Patent Application
20090143993
The present invention relates to the field of diagnosis in hepatology, and in particular relates to a method for the evaluation of the presence and/or severity of hepatic fibrosis of the liver, or the evaluation of the area of fibrosis, or the evaluation of the architecture of the liver (fibrosis score and fractal dimension).
For the purpose of the present invention, the term “evaluation of the presence of fibrosis” means that the question of whether or not a fibrosis exists in the patient tested by means of the method of the invention is investigated; the term “evaluation of the severity” means that a measurement of the degree of fibrosis is sought, this must be distinguished from the severity of the hepatic damage, which is a functional deficiency of the liver. The term “evaluation of the area of fibrosis” means that a measurement of the degree of liver lesion due to the fibrosis is sought. It is specified that the functional deficiency of the liver depends on the degree of anatomical lesion of the liver, but this is not a linear relationship.
The seriousness of chronic liver diseases lies in the fibrosis that is a scar secondary to the inflammation. The causes of fibrosing liver diseases are mainly B and C viral infections, alcohol and steatosis (fatty liver).
Up until now, the evaluation of the fibrosis was based on the liver needle biopsy (LNB). Liver fibrosis is classified, according to the LNB, by means of a semiquantitative fibrosis score. Several classifications exist, based on the observation of similar lesions. The description of these lesions is mainly qualitative according to a disturbance (or distortion) of the architecture of the elementary unit (at the functional and anatomical level) of the liver, namely the hepatic “lobe”. The fibrosis begins at the periphery of the lobe in the “portal” space (F1 stage) so as to extend within the lobe (restricted bands of fibrosis or F2 stage) and then dissect it (extensive bands of fibrosis or F3 stage) so as to be concentric and isolate the hepatic cells (F4 stage or cirrhosis). The Métavir classification described above (Bedossa et al, 1994, Hepatology, vol. 20, pages 15-20) is one of the most commonly used. It classifies liver fibrosis into five stages from F0 to F4, the F4 stage corresponding to the ultimate stage of cirrhosis. The fibrosis is said to be clinically significant when it is at stage F≧2. The fibrosis score F is used by all liver specialists throughout the world (according to different classifications). It is the most important parameter for determining the seriousness of a liver disease, its evolutive potential and the indication for treatment. It is of determining assistance in being able to prescribe a treatment or in managing a disease. This F-score classification is semiquantitative for three reasons: a) the description of the lesions is purely qualitative and therefore evaluated by a physician who is an anatomical pathologist, b) the scoring can only be given as a finite and restricted number of stages (from 4 to 6 without counting the absence of fibrosis), c) the progression of the amount of fibrosis is not linear as a function of the stages. The quantitative aspect is due to the ordered nature of the classes according to the extension of the fibrosis within the lobe.
A purely quantitative means for measuring fibrosis exists: it is the measurement of the area (or surface) of fibrosis by means of a semiautomatic technique called image analysis. The area of fibrosis, which is compared to a panel of blood markers for fibrosis, considered as a reference, has been found to be a more reliable measurement than the Métavir score (Pilette et al, 1998, J Hepatol, vol. 28, pages 439-46).
However, LNB is an expensive and invasive examination which is therefore susceptible to complications and requires at least a day's hospitalization. The current constraints of LNB (cost, invasive procedure requiring hospitalization) limit the use thereof. Resorting to this diagnostic method remains the almost exclusive use of liver specialists. As a result, current medical management of treatment concerns patients that are often at a relatively advanced stage of the disease (cirrhosis, often complicated), for which there are fewer treatment possibilities.
Several investigations clearly demonstrate that LNB is the main limiting factor of screening and of access to treatments. The development of alternatives to LNB, which is the aim of the present invention, is part of the research recommendations of the American and French consensus conferences in 2002.
Liver fibrosis, including up to the recent cirrhosis stage, is a reversible condition. Early screening for fibrosis often makes it possible to propose steps for curing the disease or at least for limiting the consequences thereof.
The alternatives to LNB are non-invasive means, at the head of which are blood markers for fibrosis. The term “blood markers for fibrosis” in fact has two meanings. For the biologist, it involves markers that reflect one of the dynamic processes of fibrosis: fibrogenesis (production of fibrosis), fibrolysis (destruction of fibrosis). For the clinician, it involves a marker for the degree of fibrosis observed upon anatomical-pathological examination (mainly “septal” fibrosis), i.e. a static image resulting from the two dynamic processes above. In addition, the clinician differentiates these indicators into direct markers when they are derived from one of the molecules involved in the extracellular matrix (fibrosis) and into indirect markers as reflections, but not an integral part, of this visible fibrosis.
The international patent application published under the number WO 02/16949 describes a method of diagnosing inflammatory, fibrotic or cancerous diseases, in which the values of biochemical markers in the serum or the plasma of a patient are measured, said values are combined by virtue of a logistical function, and the final value of said logistical function is analyzed with a view to determining the presence of fibrosis or the presence of necrotic-inflammatory lesions in the liver. This international patent application makes it possible to propose a fibrosis test. However, the markers used are conventional biochemical markers (indirect markers) which are not specific indicators of fibrosis and can vary according to other disturbances present during liver diseases. The test marketed, corresponding to the method of patent WO 02/16949 (see also Imbert-Bismut et al, Lancet 2001, Vol. 37, pages 1069-1075), called the Fibrotest sold by the company Biopredictive, has in particular the drawback that it has difficulties in correctly classifying patients having stage F0 and F4 viral hepatitis forms.
