state of a disease associated with lower abdominal pain, especially endometri- osis or ovarian cancer.
  • performance of the present method was improved even further when the level of human epididymal secretory protein E4 (HE4, also known as WAP four-disulfide core domain protein 2) was assessed together with the levels of CA125 and MDK.
  • E4 human epididymal secretory protein
  • Such a combined analysis provided an AUC value of 0.90 and, at its highest, an accuracy value of 0.886.
  • HE4 is a known biomarker whose expression is markedly increased in ovarian cancer, especially the stage I disease, and in endometrial cancer.
  • another advantage associated with the use of HE4 in combination with CA125 and MDK is the differentiation of endometriosis from endometrial cancer and ovarian cancer.
  • performance of the present method was improved further also when the level of any one of Elastin microfibril interfacer 1 (EMILIN1 ), Granulocyte-macrophage colony-stimulating factor (GM-CSF), and phosphatidylcholine acyl-alkyl C38:1 (lipid metabolite PC ae C38:1 ) or phosphatidylcholine acyl-alkyl C38:2 (lipid metabolite PC ae C38:2) were assessed together with the levels of CA125 and MDK. To be more specific, combined assessment of CA125, MDK, and EMILIN1 levels provided an accuracy value of 0.862 at its highest.
  • Elastin microfibril interfacer 1 EMILIN1
  • GM-CSF Granulocyte-macrophage colony-stimulating factor
  • phosphatidylcholine acyl-alkyl C38:1 lipid metabolite PC ae C38:1
  • the present method may, in some embodiments, further comprise determination of the level of HE4, particularly for differentiating endometriosis from endometrial cancer and ovarian cancer.
  • the present method may further comprise determination of any one or any combination of epidermal growth factor (EGF), interleukin-1 receptor antagonist (IL-1 Ra), and interleukin-17 (IL-17).
  • EGF epidermal growth factor
  • IL-1 Ra interleukin-1 receptor antagonist
  • IL-17 interleukin-17
  • the endometriosis-associated biomarkers disclosed herein namely CA125, MDK, EMILIN1 , GM-CSF, PC ae C38:1 , PC ae C38:2, EGF, IL-1 Ra, IL-17, and HE4 were identified among over 50 serum markers including for instance various cytokines, steroids, peptide hormones, and serum metabolites.
  • these biomarkers are used in the present technique or method depending on availability of reagents for assessing their expression levels.
  • PC.ae.C38.1 , PC.ae.C38.2, IL-17, and EMILIN1 were decreased or down-regulated in patients with endometriosis while CA125, IL1 RA, HE4, EGF, GMCSF, and MDK were increased or upregu- lated in patients with endometriosis.
  • the present disclosure provides a method of determining a subject's endometriosis state on the basis of assessing the level of CA125, MDK, and EMILIN1 in a sample obtained from said subject.
  • the level of at least one further marker selected from GM-CSF, PC ae C38:1 , PC ae C38:2, EGF, IL-1 Ra, IL-17, and HE4 may be used in the analysis.
  • measurements of the present biomarkers may be used alone or combined with other data obtained regarding the subject whose endometriosis state is to be determined.
  • the method per se does not involve analyzing any other data, such as clinical variables.
  • the present disclosure also provides use of the disclosed biomarker combinations in determining a subject's endometriosis state.
  • the present disclosure provides a method of determining a subject's diagnostic state by obtaining at least three features from said subject, wherein at least one of the features is a clinical variable, and at least two of the features are biomarker levels, such as levels of biomarkers selected from CA125, MDK, EMILIN1 , GM-CSF, PC ae C38:1 , PC ae C38:2, EGF, IL-1 Ra, IL-17, and HE4, or vice versa.
  • biomarker levels such as levels of biomarkers selected from CA125, MDK, EMILIN1 , GM-CSF, PC ae C38:1 , PC ae C38:2, EGF, IL-1 Ra, IL-17, and HE4, or vice versa.
  • said clinical variables are selected from menstrual pain, intercourse pain, and defecation pain, the presence of which together with the increased level of any one of CA125, IL1 RA, HE4, EGF, GM-CSF, and MDK and/or decreased level of any one of PC.ae.C38.1 , PC.ae.C38.2, IL-17 and EMILIN1 are indicative of endometriosis.
  • in- creased level of at least CA125 and/or HE4 together with the presence of relevant clinical variables may be indicative of ovarian cancer.