In addition, the international patent application published under the number WO 03/073822 concerns a method for diagnosing the presence or the severity of a liver fibrosis in a patient. This method is based on the detection of three markers, namely α-2-macroglobulin, hyaluronic acid and metalloproteinase-1 tissue inhibitor.
The object of the present invention is to propose novel tools for determining the F stages of fibrosis, in particular having a score of F≧2, and for finely quantifying the exact degree of this fibrosis, with a view to diagnosing the presence and/or severity of a liver pathology and/or for monitoring the effectiveness of a curative treatment.
The monitoring of the effectiveness of a curative treatment or a treatment that suspends the disease is important. Since most chronic liver diseases are accompanied by a fibrosis, curative treatment or treatment that suspends the disease has the effect of slowing down the progression or even of causing the fibrosis to regress. It is therefore important to be able to have tests that can evaluate this variation in fibrosis.
Contrary to the tools and methods of the prior art, the present invention relates not only to fibroses for which the cause is viral, but also to fibroses for which the cause is alcoholic and to steatosis.
Furthermore, the tools of the present invention are more reliable than those of the prior art.
These tools are: (1) a diagnostic score for the presence and severity of fibrosis, also called diagnostic score of portal and septal fibrosis, (2) a noninvasive means of quantifying the area of fibrosis, and (3) a noninvasive means of determining the fractal dimension indicating the degree of distortion of the liver due to fibrosis.
The invention therefore makes it possible to determine a noninvasive diagnostic score for portal and septal fibrosis (that reflected by the Métavir score) that is clinically significant. The score according to the invention ranges from 0 (minimal fibrosis) to 1 (maximum fibrosis) with the reference threshold fixed at 0.5 for Métavir scores F≧2. This score is calculated using a subjective semiquantitative fibrosis reference: the Métavir score. The Métavir score is determined by a physician who is an anatomical pathologist, after examination of a liver fragment under the microscope. The scale of this noninvasive score is therefore virtual since it is distorted relative to the real measurement (although itself also arbitrary and subjective) of fibrosis represented by a Métavir score of 0 to 4. The scale is virtual since it is generated by a mathematical formula and there is no unit of measurement, and this scale is distorted since there is no direct (or linear) proportionality between the Métavir and noninvasive scores. However, this score of 0 to 1 represents a finer measurement of portal and septal fibrosis since it is a quantitative variable that allows finer comparisons. Two examples of a result: an individual may evolve from a score of 0.14 to 0.28 although he or she is still at the Métavir stage F0-F1 and yet has doubled his or her fibrosis score (100% progression in relative value). Conversely, when an individual evolves from a score of 0.48 to 0.52, it could be wrongly deduced that said individual has gone from a stage F0-F1 to a stage F2-F3 (or appearance of a “clinically significant” fibrosis) whereas, in reality, the progression is only 8% (in relative value—0.48 compared to 0.52 or [(0.52−0.48)/0.52]=0.08 or 8%—or 4%—0.52−0.48=0.04—in absolute value and not clinically significant.
Furthermore, the present invention makes it possible not only to determine a diagnostic score, but also to quantify the area of fibrosis of the liver. The measurement of the area of fibrosis makes it possible to obtain results that are more accurate for calculating the percentage of the liver taken up by fibrosis than the Métavir F score for fibrosis currently used. Such a quantification was not possible, up until now, in any of the methods described. It is an index (or estimate score) of the area of fibrosis ranging from 2% to 55%, respectively minimum and maximum area of fibrosis in the reference patient population. This index is calculated with a quantitative fibrosis reference. The scale of this index is therefore real since it is the direct (non distorted) reflection of an objective and non-arbitrary real measurement. It is therefore a measurement that is both precise and meaningful since it estimates without distortion a real magnitude. Two examples of results: an individual may evolve from an estimated area of fibrosis of 8.2% to 16.4%. An individual with cirrhosis may regress from 35% to 31% then 27% and, finally, 23% of estimated area of fibrosis after the cause has been interrupted or with anti-fibrosing treatment, whereas, despite a regular decrease, said patient is still at the cirrhosis stage (F4).
Furthermore, the present invention makes it possible not only to determine a diagnostic score and to quantify the area of fibrosis of the liver, but also to determine the architecture of the liver (fractal dimension). The measurement of the architecture of the liver makes it possible to obtain results that are more accurate for evaluating the degree of liver distortion due to fibrosis than the Métavir F score for fibrosis currently used. This degree of liver distortion due to fibrosis is the fractal dimension obtained by image analysis that is based on several estimating factors including the Kolmogorov dimension (Moal F et al, 2002, Hepatology, vol. 36, pages 840-9). None of the methods of the prior art makes it possible to establish a noninvasive measurement of the fractal dimension by assaying blood markers.
In fact, the inventors have developed the following scores given in table 1 below:
The diagnostic effectiveness is the percentage of individuals correctly classified compared with the LNB. The diagnostic effectiveness of the diagnostic score of the present invention increases at the extremities of the score. The SNIFF diagnostic score does not incorrectly classify any patient with viral hepatitis for F0 and F4 (and very few for F3). In other words, this SNIFF score is very effective (100% correct responses) for two essential questions posed by the clinician: is there a risk of incorrectly classifying an individual without fibrosis or an individual with cirrhosis? The diagnostic effectiveness of an SNIFF score with five variables is 90.8% for 50.0% of the patients with the lowest and the highest values. Given the errors of LNB, especially at the low (observer error) and high (sample error) stages of fibrosis, the error rate is therefore close to 0%.
The aim of the invention is therefore in particular to determine, with greater accuracy than that allowed by the tools of the prior art, whether a patient with or without known liver disease is suffering from fibrosis, and the severity of the liver damage (degree of lesion). The test according to the invention has the advantage of being able to be carried out every 6 to 12 months, whereas the LNB can only be repeated, optionally, every 3 to 5 years according to the consensus conferences.
The method according to the invention consists in combining and in measuring various direct markers for fibrosis associated with indirect markers taken in a specific combination, said markers being called variables. These variables are measured in a sample from an individual. The choice of these variables is determined by the best overall effectiveness of the combination of variables that is obtained by statistical analysis of various mathematical models, each providing a piece of information that is statistically significant and independent of the others. In other words, it involves the best effectiveness for the least number of variables. This means that any new variable in the mathematical model provides an inventive piece of information (or gain in diagnostic effectiveness) compared to a more restricted combination that might have already been the subject of a publication.
In the context of the present invention, the term “sample” is intended to mean a sample taken from an individual prior to any analysis. This sample may be a biological medium such as blood, serum, plasma, urine or saliva from said individual or one or more cells from said individual, such as a tissue biopsy, and more particularly a liver biopsy.
The term “liver pathology” is intended to mean a liver pathology chosen from chronic hepatic fibrosis of viral origin, chronic hepatic fibrosis of alcoholic origin and chronic hepatic steatosis.
In the context of the present invention, the term “individual” is intended to mean a man, a woman or an animal, young or adult, healthy or liable to be suffering from or suffering from a liver pathology such as chronic hepatic fibrosis of viral origin, chronic hepatic fibrosis of alcoholic origin or chronic hepatic steatosis, or from any other pathology, it being possible for the affected individual to be receiving or not receiving a curative treatment against this liver pathology.
The present invention therefore relates to a method of diagnosing the presence and/or severity of a liver pathology and/or of monitoring the effectiveness of a curative treatment against a liver pathology in an individual, comprising the establishment of at least one noninvasive diagnostic score, in particular of a diagnostic score for portal and septal fibrosis, and/or a noninvasive estimate score for the area of fibrosis, and/or a noninvasive estimate score for the fractal dimension, by carrying out the following steps:
a) for determining the area of fibrosis or the fractal dimension,
measuring, in a sample from said individual, at least one variable chosen from the group consisting of α-2 macroglobulin (A2M), hyaluronic acid (HA or hyaluronate), apolipoprotein A1 (ApoA1), type III procollagen N-terminal propeptide (P3P), gamma-glutamyltranspeptidase (GGT), bilirubin, gamma-globulins (GLB), platelets (PLT), prothrombin time (PT), aspartate aminotransferase (ASAT), alanine aminotransferase (ALAT), urea, sodium (NA), glycemia, triglycerides, albumin (ALB), alkaline phosphatases (ALP), YKL-40 (human cartilage glycoprotein 39), tissue inhibitor of matrix metalloproteinase 1 (TIMP-1), matrix metalloproteinase 2 (MMP-2), ferritin,
a′) for establishing a diagnostic score for portal and septal fibrosis, measuring, in a sample from said individual, at least three variables chosen from the group consisting of α-2 macroglobulin (A2M), hyaluronic acid (HA or hyaluronate), apolipoprotein A1 (ApoA1), type III procollagen N-terminal propeptide (P3P), gamma-glutamyltranspeptidase (GGT), bilirubin, gamma-globulins (GLB), platelets (PLT), prothrombin time (PT), aspartate aminotransferase (ASAT), alanine aminotransferase (ALAT), urea, sodium (NA), glycemia, triglycerides, albumin (ALB), alkaline phosphatases (ALP), YKL-40 (human cartilage glycoprotein 39), tissue inhibitor of matrix metalloproteinase 1 (TIMP-1), matrix metalloproteinase 2 (MMP-2), ferritin; at least one of the three variables being chosen from the group consisting of platelets (PLT) and prothrombin time (PT); in the case where exactly three variables are measured, these three variables cannot together be platelets (PLT), prothrombin time (PT) and bilirubin; preferably, the at least three variables chosen do not together comprise α-2 macroglobulin (A2M), hyaluronic acid (HA or hyaluronate) and tissue inhibitor of matrix metalloproteinase 1 (TIMP-1),
b) optionally, collecting at least one clinical variable characterizing said individual;
for the diagnostic score for portal and septal fibrosis, steps a′) and b) above being such that at least 4 variables are measured or collected,
c) combining said variables in a logistic or linear function, in order to obtain a diagnostic score for portal and septal fibrosis, and/or a diagnostic estimate score for the area of fibrosis, and/or a diagnostic estimate score for the fractal dimension;
d) diagnosing the presence and/or severity of said pathology and/or the effectiveness of said treatment based on the score obtained when performing the combining of step (c).