  • normal level of any one of CA125, MDK, EMILIN1 , GM-CSF, PC ae C38:1 , PC ae C38:2, EGF, IL-1 Ra, IL-17, and HE4 together with the presence of relevant clinical variables may indicate that the subject has or is at risk of developing a lower abdominal pain-associated disease other than endometriosis or ovarian cancer.
  • the method comprises assessing any one, two or three clinical variables are selected from menstrual pain, intercourse pain, and defecation pain, and assessing the expression level of any one, two, or three biomarkers selected from CA125, MDK, and EMILIN1.
  • the combined presence of the assessed one, two, or three pain symptoms, and increased expression of CA125 or MDK, and decreased expression of EMILIN1 are indicative of the presence of or risk of developing endometriosis.
  • the term "presence of a pain symptom" may refer to a pain score X, on an intuitive scale, such as from 0 to 10.
  • the pain score of at least X is considered indicative of endometriosis.
  • the threshold value X may be determined as the value that, based on available data, maximizes the AUC value of the feature or feature combination which depends on the pain score.
  • the threshold value for the pain score may be set to an explicit value, such as at least one, at least two, at least three, at least four, at least five, at least six, at least seven, at least eight, at least nine or at least 10. Different diseases and/or different pains may have different threshold values.
  • kits for use in determining a subject's endometriosis state may comprise any reagents or test agents necessary for assessing the level of biomarker combinations disclosed herein.
  • a person skilled in the art can easily determine the reagents to be in- eluded depending on the biomarker combination in question and a desired technique for carrying out said assessment.
  • an appro- priate control sample or a threshold value may be comprised in the kit.
  • the kit may also comprise a computer readable medium, comprising computer- executable instructions for performing any of the methods of the present disclosure.
  • the serum samples from all participants were collected within 24 hours prior to the operation into non-heparinized tubes and centrifuged for 15 min at 800 g after being kept at room temperature for 30 minutes. The serum was stored at -20°C until analyzed.
  • Serum CA125 concentrations (U/ml) were evaluated using ELISA analysis (Fujirebio Diagnostics Inc, Malvern, PA, USA) according to the manufacturer's instructions.
  • concentration of Midkine MDK
  • MDK Midkine
  • EMILIN Elastin microfibril interfacer 1
  • GMCSF granulocyte-macrophage colony-stimulating factor
  • Tables 1 to 5 results obtained from different assay combinations are summarized in Tables 1 to 5 below.
  • Table 6 shows biomarker levels before and after treatment. Further assay combinations and their respective AUC values will be giv- en in Table 7.
  • the patient took the endometriosis symptom-based questionnaire to assess the risk of endometriosis. Subjecting the clinical variables to the algorithm disclosed herein gave an assessment of 82.4% chance of the disease with a 94% confidence, which result equated to a high predicted risk.
  • the patient has no history of sexually transmitted infections. She has no urinary urgency, dysuria, nocturia, dyschezia or pain with defecation. Her physical examination was normal.
  • Laparotomy performed confirms the presence of endometriosis lesion on the peritoneum and Uterosacral nodularity.
  • Case 3 is presented as an example wherein the subject's computed risk of having endometriosis was 49%. Because the computed risk alone was very close to 50%, it was insufficient for a definite diagnosis but was sufficient for triggering additional procedures, which then confirmed the presence of endometriosis.
  • case 4 is presented as an example wherein the computed risk of having endometriosis (49.2%) was very close to 50%. This finding triggered additional procedures, which in this case excluded endometriosis.
  • EMILIN 94.07 175 dps EMILIN, MDK 88.459 mps, dps, MDK 92.347 176 dps, EMILIN, ⁇ 87.571 mps, dps, 91.371 177 dps, EMILIN, EGF 85.263 mps, dps, EGF 89.631 178 dps, EMIUN.
  • GMCSF GMCSF, HE4 93.217 1681 ips, mps, CA125, MDK, *1, IL1RA 98.224