According to a first embodiment of the invention, in step a′, the at least three variables are chosen from the group consisting of α-2 macroglobulin (A2M), apolipoprotein A1 (ApoA1), type III procollagen N-terminal propeptide (P3P), gamma-glutamyltranspeptidase (GGT), bilirubin, gamma-globulins (GLB), platelets (PLT), prothrombin time (PT), aspartate aminotransferase (ASAT), alanine aminotransferase (ALAT), urea, sodium (NA), glycemia, triglycerides, albumin (ALB), alkaline phosphatases (ALP), YKL-40 (human cartilage glycoprotein 39), tissue inhibitor of matrix metalloproteinase 1 (TIMP-1), matrix metalloproteinase 2 (MMP-2), ferritin; at least one of the three variables being chosen from the group consisting of platelets (PLT) and prothrombin time (PT); in the case where exactly three variables are measured, these three variables cannot together be platelets (PLT), prothrombin time (PT) and bilirubin; preferably, the at least three variables chosen do not together comprise α-2 macroglobulin (A2M), hyaluronic acid (HA or hyaluronate) and tissue inhibitor of matrix metalloproteinase 1 (TIMP-1).
According to a second embodiment of the invention, in step a′, the at least three variables are chosen from the group consisting of α-2 macroglobulin (A2M), hyaluronic acid (HA or hyaluronate), apolipoprotein A1 (ApoA1), type III procollagen N-terminal propeptide (P3P), gamma-glutamyltranspeptidase (GGT), bilirubin, gamma-globulins (GLB), platelets (PLT), prothrombin time (PT), aspartate aminotransferase (ASAT), alanine aminotransferase (ALAT), urea, sodium (NA), glycemia, triglycerides, albumin (ALB), alkaline phosphatases (ALP), YKL-40 (human cartilage glycoprotein 39), matrix metalloproteinase 2 (MMP-2), ferritin; at least one of the three variables being chosen from the group consisting of platelets (PLT) and prothrombin time (PT); in the case where exactly three variables are measured, these three variables cannot together be platelets (PLT), prothrombin time (PT) and bilirubin; preferably, the at least three variables chosen do not together comprise α-2 macroglobulin (A2M), hyaluronic acid (HA or hyaluronate) and tissue inhibitor of matrix metalloproteinase 1 (TIMP-1).
The invention also relates to a diagnostic test for hepatic fibrosis, which implements the method of the invention. For the purpose of the present invention, the term “diagnostic” is intended to mean the establishment of the presence of a fibrosis and/or of its stage of evolution. To establish the diagnosis, the specificity of the test or of the method used is generally favored.
Advantageously, the clinical variables characterizing the individual are chosen from body weight (weight), body mass index (BMI), i.e. the weight/(size or height) 2 ratio, age (age) at the date on which the sample was collected, and cause. The term “cause” (or etiology) is intended to mean the alcoholic or viral cause. Consequently, it is clear to those skilled in the art that the “cause” clinical variable may only be used when a liver pathology such as a chronic hepatic fibrosis of viral origin or chronic hepatic fibrosis of alcoholic origin has already been diagnosed.
In the method of the invention, prior to step (c), the variables measured in step (a) or (a′) and the variables collected in step (b) can be combined with one another. Consequently, it is possible to use, in the logistic function implemented in the context of the invention, either “native variables”, also called “isolated or simple variables”, which are variables that have not undergone any modification before introduction into the logistic function, or “combinatorial variables”, which are arithmetic combinations of isolated variables with one another. By way of examples of combinatorial variables that can be used in the context of the present invention, and in a nonexhaustive manner, there are:
GAPRI=((GGT/45)/PLT)*100
GLOPRI=(GLB/PLT)*100
GLOTRI=(GLB/PT)*100
HYAPRI=(HA/PLT)*100
HYATRI=(HA/PT)*100
AMPRI=(A2M/PLT)*100
AMTRI=(A2M/PT)*100
HYAMTRI=(HA*A2M)/(PT*100)
HYAMPRI=(HA*A2M)/(A2M*100)=HA/100
HAMPRI=(HA*A2M)/(PLT*100)
HYAMPTRI=(HA*A2M)/(PLT*PT)
GHAMPRI=(GLB*HA*A2M)/(PLT*1000)
GHAMTRI=(GLB*HA*A2M)/(PT*1000)
GHAMPTRI=(GLB*HA*A2M)/(PLT*PT*10)
The acronym of these combinatorial variables uses the abbreviation of the isolated (simple) variables as a prefix and the suffix RI signifies “ratio index”.
It should be noted that a different score, but similar in its principle, called APRI (=ASAT/PLT) has been published (Wai et al, Hepatology, 2003, vol 38, pages 518-526). The ASAT/ALAT ratio, hereinafter called RAT, is also part of the prior art.
According to the present invention, the name noninvasive score for liver fibrosis (acronym: SNIFF) is given to a score composed of a combination of markers, preferably blood markers, ranging from 0 to 1, estimating the score of Métavir F type for liver diseases of viral origin (SNIFF) or alcoholic origin (SNIFFA) or the two causes (SNIFFAV) or of steatotic origin (SNIFFSA). The name noninvasive score for the area of liver fibrosis (acronym: SNIAFF) is used for a score composed of a combination of markers, preferably blood markers, ranging, in the majority of cases, from 5 to 55%. It is an estimate score for the area of liver fibrosis for liver diseases of viral origin (SNIAFF) or alcoholic origin (SNIAFFA) or the two causes (SNIAFFAV) or of steatotic origin (SNIAFFSA).
The severity of a liver pathology is the evaluation of the degree of fibrosis in the liver.
In step (a′) of the method of the invention, at least three variables, preferably 4, 5, 6 or 7 variables, are measured in a sample from said individual.
The measurements carried out in step (a) or (a′) of the method of the invention are measurements aimed either at quantifying the variable (the case for A2M, HA, bilirubin, PLT, PT, urea, NA, glycemia, triglycerides, ALB, P3P), or at quantifying the enzymatic activity of the variable (the case for GGT, ASAT, ALAT, ALP). Those skilled in the art are aware of various direct or indirect methods for quantifying a given substance or a protein or its enzymatic activity. These methods may use one or more monoclonal or polyclonal antibodies that recognize said protein in immunoassay techniques (radioimmunoassay or RIA, ELISA assays, Western blot, etc.), the analysis of the amounts of mRNA for said protein using techniques of the Northern blot, slot blot or PCR type, techniques such as an HPLC optionally combined with mass spectrometry, etc. The abovementioned protein activity assays use assays carried out on at least one substrate specific for each of these proteins. International patent application WO 03/073822 lists methods that can be used to quantify α-2 macroglobulin (A2M) and hyaluronic acid (HA or hyaluronate).
By way of examples, and in a nonexhaustive manner, a preferred list of commercial kits or assays that can be used for the measurements carried out in step (a) or (a′) of the method that is the subject of the present invention, on blood samples, is given hereinafter:
For the variables measured in step (a) or (a′) of the method that is the subject of the present invention, the values obtained are expressed in:
The clinical variables collected in step (b) of the method that is the subject of the present invention are expressed in:
The variables pinpointed with an asterisk (*) are expressed with one (*) or two (**) decimals, the others are expressed without decimals.
Advantageously, the sample from the individual used in step (a) or (a′) of the method that is the subject of the present invention is a biological medium such as blood, serum, plasma, urine or saliva from said individual or one or more cells from said individual, such as a tissue biopsy, and more particularly a liver biopsy. In the context of the present invention, it may be envisioned that the various variables measured in step (a) or (a′) are measured in different samples from the patient. By way of examples, and in a nonexhaustive manner, one variable is measured in the urine from the individual, whereas three others are measured in the blood from the same individual, the two samples (blood and urine) being taken within a relatively short period of time. However, and particularly preferably, the sample from the individual used in step (a) or (a′) of the method that is the subject of the present invention is a blood sample taken from the individual before any measurement.
According to a first embodiment of the present invention, the variables α-2 macroglobulin (A2M) and prothrombin time (PT) and at least two variables chosen from platelets (PLT), aspartate aminotransferase (ASAT), urea, hyaluronic acid (HA) and age are combined in step (c) of the method that is the subject of the present invention. Advantageously, the score obtained is a noninvasive score for liver fibrosis of viral origin, with at least four variables.
Among the preferred scores that may be obtained in this first embodiment, preference is given to the scores for which the following are combined in step (c):
The score that may thus be obtained is a noninvasive score for liver fibrosis of viral origin called SNIFF, which gives an estimate score of 0 to 1 for the score of Métavir F type, using from 4 to 7 variables.
In a second embodiment of the present invention, in addition to the prothrombin time (PT) variable, at least three variables chosen from aspartate aminotransferase (ASAT), alanine aminotransferase (ALAT) and alkaline phosphatases (ALP), age, hyaluronic acid (HA or hyaluronate) and α-2 macroglobulin (A2M) are combined in step (c). The score that may thus be obtained is a noninvasive score for liver fibrosis of alcoholic origin called SNIFFA.
Among the preferred scores that may be obtained in this second embodiment, preference is given to the scores for which the following are combined in step (c):
According to a third embodiment of the invention, at least the following 4 variables are combined in step (c) of the method: hyaluronic acid (HA or hyaluronate), gamma-glutamyltranspeptidase (GGT), bilirubin and platelets (PLT). The score thus obtained is a noninvasive estimate score (called SNIAFF with at least four variables) for the area of liver fibrosis ranging, in the majority of cases, from 5 to 35%. Preferably, in addition to the four variables described above, at least one, and preferably at least two variables, and even more preferably at least four variables, chosen from α-2 macroglobulin (A2M), urea, apolipoprotein A1 (ApoA1) and gamma-globulins (GLB), are combined in step (c).
Among the preferred scores that may be obtained in this third embodiment, preference is given to the scores for which the following are combined in step (c):
In a fourth embodiment of the present invention, a score called SNIDAFF, which is a noninvasive score for screening for liver fibrosis based on usual variables for alcoholic and viral liver pathologies, ranging from 0 to 1, can be obtained. For the purpose of the present invention, the term “screening for” should be understood to mean the search for the presence of a fibrosis regardless of its stage, either in patients with no known liver disease, or in patients with known chronic liver disease. For screening, the sensitivity of the test is a particularly important criterion.