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Medical Informatics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Biotechnology (AREA)
  • Theoretical Computer Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • Evolutionary Biology (AREA)
  • Biophysics (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Proteomics, Peptides & Aminoacids (AREA)
  • Public Health (AREA)
  • Genetics & Genomics (AREA)
  • Biomedical Technology (AREA)
  • Epidemiology (AREA)
  • Databases & Information Systems (AREA)
  • Bioethics (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Immunology (AREA)
  • Hematology (AREA)
  • Pathology (AREA)
  • Urology & Nephrology (AREA)
  • Primary Health Care (AREA)
  • Cell Biology (AREA)
  • Microbiology (AREA)
  • Food Science & Technology (AREA)
  • Medicinal Chemistry (AREA)
  • Biochemistry (AREA)
  • General Physics & Mathematics (AREA)
  • Investigating Or Analysing Biological Materials (AREA)

Abstract

La présente invention concerne le domaine des diagnostics in vitro. Plus spécifiquement, la présente invention concerne des moyens non invasifs et des procédés pour déterminer un état de diagnostic d'un sujet. L'invention concerne en particulier de nouvelles combinaisons de biomarqueurs et/ou de variables cliniques, et des processus pour les sélectionner ou les identifier afin de déterminer un état de diagnostic du sujet. Les techniques de la présente invention pour la sélection et la combinaison des biomarqueurs et variables cliniques sont applicables pour déterminer un état de diagnostic du sujet non seulement de l'endométriose mais également d'autres maladies, en particulier de maladies associées à la douleur abdominale inférieure.
PCT/FI2016/050417 2015-06-12 2016-06-10 Biomarqueurs de diagnostic, variables cliniques, et techniques permettant de les sélectionner et de les utiliser Ceased WO2016198749A1 (fr)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
FI20155455 2015-06-12
FI20155455 2015-06-12
FI20155812 2015-11-05
FI20155812 2015-11-05

Publications (1)

Publication Number Publication Date
WO2016198749A1 true WO2016198749A1 (fr) 2016-12-15

Family

ID=56292766

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/FI2016/050417 Ceased WO2016198749A1 (fr) 2015-06-12 2016-06-10 Biomarqueurs de diagnostic, variables cliniques, et techniques permettant de les sélectionner et de les utiliser

Country Status (1)

Country Link
WO (1) WO2016198749A1 (fr)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107491656A (zh) * 2017-09-04 2017-12-19 北京航空航天大学 一种基于相对危险度决策树模型的妊娠结局影响因子评估方法
CN110782996A (zh) * 2019-09-18 2020-02-11 平安科技（深圳）有限公司 医疗数据库的构建方法、装置、计算机设备和存储介质
WO2021114625A1 (fr) * 2020-05-28 2021-06-17 平安科技（深圳）有限公司 Procédé et appareil de construction de structure de réseau destinés à être utilisés dans un scénario multitâche
WO2021168040A1 (fr) * 2020-02-19 2021-08-26 Aspira Women's Health Inc. Compositions pour l'évaluation de l'endométriose présentant une spécificité améliorée
CN113782124A (zh) * 2021-09-15 2021-12-10 宁夏医科大学总医院 晚期卵巢癌满意肿瘤细胞减灭术前评估与预测模型的建立与改良

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3185828A (en) 1962-01-05 1965-05-25 Multiblitz Gerate Dr Ing D A M Photographic flashlight illumination
WO2006113747A2 (fr) * 2005-04-19 2006-10-26 Prediction Sciences Llc Marqueurs diagnostiques du traitement et de l'evolution du cancer du sein et leurs methodes d'utilisation
US20110236903A1 (en) * 2008-12-04 2011-09-29 Mcclelland Michael Materials and methods for determining diagnosis and prognosis of prostate cancer
US20140220580A1 (en) * 2011-06-16 2014-08-07 Kirk Brown Biomarker compositions and methods

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3185828A (en) 1962-01-05 1965-05-25 Multiblitz Gerate Dr Ing D A M Photographic flashlight illumination
WO2006113747A2 (fr) * 2005-04-19 2006-10-26 Prediction Sciences Llc Marqueurs diagnostiques du traitement et de l'evolution du cancer du sein et leurs methodes d'utilisation
US20110236903A1 (en) * 2008-12-04 2011-09-29 Mcclelland Michael Materials and methods for determining diagnosis and prognosis of prostate cancer
US20140220580A1 (en) * 2011-06-16 2014-08-07 Kirk Brown Biomarker compositions and methods

Non-Patent Citations (16)