The SNIDAFF score can advantageously be obtained by combining, in step (c) of the method of the present invention, at least the following four variables: platelets (PLT), prothrombin time (PT), aspartate aminotransferase (ASAT) and age. Preferably, in addition to the four variables described above, at least one, and preferably at least two variables, chosen from alkaline phosphatases (ALP), α-2 macroglobulin (A2M) and urea, are combined in step (c).
Thus, among the preferred scores that may be obtained in this fourth embodiment, preference is given to the scores for which the following are combined in step (c):
In a fifth embodiment of the present invention, a score called SNIFFSA, which is a noninvasive score for liver fibrosis for steatotic liver pathologies, ranging from 0 to 1, can be obtained. The SNIFFSA score can advantageously be obtained by combining, in step (c) of the method of the present invention, in addition to the prothrombin time (PT) variable, at least three variables chosen from aspartate aminotransferase (ASAT), triglycerides, age and glycemia.
Among the preferred scores that may be obtained in this fifth embodiment, preference is given to the scores for which the following are combined in step (c):
In a sixth embodiment of the present invention, the score called SNIFFAV, which is a noninvasive score for liver fibrosis for viral or alcoholic liver pathologies, ranging from 0 to 1, can be obtained. The SNIFFAV score can advantageously be obtained by combining, in step (c) of the method of the present invention, at least five of the following six variables: α-2 macroglobulin (A2M), platelets (PLT), prothrombin time (PT), urea, hyaluronic acid (HA or hyaluronate) or cause.
Among the preferred scores that may be obtained in this sixth embodiment, preference is given to the scores for which the following are combined in step (c):
In a seventh embodiment of the present invention, the score called SNIAFFAV, which is a noninvasive estimate score for the area of liver fibrosis for viral or alcoholic liver pathologies ranging, in the majority of cases, from 5 to 55%, can be obtained. The SNIAFFAV score can advantageously be obtained by combining, in step (c) of the method of the present invention, in addition to the prothrombin time (PT) variable, at least three, preferably at least four, or more preferably five, six or seven variables chosen from platelets (PLT), urea, hyaluronic acid (HA or hyaluronate), bilirubin, α-2 macroglobulin (A2M), gamma-glutamyltranspeptidase (GGT), gamma-globulins (GLB), aspartate aminotransferase (ASAT) and cause.
Thus, among the preferred scores that may be obtained in this seventh embodiment, preference is given to the scores for which the following are combined in step (c):
In an eighth embodiment of the present invention, the score called SNIDIFFAV, which is a noninvasive estimate score for the fractal dimension of liver fibrosis for viral or alcoholic liver pathologies ranging, in the majority of cases, from 0.7 to 1.3, can be obtained. The SNIDIFFAV score can advantageously be obtained by combining, in step (c) of the method of the present invention, at least four of the following five variables: α-2 macroglobulin (A2M), albumin (ALB), prothrombin time (PT), hyaluronic acid (HA or hyaluronate), alanine aminotransferase (ALAT), aspartate aminotransferase (ASAT) and age.
Among the preferred scores that may be obtained in this eighth embodiment, preference is given to the scores for which the following are combined in step (c):
In a ninth embodiment of the present invention, the score called SNIAFFA, which is a noninvasive estimate score for the area of liver fibrosis for alcoholic liver pathologies ranging, in the majority of cases, from 5 to 55%, can be obtained. The SNIAFFA score can advantageously be obtained by combining, in step (c) of the method of the present invention, in addition to the prothrombin time (PT) variable, at least three variables chosen from α-2 macroglobulin (A2M), hyaluronic acid (HA or hyaluronate), platelets (PLT) and weight of the individual.
Among the preferred scores that may be obtained in this ninth embodiment, preference is given to the scores for which the following are combined in step (c):
In a tenth embodiment of the present invention, the score called SNIAFFSA, which is a noninvasive estimate score for the area of liver fibrosis for steatotic liver pathologies ranging, in the majority of cases, from 5 to 35%, can be obtained. The SNIAFFSA score can advantageously be obtained by combining, in step (c) of the method of the present invention, in addition to the three variables prothrombin time (PT), gamma-globulins (GLB) and weight, at least one variable, preferably at least two variables, chosen from hyaluronic acid (HA or hyaluronate), platelets (PLT), age and BMI of the individual.
Among the preferred scores that may be obtained in this tenth embodiment, preference is given to the scores for which the following are combined in step (c):
As a variant, the present invention also relates to a method of diagnosing the presence and/or indicating the severity of a liver pathology and/or of monitoring the effectiveness of a curative treatment against a liver pathology in an individual, comprising the following steps:
a′) measuring, in a sample from said individual, at least one variable chosen from the group consisting of α-2 macroglobulin (A2M), hyaluronic acid (HA or hyaluronate), apolipoprotein A1 (ApoA1), gamma-glutamyltranspeptidase (GGT), bilirubin, gamma-globulins (GLB), platelets (PLT), prothrombin time (PT), aspartate aminotransferase (ASAT), alanine aminotransferase (ALAT), urea, sodium (NA), triglycerides, glycemia, albumin (ALB) and alkaline phosphatases (ALP), YKL-40 (human cartilage glycoprotein 39), tissue inhibitor of matrix metalloproteinase 1 (TIMP-1), matrix metalloproteinase 2 (MMP-2), ferritin;
b′) optionally, collecting at least one clinical variable characterizing said individual;
c′) combining, in a logistic or linear function, the variable(s) measured in (a′) and, optionally, the variables collected in (b′), in order to obtain a score;
d′) diagnosing the presence and/or severity of said pathology based on the score obtained when performing the combining of step (c′).
The characteristics of steps (a), (b), (c) and (d) described above (sample, assaying of variables, unit of variables) apply mutatis mutandis to steps (a′), (b′), (c′) and (d′).
According to a first embodiment of this variant of the invention, the following two variables: α-2 macroglobulin (A2M) and hyaluronic acid (HA or hyaluronate) can be measured in step (a′) so as to obtain the score called SNIAFFA 2, which is a noninvasive estimate score for the area of liver fibrosis for alcoholic liver pathologies ranging, in the majority of cases, from 5 to 55%.
It is possible to combine, with these two variables, at least one, and preferably at least two variables chosen from the weight of the individual and type III procollagen N-terminal propeptide (P3P).
Among the preferred scores that may be obtained in the first embodiment of this variant of the invention, preference is given to the scores for which the following are combined in step (c):
According to a second embodiment of this variant of the invention, the alanine aminotransferase (ALAT) variable can be measured in step (a′) of the method of the invention so as to obtain the score called SNIAH, which is a noninvasive estimate score for the necrotic-inflammatory activity of the liver for viral liver pathologies.
According to a third embodiment of this variant of the invention, the following three biological variables are measured in step (a′) of the invention: prothrombin time (PT), alanine aminotransferase (ALAT) and alkaline phosphatases (ALP). These three variables combined together in step (c′) of the present invention make it possible to obtain the SNIFFA 3 score (a non-invasive score for liver fibrosis of alcoholic origin, with three variables).
In addition, alternatively, noninvasive scores for liver fibrosis of alcoholic origin, called SNIFFA, using four variables can be used. Thus, the SNIFFA 4 score using four variables is determined by combining, in step (c′), the following variables: α-2 macroglobulin (A2M), age, hyaluronic acid (HA or hyaluronate) and alanine aminotransferase (ALAT), making it possible to obtain the SNIFFA 4a score.
The SNIFF, SNIFFA, SNIFFSA, SNIAH, SNIDAFF and SNIFFAV scores (or dependent variable) are predicted by a combination of biological or clinical markers (or independent variables). These combinations (or models) have been obtained by the statistical method called binary logistic regression with the following procedure:
Firstly, the independent variables were tested by univariable analysis.
Secondly, the independent variables that were significant in univariable analysis were tested in multivariable analysis by binary logistic regression with ascending or descending step by step selection.
The logistic regression produces the formula for each score in the form:
score=a0+a1x1+a2x2+
where the coefficients ai are constants and the variables xi are the independent variables.
This score corresponds to the logic of p where p is the probability of existence of a clinically significant fibrosis. This probability p is calculated with the following formula:
p=exp(a0+a1x1+a2x2+ . . . )/(1+exp(a0+a1x1+a2x2+ . . . )
or p=1/(1+exp(−a0−a1x1−a2x2− . . . ))
where the coefficients ai and the variables xi correspond to those of the formula for the score. The existence of a lesion (for example, clinically significant fibrosis) is determined by a probability p>0.5 (unless otherwise specified). It should be noted that the terms logistic regression “score” and SNIFF “score” do not correspond to the same term of the above equations. In clinical application, SNIFF corresponds to p.
We give below the tables for each SNIFF score with, in the first column, the name of each independent variable, in the second column, the value of the associated coefficient ai (called β in the text below and often in the literature and B in the tables below), and then its standard deviation (called S.D in the tables below) then its degree of significance (called signif in the tables below), and the last two columns give the exp(ai) confidence interval, i.e. the confidence interval (called CI in the tables below) of the corresponding odds-ratio (called exp(β) in the tables).
For each SNIFF score, as defined in the variants of the invention above, the overall predictive value of the model is reflected by the “overall percentage” of individuals correctly classified in a second table.
For each score, in the applicable equation, the coefficient βi of each independent variable xi can vary from the value β given in the table corresponding to said score ±3.3 standard deviations, a value also given in the tables. Similarly, a0 can vary from the value of the constant given in the table ±3.3 standard deviations.
By way of example and on the basis of the tables hereinafter, those skilled in the art wishing to use the SNIFF 4a score with 4 markers will employ the following formula:
p=1/(1+exp(−a0−a1(HA in μg/l)−a2(PT in %)−a3(A2M in mg/dl)−a4(AGE in years)) with
The SNIFF score is expressed in gross form (all the individuals are included) or in optimized form, and in this case, the extreme individuals, characterized by a studentized residue greater than 3, are discarded from the analysis. They are always low in number, as a rule ≦5%. For this reason, among the tables provided hereinafter, some indicated with a “o”, for instance SNIFF 4ao, provide β coefficients obtained after this optimization.