* Cited by examiner, † Cited by third party
Title
"Endometriosis", 25 January 2012, WILEY, article BEATE E SEEBER ET AL: "Diagnosis of Endometriosis: Biomarkers", pages: 309 - 323, XP055316622, DOI: 10.1002/9781444398519.ch30 *
ANNI P. VEHMAS ET AL: "Ovarian Endometriosis Signatures Established through Discovery and Directed Mass Spectrometry Analysis", JOURNAL OF PROTEOME RESEARCH., vol. 13, no. 11, 7 November 2014 (2014-11-07), US, pages 4983 - 4994, XP055314882, ISSN: 1535-3893, DOI: 10.1021/pr500384n *
BREIMAN L.: "Random forests", MACHINE LEARNING, vol. 45, no. 1, 2001, pages 5 - 32, XP019213368, DOI: doi:10.1023/A:1010933404324
BREIMAN, MACHINE LEARNING, vol. 45.1, 2001, pages 5 - 32
EVERITT, B.: "The Cambridge Dictionary of Statistics", 1998, CAMBRIDGE UNIVERSIT PRESS
GOLLAND P; LIANG F; MUKHERJEE S; PANCHENKO D.: "Permutation tests for classification", ANNUAL CONFERENCE ON LEARNING THEORY, 2005, pages 501 - 515, XP019011426
HYE WON CHUNG ET AL.: "Pleiotrophin (PTN) and midkine(MK) mRNA expression in eutopic and ectopic endometrium in advanced stage endometriosis", MOLECULAR HUMAN REPRODUCTION, OXFORD UNIVERSITY PRESS, GB - BE, vol. 8, no. 4, 1 January 2002 (2002-01-01), pages 350 - 355, XP002991729, ISSN: 1360-9947, DOI: 10.1093/MOLEHR/8.4.350 *
JOHNSON WE; LI C; RABINOVIC A.: "Adjusting batch effects in microarray expression data using empirical Bayes methods", BIOSTATISTICS, vol. 8, 2007, pages 118 - 127, XP055067729, DOI: doi:10.1093/biostatistics/kxj037
K HUHTINEN ET AL: "Serum HE4 concentration differentiates malignant ovarian tumours from ovarian endometriotic cysts", BRITISH JOURNAL OF CANCER, vol. 100, no. 8, 31 March 2009 (2009-03-31), GB, pages 1315 - 1319, XP055316614, ISSN: 0007-0920, DOI: 10.1038/sj.bjc.6605011 *
KOHAVI R; JOHN GH.: "Wrappers for feature subset selection", ARTIFICIAL INTELLIGENCE, vol. 97, no. 1-2, 1997, pages 273 - 324, XP002263359, DOI: doi:10.1016/S0004-3702(97)00043-X
LIAW A., DOCUMENTATION FOR R PACKAGE RANDOMFOREST (PDF, 16 October 2012 (2012-10-16), Retrieved from the Internet <URL:http://cran.r-project. Org/web/packages/randomForest/randomForest.pdf>
LIAW, DOCUMENTATION FOR R PACKAGE RANDOMFOREST, 16 October 2012 (2012-10-16), Retrieved from the Internet <URL:http://cran.r-project.org/web/packages/randomForest/randomForest.pdf>
NEZIRI ALBAN Y ET AL: "Correlation Between Altered Central Pain Processing and Concentration of Peritoneal Fluid Inflammatory Cytokines in Endometriosis Patients With Chronic Pelvic Pain", REGIONAL ANESTHESIA AND PAIN MEDICINE, vol. 39, no. 3, May 2014 (2014-05-01), pages 181 - 184, XP002763808 *
PIETRO SANTULLI ET AL: "Increased Serum Cancer Antigen-125 Is a Marker for Severity of Deep Endometriosis", JOURNAL OF MINIMALLY INVASIVE GYNECOLOGY, vol. 22, no. 2, 19 October 2014 (2014-10-19), NL, pages 275 - 284, XP055314871, ISSN: 1553-4650, DOI: 10.1016/j.jmig.2014.10.013 *
VARMA; SIMON, BMC BIOINFOR-MATICS, vol. 7.1, 2006, pages 91
ZHANG LI-HUI ET AL: "Differential proteomic analysis of endometriosis", CHEMICAL RESEARCH IN CHINESE UNIVERSITIES, BEIJING, CN, vol. 26, no. 1, 1 January 2010 (2010-01-01), pages 70 - 74, XP009192336, ISSN: 1005-9040 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107491656A (zh) * 2017-09-04 2017-12-19 北京航空航天大学 一种基于相对危险度决策树模型的妊娠结局影响因子评估方法
CN107491656B (zh) * 2017-09-04 2020-01-14 北京航空航天大学 一种基于相对危险度决策树模型的妊娠结局影响因子评估方法
CN110782996A (zh) * 2019-09-18 2020-02-11 平安科技（深圳）有限公司 医疗数据库的构建方法、装置、计算机设备和存储介质
WO2021168040A1 (fr) * 2020-02-19 2021-08-26 Aspira Women's Health Inc. Compositions pour l'évaluation de l'endométriose présentant une spécificité améliorée
EP4107525A4 (fr) * 2020-02-19 2024-08-14 Aspira Women's Health Inc. Compositions pour l'évaluation de l'endométriose présentant une spécificité améliorée
WO2021114625A1 (fr) * 2020-05-28 2021-06-17 平安科技（深圳）有限公司 Procédé et appareil de construction de structure de réseau destinés à être utilisés dans un scénario multitâche
CN113782124A (zh) * 2021-09-15 2021-12-10 宁夏医科大学总医院 晚期卵巢癌满意肿瘤细胞减灭术前评估与预测模型的建立与改良

Similar Documents

Publication Publication Date Title
Vodolazkaia et al. Evaluation of a panel of 28 biomarkers for the non-invasive diagnosis of endometriosis
Lopes et al. Thyroid cancer and thyroid autoimmune disease: A review of molecular aspects and clinical outcomes
WO2016198749A1 (fr) Biomarqueurs de diagnostic, variables cliniques, et techniques permettant de les sélectionner et de les utiliser
US20190004051A1 (en) Plasma autoantibody biomarkers for diagnosis of lung cancer
Hezelgrave et al. Quantitative fetal fibronectin at 18 weeks of gestation to predict preterm birth in asymptomatic high-risk women
CA2741117A1 (fr) Biomarqueurs d&#39;insuffisance cardiaque
Perelló et al. Markers of deep infiltrating endometriosis in patients with ovarian endometrioma: a predictive model
Malczewska et al. NETest is superior to chromogranin A in neuroendocrine neoplasia: a prospective ENETS CoE analysis
Grenache et al. Clinical performance of two multi-marker blood tests for predicting malignancy in women with an adnexal mass
Knific et al. Multiplex analysis of 40 cytokines do not allow separation between endometriosis patients and controls
US20230263477A1 (en) Universal pan cancer classifier models, machine learning systems and methods of use
Park et al. Utility of the immature granulocyte percentage for diagnosing acute appendicitis among clinically suspected appendicitis in adult
WO2019238944A1 (fr) Panel de biomarqueurs pour le cancer de l&#39;ovaire
Trivedi et al. Risk assessment for indeterminate pulmonary nodules using a novel, plasma-protein based biomarker assay
Florholmen et al. Discovery and validation of mucosal TNF expression combined with histological score-a biomarker for personalized treatment in ulcerative colitis
Jeong et al. Urinary DNA methylation test for bladder cancer diagnosis
Carr et al. Discriminating pheochromocytomas from other adrenal lesions: the dilemma of elevated catecholamines
CA3155044A1 (fr) Systemes et procedes pour detecter une pathologie
Rewcastle et al. Assessing the prognostic value of PAX2 and PTEN in endometrial carcinogenesis
Zeng et al. Development of a scoring system for predicting the severity of ulcerative colitis
Vallvé-Juanico et al. External validation of putative biomarkers in eutopic endometrium of women with endometriosis using NanoString technology
EP4357782A1 (fr) Panel de biomarqueurs protéiques pour le diagnostic du cancer colorectal
JP2024500575A (ja) がんの早期かつ正確な検出のための、血液ベースのタンパク質バイオマーカーパネル
JP2023531567A (ja) 進行性前立腺がんを判定するためのバイオマーカー組合せ
US20250044302A1 (en) Jup biomarker for the diagnosis of diseases or disorders of the female reproductive tract

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16733623

Country of ref document: EP

Kind code of ref document: A1

DPE1 Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101)
NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16733623

Country of ref document: EP

Kind code of ref document: A1