In addition, those skilled in the art wishing to use scores in the context of the present invention for which the various constants a0 and ai have not been provided in the present invention are capable of determining said constants. It is then necessary to have a database containing the independent variables used (as measured in step a and b) and a population of individuals having the pathology studied (alcohol and/or virus or steatosis), ideally several hundred individuals, and then to calculate the coefficients ai (or β) as indicated in step c and as explained above. The dependent variable is the lesion being sought, for example a clinically significant fibrosis defined by a Métavir score ≧2.
Firstly, the variables were tested in univariable analysis.
In a second step, the variables that were significant in univariable analysis were tested in multivariable analysis by linear regression with ascending step by step selection.
The linear statistical model is described by the following equation:
y
i
=a+β
1
x
1+β2x2+
where a is the constant, βi is the coefficient of each independent variable xi, and Yi is the dependent variable (area of fibrosis).
We give below the tables for each score with, in the first column, the name of each independent variable, in the second column, the value of the coefficient β and then its standard deviation, then the standardized coefficient β and, in the last two columns, the confidence interval at 95% for the coefficient β.
For each score, as defined in the variants of the invention above, the overall predictive value of the model is reflected, in a second table, by the coefficient R-two adjusted for each model, which is the percentage variability of yi explained by the independent variables of the model.
For each score, in the applicable equation, the coefficient βi of each independent variable xi can vary from the value β given in the table corresponding to said score ±3.3 standard deviations, a value also given in the table. Similarly, a0 can vary from the value of the constant given in the table ±3.3 standard deviations.
The caesura value is 0.500. As may be noted, the gain in effectiveness occurs not with respect to the diagnostic effectiveness (82.8 vs 83.4%), but with respect to other effectiveness indices, such as the area under the ROC curve (0.910 vs 0.890).
The SNIAFF, SNIAFFA, SNIAFFSA, SNIAFFAV and SNIDIFFAV scores were obtained by the statistical method called multiple linear regression with the following procedure:
Other advantages and characteristics of the invention will emerge from the examples that follow, given by way of illustration, and in which reference will be made to the attached drawings, in which:
A. Patients
The patient with chronic liver disease has a blood sample taken. The simple biological blood variables are determined according to good laboratory practice. The results are expressed with the units previously specified.
B. Assaying Methods
The hyaluronate concentration in a blood sample is measured by means of a radioimmunoassay technique (Kabi-Pharmacia RIA Diagnostics, Uppsala, Sweden).
The A2M concentration is determined by laser immunonephelometry using a Behring nephelometer analyzer. The reagent is a rabbit anti-human A2M antiserum.
The prothrombin time is measured from the Quick time (QT) which is determined by adding calcium thromboplastin (for example, Neoplastin CI plus, Diagnostica Stago, Asnières, France) to the plasma and the clotting time is measured in seconds. To obtain the prothrombin time (PT), a calibration line is plotted from various dilutions of a pool of normal plasmas estimated at 100%.
C. Calculation of the SNIFF Score
The results of the isolated (or simple) variables are used as they are or after conversion to combinatorial variables where appropriate. All these variables are included in the logistic regression formula. By way of example, and on the basis of the tables already described and of an example of formula use already described, those skilled in the art wishing to use the SNIFF 4a score with 4 markers will employ the following formula:
P=1/(1+exp(−a0−a1(HA in μg/l)−a2(PT in %)−a3(A2M in mg/dl)=a4(AGE in years))
i.e.
p=1/(1+exp(−2.365−(0.011×(HA in μg/l))−(−0.062×(PT in %))−(0.006×(A2M in mg/dl))−(0.030×(AGE in years)))
Two opposite examples are given:
Case 1 will be classified as having a clinically significant hepatic fibrosis and case 5 will be classified as not having any according to the caesura fixed at 0.50.
The ROC curve (
The index C obtained in the context of the tests of the invention has a value of 0.841±0.025 for the SNIFF 5 score and of 0.910±0.016 for the SNIFF 7bo score (
The box plots presented in
As could be guessed on the previous figures,
The SNIFF 7 score with 7 variables gives a lower estimation of fibrosis: r=0.769, p<10−4 than the SNIAFF 5o index with 5 variables: r=0.803, p<10−4.
This comparison shows that the SNIAFF estimate score for the area of fibrosis (
Similarly, the SNIFFA 4bo score with 4 variables gives a lower estimation of fibrosis: r=0.847, p<10−4 than the SNIAFFA 4o index with 4 variables: r=0.914, p<10−4.
This comparison also shows that the SNIAFFA estimate score for the area of fibrosis is a more reliable (accurate) indicator than the SNIFFA score for fibrosis (
The comparison of the SNIFF effectiveness and the Fibrotest effectiveness shows that the diagnostic effectiveness for Fibrotest 7 is 74.2% vs 82.1% for SNIFF 7. The AUROCs make it possible to show that the difference in effectiveness is statistically very significant (
| Number | Date | Country | Kind |
|---|---|---|---|
| 0405306 | May 2004 | FR | national |
| 0411536 | Oct 2004 | FR | national |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/FR2005/001217 | 5/13/2005 | WO | 00 | 11/20/2008 |
| Number | Date | Country | |
|---|---|---|---|
| 60622886 | Oct 2004 | US |
